Cytosolic free Ca²⁺ changes and calpain activation are required for β integrin–accelerated phagocytosis by human neutrophils

Neutrophils are “professional phagocytes” which have a remarkable capacity for phagocytosis. These cells can internalize microscopic particles (0.5–3 mm in diameter) of virtually any surface material. However, the efficiency and speed of phagocytosis is increased by coating the surface of the particles with opsonins such as antibodies or the complement component C3bi. The C3bi-accelerated phagocytosis of infecting microorganisms by neutrophils is probably the main route for the first line of defence by the innate immune system in vivo. However, the mechanism for this acceleration of phagocytosis by neutrophils is poorly understood.

However, it is clear that the accelerated phagocytosis is mediated by an interaction with the neutrophil receptor for the complement component C3bi, CD18/CD11b, CR3, or β2 integrin. There is an intricate interplay between this receptor and Ca²⁺ signalling. Although experimental strategies that immobilize β2 integrin on the neutrophil surface trigger changes in cytosolic free Ca²⁺ concentration (Jaconi et al., 1991; Ng-Sikorski et al., 1991; Petersen et al., 1993), there is also evidence that changes in cytosolic free Ca²⁺ concentration signal outwards to integrins, increasing the effectiveness of the receptor either by increasing its affinity (Li et al., 1998) or clustering ability (Hato et al., 1998; van Kooyk and Figdor, 2000).

One of the problems associated with previous attempts to study C3bi-accelerated phagocytosis in neutrophils has been the difficulty in observing the phagocytic event and of presenting the stimulus to the cell at a defined time and at a defined location on the neutrophil. In some previous studies, phagocytosis has been studied as part of the process of chemotaxis with the cell moving towards the particle before engaging it (Theler et al., 1995). In the study presented here, we avoid the possible crosstalk between chemotactic signalling and phagocytosis by presenting the particle to the neutrophil using micromanipulation. In this way, the events of binding, phagocytic cup formation, engulfment, and completion of phagocytosis were clearly defined and distinct from those involved with chemotaxis. The temporal relationship between changes in cytosolic free Ca²⁺ concentration and phagocytosis were correlated by simultaneous phase contrast and Ca²⁺ imaging. Here we show that small localized Ca²⁺ changes occur at the site of particle attachment and cup formation, but that the major effect is a global change in cytosolic free Ca²⁺ concentration. It is this latter Ca²⁺ signal which accelerates engulfment of the particle. We also show that a global rise in cytosolic free Ca²⁺ concentration liberates a fraction of β2 integrin receptors. As inhibitors of calpain activation prevent the rapid stage of phagocytosis despite the persistence of the global Ca²⁺ signal, we propose that Ca²⁺-activated calpain activation plays a key role in the acceleration of integrin-mediated phagocytosis by liberating integrin molecules remote from the initial contact site.

When permitted to send out random pseudopodia in a field of particles, neutrophils occasionally encounter particles in their near vicinity. When this occurs, four stages of phagocytosis can be distinguished: (a) pseudopod extension; (b) contact with the particle; (c) cup formation; and (d) phagosome closure and pseudopod retraction. With nonopsonised particles, the time to phagosome closure (completion of phagocytosis) was slow (2–4 min) and often did not reach this point. Changes in cytosolic free Ca²⁺ were sometimes recorded, but these did not consistently correlate with particular phases of phagocytosis, and Ca²⁺ signals were usually seen only after completion of phagocytosis. In contrast, when the particles were C3bi-opsonized, Ca²⁺ signals were consistently observed (n = 33) during this process as follows: (a) no changes in cytosolic free Ca²⁺ concentration occurred during pseudopod extension; (b) local Ca²⁺ signals were observed on contact with the particle and during cup formation; and (c) large Ca²⁺ changes throughout the neutrophil cytosol occurred immediately before rapid enclosure of the particle, phagosome closure, and pseudopod retraction (Fig. 1; Video 1, available online). In this series of experiments, the neutrophils were free to move and throw out spontaneous pseudopodia, so the possibility existed that the observed Ca²⁺ activity was contaminated by cell activities other than phagocytosis. Therefore, an approach was devised which permitted phagocytosis to be triggered on demand and in the absence of spontaneous pseudopodia formation or chemokinesis.

The method adopted used a micropipette through which slight negative pressure was applied to hold a C3bi-opsonized zymosan particle (2 mm in diameter), so that it could be presented to the cell (Fig. 2 a). With this method, the time and location of contact between the particle and the neutrophil was precisely controlled. In particular, chemotaxic cytoskeletal changes were not necessary before phagocytosis, as the neutrophil was not required to move towards the particle. Using this approach, contact between the opsonized particle and the neutrophil resulted in the formation of a phagocytic cup and complete phagocytosis. The same sequence of Ca²⁺ signalling was observed as with the nondirected phagocytosis, with localized Ca²⁺ signals occurring at the contact point during cup formation, and a large global change in cytosolic free Ca²⁺ concentration preceding rapid phagosome closure and retraction (Fig. 2 b). As both routes to phagocytosis produced similar Ca²⁺ signals, the delivery of the particles by micropipette was preferred as the individual events comprising phagocytosis could be controlled and distinguished. After presentation of the particle by the micropipette and the binding step, the formation of the phagocytic cup was slow and often appeared to temporarily arrest at this stage (Fig. 2 c). However, immediately after the global Ca²⁺ signal, there was consistently a rapid spreading of the pseudopodia around the particle (Fig. 2 c). This rapid stage resulted in an abrupt morphological transition toward roundness as the cell engulfed the particle and correlated strongly with the global Ca²⁺ signal (Fig. 2 d). The Ca²⁺ signals and phagocytosis could be provoked repetitively within an individual neutrophil. In some neutrophils, the global Ca²⁺ change had a distinctive double peak, which was also reproduced in response to subsequent phagocytic challenges (11/36 cells). Up to four particles have been “fed” to a single neutrophil, each phagocytic event being accompanied by a Ca²⁺ signal. However, successive Ca²⁺ events had a reduced magnitude (Fig. 3) and proceed more slowly (see Fig. 5). This decline in Ca²⁺ signal may have resulted from a depletion of available β2 integrin molecules (see Fig. 9) or of other signalling components coupled to the Ca²⁺ signal.

It was concluded that the Ca²⁺ events accompanying C3bi-accelerated phagocytosis were mediated by β2 integrin engagement because the events were triggered by zymosan opsonized with purified human C3bi. Also, pretreatment of the neutrophils with antibodies to either the α_m or β₂ chain of the integrin prevented both the rapid progression to completion of phagocytosis and the accompanying local and global Ca²⁺ signals triggered by either serum and C3bi opsonized particles. In contrast, antibodies to other surface proteins, including CD32 (FCRII), did not inhibit phagocytosis by either serum or purified C3bi-opsonized zymosan, and neutrophils treated with antibodies to β2 integrin were able to take up nonopsonized zymosan. As the leading edge of motile neutrophils and the pseudopodia may be enriched in β2 integrin molecules, contain polymerized actin and activated G proteins (Parent et al., 1998), and possibly have additional Ca²⁺ storage sites (Stendahl et al., 1994), it was possible that this morphological organization was crucial for phagocytosis and the accompanying Ca²⁺ signalling. To test this, particles were presented to neutrophils at locations remote from either the leading edge of motile neutrophils or the site of pseudopodia formation (Fig. 4, a and b; Video 2, available online). Presentation of opsonized particles to motile neutrophils at sites distant from the leading edge were still able to provoke the formation of phagocytic cups, signal local and global Ca²⁺ signals, and complete phagocytosis (Fig. 4 a). Similarly, presentation of particles to a site on neutrophils distant from an abortive phagocytic cup (produced by removal of a particle before completion) also resulted in complete phagocytosis and Ca²⁺ signalling in neutrophils (Fig. 4 b). There were no consistent sites on the neutrophil surface that failed to trigger Ca²⁺ and proceed to phagocytosis. Thus, it was concluded that β2 integrin, local Ca²⁺ stores, and all signalling molecules relevant for phagocytosis were available at all sites on the cell surface including sites remote from the leading edge of a motile neutrophil or phagocytic cup.

It has previously been shown that β₂-mediated Ca²⁺ signals comprise two components: Ca²⁺ store release and Ca²⁺ influx (Ng-Sikorski et al., 1991; Petersen et al., 1993; Pettit and Hallett, 1996). The Ca²⁺ signal produced by β2 engagement by phagocytic challenge was also composed of these two components, with the global Ca²⁺ change resulting from Ca²⁺ influx. As extracellular divalent cation ions are required for effective C3bi-β2 integrin binding (Li et al., 1998), experiments perfomed in the absence of extracellular Ca²⁺ may not be useful for establishing the role of Ca²⁺ influx. Instead, extracellular Ni²⁺ was used as a blocker of Ca²⁺ influx, and resulted in blockade of the β2 integrin–triggered global Ca²⁺signal (8/9 cells). This inhibition was demonstrated by both incubation with the inhibitor before challenge with a single particle (Fig. 5, a–c), and by challenge to an individual neutrophil with two particles, one in the absence and the other in presence of the inhibitor (Fig. 5, d–f). The Ca²⁺ signal was also inhibited by pretreatment with the PI-3-kinase inhibitor, LY294002 (9/12 cells). It was concluded that the initial local Ca²⁺ changes resulted from the release of Ca²⁺ from intracellular stores, whereas the global changes required Ca²⁺ influx through Ni²⁺-sensitive Ca²⁺ channels and that PI-3-kinase activity was important in its signalling (Fig. 5).

In order to establish the roles of the Ca²⁺ signals in phagocytosis, it was important to eliminate the possibility that Ca²⁺ signalling was an epiphenomenon caused by events associated with morphology changes. Indeed, it has been shown previously that deformation of the neutrophil surface by blunt micropipettes can induced localized or global Ca²⁺ changes (Laffafian and Hallett, 1995). In order to exclude this possibility, neutrophils were pretreated with cytochalasin B (5 μg/ml). This treatment totally inhibited neutrophil shape change (6/6 cells), but on contact with opsonised particles, localized and global Ca²⁺ signals were still triggered (3/6 cells) despite neither phagocytic cup formation nor any other morphological change. In the example shown (Fig. 6), a phagocytically competent neutrophil (with an internalized zymosan particle from the first challenge clearly visible) was then treated with cytochalasin B before the second challenge which provoked a Ca²⁺ signal in the absence of any detectable morphology change. These data were consistent with the Ca²⁺ signals being causal for the rapid cell shape change and phagocytosis (rather than being caused by the rapid shape change).

The global Ca²⁺ influx signal was required for maximum phagocytic rate. Under conditions in which total Ca²⁺ signalling and Ca²⁺ influx alone were prevented (LY294002 or Ni²⁺), binding and cup formation occurred, but completion of phagocytosis was slow (3/9 cells) or inhibited entirely (6/9 cells). As the rate of phagocytosis was also reduced by omission of extracellular Ca²⁺ (see Fig. 8 a), this led to the conclusion that the global Ca²⁺ signal, which resulted from Ca²⁺ influx, was the key step in triggering the rapid phase of completion of phagocytosis.

Calpain has previously been suggested to have a role in untethering integrin molecules from their cytoskeletal anchors (Stewart et al., 1998; Leitinger et al., 2000; Hogg and Leitinger, 2001). Therefore, the possibility existed that the global Ca²⁺ signal observed in neutrophils undergoing phagocytosis freed integrins from sites distant from the contact point to participate in the internalization process. In order to test this, the mobility of αmβ2 (fluorescent antibody-labelled CD18/CD11b) was determined by laser FRAP. In resting neutrophils, β2 integrin molecules were poorly mobile (Fig. 7 a, i) with only 15.5% (±5%, n = 4) of the fluorescence molecules free to diffuse. However, after activation of calpain with high cytosolic free Ca²⁺ concentrations (>5 μM average cytosolic concentration), or moderately elevated (0.8–1 μM), the fraction of mobile integrin molecules increased to 100 and 43% (±16%, n = 7), respectively (Fig. 7 a, ii and iii). Elevated cytosolic free Ca²⁺ significantly increased the fraction of mobile molecules (P < 0.001), but did not significantly affect the rate of recovery. The increase in the mobile fraction induced by raising cytosolic free Ca²⁺ concentration was prevented by the calpain inhibitor (PD 150606), with the mobility of β2 integrin molecules being reduced to below that of the resting neutrophils (Fig. 7 a, iv). As the rate of fluorescent recovery after elevation of cytosolic free Ca²⁺ was similar under these conditions, but the fractional recovery was altered, this suggested that once liberated, β2 integrin molecules moved at a similar rate as would be expected for simple diffusion. These data were consistent with elevated cytosolic free Ca²⁺ releasing tethered integrin molecules for diffusion by a calpain-dependent step.

In order to determine whether such β2 integrin untethering occurred during phagocytosis, part of the neutrophil surface was labelled with fluorescent antibody using a wide mouthed micropipette containing the antibody (Fig. 7 b, i). As demonstrated by FRAP, the labelled β2 integrin molecules were essentially immobilised and did not diffuse around the cell periphery (Fig. 7 b, ii). However, after challenge with a C3bi-opsonized particle, these labelled β₂ integrin molecules became again free to diffuse (Fig. 7 b, iii).

In order to probe whether the calpain-sensitive step was involved in the rapid completion of phagocytosis, pharmacological inhibitors of calpain were employed. An inhibitor of the Ca²⁺ binding site on calpain (PD 150606; Lin et al., 1997) inhibited the rapid (opsonin-dependent) uptake of particles and reduced the number of particles which were internalized (Fig. 8 a). The effect was specific for the Ca²⁺ influx route, as PD150606 did not inhibit phagocytosis in the absence of extracellular Ca²⁺. Inhibitors of the protease activity of calpain (calpeptin; Tsujinaka et al., 1988) and ALLN had similar inhibitory effects on phagocytosis, reducing both the rate and capacity for phagocytosis. Treatment of individual neutrophils with PD150606 or calpeptin also reduced the speed of phagocytosis in some neutrophils (4/9 with calpeptin) and prevented phagocytosis entirely in the remainder (3/3 cells with PD 150606; 5/9 cells with calpeptin). In all cases, the intact Ca²⁺ signalling sequence was evident on presentation of the opsonized particle (Fig. 8 b). After inhibition of calpain, binding of the particle and formation of the phagocytic cup could still occur, but progression to complete phagocytosis was prevented (Fig. 8 b, i). The consequent abortive phagocytic cups were deficient in antibody-accessible β2 integrin (Fig. 9, b and c), despite being uniformly distributed on the rest of the cell surface. In contrast, calpain-competent cells had an abundance of antibody-detectable β2 integrin at the phagocytic cup (Fig. 9 a). Together, these data pointed to a crucial role for Ca²⁺ activation of calpain activity as an important step in liberating β2 integrin for facilitation of phagocytosis.

In this paper, we have shown that phagocytosis of C3bi-opsonized particles by neutrophils is accelerated by a global increase in cytosolic free Ca²⁺ concentration which follows localized integrin engagement at the point of contact between the particle and the cell. The global Ca²⁺ change resulted from Ca²⁺ influx and was responsible for increased mobility of β2 integrin molecules distant from the contact site. The mechanism for this latter stage was dependent on calpain activation via its Ca²⁺ binding sites. As inhibition of calpain also inhibited C3bi-accelerated phagocytosis, we propose the following model to explain how C3bi opsonization facilitates phagocytosis. The initial contact between the C3bi-opsonized particle and the neutrophil causes crosslinking of local β2 integrin molecules which results in localized Ca²⁺ release from storage sites near the plasma membrane and formation of the phagocytic cup. The rate of further β2 integrin binding to the particle is limited by the supply of β2 integrin molecules which are tethered to cytoskeletal components. However, depletion of localized Ca²⁺ stores generates a diffusible signal which opens Ca²⁺ channels in the plasma membrane distant from the initial contact site and Ca²⁺ influx occurs across the entire neutrophil surface. This Ca²⁺ influx activates calpain which releases distant β2 integrin molecules from their tethers and permits their diffusion to the contact site to complete the phagocytic event.

It is well established that the β2 integrin is coupled to cytosolic free Ca²⁺ signalling when it is immobilized or crosslinked. Thus, antibody crosslinking with anti-β2 antibodies (Ng-Sikorski et al., 1991) or permitting neutrophils to sediment onto anti-β2 antibody coated surfaces triggers (Pettit and Hallett, 1996) release of Ca²⁺ from internal stores and influx of Ca²⁺ from the extracellular environment. Using confocal z sectioning, it was shown that the release of intracellular Ca²⁺ occurred only at the points of cell contact with the anti-β2 integrin (Pettit and Hallett, 1996), whereas Ca²⁺ influx occurred from all surfaces. It was also observed that the global Ca²⁺ signal was also followed by rapid neutrophil spreading onto the integrin-engaging surface. As Ca²⁺ signals often precede neutrophil spreading (Kruskal et al., 1986; Marks and Maxfield, 1990), the possibility that global Ca²⁺ signals caused neutrophil spreading was investigated by releasing caged Ca²⁺ in neutrophils which had newly sedimented onto a surface but had yet to spread. These experiments showed that there was a linkage between globally elevated Ca²⁺ and neutrophil spreading but only onto β2 integrin engaging surfaces (Pettit and Hallett, 1998). The data we present here now provides the explanation for those experiments, as neutrophil spreading onto β2-integrin engaging surfaces and C3bi-mediated phagocytosis are essentially equivalent cell events. The increased mobility of β2 integrin molecules may also underlie the changes in receptor affinity and avidity which occur on cell activation (Hato et al., 1998; Li et al., 1998; van Kooyk and Figdor, 2000). Thus, the positive feedback provided by the loop of integrin engagement signalling Ca²⁺ and Ca²⁺ permitting further integrin engagement may be the motor for phagocytic acceleration. The second phase of integrin binding may be responsible for the second Ca²⁺ peak observed in some neutrophils. It possible that the two Ca²⁺ events that are visible in these cells are present in all cells, but that they manifest as a single Ca²⁺ peak (Fig. 5 e), a Ca²⁺ peak with shoulder (Fig. 5 a), or double Ca²⁺ peak (Fig. 5 f) depending on the kinetics of cytosolic free Ca²⁺ reuptake. A recent mathematical model suggests that such differences may arise in individual cells as a result of differences in the amounts of their Ca²⁺ storage proteins (Baker et al., 2002).

However, some important questions remain unresolved. The first relates to the mechanism by which calpain liberates β2 integrin molecules in neutrophils. It is not clear which is the crucial calpain substrate in neutrophils and whether this is the linkage between β2 integrin and the cytoskeleton. However, talin, a known cytoskeletal linker, is a strong candidate as it is established that an elevation of cytosolic free Ca²⁺ concentration in neutrophils results in the cleavage of this molecule (Sampath et al., 1998). Another question relates to the specificity of the Ca²⁺ signal for calpain activation. It is clear that neutrophils are able to complete multiple tasks, such as phagocytosis, oxidase activation, and exocytosis and that changes in cytosolic free Ca²⁺ have been implicated in each of them (Scharff and Foder, 1993). Although it has not yet been possible to establish how specific outcomes can arise from Ca²⁺ signalling, for calpain activation, the explanation may be clearer. The two forms of calpain (μ and m) differ in their affinity for Ca²⁺, with μ-calpain having an apparent dissociation constant for Ca²⁺ of ∼25 μM (Michetti et al.,1997). As bulk cytosolic free Ca²⁺ never reaches this level under physiological conditions, calpain is essentially held in an inactive form. However, during Ca²⁺ influx, the concentration of free Ca²⁺ just beneath the plasma membrane rises to at least 30 μM (Davies and Hallett, 1999). Thus, μ-calpain would be activated specifically just beneath the plasma membrane at the strategic site required for talin cleavage. Other questions which arise include whether similar mechanisms accompany IgG-mediated phagocytosis by neutrophils. Although it has been suggested that the mechanisms of phagocytosis by the two routes may differ with respect to pseudopod protrusion (May and Machesky, 2001), with the micropipette method used here, presentation of either C3bi or IgG-opsonized zymosan particles resulted in similar pseudopod extension. A final question concerns the role of the initial Ca²⁺ release event. Although it may be a necessary prelude for Ca²⁺ influx, the possibility exists that it may also be important for local oxidase activation or phagosome–lysosome fusion. In a recent paper (Müller-Taubenberger et al., 2001), it has been shown in Dictyostelium lacking the Ca²⁺ storage proteins calreticulin and calnexin, that phagocytosis was impaired as a result of a defect in actin polymerization. These authors suggested that the Ca²⁺ storage of these cells was directly linked to actin regulation. If a similar situation exists in neutrophils, then the initial local Ca²⁺ signal may “seed” actin polymerization for phagocytic cup formation. Clearly, further work will be required to establish the answers to these and other questions. However, the data presented here lays the foundation for understanding the link between C3bi engagement, Ca²⁺ signals and phagocytosis.

Neutrophils were isolated from heparinized blood of healthy volunteers as described previously (Hallett et al., 1990). After dextran sedimentation, centrifugation through Ficoll-Paque (Amersham Biosciences) and hypotonic lysis of red cells, neutrophils were washed and resuspended in Krebs buffer (120 mM NaCl, 4.8 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 1.3 mM CaCl₂, 25 mM Hepes and 0.1% bovine serum albumin, adjusted to pH 7.4 with NaOH).

Neutrophils were loaded with fura-2 from the acetoxy-methyl ester for ratio imaging measurements as previously described (Hallett et al., 1990). Excitation wavelengths (340 and 380 nm) selected using a rapid monochromator (Delta RAM; PTI), which was connected to a Nikon Eclipse inverted microscope. The cells were maintained at 37°C using a temperature controlled microscope stage heater. The images at each excitation wavelength were collected using an intensified CCD camera (IC100; PTI) and the ratio image calculated using Image Master software (PTI). All images were captured using an oil immersion 100× objective. Phase contrast images were taken under far red illumination simultaneously with acquisition of Ca²⁺ images. This was achieved using an appropriate dichroic mirror and a second red-sensitive CCD camera. Fura-2 ratio images were collected from adherent neutrophils while a suspension of C3bi-opsonized zymosan particles was added. 340/380nm ratio fluorescence images were acquired with 16-frame averaging and threshold background subtraction at a rate of at least 0.6 frames/second. The ratiometric (Ca²⁺) images were pseudocolored according to the scale shown in Fig. 1, and the average ratio value of the pixels was calculated and plotted over the time course. The morphological changes (e.g., cell area, roundness, and perimeter contact between particle and cell) were determined using purpose-written software (A. Hoppe, University of Glamorgan, Wales, UK) and Optimas. The stimulus was added to the cells under view while recording continuously. All experiments were repeated on separate occasions with neutrophils isolated from different donors.

Zymosan particles (10 mg/ml) were opsonized either by incubation with human serum (50% diluted, 30 min, 37°C) or with purified human C3bi (1 mg/ml; 30 min, 4°C). The particles were then washed by centrifugation and resuspension to remove unfixed C3bi and used immediately or stored at −20°C. Human neutrophils were allowed to adhere to glass coverslips for 1 to 2 min before presentation of C3bi-opsonized zymosan particles (∼2 mm in diameter). Zymosan particles were allowed to sediment among the cells. A micropipette (tip diameter, 1–1.5 μm; WPI), with slight negative pressure applied, was used to pick up and hold a single zymosan particle. This was brought to the target neutrophil and contact between the neutrophil and the particle made. When adhesion between the opsonized particle and the neutrophil had occurred, and the particle was released from the micropipette and phagocytosis was allowed to proceed (Fig. 2 a).

The β chain (CD18) and α chain (CD11b) of the β2 integrin were labelled with fluorescently labelled antibody (phycoerthythrin or fluorescein) by incubation of the cells with excess antibody. A confocal plane was chosen which approximately bisected the cell to provide an image of the cell equator at low laser power. The voltage on the photomultiplier tube was set to maximum and line averaging was used. The portion of the cell to be photobleached was subjected to high power laser scanning for 10–30 s. Confocal images were then acquired as the fluorescence within the bleached area recovered. The ratio of the intensities of the bleached to nonbleached regions of the cell (Ir) were measured, at time intervals (3–30 s) over the following 2 min. The rate of recovery after photobleaching (Ir) and the fraction of mobile molecules (Mf) was calculated as Ir = Mf(1 − e^−kt), where k = 1/ln₂t1/2 and t = time after bleaching.

In experiments in which the location of β2 integrin was determined during phagocytosis, it was not possible to preincubate the cells with antibody to either CD18 or CD11b, as this inhibited accelerated phagocytosis. Therefore, experiments were performed in two ways. The first approach was to label part of the neutrophil surface remote from the phagocytic event. This was achieved by using a wide micropipette (∼5 μm; WPI) containing the fluorescent antibody. A neutrophil was drawn into the mouth of the micropipette with slight negative pressure and held for ∼20 s, before it was expelled by restoring zero intrapippette pressure. This ejected the partially β2 integrin–labelled neutrophil without expelling antibody into the surrounding medium. The second approach was to label β2 integrin after initiating phagocytosis in neutrophils by adding a pulse of excess fluorescently labelled antibody to the cell at the point of phagocytic cup formation. This permitted CD18/CD11b labelling and washing of nonadherent antibody to be achieved within ∼30 s of the onset of the Ca²⁺ signal.

Fura-2–AM was purchased from Molecular Probes. Purified human C3bi, ALLN, and PD150606 were purchased from Calbiochem, LY294002 from Sigma-Aldrich, and antibody to CD18 and CD11b from Dako. All other standard chemicals were purchased from Sigma-Aldrich.

Video clips of the full data from which Figs. 1 and 4 were derived are available online. These video clips show the cytosolic free Ca²⁺ concentration changes which occur on contact between the C3bi-coated particle and the neutrophil, either spontaneously (Video 1) or on presentation with a micropipette (Video 2). In the Video 2, the phase contrast image is also shown together with the Ca²⁺ measurement, so that the correlation between Ca²⁺ elevation and the accelerated phase of phagocytosis can be clearly seen.

We are grateful to Dr. Andreas Hoppe for writing and providing the neutrophil morphometry software used.

This work was supported by the Max Hanss Fund (BBSRC) and UWCM.