Phagosomal Acidification Is Mediated by a Vacuolar-type H+-ATPase in Murine Macrophages”

The mechanism underlying phagosomal acidification was studied in thioglycolate-elicited murine macrophages. The pH of the phagosomal compartment (pHp) was measured fluorimetrically in macrophage suspensions following ingestion of fluorescein isothiocyanate-labeled Staphylococcus aureus. At 37 degrees C, pHp decreased rapidly, reaching a steady state value of 5.8-6.1, while the cytoplasmic pH remained near neutrality, pH 7.1. The phagosome to cytosol pH gradient could be collapsed by addition of nigericin, monensin, or weak bases. The substrate dependence and inhibitor sensitivity profile of phagosomal acidification were investigated in intact and permeabilized cells. Phagosomal acidification was inhibited when ATP was depleted using metabolic inhibitors or permeabilizing the plasma membrane by electroporation. In permeabilized cells, acidification could be initiated by readdition of both Mg2+ and ATP. Neither adenosine 5'-(beta,gamma-imido)triphosphate nor adenosine 5'-(gamma-thio)triphosphate supported phagosomal acidification. Inhibitors of F1F0-type H(+)-ATPase such as oligomycin and azide, and the E1E2-type H(+)-ATPase inhibitor vanadate had no effect on phagosomal acidification. In contrast, the rate of phagosomal acidification was reduced by micromolar concentrations of N-ethylmaleimide and N,N'-dicyclohexylcarbodiimide. In permeabilized cells, nitrate inhibited the acidification with an apparent Ki of 25 mM. Phagosomal acidification was also effectively blocked by the macrolide antibiotic bafilomycin A1, with an apparent Ki of approximately 3 mM in both intact and electroporated cells. In this concentration range, bafilomycin A1 selectively inhibits vacuolar H(+)-ATPases. The substrate requirement and inhibitor susceptibility profile of phagosomal acidification strongly suggest that proton translocation across the phagosomal membrane is mediated by a vacuolar-type H(+)-ATPase.

The mechanism underlying phagosomal acidification was studied in thioglycolate-elicited murine macrophages. The pH of the phagosomal compartment (pH,) was measured fluorimetrically in macrophage suspensions following ingestion of fluorescein isothiocyanatelabeled Staphylococcus aureus.
At 37 "C, pH, decreased rapidly, reaching a steady state value of 5.8-6.1, while the cytoplasmic pH remained near neutrality, pH 7.1. The phagosome to cytosol pH gradient could be collapsed by addition of nigericin, monensin, or weak bases. The substrate dependence and inhibitor sensitivity profile of phagosomal acidification were investigated in intact and permeabilized cells. Phagosomal acidification was inhibited when ATP was depleted using metabolic inhibitors or permeabilizing the plasma membrane by electroporation.
In permeabilized cells, nitrate inhibited the acidification with an apparent Ki of 25 mM. Phagosomal acidification was also effectively blocked by the macrolide antibiotic bafilomycin Al, with an apparent Ki of ~3 nM in both intact and electroporated cells. In this concentration range, bafilomycin A1 selectively inhibits vacuolar H+-ATPases. The substrate requirement and inhibitor susceptibility profile of phagosomal acidification strongly suggest that proton translocation across the phagosomal membrane is mediated by a vacuolar-type H'-ATPase.
The ability of phagocytic leukocytes to ingest and kill microorganisms is one of the major mechanisms whereby microbial invasion is thwarted by the host. Following attachment of the opsonized microorganisms to the cell surface, extension of the phagocytic cell membrane around the particle results in total enclosure and formation of a sealed intracellular compartment, the phagosome. Using pH-sensitive dyes, several investigators have demonstrated the occurrence of progressive acidification of the phagosomal compartment to levels as low as pH 5.5-6.5 (1,6). This ability of the cell to lower phagosomal pH appears to be crucial to its microbicidal functions. The elevated proton concentration is directly lethal for certain microorganisms (7). In addition, low intraphagosomal pH promotes the spontaneous dismutation of superoxide to hydrogen peroxide (8), provides optimal conditions for the activity of certain hydrolytic enzymes (9), and appears to be prerequisite for the process of phagosome-lysosome fusion (6). Indeed, certain microorganisms such as Legionella pneumophila (10) and Toxoplasma gondii (6) appear capable of preventing phagosomal acidification, thereby evading killing by phagocytic cells of the host.
In the present study phagosomal pH was measured by monitoring the fluorescence of FITC'-labeled, opsonized Staphylococcus aureus that were ingested by thioglycolateelicited peritoneal macrophages. Using this technique the putative role of H+-pumping ATPases in phagosomal acidification was investigated. A direct examination of this mechanism was made possible by two recent developments: 1) the availability of bafilomycin Al, a macrolide antibiotic isolated from Streptomyces sp., which is a potent and, at low concentrations very selective inhibitor of vacuolar-type H+-ATPases (19) and 2) the implementation of electropermeabilization techniques (20), which can render the plasma membrane leaky without affecting the phagosomal membrane. This procedure provides direct access to the cytoplasmic face of the phago- to quench all the exposed fluorophores in FITC-labeled bacteria.
The number of bacteria used in Fig. 1

RESULTS
Phagosomal Acid$cation-To measure phagosomal acidification, macrophages were allowed to phagocytose FITClabeled bacteria and were then resuspended in Na' medium in the presence of antifluorescein-IgG to quench extracellular dye ( Fig. 2A) consistently recorded. The following observations indicate that this reduction in fluorescence represents a true acidification of the phagosomal compartment rather than leaking, quenching, or degradation of the dye or the result of ongoing phagocytosis of adherent bacteria. First, addition of the alkali cation-proton exchangers nigericin and monensin increased the fluorescence intensity, presumably through dissipation of the pH gradient across the phagosomal membrane, elevating pH, (Fig. 2A). Second, the fluorescence was increased by addition to the medium of weak bases such as ammonium chloride (Fig. 2B), methylamine, or chloroquine (not shown), as their hydrophobic unprotonated form diffused into the phagosomal compartment and associated with protons, thereby increasing pH, (27). Third, the spontaneous fluorescence decline was abolished when monensin and nigericin were present from the onset of the recording (Fig. 2C). It is noteworthy that addition of nigericin plus monensin after spontaneous acidification restored the fluorescence intensity to the level noted in cells that were prevented from acidifying by the continuous presence of the ionophores (cfi Fig. 2, A and C).
After ingestion of bacteria, and prior to the measurement of fluorescence, the macrophages were stored at 4 "C to prevent further acidification and phagosome-lysosome fusion. When cells were allowed to carry out phagocytosis for only 2 min, the initial pH, recorded upon rewarming to 37 "C was 7.4 (similar to the extracellular pH). In contrast, when cells were allowed to phagocytose for 10 or 20 min prior to the measurement, the initial pH, recorded was 6.74 -I-0.15 (n = 28) and 6.47 f 0.11 (n = S), respectively, indicating that phagosomal acidification had already taken place and was (at least partially) preserved during the subsequent washes and incubation in the cold. Because the intervening incubation on ice was up to 4 h long, dissipation of the transphagosomal membrane pH gradient during storage appears to be comparatively slow. Indeed, to attain complete equilibration of pH, with the cytoplasmic pH, cells had to be incubated overnight on ice after phagocytosis. Unless otherwise specified, a phagocytosis period of 10 min was chosen for the experiments described hereafter. Even though the initial acidification was smaller when using shorter periods, the low absolute level of fluorescence internalized made detection of pH, inaccurate. Upon rewarming to 37 "C, pH, decreased at an initial rate of 0.20 f 0.4 pH unit/min (n = 20) and reached a steady state at 5.89 +-0.18 (n = 28) in 6-8 min. In parallel experiments, changes in the cytosolic pH were measured during this period of phagosomal acidification, using BCECF. To maintain parallel experimental conditions and to avoid interference with the cytosolic pH measurements by bacterial FITC fluorescence, the macrophages were allowed to ingest opsonized, unlabeled S. aureus prior to loading with BCECF. The cytosolic pH slightly increased from 7.0 to 7.1 after resuspension of the cells (not shown). These observations confirm the independence of the two compartments. ATP-dependence of Phagosomal Acidification-Based on the well defined role of vacuolar-type H+-ATPases in the acidification of intracellular organelles (see Refs. 11 and 28-31 for reviews), it was hypothesized that phagosomal acidification was mediated via a H+-ATPase, likely of the vacuolar type. We first examined the ATP dependence of the process. When ATP synthesis was partially inhibited by substituting the medium glucose with the nommetabolized glucose analog 2-deoxy-D-glucose, the rate of acidification was reduced but not abolished and the pH, at steady state was higher ( ml) or rotenone (10 PM) caused a further diminution in the rate and extent of phagosomal acidification (Fig. 3B). Notice that inhibition of acidification was only partial. This is likely due to the fact that depletion of ATP was incomplete, as it was initiated simultaneously with the initiation of the recordings. Prior depletion was not feasible because phagocytosis is an ATP-dependent process and because depletion occurs extremely slowly on ice. The metabolic inhibitors did not interfere with either the pH sensitivity of the fluorophore (data not shown) or the total fluorescence attained after addition of monensin and nigericin (Fig. 3B uersu.s A). Therefore, the reduced rate of fluorescence change reflects a slower acidification, likely due to a significant yet incomplete ATP depletion.
To attain a more efficient and reversible ATP depletion, the plasma membrane of macrophages was permeabilized by electrical discharges. The electroporation capacitance and voltage were selected empirically, based on their ability to permeabilize the plasma membrane, without disrupting the phagosomal membrane. In preliminary experiments, the appropriate voltage range was selected by electronically measuring the colloidosmotic swelling of macrophages that occurs when the plasma membrane is permeabilized in a Na+-containing medium. With an optimal electric field (2.3 kV/cm), about 80% of the cells lost their ability to exclude trypan blue from the cytoplasm, meanwhile preserving their content of FITC-labeled bacteria. Successful permeabilization of the plasma membrane was confirmed examining the loss of the normally impermeant fluorescent indicator BCECF from the cytoplasm. For these studies the cytoplasmic compartment of intact macrophages was loaded with BCECF, the cells were next allowed to phagocytose unlabeled S. aureus, and then permeabilized under the conditions described above. Ninety s after electroporation, the BCECF content of the cells had already decreased to 29.2 f 3.5%, n = 5, and to 19.8 +_ 2.6% after 11 min. In contrast, intact cells lost only a negligible amount of BCECF after 11 min (data not shown). The incomplete depletion of BCECF in the permeabilized cells might be due to the presence of a small fraction of intact cells and/or to the accumulation of BCECF in intracellular compartment(s) other than the cytosol (see Ref. 32). These results suggest that electroporation must have rendered the plasma membrane permeable to Mg2+, ATP, and related nucleotides, since these molecules are smaller than BCECF (M, 520) and trypan blue (M, 960.8), the markers used. Parallel studies were performed to ensure that, under the conditions selected, electroporation did not effect permeabilization of intracellular compartments. Lucifer yellow (M, 457) which had been previously taken up by pinocytosis (for 10 min), did not leak from the endosomal compartment following electroporation (not illustrated). More importantly, the phagosomal compartment seemed to be largely unaffected by the electrical discharges. FITC-labeled bacteria remained within the phagosomes following electroporation and, as shown below, the phagosomes retained their ability to acidify in an ATP-dependent manner.
Following electroporation, macrophages were kept on ice for 5 min in a medium devoid of ATP and Me. In contrast to intact cells, the pH, of porated cells remained near neutrality (7.01 + 0.06; n = 8) after resuspension at 37 "C, presumably due to depletion of ATP and/or M< through leakage (Fig. 4A). Reintroduction of ATP and Me initiated a rapid drop in FITC fluorescence. This fluorescence decrease was also attributable to phagosomal acidification, as it was reversed by the addition of nigericin plus monensin or of ammonium chloride (Fig. 4). The considerable acidification rate obtained in the presence of both agents was not observed when either ATP or Mg2+ alone were added, suggesting that the simultaneous presence of the nucleotide and the divalent cation is required for phagosomal acidification. The rate of Mg-ATP-stimulated acidification in permeabilized cells averaged 0.16 ?Z 0.04 pH/min (n = 8) and the final pH, reached was 6.16 -+ 0.14 (n = 8). These figures resemble the properties of the spontaneous acidification observed in intact cells (see above), suggesting a common mechanism.
The hydrolysis of the high energy y-phosphate bond of ATP appeared to be required for H' (equivalent) translocation, since the nonhydrolyzable ATP analog AMP-PNP failed to support the acidification (Fig. 4B). In addition, ATPrS which can be utilized as a substrate by some protein and lipid kinases, did not support phagosomal acidification (Fig. 4B). Moreover, these analogs antagonized the effect of ATP on pH,, presumably by interacting competitively with the nucleotide-binding site of the transporter (Fig. 48). Other nucleotides and divalent cations were also examined for their ability to restore phagosomal acidification in depleted cells. GTP, ITP, and UTP also supported acidification, albeit less effectively than ATP. Precise determination of the substrate specificity of the proton pump from such measurements was hampered by the nucleoside diphosphokinase activity of the cells. Substitution of M< by Ca2+ greatly diminished the rate of acidification (data not shown). When considered together, these data are consistent with the notion that a H+pumping ATPase is involved in acidification of the phagosome. To further define the nature of this ATP-dependent mechanism, the inhibitor sensitivity profile of phagosomal acidification was examined.

Inhibitor
Sensitivity of Phagosomal Acidification in Intact and Ekctroporated Macrophuges-There are three general groups of proton pumping ATPases: those that form a phosphorylated intermediate or E1E2-ATPases, the mitochondrialtype FIF,,-ATPases, and the vacuolar-type ATPases (34, 35). The three families of pumps differ in their pharmacological properties (see Refs. 11 and 28-31 for reviews): EIEP-type ATPases are inhibited specifically by vanadate, FIFO-ATPases by oligomycin and azide, and vacuolar-type H+-ATPases by NEM, nitrate, and by low nanomolar concentrations of bafilomycin A1 (19). The lipid-soluble carbodiimide DCCD is a potent inhibitor of both vacuolar and FIFO-ATPases (36, 37). We have exploited the substantially different inhibitor sen-   5A). This is similar to the range of concentrations reported to block vacuolar ATPases (lo-50 pM; see references in Refs. 11 and 30), but substantially higher than the concentrations needed to inhibit the FIFO-ATPase (36). NEM has been reported to inhibit vacuolar H+-ATPases at micromolar concentrations (11,(32)(33)(34). In intact cells, however, substantial inhibition was detected only at millimolar concentrations of this agent (Table I Fig. 6A. Phagosomal acidification was measured in macrophages that were allowed to ingest bacteria for 2,5,10, or 20 min. The results obtained at different times were pooled, as there was no difference in the sensitivity to bafilomycin A,. Data are means f S.E. of 8 experiments. Where absent, error bars were smaller than the symbol. apparent K, of 3 nM (Fig. 7), a value that is similar to the K, reported for inhibition of the pump in isolated chromaffin granules, Neurospora crassa vacuoles, Golgi vesicles, and yeast vacuoles (19,41,42).
The acidification initiated by addition of ATP plus M$ to permeabilized cells was also found to be sensitive to bafilomycin Ai (Fig. 4C) and the apparent K; was 5 nM, very similar to that determined in intact cells (not shown). In addition, pretreatment of macrophages with bafilomycin A1 caused irreversible inhibition of the acidification process (not shown). Taken together, the sensitivity of phagosomal acidification to bafilomycin A1, NEM, nitrate, and DCCD and its resistance to vanadate, oligomycin, and azide, are fully compatible with the notion that phagosomal proton transport is mediated by a vacuolar-type H'-ATPase. When added to the cells after pH, had reached a steady state, maximally inhibitory doses of bafilomycin A1 (500 nM) evoked a slow dissipation of the pH gradient, indicating the existence of a measurable proton "leak" permeability of the phagosomal membrane. This implies that continuous proton pumping is required for the maintenance of steady state pH, The rate of the bafilomycin-induced alkalinization was substantially accelerated by addition of the protonophore car-bony1 cyanide m-chlorophenylhydrazone (Fig. 6B). This finding indicates that "leakage" is limited by the proton (equivalent) permeability and not by the counterion conductance. The Origin of the Phagosomal Pumps-Phagosomes are known to fuse with lysosomes and other vesicular organelles (1,43,44). It is therefore conceivable that delivery of the acidic contents of organelles such as lysosomes, endosomes, or tertiary granules into the lumen of the phagosome contributes to phagosomal acidification. A limited number of experiments were performed to determine the origin of the phagosomal H+-ATPases. The phagosomal pumps might be an intrinsic component of the plasma membrane (45, 46) that becomes internalized during phagocytosis (see "Discussion").
Alternatively, the pumps may become incorporated into the phagosome by fusion of an acidic endomembrane compartment. Earlier reports have indicated that substantial phagolysosomal fusion is apparent after -15 min (1,2,47). We therefore examined whether significant fusion of lysosomes with phagosomes occurred under our experimental conditions. Morphological identification of lysosomes of thioglycolate-elicited macrophages was facilitated by the swelling of these organelles after accumulation of the osmotically active, indigestible components of the thioglycolate medium (47). Such swollen lysosomes could be visualized by fluorescence microscopy following pinocytosis of rhodamine B isothiocyanate/dextran or Lucifer yellow and also by accumulation of acridine orange, indicating their acid nature (data not shown). The lysosomes filled with thioglycolate can also be readily identified by electron microscopy, due to their size and content of electron-dense fibrillar material, likely the agar present in the thioglycolate broth (not shown; the swollen structures containing fibrillar material were absent in resident macrophages).
After 10 min of phagocytosis, 95% of the internalized bacteria are present in a compartment that is separate from the large lysosomes, as evidenced by the small size of the phagosome and by the absence of dense fibrillar material in their lumen. These observations suggest that no substantial fusion of phagosomes and large acidic lysosomes had occurred during the incubation period utilized to measure the intraphagosomal pH. These experiments, however, do not exclude the possibility that smaller thioglycolate-free lysosomes fuse to the phagosome.
It was recently demonstrated that the elevation of [Ca2+li that accompanies phagocytosis is required for successful phago-lysosome fusion (48). This observation enabled us to assess the role of phago-lysosome fusion in phagosomal acidification.
The rate of acidification was determined under conditions where the [Ca2+lr increase normally associated with phagocytosis was prevented by omission of extracellular calcium and by loading the cytosol with BAPTA, a calcium chelator. That BAPTA markedly enhanced the cellular calcium buffering capacity was demonstrated by suspending cells in calcium-free medium with EGTA and pulsing with ionomycin, a divalent cation ionophore.
Ionomycin induced a modest [Ca'+]i increase (20-30 nM) in BAPTA-buffered cells, whereas an increment of 300-500 nM was observed in unbuffered cells. The combination of BAPTA and calcium omission also effectively precluded [Ca*+]i increases during phagocytosis. In samples taken at various intervals during and after phagocytosis, [Ca'+]; remained 30-40 nM below the normal resting level of 130 nM as determined with indo-1. Despite the obliteration of the [Ca*+]i increase, there was no substantial difference in the initial rate of acidification (0.2 f 0.01 pH/min uers'sus 0.18 f 0.003 pH/min; six determinations from three independent experiments) or in the pH, recorded upon resuspension after 10 min of phagocytosis (6.55 + 0.07 uersus 6.61 + 0.04) in BAPTA loaded and control cells, respectively. These results further support the notion that phagosomelysosome fusion does not play a major role in the initial phase of phagosomal acidification.

DISCUSSION
The process of phagosomal acidification following ingestion of an infecting microorganism has been well documented and appears to be a crucial step in the normal degradative process characteristic of phagocytic cells (l-6). Despite its important role, however, the mechanism underlying the acidification has been incompletely studied (e.g. Ref. 11.) The present study provides evidence supporting the concept that phagosomal acidification is mediated via a vacuolar-type H+-ATPase active in the phagosomal membrane, which translocates protons from the cytoplasm into the intraphagosomal space. This conclusion was based on two major approaches. First, the Mg-ATP dependence of the process was demonstrated directly. ATP depletion of intact cells by incubation with P-deoxy-Dglucose plus antimycin A or rotenone inhibited phagosomal acidification.
Similarly, when ATP was depleted by leaching electroporated macrophages, acidification was also obliterated. In the latter case, acidification could be restored by the addition of exogenous Mg-ATP to the permeabilized cell suspension. That acidification required ATP hydrolysis was indicated by the failure of AMP-PNP or ATPyS to support acidification.
The second line of evidence indicating involvement of proton pumps was mainly pharmacological.
In addition, this approach revealed that the phagosomal pumps are of the vacuolar type. Briefly, the acidification was insensitive to oligomycin and azide, ruling out the involvement thought to provide 7-chloro-4-nitrobenz-2-oxa-1,3-diazolesensitive proton pumps for the acidification of phagosomes and pinosomes (50). The presence of a similar organelle in macrophages, capable of rapid fusion and proton pump delivery to the phagosomal compartment is a definite possibility, but direct evidence is presently lacking.
The ability of intracellular parasites such as L. pneumophila (10) and T. go&i (6) to prevent phagosomal acidification has been implicated as a mechanism whereby they are able to survive intracellularly.
The present studies demonstrating the role of vacuolar pumps in phagosomal acidification suggests that these pathogens might diminish acidification by directly interacting with the vacuolar H'-ATPase. Alternatively, the parasites or their products may decrease the permeability of the counterions required for net H' translocation, or increase the leakage of acid out of the phagosomal space. Further characterization of the regulation of phagosomal acidification is clearly required, as it might suggest alternative approaches to the treatment of the diseases caused by these microorganisms.