Essential Role for Cholesterol in Entry of Mycobacteria into Macrophages

See allHide authors and affiliations

Science  02 Jun 2000:
Vol. 288, Issue 5471, pp. 1647-1651
DOI: 10.1126/science.288.5471.1647


Mycobacteria are intracellular pathogens that can invade and survive within host macrophages, thereby creating a major health problem worldwide. The molecular mechanisms involved in mycobacterial entry are still poorly characterized. Here we report that cholesterol is essential for uptake of mycobacteria by macrophages. Cholesterol accumulated at the site of mycobacterial entry, and depleting plasma membrane cholesterol specifically inhibited mycobacterial uptake. Cholesterol also mediated the phagosomal association of TACO, a coat protein that prevents degradation of mycobacteria in lysosomes. Thus, by entering host cells at cholesterol-rich domains of the plasma membrane, mycobacteria may ensure their subsequent intracellular survival in TACO-coated phagosomes.

Macrophages play an important role in the host defense against bacteria because of their capacity to phagocytose and to degrade microbes in lysosomes. Mycobacteria, which are highly successful pathogens, resist delivery to lysosomes and instead survive within a specialized vacuole, the mycobacterial phagosome (1–3). The bacteria survive intracellularly because they are able to actively recruit and retain TACO (for tryptophane aspartate–containing coat protein) at the mycobacterial phagosome, where it prevents fusion of phagosomes with lysosomes (4). TACO lacks a putative transmembrane domain, which suggests that it is a peripherally associated protein. To analyze the association of TACO with the phagosomal membrane, murine macrophages were infected with Mycobacteria bovis Bacille Calmette-Guérin (BCG). Subjection of isolated phagosomes to conditions known to release membrane-associated proteins, such as high salt (1 M NaCl), sodium carbonate (pH 11), and Triton X-100 (1%), did not solubilize TACO (Fig. 1A), although some background release was apparent because of the fragility of isolated phagosomes. However, extraction of the phagosomes in 0.02% digitonin, a cholesterol-sequestering reagent (5), resulted in complete solubilization of TACO, to the same degree as disruption of the phagosomes in SDS or urea (Fig. 1A). Similarly, digitonin, but not Triton X-100, disrupted TACO association with membranes from uninfected cells (Fig. 1C). Furthermore, in cells depleted of cholesterol (6), TACO was not associated with the membrane fraction (Fig. 1D). TACO was completely solubilized after Triton X-100 treatment at 25°C (Fig. 1E), which suggests that incomplete extraction was not because of cytoskeletal association (7). Thus, TACO is associated with membranes in a cholesterol-dependent manner.

Figure 1

Association of TACO with the (A) phagosomal and (C through E) plasma membrane. Phagosomes (A) or membranes (C through E) were isolated from J774A.1 macrophages as described (4,24, 25) and incubated in extraction buffer (20 mM Hepes, pH 7, 5 mM MgCl2, and 100 mM NaCl) supplemented as indicated for 30 min at 4°C (A, C, and D) or 25°C (E). After sedimentation at 20,000g for 15 min, pellets (P) and supernatants (S) were analyzed for the presence of TACO by SDS–polyacrylamide gel electrophoresis (10%) and immunoblotting. (B) Distribution of cholesterol during mycobacterial uptake. J774A.1 macrophages were incubated with M. bovis BCG expressing GFP (26) for 1 hour, washed, and chased for an additional 2 hours. Cholesterol was localized after fixation of the cells in 3% paraformaldehyde at 37°C for 10 min followed by incubation in 0.05% (w/v) filipin (blue) (Sigma) (10) in PBS for 4 hours at room temperature. Scale bar, 10 μm.

To analyze the distribution of cholesterol during phagocytosis, macrophages that had taken up M. bovis BCG expressing the green fluorescent protein (GFP) (8) for 1 hour, followed by a 2.5-hour chase, were fixed and analyzed by fluorescence microscopy. Cholesterol was visualized by using the cholesterol-binding compound filipin (9, 10). Whereas in uninfected macrophages cholesterol was distributed at the plasma membrane, as well as in intracellular stores (Fig. 1B), infection of macrophages with mycobacteria resulted in relocation and accumulation of cholesterol at the site of bacterial entry in all phagocytosing cells examined (Fig. 1B).

To further analyze the role of cholesterol in mycobacterial uptake, cellular cholesterol levels were reduced by pharmacological inhibition of cholesterol synthesis followed by extraction of residual cholesterol from the plasma membrane (6). Cholesterol depletion led to a specific decrease in total cellular cholesterol levels (∼60 to 70%) as determined by thin-layer chromatography (TLC) and densitometry (Fig. 2, C and D). Cholesterol-depleted macrophages were infected with metabolically labeled M. bovis BCG, and the degree of uptake was analyzed. Depletion of plasma membrane cholesterol resulted in a reduction of mycobacterial uptake of ∼85 to 90% by both the macrophage cell line J774A.1 and bone marrow–derived macrophages (Fig. 2A). In addition, cholesterol depletion caused a similar reduction ofM. tuberculosis uptake of ∼85% (Fig. 2B). To test whether inhibition of uptake was specific for mycobacteria or occurred as a result of a general inhibition of all phagocytic events, the ability of depleted macrophages to phagocytose various other microorganisms was analyzed. Depleted macrophages were still capable of internalizing Escherichia coli,Yersinia pseudotuberculosis,Salmonella typhimurium, andLactobacillus casei (Fig. 2B).

Figure 2

Effect of cholesterol depletion on mycobacterial entry. (A) Control (black bars) and cholesterol-depleted (6) (gray bars) subconfluent monolayers of J774 cells or bone marrow–derived macrophages (BMM) (27) were infected with metabolically labeled M. bovisBCG (OD600 = 3 in DMEM) for 1 hour, followed by a 2.5-hour chase, after which the cell-associated radioactivity was determined. The relative amount of uptake was determined by dividing counts per minute by the cell number in every sample. Results shown represent mean values (±SD) from three experiments and are corrected for background adherence occurring at 4°C. Au, arbitrary units. (B) Control and cholesterol-depleted J774 cells were allowed to internalize the different GFP-expressing (OD600 = 2) or metabolically labeled (OD600 = 3) microorganisms (28). Phagocytosis was corrected for background adherence in the cold and is expressed as percent phagocytosed bacilli compared with macrophages that were not depleted. (C and D) Analysis of cholesterol content in control and depleted J774 macrophages by TLC (C) and fluorescence microscopy (D). TLC analysis was performed as described (29), and visualized using 3% cupric acetate in 15% aqueous phosphoric acid and exposure of the TLC plate at 120°C for 1 hour.

To visualize phagocytosis events in normal and cholesterol-depleted macrophages, cells were followed by time-lapse video microscopy after the addition of bacteria. Mycobacteria were readily phagocytosed by control macrophages (Fig. 3A and see video, whereas depletion of cellular cholesterol resulted in an abrogation of mycobacterial uptake (Fig. 3A and see video). Instead of being internalized, the bacilli became loosely associated with macrophage membranes (Fig. 3A). Quantification of phagocytic events observed after video microscopy revealed that the percentage of phagocytosing cells was reduced to ∼15% after cholesterol depletion (Fig. 3B).

Figure 3

Time-lapse video microscopy. Control and cholesterol-depleted J774 macrophages grown to 20% confluency were incubated with mycobacteria (OD600 = 0.1), and 1-hour sequences were recorded by video microscopy using a digital camera (Hamamatsu Photonics, Japan) connected to a Zeiss Axiophot equipped with a thermostated (37°C) and CO2 (5%) equilibrated chamber and OpenLab software (version 2.0). (A) Example of a phagocytic event from control (upper panels) and cholesterol-depleted (lower panels) macrophages. Bacteria are indicated by arrows. Time points are indicated in minutes.seconds. Scale bar, 10 μm. Full video sequences can be accessed at (B) Quantification of phagocytic events (control, 95 cells recorded from seven sequences; depleted, 89 cells recorded from eight sequences). Shown are mean values (±SD). (C andD) Effect of cholesterol depletion on membrane ruffling (C) and cell locomotion (D) determined from the sequences used in (B). Active movement of cells over a distance of 50 μm within 1 hour was considered locomotion. (E) Protein biosynthesis in control and cholesterol-depleted J774 cells. Cells were metabolically labeled for 1 hour with [35S]methionine and cysteine, washed, and lysed as described (30). Shown are mean values (±SD) from three experiments.

As cholesterol is involved in the regulation of membrane fluidity (11, 12), depletion may affect other cellular functions besides mycobacterial uptake. To analyze whether depletion of surface cholesterol caused a general malfunctioning of the plasma membrane, the overall motility, as well as the membrane ruffling, of control and depleted cells was analyzed. All cells under control and cholesterol-depleted conditions showed membrane ruffling (Fig. 3C), although qualitative differences were apparent, most likely because of altered membrane fluidity (see video). Consistent with a role for cholesterol in providing rigidity to the cell surface (12), cellular motility was moderately enhanced after cholesterol depletion (Fig. 3D). In addition, the effect of cholesterol depletion on protein biosynthesis was analyzed and found to be equal in control and depleted cells (Fig. 3E).

To better understand the role of cholesterol in mycobacterial entry, the receptor molecules involved were analyzed. Whereas mannan did not block mycobacterial uptake (13), antibodies against complement receptor type 3 (anti-CR3) (14–17) blocked the entry of mycobacteria into J774A.1 cells, as well as into bone marrow–derived macrophages (Fig. 4A). To define the role of cholesterol in CR3-mediated uptake per se, macrophages were depleted of cholesterol and the capacity of control and depleted macrophages to phagocytoseL. casei in a CR3-dependent manner was analyzed. Cholesterol-depleted and control cells showed an equal capacity to internalize L. casei, and the uptake ofL. casei in both depleted and control macrophages was inhibited by anti-CR3 antibodies (Fig. 4B). Thus, cholesterol is not required for complement receptor function in general but specifically for mycobacterial entry.

Figure 4

(A) Effect of anti-CR3 antibodies (Serotec) on uptake of metabolically labeled mycobacteria by J774 and bone marrow–derived macrophages (BMM). (B) Effect of cholesterol depletion on CR3-mediated uptake of serum opsonized GFP-expressing lactobacilli by J774 macrophages. Cells were preincubated with the anti-CR3 antibodies for 10 min followed by the addition of bacteria for a period of 40 min. AU, arbitrary units. In (A) monoclonal antibodies M1/70 and 5C6 (80 μg/ml each) were used as a mixture; in (B) only 5C6 (20 μg/ml) was tested. (C) Cholesterol binding capacity of various bacteria. Bacteria (6 × 106) were incubated with the indicated amounts of [4-14C]cholesterol (NEN Life Sciences) for 1 hour at room temperature in 400 μl of DMEM. Bound and free cholesterol were separated by centrifugation of the bacterial suspensions through 0.5 M sucrose cushions, and the amount of cholesterol associated with the bacterial pellets was determined by liquid scintillation counting. Shown are mean values (±SD) from two experiments corrected for background values from samples without bacteria.

Because cholesterol is not required for complement receptor function, but is nevertheless crucial for mycobacterial entry into macrophages, we next explored the possibility that mycobacteria could directly bind to cholesterol. Mycobacteria displayed a high binding capacity for cholesterol (Fig. 4C), whereas none of the other microorganisms analyzed displayed a similar binding of cholesterol, with the exception of Yersinia; the significance of this affinity is presently unclear, as cholesterol does not play a role inYersinia uptake (Fig. 2B).

Together, our results indicate that plasma membrane cholesterol is required for a stable physical interaction of the mycobacteria with the plasma membrane, as a result of which the mycobacteria can be efficiently internalized. What may be the benefit of cholesterol-mediated entry for mycobacteria? Because TACO associates with the phagosomal membrane in a cholesterol-dependent manner, mycobacteria that enter macrophages at cholesterol-containing plasma membrane domains are subsequently sequestered in TACO-coated phagosomes, preventing lysosomal delivery and ensuring intracellular survival.

One of the hallmarks of mycobacteria is the extremely glycolipid-rich cell wall (18), which may contain compounds involved in cholesterol-mediated entry into macrophages. The availability of a blueprint for the biochemical pathways operating inM. tuberculosis through the elucidation of its complete genome sequence (19, 20) might allow the identification of components that may prove to be targets for anti-mycobacterial reagents.

  • * To whom correspondence should be addressed. E-mail: pieters{at}


View Abstract

Navigate This Article