pH-dependent Perforation of Macrophage Phagosomes by Listeriolysin O from Listeria monocytogenes

All reagents were purchased from Sigma Chemical Co. (St. Louis, MO) or Molecular Probes (Eugene, OR). Mouse bone marrow–derived macrophages were obtained from the femurs of female C3H/HeJ mice (The Jackson Laboratory, Bar Harbor, ME) and cultured in vitro as described (19). Strains of L. monocytogenes were kindly provided by Dr. Daniel Portnoy (University of Pennsylvania, Philadelphia, PA) and were grown to stationary phase in Luria-Bertani broth. Just before use, bacteria were washed three times in Ringer's buffer (RB; 155 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 2 mM NaH₂PO₄, 10 mM Hepes, 10 mM glucose, pH 7.2). Fresh bacteria were prepared every 2 h.

Macrophages were kept on a temperature controlled stage set to 37°C, and coverslips containing 1.5–2.0 × 10⁵ cells were incubated with 1 ml 5 mM HPTS in RB and 5–10 μl of a washed L. monocytogenes culture (∼10 CFU/cell). The macrophages were incubated for up to 10 min, during which time an L. monocytogenes–infected cell was located. The coverslip was then quickly washed six to seven times with RB, and pH recordings were begun immediately (time 0). Washes with RB were repeated as necessary to reduce background fluorescence.

To inhibit vacuole acidification, 10 mM NH₄Cl was included in the first RB wash and all subsequent solutions. Alternatively, 0.5 μM bafilomycin A₁ was added to the cells 30 min before infection and throughout the experiment.

Fluorescent images were collected on an IM-35 inverted microscope (Carl Zeiss Inc., Thornwood, NY) equipped for epifluorescence using an ×100 lens, numerical aperture 1.32. The fluorescence excitation system included a 50 W mercury arc lamp and filter changer (Stephen Baer, Photome, Cambridge, MA), which provided alternating 405 and 440 nm excitation wavelengths (exc.); emission was measured at 520 nm. Neutral density filters were sometimes added to reduce photodamage. Images were collected through a video camera, intensified through a multichannel plate intensifier (Video Scope Int. Ltd., Washington, DC). During recording, three images were taken every 15 sec: a 405 nm image, a 440 nm image, and a phase-contrast image. Digital images were assembled into video sequences.

Images were analyzed with a Metamorph image analysis system (Universal Imaging Corp., West Chester, PA). Two excitation wavelengths (440 and 405 nm) were used to calculate a ratio value to correct for differences in local dye concentration, optical pathlength, illumination intensity, and photobleaching of samples (20). 440:405 nm ratios were calculated by thresholding the 405 nm image to restrict measurement to the vacuoles of interest. The intensities for identical masked areas were obtained from each of the two fluorescent images and divided to obtain a ratio value. These values were calibrated using HPTS-labeled macrophages incubated with 10 μM nigericin in isotonic potassium buffers of known pH (130 mM KCl, 1 mM MgCl₂, 15 mM Hepes, 15 mM MES, 0.02% sodium azide). A standard curve was generated by collecting 405 and 440 nm images of fluorescent vacuoles equilibrated at each pH (5.0, 5.5, 6.0, 6.5, 7.0, and 7.5).

L. monocytogenes were internalized into spacious phagosomes by a process resembling macropinocytosis. This was different from the reported mechanism of L. monocytogenes entry into epithelial cells (21), but similar to the way Salmonella typhimurium enter macrophages (22). Immediately after phagocytosis, individual bacteria could be seen by phase-contrast microscopy as phase-dense rods moving freely inside phase-bright phagosomes; these phagosomes also contained HPTS which was included during the infection (see Fig. 1,a, vesicle 2; Fig. 2,a, vesicle 1; and Fig. 3 a, vesicle 3). No tight-fitting phagosomes were observed, although older phagosomes became smaller and restricted the movement of bacteria.

The pH of L. monocytogenes–containing vacuoles and of phase-bright, fluorescent vacuoles that did not contain bacteria was monitored by collecting phase contrast and fluorescence images at regular intervals (Fig. 1). Fluorescent images of HPTS in vacuoles were taken at exc. 405 and 440 nm. The ratio of fluorescence from these two recordings indicated the pH of the probe environment (18): bright fluorescence at exc. 440 nm and dim fluorescence at exc. 405 nm indicated higher pH (>7.2), whereas bright fluorescence at exc. 405 nm and dim fluorescence at 440 nm indicated low pH (<5.0).

Phagosomes began to acidify soon after phagocytosis, with larger ones acidifying more slowly. A typical infection with wild-type L. monocytogenes (Fig. 1), showed one or more macropinosomes (Fig. 1, a–c, vesicle 1) and phagosomes (Fig. 1, a–c, vesicle 2) labeled with HPTS. Whereas the macropinosome remained acidic throughout the period of observation, acidified phagosomes often showed a sudden increase of fluorescence at exc. 440 nm, reflecting an increase in pH (Fig. 1, b and c, frame 3, vesicle 2). Sometimes a slow decrease in pH was followed by an abrupt alkalinization of the phagosome (see Fig. 3, a–c, vesicle 2). The increases in pH were often quickly followed by loss of the fluorescence from the vacuole (Fig. 2, a–c, vesicle 1; Fig. 3, a–c, vesicles 1 and 2). Concomitant increases in the 440 nm fluorescence intensity in the nucleus and cytoplasm of the same cell were usually seen (Fig. 2). Inasmuch as L. monocytogenes escapes from the phagocytic vacuoles of macrophages in an LLO-dependent process (8–11, 13), and as HPTS has been shown to be released from LLO-destabilized compartments (23), the sudden pH increases likely resulted from equilibration of the pH across the vacuolar membrane after perforation by LLO. Further destabilization of the membrane released HPTS into the cytoplasm.

To calculate the pH of individual phagosomes, digital images were recorded at exc. 440 and 405 nm, and the intensity values from corresonding areas in the two images were divided to obtain ratio values (I₄₄₀:I₄₀₅). The ratios were calibrated using pH-clamped cells. Fig. 4, A and B, shows representative changes of pH for vacuoles in cells infected with wild-type L. monocytogenes. The pH of the bacterium-containing vacuoles decreased and then abruptly increased by 0.2–1.6 pH U. Increases in pH were observed in 24 out of 52 vacuoles containing L. monocytogenes (Table 1). The increases were rapid and usually accompanied by release of dye into the cytoplasm. Release of HPTS did not always follow the rise in pH. In fact, sometimes vacuoles reacidified (Fig. 4 A). This reacidification was often followed by another sudden increase in pH and more complete release of the HPTS into the cytosol.

Sudden increases in pH were sometimes seen in macropinosomes of cells infected with the wild-type bacterium (Figs. 3 and 4 B; 7 out of 36 vacuoles). These increases occurred in a pH range similar to that of phagosomes and probably resulted from soluble LLO pinocytosed from the medium.

Two mutants deficient in LLO failed to show any sudden increases in vacuolar pH. The first mutant, DP-L2864 (hly-, plca-, plcb-, mpl-) lacks several virulence factors implicated in vacuolar perforation, including LLO, a phosphatidylinositol-specific phospholipase C, a phosphatidylcholine-preferring phospholipase C, and a metalloprotease (24–27). The second mutant, DP-L2161 (28), has a single, in-frame deletion of hly, the structural gene for LLO. Vacuoles containing either of these acidified and maintained a low pH. Sudden increases in pH and release of dye into the cytoplasm were never observed with either mutant (Table 1), indicating that LLO was required for perforation.

To determine the optimal pH for LLO activity in vivo, we measured the lowest pH achieved by each vacuole before perforation. Perforation occurred over a range of acidic pH values from 4.9 to 6.7, with a mean near 6.0 (mean = pH 5.94, SD = 0.45; Fig. 5), consistent with earlier suggestions that vacuole acidification creates optimal conditions for LLO activity in vivo (15–17).

There are conflicting reports as to whether inhibitors of vacuole acidification prevent L. monocytogenes from moving to the cytosol (5, 29, 30). To determine whether low pH was required for LLO to induce vacuolar perforation, macrophages were infected with wild-type L. monocytogenes in the presence of NH₄Cl, a weak base that raises the pH of acidic vesicles (31), or bafilomycin A₁, an inhibitor of vacuolar proton ATPases (32). In the presence of bafilomycin A₁, phagosomes maintained a steady, alkaline pH (Fig. 4,C). Vacuolar perforation, as monitored by leakage of the HPTS to the cytoplasm, was completely prevented by bafilomycin A₁ (0 out of 16; Table 1) and was less frequent in NH₄Cl (3 out of 16). Consistent with the results of Okhuma and Poole (33), 10 mM NH₄Cl elevated pH but did not completely prevent acidification of macrophage vacuoles. 65% of L. monocytogenes–containing vacuoles in NH₄Cl-treated cells were pH 7 or less, as opposed to 33% of vacuoles in bafilomycin A₁–treated cells. The vacuoles that showed perforation in the presence of NH₄Cl were within the range of pH at which perforation occurred without this compound. These results show that the level of acidification corresponded with the amount of LLO activity observed (46% for buffer, 19% for NH₄Cl, and 0% for bafilomycin A₁; Table 1). This is consistent with the fact that weak bases cause a slight decrease in intracellular growth of L. monocytogenes (5, 29), as the number of bacteria that could escape to the cytosol was reduced in the presence of NH₄Cl.

Although it is established that L. monocytogenes escapes its vacuole, the mechanism of this process remains ill-defined. Evidence presented here shows that one of the earliest detectable steps in vacuole lysis is perforation of the membrane, indicated by an abrupt increase in vacuolar pH and by leakage of its contents. These changes required LLO and acidic pH. We conclude that acidic pH facilitates LLO activity in vivo through an undefined mechanism.

The authors wish to thank Dr. Daniel Portnoy for providing the strains used in this paper and for critical reading of the manuscript, Dr. Wayne Lencer for providing the bafilomycin A₁ used in the initial experiments, David Garber for critical reading of the manuscript, and members of the Swanson lab for helpful suggestions.

This work was supported by National Institutes of Health grants AI-35950 (to J.A. Swanson) and AI-22021 (to R.J. Collier).