Higher order Rab programming in phagolysosome biogenesis

Dynamics of group V Rabs on phagosomes analyzed by quantitative four-dimensional (4D) confocal microscopy

We investigated the members of the group V Rabs (Pereira-Leal and Seabra, 2001), Rab5, Rab21, and Rab22a, by live microscopy using a previously published approach (Chua and Deretic, 2004; Vergne et al., 2005). The entry of a phagocytosed particle was identified as previously described (Chua and Deretic, 2004), and live imaging was initiated to record EGFP-Rab dynamics on nascent and maturing phagosomes containing latex beads (Fig. 1, A–C), followed by quantification using ratiometric analysis of intensities comparing phagosome and cytosol fluorescence values (R_φ/c; Fig. 1, G–I; Chua and Deretic, 2004; Vergne et al., 2005). The majority of the group V Rabs were transiently recruited to latex bead phagosomes (Fig. 1, A–C), with Rab5 and Rab21 (Fig. 1, A and B) desorbing from the phagosomes by 10 min after the uptake (Fig. 1, G and H), and with Rab22a showing a diminutive initial peak (Fig. 1, C and I). Because of a very low-level EGFP-Rab22a recruitment to latex bead phagosomes, we wondered whether macrophages expressed Rab22a. Rab22a expression was confirmed by RT-PCR (Fig. S1 A, available at http://www.jcb.org/cgi/content/full/jcb.200603026/DC1). In addition, endogenous Rab22a was detected in macrophages by immunofluorescence (Fig. S1 B). Hence, low levels of Rab22a on latex bead phagosomes cannot be explained by a lack of Rab22a expression in macrophages. Thus, the group V Rabs are transiently recruited in small amounts to latex bead phagosomes during early time points after phagocytosis.

Mycobacterial phagosomes recruit and retain copious amounts of Rab22a

We next tested group V Rab dynamics on mycobacterial phagosomes. Rab5 and Rab21 followed similar kinetics on both mycobacterial and latex beads phagosomes (Fig. 1, D and E, G and H). However, mycobacterial phagosomes displayed a marked difference relative to latex bead phagosomes by recruiting and retaining high quantities of Rab22a (Fig. 1, F and I). Enumeration of Rab5, Rab21, and Rab22a profiles (Fig. 1, J and K) confirmed that Rab22a was persistently accumulating on mycobacterial phagosomes. This was accompanied by diminishing levels of EGFP-Rab22aWT fluorescence in other parts of the cell, a phenomenon that was augmented in macrophages infected with more than one bacillus. EGFP-Rab22aWT–positive profiles were observed to tether and fuse with mycobacterial phagosomes, increasing EGFP-Rab22aWT levels on these organelles (Video 1, available at http://www.jcb.org/cgi/content/full/jcb.200603026/DC1). These observations were confirmed by immunofluorescence detection of endogenous Rab22a on bacillus Calmette-Guérin (BCG) phagosomes (Fig. S1 B). The differential distribution of EGFP-Rab22aWT was not caused by phagosome size difference because 3-μm latex beads (Fig. S1 C) behaved similarly to the 1-μm beads. Thus, Rab22a is specifically enriched on phagosomes containing mycobacteria.

Rab22a is an early endocytic Rab in macrophages

Rab22a has been implicated in early endosomal and recycling pathways in nonphagocytic cells (Kauppi et al., 2002; Weigert et al., 2004). We tested Rab22a localization in macrophages and found that both EGFP-Rab22aWT and endogenous Rab22a overlapped with the early endosomal marker EEA1 (Fig. S1, D and E). This is in keeping with the reported early endosomal localization of Rab22a in other cells (Kauppi et al., 2002). Immunofluorescence analysis using GM130, syntaxin 6, and TGN38 showed that in macrophages Rab22a was not on Golgi organelles (Fig. S1 F), and Golgi vesiculation did not occur in cells transfected with EGFP-Rab22aQ64L (Fig. S1 F), in contrast to a report that the Rab22a mutant vesiculates Golgi in CHO cells (Kauppi et al., 2002). The early endosomal localization of Rab22a, and the increased fusion of early endosomal organelles with mycobacterial phagosomes stimulated by Rab14 (unpublished data), may partially explain Rab22a enrichment on BCG phagosomes.

Rab22a affects phagosomal maturation

We examined whether expression of constitutively active Rab22a (EGFP-Rab22aQ64L) affected phagosomes harboring dead mycobacteria. Heat inactivation of M. tuberculosis incapacitates it to block phagolysosome biogenesis (Armstrong and Hart, 1971; Chua and Deretic, 2004; Vergne et al., 2005). The constitutively active mutant of Rab22a accumulates on mycobacterial phagosomes (Fig. S1 G) in a manner similar to wild-type Rab22a. Expression of EGFP-Rab22aQ64L inhibited maturation of phagosomes containing dead BCG, as follows: (a) heat-killed BCG phagosomes showed reduced colocalization with the acidotropic dye LysoTracker Blue (Fig. 2, A–F), indicating impaired acidification; (b) phagosomes with dead BCG showed lower proteolytic activity in macrophages transfected with EGFP-Rab22aQ64L, as indicated by lower staining with DQ-Red BSA, which is an endocytic protease substrate whose fluorescence dequenches upon proteolysis (Fig. 2, G–L); (c) although phagosomes harboring dead mycobacteria normally do not retain transferrin receptor (Fig. 3, A–D and Q), transfection with EGFP-Rab22aQ64L caused significant presence of transferrin receptor on dead mycobacterial phagosomes (Fig. 3, E–H and Q); (d) A similar effect was observed with another previously mapped (Fratti et al., 2003b) early/recycling endocytic marker, syntaxin 13 (Fig. 3, I–P and R); and (e) Rab11, a GTPase that controls transferrin receptor (TfR) recycling (Ullrich et al., 1996; Ren et al., 1998), accumulated on dead mycobacterial phagosomes in cells expressing EGFP-Rab22aQ64L when compared with control untransfected cells (34 ± 5% vs. 14 ± 4% colocalization; P = 0.04; Fig. S1 H), indicating that the constitutively active mutant of Rab22a conferred recycling endosomal characteristics upon dead mycobacterial phagosomes.

Knockdown of Rab22a overcomes the mycobacterial phagosome maturation block

In lieu of experiments with a dominant-negative Rab22a, which was found to cause macrophage detachment, we resorted to siRNA knockdown of Rab22a and examined its effects on maturation of phagosomes harboring live M. tuberculosis variant bovis BCG. Rab22a knockdown (Fig. 4 A) caused a threefold increase in the colocalization of live mycobacterial phagosomes with the most robust late endocytic marker CD63 (Fig. 4, B and E; Fratti et al., 2003b), indicating that live mycobacterial phagosomes were maturing into phagosomes with late endosomal characteristics. Live mycobacterial phagosomes also showed a doubling of V_o H⁺ATPase association with phagosomes containing live mycobacteria (Fig. 4, C and F; Sturgill-Koszycki et al., 1994; Fratti et al., 2003b). Rab22a knockdown did not cause indiscriminate mixing of endosomal markers (Fig. S2, A–C, available at http://www.jcb.org/cgi/content/full/jcb.200603026/DC1). Furthermore, colocalization of TfR with live mycobacterial phagosomes was diminished upon Rab22a siRNA knockdown compared with scrambled siRNA control (Fig. 4, D and G).

The effects of Rab22a knockdown with SMARTpool (Dharmacon) Rab22a siRNA (a combination of four Rab22a-specific siRNA duplexes) was confirmed using individual siRNA duplexes (Fig. S2 D), which also caused an increase in live mycobacterial phagosome maturation (Fig. S3, A–D, available at http://www.jcb.org/cgi/content/full/jcb.200603026/DC1). Furthermore, transfection with siRNA against the closely related Rab22b, which is also expressed in macrophages (Fig. S2, E and F), did not alter mycobacterial phagosomes (Fig. S3, A–D). The effects of Rab22a knockdown on mycobacterial survival (Fig. S3, E–G) were mild within the period investigated, and although a trend was observed, no statistically significant differences could be established (Fig. S3, F and G). We conclude that Rab22a is necessary to maintain M. tuberculosis phagosome maturation block, but that maturation block override does not automatically translate into direct bacterial elimination by macrophages, in keeping with the early observations by Armstrong and Hart (1975).

Rab22a knockdown leads to Rab7 conversion on M. tuberculosis phagosomes

A prequel to endosomal maturation into late endosomal/lysosomal organelles is Rab conversion (Rink et al., 2005). This term describes a process whereby an organelle synchronously sheds off early endosomal Rab(s) and concomitantly receives the late endosomal Rab, Rab7 (Rink et al., 2005). The signals for this transition are currently unknown (Deretic, 2005). We wondered whether Rab conversion applies to phagosomes, and whether Rab22a, as a candidate terminal recycling Rab involved in cargo and membrane sorting from the early endosome (Mesa et al., 2001; Weigert et al., 2004), could supply or contribute to such signals. To test this, we examined Rab7 acquisition by the mycobacterial phagosome, which was previously shown to exclude this critical late endocytic Rab (Via et al., 1997). Unlike in cells treated with control scrambled siRNA, Rab7 acquisition was increased to 80% on live mycobacterial phagosomes in macrophages in which Rab22a was knocked down by siRNA (Fig. 5). These findings are consistent with a functional role for Rab22a in mycobacterial phagosome maturation block. More generally, the conversion of live mycobacterial phagosomes into the Rab7 stage upon Rab22a knockdown suggests that Rab22a supplies signals preventing acquisition of Rab7 and precluding organellar maturation into a late endosomal/lysosomal compartment. Thus, Rab22a functions not only as a recycling Rab involved in cargo and membrane sorting from the early endosome (Mesa et al., 2001; Weigert et al., 2004), but it also acts as a coordinator of Rab succession. In keeping with our findings with phagosomes, there is a lack of endosomal EGF degradation in Rab22aQ64L-transfected Hep2 cells (Kauppi et al., 2002). In CHO cells, the expression of Rab22aQ64L causes endocytic tracers to remain in Rab22a-positive vesicles (Mesa et al., 2005). We propose that Rab22a is a central regulator of the transition to late endocytic organelles by signaling “all clear” and allowing the leftover sorting endosomal or phagosomal organelle to transit from an early compartment to a degradative organelle controlled by Rab7.

Cell culture and preparation of Texas red–labeled latex beads and M. tuberculosis variant bovis BCG

RAW264.7 macrophages and M. tuberculosis variant bovis BCG were maintained as previously described (Chua and Deretic, 2004). Mycobacteria were heat killed by incubation at 90°C for 5 min before labeling. Both live and dead mycobacteria were labeled with 0.5 mg/ml Texas red–succinimidyl ester and prepared as previously described (Chua and Deretic, 2004). Dead mycobacteria were also labeled with 0.25 mg/ml Alexa Fluor 647–succinimidyl ester in PBS for 1 h. Streptavidin-conjugated 1-μm polystyrene beads (Sigma-Aldrich) were labeled and prepared as previously described (Chua and Deretic, 2004). 3-μm polystyrene beads were opsonized in DME supplemented with 10% FBS before use.

Plasmids and transfection

The plasmids pEGFP-Rab21WT, pEGFP-Rab22aWT, and pEGFP-Rab22aQ64L were obtained from J. Donaldson (National Institutes of Health, Bethesda, MD), pEGFP-hRab5WT was obtained from P. Stahl (Washington University, St. Louis, MO), and pEGFP-Rab7WT was obtained from A. Wandinger-Ness (University of New Mexico, Albuquerque, NM). For transfection, 5 × 10⁶ RAW264.7 cells were resuspended in a nucleoporator buffer supplied by the manufacturer (Amaxa Biosystems) with 5 μg of plasmid DNA. Cells were nucleoporated according to the manufacturer's protocol and allowed to express the construct for 24 h before the imaging experiments.

Antibodies and endocytic tracers

Rabbit polyclonal antibody to Rab22a was obtained from J. Donaldson, and polyclonal antibody to syntaxin 13 was obtained from R. Scheller (Genentech, South San Francisco, CA). Rabbit polyclonal antibodies to transferrin receptor and CD63 were purchased from Santa Cruz Biotechnology, Inc. Antibody to V₀ was used as previously described (Fratti et al., 2003a,b). Mouse monoclonal antibody to transferrin receptor was purchased from Zymed Laboratories. Monoclonal antibodies against GAPDH, GM130, and TGN38 were obtained from Abcam. Lysotracker Blue, DQ Red BSA, and secondary antibodies conjugated to Alexa Fluor 488, 568, and 647 were purchased from Invitrogen. The acidotropic dye Lysotracker Blue was diluted in DME (1:10,000) and preloaded into macrophages for 2 h. DQ Red BSA was preloaded at 10 μg/ml for 3 h before infection. Cells were subsequently fixed and viewed using immunofluorescence microscopy.

Rab siRNA knockdowns and immunoblotting

Rab22a and Rab22b knockdowns were achieved by using siGENOME SMARTpool reagent (Dharmacon) specific for Mus musculus Rab22a and Rab22b (Dharmacon). All effects of Rab siRNAs were compared with siCONTROL Nontargeting siRNA pool (Dharmacon), which is labeled as scrambled siRNA in figures. RAW264.7 cells were transfected with 1.5 μg siRNA by nucleoporation. Immunoblotting (30 μg of total protein) was performed as previously described (Fratti et al., 2003a,b). GAPDH immunoblotting was used as a loading control. Rab22a single siRNA duplexes used individually were as follows: duplex 1, sense (CAGCAGCCAUCAUCAUCGUUUAUU) and antisense (5′-PUAAACGAUGAUGAUGGCUGCUGUU); duplex 2, sense (GGGAACAAGUGCGAUCUUAUU) and antisense (5′-PUAAGAUCGCACUUGUUCCCUU); duplex 3, sense (GAGAUUAGUCGAAGAAUUCUU) and antisense (5′-PGAAUUCUUCGAAGAAUUCUU); and duplex 4, sense (GGAUACGGGUGUGGGUAAAUU) and antisense (5′-PUUUACCCACACCCGUAUCCUU).

Immunofluorescence laser scanning confocal microscopy

Imaging of 1-μm-thick optical sections was performed using an Axiovert 200M microscope with an Axioscope 63× oil objective and LSM 5 Pascal or LSM 510 META systems (Carl Zeiss MicroImaging, Inc.). At least 200 phagosomes from three independent experiments were analyzed for colocalization studies.

4D confocal microscopy

A rotating disk confocal microscope (UltraView; PerkinElmer) that affords low photocytotoxicity and low photobleaching was applied for 4D imaging, as previously described (Chua and Deretic, 2004; Vergne et al., 2005). For ratiometric quantitative analysis (Chua and Deretic, 2004; Vergne et al., 2005) of a volume over time, z sections were collapsed into a single projection according to the published procedure (Gerlich et al., 2001). Transfected RAW264.7 cells were synchronously infected by centrifugation of bacteria or beads onto macrophages adherent to coverslips at 1,000 rpm for 5 min. Coverslips were mounted into a perfusion chamber (Harvard Apparatus) set at 37°C. Identification of mycobacterial entry and image acquisition was performed as previously described (Chua and Deretic, 2004). To measure R_Φ/C, fluorescence intensity of the phagosomal membrane was divided by background cytosolic fluorescence.

Mycobacterial survival assay

RAW264.7 macrophages were seeded at 2.0 × 10⁵ cells/well in 12-well plates after transfection with either siGENOME SMARTpool Rab22a siRNA or scrambled siRNA. Cells were incubated for 24 h. Macrophages were infected with live M. tuberculosis variant bovis BCG or live M. tuberculosis H37Rv (preincubated for 30 min at 37°C in DME), at a nominal multiplicity of infection of 10, followed by four washes using complete DME. Macrophages were hypotonically lysed using cold sterile water after a 2, 4, or 24 h chase period. Mycobacteria were plated for colony forming units on Middlebrook 7H11 agar (Difco) and incubated at 37°C for 2.5 wk. Bacterial viability was expressed as percentage of survival relative to scrambled siRNA control. Experiments were performed in triplicate.

Statistical analysis

Results are from experiments performed in triplicate. All statistical analyses were calculated using Fisher's protected least significant difference post hoc test (analysis of variance, ANOVA) (SuperANOVA 1.11; Abacus Concepts). P values of ≤ 0.05 were considered significant.

Online supplemental material

Fig. S1 shows that endogenous Rab22a and EGFP-Rab22aQ64L in macrophages is recruited to mycobacterial phagosomes and that Rab22a colocalizes with early endosomes, but not Golgi organelles. Fig. S2 shows an analysis of Rab22a knockdown effects on early and late endosomes, characterization of single duplex Rab22a siRNA knockdowns, and expression of Rab22b. Fig. S3 shows the effects of single-duplex siRNA Rab22a knockdown on mycobacterial phagosomal maturation and Rab22a knockdown on intracellular survival of mycobacteria. Video 1 shows an EGFP-Rab22aWT–transfected macrophage infected with live Texas red–labeled M. tuberculosis variant bovis BCG.