Phagosomal oxidative activity during β2 integrin (CR3)-mediated phagocytosis by neutrophils is triggered by a non-restricted Ca²⁺signal: Ca²⁺ controls time not space

A key event in combating infection is mediated by a programme of activity by circulating neutrophils, which includes neutrophil extravasation,chemotaxis and phagocytosis. It can be argued that the purpose of this series of separate cellular events is to bring the non-mitochondrial oxidase of the neutrophil in close proximity to infecting microbes. This results in the killing of the ingested microbe either directly by oxidase-generated reactive oxygen species, indirectly by oxidants generated subsequently by intraphagosomal myeloperoxidase or by regulation of the intraphagosomal pH or K⁺ ion concentration to activated proteases(Reeves et al., 2002). Oxidase activation results after the assembly of its components in the phagosomal membrane (Segal, 1996)signalled by binding of phospahtidylinositol 3-phosphate [PtdIns(3)P]through PX domains on p40phox and p47phox(Ellson et al., 2001a; Kanai et al., 2001). In vivo,particulate phagocytic stimuli are rapidly opsonised by the complement component C3bi, for which there is a receptor on neutrophils, CR3, theα _mβ₂ integrin (also called CD11b/CD18). Experimentally, engagement of CR3 has been shown to trigger changes in cytosolic free Ca²⁺(Ng-Sikorski et al., 1991; Jaconi et al., 1991; Petersen et al., 1993). Because, experimentally, an elevation in cytosolic free Ca²⁺ is well known to trigger activation of the oxidase (e.g. Lew et al., 1984; Al-Mohanna and Hallett, 1988),it has been suggested that the CR3-mediated Ca²⁺ signal would trigger oxidase activation during CR3-mediated phagocytosis. However, the temporal and spatial relationship between particle binding to the neutrophil by CR3, the Ca²⁺ elevation and subsequent oxidase activation has been unclear. In part, this is because of the difficulty in studying the time course of individual phagocytic events in individual neutrophils. However,here we have used a micromanipulation technique to present C3biopsonised particles to the neutrophil under observation(Dewitt and Hallett, 2002). The moment of contact, pseudopod formation and internalisation has been established with this technique and cytosolic free Ca²⁺ and oxidation of labelled particles determined simultaneously. Using this approach, we have found that the Ca²⁺ signal, which is triggered by CR3-mediated phagocytosis, can be resolved into two temporally separated components. The first Ca²⁺ signal occurs during integrin engagement but does not trigger oxidation of the attached particle. The second global Ca²⁺ signal, which is triggered about the time of phagosomal closure, causes an abrupt activation of the oxidase. This second Ca²⁺ signal was not restricted to the region of the phagosome, yet only triggered the oxidase activation locally in the phagosome, with no evidence of activation at other sites in the neutrophil. This points to a dual control of oxidase activation, with Ca²⁺ controlling the timing of oxidase activation, but slower and more localised molecular events, perhaps involving oxidase assembly and PtdIns(3)P generation, determining the site of oxidase activation.

Neutrophils were isolated from the heparinised blood of healthy or myeloperoxidase (MPO)-deficient volunteers as previously described(Hallett et al., 1990). A single MPO-deficient volunteer was recruited with a well-characterised MPO-deficiency (Clark et al.,2002). Informed consent was obtained from all volunteers. Following dextran separation, hypotonic lysis of red cells and centrifugation through Ficoll-Paque, neutrophils were resuspended in Krebs buffer (120 mM NaCl, 4.9 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).

Zymosan particles (10 mg/ml) were opsonised either by incubation with human serum (50% diluted, 30 minutes, 37°C) or with purified human C3bi (1 mg/ml; 30 minutes, 4°C). The particles were then washed by centrifugation and resuspension to remove unfixed C3bi and either labelled immediately or stored at -20°C. Opsonised zymosan labelling was achieved in bicarbonate buffer (0.1 M, pH 8.3, 10 mg/ml) (NaHCO₃ and HEPES buffer-titrate). Succimidyl ester of 2′7′dichlorodihydrofluorescein diacetate (1 mg, Molecular Probes, Eugene, Oregon) was dissolved in 100 μl DMSO(anhydrous), and added dropwise to zymosan (1 ml), mixing between each drop. The solution was then drawn into a foil-coated syringe, capped to exclude air and allowed to react overnight at 4°C. This resulted in coupling between the acetylated, oxidant-insensitive probe and the zymosan. In order to de-acetylate the product to generate the oxidant-sensitive dichlorodihydrofluorescein (DCDHF) for use, 100 μl hydroxylamine (105 mg/ml) was added to the suspension, which was then withdrawn into a syringe and capped for a further 3 hours. The DCDHF-labelled zymosan suspension was washed several times by centrifugation and resuspension with BSS (balanced salt solution), aliquotted and stored at -20°C. Fluorescein isothiocyanate(FITC) was conjugated to zymosan by over-night incubation, at 4°C at pH 9.0 as previously described (Morris et al., 2003).

Neutrophils were allowed to adhere to glass coverslips for 1-2 minutes. Zymosan particles were allowed to sediment among cells and a micropipette (tip diameter 1 μm) was brought into the field of view. A particle was drawn into the mouth of the micropipette by applying slight negative pressure, and was lightly placed in contact with a neutrophil. After adherence between cell and particle was established, the zymosan was released by removing the negative pressure and phagocytosis allowed to proceed(Dewitt and Hallett, 2002).

Neutrophils were loaded with the Ca²⁺ indicator, fura-2-AM as previously described (Hallett et al.,1996) or fura2-dextran by micro-injection and allowed to adhere to glass coverslips maintained at 37°C with a temperature-controlled microscope stage heater. Light was transmitted to an inverted microscope(Nikon Eclipse) with an oil immersion 100× objective using a rapid monochromator (Delta RAM, PTI, Surbiton, UK) at three excitation wavelengths -340 nm, 380 nm (for fura2) and 490 nm (for DCDHF) - delivered sequentially(see below). Phase contrast images were simultaneously taken under far red illumination (690 nm) using an appropriate dichroic mirror and a red-sensitive CCD camera. The fluorescent images were collected using a CCD camera (IC100,Photon Technology International, Surbiton, UK), and 340/380 nm ratio images and 490 nm intensity images were calculated and captured using ImageMaster(PTI, UK) software. The sequence of the excitation wavelengths was 340 nm, 490 nm and 380 nm, thus the two data sets for a single Ca²⁺ ratio image and the DCDHF images were overlapping temporally. Images were acquired through the same emission filter with maximum transmission at 505 nm, with 16-frame averaging, threshold background subtracted and ratio or intensity calculated and pseudocoloured. The sum intensity (SI) (rather than average intensity) of the zymosan particles was used to monitor oxidase activity to avoid errors associated with the periphery of the zymosan particle where the nonoxidised indicator often had subthreshold fluorescence. The SI was therefore calculated as SI=Nø.Iav, where Nø is the number of non-zero pixels and Iav is the average intensity of the pixels.

The large molecular weight conjugate of fura2, fura2-dextran (molecular mass 10 kDa), was used to measure cytosolic free Ca²⁺ concentration with reduced diffusion of the fura2-Ca²⁺ complex within the cell. The fura2-dextran (Molecular Probes) was micro-injected into neutrophils by the simple lipid assisted micro-injection technique (SLAM) as previously described (Laffafian and Hallett,1998). The probe was dissolved in intracellular medium (KCl, 150 mM, HEPES, 25 mM, pH 7.0) to give a final concentration of 500 μM and either loaded into a micropipette (tip diameter 0.5 μm), before tip lipid coating, as described previously(Laffafian and Hallett, 1998),or loaded into a pre-lipid-coated micropipette (Cell Engineering, Swansea,UK). On contact between the micropipette and the neutrophil, transfer of fura2-dextran into the cell was monitored by an increase in fluorescence at 360 nm to give intracellular concentrations of fura2-dextran of between 10-50μM. After successful micro-injection, the neutrophils remained fully functional and able to undergo phagocytosis in response to challenge

Following presentation of C3bi-opsonised zymosan particles (diameter 2μm) to adherent neutrophils using a micropipette, binding to the cell and phagocytic cup formation begins before a large change in cytosolic free Ca²⁺ is observed. This sequence has been previously reported(Suzuki et al., 1997; Dewitt and Hallett, 2002) and shown previously to result from Ca²⁺ influx through Ni²⁺-sensitive Ca²⁺ channels and to precede and accelerate phagosome closure (Dewitt and Hallett, 2002). The global Ca²⁺ change often displayed a distinctive double Ca²⁺ peak, or 'shouldered peak'(Fig. 1a). Analysis of several Ca²⁺ signals suggested that there were two separate Ca²⁺signals within each response and that the shape of the peak observed depended on the temporal separation between the two components of the Ca²⁺signal or the rate of the down slope (Ca²⁺ reuptake). The qualitative shape of the Ca²⁺ signal was constant within an individual neutrophil, with individual neutrophils displaying the same double peak, 'shouldered' or single Ca²⁺ in response to subsequent phagocytic stimuli (Fig. 1b). With nonopsonised zymosan and with neutrophils on which CR3 was blocked by antibody to CD11b/CD18 challenged with C3bi-opsonised zymosan, the characteristic Ca²⁺ signal was absent, although sometimes a rise in Ca²⁺ occurred after completion of phagocytosis(Dewitt and Hallett, 2002).

DCDHF-labelled zymosan particles triggered the same typical sequence of phagocytic events with particle binding, cup formation and local Ca²⁺ signalling, giving rise to a global Ca²⁺ rise due to influx (and inhibitable by Ni²⁺) and then phagosome closure(Fig. 2). The global Ca²⁺ change was also observed as a single transient increase, with a shoulder or double peak. The DCDHF-labelled zymosan particles were therefore a useful model for studying phagocytosis by neutrophils.

The simple addition of hydrogen peroxide (H₂O₂) to a suspension of DCDHF-labelled particles produced an increase in particle-associated fluorescence excited at 490 nm(Fig. 3a). The rate of H₂O₂-induced increase in intensity was greatly enhanced by peroxidase activity (Fig. 3c). After phagocytosis, C3bi-opsonised DCDHF-labelled zymosan particles that had been engulfed by adherent neutrophils were significantly brighter than those which had not (Fig. 3b). This was not a general effect of the phagosomal environment on the 'parent molecule', fluorescein, as fluorescein-labelled particles,which were insensitive to oxidation, did not increase in intensity during phagocytosis (4/4 cells) and often decreased in intensity (3/4 cells; Fig. 3d). Thus, the increase in fluorescence intensity of DCDHF-labelled zymosan during phagocytosis was used as a monitor of oxidative activity, particularly in the presence of peroxidase activity.

In order to establish the time sequence of oxidase activation and Ca²⁺ signalling during phagocytosis, phase contrast images, the DCDHF signal from the particle and Ca²⁺ signals from the cell were recorded simultaneously. Using this simultaneous imaging approach, it was immediately evident that the Ca²⁺ signal in the neutrophil began many seconds before the onset of oxidase activation detectable by increased DCDHF fluorescence (42/47 cells). To determine whether the onset correlated with any of the phases of the complex Ca²⁺ signal, neutrophils that displayed distinct double Ca²⁺ peaks were chosen for analysis because the separate Ca²⁺ phases could be more easily distinguished and timed. In the majority of the cells studied (18/25 cells), the onset of DCDHF oxidation (increased fluorescence) was shown to correlate with the second of the two Ca²⁺ peaks(Fig. 4a). It was difficult to time the onset of oxidation accurately in the other cells tested (7/25),either because the particle was moving in and out of focus during phagocytosis or because there was little oxidative increase after phagocytosis. The relationship between oxidation and the later Ca²⁺ signal was not unique to these 'twin peak' cells, as it was also evident in cells with less obvious separation of the two signals, such as those giving a Ca²⁺shouldered response (Fig. 4b).

The second Ca²⁺ event usually occurred around the same time as phagosomal closure. In the majority of cells, phagosomal closure occurred just after the second peak in the Ca²⁺ signal (21/33); in some cells(8/33) phagosomal closure occurred while the second Ca²⁺ signal was subsiding, but in one cell, phagosomal closure occurred just before the second peak, and in the remainder (3/33), phagosome closure occurred after the return of the second Ca²⁺ peak to base-line. As oxidative activation also occurred near the time of phagosome closure and the second Ca²⁺peak (see Fig. 4a,b), we considered the possibility that the oxidase was actually activated and peroxidase secreted before phagosome closure, but that the local oxidant and peroxidase concentrations were too low (as a result of diffusional dilution before phagosome closure) to be effective in causing oxidation of DCDHF on the zymosan until the phagosome closed, preventing dilution. However, this explanation seemed unlikely to account for the abrupt increase in oxidation within the phagosome because first, this mechanism would be expected to cause a gradual increase in DCDHF fluorescence, in contrast to the abrupt changes observed (Fig. 4a,b, Figs 5 and 6); second, there was no evidence that the base of the zymosan particle deepest in the phagocytic cup,where the diffusion path to the outside was longest, became oxidised earlier or before phagosome closure (e.g. Fig. 4a,b, Fig. 6); and third, in some cells (e.g. Fig. 4a,b), phagosome closure occurred several seconds before the onset of detectable oxidation, there being no sudden effect of oxidation at the time of phagosomal closure. It was therefore concluded that during phagocytosis,activation of the oxidase and triggering of the release of peroxidase into the phagosome did not occur during the first global change in Ca²⁺signal but correlated temporally with the second phase of the Ca²⁺signal.

To establish the role for the phagocytic-induced Ca²⁺ signal in oxidase activation, two pharmacological inhibitors of the Ca²⁺signal were used. Ni²⁺ (2 mM), which blocked Ca²⁺influx, totally prevented the global Ca²⁺ signal and slowed phagocytosis (Dewitt and Hallett,2002). Blocking Ca²⁺ influx in this way also reduced the increase in DCDHF-zymosan fluorescence, the oxidative response being significantly reduced by Ni²⁺ (P<0.05, n=6, Fig. 5). The phosphatidylinositol 3-kinase [PtdIns(3) kinase] inhibitor, LY294002 (50μM), also inhibited the Ca²⁺ signal and also slowed phagocytosis(Dewitt and Hallett, 2002). Again, there was a significant decrease in the oxidative response(Fig. 5). It was therefore concluded that during phagocytosis, the global Ca²⁺ signal was obligatory for maximum phagosomal oxidation (during the second Ca²⁺phase), but it was not sufficient alone to trigger oxidative activation (in the first Ca²⁺ phase).

As myeloperoxidase (MPO), which is responsible for the generation of HOCl from H₂O₂ in neutrophil phagosomes, is delivered to the phagosome by granule-phagosome fusion, and has a major effect on DCDHF-zymosan oxidation (Fig. 3), it was important to consider the role of this enzyme in the oxidation signal observed. DCDHF-labelled zymosan was therefore delivered to neutrophils in the presence of the potent MPO inhibitor, azide (10 mM N^3-)_.Under these conditions, Ca²⁺ signalling was normal, but the fluorescent signal from internalised particles was significantly reduced(P<0.01), although it could not be totally inhibited(n=6) (Fig. 5). This may have resulted from an inability of azide to cause total peroxidase inhibition, as the residual response could be inhibited further by the Ca²⁺ channel blocker, Ni²⁺ (P<0.001, n=2, Fig. 5). This reduced the response to that observed in MPO-deficient neutrophils(P<0.001, Fig. 5). In these latter cells, the oxidative response was restored to normality by addition of extracellular MPO (5.5 nM, by haem content) or horseradish peroxidase (0.25 units/ml) before phagocytosis. It was therefore concluded that delivery of MPO to the phagosome played an important role in the DCDHF-oxidation signal.

In the absence of MPO-activity (MPO-deficiency), although the DCDHF signal was reduced and slowed as with N₃^- treatment, a detectable response was observed which commenced at the second phase of the Ca²⁺ signal. As this suggested that the second phase of the Ca²⁺ signal was also important in activating the oxidase (in the absence of peroxidase degranulation), this proposal was tested by measuring oxidation of DCDHF-zymosan during phagocytosis in the presence of extracellular MPO or HRP (0.25 units/ml). This strategy, which would eliminate the dependence of degranulation of peroxidase into the phagosome, had no effect on the timing of the DCDHF response. It was therefore concluded that the oxidase (or delivery of preformed oxidant) was not activated earlier than the second phase of the Ca²⁺ signal and that oxidase activation occurred with the second Ca²⁺ signal.

As the second Ca²⁺ signal triggered oxidative activity but was not restricted within the cytosol, the possibility that oxidase and peroxidase release were also triggered at nonphagosomal sites was investigated. Neutrophils were chosen that had oxidant-detecting zymosan particles close by or touching, which would act as a sentinel reporting extracellular oxidation during phagocytosis. In these experiments, no oxidants were detected at extracellular sites around the cell (47/47) even if the sentinel zymosan particles were near the open mouth of the phagosome (e.g. Fig. 6a, particle 3). Sentinel particles near or touching the outer face of the phagocytic cup(Fig. 6a, particle 2) also failed to become oxidised. Thus, it appeared that the second phase Ca²⁺ rise triggered oxidase assembly and granule fusion restricted to the inner face of the forming phagosome. However, the concentration of oxidants near extracellular zymosan particles (i.e. not within phagosomes) was expected to be less as a result of dilution and so detection of oxidation of DCDHF might have been preferential within the phagosome. This was confirmed in neutrophils in which pseudopodia formation and internalisation of particles was prevented by pretreatment with cytochalasin B (5 μg/ml). Although this treatment did not inhibit the Ca²⁺ signal(Dewitt and Hallett, 2002), no oxidation of the extracellular particle was detected(Fig. 6b). Cytochalasin B does not inhibit but enhances oxidase-mediated oxygen consumption(Al-Mohanna and Hallett, 1987)and causes nondirected MPO release, which was consistent with extracellular dilution of oxidants preventing the detection of extracellular oxidation by this approach. This conclusion was also supported by the failure of phorbol myristate acetate (PMA), a non-phagocytic stimulus of the oxidase, but not of MPO release, to oxidise extracellular DCDHF-zymosan. A second indicator of oxidase activity, independent of MPO, was therefore used. Nitroblue teterazolium (NBT) is reduced by O₂^- (and may also accept electrons directly from the oxidase) to form insoluble formazan. As the precipitate cannot diffuse, it provides spatial information of oxidase activity. This indicator reported strong oxidation within the phagosome,beginning on closure of the phagosome (n=3), but produced no evidence of oxidase activity at adjacent plasma membrane sites(Fig. 6c). There was evidence of oxidation of NBT within the cytosol near the phagosome, perhaps as a result of diffusion of oxidants from localised oxidase activity within the phagosome(Fig. 6c) (see also Sullivan, 2003). There was,however, no clear formazan precipitation at membranes other than within the phagosome (Fig. 6c).

Whether or not activation of the oxidase was restricted absolutely to the inner phagosomal membrane, formazan precipitation clearly showed it was localised at least to that part of the cell undergoing phagocytosis. As the activation of the oxidase appeared to be spatially restricted, yet Ca²⁺ signalling was unrestricted throughout the cytosol, we examined the possibility that the apparent global distribution of elevated Ca²⁺ was generated artefactually as a result of the rapid diffusion of the fura2-Ca²⁺ complex away from a more restricted region of the cytosol. This possibility was considered because the initial small localised Ca²⁺ event, which occurs on initial contact between an opsonised particle and a neutrophil, is best observed when the geometry of the phagocytic cup includes a narrowing before the main body of the cell(Dewitt and Hallett, 2002). This would be explained if the narrowing at the base of some phagocytic cups imposed a 'diffusion resistance' to the fura2-Ca²⁺ complex and so more accurately reported the site of elevated Ca²⁺. The possibility that the oxidase-triggering second Ca²⁺ signal was also restricted to a zone around the phagosome was therefore tested by using a larger molecular weight dextran conjugate of fura2 (molecular mass 10 kDa) as a more slowly diffusible Ca²⁺ reporter. After micro-injection with fura2-dextran (50-100 μM), the phagocytic event was identical to that seen in noninjected neutrophils or those micro-injected with an irrelevant molecule(lucifer yellow). Furthermore, apart from the initial Ca²⁺ event,which was observed as localised in 1/3 phagocytotic events, fura2-dextran also reported synchronised global Ca²⁺ signals (3/3), with no evidence of a time delay between the periphagosomal region and elsewhere(Fig. 7). Also, fast confocal imaging (33 mseconds resolution) failed to reveal a wave of Ca²⁺originating from the phagosome (data not shown). There was thus no evidence that the global Ca²⁺ changes originated from a localised source of Ca²⁺ in the phagocytic region or that the cytosolic free Ca²⁺ concentration was higher there (in Ca²⁺ ratio images, the DCDHF signal contaminates the Ca²⁺ image as it gets brighter after phagocytosis and give an artefactually higher ratio value at that location). It was therefore concluded that the Ca²⁺ signal that triggered the oxidation response was the result of influx of Ca²⁺ across the plasma membrane around the whole cell perimeter. However, as this elevated Ca²⁺ alone was insufficient to trigger oxidase activity in non-phagosomal locations in the cell, it was concluded that other key factors required for activation of the oxidase and peroxidase release by Ca²⁺ were limited to the phagosomal membrane.

The work presented here has shown that the Ca²⁺ signals that accompany phagocytosis are composed of at least two components but that only the second phase is associated with oxidative activity in the phagosome, the earlier Ca²⁺ changes failing to trigger oxidase activation. It is therefore proposed that elevated cytosolic free Ca²⁺ concentration alone is insufficient to activate the oxidase and that other obligatory but slower events, which are complete by the time of the second phase of the Ca²⁺ signal, are required. It is probable that the slower events include the assembly of oxidase components(Segal, 1996) and PI(3)P generation (Ellson et al.,2001a; Kanai et al.,2001) in the phagosomal membrane. The data here also shows that while the Ca²⁺ signal controls the timing of oxidase activation, it plays no role in restricting the activity to the phagosome.

This study raises some important questions. The first set of questions concern the mechanism underlying the complexity of the phagocytic Ca²⁺ signal. We have recently shown that the initial β2 integrin-mediated Ca²⁺ signal causes an increase in mobility ofβ2 integrin molecules distant from the contact site, which results in an acceleration of phagocytosis (Dewitt and Hallett, 2002). The second phase of β2 integrin binding may be responsible for the second Ca²⁺ peak observed in some neutrophils. It is probable that the same events exist in all neutrophils, but appeared as a single Ca²⁺ peak, a Ca²⁺ peak with shoulder or a double Ca²⁺ peak, depending on the some critical values for key kinetic parameters of integrin mobilisation or Ca²⁺homeostasis. A recent mathematical model suggests that splitting of Ca²⁺ signals may arise in individual cells as a result of differences in the number of their Ca²⁺ storage proteins(Baker et al., 2002) or in Ca²⁺ reuptake rates. The question also arises about the mechanism by which the global Ca²⁺ signal is generated. Clearly, with the technique used here, the stimulus delivered to the cell was very localised,yet no evidence for localisation of the 'global' Ca²⁺ signal was found (Fig. 4a,b, Fig. 7). As the global Ca²⁺ signal was driven by Ca²⁺ influx, three possibilities exist: first, a factor was released into the medium that diffused extracellularly to trigger remote Ca²⁺ influx; second, a factor was released into the cytosol that diffused intracellularly to trigger remote Ca²⁺ influx; or third, a factor was released into the plasma membrane that diffused around the cell membrane to trigger remote Ca²⁺ influx. There is some evidence in support of either of the latter two possibilities, as a water-soluble factor (Ca²⁺ influx factor) has been described (Randriamampita and Tsein, 1993) and can be isolated from neutrophils(Davies and Hallett, 1995), and inhibitors of PI(3) kinase inhibit Ca²⁺ signals triggered byβ2 integrin (Dewitt and Hallett,2002) and phosphatidylinositol-3,4,5-trisphosphate[PtdIns(3,4,5)P₃] itself can induce Ca²⁺signals in neutrophils. For any of the three mechanisms, the Ca²⁺concentration may be elevated nearer the generation site of the'Ca²⁺ influx signal' as it diffuses down a concentration gradient. In an earlier study, a wave of Ca²⁺ (17 μm/s) originating near the phagosome (Schwab et al.,1992) was indeed reported. However, we and others(Theler et al., 1995) have not been able to show such a Ca²⁺ wave.

The Ca²⁺ influx signal was responsible for abrupt oxidation of internalised zymosan (Fig. 5). This may be attributed to the abrupt delivery to the phagosome of MPO by granulephagosome fusion, the abrupt activation of the oxidase in the phagosome membrane, the delivery of long-lived oxidants preformed in a granule (that perhaps also contain the MPO) or a combination of these. It has previously been shown that high Ca²⁺ levels are required for MPO-granule-phagosome fusion, with K₅₀ values estimated by quin2 buffering as 2.6 μM (Lew et al.,1986; Jaconi et al.,1990) or by internal perfusion at near 100 μM(Nüsse et al., 1998). As these levels of Ca²⁺ are higher than the bulk Ca²⁺ level(approximately 700 nM), either β2 integrin engagement increased the efficacy of Ca²⁺ for degranulation or Ca²⁺ just under the phagosomal membrane reached much higher levels. Although two early papers reported a long-lived localised Ca²⁺ signal around the phagosome(Sawyer et al., 1985; Murata et al., 1987), this could not be shown here. However, Ca²⁺ influx in neutrophils has been shown to cause high submembrane changes in Ca²⁺, to give a Ca²⁺ concentration of at least 50 μM(Davies and Hallett, 1998),which has been suggested to mediate subplasma membrane calpain activation(K_d 30 μM) (Dewitt and Hallett, 2002). As the concentration of Ca²⁺ within the phagosome decreases after phagocytosis(Lundqvist-Gustafsson et al.,2000), the possibility exists that Ca²⁺ channel opening in the phagosomal membrane may elevate Ca²⁺ concentrations to very high levels immediately around the phagosome. However, it is clear that this Ca²⁺ signal alone is insufficient to activate the oxidase as its activity was not triggered by the initial Ca²⁺ change. The possibility exists that between the first and second Ca²⁺ event,the oxidase was assembled in a form awaiting the Ca²⁺ signal for full activation. It is known that components of the oxidase p40phox and p47phox must be assembled (Segal,1996) and that PtdIns(3)P binding and generation occurs(Ellson et al., 2001a; Kanai et al., 2001). If PtdIns(3)P is localised to the phagosomal membrane in neutrophils, as it is in RAW 264.7 cells (Ellson et al.,2001b), the effect of the global Ca²⁺ signal would only be evident locally within the phagosome. Although the oxidase complex may be partially active without the Ca²⁺ signal, and thus account for the inability to completely inhibit oxidative activity by blocking the Ca²⁺ signal (Fig. 5), the Ca²⁺ signal was associated with a massive increase in oxidative activity in the phagosome. As it is well documented that activators and inhibitors of protein kinase C (PKC) activity also activate and inhibit oxidase activation (e.g. Cooke and Hallett, 1985), and PKC associates with the oxidase component p40phox (Reeves et al., 1999)the possibility exists that the action of Ca²⁺ on the oxidase may be mediated by PKC activity.

An issue important for pathology raised in this study is whether leakage of oxidants into the extracellular medium can occur from the activated oxidase. No leakage was detected by zymosan particles near by or even at the mouth of the phagosome. However, this may have arisen artefactually as a result of dilution of oxidants in the extracellular medium (an event which cannot occur within the enclosed phagosome). In some cells, efflux of oxidants from the forming phagosome may thus remain a possibility, because although there was a good correlation between the second phase on the Ca²⁺ signal and the activation of oxidative activity, this was not tightly coupled to the time of phagosomal closure. This means that in some cells, the oxidase was activated before complete closure of the phagosome and that, theoretically,extracellular oxidation would be possible.

The work here has therefore provided a role for the second of the two Ca²⁺ signal phases triggered during β2 integrin-mediated phagocytosis - namely, in activation of the oxidase and triggering granule-phagosome fusion. Previously, we have shown that the first phase was required for the liberation of immobile β2 integrin to cause acceleration of phagocytosis. However, it still remains to be established whether the first localised Ca²⁺ signal fulfils any role and how the Ca²⁺signal interacts with other events before a full explanation of the roles of the complex Ca²⁺ signalling of phagocytosis by neutrophils can be given.

We thank Carolyn Cole (initial experiments with DCDHF-zymosan coupling),Stephen Clark and Valerie O-Donnell (access to the MPO-deficient volunteer)and to all the blood donors, especially the MPO-deficient volunteer. We are also grateful to UWCM, BBSRC (Maxime Hanss Fund) and Wellcome Trust for support.