SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Anaplasma phagocytophilum, the causative agent of human granulocytic anaplasmosis, is an obligatory intracellular pathogen. After entry into host cells, the bacterium is diverted from the endosomal pathway and replicates in a membrane-bound compartment devoid of endosomal or lysosomal markers. Here, we show that several hallmarks of early autophagosomes can be identified in A. phagocytophilum replicative inclusions, including a double-lipid bilayer membrane and colocalization with GFP-tagged LC3 and Beclin 1, the human homologues of Saccharomyces cerevisiae autophagy-related proteins Atg8 and Atg6 respectively. While the membrane-associated form of LC3, LC3-II, increased during A. phagocytophilum infection, A. phagocytophilum-containing inclusions enveloped with punctate GFP-LC3 did not colocalize with a lysosomal marker. Stimulation of autophagy by rapamycin favoured A. phagocytophilum infection. Inhibition of the autophagosomal pathway by 3-methyladenine did not inhibit A. phagocytophilum internalization, but reversibly arrested its growth. Although autophagy is considered part of the innate immune system that clears a variety of intracellular pathogens, our study implies that A. phagocytophilum subverts this system to establish itself in an early autophagosome-like compartment segregated from lysosomes to facilitate its proliferation.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Human granulocytic anaplasmosis, an emerging zoonosis in the USA and other parts of the world, is caused by Anaplasma phagocytophilum, a Gram-negative obligate intracellular bacterium in the order Rickettsiales (Goodman et al., 1996; Dumler and Bakken, 1998; Demma et al., 2005). This pathogen infects and replicates in granulocytes and endothelial cells (Goodman et al., 1996; Munderloh et al., 2004; Herron et al., 2005). A key feature in the pathogenesis of granulocytic anaplasmosis is the ability of A. phagocytophilum to avoid being processed by the endocytic machinery of the host cell. After endocytosis, the pH of the membrane-bound vacuole in which A. phagocytophilum resides remains neutral, and this vacuole is not stainable with endosomal or lysosomal markers (Webster et al., 1998; Mott et al., 1999). Webster et al. (1998) reported that some inclusions are mannose 6-phosphate receptor-positive. Mannose 6-phosphate receptor is considered as a late endosome marker, but it may be also present in autophagosomes (Dunn, 1990; Dorn et al., 2002; Eskelinen et al., 2002). In order to replicate within the vacuole, A. phagocytophilum is obliged to usurp and acquire various components from its host cytoplasm, as it has a limited number of genes for de novo amino acid biosynthesis and central intermediary metabolism (Hotopp et al., 2006). Despite the importance of modulation of vesicular traffic and access to host cell metabolites in the course of A. phagocytophilum infection, very little is known about the biogenesis of the A. phagocytophilum replicative vacuole.

Autophagy is an evolutionary conserved and regulated intracellular catabolic mechanism that mediates the degradation of cytosolic components, including proteins, large protein complexes and damaged organelles in a lysosome-dependent fashion (Levine and Klionsky, 2004). The hallmark of autophagy is the formation of a double-membrane cytosolic vesicle, the autophagosome, which sequesters cytoplasm and delivers it to the lysosome for degradation (Levine and Klionsky, 2004). Multiple steps are involved in the formation and maturation of autophagosomes: autophagy induction, vesicle nucleation, vesicle expansion and completion, retrieval of autophagosome-forming proteins for recycling, and fusion with lysosomes to form autolysosomes (Klionsky, 2005; Levine and Yuan, 2005). Many proteins participate in this process. Genetic analysis in yeast has identified 27 autophagy-related (Atg) genes that encode components of the autophagic machinery (Yorimitsu and Klionsky, 2005). Homologues of many of them have also been identified in mammalian cells, such as Atg6 (Beclin 1), Atg7 and Atg8 (microtubule-associated protein light chain 3; LC3) (Klionsky and Emr, 2000). Most of the Atg proteins are soluble and transiently associate with the vesicle, cycling off the vesicle after completion of its function (Klionsky, 2005). Of these proteins, only two, Atg8, when conjugated to phosphatidylethanolamine, and Atg19, are known to remain associated with the autophagosomes in yeast (Kabeya et al., 2000; Scott et al., 2001).

Autophagy serves to help clear intracellular infection and process non-self- and self-antigens in the host cytoplasm as part of the innate and adaptive immune system (Amano et al., 2006; Ling et al., 2006; Schmid et al., 2006). Autophagosome formation facilitates killing of several intracellular bacteria, including Rickettsia conorii, Mycobacterium tuberculosis and Shigella flexneri (Walker et al., 1997; Gutierrez et al., 2004; Ogawa and Sasakawa, 2006). However, diversion of phagosomes containing bacteria to autophagosomes favours survival of other intracellular bacteria, such as Legionella pneumophila and Coxiella burnetii, perhaps by delaying the fusion of pathogen-containing vacuoles with lysosomes, thereby giving them time to develop into a more acid- and protease-resistant stage (Kirkegaard et al., 2004; Amer et al., 2005; Colombo, 2005; Gutierrez et al., 2005). For Francisella tularensis, autophagy induction helps the bacterium to re-enter a membrane-bound compartment after replication in the cytoplasm (Checroun et al., 2006). However, the role of autophagosomes in A. phagocytophilum infection is unknown.

Here, we explore the possibility of the involvement of autophagosomes in A. phagocytophilum infection. Our results demonstrate a new cellular process by which this intracellular bacterium creates a replicative compartment which does not mature into an autolysosome.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Anaplasma phagocytophilum inclusions are enveloped by double-lipid bilayer membranes

The hallmark of early autophagy is the formation of double-lipid bilayer membrane autophagosomes (Kirkegaard et al., 2004). After fusion with lysosomes, autolysosomes have a single-lipid bilayer membrane because the inner membrane disintegrates (Kirkegaard et al., 2004). Transmission electron microscopy (TEM) revealed that many of inclusions containing replicating A. phagocytophilum in human myelocytic HL-60 cells are at least partially enveloped with double-lipid bilayer membranes (Fig. 1A and B), indicating that the inclusions have a defining property of autophagosome. The percentage of inclusions with double-lipid bilayer membrane increased from 50% at 40 h post infection (p.i.) to 78% at 60 h p.i. (Fig. 1D). As a control, the membrane of phagosomes harbouring Escherichia coli in HL-60 cells was a single-lipid bilayer (Fig. 1C).

image

Figure 1. Transmission electron micrograph of an A. phagocytophilum replicative inclusion. A. An A. phagocytophilum replicative inclusion in HL-60 cells at 40 h p.i. Double-lipid bilayer is highlighted with arrows. B. An enlarged part of a replicative inclusion. Double-lipid bilayer is indicated with arrows. C. A phagosome containing E. coli in HL-60 cells. Scale bar = 0.25 μm in A and C, and 0.125 μm in B. D. Percentages of replicative inclusions containing double-lipid bilayer membranes at three time points post infection. One hundred inclusions were scored at each time point.

Download figure to PowerPoint

Microtubule-associated protein light chain 3, an essential component of cellular autophagosomes, colocalizes with A. phagocytophilum replicative inclusions

Upon induction of autophagy, cytosolic LC3, a protein essential for autophagosome formation, is conjugated to phosphatidylethanolamine through a C-terminal glycine and associates with the membrane (Ichimura et al., 2000; Kabeya et al., 2000). To determine whether A. phagocytophilum replicative inclusions are autophagosomes, we examined the localization of LC3 throughout the time-course of infection. A plasmid encoding GFP-tagged LC3 has been developed and widely used to study LC3 processing and distribution, as the cellular amount of native LC3 is too low (Kabeya et al., 2000). Because HL-60 cells are difficult to transfect, and A. phagocytophilum infection progresses in a similar fashion in both human HL-60 cells and monkey endothelial RF/6A cells (Munderloh et al., 2004), RF/6A cells were used for the study of LC3. RF/6A cells were transfected either with a plasmid encoding GFP-LC3 or with a plasmid encoding the non-functional mutant GFP-LC3ΔC22,G120A (GFP-LC3ΔC22), which is unable to be lipidated at the C-terminus as a control. By fluorescence microscopy, a cytosolic LC3 has a diffuse distribution pattern, whereas membrane-associated LC3 has a punctate pattern corresponding to nascent autophagosomes (Mizushima et al., 2004). Uninfected transfected RF/6A cells showed predominantly diffuse GFP-LC3 fluorescence distribution throughout the cytoplasm, indicating little autophagosome formation (Fig. 2G). In contrast, during infection, GFP-LC3 became increasingly punctated, and most of the punctate GFP-LC3 was associated with A. phagocytophilum inclusions at 48 and 72 h p.i. (Fig. 2B, C and H). Especially at 48 h p.i., A. phagocytophilum inclusions were individually encased by the punctuate GFP-LC3 green circle (Fig. 2B). Neither GFP-LC3 punctation nor colocalization with the inclusion was evident at 14 h p.i. (Fig. 2A and H). GFP-LC3ΔC22 did not make the punctate pattern during infection and was not specifically concentrated on A. phagocytophilum-containing vacuoles, although it was squeezed in between growing inclusions, and thus appeared colocalized with some inclusions at 48 and 72 h p.i. (Fig. 2D–F). Infectivity did not differ between RF/6A cells transfected with the wild-type or mutant LC3 constructs.

image

Figure 2. Colocalization of A. phagocytophilum inclusions with GFP-LC3 and conversion of LC3-I to LC3-II. Immunofluorescence micrographs of RF/6A cells transfected with GFP-LC3 (A, B and C) or GFP-LC3ΔC22,G120A (D, E and F) and infected with A. phagocytophilum for 14 h (A and D), 48 h (B and E) and 72 h (C and F). Infected cells were fixed and stained with horse anti-A. phagocytophilum and Cy3-conjugated goat anti-horse IgG. Scale bar = 10 μm. Insets are images taken under a confocal microscope. Typical A. phagocytophilum inclusions were circled with dashed white line. G. Uninfected RF/6A cells transfected with GFP-LC3 as a control. Scale bar = 10 μm. H. The percentage of A. phagocytophilum inclusions that were colocalized with punctate GFP in RF/6A cells transfected with GFP-LC3 or GFP-LC3ΔC22,G120A at three time points post infection. One hundred GFP-positive and infected RF/6A cells in each group were scored, and the percentage of punctate GFP colocalization with A. phagocytophilum inclusions was determined. Data are presented as the means and standard deviations of triplicate samples. The asterisk indicates a significant difference compared with GFP-LC3ΔC22,G120A by Student's t-test (P < 0.01) within each group at each time point. I. A. phagocytophilum infection induces LC3-I processing to LC3-II. GFP-LC3-transfected RF/6A cells were infected with A. phagocytophilum. Cells were harvested at 2 and 3 days p.i. for Western blot analysis using mouse monoclonal anti-GFP. Uninfected RF/6A cells transfected with GFP-LC3 were used as a control. The data are the representative of three independent experiments. J. A. phagocytophilum growth curve in transfected RF/6A cells as determined by real-time PCR. The genomic DNA extracted from infected RF/6A transfected with GFP-LC3 plasmid during different time points post infection, was subjected to real-time PCR analysis. Data are presented as the means and standard deviations of triplicate samples. K. Lack of fusion of A. phagocytophilum autophagosomes with lysosomes. Inclusions containing DAPI-stained A. phagocytophilum were surrounded with GFP-LC3 in the punctate pattern at 48 h p.i. Lysosomes were labelled with PE-conjugated anti-LAMP-3 antibody. The nucleus of RF/6A cell was marked ‘N’ and with dashed lines. Scale bar = 10 μm.

Download figure to PowerPoint

A cytosolic LC3 form is called LC3-I, and its membrane-associated lipidated form is called LC3-II (Kabeya et al., 2000). By SDS-PAGE followed by immunoblotting, GFP-LC3-II with an apparent molecular mass (Mr) of 43 kDa can be distinguished from GFP-LC3-I with the Mr of 45 kDa. To confirm modification of LC3-I to LC3-II during A. phagocytophilum infection, Western blot analysis was performed on A. phagocytophilum-infected RF/6A cells transfected with GFP-LC3. A weak GFP-LC3-II band was detected in A. phagocytophilum-infected RF/6A cells at 2 days p.i., and the amount of GFP-LC3-II increased at 3 days p.i. (Fig. 2I). The ratios of GFP-LC3-II to GFP-LC3-I were 1:33 at 2 days, and 1:10 at 3 days p.i. Thus, A. phagocytophilum infection induced the conversion of LC3-I to LC3-II, a signature of autophagosome formation (Kabeya et al., 2000).

To examine whether autophagosome formation in RF/6A cells was positively associated with bacterial load, real-time PCR analysis targeting A. phagocytophilum pleD (a single-copy gene) was performed to determine bacterial numbers in the infected cells. During the lag phase (up to 24 h p.i.) of A. phagocytophilum growth, autophagosome formation was undetectable, whereas, autophagosomes were formed during the rapid growth (from 24 to 72 h p.i.) (Fig. 2J). Thus, the increasing bacterial load was associated with the induction of autophagosome formation, suggesting its induction in response to infection.

Because autophagy formation leads to destruction of several bacterial pathogens in autolysosomes, we examined whether A. phagocytophilum inclusions surrounded with punctate LC3 were undergoing lysosomal fusion by triple fluorescence labelling. These A. phagocytophilum inclusions did not fuse with lysosomes as revealed by the absence of LAMP-3 (lysosome-associated membrane protein 3), a late endosomal and lysosomal marker (Fig. 2K). Furthermore, the A. phagocytophilum inclusions surrounded with GFP-LC3-II increased in number and in size during logarithmic growth (Fig. 2H). Taken together, these results imply that A. phagocytophilum infection induces autophagy and A. phagocytophilum in the autophagosome is not on the way to destruction.

Beclin 1, another essential component of cellular autophagosomes, colocalizes with A. phagocytophilum replicative inclusions

Beclin 1, a tumour suppressor, forms a complex with the mammalian class III phosphatidylinositol 3-kinase, Vps34, that is essential for induction of autophagosome formation (Liang et al., 1999). Beclin 1, which is not required for the conventional endosomal/lysosomal pathway (Zeng et al., 2006), is only present on autophagosomes during formation, as it dissociates from the membrane prior to autophagosome maturation (Klionsky, 2005). To further assess the ability of A. phagocytophilum to induce autophagosomes and whether A. phagocytophilum inclusions are autophagosomes, Beclin 1 localization was examined in infected HL-60 cells by confocal double immunofluorescence labelling using mouse monoclonal anti-human Beclin 1 and horse anti-A. phagocytophilum antiserum. The horse antibody does not react with uninfected HL-60 cells, as shown in previous studies (Niu et al., 2006), and the normal horse serum control and mouse isotype control did not label infected or uninfected HL-60 cells (data not shown). Beclin 1 did not colocalize with the internalized individual bacteria at 3 h p.i., or inclusions of the early replication stage at 20 h p.i. (Fig. 3A, B and E), but did colocalize with approximately 15% of A. phagocytophilum replicative inclusions at 32 h p.i. and with 5% A. phagocytophilum replicative inclusions at 50 h p.i. (Fig. 3C–E). Figure S1 shows a conventional fluorescence microscopy image of Beclin 1 colocalization. This result corroborates with the GFP-LC3 and TEM results that A. phagocytophilum induces autophagosome formation and that the replicative inclusion has the properties of an early autophagosome.

image

Figure 3. Colocalization of A. phagocytophilum inclusions with Beclin 1 by confocal double immunofluorescence microscopy. A. phagocytophilum-infected HL-60 cells harvested at 3 h (A), 20 h (B), 32 h (C) and 50 h (D) p.i. were double immunofluorescence labelled with horse anti-A. phagocytophilum (Anaplasma) (FITC, green) and mouse monoclonal anti-Beclin 1 (Cy3, red). Scale bar = 5 μm. E. The percentage of A. phagocytophilum inclusions that colocalized with Beclin 1. The percentage of colocalization was scored in 100 inclusions. Data are presented as the means and standard deviations of triplicate samples. The asterisk and double asterisks indicate a significant difference compared with the sample at 3 h p.i. by anova and Tukey honestly significant differences (HSD) (P < 0.05).

Download figure to PowerPoint

Influence of an inhibitor and a stimulator of cellular autophagy on A. phagocytophilum infection

Because A. phagocytophilum infection induces autophagosome formation, but lysosomal fusion with the autophagosome containing A. phagocytophilum was not detected, we further examined whether autophagy hinders or facilitates A. phagocytophilum replication. 3-methyladenine (3-MA) is a widely used pharmacological inhibitor of autophagy, which inhibits the activity of class III phosphatidylinositol 3-kinase (Seglen and Gordon, 1982; Lindmo and Stenmark, 2006). A dose–response study showed that 2 mM 3-MA significantly inhibited A. phagocytophilum infection, and 3-MA at 5 and 10 mM completely blocked A. phagocytophilum infection in HL-60 cells (Fig. 4A). 3-MA at 10 mM concentration was widely used for inhibition of autophagy in many types of cells (Amer et al., 2005). At 10 mM, 3-MA was not toxic to the HL-60 cells, as no morphologic and viability changes were seen in HL-60 cells and cells treated at this concentration maintained the ability to replicate at a rate similar to non-treated cells (data not shown). When A. phagocytophilum was pretreated with 3-MA and then incubated with HL-60 cells in the absence of 3-MA, there was no inhibitory effect on its infection of HL-60 cells: with 3-MA pretreated A. phagocytophilum % infected cells was 80 ± 5%, and 82 ± 6% with untreated A. phagocytophilum control at 2 days p.i. (n = 3 independent experiments). Thus 3-MA likely inhibited A. phagocytophilum infection by preventing cellular autophagy induction, rather than by a direct toxic effect on the bacteria or host cells. This result also implies that autophagy is required for A. phagocytophilum infection. To further investigate the effect of 3-MA on the binding, internalization or growth of A. phagocytophilum, 3-MA was added to infected HL-60 cells at different time p.i. When 2 mM 3-MA was added to HL-60 cells at 5, 10, or 20 h p.i., A. phagocytophilum infection was significantly inhibited as determined at 2 days p.i. (Fig. 4B). 3-MA did not inhibit binding or internalization (Fig. 5B and E). These results suggest that internalized bacteria could not replicate when autophagy was blocked by 3-MA. In fact, A. phagocytophilum inclusions did not enlarge in 3-MA-treated cells, but bacterial growth and expansion of the inclusion size resumed after removal of 3-MA from the culture medium (Fig. 5G–L). Furthermore, LAMP-1 did not colocalize with growth-arrested inclusions in 3-MA-treated cells (Fig. 5D and F). Taken together, these results support autophagy facilitates the A. phagocytophilum replication.

image

Figure 4. 3-MA inhibits A. phagocytophilum replication. A. Dose–response of 3-MA on A. phagocytophilum infection. HL-60 cells pretreated with the indicated concentrations of 3-MA were infected with A. phagocytophilum, and the bacteria were allowed to replicate in the presence of 3-MA until 2 days p.i. The percentage of infected HL-60 cells was determined based on the presence of inclusions in 100 HL-60 cells. Data are presented as the means and standard deviations of triplicate samples. The asterisk indicates a significant difference compared with no 3-MA control by the Tukey HSD test (P < 0.01). B. Temporal effect of 3-MA addition on A. phagocytophilum infection. HL-60 cells were infected with A. phagocytophilum for 5, 10 and 20 h. 3-MA was added to the culture at a final concentration of 2 mM. The cultures were allowed to further incubate until 2 days p.i. Data are presented as the means and standard deviations of triplicate samples. The asterisk indicates a significant difference compared with the RPMI control by the Tukey HSD test (P < 0.01).

Download figure to PowerPoint

image

Figure 5. 3-MA neither alters internalization of A. phagocytophilum nor induces the colocalization of A. phagocytophilum with LAMP-1, but reversibly inhibits A. phagocytophilum growth. Immunofluorescence labelling was performed to examine the internalization of A. phagocytophilum into HL-60 cells in the control medium (RPMI) (A) or in the presence of 3-MA (B) at 2 h p.i. Extracellular bacteria were stained prior to permeabilization, with horse anti-A. phagocytophilum and Cy3 (red)-conjugated goat anti-horse IgG. Total bacteria were stained after permeabilization with saponin, with horse anti-A. phagocytophilum and FITC (green)-conjugated goat anti-horse IgG. Intracellular bacteria were indicated with white arrows. Scale bar = 5 μm. Lack of colocalization of A. phagocytophilum with LAMP-1 after treatment with 3-MA as determined by confocal double immunofluorescence microscopy. A. phagocytophilum-infected HL-60 cells treated with RPMI medium (C) or with 3-MA (D) were labelled with two primary antibodies (horse anti-A. phagocytophilum, and mouse monoclonal anti-LAMP-1) and two secondary antibodies (FITC-conjugated goat anti-horse IgG, and Cy3-conjugated goat anti-mouse IgG), and observed under confocal immunofluorescence microscopy. Scale bar = 5 μm. E. The percentage of intracellular A. phagocytophilum at 2 h p.i. One hundred A. phagocytophilum in each group were scored, and the percentage of intracellular A. phagocytophilum was determined. Data are presented as the means and standard deviations of triplicate samples. There is no significant difference between groups treated with 3-MA or RPMI medium control by Student's t-test (P > 0.05). F. The percentage of A. phagocytophilum inclusions that did not colocalize with LAMP-1. One hundred A. phagocytophilum inclusions in each group were scored, and the percentage of A. phagocytophilum inclusions which did not colocalize with LAMP-1 was determined. Data are presented as the means and standard deviations of triplicate samples. There is no significant difference between groups treated with 3-MA or RPMI medium control by Student's t-test (P > 0.05). G–L. The inhibition of the replication of A. phagocytophilum by 3-MA is reversible. G. A. phagocytophilum in HL-60 cells at 20 h p.i., without 3-MA treatment. H. A. phagocytophilum in HL-60 cells at 25 h p.i., without 3-MA treatment. I. A. phagocytophilum in HL-60 cells at 44 h p.i., without 3-MA treatment. J. A. phagocytophilum in HL-60 cells was incubated for 20 h without 3-MA, then incubated for 5 h in the presence of 10 mM 3-MA. K. A. phagocytophilum in HL-60 cells was incubated for 20 h without 3-MA, then incubated for 5 h in the presence of 10 mM 3-MA, and finally 3-MA was removed and continuously incubated for additional 19 h. L. A. phagocytophilum in HL-60 cells was incubated for 20 h without 3-MA, then incubated for 24 h in the presence of 10 mM 3-MA. Arrows indicate A. phagocytophilum inclusions. Scale bar = 5 μm.

Download figure to PowerPoint

In yeast, target of rapamycin (TOR), a serine/threonine phosphatidylinositol 3-kinase-related kinase, controls autophagy as a negative regulator of Atg1 (Noda and Ohsumi, 1998). Inhibition of mammalian TOR (mTOR) by rapamycin stimulates autophagy (Klionsky and Emr, 2000). Therefore, we tested whether TOR was involved in A. phagocytophilum infection, and found that infection by the bacteria increased approximately twofold when cells were pretreated with 50 ng ml−1 rapamycin (Fig. 6A). At this concentration, rapamycin was not toxic to HL-60 cells (data not shown). The increase in infectivity by rapamycin treatment was also supported by data from Western blot analysis (Fig. 6B). Compared with the DMSO-solvent control, the band density of the major outer-membrane protein of A. phagocytophilum, P44, as detected with the P44-specific monoclonal antibody 5C11 (Kim and Rikihisa, 1998), increased by 2.2-fold in rapamycin-treated cells at 2 days p.i. after normalization to the α-tubulin band density. Infection rates of control cultures with or without DMSO were almost identical (data not shown). Therefore, during A. phagocytophilum infection, the activity of the autophagic pathway correlates with productive A. phagocytophilum replication, not with A. phagocytophilum destruction.

image

Figure 6. Rapamycin enhances A. phagocytophilum infection. A. Rapamycin enhanced A. phagocytophilum infection in HL-60 cells. HL-60 cells pretreated with rapamycin at 50 ng ml−1 for 3 h were infected with A. phagocytophilum in the continued presence of rapamycin until 2 days p.i. Data are presented as the means and standard deviations of triplicate samples. The asterisk indicates a significant difference compared with the DMSO-solvent alone control by Student's t-test (P < 0.01). B. Western blot analysis of A. phagocytophilum infection in rapamycin-treated cells. HL-60 cells pretreated with rapamycin at 50 ng ml−1 for 3 h were infected with A. phagocytophilum in the continued presence of rapamycin until 2 days p.i. The infected cells (4.5 × 105) were used for Western blot analysis with mouse monoclonal anti-P44. The loading amount was normalized using α-tubulin. Results shown are representative of three independent experiments.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

In this work, we have shown that several hallmarks of autophagosomes, including a double-lipid bilayer membrane, and colocalization with GFP-tagged LC3 and Beclin 1 can be observed in A. phagocytophilum inclusions. Does the autophagosome formation help A. phagocytophilum infection, or is it a part of anti-rickettsial response? Based on observation of the absence of lysosome fusion with A. phagocytophilum inclusions surrounded with GFP-LC3-II, increase in the number and the size of GFP-LC3-II-labelled inclusions during rapid bacterial growth, and the effects of pharmacologically disrupted or induced autophagy on infection of A. phagocytophilum, we argue that the autophagosome formation favours A. phagocytophilum infection. Taken together with previous studies that showed an absence of endosomal and lysosomal markers, and lack of acidification in A. phagocytophilum inclusions (Webster et al., 1998; Mott et al., 1999), the present study indicates that the maturation of autophagosomes harbouring A. phagocytophilum to late autophagosomes and the fusion with lysosomes are arrested. These early autophagosomes, therefore, can provide a safe haven for intracellular A. phagocytophilum to survive and replicate. This mechanism is distinct from other intracellular bacteria for which autophagosome induction favours their infection (Colombo, 2005), as these pathogens replicate in acidic LAMP-1-positive late autophagosomes (Porphyromonas gingivalis and Brucella abortus) or autolysosomes that are positive for the lysosomal marker cathepsin D (L. pneumophila, C. burnetii and F. tularensis) (Pizarro-Cerda et al., 1998; Sturgill-Koszycki and Swanson, 2000; Dorn et al., 2001; Checroun et al., 2006; Voth and Heinzen, 2007).

Anaplasma phagocytophilum is the first bacterium colocalized with Beclin 1, which acts favourably for its growth. There is a limited body of literature describing percentage of Beclin 1 colocalization with intracellular pathogens. Induction of autophagy inhibits mycobacterial survival in infected macrophages. Upon induction of autophagy by starvation, 51.4 ± 8.3% of Mycobacterium-containing phagosome colocalizes with LysoTracker, versus 26.5 ± 6.3% in the full nutrient condition control samples. Upon induction of autophagy, colocalization of Beclin 1 with Mycobacterium-containing phagosome increases from 37.6% to 57.9% (Gutierrez et al., 2004). In this paper, macrophage transiently transfected with FLAG epitope-tagged human Beclin 1 was used. The 37.6% of Beclin 1 colocalization under non-starved condition is higher than the result of our study, which used the specific antibody to detect native Beclin 1. The different colocalization rates between the studies of Gutierrez et al. (2004) and ours may be in part due to the technique used. Liang et al. (1999) showed that Beclin 1 gene transfer in MCF7 cells increases the basal level of the autophagy. In addition, compared with LC3 colocalization, the lower percentage of colocalization of Beclin 1 with replicative inclusions of A. phagocytophilum may be due to the transient nature of the interaction of Beclin 1 with the autophagosome (Klionsky, 2005). Approximately 15% and 5% A. phagocytophilum inclusion was Beclin 1-positive at 32 and 50 h p.i. respectively. The decreasing colocalization of Beclin 1 may also be due to its transient nature. The similar phenomenon was observed for Atg7, another transient component in autophagosome formation, in L. pneumophila-infected macrophage. About 50% of phagosome containing L. pneumophila was Atg7-positive immediately after internalization. However, Atg7 was rarely detected near vacuoles at 3 h p.i. (Amer and Swanson, 2005).

The timing of autophagy in A. phagocytophilum inclusion biogenesis is unique compared with other intracellular bacteria. Vacuoles containing L. pneumophila colocalize with LC3 by 4 h p.i. (Amer et al., 2005). C. burnetii inclusions acquire LC3 at 5 min p.i. (Romano et al., 2007). In contrast, A. phagocytophilum inclusions did not colocalize with LC3 and Beclin 1 at 14 and 20 h p.i. respectively. This relatively slow induction of autophagy during exponential growth is in agreement with the fact that 3-MA, which inhibits autophagy at the induction stage, was effective in inhibiting A. phagocytophilum infection, even when added at 20 h p.i. Autophagy is mainly regulated at the post-translation level (Klionsky, 2005). In fact, the amount of Beclin 1 is not different between uninfected and A. phagocytophilum-infected HL-60 cells (unpubl. data). This result is consistent with the finding by de la Fuente et al. (2005) that mRNAs of Beclin 1 and LC3 are unchanged in A. phagocytophilum-infected HL-60 cells.

The present data led to a new question: Are autophagosomes recruited to the A. phagocytophilum-containing vacuole, or some other mechanism leads to recruitment of autophagosome-inducing proteins or autophagosome components to generate an early autophagosome enclosing the bacteria? Punctate staining of LC3-II is currently viewed as a definitive marker for activation of the autophagic pathway (Kirkegaard et al., 2004). Because numerous LC3- and Beclin 1-positive punctate structures were observed in the infected host cytoplasm, we propose that small autophagosomes are induced by A. phagocytophilum and recruited to the vacuole enclosing the bacteria. Another fundamental question is: Is autophagy a host response specific to a bacterial product or a host starvation response due to deprivation of nutrients by A. phagocytophilum? In this regard, recently several bacterial products have been shown to induce autophagosome formation, such as Vibrio cholerae cytolysin, SipB, an effector protein of the type III secretion system of Salmonella enterica, and VirG of S. flexneri (Hernandez et al., 2003; Ogawa et al., 2005; Gutierrez et al., 2007). Induction of autophagy in L. pneumophila-infected macrophages also depends on its type IV secretion system (Amer and Swanson, 2005). A. phagocytophilum has a functional type IV secretion system that can deliver protein substrates into the host cytoplasm (Niu et al., 2006; Lin et al., 2007). Thus, we anticipate that A. phagocytophilum might also induce autophagy by secreted products.

Our study showed that induction of autophagy with rapamycin facilitates A. phagocytophilum infection, suggesting that autophagy not only shields the bacterium from endosomal and lysosomal pathway, but also enhances replication of the bacterium. In fact, 3-MA did not inhibit binding or internalization, but internalized bacteria could not replicate. The Lamp-1 did not colocalize with A. phagocytophilum inclusions in 3-MA-treated cells. The lack of colocalization may be due to the incomplete inhibition of autophagosome formation by 3-MA. 3-MA at 5 mM suppresses the autophagosome formation by 80% in hepatocytes (Seglen and Gordon, 1982). The remaining activity of autophagosome formation may be sufficient for A. phagocytophilum survival and preventing lysosomal fusion, but may be insufficient for its replication. The result of reversibe inhibition of A. phagocytophilum replication by 3-MA supports this assumption. A. phagocytophilum has a limited number of genes for biosynthesis and central intermediary metabolism (Hotopp et al., 2006). Because autophagy engulfs cytosolic components (Klionsky and Emr, 2000), we speculate that the autophagosome could provide A. phagocytophilum with the direct access to host cytosolic nutrients without the need for transport across the inclusion membrane.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Cell culture and TEM

Anaplasma phagocytophilum HZ strain was propagated in HL-60 cells (ATCC, Manassas, VA) in RPMI 1640 medium supplemented with 10% fetal bovine serum and 2 mM l-glutamine. Monkey endothelial cell line RF/6A (ATCC) was cultured in advanced MEM (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum and 2 mM l-glutamine for infection with A. phagocytophilum. Cultures were incubated at 37°C in a humidified 5% CO2/95% air atmosphere. The degree of bacterial infection in host cells was assessed by Diff-Quik staining (Baxter Scientific Products, Obetz, OH) of cytocentrifuged preparations. No antibiotic was used throughout the study. Host cell-free A. phagocytophilum was prepared by sonication, as described elsewhere (Mott et al., 2002).

Anaplasma phagocytophilum-infected HL-60 cells were processed for TEM as described (Rikihisa et al., 1991). As a control, the pellet of 1 ml of overnight culture of E. coli DH5α was incubated with 1 × 107 HL-60 cells for 2 h, and processed for TEM as A. phagocytophilum-infected HL-60 cells.

LC3 localization and conversion

GFP-LC3 and GFP-LC3ΔC22,G120A plasmids were kindly provided by Dr Tamotsu Yoshimori at The Research Institute for Microbial Diseases, Osaka University, Japan, through Dr Jean Celli at the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, MT. Plasmids were purified using the EndoFree Plasmid Maxi Kit (Qiagen, Valencia, CA) according to the manufacturer's instruction. Transfection of RF/6A endothelial cells with these plasmids was performed using electroporation. Briefly, RF/6A cells were washed once using PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 2 mM KH2PO4, pH 7.4) and resuspended in RPMI 1640 medium (no serum) at a final density of 2 × 107 cells ml−1. RF/6A cells (80 μl) were mixed with 5 μg plasmid and subjected to a pulse from a Gene Pulser Xcell System (Bio-Rad, Hercules, CA) in a 0.2 cm cuvette. The setting was 100 V, 1000 μF, which resulted in a pulse time of 50 ms. For localization of GFP-LC3, 1.6 × 106 RF/6A cells transfected with GFP-LC3 or GFP-LC3ΔC22 plasmid were cultured on coverslips in a 6-well plate for 1 day post transfection, and host cell-free A. phagocytophilum in 2 ml of advanced MEM with serum was added at the approximate ratio of host cell to bacteria of 1:100. Infected cells were washed twice using advanced MEM with serum (3 ml each time) after 4 h incubation and continued to incubate for the designated time periods. Coverslips were washed three times with PBS, fixed with 2% paraformaldehyde and subjected to immunofluorescence labelling after saponin permeabilization using horse anti-A. phagocytophilum (1:200 dilution) as primary antibody and Cy3-conjugated goat anti-horse IgG (diluted 1:100, Jackson ImmunoResearch Laboratories) as secondary antibody as described (Niu et al., 2006). For triple labelling, GFP-LC3-transfected RF/6A cells were infected with A. phagocytophilum for 48 h and fixed with 2% paraformaldehyde on the coverslip. Cells were subjected to immunofluorescence labelling after permeabilization with cold methanol for 3 min, using mouse monoclonal anti-LAMP-3 conjugated with PE (BioLegend, San Diego, CA). Immediately prior to viewing under a fluorescence microscope, the cells were counterstained for the nucleus of RF/6A cells and A. phagocytophilum with 300 nM 4′,6-diamidino-2-phenylindole, dilactate (DAPI, Molecular Probes) for 5 min. Fluorescence images were analysed by a Nikon Eclipse E400 fluorescence microscope with a xenon-mercury light source (Nikon Instruments, Melville, NY) or LSM 510 laser-scanning confocal microscope (Carl Zeiss, Thornwood, NY). The original colour emitted by excited DAPI (blue) was transformed to grey pseudocolour for clear viewing with Photoshop 7.0 software (Adobe, San Jose, CA).

Real-time PCR analysis

The RF/6A cells were transfected with GFP-LC3 plasmid, and infected with A. phagocytophilum as described above. The cells were harvested with the cell scraper after incubation for designated time periods, and homogeneously dispersed by pipetting using 1 ml pipette tip for 20 times. The pellets from 1 ml culture of each sample were subjected to genomic DNA extraction (QIAamp DNA Blood Mini Kit, Qiagen) after centrifugation at 12 000 g, 4°C for 10 min. The genomic DNA was eluted in 50 μl of elution buffer (10 mM Tris-HCl, 0.5 mM EDTA, pH 9.0). pleD copy number in each sample was quantified by a real-time, quantitative PCR instrument, the Stratagene Mx3000P (Stratagene, La Jolla, CA). The quantitative PCR was performed in the presence of SYBR Green, using a pair of primer (forward primer: 5′-ACATGCGTACAAACCCTGCCATTG-3′; and reverse primer: 5′-AAATCATCTGCACCAGCACTCAGC-3′). The data analysis was performed using the integrated analysis software (Stratagene). A standard curve was made by amplification of a serial 10-fold dilution of pET-33b (+) plasmid containing pleD (T.-H. Lai and Y. Rikihisa, unpubli. data).

Western blot analysis

To examine LC3 forms I and II, protein was extracted from A. phagocytophilum-infected or uninfected GFP-LC3-transfected RF/6A cells using cold lysis buffer [20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% (v/v) Triton X-100] containing protease inhibitors (Cocktail Set III, 1:50 dilution, EMD Chemicals, San Diego, CA). Western blot analysis was performed using anti-GFP (Santa Cruz Biotechnology, Clone B-2, diluted 1:1000), and peroxidase-conjugated goat anti-mouse IgG (KPL, Gaithersburg, MD, diluted in 1:1000). The bound antibody was detected by ECL chemiluminescence (Pierce Biotechnology, Rockford, IL). The images were captured by a CCD camera (Fuji LAS-3000 imaging system), and band density was measured by Fujifilm MultiGauge software.

Beclin 1 localization

HL-60 cells (1.5 × 107) were incubated with A. phagocytophilum derived from 2 × 107 infected HL-60 cells in 15 ml of RPMI medium with serum (the approximate ratio of host cell to bacteria was 1:100). After 2 h of incubation at 37°C, infected cells were washed twice with PBS to remove bacteria not associated with host cells, and then incubated at 37°C to allow A. phagocytophilum to replicate for the designated time periods. A. phagocytophilum-infected HL-60 cells were fixed with 2% paraformaldehyde and subjected to double immunofluorescence labelling after saponin permeabilization. Horse anti-A. phagocytophilum or preimmune horse serum (1:200 dilution) and mouse monoclonal anti-Beclin 1 (Santa Cruz Biotechnology, clone E-8, 1:50 dilution) were used as primary antibodies, and FITC-conjugated goat anti-horse IgG (diluted 1:50, Jackson ImmunoResearch) and Cy3-conjugated goat anti-mouse IgG (diluted 1:100, Jackson ImmunoResearch) were used as secondary antibodies. An unrelated mouse monoclonal antibody of the same isotype (IgG2a) as anti-Beclin 1 was used as a negative control.

3-MA and rapamycin treatment

For the study of the dose–response effect of 3-MA, HL-60 cells (1.5 × 105 per well) in 12-well plates were incubated with host cell-free A. phagocytophilum at the approximate ratio of host cell to bacteria of 1:100 in 1 ml of RPMI 1640 medium with serum containing the indicated concentration of 3-MA. 3-MA (Sigma, St Louis, MO) was prepared from a 20 mM stock solution in RPMI 1640 medium, and added to the HL-60 cell culture at 2 h preinfection. For the time-course study, 3-MA at a final concentration of 2 mM was added to A. phagocytophilum-infected HL-60 cells at 5, 10 and 20 h p.i., and A. phagocytophilum infection was determined 2 days p.i. To study the effect of rapamycin, 1.5 × 105 HL-60 cells at the logarithmic stage of growth were pre-incubated with 50 ng ml−1 of rapamycin (Sigma) for 3 h, and then host cell-free A. phagocytophilum was added at the approximate ratio of host cell to bacteria of 1:2. Rapamycin was added from a stock solution of 50 μg ml−1 in DMSO. At 2 days p.i., the degree of bacterial infection in host cells was assessed by Diff-Quik staining and by Western blot analysis using anti-P44 (mAb5C11, diluted at 1:2000) and anti-α-tubulin (Santa Cruz Biotechnology, clone B-7, diluted at 1:2000). To determine the toxicity of 3-MA and rapamycin to HL-60 cells, 1.0 × 106 HL-60 cells in 2 ml complete RPMI 1640 medium were incubated at 37°C for 2 days in the presence of 10 mM 3-MA or 50 ng ml−1 of rapamycin. The cell number was scored, morphology was observed by phase-contrast microscopy, and the cell viability was determined with the trypan blue dye exclusion test (Altman et al., 1993).

To determine the toxicity of 3-MA, host cell-free A. phagocytophilum from 1.0 × 107 infected HL-60 cells was first incubated with 10 mM 3-MA or culture medium for 1.5 h on ice. After washing to remove 3-MA, A. phagocytophilum was incubated with 1.0 × 106 HL-60 cells in 2 ml complete RPMI 1640 medium. Bacterial infection was determined at 48 h p.i. by Diff-Quik staining. For determination of the effect of 3-MA on the A. phagocytophilum internalization into HL-60 cells, host cell-free A. phagocytophilum from 1.0 × 107 infected HL-60 cells was incubated with 1.0 × 106 HL-60 cells for 2 h in the presence of 10 mM 3-MA. The cells were washed to remove unbound bacteria, and fixed with 2% paraformaldehyde. The extracellular and intracellular bacteria were distinguished by immunostaining using horse anti-A. phagocytophilum and Cy3-conjugated goat anti-horse IgG before permeabilization, and horse anti-A. phagocytophilum and FITC-conjugated goat anti-horse IgG after permeabilization (Niu et al., 2006). To determine whether A. phagocytophilum colocalizes with LAMP-1 after the treatment of 3-MA, HL-60 cells infected with host cell-free A. phagocytophilum were incubated for 20 h to allow the appearance of small inclusions. 3-MA was added into cell culture to final concentration of 10 mM, and continued to incubate for additional 5 h. HL-60 cells were fixed and immunostained with horse anti-A. phagocytophilum and mouse monoclonal anti-LAMP-1 (1D4B; Developmental Hybridoma Bank), and the appropriate secondary fluorochrome-conjugated antibody and observed under a LSM 510 laser-scanning confocal microscope. To determine reversibility of 3-MA inhibition of the A. phagocytophilum replication, A. phagocytophilum-infected HL-60 cells were first incubated at 37°C for 20 h, then a portion of cells were incubated in the presence of 10 mM 3-MA for 5 or 24 h. The cells which were treated with 3-MA for 5 h, were washed to remove 3-MA, and continued to incubate for additional 19 h. The A. phagocytophilum inclusions in HL-60 cells were observed under a light microscope after Diff-Quik staining.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was supported by Grants R01 AI054476 and R01 AI30010 from the National Institutes of Health. We thank Dr Tamotsu Yoshimori at The Research Institute for Microbial Diseases, Osaka University, Japan, and Dr Jean Celli at the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, MT, USA for sharing with us GFP-LC3 and its mutant plasmids. We also thank Dr Mingqun Lin for helpful discussions, and Tzung-Huei Lai for providing reagents and protocol for real-time PCR analysis.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1 Double immunofluorescence microscopy of A. phagocytophilum inclusions with Beclin 1. A. phagocytophilum-infected HL-60 cells harvested at 3 (A), 20 h (B), 32 h (C) and 50 h (D) p.i. were labelled for double immunofluorescence with horse anti-A. phagocytophilum (FITC, green) and mouse monoclonal anti-Beclin 1 (Cy3, red). Scale bar = 5 µm.

FilenameFormatSizeDescription
CMI_1068_sm_Fig_S1_Legend.doc24KSupporting info item
CMI_fig_s1_legend.doc24KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.