Mycobacterial infection of macrophages results in membrane-permeable phagosomes

Abstract

Cell-mediated immunity is critical for host resistance to tuberculosis. T lymphocytes recognizing antigens presented by the major histocompatibility complex (MHC) class I and class II molecules have been found to be necessary for control of mycobacterial infection. Mice genetically deficient in the generation of MHC class I and class Ia responses are susceptible to mycobacterial infection. Although soluble protein antigens are generally presented by macrophages to T cells through MHC class II molecules, macrophages infected with Mycobacterium tuberculosis or bacille Calmette-Guerin have been shown to facilitate presentation of ovalbumin through the MHC class I presentation pathway via a TAP-dependent mechanism. How mycobacteria, thought to reside within membrane-bound vacuoles, facilitate communication with the cytoplasm and enable MHC class I presentation presents a paradox. By using confocal microscopy to study the localization of fluorescent-tagged dextrans of varying size microinjected intracytoplasmically into macrophages infected with bacille Calmette-Guerin expressing the green fluorescent protein, molecules as large as 70 kilodaltons were shown to gain access to the mycobacterial phagosome. Possible biological consequences of the permeabilization of vacuolar membranes by mycobacteria would be pathogen access to host cell nutrients within the cytoplasm, perhaps contributing to bacterial pathogenesis, and access of microbial antigens to the MHC class I presentation pathway, contributing to host protective immune responses.

Figure 1

Mycobacteria reside within permeable phagosomes admitting molecules up to 70 kDa. Macrophages infected with GFP-BCG for 18 hours were microinjected with Texas Red-tagged dextrans (3 kDa) and immediately were visualized by confocal microscopy. A and D show fluorescence in the green channel only, B and E show fluorescence in the red channel only, and C, F, G, and H are merged images. Colocalization of green fluorescent bacilli (A, yellow arrow) with 3-kDa-sized Texas Red-tagged dextrans (B, yellow arrow) manifested as yellow bacilli upon merging of the two channels (C, yellow arrow). Not all GFP-BCG (A, white arrow) are found within membrane-permeable vesicles (B, white arrow), and therefore did not exhibit signal colocalization (C, white arrow). Microinjection of 2,000-kDa-sized dextrans revealed green bacilli (D, white arrow) within vesicles clearly impermeable to the infected dextrans (E and F, white arrows). Mycobacterial phagosomes are permeable to injected proteins, within the appropriate size range, as ovalbumin accesses the mycobacterial phagosome (G, yellow arrow). Molecules of up to 70 kDa in size access the mycobacterial phagosome, as evidenced by the presence of yellow bacilli indicating signal colocalization (H, yellow arrow). Colocalization is not a limit of resolution artefact as Oregon-green-tagged beads ≤0.5 μm are readily resolved from the surrounding Texas-red-tagged dextrans (I, white arrow).

Figure 2

Signal colocalization is not a plane of section artifact. BCG-GFP-infected macrophages were microinjected with 40-kDa-sized Texas Red-tagged dextrans and immediately were visualized by confocal microscopy. Successive optical sections were collected along the z axis, with a step of 0.9 μm. Three optical planes taken from the substratum up—A–C, D–F, and G–I—are shown. A, D, and G show fluorescence in the green channel only, B, E, and H show fluorescence in the red channel only, and C, F, and I are merged for RGB color. Numerous bacilli (A, white arrow) remained within clearly discernable vesicles impenetrable to the dextrans (B, white arrow), thereby precluding signal colocalization. Occasionally, yellow bacilli seemingly within compartments accessible to the dye appeared within an impermeable vesicle only after evaluation of several successive optical sections (compare bacteria in I, white arrow, with C). A significant number of BCG-GFP, however, facilitated dextran signal colocalization at every plane of section (D–F, yellow arrows).

Figure 3

BCG reside within vesicles permeable to cytosolic macromolecules of up to 70 kDa in size. Macrophages infected with BCG-GFP and microinjected with 3- or 40-kDa Texas Red-tagged dextrans or tagged-ovalbumin showed appreciable colocalization of the two signals. Microinjection of 70-kDa tagged dextrans into macrophages infected with live BCG-GFP demonstrates reduced colocalization, probably indicating an upper limit of accessibility to the phagosome. Injection of 2,000 kDa failed to reveal any appreciable signal colocalization.

Figure 4

Listeria strains expressing the listeriolysin O reveal colocalization with 2,000-kDa-sized macromolecules. Macrophages were infected with LLO⁺-GFP (A–C) for 3 hours before microinjection with 2,000-kDa Texas Red-tagged dextrans and were visualized by confocal microscopy. LLO⁺-GFP (A, yellow arrow) were readily observed within regions completely accessible to the large dextran molecules (B), thereby enabling colocalization of the two fluorescent signals (C). Macrophages infected with LLO-GFP (D, white arrow) then microinjected with 3-kDa Texas Red-tagged dextrans (E) failed to reveal signal colocalization (F).

Figure 5

Only viable mycobacterial strains of the tuberculosis complex, as well as listeriolysin O expressing L. monocytogenes strains, efficiently access the cytoplasm. Macrophages infected with killed BCG-GFP and microinjected with 3 kDa revealed no appreciable colocalization of the two signals, and neither did macrophages infected with live SMEG-GFP injected with 3- or 40-kDa-sized tagged dextrans. Intracytoplasmic microinjection of 3-kDa tagged dextrans in macrophages infected with LLO-GFP similarly demonstrate negligible colocalization, indicating bacterial residence within a membrane-impermeable vesicle. Injection of 2,000-kDa Texas Red-tagged dextrans in LLO+GFP revealed substantial signal colocalization (50%), indicating free access of the bacteria to cytoplamically located macromolecules.