The full-of-bacteria gene is required for phagosome maturation during immune defense in Drosophila

Early steps in phagocytosis do not require Fob

Another facet of host defense is the phagocytosis of invading microbes. After injection with pHrodo-labeled bacteria into their abdomen, wild-type flies displayed a characteristic accumulation of fluorescence in sessile hemocytes in the thorax (Fig. 2 A). This reflects the uptake of bacteria into hemocyte phagosomes and their subsequent acidification (Cuttell et al., 2008). In comparison, fob¹ flies exhibited weaker fluorescence, but the number of immobilized bacteria did not seem drastically reduced, which is consistent with the presence of active phagocytic sessile hemocytes. In contrast, in eater flies, the number of bacteria appeared reduced but the signal strength of individual punctae was similar to wild type (Fig. 2 A). The loss of one of several phagocytic receptors in eater flies reduces phagocytic uptake but not phagosome maturation and acidification (Kocks et al., 2005). In contrast, the reduced pHrodo fluorescence in fob¹ flies suggested a defect in phagosomes acidification.

Open in a separate window

Figure 2.

fob mutants exhibit a defect in bacterial clearance. (A) 2 h after injection, pHrodo-labeled E. coli (arrows) were visible around the dorsal vessel in the thorax of wild-type and eater flies (Df(3R)D605/Df(3R)Tl-I). In contrast, diffuse weaker signals appeared in fob¹ flies. (B) Hemocytes were allowed to engulf FITC-labeled E. coli for 15 min, and, after quenching fluorescence of external bacteria with Trypan blue, the fluorescence of phagocytosed bacteria was visible in wild-type (B) and fob¹ (C) cells visualized by differential interference contrast microscopy (B′ and C′). (D and E) After 45 min of chase, bacteria were cleared from wild type (D) but accumulated inside fob¹ hemocytes (E). (F) Box and whisker plots display the number of bacteria detected in hemocytes of indicated genotypes after 45 min of chase. Bars: (A) 0.5 mm; (B and C) 25 µm; (D and E) 10 µm.

To address this issue at higher resolution, we analyzed phagocytosis in primary hemocytes from third instar larvae (Pearson et al., 2003). Compared with wild type, fob¹ hemocytes did not exhibit a defect in initial phagocytic uptake of bacteria (Fig. 2, B and C). Subsequently, wild-type cells efficiently digested bacteria (Fig. 2 D), and after a 45-min chase contained only 4.8 ± 3 bacteria (n = 6). In contrast, at this time point, fob¹ cells (Fig. 2 E) were still full of bacteria (26 ± 16, n = 6, P < 0.0001). Similarly, elevated levels of bacteria were observed after knockdown of Vps33B or Vps16A (Fig. 2 F). Inability to digest bacteria was also observed in vivo after injection with E. coli. In fob¹ flies, the bacterial burden remained elevated (Fig. 1 G), as opposed to the efficient clearance of the majority of bacteria in wild-type flies after just 1 d. Together, these data suggest that the poor survival of fob¹ flies after injections with nonpathogenic E. coli reflects a defect in phagosome maturation.

Phagosome maturation requires Fob

A necessary step in the acquisition of the full degradative potential of phagosomes is their acidification, which can be monitored by imaging the fluorescence ratio of Oregon green/Texas red doubly labeled phagocytosed bacteria. After a 30–45-min chase in wild-type hemocytes, the majority of internalized bacteria appeared degraded as judged by the diffuse appearance of remaining fluorescence, but the few phagosomes that still contained well-defined bacteria had acidified to a mean pH of 5.5 ± 0.15 compared with a mean starting pH of 6.5 ± 0.15 (Fig. 3 A and Fig. S1). In contrast, fob¹ phagosomes acidified only minimally, if at all: the starting mean pH of 6.6 ± 0.12 was only lowered to pH 6.3 ± 0.13 after a 30-min chase (Fig. 3 B). In accordance with this finding, electron microscopy of isolated fob¹ hemocytes revealed that after a 30-min chase, phagocytosed bacteria were predominantly present in late phagosomes (Fig. 3, C and D) that were characterized by a mix of undigested or mildly degraded bacteria (Fig. 3 C′). In contrast, after a 30-min chase in wild-type hemocytes, the majority of bacteria had reached phagolysosomes (Fig. 3, C′′ and D) characterized by strongly degraded content (Fig. 3 C′′).

Open in a separate window

Figure 3.

fob phagosomes fail to mature. Double-labeled bacteria were allowed to internalize for 10 min (broken line) or 30 min (solid line), and images were captured for 15 min. The distribution of fluorescence ratios is shown for Ore-R (A) and fob¹ (B). The fluorescence ratio relates to pH as shown in Fig. S1. (C) Electron micrographs of phagosomes detected after a 30-min chase of phagocytosed E. coli. Phagosomal structures were broadly classified in three categories based on their ultrastructural appearance: phagosome (C), late phagosome (C′) and phagolysosomes (C′′). Bar, 1 µm. (D) Relative distribution of different categories of phagosomes in Ore-R and fob1. Data were collected from two independent sets of experiments with equivalent results. Quantification was performed in triplicate with a representative example shown in D.

Phagosomes mature by interacting with endosomal compartments (Desjardins et al., 1994). Thus, we explored endosomal markers, including Rab GTPases (Smith et al., 2007), as indicators of phagosome maturation. We found that a Rab-5 effector, Rabenosyn-5 (Rbsn-5), was present on fob¹ and wild-type phagosomes (Fig. 4, A and F). Rbsn-5 is a FYVE domain protein whose phagosomal localization depends on activation of phosphatidylinositol 3-kinases (PI3K) and Rab5 (Stenmark et al., 1995; Vieira et al., 2003). These data indicated that early steps in phagosome maturation, including the generation of 3-phosphoinositides and subsequent recruitment of Rab5 effectors, are normal in fob¹ mutant phagosomes. This conclusion was further supported by the presence of Avalanche (Avl), an early endosomal SNARE (Lu and Bilder, 2005), on Rbsn-5–positive phagosomes in fob¹ and wild-type hemocytes (Fig. 4, A and F). This indicates that fob¹ mutants have normal early endosome–phagosome fusion.

Open in a separate window

Figure 4.

fob is required for the fusion of phagosomes with lysosomes. Micrographs show hemocytes isolated from wild-type (A–E) or fob¹ (A’–E’) larvae. (A–C) Hemocytes were allowed to phagocytose FITC-labeled E. coli, and were immunostained for Avl (red) and Rbsn-5 (blue; A), Rab7 (B), or Hook (C). (D) Hemocytes were allowed to internalize dextran–Alexa Fluor 488 (10 kD), which after a 90-min chase partially colocalized with LysoTracker in wild type (D) and fob¹ (D’), indicating functional labeling of lysosomes by dextran. (E) Hemocytes with lysosomes preloaded with internalized Texas red–dextran were allowed to phagocytose GFP-labeled E. coli. After 30–45 min of chase, the bacteria colocalized with dextran in wild-type (E) but not fob¹ hemocytes (E′). For display, images were imported into Photoshop (Adobe) and adjusted for gain, contrast, and gamma settings. Bar, 5 µm. (F) Bar graphs indicate percentages of bacteria in phagosomes positive for the indicated markers or the percentage of dextran in LysoTracker-positive structures (error bars indicate standard deviation). Hemocytes were harvested from larvae with indicated genotypes.

A subsequent step in phagosome maturation involves the exchange of Rab5 to Rab7, similar to their exchange observed in endosomes (Vieira et al., 2002; Rink et al., 2005). 62 ± 26% of phagosomes in fob¹ hemocytes were decorated by Rab7 compared with 36 ± 14% in wild type, which indicates that Rab5-to-Rab7 conversion was not inhibited in fob¹ (Fig. 4, B, B′, and F). This significantly increased presence of Rab7 on phagosomes in fob¹ cells (P < 0.0001) suggests that phagosomes are stalled at this stage. This is reminiscent of the dramatic increase of Rab7 on late endosomes in car-null cells (Fig. S2 K; Akbar et al., 2009) and is consistent with Rab7 recruitment not being sufficient to induce fusion with lysosomes (Vieira et al., 2003). We explored other markers and found that Hook was present on 31 ± 8% of wild type but only on 6 ± 2% of fob¹ phagosomes, without ever decorating entire phagosomes as we observed in wild-type cells (Fig. 4 C). Interestingly, Drosophila Hook has been implicated in the maturation of multivesicular bodies (Sunio et al., 1999), which are involved in phagosome maturation (Philips et al., 2008). Considering the connection between endosomal and phagosomal maturation pathways, our data suggest that fob¹ phagosomes failed to acquire late endosomal/lysosomal characteristics due a loss of fusion with those compartments.

Infection experiments

E. coli (DH5α, amp resistance, GFP) and E. faecalis cultures were grown overnight in Luria Bertani (LB) or brain heart infusion medium (BHI) medium at 37°C. Female virgin flies (5 d old) were injected (Schneider et al., 2007) with 80 nl PBS containing a mean of 1,600 E. coli (OD₆₀₀ = 0.1) or 200 E. faecalis (OD₆₀₀ = 0.005). Sterile PBS was injected as a control. Injected flies (20 flies per vial) were reared at 25°C, 65% humidity, on yeast-agar-molasses food. Injections were performed with a pico-injector (model PLI-188; Nikon) fitted with glass capillary needles. Injections were performed in triplicate (total of 60 flies) for each group with either of the indicated microbes and PBS control on the same day. All injection experiments were repeated 8–10 times. For each survival curve, flies were counted every 24 h, and bars represent mean values with standard deviation. Data were analyzed using the SAS software (SAS Institute, Inc).

To determine bacterial load, flies were injected with E. coli (DH5α, kanamycin resistant, five flies per data point) and homogenized after the indicated time (Schneider et al., 2007). Serial dilutions were plated and colonies were counted for each time point. Data are plotted as boxes with whiskers. The mean is indicated with a diamond. The boxes indicate 25th and 75th percentiles; the bold line is the 50th percentile, whereas the whiskers show the complete range of the data.

pHrodo-E. coli bioparticles (Invitrogen) were suspended according to manufacturer’s instructions, and 80 nl were injected. After 2 h, flies were mounted and imaged on a microscope with 1.5× magnification (SZX12; Olympus). During imaging, exposure parameters were set such that for Oregon-R the brightest spots were not saturated. FITC-E. coli (catalogue No. F6694; Invitrogen) were used for phagocytosis and immunostaining experiments in hemocytes.

For qRT-PCR experiments, RNA was isolated using TRIzol (Invitrogen) according to the manufacturer’s instructions. For anti-microbial peptide measurements, RNA was isolated from five flies after injection (6 h for E. coli and 12 h for E. faecalis). qRT-PCR was performed using a DNA-free, high-capacity cDNA reverse transcription kit (Fast SYBR Green master mix; Applied Biosystems) and a Fast Real-Time PCR System (7500; Applied Biosystems). Each data point was repeated three times beginning from injection. Values were normalized first with rp49 as an internal control and then expressed as fold change compared with flies injected with PBS as control. The following primer sets were used for amplification: fob left, 5′-TATTGGAACCGATCCTCTCG-3′; fob right, 5′-CACCAGTTTCAATGCCTCCT-3′; Ca left, 5′-CCATATCAGCCGCATTTCTT-3′; Ca right, 5′-AAGCTGGCATCGTTCTGACT-3′; CG7829 left, 5′-CAGGAACCTACTGGGCAAAA-3′; CG7829 right, 5′-AGTAGACTCCCGGCTTGTCC-3′; CG7802 left, 5′-GTCGCGACATCGACACTTC-3′; CG7802 right, 5′-CGTTGGCAGTGAATGTGGT-3′; Attacin A left, 5′-TGCAGAACACAAGCATCCTAA-3′; Attacin A right, 5′-TAAGGAACCTCCGAGCACCT-3′; Cecropin A1 left, 5′-TCTTCGTTTTCGTCGCTCTC-3′; Cecropin A1 right, 5′-ACATTGGCGGCTTGTTGAG-3′; Defensin left, 5′-GATGTGGATCCAATTCCAGA-3′; Defensin right, 5′-CTTTGAACCCCTTGGCAAT-3′; Diptericin left, 5′-ACCGCAGTACCCACTCAATC-3′; Diptericin right, 5′-CCATATGGTCCTCCCAAGT-3′; G Drosocin left, 5′-TTCACCATCGTTTTCCTGCT-3′; Drosocin right, 5′-GGCAGCTTGAGTCAGGTGAT-3′; Drosomycin left, 5′-GTACTTGTTCGCCCTCTTCG-3′; Drosomycin right, 5′-ACTGGAGCGTCCCTCCTC-3′; rp49 left, 5′-ATCGGTTACGGATCGAACAA-3′; and rp49 right, 5′-GACAATCTCCTTGCGCTTCT-3′.

Hemocyte isolation and phagocytosis experiments

Hemocytes were collected from 60–80 wandering third instar larvae in Schneider’s Drosophila medium (10% heat-inactivated FBS) containing glass-bottom culture dishes. Cells were allowed to settle down for 15 min and washed with Schneider’s media followed by incubation with the indicated bacteria or dextran.

After a 15-min incubation at 4°C, unbound bacteria were washed out and phagocytosed bacteria were chased for the various times. (a) 15 min to measure initial uptake, after which extracellular fluorescent bacteria were quenched with Trypan blue. (b) To visualize late stage phagosomes in life cells, we chased for 30 min, after which we collected images for 15 min. We call this a “30–45 min chase.” (c) To capture early or late stage phagosomes by immunofluorescence staining, we fixed after 15–25 min or 30–45 min of chase. (d) We chased for 45 min to analyze bacterial persistence in cells, as at that time most bacteria were digested in wild-type hemocytes. Subsequently, cells were fixed (4% PFA) and counterstained with Phalloidin–Alexa Fluor 546 (Invitrogen). Oregon-R was used as control in all assays. (e) For electron microscopy analysis of phagosome maturation, we chased for 30 min. Cells were fixed with a mixture of 1.5% glutaraldehyde and 2.5% PFA for 2 h, then stained with osmium tetroxide, dehydrated, and embedded in epon. Sections (60–70 nm) were stained with uranyl acetate, dried, and viewed under a transmission electron microscope (120kV; Technai G2 spirit; FEI) with an 11-megapixel Morada camera (Olympus). Data were collected from two independent sets of experiments with equivalent results. Quantification was performed in triplicate, with a representative example shown in Fig. 3 D.

For phagocytosis experiments, n refers to the number of independent experiments that were quantified.

pH measurement

To measure acidity (pH), heat-killed E. coli (DH5α) were colabeled with the pH-sensitive fluorophore Oregon green 488 and the pH-insensitive carboxy-tetramethylrhodamine (Vergne et al., 1998). To calibrate their pH-dependent fluorescence, bacteria were suspended in phosphate-citrate buffer, pH 2.5–7.0, and imaged using a 63×, NA 1.4 lens on a confocal microscope (LSM510; Carl Zeiss, Inc.). Fluorescence ratios were determined for individual bacteria using ImageJ. To measure changes in phagosome pH, fob¹ or wild-type hemocytes were incubated with double-labeled bacteria (2–4 × 10⁷ ml⁻¹) at 4°C for 15 min. After a 10-min or 30-min chase, fluorescence ratios were measured from intact bacteria that did not appear degraded as indicated by diffuse fluorescence.

Immunofluorescence and dextran internalization

Hemocytes were incubated with dextran–Alexa Fluor 488 or dextran–Alexa Fluor 594 (10 kD, 1 mg/ml) for 5 min in PBS, pH 7.4, for monitoring fluid phase endocytosis. Free dextran was removed by washing extensively, and cells were chased for 90 min in Schneider’s Drosophila medium with 10% heat-inactivated FBS. After that chase, cells were either incubated with LysoTracker (GFP-certified Lyso-ID red lysosomal detection kit; Enzo Biochem, Inc.) or allowed to phagocytose GFP-tagged E. coli or latex beads to measure fusion of lysosome to phagosome.

For immunofluorescence staining after phagocytosis, hemocytes were fixed with 4% PFA for 45 min, then washed with PBS with 0.3% saponin. Samples were stained with the indicated primary antibodies rabbit anti-Hook (1:250; Sunio et al., 1999), rabbit anti-Avl (1:1,000; Pulipparacharuvil et al., 2005), rabbit anti–Rbsn-5 (1:5,000; a gift from A. Nakamura; Institute of Physical and Chemical Research Center for Developmental Biology, Kobe, Japan; Tanaka and Nakamura, 2008), rabbit anti-Rab7 (1:3,000; a gift of P. Dolph; Dartmouth College, Hanover, NH; Chinchore et al., 2009); and secondary antibodies labelled with Alexa Fluor 568 and Alexa Fluor 647 and mounted in Vectashield (Vector Laboratories).

Fluorescence images were captured with a 63×, NA 1.4 lens on an inverted confocal microscope (LSM510 Meta; Carl Zeiss, Inc.) at room temperature (21°C). All digital images were imported into Photoshop (Adobe) and adjusted for gain, contrast, and gamma settings.

Statistics

The LIFETEST procedure of SAS 9.2 (SAS Institute Inc.) was used to analyze survival curves by the Kaplan-Meier method. Logrank comparisons were used to assess significance of differences between curves. Student’s t tests (two-tailed) were used to determine the statistical significance of differences between colony-forming unit counts and changes of Avl/Rbsn-5, Rab7, and Hook localization to phagosomes was identified by bacterial content. At least 50 cells with phagosome structures containing bacteria were used for each of the quantifications. Phagosomes in >100 cells were used for the quantification of dextran/LysoTracker or dextran/bacteria.

We thank Drs. David Schneider and John Abrams for critical reading of the manuscript and Drs. Akira Nakamura and Patrick Dolph for antibodies. We are grateful to Sanchali Ray for technical assistance. We would like to thank the Bloomington Stock Center for fly stocks, the Berkeley Drosophila Genome Project for producing DNA clones, the Drosophila Genomics Resource Center for their distribution, and the Developmental Studies Hybridoma Bank at the University of Iowa for antibodies. FlyBase provided important information used in this work.

This work was supported by grants from the Canadian Institutes of Health Research (CIHR; MOP-81208) to W.H.A. Kahr, the National Institutes of Health (EY10199) to H. Krämer, and the core grant EY020799.