Phagocytosis of infected pathogens by macrophages plays an important role in the early stages of innate immunity. Phagocytosed pathogens are incorporated into phagosomal vacuoles. These phagosomes then interact with endosomal and lysosomal vesicles in a process referred to as phagolysosome biogenesis. During phagolysosome biogenesis, phagosomes acquire degradative and microbicidal properties, leading phagocytosed pathogens to be killed and degraded.
M.tb, the causative bacterium of tuberculosis, infects more than one-third of the human population. M.tb is able to survive and proliferate within phagosomes of the host's macrophages by inhibiting phagolysosome biogenesis (1, 2). However, the exact process by which M.tb blocks phagolysosome biogenesis is not fully understood. Recently, it was reported that phagosomes containing M.tb (M.tb phagosomes) within dendritic cells are associated with lysosomes in the early stages of infection (3). In addition, we have previously demonstrated that LAMP-2, but not cathepsin D, is recruited to M.tb phagosomes in macrophages (4). These results suggest that M.tb phagosomes selectively fuse with lysosomal vesicles which have distinct characteristics. To investigate this possibility, we further examined the localization of two lysosomal marker proteins, CD63 and RILP, on M.tb phagosomes in this study.
Raw264.7 macrophage was obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained in Dulbecco's modified Eagle's medium supplemented with 10% FBS (Invitrogen, Carlsbad, CA, USA), 25 μg/ml penicillin G, and 25 μg/ml streptomycin at 37°C in 5% CO2. M.tb strain H37Rv and Mycobacterium smegmatis mc2155 were grown in 7H9 medium supplemented with 10% Middlebrook ADC (BD Biosciences, San Jose, CA, USA), 0.5% glycerol, 0.05% Tween 80 (mycobacteria complete medium) at 37°C. M tb strain H37Rv transformed with a plasmid encoding DsRed (5) was grown in mycobacteria complete medium with 25 μg/ml kanamycin at 37°C. To construct the plasmids encoding CD63-EGFP and EGFP-RILP, PCR was carried out using cDNA derived from HeLa cells as a template and the following primer sets: human CD63 (5′-CCTCGAGCCACCATGGCGGTGGAAGGAGGAATGAAATG-3′ and 5′-CGGATCCCCATCACCTCGTAGCCACTTCTGATAC-3′), and human RILP (5′-CAGATCTATGGAGCCCAGGAGGGCGGC-3′ and 5′-CGAATTCTCAGGCCTCTGGGGCGGCTG-3′). The PCR products of CD63 and RILP were inserted into pEGFP-N2 and pEGFP-C1 vectors (Clontech, Mountain View, CA, USA), respectively. Transfection of macrophages with plasmids, infection of bacteria with transfected macrophages, CLSM, immunofluorescence microscopy, and isolation of mycobacterial phagosomes were performed as described previously (4). For immunofluorescence microscopy, macrophages were stained with rat anti-CD63 monoclonal antibody (1:30 v/v, MBL, Nagoya, Japan) and Alexa488-conjugated anti-rat IgG antibody (1:1000 v/v, Invitrogen). For immunoblotting analysis, aliquots of 40 μg of cell lysates from Raw264.7 and 15 μg of phagosomal fraction proteins were separated by SDS-PAGE and then subjected to immunoblotting analysis using rat anti-CD63 monoclonal antibody (1:100 v/v, MBL). The unpaired two-sided Student's t-test was used to assess the statistical significance of the differences between the two groups.
CD63 has been shown to be localized to the phagosome during phagolysosome biogenesis (2, 6), but its localization on live mycobacterial phagosomes is still controversial (2, 3, 7). CD63 was originally identified as a platelet activation marker (8) and has also been used as a marker for late endosomes and lysosomes because of its function in phagosome acidification (9–12). We therefore re-assessed CD63 localization on M.tb phagosomes in infected macrophages (Fig. 1). Raw264.7 macrophages transfected with a plasmid encoding CD63-EGFP were infected with M.tb expressing DsRed. Infected cells were fixed and observed by CLSM. Clear CD63 localization was observed on more than 60% of M.tb phagosomes at 30 min and 6 hr post infection (Fig. 1a, b). To rule out the possibility that CD63 localization on M.tb phagosomes is caused by exogenous expression of CD63-EGFP, immunofluorescence microscopy with anti-CD63 antibody was performed (Fig. 1c). We found that endogenous CD63 was also localized to about 60% of M.tb phagosomes at 6 hr post infection. To confirm the recruitment of CD63 to live M.tb phagosomes biochemically, we carried out immunoblotting analysis for CD63 in isolated mycobacterial phagosome fractions (Fig. 1d). Raw264.7 macrophages were allowed to phagocytose heat-inactivated M. smegmatis or infected with M.tb for 6 hr, and the phagosomal fractions isolated as described previously (4, 13). Proteins extracted from isolated phagosomal fractions were subjected to immunoblotting analysis using anti-CD63 antibody. Immunoblotting analysis revealed that CD63 is recruited to live M.tb phagosomes as well as to heat-inactivated M. smegmatis phagosomes. These results suggest that M.tb phagosomes fuse with CD63-positive lysosomal vesicles.
RILP interacts with the active form of Rab7 and mediates the fusion of endosomes with lysosomes (14, 15). RILP is also reported to be localized to the phagosome and to recruit the minus-end motor complex dynein-dynactin to the phagosome, resulting in migration of the phagosome to the MTOC where late endosomal and lysosomal vesicles accumulate (16). In the process of recruitment of RILP to the phagosome, tubular vesicles expressing RILP have been observed to be elongated from the MTOC, fusing with the phagosome (16). RILP has been reported to be absent from the Mycobacterium bovis strain BCG phagosome despite Rab7 localization (17). We have previously shown that Rab7 is transiently recruited to, and subsequently released from, M.tb phagosomes (4), but the interaction of RILP with M.tb phagosomes has not been previously reported. We examined the subcellular localization of EGFP-RILP in macrophages infected with M.tb (Fig. 2). In M.tb-infected macrophages, RILP-positive phagosomes appeared and increased to 30% of M.tb phagosomes up until 30 min post infection (Fig. 2a, c). No further increase was seen after this time (Fig. 2b, c). On the other hand, the proportion of RILP-positive Staphylococcus aureus phagosomes continued to increase beyond 30 min post infection (Fig. 2c). We also found that the proportion of RILP-positive phagosomes containing heat-inactivated M.tb reached more than 80% at 6 hr post infection. These results suggest that further recruitment of RILP to phagosomes containing live M.tb after 30 min post infection might be actively inhibited.
Next, we examined whether recruitment of CD63 and RILP to phagosomes depends on the function of Rab7 in macrophages. Raw264.7 macrophages transfected with two plasmids encoding either EGFP-fused CD63 or RILP and a dominant-negative form of Rab7, Rab7T22N, were allowed to phagocytose latex-beads for 2 hr and were then examined by CLSM for localization of lysosomal proteins on the phagosomes. Both lysosomal markers were localized to latex-bead-containing phagosomes in the control cells (Fig. 3a-1, b-1). CD63 was found on the majority of latex-bead-containing phagosomes in the cells expressing Rab7T22N (Fig. 3a-2, a-3), as well as in the control cells. However, consistent with previous findings, RILP was not present on latex-bead phagosomes in cells expressing Rab7T22N (Fig. 3b-2, b-3) (17). We also found that clustering of RILP in the perinuclear regions was disrupted and diffused by the expression of Rab7T22N. Collectively, our data demonstrate that Rab7 is vital for recruiting RILP to phagosomes during the maturation process, but not for recruiting CD63.
How M.tb escapes the effects of the bactericidal components within the phagosome while still acquiring nutrients for growth is very important question. It has been suggested that mycobacterial phagosomes arrest their maturation at an early stage and completely avoid fusion with lysosomes (18, 19). However, we have shown the localization of CD63 (Fig. 2) and LAMP-2 (4) on M.tb phagosomes in macrophages. It has been proposed that phagolysosome biogenesis is achieved by a series of fusions with heterogeneous lysosomes (20). This model is supported by a report demonstrating the existence of sub-populations of lysosomes in macrophages (6). Our previous and current studies demonstrating the alternative localization of lysosomal markers on M.tb phagosomes further support this model. From these observations, it seems that dissociation of Rab7 from M.tb phagosomes selectively inhibits fusion with harmful lysosomes despite continued fusion with non-microbicidal lysosomes.
In conclusion, based on our findings we propose the following model for M.tb-induced inhibition of phagolysosome biogenesis: Early M.tb phagosomes are capable of recruiting Rab7 and can potentially fuse with lysosomes. RILP is also recruited to M.tb phagosomes, which form the Rab7-RILP-dynein/dynactin protein complex followed by promotion of phagolysosome biogenesis. However, viable M.tb is able to release Rab7 from phagosomes, resulting in inhibition of further fusion with lysosomal vesicles and disassembly of the RILP-phagosome complex. This causes the blocking of subsequent phagolysosome biogenesis. On the other hand, non-microbicidal vesicles expressing CD63 and/or LAMP-2 continuously fuse with M.tb phagosomes despite Rab7 dissociation, and this fusion would support the acquisition of nutrients for mycobacterial proliferation within the phagosome.