Increased Expression of Rab5a Correlates Directly with Accelerated Maturation of Listeria monocytogenesPhagosomes*

Previous studies have shown that Listeria monocytogenes (LM) modulates phagocytic membrane traffic. Here we explore whether Rab5a, a GTPase associated with phagosome-endosome fusion, is related to phagosome maturation and to the intracellular survival of LM. Stable transfection of Rab5a cDNA into macrophages accelerates intracellular degradation of LM. Morphological studies confirmed that phagosome maturation and phagosome-lysosome fusion is enhanced by overexpression of Rab5a. Down-regulation experiments using antisense oligonucleotides targeted to the Rab5a mRNA efficiently reduced Rab5a synthesis, reduced phagosome-endosome traffic, blocked phagosome-lysosome fusion, and extended intraphagosomal survival of LM. Down-regulation of Rab5a had no effect on LM internalization. Down-regulation of Rab5c had no effect on phagosome maturation and phagosome-lysosome fusion. The results indicate that Rab5a controls early phagosome-endosome interactions and governs the maturation of the early phagosome leading to phagosome-lysosome fusion.


Previous studies have shown that Listeria monocytogenes (LM) modulates phagocytic membrane traffic.
Here we explore whether Rab5a, a GTPase associated with phagosome-endosome fusion, is related to phagosome maturation and to the intracellular survival of LM.

Stable transfection of Rab5a cDNA into macrophages accelerates intracellular degradation of LM. Morphological studies confirmed that phagosome maturation and phagosome-lysosome fusion is enhanced by overexpression of Rab5a.
Down-regulation experiments using antisense oligonucleotides targeted to the Rab5a mRNA efficiently reduced Rab5a synthesis, reduced phagosome-endosome traffic, blocked phagosome-lysosome fusion, and extended intraphagosomal survival of LM. Down-regulation of Rab5a had no effect on LM internalization. Down-regulation of Rab5c had no effect on phagosome maturation and phagosome-lysosome fusion. The results indicate that Rab5a controls early phagosome-endosome interactions and governs the maturation of the early phagosome leading to phagosome-lysosome fusion.
Phagocytosis is a complex process required for host defense and tissue remodeling. The uptake of pathogens and the activation of membrane trafficking and other events that lead to killing and disposal is key to an efficient host defense strategy. Listeria monocytogenes (LM), 1 a human pathogen that infects a variety of cell types including macrophages (MØ), has served as an excellent model for examining membrane trafficking events involved in intracellular killing (1). MØ, unlike other host cells, readily clear intracellular infections (2,3). Mounting a successful defense against internalized pathogens appears to require maturation of the phagosome leading to phagosome-lysosome fusion and the discharge of lysosomal contents into the phago-some. Thus inhibition of phagosome-lysosome fusion represents a common strategy for sustained intracellular growth. Variations of this strategy can be found in Mycobacterium tuberculosis, Listeria monocytogenes, and Salmonella typhimurium among others.
Virulent strains of LM have been shown to access the cytoplasm where bacterial growth flourishes. However, using a nonhemolytic mutant of LM (LM hlyϪ ), which is retained within the phagosome, we have made several observations that are pertinent to intracellular survival of the pathogen including delayed phagosome maturation and fusion with lysosomes. The choice of the LM hlyϪ mutant was fortuitous, since it allowed us to unmask bacterial targets, which modulate intracellular trafficking, that would not have been possible using the virulent organism (4).
Rab5, the rate-limiting GTPase for endocytosis (5)(6)(7), is expressed as three different isoforms (a, b, c) that appear to have overlapping intracellular distributions (8). Rab5 isoforms in the pathogen Trypanosoma brucei appear to have different localization and functions (9). Although Rab5a has been localized on phagosomes containing different particles (10 -13), its role in mediating phagosome maturation has not been extensively investigated.
Our results indicate that Rab5a and Rab5c play different roles in the phagocytic pathway.
Immunoprecipitations and PT-antisense Oligonucleotide Treatment-Immunoprecipitation was carried out as described (18). PT-oligonucleotides (10 g) were incubated with 10 g/ml of Lipofectin in 200 l of Opti-MEM for 15 min at room temperature before being added to Opti-MEM-treated HMØ for 4 h at 37°C. For electroporation, HMØ (5 ϫ 10 6 /ml) were incubated with 10 g of PT-oligonucleotides on ice for 10 min. Electroporation was performed with a Baxter BTX-600 electroporator using 2-mm gap cuvette chambers with the following settings: 220 V, 800 microfarads, 72 ⍀. Cells were placed on ice immediately, and complete medium was added. Cells were set onto culture plates for 2 h at 37°C and extensively washed before use.
LM Phagocytosis-LM infection was performed according to standard protocols (4, 11) at a 10:1 bacteria/cell ratio. Cells were incubated at 37°C for 15 min to allow for bacterial uptake followed by a 45-min * This work was supported by National Institutes of Health grants (to P. D. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Recipient of a Postdoctoral Fellowship from the Formación de Personal Investigador, Ministerio de Educación y Ciencia, Madrid, Spain.
incubation in medium containing gentamicin (5 g/ml) to kill extracellular bacteria. For kinetic experiments, this time period was used as the zero time point. For other time points, infected cells were incubated at 37°C in complete medium containing gentamicin (5 g/ml) for the indicated time and then washed with phosphate-buffered saline. Cells were fixed in 2.5% glutaraldehyde and processed for electron microscopy. For kinetic studies, cells were lysed and plated onto brain heart infusion agar plates (37°C, 24 h). The number of live bacteria was estimated by counting cfu (colony forming units). For phagosome-endosome or phagosome-lysosome fusion assays, infections were allowed for only 10 min (1 h, 60°C (11)). Phagosomes from J774 cells were isolated as described with minor modifications (4,11,19).
LM hlyϪ Uptake and Catabolism Assays-Phagocytosis and catabolism assays were performed as described (4). Bacteria were labeled with Tran 35 S-label. Dead (1 h, 60°C (11)) and live LM (3 ϫ 10 5 cpm/well) were added to 2 ϫ 10 6 HMØ pretreated with PT-oligonucleotides as described above. After 20 min of internalization, cells were washed and lysed to quantify LM uptake. To measure LM catabolism, cells were incubated for 1.6 h before lysis. Cells were solubilized in 1% Triton X-100 and proteins were precipitated with 10% trichloroacetic acid.

RESULTS AND DISCUSSION
Previous studies have shown that LM hlyϪ interferes with phagocytic trafficking and phagosome maturation (4,11). Here, we analyze the role of Rab5a in phagosome maturation. To demonstrate that Rab5a, or a Rab5a regulatory factor plays a role in phagosome maturation, we stably expressed Rab5a in the mouse MØ cell line, J774, and investigated whether elevated levels of functional Rab5a could override the inhibitory effect of the bacterium. Stable transfection of J774 with Rab5a cDNA cloned into a pcDNA3 vector (Fig. 1A) increased Rab5a levels by approximately 5-10-fold. Intracellular killing of internalized LM hlyϪ in Rab5a-transfected cells was clearly higher than in cells transfected with the vector alone or in nontransfected cells (Fig. 1A). To confirm that overexpression of Rab5a overrides the inhibition of phagosome maturation caused by LM hlyϪ , we recorded the interactions of phagosomes containing live LM hlyϪ with the endosomal and lysosomal compartments (Fig. 1B). Whereas LM hlyϪ -containing phagosomes in control cells interacted extensively with the endosomal compartment and minimally with lysosomes, in Rab5a-overexpressing cells, phagosome access to both endosomal and lysosomal markers was enhanced. Phagosome-lysosome fusion was particularly enhanced in Rab5a-transfected cells (from 5% in control cells to 45% in cells overexpressing Rab5a). Phagosome . Infection with LM hlyϪ was performed as described previously (4). The ratio of live bacteria (i.e. cfu) recovered at time 0 divided by the cfu recovered at 8 h was used as an index of intracellular killing. Cfu recovered at time 0 for control and for Rab5atransfected cells were 2.5 ϫ 10 6 Ϯ 267 cfu and 9 ϫ 10 5 Ϯ 189 cfu, respectively. At 8 h, cfu recovered from control and from Rab5a-transfected cells was 1 ϫ 10 6 Ϯ 122 cfu and 2.7 ϫ 10 4 Ϯ 64 cfu, respectively. Results are the mean Ϯ S.D. of four different experiments. B, J774 cells transfected with Rab5a or with pcDNA3 vector alone (control cells) were incubated with BSA-gold (10-nm particles, 1 mg/ml) for 10 min and chased overnight to label lysosomes. Another set of cells was incubated the following day with BSA-gold for 10 min to label endosomes. All cells were then infected with LM hlyϪ for 10 min. Total gold particles per cell were quantified in each case from a total of 200 cells, as well as the number of gold particles found in phagosomes. Results are expressed as the percentage of gold particles found in phagosomes compared with total gold. Results are representative of at least three different experiments. C, cells were infected with radiolabeled live LM hlyϪ (200,000 cpm/well) as reported previously (4), incubated at 37°C for 20 min, washed, and solubilized with 1% Triton X-100. Proteins were precipitated from cell lysates with 10% trichloroacetic acid on ice. Results correspond to bacteria uptake (counts/min) after a 20-min incubation.
Results are the mean of triplicates Ϯ S.D. of four different experiments.

FIG. 2.
Overexpression of Rab5a accelerates the transformation of the LM hly؊ -phagosomal compartment to a "phago-lysosome"-like compartment. J774 cells transfected with Rab5a or with pcDNA3 vector alone (control cells) were offered BSA-gold (10 nm, 1 mg/ml) as in Fig. 1 to label endosomes or lysosomes. The cells were infected as above. a shows a representative control cell that exhibits the typical swollen appearance of live LM hlyϪ phagosomes (4). b and c show two representative Rab5a-transfected cells where the lysosomal compartment has been labeled with BSA-gold. These phagosomal compartments exhibit tightly apposing membranes and appear full of dense material and of multivesicular bodies. Phagosomes are heavily labeled with lysosomal BSA-gold. d shows a representative Rab5a-transfected cell where the endosomal compartment has been labeled with BSA-gold. maturation was also confirmed by the acquisition of typical lysosomal proteins such as Lamp-1 and the mature form of cathepsin-D (Ref. 4 and data not shown). Rab5a appears to operate principally at the phagosomal compartment rather than in phagosome formation, since overexpression of Rab5a caused only a moderate increase in LM hlyϪ uptake (1.2-1.5fold) (Fig. 1C).
The dramatic effect of Rab5a overexpression on phagosome maturation is also reflected in the morphological characteristics of the LM hlyϪ phagosomal compartment. LM hlyϪ phagosomes in control cells appear as swollen, "endosomal-like" compartments (see Fig. 2a). In Rab5a-transfected cells, LM hlyϪ phagosomes display tightly apposed limiting membranes with dense material and membrane inclusions, typical of multivesicular and lysosomal compartments (BSA-gold-marked lysosomes are shown in Fig. 2, b and c, and BSA-gold marked endosomes are shown in Fig. 2d). Interestingly, this compartment resembled phagolysosomes containing dead bacteria (4). All of these findings point to a linkage between phagosome maturation and LM hlyϪ killing in MØ and to Rab5a as an essential regulator of both processes.
Conclusions based on overexpression of proteins alone may not rule out compensatory expression of other factors that could account for the dramatic effects observed above. To corroborate our findings with Rab5a overexpression, we carried out experiments to determine whether down-regulation of Rab5a would alter the killing capabilities of the MØ and allow for LM hlyϪ survival.
Antisense phosphothioate oligonucleotides (PT-oligonucleotides) were chosen due to their longer intracellular half-lives (17,18,20). Transfection with PT-oligonucleotides was carried out by two methods with similar results (Lipofectin treatment or electroporation) (21). Transfection with antisense PT-oligonucleotides directed to the AUG translation initiation codon of the mRNA sequence of Rab5a and Rab5c was carried out on HMØ, because the sequences of human Rab5 isoforms (Rab5a and Rab5c) are known, whereas the mouse sequences are not available. The intracellular degradation of LM hlyϪ in HMØ resembles that found in the mouse MØ cell line, J774 (22)(23)(24). Several PT-oligonucleotides, designed to hybridize with specific Rab5a mRNA sequences (e.g. effector domain, 3Ј-untranslated regions, or translation initiation codon), were examined. However, only those oligonucleotides hybridizing with the translation initiation codon of Rab5a mRNA blocked Rab5a synthesis (approximately 85-90%, as determined by immunoprecipitation (Fig. 3A)). Rab5c, which shares more than 80% identity with Rab5a (8,25), was unaltered by Rab5a-antisense treatment indicating that the PT-oligonucleotide effects were highly specific. Rab5a-antisense-treated HMØ permitted intracellular growth of LM hlyϪ mutant, while in sense or control cells, the bacteria were destroyed (Fig. 3B) (biosynthetic levels of Rab5a   FIG. 3. Down-regulation of Rab5a selectively impairs LM hly؊ killing without affecting internalization rates. A, down-regulation of Rab5a or Rab5c with PT-antisense oligonucleotides. HMØ were treated with PT-oligonucleotides. Rab5 from antisense (A lanes)-or sense (S lanes)-treated cells was immunoprecipitated with monoclonal anti-Rab5a antibody (4F11) (Rab5a lanes) or with rabbit polyclonal antibody anti-Rab5c (Rab5c lanes). Upper panels show Rab5a and Rab5c immunoprecipitates after treatment with Rab5a-PT-antisense or sense oligonucleotides. Lower panels show Rab5a and Rab5c immunoprecipitates after treatment with Rab5c-PT-antisense or -sense oligonucleotides. B, HMØ were treated with Rab5a or Rab5c PT-oligonucleotides (antisense (A) or sense (S)) or nontreated (control (C)) as in A and plated at 2 ϫ 10 6 cells/ml. Infection with LM (LM hlyϪ ) was performed as described with a 10:1 bacteria:cell ratio (4,9). Bacteria recovered from Rab5a-sense-treated cells at 0 h and at 8 h were 10 4 Ϯ 369 cfu and 3 ϫ 10 3 Ϯ 175 cfu, respectively. In Rab5a-antisense-treated cells, recovery at 0 h was 1.2 ϫ 10 4 Ϯ 305 cfu and at 8 h, 4.5 ϫ 10 4 Ϯ 500 cfu. Bacteria recovered from Rab5c-sense-treated cells at 0 h was 1.5 ϫ 10 4 Ϯ 218 cfu, and at 8 h recovery was 2.5 ϫ 10 3 Ϯ 106 cfu. In Rab5c-antisense-treated cells recovery at 0 h was 1.6 ϫ 10 4 Ϯ 303 cfu and at 8 h, recovery was 2.8 ϫ 10 3 Ϯ 182 cfu. Infection was allowed to proceed for 4 and 8 h (zero time refers to the 2-h postinfection time point). After cell lysis, bacteria were plated onto brain heart infusion agar plates, incubated at 37°C for 24 h, and the number of live bacteria was estimated by cfu counting. Individual data points were carried out in triplicate, and results are expressed as the mean Ϯ S.D. of three independent experiments. Insets correspond to Rab5a (upper plot) and Rab5c (lower plot) levels measured by immunoprecipitation after treatment for 24 h with Rab5a-or Rab5c-antisense or -sense PT-oligonucleotides, respectively. C, HMØ were treated with Rab5a PT-oligonucleotides (antisense or sense) or were untreated (control as in A). Bacteria were radiolabeled with Tran 35 S-label and internalized as described (4). Dead or live LM hlyϪ (300,000 cpm/well) were sedimented onto cells to synchronize infection. A total of 2 ϫ 10 6 cell-associated radioactive bacteria were offered to 2 ϫ are shown as an inset in Fig. 3B). LM hlyϪ growth occurred inside the phagosomes, since this mutant lacks the protein necessary for lysis of the phagosomal membranes (as detected by electron microscopy, data not shown). The growth of LM hlyϪ in Rab5aantisense-treated cells was heterogeneous (Table I) It has been reported that the Rab5 isoforms (a, b, and c) share similar functions and colocalize to the same compartment (8,(25)(26)(27). However, recent results demonstrating differential expression of Rab5a following lymphokine signaling in MØ suggests specialized functions for the Rab5 isoforms (19).
Rab5c-antisense treatment was efficient and selective in blocking Rab5c synthesis (Ͼ88% inhibition, see Fig. 3A), since Rab5a synthesis remained unaltered. Analysis of LM hlyϪ infection in Rab5c-antisense-treated cells revealed no differences in LM hlyϪ destruction (Fig. 3B). These findings suggest that Rab5a is the predominant regulatory Rab GTPase in the phagocytic pathway. Rab5c may play a minor role or function elsewhere.
Rab5a appears to play virtually no role in LM hlyϪ internalization, since transport from the plasma membrane to the phagosomes, as detected by following the internalization of radiolabeled LM hlyϪ in antisense-treated cells (both Rab5a and Rab5c), was unaffected. However, Rab5a is clearly involved in the degradation of pre-internalized dead LM hlyϪ . The percentage of radiolabeled bacteria remaining after a 20-min uptake and 100-min chase in Rab5a-antisense-treated cells was significantly higher (33% higher) than in Rab5a-sense-treated cells or control cells (Fig. 3C).
Overexpression of Rab5a only marginally increased bacterial phagocytosis. However, following phagosome formation, Rab5a appears to play at least two roles: (i) by mediating fusion events within the phagosomal-endosomal compartment and (ii) by facilitating or initiating phagosome maturation culminating in phagosome-lysosome fusion. The former is supported by in vitro reconstitution studies (11) and by the observations presented here that the intermingling of phagosomes and endosomes is reduced in Rab5a-antisense-treated cells. The latter is supported by the observation that fusion of phagosomes, containing dead LM hlyϪ bacteria, with lysosomes was impaired in Rab5a-antisense treated cells. The finding that phagosomes containing dead LM hlyϪ are unable to mature in Rab5a-antisense-treated cells and the observation that live LM hlyϪ phagosomes remain as immature endosomal-like compartments point to key, perhaps common, regulatory steps that are required to initiate phagosome maturation (28).
Phagosome maturation is a complex process, and results reported to date suggest that the nature of the internalized particle plays a role in modulating the rate and perhaps the quality of the process. For example, a recent report using latex beads as model particles in J774 cells demonstrated that phagosomes, which had been internalized for several hours, fused with early and late endosomes in vitro in a Rab5-dependent manner (13). Indeed, earlier work (29) demonstrated that the fusion capacity of phagosomes, containing Staphylococcus A particles internalized via the Fc receptor, is restricted to early endosomes in an in vitro phagosome-endosome fusion. Thus, it is likely that both the receptor that mediates particle internalization and the nature of the internalized particle (e.g. digestible versus nondigestible, live versus dead etc.) play important roles in phagosome maturation and phagosome-lysosome fusion. Interestingly, interferon-␥, a lymphokine known to accelerate intracellular killing of pathogenic LM (LM hlyϩ ) and other intracellular pathogens (30), specifically induces Rab5a biosynthesis and processing (18).

TABLE I
Distribution of bacteria in Rab5a-antisense-treated cells HMØs were treated with PT-oligonucleotides targeted to the AUG codons of Rab5-mRNA (see "Experimental Procedures"). Cells were treated for 4 h with PT-oligonucleotides and then infected with LM hlyϪ as in Fig. 1. At different time points cells were fixed and processed for EM. A total of 200 cells were examined per time point and condition. Results represent the mean of bacteria in cells containing bacteria and the percentage of cells containing 0, 1-2, 3-5, and Ն5 bacteria.