Cell, Vol.

97, 435447, May 14, 1999, Copyright 1999 by Cell Press

A Coat Protein on Phagosomes Involved in the Intracellular Survival of Mycobacteria

Giorgio Ferrari,* Hanno Langen, Makoto Naito, and Jean Pieters* * Basel Institute for Immunology Grenzacherstrasse 487 CH-4005 Basel Switzerland Hoffmann-La Roche CH-4002 Basel Switzerland Department of Pathology Niigata University School of Medicine Niigata 951-8510 Japan events along the phagosomal and endocytic pathway to lysosomes, where such material is degraded (Desjardins et al., 1994; Russell, 1995). In the case of mycobacteria, early observations suggested that phagosomes containing living mycobacteria resist fusion with organelles of the endosomal/lysosomal system (Armstrong and DArcy Hart, 1971). This inhibition depends on active processes exerted by the mycobacteria, as dead mycobacteria are readily transported to endosomal/lysosomal organelles (Armstrong and DArcy Hart, 1971; Barker et al., 1997; Hasan et al., 1997). This resistance of the mycobacterial phagosome to fusion with endosomal/ lysosomal organelles is thought to be related to the ability of mycobacteria to escape from macrophage bactericidal mechanisms (Small et al., 1994; Russell, 1995). Morphological characterization of the mycobacterial phagosome showed that this organelle contains markers of the early but lacks markers of later stages of the endocytic pathway, such as late endosomes and lysosomes (Sturgill-Koszycki et al., 1994; Clemens and Horwitz, 1995; Barker et al., 1997). Interestingly, the mycobacterial phagosome selectively excludes the proton-ATPase that is normally responsible for acidification along the endocytic pathway (Mellman et al., 1985; Sturgill-Koszycki et al., 1994). The resulting limited acidification of the mycobacterial phagosome is thought to represent a key mechanism allowing mycobacterial survival. However, the molecular mechanisms by which the mycobacterial phagosome resists fusion with or maturation into lysosomes remain largely unknown. As an approach toward the identification of components involved in these processes, we have developed a system in which phagosomes containing living mycobacteria could be physically separated from phagosomes containing dead bacilli. We reasoned that biochemical analysis of these two populations of vesicles might allow the identification of factors that contribute to the intraphagosomal survival of mycobacteria. We here report the identification and characterization of a WD repeatcontaining molecule, termed TACO (for tryptophane aspartatecontaining coat protein), that was associated with phagosomes containing live, but not with those containing killed, bacteria. TACO represents a coat protein that transiently associated with phagosomes but was actively retained on mycobacterial phagosomes as long as the bacteria were viable. The capacity of viable mycobacteria to retain TACO on the phagosomal membrane, thereby preventing their delivery to lysosomes, thus represents a mechanism by which these pathogens are able to escape host defense mechanisms.

Summary Mycobacteria are intracellular pathogens that can survive within macrophage phagosomes, thereby evading host defense strategies by largely unknown mechanisms. We have identified a WD repeat host protein that was recruited to and actively retained on phagosomes by living, but not dead, mycobacteria. This protein, termed TACO, represents a component of the phagosome coat that is normally released prior to phagosome fusion with or maturation into lysosomes. In macrophages lacking TACO, mycobacteria were readily transported to lysosomes followed by their degradation. Expression of TACO in nonmacrophages prevented lysosomal delivery of mycobacteria and prolonged their intracellular survival. Active retention of TACO on phagosomes by living mycobacteria thus represents a mechanism preventing cargo delivery to lysosomes, allowing mycobacteria to survive within macrophages.

Introduction Pathogenicity of microorganisms is often related to their ability to survive and replicate within their mammalian hosts. After entry, a pathogen should find a suitable site for survival and replication within the cell while at the same time evade or circumvent host defense mechanisms. As many pathogens have coevolved with the organisms that they infect, they may use host cellular components for their entry, survival, and growth (Falkow, 1991; Bliska et al., 1993; Beverley, 1996; Galan and Bliska, 1996; Finlay and Cossart, 1997). One example of a successful exploitation of cellular mechanisms to survive is provided by the interaction of Mycobacterium species with macrophages. Whereas the normal function of macrophages is to engulf and destroy microorganisms, mycobacteria have evolved ways to circumvent the hostile environment of the macrophage (Russell, 1995). Phagocytosed material is normally transported via a series of fission and fusion
To whom correspondence should be addressed (e-mail: pieters@

Results Biochemical Characterization of Mycobacterial Phagosomes To study the organelles in which mycobacteria reside inside macrophages, the interaction of M. bovis BCG

bii.ch).

Cell 436

Figure 1. Detection of a 50 kDa Protein Present in Phagosomes Containing Viable Mycobacteria (a and b) Subcellular fractionation of phagosomes containing viable (a) and heat-killed (b) mycobacteria. The murine macrophage cell line J774 was infected for 1 hr with M. bovis BCG, followed by a 12 (live) or 2 (dead) hr chase, after which a homogenate was prepared. Organelles were subsequently separated by organelle electrophoresis, and the distribution of the organelle-specific markers as indicated were determined (upper panels). From each fraction, the amount of mycobacteria was quantitated by direct observation by light microscopy after acid fast staining (lower panels). (c) Analysis of proteins present in subcellular fractions isolated from metabolically labeled macrophages. From infected macrophages (upper panels; left, live; right, dead), bacteria containing fractions after organelle electrophoresis were pooled and phagosomes were prepared by low-speed centrifugation as described in Experimental Procedures. Separately, from uninfected macrophages a late endosomal/lysosomal fraction as well as a lysosomal fraction (from pooling exclusively the most anodally -hexosaminidase containing shifted membranes) was prepared by organelle electrophoresis and high-speed centrifugation (lower panels). Proteins from phagosomes were separated by two-dimensional IEF/SDSPAGE and analyzed by fluorography. Acidic, left; basic, right. The 50 kDa protein specifically associated with phagosomes containing live bacteria is indicated by an arrow. The major spot in phagosomes containing viable mycobacteria represents actin.

with the murine macrophage cell line J774A.1 was investigated. To analyze the molecular composition of mycobacteria-containing phagosomes, subcellular fractionation was utilized. Macrophages that had been metabolically labeled with 35S-methionine/cysteine were infected with live BCG for 1 hr, washed, and chased for 12 hr prior to homogenization and subcellular fractionation. To separate the different subcellular compartments, organelle electrophoresis was used (Tulp et al., 1994). During such a separation procedure, late endosomes and lysosomes shift toward the anode (Figures 1a and 1b, upper panels), whereas mycobacterial phagosomes remain in unshifted fractions, as analyzed by the presence of bacteria (Figure 1a, Live, lower panel). Killing of the mycobacteria

by heat treatment prior to uptake by macrophages for 1 hr followed by a 2 hr chase (instead of 12 hr, to prevent excessive degradation of the bacilli) resulted in their recovery in shifted fractions, together with late endosomal/lysosomal membranes (Figure 1b, Dead, lower panel). Thus, using this method, vacuoles containing live mycobacteria could be physically separated from those containing dead mycobacteria. To analyze the protein composition of the phagosomes containing viable as well as those containing heat-killed mycobacteria, the bacteria-containing fractions were pooled and sedimented at low speed to prepare a phagosomal fraction as described in Experimental Procedures. Separately, a late endosomal/lysosomal

Mycobacterial Subversion of Phagocytosis 437

fraction as well as a lysosomal fraction was prepared from uninfected macrophages by organelle electrophoresis and high-speed sedimentation. Protein patterns from these organelles were analyzed by two-dimensional IEF/SDSPAGE followed by fluorography (Figure 1c). Most polypeptides present in phagosomes containing living mycobacteria (Figure 1c, Phagosomes (Live)) could also be detected in endosomal/lysosomal organelles from uninfected macrophages (Figure 1c, Late endosomes/Lysosomes). However, in phagosomal membranes containing live mycobacteria, a 50 kDa polypeptide was present (Figure 1c, arrow) that was not detected in any of the other subcellular fractions analyzed, including phagosomes containing killed mycobacteria (Figure 1c, Phagosomes (Dead)). Phagosomes containing killed bacteria were enriched for a subset of lysosomal proteins (compare Figure 1c, right hand panels), all of which were present in lysosomes prepared from an uninfected homogenate (Figure 1c, Lysosomes). The 50 kDa polypeptide was absent from fluorographs prepared from a similar subcellular fraction in the absence of infection, as well as from metabolically labeled bacteria (data not shown), and could be detected in total homogenates from uninfected macrophages, indicating that it represented a host protein. Cloning and Sequencing of the Phagosomal Protein of 50 kDa To identify the 50 kDa protein, membrane fractions were subjected to two-dimensional IEF/SDSPAGE, and after Coomassie blue staining the 50 kDa polypeptide was excised and subjected to lys-C protease digestion and mass spectrometry (Figure 2a). Nine out of 11 peptides matched the digestion profile of entry P31146 (SwissProt), representing a 57 kDa protein previously identified in human cells (Suzuki et al., 1995). Based on this homology, the cDNA encoding the mouse homolog of this protein was cloned from a macrophage library as described in Experimental Procedures. Sequencing of positive clones revealed an open reading frame of 1386 base pairs encoding a 461 amino acid residue polypeptide with a predicted sequence shown in Figure 2b. To confirm that this reading frame encoded the 50 kDa protein that accumulated in mycobacterial phagosomes, a lys-C digest of the polypeptide isolated from the twodimensional gel was separated by HPLC. Sequence analysis of the peptide eluting at 21 min by Edman degradation revealed the sequence FRHV, matching amino acid residues 1114 (Figure 2b). In addition, antibodies raised against the C-terminal amino acid sequence were used to immunoprecipitate the 50 kDa protein from lysates of noninfected macrophages that had been metabolically labeled. Analysis by two-dimensional IEF/SDSPAGE and coelectrophoresis with phagosomal organelles showed that the immunoprecipitated protein migrated at the identical position as the phagosomal 50 kDa protein (Figure 2c). Interestingly, several additional molecules of the same molecular weight but differing in isoelectric point were detected on the fluorograph, possibly representing post-translationally modified forms or, alternatively, closely related proteins. The primary sequence of the 50 kDa polypeptide contains five WD repeats (underlined in Figure 2b). WD repeats are characterized by a 3040 amino acid residue

Figure 2. Identification of the 50 kDa Protein (a) The polyacrylamide gel piece containing the 50 kDa protein (see Figure 1) was excised and subjected to lys-C digestion and mass spectrometry. Shown is the spectrum obtained after mass spectrometry. (b) Predicted amino acid sequence of the 50 kDa protein. The five WD repeats present in the sequence are underlined. (c) Immunoprecipitation using anti-TACO antibodies of cell lysates from metabolically labeled macrophages followed by two-dimensional IEF/SDSPAGE. The gel was coelectrophoresed with different organelle fractions to allow alignment of the phagosomal 50 kDa spot (arrow).

segment bordered by Gly-His (GH) and Trp-Asp (WD) dipeptide residues, but their function remains unknown (Neer et al., 1994). A BLAST search with the 50 kDa protein sequence revealed homology to IR10 (40% identity), a human WD repeat protein of unknown function, as well as to a WD repeat protein identified in Dictyostelium discoideum, termed coronin (35% identity), an actinbinding protein involved in actin-based cytoskeletal rearrangements. Furthermore, analysis of the primary sequence using Prosite revealed the lack of a signal sequence for targeting to the endoplasmic reticulum as

Cell 438

well as the presence of a coiled-coil region within the 32 C-terminal amino acid residues (Berger et al., 1995). Because of the presence of the WD repeats as well as its morphological characterization (cf. later), we have named this protein TACO for tryptophane aspartate containing coat protein. Tissue Distribution of TACO Several members of the WD repeat family are involved in regulating cytoskeletal assembly and membrane traffic, thus carrying out functions that are essential for normal cellular growth and therefore expected to be expressed in all cell types (Neer et al., 1994). To analyze the tissue distribution of TACO, proteins from a homogenate prepared from various tissues were separated by SDS PAGE and transferred to nitrocellulose (Figure 3a). After staining of the membrane using anti-TACO antibodies, immunoreactivity was found to be present in spleen, lymph nodes, thymus, and brain, as well as a small amount in lung, but not in any other tissue examined (Figure 3a, upper panel). As a control, the immunoblot was stripped and stained for actin (Figure 3a, middle panel) as well as for tubulin (Figure 3a, lower panel). In addition, bone marrowderived and peritoneal macrophages as well as all macrophage cell lines analyzed expressed TACO (data not shown), with the notable exception of Kupffer cells (see later). Thus, TACO is expressed in cells of the lymphoid/myeloid lineage, consistent with a function in macrophages and macrophage-like cells. Subcellular Localization of TACO The subcellular distribution of TACO was analyzed morphologically using confocal immunofluorescence microscopy. In noninfected macrophages, a rat monoclonal antibody raised against TACO predominantly labeled the cortex of the cells and colocalized with tubulin, but not with actin (Figure 3b). Double labeling of cells for actin and tubulin showed only a partial colocalization (Figure 3b), indicating the presence of two distinct cortical cytoskeletal elements in these cells. Together these results suggest that in noninfected macrophages, TACO is associated with the cortical microtubule network. Localization of TACO in Infected Cells Since TACO was identified as a component of phagosomes containing viable mycobacteria, the subcellular distribution in cells infected with mycobacteria was analyzed. To visualize mycobacteria-containing phagosomes, M. bovis BCG expressing green fluorescent protein (BCG-GFP) was utilized. Cells were allowed to internalize mycobacteria for 1 hr, washed and trypsinized to remove adherent bacteria, and cultured at 37 C for the times indicated in Figure 4a. Immediately after infection (chase 0 hr), TACO was localized around the mycobacterial phagosomes, in addition to a cortical localization. Within the first 2 hr, TACO became almost completely relocalized to the phagosomal membrane and remained associated with the phagosome up to prolonged periods of time (Figure 4a). When heat-killed bacteria were allowed to be internalized by macrophages, TACO accumulated initially at the site of bacterial uptake but redistributed to the cell

Figure 3. Distribution and Localization of TACO (a) Tissue distribution of TACO. The different tissues as indicated were isolated from a mouse, homogenized, and total proteins (15 g) separated by SDSPAGE and immunoblotted using anti-TACO antiserum (upper panel). After stripping, the blot was reincubated with anti-actin antibodies (middle panel) or anti-tubulin antibodies (lower panel). (b) Subcellular localization of TACO in uninfected cells. Macrophages were grown on coverslips, fixed, permeabilized, and stained for tubulin (upper left) followed by Texas redlabeled secondary antibodies or F-actin using phalloidin-Texas red (upper right) and double labeled for TACO using a rat monoclonal antibody followed by FITC-labeled secondary antibodies. (Lower left) Localization of actin (Texas red) and tubulin (FITC). (Lower right) Inclusion of the immunizing peptide blocked TACO staining. Bar, 5 m.

cortex over the next few hours, resulting in a staining pattern as in noninfected cells (Figure 4b). Quantitation of TACO localization at the phagosome revealed that 80%90% of the mycobacterial phagosomes stained positive for TACO in case of viable bacilli at all time points. In contrast, TACO dissociated from phagosomes containing killed bacilli within the first 2 hr and redistributed to the cell cortex during this time period (Figure 4c). In noninfected macrophages, TACO strongly colocalized with tubulin (Figure 3), and as tubulin has been implicated in phagosomelysosome fusion (Desjardins

Mycobacterial Subversion of Phagocytosis 439

Figure 4. Subcellular Localization of TACO in Infected Cells Macrophages were grown on coverslips and infected with live (a) or heat-killed (b) M. bovis BCG expressing the green fluorescent protein (BCG-GFP) for 1 hr followed by a chase for the times indicated. After fixation and permeabilization, the cells were stained for TACO using mAb GK3 followed by Texas redlabeled secondary antibodies. For each field, a bright field image is shown (left side). Bar, 5 m. (c) For quantitation, cells were scored for the presence of TACO at the mycobacterial phagosome. Shown are mean values ( SD) from three experiments (n 50). In (d), cells were infected with live (left) or heat-killed (right) mycobacteria followed by fixation and staining using antitubulin antibodies and Texas redlabeled secondary antibodies. Bar, 5 m.

Cell 440

Figure 5. Effect of Rifampicin and Alkalinization on TACO Distribution (a) Cells were infected for 1 hr with live mycobacteria, followed by a 2 hr chase in the absence (upper panels) or presence (lower panels) of rifampicin (1 mg/ml), fixed, and stained for TACO as described. Quantitation of TACO recruitment (as in Figure 4) revealed a strong reduction of TACO colocalization with mycobacteria (right side). The difference in TACO recruitment between phagosomes containing heat- and rifampicin-killed mycobacteria might be due to differences in bacterial treatment. (b) Cells were alkalinized as described in Experimental Procedures and subsequently infected with live (upper panels) or heat-killed (lower panels) mycobacteria for 1 hr, followed by a 2 hr chase. After fixation, cells were stained for TACO as described. Quantitation (right side) revealed that whereas 90% of the phagosomes containing viable bacilli had recruited TACO, virtually none of those containing killed bacilli colocalized with TACO. Bar, 5 m.

et al., 1994), the distribution of tubulin in macrophages infected with live as well as with heat-killed mycobacteria was analyzed. As shown in Figure 4d, whereas tubulin remained present at all times around phagosomes containing killed bacilli, in cells infected with live mycobacteria tubulin rapidly dissociated from the mycobacterial phagosome. Besides tubulin, as expected, also LAMP was recruited to the phagosomes containing killed bacilli (results not shown), consistent with a role for microtubules in phagosomelysosome fusion (Desjardins et al., 1994; Blocker et al., 1998). To analyze whether the recruitment and retention of TACO required viability of the mycobacteria, cells infected with live bacteria were incubated for 2 hr in medium containing the anti-mycobacterial drug rifampicin. Whereas in nontreated cells 90% of the mycobacterial phagosomes contained TACO (Figure 5a, upper panels), the in situ killing of the bacteria by rifampicin resulted in a rapid release of TACO from the phagosomal membrane (Figure 5a, lower panels). Retention of TACO on phagosomes may occur as a result of the elevated pH in phagosomes as opposed to lysosomes (Crowle et al., 1991; Sturgill-Koszycki et al., 1994) or, alternatively, be dictated by processes exerted by viable bacteria. To distinguish between these possibilities, phagosomes containing heat-killed mycobacteria formed in the presence of NH4Cl were analyzed. Ammonium chloride inhibits phagosomelysosome fusion (Gordon et al., 1980; Kielian and Cohn, 1982; Hart and Young, 1991) without altering the morphology or

distribution of phagosomes (Heuser, 1989). When organelles from NH4Cl-treated macrophages infected with heat-killed bacteria were separated by organelle electrophoresis, virtually all bacilli were retrieved in unshifted fractions, confirming the ability of NH4Cl to inhibit phagosomelysosome fusion (data not shown). Immunofluorescence analysis of these cells revealed that while ammonium chloride treatment did not affect TACO recruitment on phagosomes containing living mycobacteria (Figure 5b, upper panels), TACO was not retained on phagosomes containing heat-killed bacilli (Figure 5b, lower panels). This further indicates that lack of acidification per se is not responsible for TACO retention but rather processes exerted by the mycobacteria themselves.

Localization of Mycobacteria in TACO-Negative Kupffer Cells In contrast to all other macrophage-containing tissue, the liver did not show any expression of TACO (Figure 3a), despite the presence of large numbers of macrophages, the Kupffer cells (Wardle, 1987). The absence of TACO from Kupffer cells was confirmed by immunoblotting of a lysate prepared from isolated Kupffer cells as well as by immunohistochemistry on liver sections (Figures 6a and 6c). To directly analyze phagocytosis of mycobacteria in TACO-negative Kupffer cells, these cells were isolated from the liver and infected with mycobacteria for 1 hr followed by a 2 hr chase. Cells

Mycobacterial Subversion of Phagocytosis 441

Figure 6. TACO Expression and Mycobacterial Entry in Liver Macrophages (a and b) Localization of mycobacteria in TACO-negative Kupffer cells. (a) Kupffer cells were isolated from liver as described in Experimental Procedures, lysed, and total proteins (10 g) separated by SDSPAGE followed by immunoblotting using anti-TACO antibodies. As a control, a J774 lysate (10 g) was used. (b) Isolated Kupffer cells were incubated for 1 hr with M. bovis BCG expressing the green fluorescent protein (BCG-GFP). Cells were washed and incubated for 2 hr (upper panels) or 16 hr (lower panels) prior to fixation and staining for lysosomal glycoprotein followed by Texas redlabeled secondary antibodies (LAMP-TR). Bar, 5 m. (c and d) Immunohistochemistry of TACO expression in mouse liver. Mice were untreated (c) or infected (d) with M. bovis BCG, sacrificed, and sections were obtained from the liver. Immunohistochemistry was performed using the rat mAb GK3 followed by HRP-coupled secondary antibodies (left panel), or double labeling was performed using mAb GK3 and Ziehl-Neelsen staining.

were subsequently fixed and stained for lysosomes using antibodies against lysosomal-associated membrane glycoprotein (LAMP) and analyzed by confocal immunofluorescence microscopy (Figure 6b, upper panels). All of the infected Kupffer cells observed contained large vesicular structures that strongly stained positive for LAMP (Figure 6b, LAMP-TR). All phagocytosed bacilli were present in lysosomal organelles, and no mycobacteria could be detected in LAMP-negative vacuoles. In fact, in most if not all phagocytosing Kupffer cells, lysosomal vacuoles could be found recruited to and concentrated at the site of bacterial uptake. Interestingly, whereas after a 2 hr chase phagolysosomes could be detected harboring both intact bacilli as well as partially degraded mycobacteria (as judged from the green fluorescent protein distribution pattern), after an overnight chase virtually all bacilli were degraded (Figure 6b, lower

panels). Thus, in macrophages lacking TACO, mycobacteria were rapidly taken up in lysosomes, followed by a complete destruction of the bacilli. Immunohistochemical Analysis of TACO Expression Although the liver Kupffer cells are known to completely eradicate bacterial infections (North, 1974), one consequence of a mycobacterial challenge in vivo can be the formation of liver granulomas. Granulomas are macrophage-rich lesions harboring viable bacilli, thought to be involved in restricting the bacterial spread (Gordon et al., 1994). When liver sections obtained from BCGinfected mice were analyzed, the granuloma regions stained positive for TACO, and these macrophages harbored mycobacteria as revealed by Ziehl-Neelsen staining (Figure 6d). TACO-positive macrophages represented exclusively infiltrated macrophages, as they stained posi-

Cell 442

Figure 7. Differential Trafficking and Survival of Mycobacteria in TACO-Transfected Cells Mel JuSo cells transfected with cDNA encoding TACO in the sense (TACO-transfected) or antisense (Control) orientation were lysed and total proteins (10 g) separated by SDSPAGE followed by immunoblotting using anti-TACO antibodies (a). Control (b) or TACO-transfected (c) cells were incubated with mycobacteria for 1 hr followed by a 3 hr chase, homogenized, and subjected to organelle electrophoresis, and the distribution of the organelle-specific markers as indicated was determined (b and c, upper panels). From each fraction, the amount of mycobacteria was analyzed by direct observation by light microscopy after acid fast staining (b and c, lower panels). Shown is a representative experiment out of three. (d) TACO-transfected or control cells were infected with metabolically labeled mycobacteria for 1 hr at 37 C, washed, and lysed, and the total uptake was determined by liquid scintillation counting of the lysate. (e) TACO-transfected or control cells were incubated with mycobacteria, washed, and chased. At the chase times indicated, cells were lysed in PBS containing 0.5% Triton X-100, and bacteria were collected by centrifugation and resuspended in 7H9 medium. Serial dilutions were plated out and the number of colonies determined from the appropriate dilution. Shown are mean values ( SD) from five experiments (d and e).

tive both for F4/80 and Mac-I antigens, the former staining all macrophage cell types, the latter being not reactive with Kupffer cells (Hirsch et al., 1981; Nibbering et al., 1989). Differential Mycobacterial Trafficking and Survival in Nonphagocytes Expressing TACO The recruitment of TACO on phagosomes containing live mycobacteria together with the resistance of these mycobacterial phagosomes to fuse with or mature into lysosomes suggests that TACO functions as a component of the phagosomal coat that, when present, prevents fusion with lysosomes. To directly analyze this, the nonphagocytic melanoma cell line Mel JuSo, which does not express TACO, was transfected with cDNA encoding TACO, and stably transfected cell lines were selected (Figure 7a). Transfection of TACO did not affect

mycobacterial uptake, which was in fact three orders of magnitude lower when compared to J774 macrophages (Figure 7d and data not shown). When Mel JuSo cells were infected for 1 hr with mycobacteria followed by a 3 hr chase, after homogenization and subcellular fractionation by electrophoresis 15% of the internalized bacilli were located in shifted fractions, indicating a limited ability of these cells to target mycobacteria to late endosomal/lysosomal organelles (Figure 7b). However, in TACO-transfected Mel JuSo cells, transfer of the mycobacteria to late endosomes was completely abrogated, as shown by the absence of mycobacteria from shifted fractions (Figure 7c). In accordance with this, in control Mel JuSo cells a subset of intracellularly residing bacteria colocalized with LAMP ( 10%), whereas no colocalization at all was observed in TACO-transfected cells (data not shown).

Mycobacterial Subversion of Phagocytosis 443

To analyze mycobacterial survival in TACO-transfected versus control Mel JuSo cells, these cells were infected with mycobacteria for 1 hr, washed, and further incubated for the times indicated in Figure 7e. At each time point, cells were lysed and the amount of viable mycobacteria determined by counting the formation of colonies after plating out different dilutions on agar. Strikingly, in TACO-transfected Mel JuSo cells survival of mycobacteria was 3-fold higher within the first 3 hr after infection and increased to 8-fold after an overnight chase (Figure 7e). Thus, also in nonphagocytic but TACO-expressing cells, mycobacteria have the capacity to resist their delivery to late endosomal/lysosomal organelles, thereby prolonging their survival. Discussion A hallmark of the mycobacterial phagosome is its resistance to fusion with other stages of the endosomal/ lysosomal pathway (Armstrong and DArcy Hart, 1971; Clemens and Horwitz, 1995; Russell, 1995). As a result, delivery of mycobacteria to lysosomes is prevented, which is thought to represent a key mechanism by which mycobacteria manage to survive intracellularly. In a search for host cell components that are involved in the intracellular survival of mycobacteria, we have identified TACO, a protein associated with phagosomes containing viable but not killed mycobacteria. Retention of TACO on the phagosomal membrane was strictly dependent on mycobacterial viability and not due to impaired acidification. Moreover, in TACO-negative Kupffer cells all bacilli were located in lysosomes followed by their degradation. Expression of TACO in a nonmacrophage cell line was sufficient to prevent lysosomal fusion of the mycobacterial phagosome, allowing survival of mycobacteria. Together these results suggest that TACO is actively retained on mycobacterial phagosomes by processes exerted by the mycobacteria themselves and is one of the factors providing a niche for mycobacteria to survive within macrophages. TACO Is a WD Repeat Protein The predicted amino acid sequence of TACO showed that it is a member of the WD repeat protein family (Neer et al., 1994; Suzuki et al., 1995). WD repeatcontaining proteins are involved in a variety of processes, including signal transduction, motility (King et al., 1995), cytokinesis (Shirayama et al., 1998), cytoskeletal organization, and vesicle fusion (Pryer et al., 1993; Salama et al., 1993). Members of this family are present throughout eukaryotic evolution and thus far have not been found to be present in prokaryotic organisms (Neer et al., 1994). Although the precise function of the WD repeat remains unknown, the almost complete colocalization of TACO with tubulin in noninfected macrophages suggests that at least some part of TACO either directly or indirectly binds to microtubules. Given the role of several, otherwise distinct WD repeat proteins in processes requiring microtubule-based activity (Li and Suprenant, 1994; Wilkerson et al., 1995), it is tempting to speculate that the WD repeat represents a microtubule-binding domain. Microtubule binding might be regulated by posttranslational modifications, and the predicted protein kinase
Figure 8. Model Describing the Role of TACO in Intracellular Survival of Mycobacteria Phagocytosis triggers the recruitment of TACO (yellow) around the particle to be ingested, possibly as a result of its initial association with microtubules (black lines) at the cell cortex. In the case of viable bacteria (violet), TACO is actively retained on the phagosomal membrane, while it dissociates from the microtubule network, perhaps as a consequence of a conformational change. In contrast, during phagocytosis of killed bacteria (black) or after in situ killing of the bacilli, TACO initially associates with phagosomes but is rapidly released concomitant with lysosomal delivery of the bacilli.

C phosphorylation consensus sites present in TACO are possibly involved in such regulation. Most studies to date have implicated the actin cytoskeleton in phagocytosis of a wide variety of living and inert particles (Greenberg et al., 1990; Rathman et al., 1997). In agreement with these reports, we found that phagosomes containing viable as well as those containing dead bacilli were associated with actin. In contrast to actin, tubulin remained associated exclusively with phagosomes containing killed bacteria, consistent with previous reports showing that fusion of phagosomes with lysosomes is a microtubule-dependent process (Desjardins et al., 1994; Blocker et al., 1998). Both the actin and tubulin cytoskeleton may play a role during the initial phase of bacterial internalization to generate the forces required for the actual internalization process. However, after the internalization of living mycobacteria, tubulin rapidly dissociated from TACOcoated phagosomes containing viable mycobacteria. Interestingly, TACO shows 35% homology to a WD repeat, actin-binding protein identified in the slime mold Dictyostelium discoideum, termed coronin (de Hostos et al., 1991; de Hostos et al., 1993). Deletion of the gene encoding this protein in Dictyostelium results in an impaired functioning of actin-based processes, one of which leads to a reduced capacity to internalize yeast particles (Maniak et al., 1995). This indicates that in this organism, one function of this protein is associated with particle uptake (Maniak et al., 1995; Rauchenberger et al., 1997). Cellular slime molds such as Dictyostelium live on soil bacteria, including mycobacteria, and it is conceivable that during the 15002000 million years that separate Dictyostelium from mammals (Knoll, 1992), mycobacteria have acquired the ability to actively recruit

Cell 444

and retain host cell components, including TACO, that ensure their survival, rather than being delivered to lysosomes for degradation (see also Figure 8). Retention of TACO by Mycobacteria to Prevent Lysosomal Delivery The distribution of TACO during phagocytosis suggests that it represents a coat protein of the mycobacterial phagosome. Vesicular coat proteins are transiently recruited onto membranes and required in a defined budding or fusion step or during organelle maturation to transfer cargo from one stage to another (Rothman and Wieland, 1996; Schekman and Orci, 1996). Thus far, three types of coats have been defined to function in distinct steps of vesicular traffic. Clathrin-based coats function in cargo delivery to the endocytic pathway as well as to post-Golgi organelles (Kirchhausen et al., 1997), while COPI- (for coat protein I) and COPII-based coats are involved in traffic between the endoplasmic reticulum and Golgi organelles (Bednarek et al., 1995; Cosson and Letourneur, 1997). Instead of functioning in cargo selection and/or budding processes, TACO might in fact prevent the budding and fusion machinery to be recruited onto the plasma membrane and therefore identify this membrane as nonfusogenic with lysosomes. Once material is being phagocytosed, release of TACO would be required to trigger lysosomal fusion. It is precisely at this last step in the normal maturation pathway of phagocytosed material that mycobacteria intervene. To avoid lysosomal delivery, mycobacteria have gained the capacity to actively retain TACO on the phagosomal membrane, thereby possibly preventing the recruitment of the machinery necessary for cargo delivery to lysosomes (Figure 8). As transfection of TACO in nonphagocytic cells also inhibits mycobacterial delivery to late endosomes, this function of TACO appears to be independent from the phagocytic machinery with which macrophages are equipped. Interestingly, another WD repeat protein, the beige/ lyst gene product, might have a very similar scaffolding function for lysosomes (Perou et al., 1996). Also this protein colocalizes with microtubules (Faigle et al., 1998), and cells lacking an intact beige/lyst product are characterized by the presence of abnormally large lysosomes whose function is impaired (Burkhardt et al., 1993). In fact, one function of the WD repeat proteins might be to provide a scaffold for the different organelles to maintain their integrity. Recruitment and retention of TACO on the phagosomal membrane might be regulated through transmembrane signal transduction processes induced by the living bacteria. How mycobacteria achieve this remains to be established, and it is possible that phagosomal transmembrane proteins are involved in the regulation of TACO retention. Defining the binding partners of TACO at the phagosomal membrane as well as in the cytosol might identify the complexes involved in such signal transduction processes. In human neutrophiles, a TACO homolog was found to copurify with the p40 protein thought to play a role in the NADPH oxidase (Grogan et al., 1997). However, the contribution of p40 in the NADPH oxidase is unclear (Tsunawaki et al., 1996;

Sathyamoorthy et al., 1997), and copurification may have resulted from the colocalization of both proteins at the cortical cytoskeleton (Woodman et al., 1991). A Balance between the Host and the Pathogen Evidence for a role for TACO in vivo in the containment of mycobacteria in a viable state was provided by the finding that infiltrated macrophages in liver granulomas expressed TACO, whereas the resident liver macrophages, the Kupffer cells, did not. The liver is the prime target for infectious agents that have accessed an organism, but as an organ it is rarely infected (Brandborg and Goldman, 1990). Clearance of bacteria in the liver is attributed to the presence of large amounts of Kupffer cells (North, 1974; Wardle, 1987). Indeed, when isolated Kupffer cells were infected with mycobacteria, all of the bacilli were rapidly transferred to lysosomes and degraded. First, this indicates that TACO is not absolutely essential for phagocytosis. Second, these results suggest that at the level of the whole organism rather than at the single cell, evolutionary pressure may have ensured downregulation of TACO expression in Kupffer cells to provide an efficient and complete mycobacterial clearance site. Why have mammalian cells retained a molecule that allows mycobacteria to survive within macrophages? One possibility is that TACO performs another, as yet unknown function in macrophages. For example, as the processes involved in cell locomotion and phagocytosis are quite similar (Berlin et al., 1978; Valerius et al., 1981), TACO may be a crucial component of the cytoskeleton of highly motile cells, functioning both in the invagination of large pieces of plasma membrane, as well as in forming protrusions of the plasma membrane involved in cell locomotion (Valerius et al., 1981). The homology of TACO with Dictyostelium coronin would indeed argue in favor of such a role. The fact that mycobacteria, in the course of their coevolution with higher eukaryotes, have made use of the normal cellular processes involved in their own uptake to circumvent host defense responses is yet another striking example of the ingenuity of pathogenic microorganisms.
Experimental Procedures Cells, Bacteria, and Antibodies The murine macrophage cell line J774A.1 was obtained from ATCC and cultured in DMEM containing 10% heat-inactivated fetal calf serum at 37 C in 5% CO2. The human melanoma cell line Mel JuSo (Johnson et al., 1981) was grown in RPMI-1640 containing 5% heatinactivated fetal calf serum. Mice (BALB/c) were obtained from IffaCredo (Basel, Switzerland) or from Charles River, Inc. (Tokyo, Japan). Kupffer cells were isolated according to Lee et al. (1986) with the following modifications. Mice were sacrificed and the thoracic cavity exposed followed by canulation of the inferior vena cava via the right atrium. After opening of the portal vein to allow efflux, the liver was perfused with 10 ml of PBS followed by perfusion of a collagenase solution (5 mg/ml collagenase D [Boehringer Mannheim] in RPMI-1640, prewarmed at 37 C). The liver was removed and mechanically disrupted prior to incubation with 10 ml of 1 mg/ ml of the same collagenase solution at 37 C. After 15 min, DNAse (Sigma) from a 5 mg/ml stock solution in PBS was added to a final concentration of 0.5 mg/ml. The resulting cell suspension was filtered through dressing sponges (Johnson and Johnson, Arlington, TX), and the filtrate was diluted to 50 ml in RPMI containing 10% heat-inactivated FCS. Cells were centrifuged five times (30 g, 4

Mycobacterial Subversion of Phagocytosis 445

min) at room temperature, and the pellets were discarded. After two final washes in RPMI-FCS, cells were plated out, left to adhere for 60 min at 37 C, after which the nonadherent cells were removed by five washes in RPMI. The resulting cell population consisted of 90% macrophages as revealed by F4/80 staining. Mycobacterium bovis BCG, strain Montreal was kindly provided by Dr. Thole (TNO, Leiden), and M. bovis BCG harboring phsp60gfp (BCG-GFP) (Dhandayuthapani et al., 1995) was kindly provided by Dr. Deretic (University of Michigan, Ann Arbor, MI) and propagated in 7H9 mycobacterial medium (Difco) supplemented with 10% OADC Middlebrook supplement (Difco). Killing of mycobacteria was carried out by heat treatment for 30 min at 80 C. Infection of macrophages was essentially as described (Hasan et al., 1997). Quantitation of the total mycobacterial uptake was performed by incubating cells for 1 hr with 35S-methionine/cysteine-labeled mycobacteria. After extensive washing, cells were lysed in Dulbeccos phosphate-buffered saline (D-PBS) containing 1% Triton X-100 and analyzed by liquid scintillation counting. Mycobacterial survival was analyzed by lysing cells in D-PBS containing 0.5% Triton X-100 and sedimentation of the bacteria at 2700 g for 10 min. The pellet was resuspended in 7H9 medium, and different dilutions were plated out on 7H11 agar (Difco), supplemented with 10% OADC Middlebrook and glycerol (5 ml/l). After 3 weeks, the number of colonies formed was determined from the appropriate dilutions. In vivo infection was performed by injecting 8-week-old BALB/c mice intravenously with 0.2 ml of M. bovis BCG (Japan BCG Inc., Tokyo, Japan), dissolved in physiological saline at a concentration of 10 mg/ml. The liver was removed at various intervals after BCG injection and processed for immunohistochemistry. Polyclonal antisera were raised against KLH-coupled peptides in New Zealand white rabbits, and antibodies were affinity purified by affinity chromatography using the immunizing peptides as ligands. Monoclonal antibodies were generated by immunizing rats (Wistar) with the same peptides and fusion of the spleen and lymph nodes with SP2/0 myeloma cells. Antibodies against actin (clone C4; mouse IgG1) and -tubulin (mouse IgG1) were obtained from Boehringer Mannheim and Amersham, respectively. Rat anti-Mac-1 antibody (IgG2b) was from CALTAG (San Francisco, CA), and the rat mAb F4/80 (IgG2b) was kindly provided by Dr. S. Gordon (Oxford, UK). The rat anti-murine LAMP-1 antibody 1D4B (IgG2a) was developed by T. August and obtained from the Developmental Studies Hybridoma Bank at the University of Iowa. Alkalinization was performed as described in Heuser (1989). Briefly, cells were washed three times in Ringers solution without NH4Cl, followed by equilibration of the cells (two times for 15 min each) in Ringers solution containing 30 mM NH4Cl (alkalinization medium). Bacteria were washed in the same alkalinization medium, and the infection was performed as described above. Subcellular Fractionation and Phagosome Isolation Cells were homogenized and organelles prepared essentially as described (Tulp et al., 1994; Ferrari et al., 1997; Hasan et al., 1997). The membrane pellet was resuspended in 6% Ficoll-70 in homogenization buffer, and organelle electrophoresis as well as the determination of organelle specific markers was carried out as described (Ferrari et al., 1997). The presence of mycobacteria was determined by centrifugation of the fractions onto glass coverslips followed by acid fast staining (Difco) and quantitation by direct observation of the coverslips. Phagosomes were isolated after pooling of the bacteria-containing fractions followed by sedimentation at 100,000 g. The membrane pellet was resuspended in homogenization buffer, and phagosomes were collected by low-speed centrifugation (30 min, 380 g). Protein Analysis and Mass Spectrometry Proteins were analyzed using one-dimensional (10%) SDSPAGE according to Laemmli (1970) or two-dimensional IEF/SDSPAGE as described (Lefkovits et al., 1985). For mass spectrometry, twodimensional gels were fixed and stained using colloidal Coomassie (Novex, San Diego CA). The protein spot of interest was excised, digested in situ, and characterized by matrix-assisted laser desorption ionization mass spectrometry (Fountoulakis and Langen, 1997).

In short, excised spots were destained with 30% acetonitrile in 0.1 M ammonium bicarbonate and washed with water. The gel pieces were reswollen with 5 l of 3 mM Tris-HCl (pH 8.8), containing 1 g of endopeptidase Lys-C (Wako) after drying in a speedvac evaporator. The digestion was performed for 312 hr at room temperature. Three microliters of 30% acetonitrile containing 0.1% trifluoroacetic acid was added, and the content was vortexed, centrifuged for 2 min, and sonicated for 5 min. One microliter from the separated liquid was applied onto the dried matrix spot. The matrix consisted of 15 mg of nitrocellulose (Bio-Rad) and 20 mg of -cyano 4-hydroxycinnamic acid (Sigma) dissolved in 1 ml of acetone isopropanol 1:1 (v/v). Matrix solution (0.5 l) was applied on the sample target. The samples were analyzed in a time-of-flight Perseptive Biosystems mass spectrometer (Voyager Elite) equipped with a reflector and delayed extraction. An acceleration voltage of 20 kV was used. Calibration was internal to the samples with des-Arg-1 Bradykinin (Sigma) and adrenocorticotropic hormone fragment 18-39 (Sigma). For the protein search, monoisotopic peptide masses were used, and a mass tolerance of 0.006% was allowed. cDNA Cloning and Sequencing To obtain the cDNA encoding TACO, a mouse macrophage ZAP II library (from peritoneal macrophages; Stratagene) was screened with a probe amplified from human spleen cDNA (Clonetech) using the following primers: 5 -CCAGTGCTATGAAGATGTGCGCG-3 (forward primer) and 5 -ATGATGCGCACGCGCTTGTCACGGC-3 (reverse primer). Positive clones were expanded, and the pBluescript phagemid containing the cloned DNA inserts was excised by coinfection with helper phage in XL1-Blue MRF to yield TACOpBL-SK. Sequencing was performed using the ABI PRISM Dye terminator cycle sequencing reaction (Perkin Elmer, Foster City, CA) and analyzed using an ABI 373 DNA sequencer. For the construction of a eukaryotic expression construct, a TACO-encoding fragment was subcloned into pBL-SK after digestion of TACOpBL-SK with XhoI and XbaI, and the purified fragment ligated into pBK-CMV (Stratagene) digested with the same enzymes. A TACO-encoding fragment was generated using EcoRI and ligated into the mammalian expression vector pUC-CAGGS (Niiwa et al., 1991) either in the antisense or sense orientation. cDNAs were introduced into Mel JuSo cells by electroporation using the BioRad Gene Pulser (400 V/125 F), and clones were selected using G418. Metabolic Labeling and Immunoprecipitation Prior to labeling, macrophages were grown for 30 min in methionine/ cysteine-free medium containing 10% dialyzed FCS. Labeling was performed by placing the cells in methionine/cysteine-free medium containing 10% dialyzed FCS and 0.15 mCi/ml 35S-methionine/cysteine mix (Promix, Amersham). Mycobacteria were labeled in 7H9 medium supplemented with 10% OADC Middlebrook containing 5 Ci/ml 35S-methionine/cysteine mix (Promix). Immunoprecipitations were performed as described (Pieters et al., 1991). Immunofluorescence Microscopy and Immunohistochemistry Cells were grown on glass coverslips and infected as described above. After incubation, cells were fixed in 3% paraformaldehyde (at 37 C) for 10 min, washed in PBS, and permeabilized in 0.1% saponin in PBS for 30 min at room temperature. Saponin (0.1%) was included in all subsequent incubations. Alternatively, cells were fixed in methanol (4 min at 20 C). Coverslips were incubated with primary antibodies for 40 min at room temperature, washed in PBS containing 2% BSA, and then incubated with FITC and Texas red coupled goat anti-rabbit and/or goat anti-mouse antibodies (Southern Biotechnology Associates, Birmingham, AL). Actin was visualized using phalloidin-Texas red (Molecular Probes, Leiden, The Netherlands). Coverslips were mounted in FluoroGuard Antifade Mounting Reagent (Bio Rad) and analyzed using a confocal laser scanning microscope system (MRC-600; Bio Rad laboratories, Hercules, CA) attached to an Axiovert 35M microscope (Carl Zeiss, Inc., Thornwood, NY). For each condition as well as for quantitation, 150 bacteria-containing cells were analyzed, and a representative field is shown. For immunohistochemistry, the liver was fixed for 4 hr at 4 C in

Cell 446

periodate-lysine-paraformaldehyde, washed for 4 hr with phosphate-buffered saline (PBS) containing 10%, 15%, and 20% sucrose, embedded in OCT compound (Miles, Elkhart, IN), frozen in dry ice-acetone, and cut by a cryostat (Bright, Huntington, UK) into 6 m thick sections. After inhibition of endogenous peroxidase activity as described by Isobe et al. (1977), immunohistochemistry was performed using the antibodies described above, with anti-rat Ig or sheep anti-rabbit Ig-horseradish peroxidase-linked F(ab)2 fragment (Amersham, Little Chalfont, UK) as secondary antibodies. Antibodies were visualized with 3,3 -diaminobenzidine (DAB; Dojin Chemical, Kumamoto, Japan), and nuclei were stained with methylene green or hematoxyline. Acknowledgments We are grateful to L. Kuhn and I. Lefkovits for support in the twodimensional gel analysis and M. Dessing for his help in confocal microscopy. We thank T. Miyazaki for his advice on cloning and V. Deretic, Z. Hasan, and J. Thole for discussion as well as bacterial strains. We thank T. Miyazaki and M. Kopf for discussion and critical evaluation of the manuscript. The Basel Institute for Immunology was founded and is supported by F. Hoffmann-La Roche and Co., Ltd., Basel, Switzerland. Received November 25, 1998; revised April 7, 1999. References Armstrong, J., and DArcy Hart, P. (1971). Response of cultured macrophages to Mycobacterium tuberculosis, with observations on fusion of lysosomes with phagosomes. J. Exp. Med. 134, 713740. Barker, L.P., George, K.M., Falkow, S., and Small, P.L. (1997). Differential trafficking of live and dead Mycobacterium marinum organisms in macrophages. Infect. Immun. 65, 14971504. Bednarek, S.Y., Ravazzola, M., Hosobuchi, M., Amherdt, M., Perrelet, A., Schekman, R., and Orci, L. (1995). COPI- and COPII-coated vesicles bud directly from the endoplasmic reticulum in yeast. Cell 83, 11831196. Berger, B., Wilson, D.B., Wolf, E., Tonchev, T., Milla, M., and Kim, P.S. (1995). Predicting coiled coils by use of pairwise residue correlations. Proc. Natl. Acad. Sci. USA 92, 82598263. Berlin, R.D., Oliver, J.M., and Walter, R.J. (1978). Surface functions during mitosis I: phagocytosis, pinocytosis and mobility of surfacebound Con A. Cell 15, 327341. Beverley, S.M. (1996). Hijacking the cell: parasites in the drivers seat. Cell 87, 787789. Bliska, J.B., Galan, J.E., and Falkow, S. (1993). Signal transduction in the mammalian cell during bacterial attachment and entry. Cell 73, 903920. Blocker, A., Griffiths, G., Olivo, J.C., Hyman, A.A., and Severin, F.F. (1998). A role for microtubule dynamics in phagosome movement. J. Cell Sci. 111, 303312. Brandborg, L., and Goldman, I.S. (1990). Bacterial and miscellaneous infections of the liver. In Hepatology: A Textbook of Liver Disease, D. Zakim and T.D. Boyer, eds. (Philadelphia: W.B. Saunders), pp. 10861098. Burkhardt, J.K., Wiebel, F.A., Hester, S., and Argon, Y. (1993). The giant organelles in beige and Chediak-Higashi fibroblasts are derived from late endosomes and mature lysosomes. J. Exp. Med. 178, 18451856. Clemens, D.L., and Horwitz, M.A. (1995). Characterization of the Mycobacterium tuberculosis phagosome and evidence that phagosomal maturation is inhibited. J. Exp. Med. 181, 257270. Cosson, P., and Letourneur, F. (1997). Coatomer (COPI)-coated vesicles: role in intracellular transport and protein sorting. Curr. Opin. Cell Biol. 9, 484487. Crowle, A.J., Dahl, R., Ross, E., and May, M.H. (1991). Evidence that vesicles containing living, virulent Mycobacterium tuberculosis or Mycobacterium avium in cultured human macrophages are not acidic. Infect. Immun. 59, 18231831.

de Hostos, E.L., Bradtke, B., Lottspeich, F., Guggenheim, R., and Gerisch, G. (1991). Coronin, an actin binding protein of Dictyostelium discoideum localized to cell surface projections, has sequence similarities to G protein beta subunits. EMBO J. 10, 40974104. de Hostos, E.L., Rehfuess, C., Bradtke, B., Waddell, D.R., Albrecht, R., Murphy, J., and Gerisch, G. (1993). Dictyostelium mutants lacking the cytoskeletal protein coronin are defective in cytokinesis and cell motility. J. Cell Biol. 120, 163173. Desjardins, M., Huber, L.A., Parton, R.G., and Griffiths, G. (1994). Biogenesis of phagolysosomes proceeds through a sequential series of interactions with the endocytic apparatus. J. Cell Biol. 124, 677688. Dhandayuthapani, S., Via, L.E., Thomas, C.A., Horowitz, P.M., Deretic, D., and Deretic, V. (1995). Green fluorescent protein as a marker for gene expression and cell biology of mycobacterial interactions with macrophages. Mol. Microbiol. 17, 901912. Faigle, W., Raposo, G., Tenza, D., Pinet, V., Vogt, A.B., Kropshofer, H., Fischer, A., de Saint-Basile, G., and Amigorena, S. (1998). Deficient peptide loading and MHC class II endosomal sorting in a human genetic immunodeficiency disease: the Chediak-Higashi syndrome. J. Cell Biol. 141, 11211134. Falkow, S. (1991). Bacterial entry into eukaryotic cells. Cell 65, 1099 1102. Ferrari, G., Knight, A.M., Watts, C., and Pieters, J. (1997). Distinct intracellular compartments involved in invariant chain degradation and antigenic peptide loading of major histocompatibility complex (MHC) class II molecules. J. Cell Biol. 139, 14331446. Finlay, B.B., and Cossart, P. (1997). Exploitation of mammalian host cell functions by bacterial pathogens. Science 276, 718725. Fountoulakis, M., and Langen, H. (1997). Identification of proteins by matrix-assisted laser desorption ionization-mass spectrometry following in-gel digestion in low-salt, nonvolatile buffer and simplified peptide recovery. Anal. Biochem. 250, 153156. Galan, J.E., and Bliska, J.B. (1996). Cross-talk between bacterial pathogens and their host cells. Annu. Rev. Cell Dev. Biol. 12, 221255. Gordon, A.H., Hart, P.D., and Young, M.R. (1980). Ammonia inhibits phagosome-lysosome fusion in macrophages. Nature 286, 7980. Gordon, S., Keshav, S., and Stein, M. (1994). BCG-induced granuloma formation in murine tissues. Immunobiology 191, 369377. Greenberg, S., Burridge, K., and Silverstein, S.C. (1990). Colocalization of F-actin and talin during Fc receptor-mediated phagocytosis in mouse macrophages. J. Exp. Med. 172, 18531856. Grogan, A., Reeves, E., Keep, N., Wientjes, F., Totty, N.F., Burlingame, A.L., Hsuan, J.J., and Segal, A.W. (1997). Cytosolic phox proteins interact with and regulate the assembly of coronin in neutrophils. J. Cell Sci. 110, 30713081. Hart, P.D., and Young, M.R. (1991). Ammonium chloride, an inhibitor of phagosome-lysosome fusion in macrophages, concurrently induces phagosome-endosome fusion, and opens a novel pathway: studies of a pathogenic mycobacterium and a nonpathogenic yeast. J. Exp. Med. 174, 881889. Hasan, Z., Schlax, C., Kuhn, L., Lefkovits, I., Young, D., Thole, J., and Pieters, J. (1997). Isolation and characterization of the mycobacterial phagosome: segregation from the endosomal/lysosomal pathway. Mol. Microbiol. 24, 545553. Heuser, J. (1989). Changes in lysosome shape and distribution correlated with changes in cytoplasmic pH. J. Cell Biol. 108, 855864. Hirsch, S., Austyn, J.M., and Gordon, S. (1981). Expression of the macrophage-specific antigen F4/80 during differentiation of mouse bone marrow cells in culture. J. Exp. Med. 154, 713725. Isobe, S., Chen, S.-T., Nakane, P.K., and Brown, W.R. (1977). Studies on translocation of immunoglobulins across intestinal epithelium. I. improvements to study the peroxidase-labeled antibody method for application to study of human intestinal. Acta Histochem. Cytochem. 10, 161171. Johnson, J.P., Demmer-Dieckmann, M., Meo, T., Hadam, M.R., and Riethmuller, G. (1981). Surface antigens of human melanoma cells defined by monoclonal antibodies. I. biochemical characterization of two antigens found on cell lines and fresh tumors of diverse tissue origin. Eur. J. Immunol. 11, 825831.

Mycobacterial Subversion of Phagocytosis 447

Kielian, M.C., and Cohn, Z.A. (1982). Intralysosomal accumulation of polyanions. II. polyanion internalization and its influence on lysosomal pH and membrane fluidity. J. Cell Biol. 93, 875882. King, S.M., Patel-King, R.S., Wilkerson, C.G., and Witman, G.B. (1995). The 78,000-M(r) intermediate chain of Chlamydomonas outer arm dynein is a microtubule-binding protein. J. Cell Biol. 131, 399409. Kirchhausen, T., Bonifacino, J.S., and Riezman, H. (1997). Linking cargo to vesicle formation: receptor tail interactions with coat proteins. Curr. Opin. Cell Biol. 9, 488495. Knoll, A. (1992). The early evolution of eukaryotes: a geological perspective. Science 256, 622627. Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685. Lee, S.H., Crocker, P., and Gordon, S. (1986). Macrophage plasma membrane and secretory properties in murine malaria. Effects of Plasmodium yoelii blood-stage infection on macrophages in liver, spleen, and blood. J. Exp. Med. 163, 5474. Lefkovits, I., Young, P., Kuhn, L., Kettman, J., Gemmell, A., Tollaksen, S., Anderson, L., and Anderson, N. (1985). Use of large-scale two-dimensional ISODALT gel electrophoresis systems in immunology. In Immunolgical Methods III, I. Lefkovits and B. Pernis, eds. (Orlando, FL.: Academic Press), pp. 163. Li, Q., and Suprenant, K.A. (1994). Molecular characterization of the 77-kDa echinoderm microtubule-associated protein. Homology to the -transducin family. J. Biol. Chem. 269, 3177731784. Maniak, M., Rauchenberger, R., Albrecht, R., Murphy, J., and Gerisch, G. (1995). Coronin involved in phagocytosis: dynamics of particle-induced relocalization visualized by a green fluorescent protein tag. Cell 83, 915924. Mellman, I., Fuchs, R., and Helenius, A. (1985). Acidification of endocytic and exocytic pathways. Annu. Rev. Biochem. 55, 663700. Neer, E.J., Schmidt, C.J., Nambudripad, R., and Smith, T.F. (1994). The ancient regulatory-protein family of WD-repeat proteins. Nature 371, 297300. Nibbering, P.H., van der Heide, G.A., and van Furth, R. (1989). Immunocytochemical analysis of cellular responses to BCG. Clin. Exp. Immunol. 75, 147154. Niiwa, H., Yamamura, K., and Miyazaki, T. (1991). Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108, 193199. North, R.J. (1974). T cell dependence of macrophage activation and mobilization during infection with Mycobacterium tuberculosis. Infect. Immun. 10, 6671. Perou, C.M., Moore, K.J., Nagle, D.L., Misumi, D.J., Woolf, E.A., McGrail, S.H., Holmgren, L., Brody, T.H., Dussault, B.J., Jr., Monroe, C.A., et al. (1996). Identification of the murine beige gene by YAC complementation and positional cloning. Nat. Genet. 13, 303308. Pieters, J., Horstmann, H., Bakke, O., Griffiths, G., and Lipp, J. (1991). Intracellular transport and localization of major histocompatibility complex class II molecules and associated invariant chain. J. Cell Biol. 115, 12131223. Pryer, N.K., Salama, N.R., Schekman, R., and Kaiser, C.A. (1993). Cytosolic Sec13p complex is required for vesicle formation from the endoplasmic reticulum in vitro. J. Cell Biol. 120, 865875. Rathman, M., Barker, L.P., and Falkow, S. (1997). The unique trafficking pattern of Salmonella typhimurium-containing phagosomes in murine macrophages is independent of the mechanism of bacterial entry. Infect. Immun. 65, 14751485. Rauchenberger, R., Hacker, U., Murphy, J., Niewohner, J., and Maniak, M. (1997). Coronin and vacuolin identify consecutive stages of a late, actin-coated endocytic compartment in Dictyostelium. Curr. Biol. 7, 215218. Rothman, J.E., and Wieland, F.T. (1996). Protein sorting by transport vesicles. Science 272, 227234. Russell, D.G. (1995). Mycobacterium and Leishmania: stowaways in the endosomal network. Trends Cell Biol. 5, 125128. Salama, N.R., Yeung, T., and Schekman, R.W. (1993). The Sec13p complex and reconstitution of vesicle budding from the ER with purified cytosolic proteins. EMBO J. 12, 40734082.

Sathyamoorthy, M., de Mendez, I., Adams, A.G., and Leto, T.L. (1997). p40(phox) down-regulates NADPH oxidase activity through interactions with its SH3 domain. J. Biol. Chem. 272, 91419146. Schekman, R., and Orci, L. (1996). Coat proteins and vesicle budding. Science 271, 15261533. Shirayama, M., Zachariae, W., Ciosk, R., and Nasmyth, K. (1998). The Polo-like kinase Cdc5p and the WD-repeat protein Cdc20p/fizzy are regulators and substrates of the anaphase promoting complex in Saccharomyces cerevisiae. EMBO J. 17, 13361349. Small, P.L.C., Ramakrishnan, L., and Falkow, S. (1994). Remodeling schemes of intracellular pathogens. Science 263, 637639. Sturgill-Koszycki, S., Schlesinger, P.H., Chakraborty, P., Haddix, P.L., Collins, H.L., Fok, A.K., Allen, R.D., Gluck, S.L., Heuser, J., and Russell, D.G. (1994). Lack of acidification in Mycobacterium phagosomes produced by exclusion of the vesicular proton-ATPase. Science 263, 678681. Suzuki, K., Nishihata, J., Arai, Y., Honma, N., Yamamoto, K., Irimura, T., and Toyoshima, S. (1995). Molecular cloning of a novel actinbinding protein, p57, with a WD repeat and a leucine zipper motif. FEBS Lett. 364, 283288. Tsunawaki, S., Kagara, S., Yoshikawa, K., Yoshida, L.S., Kuratsuji, T., and Namiki, H. (1996). Involvement of p40phox in activation of phagocyte NADPH oxidase through association of its carboxyl-terminal, but not its amino-terminal, with p67phox. J. Exp. Med. 184, 893902. Tulp, A., Verwoerd, D., Dobberstein, B., Ploegh, H.L., and Pieters, J. (1994). Isolation and characterization of the intracellular MHC class II compartment. Nature 369, 120126. Valerius, N.H., Stendahl, O., Hartwig, J.H., and Stossel, T.P. (1981). Distribution of actin-binding protein and myosin in polymorphonuclear leukocytes during locomotion and phagocytosis. Cell 24, 195202. Wardle, E.N. (1987). Kupffer cells and their function. Liver 7, 6375. Wilkerson, C.G., King, S.M., Koutoulis, A., Pazour, G.J., and Witman, G.B. (1995). The 78,000 M(r) intermediate chain of Chlamydomonas outer arm dynein is a WD-repeat protein required for arm assembly. J. Cell Biol. 129, 169178. Woodman, R.C., Ruedi, J.M., Jesaitis, A.J., Okamura, N., Quinn, M.T., Smith, R.M., Curnutte, J.T., and Babior, B.M. (1991). Respiratory burst oxidase and three of four oxidase-related polypeptides are associated with the cytoskeleton of human neutrophils. J. Clin. Invest. 87, 13451351. GenBank Accession Number The sequence reported in this paper has been deposited under accession number AF047388.