the early secretory pathway contributes to the growth of ... · mids (1 mg ml1) using the...

INFECTION AND IMMUNITY, Jan. 2011, p. 402–413 Vol. 79, No. 10019-9567/11/$12.00 doi:10.1128/IAI.00688-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

The Early Secretory Pathway Contributes to the Growth of theCoxiella-Replicative Niche�†

Emanuel Martín Campoy, Felipe Carlos Martín Zoppino, and María Isabel Colombo*Laboratorio de Biología Celular y Molecular, Instituto de Histología y Embriología, Facultad de Ciencias Medicas,

Universidad Nacional de Cuyo-CONICET, Mendoza, Argentina

Received 25 June 2010/Returned for modification 14 July 2010/Accepted 28 September 2010

Coxiella burnetii is a Gram-negative obligate intracellular bacterium. After internalization, this bacteriumreplicates in a large parasitophorous vacuole that has features of both phagolysosomes and autophagosomalcompartments. We have previously demonstrated that early after internalization Coxiella phagosomes interactwith both the endocytic and the autophagic pathways. In this report, we present evidence that the Coxiella-replicative vacuoles (CRVs) also interact with the secretory pathway. Rab1b is a small GTPase responsible forthe anterograde transport between the endoplasmic reticulum and the Golgi apparatus. We present evidencethat Rab1b is recruited to the CRV at later infection times (i.e., after 6 h of infection). Interestingly, knockdownof Rab1b altered vacuole growth, indicating that this protein was required for the proper biogenesis of theCRV. In addition, overexpression of the active GTPase-defective mutant (GFP-Rab1b Q67L) affected thedevelopment of the Coxiella-replicative compartment inhibiting bacterial growth. On the other hand, disrup-tion of the secretory pathway by brefeldin A treatment or by overexpression of Sar1 T39N, a defectivedominant-negative mutant of Sar1, affected the typical spaciousness of the CRVs. Taken together, our resultsshow for the first time that the Coxiella-replicative niche also intercepts the early secretory pathway.

Coxiella burnetii, the etiologic agent of Q fever in humans, isan airborne, Gram-negative bacterium, which develops andmultiplies in large, acidified, hydrolase-rich vacuoles withphagolysosome-like characteristics. These Coxiella-replicativevacuoles (CRVs) are highly fusogenic and capable of interact-ing with both the endocytic and the phagocytic pathways (16).We have previously shown that the parasitophorous vacuolesalso have the hallmarks of an autophagosomal compartment(4) and that autophagy induction is beneficial for the replica-tion and survival of Coxiella (6, 13). Interestingly, we haverecently reported that the interaction with the autophagicpathway takes place early after infection (i.e., few minutes) andthat the autophagic protein LC3 is actively recruited by Cox-iella to its phagosomes long before the formation of the CRV(30). Therefore, since the bacterium-customized replicativevacuole that shelter Coxiella also displays lysosomal markerssuch as Lamp1 proteins and lysosomal enzymes (16), we inferthat C. burnetii resides in an autophagolysosome-like compart-ment.

Rab proteins are small GTPases that function as molecularswitches regulating important steps in membrane traffickingsuch as budding, microtubule-mediated transport, and tether-ing of the vesicle with the target compartment prior to fusion(49). These proteins serve as scaffolds recruiting differentdownstream effector molecules to integrate vesicular transport

and signaling (33, 49). More than 70 Rab and Rab-like proteinshave been identified in humans, and the function of many ofthem has been determined (reviewed in reference 34). In gen-eral, each Rab protein is enriched in a specific compartment,and although several Rab proteins can be present in the sameorganelle, it is believed that they occupy discrete membranedomains. For instance, Rab5 is present in early endosomes,whereas Rab7 is enriched in late endocytic/lysosomal compart-ments.

We have recently demonstrated that Coxiella-containingphagosomes acquire Rab5 and Rab7 sequentially at early timesafter infection (30). Rab proteins are known to cycle betweena GTP-bound active conformation and a GDP-bound inactiveconformation. Dominant-negative mutants that preferentiallyremain in the GDP inactive estate have been generated formost of the Rab proteins. In our previous studies we haveobserved that overexpression of a dominant-negative mutantform of Rab5, but not of Rab7, hampered Coxiella entry,whereas both Rab5 and Rab7 dominant-negative mutants in-hibited replicative vacuole formation, indicating that interac-tion with the endocytic pathway is a critical event for thebiogenesis of the Coxiella vacuole (30).

Rab1 is a member of the Rab family that preferentiallylocalizes at the endoplasmic reticulum (ER)-Golgi intermedi-ate compartment (12, 32). It has been reported that Rab1regulates anterograde transport between ER-Golgi compart-ments (25, 27), mediating the docking of ER-derived vesicleswith the cis-Golgi compartment (27). Rab1 has been shown tointeract with the Golgi membrane proteins GM130 andGRASP (25) contributing to the docking process. The intra-cellular pathogen Legionella pneumophila, after internaliza-tion, recruits early secretory vesicles from the host cell delayingthe fusion with the lysosomal compartment (17, 21). Interest-ingly, Rab1b was recruited to the Legionella-containing vacuole

* Corresponding author. Mailing address: Laboratorio de BiologíaCelular y Molecular, Instituto de Histología y Embriología (IHEM)-CONICET, Facultad de Ciencias Medicas, Universidad Nacional deCuyo, Casilla de Correo 56, Centro Universitario, Parque General SanMartín, 5500 Mendoza, Argentina. Phone: 54-261-4494143, ext. 2690.Fax: 54-261-4494117. E-mail: [email protected].

† Supplemental material for this article may be found at http://iai.asm.org/.

� Published ahead of print on 11 October 2010.

402

on January 15, 2021 by guesthttp://iai.asm

.org/D

ownloaded from

http://iai.asm.org/

(LCV) within minutes after internalization, and this Rab1brecruitment was important for the subsequent interaction ofER-derived vesicles with the LCV and for Legionella intracel-lular growth (22). In the present study, we show that at laterinfection times Rab1b decorates the Coxiella-containing vacu-ole, suggesting that this compartment intercepts with the earlysecretory pathway. Knockdown of Rab1b or overexpression ofa Rab1 dominant-negative mutant affected vacuole growth,suggesting that this protein was involved in the biogenesis ofthe CRV. In addition, we present evidence that disruption ofthe secretory pathway by brefeldin A treatment or by overex-pressing a dominant-negative mutant of Sar1 alters the forma-tion and maintenance of the single large and spacious Coxiellavacuole. Interestingly, overexpression of the green fluorescentprotein (GFP)-Rab1b GTPase-defective mutant affected thebiogenesis of the CRV. Our results indicate that this mutantaltered the normal growth of the Coxiella vacuole, at least inpart, by changing its fusogenic capacity with endocytic com-partments. Taken together, our results show the functionalconsequences of altering the secretory pathway on the Coxiella-replicative niche formation.

MATERIALS AND METHODS

Materials. Minimum essential alpha medium (�-MEM) and Dulbecco modi-fied Eagle medium (DMEM) were obtained from Gibco Laboratories (Invitro-gen, Argentina) and fetal bovine serum (FBS) was obtained from NATOCORS.RL, Cordoba, Argentina). LysoTracker and DQ-BSA were from MolecularProbes (Invitrogen). All other chemicals were from Sigma (Buenos Aires, Ar-gentina). Plasmids encoding Rab1b-myc and their mutants were kindly providedby Cecilia Alvarez, CIBICI (UNC-CONICET)-Fac. Cs. Químicas, UNC, Cor-doba. These plasmids were subcloned in the vector pEGFP in our laboratory.Rabbit polyclonal anti-Coxiella antibody was generously provided by RobertHeinzen (Rocky Mountain Laboratories, NIAID, NIH, Hamilton, MT).pIRES2-DsRed2 bicistronic vectors encoding Sar1, Sar1 H79G, and Sar1 T39Nwere a gift from Hirofumi Kai and Akiko Niibori (Faculty of PharmaceuticalSciences, Kumamoto University, Japan). Rab1b and control siRNAs were ob-tained from BIONEER (Korea).

Cell culture. Chinese hamster ovary cells (CHO), Vero cells, HeLa cells andRaw macrophages were grown on coverslips in �-MEM or DMEM supple-mented with 15% FBS at 37°C in an atmosphere of 95% air and 5% CO2 in24-well plates to 80% confluence. Stably transfected CHO cells overexpressingEGFP-Rab1b and its respective mutants were grown in the same medium sup-plemented with 0.2 mg of Geneticin ml�1. All antibiotics were removed 24 hbefore infection with Coxiella.

Cell transfection. CHO, HeLa, and Raw cells were transfected with the plas-mids (1 mg ml�1) using the Lipofectamine 2000 reagent (Invitrogen) accordingto the instructions supplied by the manufacturer. Transfected cells were incu-bated for 24 h in �-MEM supplemented with 15% FBS in 24- and 12-well platesto 80% confluence. Stably transfected cells were generated by selection with 0.5mg of Geneticin ml�1. In some cases, the cells were subsequently cloned andmaintained with 0.1 mg of Geneticin ml�1. Cotransfection experiments werecarried with Lipofectamine 2000 (Invitrogen).

Propagation of C. burnetii phase II. The clone 4 phase II Nine Mile strain ofC. burnetii bacteria, which is infective for cells but not for mammals, was pro-vided by Ted Hackstadt (Rocky Mountain Laboratories, National Institute ofAllergy and Infectious Disease, National Institutes of Health, Hamilton, MT)and handled in a biosafety level II facility (13a). Infective inocula were preparedas described previously (48a). Nonconfluent CHO cells were cultured in T75flasks at 37°C in �-MEM supplemented with 5% FBS, 0.22 g of sodium bicar-bonate liter�1, and 20 mM HEPES (pH 7) (MfbH). Cultures were infected withC. burnetii phase II suspensions at 37°C in an air (CO2) atmosphere. After 6 days,the cells were scrapped and passed 20 times through a 27-gauge needle con-nected to a syringe. Cell lysates were centrifuged at 1,800 � g for 10 min at 4°C.The supernatants were centrifuged at 25,000 � g for 30 min at 4°C, and pelletscontaining C. burnetii were resuspended in �-MEM and passed 10 times througha 27-gauge needle connected to a syringe. Afterward, the bacterial suspensionswere divided into aliquots and kept frozen at �70°C.

Infection of cells with C. burnetii. CHO and HeLa cells plated in T25 flaskswere washed two to three times with phosphate-buffered saline (PBS) and de-tached by using trypsin-EDTA. After resuspension in �-MEM, cells were platedon coverslips distributed in 6-, 12-, or 24-well plates. For infection, a 50-�laliquot of C. burnetii suspension was diluted with 10 ml of �-MEM, and 0.5 to 1ml of this dilution was added per well. In all experiments, host cells were infectedwith C. burnetii at a multiplicity of infection (MOI) of �20. Cells were incubatedfor different periods of time at 37°C in an atmosphere of 95% air and 5% CO2.The same protocol was used for the infection of CHO cells overexpressingpEGFP-Rab1b and the corresponding mutants.

Confocal microscopy. pEGFP-Rab1b-, pEGFP-, or pRFP-LC3-transfectedCHO cells were analyzed by confocal microscopy using a Nikon C1 confocalmicroscope system or an Olympus FluoViewTM FV1000 confocal microscope(Olympus, Argentina), with the EZ-C1 software (Nikon, Japan) or the FV10-ASW (version 01.07.00.16) software, respectively. Images were processed byusing Adobe CS3 (Adobe Systems).

Indirect immunofluorescence. CHO, HeLa, and Raw cells were fixed with 3%paraformaldehyde solution in PBS for 10 min at room temperature, washed withPBS, and blocked with 50 mM NH4Cl in PBS. Subsequently, cells were perme-abilized with 1% saponin in PBS containing 1% bovine serum albumin (BSA),and then incubated with the primary antibody against Coxiella (1:1,000). Boundantibodies were detected by incubation with a secondary antibody conjugatedwith Texas Red or Cy5 (1:800; Jackson Immunoresearch Laboratories). Cellswere mounted with Mowiol (plus Hoechst stain) and examined by confocalmicroscopy.

Measurement of the percentage of infected cells and the number and size ofC. burnetii RVs. At different times postinfection, the cells were fixed for 10 minin 3% paraformaldehyde. About 500 cells per coverslip (in triplicates) werescored for the presence or absence of large C. burnetii vacuoles using a confocalmicroscope (Nikon and Olympus) with a 60� objective lens. Infected cells weredefined as those with at least one large parasitophorous vacuole (i.e., �2 �m)with clear identifiable bacteria inside. The size of the vacuoles was determined bya morphometric analysis using Metamorph and FV10-ASW software.

Endocytosis and phagocytosis assays. CHO cells overexpressing EGFP,EGFP-Rab1b wild-type (wt), or Rab1b Q67L were plated at a confluence of 50to 60% the day before the experiment. After 24 h, the cells were infected with C.burnetii. At 48 h postinfection, the cells were allowed to internalize heat-inacti-vated Staphylococcus aureus-rhodamine or 5 �g of dextran-rhodamine/ml to labelthe phagocytic and endocytic pathways, respectively. The 24-well plate was cen-trifuged for 10 min (1,500 rpm) in order to induce the contact of the S. aureusparticles to the cell surface. After 1 h of incubation, the cells were washed threetimes with PBS, fixed with 3% paraformaldehyde for 12 min, and quenched with50 mM NH4Cl for 15 min. The cells were subjected to indirect immunofluores-cence using a polyclonal antibody against C. burnetii. The glass slides weremounted using Mowiol and analyzed by confocal microscopy.

Phagocytosis of latex beads. CHO cells overexpressing enhanced green fluo-rescent protein (EGFP), EGFP-Rab1b wt, or Rab1b Q67L were plated at aconfluence of 50 to 60% the day before the experiment. The cells were allowedto internalize 4-�m latex beads (with an MOI of 10) for 24 h. The cells were thenwashed three times with PBS, fixed with 3% paraformaldehyde for 12 min, andquenched with 50 mM NH4Cl for 15 min. Samples were mounted on glass slidesusing Mowiol and analyzed by confocal microscopy.

Fluorescent infectious FFU assay. C. burnetii replication during its develop-mental cycle was quantified by using a replating fluorescent infectious focus-forming unit (FFU) assay (18). Confluent CHO cells stably overexpressingEGFP-Rab1b Q67L or EGFP (as a control) were grown in individual wells of asix-well culture dish. CHO cells were incubated with C. burnetii inocula for 48 hat 37°C. Infected cells were harvested by scraping and then disrupted by gentlesonication. Cell lysates with released C. burnetii were used to infect fresh con-fluent Vero cells in an individual well of a 24-well culture dish. Vero cells wereincubated with cell lysates containing C. burnetii inoculum for 3 h at 37°C to allowinternalization. Cells were washed several times to eliminate extracellular bac-teria and incubated with fresh �-MEM (supplemented with 5% FBS). After 72 hof incubation, the cells were fixed, and the FFU were determined by indirectimmunofluorescence using a polyclonal antibody against C. burnetii. FFU countswere scored by confocal microscopy in an Olympus FluoViewTM FV1000 con-focal microscope with the FV10-ASW software.

siRNA silencing of Rab1b. Purified small interfering RNA (siRNA) againsthuman Rab1b and control siRNA were purchased from BIONEER (Korea).Confluent HeLa cells were transfected with siRNA Rab1b or with controlsiRNA, both prepared at a final concentration of 20 nM in 400 ml of �-MEMwithout serum and with Lipofectamine 2000 reagent. The mix was added to a24-well culture dish. After 5 h, the transfection mix was replaced by fresh

VOL. 79, 2011 COXIELLA BURNETII AND THE SECRETORY PATHWAY 403


.org/D

ownloaded from

http://iai.asm.org/

�-MEM with 10% FBS. At 48 h posttransfection, the medium was removed, andthe same transfection protocol with the corresponding siRNA was applied again(second hit). When the second transfection mix was removed, the HeLa cellswere infected with C. burnetii for 72 h at 37°C.

Western blotting. HeLa cells were washed twice, scraped, and resuspended insample buffer containing 1% 2-mercaptoethanol and then sonicated for 10 minat 4°C. For Western blot analysis, protein extracts were subjected to electro-phoresis in SDS–12.5% PAGE gels and transferred to a Hybond-enhancedchemiluminescence nitrocellulose membrane (Amersham GE, Buenos Aires,Argentina). Membranes were blocked with blotto (PBS with 5% nonfat milk and0.1% Tween 20) for 1 h at room temperature. Subsequently, the membraneswere washed twice with 0.1% PBS-Tween 20 and incubated with the primaryantibody. To detect Rab1, membranes were incubated overnight at 4°C withrabbit anti-Rab1 1:100 (Santa Cruz Biotechnology, Santa Cruz, CA), washed,and incubated with horseradish peroxidase-conjugated anti-rabbit antibody(Jackson Immunoresearch Laboratories) at a final dilution of 1:10,000. Equiva-lent protein loadings were confirmed by incubating membranes with a mouseanti-actin antibody (1:300) for 1 h at room temperature. The bands were visu-alized using the ECL reagent from GE. Images of the bands were obtained byusing a LAS-4000 luminescent image analyzer (Fujifilm).

RESULTS

Rab1b decorates the Coxiella-replicative compartment atlater infection times. As indicated in the introduction, Legio-nella-containing vacuoles recruit Rab1b within few minutesafter internalization (8). Given certain similarities between theintracellular behavior of Legionella and Coxiella, we decided toanalyze whether this Rab protein could be recruited to theCoxiella phagosomes. For this purpose, stably transfectedCHO cells overexpressing EGFP-Rab1b wt or the GTPasedefective mutant EGFP-Rab1b Q67L were infected with C.burnetii. Cells overexpressing EGFP alone were used as con-trol. The cells were fixed at different infection times (from 10min to 24 h) and subjected to immunofluorescence using an

antibody against C. burnetii. As shown in the confocal images,Rab1b presented a typical perinuclear Golgi distribution (Fig.1). However, no colocalization between Coxiella and Rab1b(wt or mutant) was observed between 10 and 60 min after theuptake. By 6 or 12 h postinfection, only in cells overexpressingthe mutant Rab1b Q67L were some of the incipient Coxiella-containing vacuoles decorated by this protein. This colocaliza-tion was more evident by 24 h postinfection, and at this timepoint it was also visible in cells expressing the wt protein. Asshown in Fig. 2, by 48 h postinfection, when the large andspacious Coxiella niche is generated, the vacuoles were clearlydecorated by Rab1b Q67L, whereas in cells overexpressingRab1b wt, the protein, although present, presented a morediscontinuous distribution showing a patchy pattern (see Fig. 2,insets). The bottom panels show the quantification of the flu-orescence intensity along the yellow line scan across the CRVdepicted in the insets. The graphs show accumulation of Rab1bwt or Rab1b Q67L mutant (green) at the surrounding CRVmembrane.

The results point out that at later infection times the Cox-iella-replicative niche recruits Rab1b or interacts with a Rab1b-labeled compartment. In order to demonstrate the specificityof the Rab1 recruitment to the Coxiella vacuole, we analyzedthe interaction with Rab11, another member of the Rab familyof GTPases. Rab11, is a well-known marker for recycling en-dosomes (RE) and plays a key role in regulating the transportof vesicles/proteins from REs and early endosomes to thetrans-Golgi network and plasma membrane through recyclingpathways (28, 41, 46). In contrast to Rab1, no colocalizationwith either Rab11wt or the GTPase-deficient mutant(Rab11Q70L) was observed (see Fig. S1A in the supplemental

FIG. 1. Kinetic of infection of C. burnetii under Rab1b overexpression. Stably transfected CHO cells overexpressing EGFP alone, EGFP-Rab1bwt, or the GTPase-defective mutant EGFP-Rab1b Q67L were infected with C. burnetii. The cells were fixed at different infection times (10 min,30 min, 1 h, 6 h, 12 h, and 24 h), subjected to indirect immunofluorescence using an antibody against C. burnetii (red), and visualized by confocalmicroscopy. The images show a typical Rab1b distribution in a perinuclear Golgi pattern; however, no colocalization of Rab1b (wt or mutant) withC. burnetii containing compartment was observed at early times (10, 30, or 60 min). Nevertheless, a clear recruitment of EGFP-Rab1b Q67L isobserved at later infection times (6, 12, and 24 h). Bars, 10 �m.

404 CAMPOY ET AL. INFECT. IMMUN.


.org/D

ownloaded from

http://iai.asm.org/

material), even at more than 48 h postinfection, suggesting thatthe presence of Rab1 on the vacuole was not just due tooverexpression of any Rab protein. In addition, with the pur-pose of demonstrating that the recruitment of Rab1 wt orQ67L mutant was specific to the Coxiella-vacuole, CHO cellsoverexpressing EGFP, EGFP-Rab1b wt, or EGFP-Rab1bQ67L were allowed to phagocytose latex beads for 24 h (seeMaterials and Methods). As expected, no EGFP-Rab1b wt orEGFP-Rab1b Q67L was found decorating the bead phagoso-mal membrane (see Fig. S1B in the supplemental material),indicating that Rab1 recruitment was limited to the pathogen-containing compartment.

In order to demonstrate that the association with Rab1 wasalso observed in another cell type, we next assessed the pres-

ence of Rab1 on the CRV using a macrophage cell line. Similarto the results obtained with CHO cells, in Raw macrophagesEGFP-Rab1b Q67L was also recruited to the vacuole mem-brane since 6 h after infection, whereas in Rab1b wt-overex-pressing cells the recruitment was only observed by 24 h postin-fection (see Fig. S2 in the supplemental material). Theseresults indicate that no major differences in the kinetics ofRab1 recruitment are observed between the two cell lines used(i.e., CHO and Raw macrophages). An interesting observationis that in both CHO and Raw macrophages overexpressing theGTPase-defective mutant (Rab1bQ67L) the infection patternwas different from that of cells overexpressing GFP alone orRab1b wt. Instead of the one or two large spacious vacuoletypically observed by 48 h postinfection, numerous but smallervacuoles per cell were present, suggesting that the overexpres-sion of this Rab1 mutant somehow alters the normal develop-ment of the Coxiella-vacuole.

Taken together, these results indicate that, unlike Legio-nella, Coxiella phagosomes do not interact with Rab1b at earlytimes after internalization (i.e., 10 to 60 min), whereas at latertimes postinfection the CRV becomes decorated by Rab1,mainly when the protein is in its GTP-bound state (i.e., Rab1bQ67L mutant). The association of Rab1b with the limitingmembrane of the Coxiella-vacuole remains even after 48 h,when the vacuole is fully developed.

Overexpression of Rab1b N121I or Rab1b silencing alterCoxiella-vacuole growth. Given the observation that bothEGFP-Rab1b wt and the constitutively active mutant (Rab1bQ67L) were present at the Coxiella-vacuole membrane, wedecided to analyze the effect of overexpressing a dominant-negative mutant of Rab1b (Rab1b N121I) on CRV growth.The N121I mutation renders a Rab1 protein defective in bind-ing to either GDP or GTP (i.e., nucleotide empty mutant). TheN121I mutant presents a diffuse cellular pattern without asso-ciating with a clearly defined compartment (2). Expression ofthis mutant in cells blocks the exit of cargo vesicular stomatitisvirus G protein from the ER and causes the complete disas-sembly of the Golgi apparatus. Because prolonged overexpres-sion of this mutated protein is lethal to CHO cells, we applieda different experimental approach to assess the impact ofRab1b N121I on Coxiella-vacuole growth. As shown in Fig. 3A(experimental procedure 1), CHO cells were infected for 24 hbefore the transfection process and fixed after 48 h of contin-uous infection. Measurements of the vacuole diameter underthe overexpression of Rab1b N121I showed a decrease in thelarge vacuole population (�10 �m), in contrast to the controlcondition (Fig. 3B). These results indicate that a functionalRab1b is required for the normal growth of the CRV.

In order to confirm the requirement of Rab1b for the gen-eration of a spacious Coxiella-replicative vacuole, a knockdownassay was performed. HeLa cells were transfected with asiRNA against human Rab1b or a control siRNA. A double-hitprotocol, as depicted in Fig. 3A (experimental procedure 2),was applied to maintain the protein silencing effect during theentire infection process. HeLa cells were transfected twice withthe siRNA during the 48-h period after cell seeding. Immedi-ately after the second transfection, the cells were infected withC. burnetii and cultured for an additional 72 h to allow thegrowth of the large CRV. It has been demonstrated that Rab1bdepletion results in ER-to-Golgi trafficking defect and disrup-

FIG. 2. Rab1b decorates the Coxiella-replicative niche at latertimes of infection. CHO cells stably overexpressing EGFP, EGFP-Rab1b wt, or the active mutant of Rab1 (EGFP-Rab1b Q67L) wereinfected for 48 h with C. burnetii. After fixation, cells were subjected toindirect immunofluorescence using an antibody against C. burnetii(red) and analyzed by confocal microscopy. As depicted in the insets,the large Coxiella-containing vacuole was strongly labeled by Rab1bQ67L and in a patchy pattern by Rab1b wt. Insets also show a yellowline scan in order to visualize the localization of the Rab1b wt and itsdefective dominant-active mutant (Q67L) in the vacuole membrane.Images are representative of at least three independent experiments.Bars, 10 �m.



.org/D

ownloaded from

http://iai.asm.org/

tion of the Golgi apparatus (19). Thus, in order to corroboratethat Rab1b was effectively knocked down, the distribution ofthe Golgi marker GM130 was analyzed. The results of anindirect immunofluorescence assay for the detection of GM130(a cis-Golgi protein) (red, arrowheads) and C. burnetii (green)are shown in Fig. 3C. As expected, in cells treated with the

siRNA against Rab1b, the Golgi apparatus was disassembled.A disperse distribution of the Golgi apparatus was observed, incontrast to the normal perinuclear pattern observed under thecontrol condition (Fig. 3C, arrowheads), indicating disruptionof the organelle. A Western blot assay to detect Rab1 was alsoperformed in order to verify the silencing of the protein (Fig.3E). As shown in the figure, in cells treated with the siRNAagainst Rab1b the Coxiella-vacuoles (surrounded by a yellowdashed line) were smaller than in cells treated with the controlsiRNA (arrows). The vesicle size was quantified, and a markeddecrease in the diameter of CRVs was clearly observed in cellswithout normal expression levels of Rab1b (Fig. 3D). Theseresults indicate that Rab1b silencing affects vacuole growth,suggesting that this protein has a critical role in the biogenesisof the Coxiella-replicative niche.

Disruption of the secretory pathway alters the size of thespacious Coxiella-vacuole. As indicated above, Rab1b N121I, amutant that leads to disruption of the early secretory pathwaywith the consequent Golgi disassembly, altered CRV growth.To confirm the participation of the secretory pathway in thegeneration of the CRV, we used two different strategies. Thefirst strategy involved the Ras-related protein ADP-ribosyla-tion factor 1 (Arf1), which is a low-molecular-weight GTP-binding protein, that in its GTP state supports the binding ofcoatomer, a cytosolic coat protein complex, to Golgi mem-branes (9). Brefeldin A (BFA) is a noncompetitive inhibitor ofArf1 activation. Cell treatment with BFA results in specificinhibition of the Arf1 guanine-nucleotide exchange factor(GEF) and in efficient and reversible blocking of membranetraffic at the Golgi apparatus (26). In order to analyze theeffect of BFA on the maintenance of the spacious CRV, CHOcells were transfected with EGFP or EGFP-Rab1b Q67L, theGTPase-defective mutant of Rab1b. After transfection, cellswere infected with C. burnetii for 48 h and incubated for 2 or6 h in the presence or absence of 200 nM BFA. The cells werefixed and subjected to indirect immunofluorescence for thedetection of both the Golgi protein GM130 and Coxiella usingspecific antibodies. As expected, after 2 h of BFA treatment,the Golgi apparatus showed a disperse pattern, indicating thatthis organelle has been disassembled (Fig. 4A, lower panels,arrowheads). BFA caused a significant decrease in the CRVsize compared to untreated cells (Fig. 4A, arrows pointing to aCRV encircled with a yellow line). The quantification of theCRV diameter is depicted in Fig. 4B. In cells overexpressingRab1b Q67L and treated with BFA, the CRVs remained dec-orated with this Rab1b mutant (data not shown), and no dif-ferences in vacuole diameter were observed compared to un-treated cells (i.e., no BFA). Our results are in agreement withpreviously published data indicating that overexpression ofRab1b Q67L conferred resistance to BFA (24). The reductionin vacuole size after BFA treatment indicates that a continuousprovision of membranes from the ER is a requirement for thepersistence of the spacious Coxiella vacuole.

Sar1 is a small GTPase involved in the formation of COPII-coated transport vesicles that bud from the endoplasmic retic-ulum (reviewed by (37). Overexpression of both Sar1 T39N, aGDP-restricted mutant of Sar1 (23), or Sar1 H79G, a GTP-restricted mutant, blocks ER exit sites disrupting the secretorypathway with the consequent disassembly of the Golgi appa-ratus (48). Thus, the second strategy consisted in disrupting the

FIG. 3. Overexpression of Rab1b N121I and knockdown of Rab1baffect CRV growth. (A) Outline of the experimental procedures ap-plied in the different assays. (B) CHO cells were infected for 24 h withC. burnetii and subsequently transfected with pEGFP-Rab1 N121I orpEGFP (control). At 48 h postinfection the cells were fixed and sub-jected to indirect immunofluorescence to detect Coxiella using specificantibodies. Quantification of the vacuole diameter in cells overexpress-ing EGFP-Rab1b N121I was compared to the control condition (P �0.05). (C) HeLa cells were transfected with siRNA against Rab1b or anirrelevant siRNA as a negative control. After 48 h, cells were trans-fected for a second time, infected with C. burnetii, and cultured for anadditional 72-h period to allow the development of the large CRV.Subsequently, the cells were fixed and subjected to indirect immuno-fluorescence for the detection of both Golgi protein GM130 and Cox-iella by using specific antibodies. Images were captured by confocalmicroscopy. The panels show that in contrast to control cells, in cellstreated with the siRNA against Rab1b the Golgi apparatus was disas-sembled (punctate distribution, arrowheads) in the majority of thecells, indicating that Rab1b was effectively depleted. Furthermore, thesizes of the Coxiella-vacuoles were markedly decreased in cells treatedwith the siRNA against Rab1b (yellow dashed line). (D) Quantifica-tion of the vacuole diameter in cells in which Rab1b was depletedcompared to the control condition. The data represent the means �the standard errors of the mean (SEM) of at least three independentexperiments in which at least 200 vacuoles were scored in each exper-iment (P � 0.001). The data represent the means � the SEM of at leastthree independent experiments. (E) Western blot of the assay de-scribed in panel C and quantification of intensity of the Rab1 bandsrelative to actin. The data represent one of two independentexperiments.



.org/D

ownloaded from

http://iai.asm.org/

secretory pathway with Sar1 T39N. Under overexpression ofSar1 T39N, the protein GM130 is redistributed into punctatecytoplasmic structures, namely, Golgi remnants. When Sar1T39N is overexpressed the COPII coat protein Sec13p is dis-placed from the ER exit sites (35), blocking ER-Golgi trans-port. To analyze the Sar1 T39N effect on the generation of thespacious CRV, CHO cells were infected with C. burnetii, ac-cording to the experimental procedure 1 described in Fig. 3A.At 24 h postinfection, the cells were transfected with a bicis-tronic plasmid encoding Sar1 T39N and the protein Ds-Red(pIRES-DsRed Sar1 T39N) to visualize the transfected cells.Subsequently, at 24 h posttransfection, the cells were fixed,subjected to immunofluorescence, and analyzed by confocalmicroscopy. Figure 4C (inset) shows two cells, one of themoverexpressing Sar1 T39N (ds-red) with a disassembled Golgiapparatus (green) and an untransfected cell with the typicalperinuclear Golgi distribution (see arrowheads). The lowerpanel shows that the size of the vacuoles (surrounded by ayellow dashed line) is smaller in Sar1 T39N-overexpressing

cells than in control untransfected cells (arrows). Quantifica-tion of the CRV diameter is depicted in Fig. 4D, which showsthat the percentage of larger vacuoles (�10 �m) was dramat-ically decreased under Sar1 T39N overexpression compared tothe control condition. Taken together, both the BFA effect andthe results obtained with the Sar1 mutant confirm that disrup-tion of the early secretory pathway alters normal CRV growth.

In cells transfected with the GTPase-deficient mutant ofRab1b, the biogenesis of the C. burnetii-containing compart-ment is altered. In general, at 48 h postinfection with C. bur-netii a very large CRV vacuole is generated. At this infectiontime, it is very common to observe only one or two spaciousvacuoles per cell, likely as a result of homotypic fusion events.An interesting observation from the images shown in Fig. 2 isthat, apparently, in cells overexpressing the GTPase-defectivemutant Rab1b Q67L the size of the CRVs was decreased,albeit with a concomitant increase in vesicle number. Pub-lished work has shown that expression of the constitutivelyactive Q67L mutant has no detectable effect on ER-Golgitrafficking and Golgi structure (2, 39). Thus, the diminishedsize is not likely attributed to a deficient delivery of earlysecretory vesicles to the CRV. Therefore, we were interested indetermining why the size and number of the CRVs were al-tered in Rab1b Q67L-expressing cells. To begin addressing thispoint, we first decided to carefully analyze both aspects (i.e.,vesicle size and number) in cells transfected with pGFP-Rab1bwt or pGFP-Rab1b Q67L. Transfected cells were infected withC. burnetii for 48 or 72 h and subjected to immunofluorescenceto detect the bacteria. As shown in Fig. 5A, in control cellsoverexpressing EGFP alone the typical formation of one ortwo very large CRV containing C. burnetii is depicted (arrows).A similar infection pattern was observed in cells overexpress-ing EGFP-Rab1b wt. In contrast, numerous CRVs werepresent in cells overexpressing GFP-Rab1b Q67L at either 48or 72 h postinfection. Both parameters, vesicle size and num-ber, were quantified and, as indicated in Fig. 5B and C, anoticeable increase in the number of CRVs was clearly ob-served in cells overexpressing the Rab1b mutant with a markeddecrease in vacuole size. These results indicate that overex-pression of the GTPase-defective mutant of Rab1b alters theinfection pattern of C. burnetii modifying the generation of theCoxiella-replicative compartment.

Another important point was to determine whether thechange in the typical infection pattern described above inter-fered with the replicative properties of C. burnetii. Therefore,with the purpose of determining this aspect, we performed afluorescent infectious FFU assay (see Materials and Methods).CHO cells stably overexpressing EGFP (as control) or EGFP-Rab1b Q67L were infected with C. burnetii for 48 h. The cellswere then harvested by scraping and disrupted by sonication.Afterward, Vero cells were incubated with the cell lysates con-taining C. burnetii for 3 h (uptake) at 37°C to allow internal-ization. After 72 h of incubation, the cells were fixed andsubjected to indirect immunofluorescence, and FFU countswere determined by using confocal microscopy. As shown inFig. 5D, there was a marked decrease (ca. 50%) in the bacterialreplication of Rab1b Q67L compared to the control condition(EGFP) or to cells overexpressing GFP-Rab1wt (not shown).

In conclusion, these results show that the vacuole size was

FIG. 4. BFA treatment or overexpression of a dominant-negativemutant of Sar1 alter CRVs growth. CHO cells transiently overexpress-ing EGFP were infected with C. burnetii. After 48 h of continuousinfection, the cells were treated with BFA for 2 or 6 h. After treatment,the cells were fixed and subjected to indirect immunofluorescence withspecific antibodies against C. burnetii and GM 130 as a Golgi marker.(A) The images show smaller CRVs (blue) under BFA treatment (6 h)compared to untreated control cells (yellow dashed line). A partialdisassembled Golgi (red) is shown in the lower panels (arrowheads).(B) Quantification of the vacuole size (10 �m) of the experimentpresented in panel A (P � 0.05). (C) CHO cells infected with C.burnetii for 24 h were transfected with pIRES-DsRed Sar1 T39N. Afterfixation, the cells were subjected to immunofluorescence using anti-bodies against C. burnetii (blue) and GM130 as a Golgi marker(green). A Golgi apparatus totally disassembled is shown (arrow-heads). The figure shows the reduction in the CRV size (yellow dashedline) under Sar1 T39N overexpression (white dashed line). (D) A bargraph shows a quantification of the �10-�m CRVs in control or Sar1T39N-overexpressing cells. The data represent the means � the SEMof at least three independent experiments (P � 0.001). Bars, 15 �m.



.org/D

ownloaded from

http://iai.asm.org/

altered by overexpression of Rab1b Q67L and, more impor-tantly, that the C. burnetii replication rate was affected.

The Coxiella-containing compartments generated in cellsoverexpressing the mutant Rab1b Q67L retain its acidic anddegradative characteristics. As indicated in the introduction, itis well known that the large vacuoles generated by C. burnetiiare acidic and have features of phagolysosomes (reviewed inreference 15). Since in cells overexpressing the Rab1b Q67Lmutant no generation of the large CRVs was observed, wewere interested in determining whether the smaller vacuolesconserved the other features of normal CRVs. To assesswhether these Coxiella-containing vesicles were indeed acidic,CHO cells stably overexpressing GFP, Rab1wt, or Rab1bQ67L were infected with Coxiella for 48 h and labeled with the

acidotropic marker LysoTracker. As shown in Fig. 6A, theCoxiella-containing vacuoles generated in all of the conditionswere clearly labeled with LysoTracker (see insets), indicatingthat even in cells overexpressing the Q67L mutant the bacte-rium-containing compartments preserve the acidic propertiestypical of the Coxiella-replicative niche.

We next assessed whether these Coxiella-containing vacuoleshad lysosomal features (i.e., hydrolytic compartment). For thispurpose, infected cells were incubated for 6 h with DQ-BSA(Molecular Probes), a BSA derivative so heavily labeled thatthe fluorophore is self-quenched. Proteolysis of this compoundresults in dequenching and the release of brightly fluorescentfragments. Thus, the use of DQ-BSA is a valuable tool for thevisualization of a proteolytic compartment. Similar to the con-trol condition (i.e., cells overexpressing the vector alone), co-localization of Coxiella and DQ-BSA in a EGFP-Rab1b wt orQ67L-decorated vacuole is depicted in Fig. 6B, indicating thedegradative characteristics of the vacuole. As shown in panelsC and D, practically no differences were observed in the per-centage of colocalization with LysoTracker or DQ-BSA andCoxiella-vacuole in cells overexpressing the different con-structs.

We have previously demonstrated that the Coxiella-replica-tive niche is decorated by the protein LC3, a hallmark ofautophagic vacuoles (4, 13). Furthermore, we have shown thatautophagy induction by amino acid deprivation favors the de-velopment of the CRV (13) and that LC3 recruitment is ac-tively modulated by Coxiella itself (30). Therefore, we wereinterested in determining whether the Coxiella-containingvacuoles generated in cells overexpressing the Rab1b mutantalso conserved the capability of recruiting the autophagic pro-tein LC3. Stably transfected CHO cells overexpressing EGFP-Rab1b Q67L were transiently transfected with pRFP-LC3. Thecells were subsequently infected with C. burnetii for 72 h, sub-jected to immunofluorescence, and analyzed by confocal mi-croscopy. In the insets in Fig. 6E, colocalization of Rab1bQ67L (green) and LC3 (red) at the membrane of the Coxiella-containing vacuoles (blue) is depicted, indicating that the in-teraction with the autophagy pathway is also maintained.

Taken together, our results suggest that despite the reducedvesicle size of the vacuoles that shelter Coxiella in cells over-expressing the mutant Rab1b Q67L, the vacuoles retain theacidic and autophagolysosome-like features of these compart-ments.

The fusogenic and replicative properties of the Coxiella com-partments generated in cells overexpressing Rab1b Q67L arealtered. As indicated above, the expression of the constitu-tively active Q67L mutant does not inhibit ER-Golgi trans-port (2, 39). Since the CRVs generated in Rab1b Q67L-expressing cells have a reduced vesicle size, we wonderedwhether the fusion capacity of the CRVs was altered. It hasbeen previously demonstrated that molecules internalizedby fluid-phase endocytosis are found in vacuoles containingC. burnetii (15) and that these vacuoles also fuse with othercompartments of the phagocytic/endocytic system (43, 44).To analyze the fusion capacity of the CRVs generated incells overexpressing the Rab1b mutant Q67L, CHO cellsoverexpressing EGFP alone (control) or EGFP-Rab1bQ67L were infected with C. burnetii for 48 h to allow thedevelopment of the Coxiella vacuole. Subsequently, cells

FIG. 5. Overexpression of Rab1b Q67L alters the infection profileof C. burnetii. CHO cells transfected with pEGFP plasmids encodingRab1b wt or the active mutant Rab1b Q67L were infected with C.burnetii for 48 or 72 h (see Materials and Methods), subjected toindirect immunofluorescence to detect the bacteria, and analyzed byconfocal microscopy. (A) Confocal images showing the typical forma-tion of one or two large vacuoles containing C. burnetii (yellow dashedline) in CHO cells overexpressing EGFP alone or EGFP-Rab1b wt. Incontrast, in cells overexpressing EGFP-Rab1b Q67L, the Coxiella-containing vacuoles are smaller than those of the control condition butare present in large numbers (arrowheads). (B) Quantification of thenumber of vacuoles per cell at 48 and 72 h of infection from theexperiment presented in panel A. (C) Quantification of the size ofvacuoles at 48 h postinfection from the experiment presented in panelA. The data represent the means � the SEM of at least three inde-pendent experiments. (D) CHO cells stably overexpressing EGFP (ascontrol) or EGFP-Rab1b Q67L were incubated with C. burnetii. After48 h of infection, the cells were lysed. Cell lysates with released C.burnetii were used to infect Vero cells. After 72 h of incubation (chase),cells were fixed, and the fluorescent infectious FFU were determinedby indirect immunofluorescence, and FFU counts were obtained usingconfocal microscopy. The data represent the means � the SEM of atleast three independent experiments where at least 200 cells werescored in each experiment (P � 0.05). Bars, 15 �m.



.org/D

ownloaded from

http://iai.asm.org/

were incubated for 1 h with heat-inactivated Staphylococcusaureus-rhodamine or dextran-rhodamine to label the phago-cytic and endocytic pathways, respectively. After fixation,cells were labeled by immunofluorescence with a specificantibody against C. burnetii and analyzed by confocal mi-croscopy. As shown in Fig. 7A, upper panels, colocalizationof S. aureus-rhodamine (red) and C. burnetii (blue) in cellsoverexpressing EGFP is clearly visualized in most of theCRVs (see also the quantification in Fig. 7C). In contrast, inCHO cells overexpressing EGFP-Rab1b Q67L very limitedcolocalization is observed (Fig. 7A, lower panels, and Fig.7C), suggesting that these compartments have its fusogenicproperties with the phagocytic pathway altered. Similar re-sults were obtained when the endosomal compartment wasloaded with rhodamine-dextran (Fig. 7B, lower panels), in-dicating that the fusogenic capacity of the CRVs with theendocytic pathway was also affected (see the quantificationin Fig. 7C).

Taken together, these results suggest that the GTPase-defective mutant of Rab1b alters normal growth of the Cox-iella vacuole by changing the fusogenic properties of thiscompartment, since fusion between the Coxiella vacuole andboth the endocytic and the phagocytic pathways was ham-pered.

The recruitment of Rab1b Q67L to the Coxiella-vacuolemembrane depends on a functional secretory pathway. Wehave presented evidence that Rab1b Q67L is decorating theCoxiella-vacuole membrane (Fig. 1 and 2). We have alsoshown that vacuole growth requires a functional secretorypathway, since the overexpression of the negative mutant ofSar1 affected CRV enlargement. Therefore, our next aimwas to elucidate whether the recruitment of the GTPasedefective mutant was a process mediated by vesicular trans-port from the ER or whether the protein was directly re-cruited from the cytosol. For this purpose, we blocked theearly secretory pathway by overexpressing the GDP-re-stricted mutant of Sar1 (Sar1 T39N) which strongly inhibitsvesicle budding from the ER. Because the overexpression ofSar1 T39N is extremely toxic, cells were infected prior to thetransfection procedure (see Fig. 3A, experimental proce-dure 1). At 24 h postinfection, cells were cotransfected withRab1b Q67L and Sar1 T39N for 24 h. After fixation, cellswere subjected to immunofluorescence and analyzed by con-focal microscopy. As previously observed in cells expressingRab1b Q67L alone, this mutant protein was recruited to theCRV (Fig. 8A,). In contrast, in cells coexpressing the dom-inant-positive mutant of Rab1b (Q67L) and the dominant-negative mutant of Sar1 (T39N) the CRVs were not deco-rated with Rab1b Q67L (Fig. 8A, inset). As expected, Fig.8A (arrowheads) also shows that the Golgi integrity is af-fected in cells overexpressing Sar1 T39N (delineated cell).Likewise, when the dominant-positive mutant of Sar1 (Sar1H79G), which also inhibits the ER-Golgi transport, wascoexpressed with Rab1b Q67L, the Coxiella vacuoles werenot decorated by this Rab1b mutant (data not shown).Quantification of the CRVs size also revealed that, similarto the effect of Sar1 T39N, the GTPase-defective mutantSar1 H79G also affected Coxiella-vacuole growth (Fig. 8Band C).

Taken together, these results highlight the important con-

FIG. 6. Coxiella-containing compartments labeled with Rab1b Q67Lretain its acidic and degradative characteristics and are also labeled byLC3. CHO cells stably overexpressing EGFP, EGFP-Rab1wt, or EGFP-Rab1b Q67L were infected with C. burnetii. At 48 h of infection, the cellswere incubated for 30 min with the acidotropic probe LysoTracker orDQ-BSA for 6 h. After fixation, the cells were examined by indirectimmunofluorescence with a specific antibody against C. burnetii and an-alyzed by confocal microscopy. (A) The insets show the colocalization ofLysoTracker (red) and C. burnetii (blue) inside the Coxiella vacuole in allcell types, suggesting that these compartments preserve the acidic prop-erties typical of the Coxiella-replicative niche. (B) Confocal images showthe colocalization of C. burnetii (blue) and DQ-BSA (red), indicating thedegradative characteristics of the vacuoles. Bars, 20 �m. (C) Quantifica-tion of the percentage of colocalization of C. burnetii and LysoTracker.(D) Quantification of the percentage of colocalization of C. burnetii andDQ-BSA. A total of 50 cells were counted in each condition. (E) CHOcells stably overexpressing EGFP-Rab1b Q67L were transiently trans-fected with pRFP-LC3. At 24 h posttransfection, the cells were infectedwith C. burnetii for 72 h. Cells were fixed, analyzed by indirect immuno-fluorescence, and examined by confocal microscopy. Insets show the co-localization of Rab1b Q67L (green) and LC3 (red) at the membrane ofthe vacuoles containing Coxiella (blue). This result suggests that the in-teraction with the autophagic pathway is maintained under overexpres-sion of Rab1b Q67L. Bars, 15 �m.



.org/D

ownloaded from

http://iai.asm.org/

tribution of the secretory pathway to the biogenesis of the C.burnetii-replicative niche, since disruption of this essentialpathway affects the enlargement of the vacuole. Moreover,the evidence also indicates that the recruitment of Rab1b

Q67L to the vacuole membrane depends on a functionalsecretory pathway.

DISCUSSION

C. burnetii, like several other intracellular pathogens, repli-cates in a membrane-bound compartment within the host cell(reviewed in reference 45). After 48 to 72 h of infection, Cox-iella generates a large replicative compartment (i.e., CRV) thatin many cases occupies the whole cytoplasm, displacing themajority of the organelles and even the nucleus. It is assumedthat for the generation of such a large vacuole the microor-ganism manipulates several intracellular trafficking pathwaysin order to obtain enough membrane. Cumulative evidenceindicates that both the endocytic and the phagocytic pathways

FIG. 7. Overexpression of Rab1b Q67L alters the fusogenic and replica-tive properties of the C. burnetii compartments. CHO cells stably overexpress-ing EGFP alone as a control or EGFP-Rab1b Q67L were infected with C.burnetii. At 48 h postinfection, the cells were incubated for 1 h with heat-inactivated Staphylococcus aureus-rhodamine or with 5 �g of dextran-rhoda-mine/ml to label the phagocytic and endocytic pathways, respectively. (A) Theupper panels show colocalization (see inset) of C. burnetii (blue) and S.aureus-rhodamine (red). In contrast, no colocalization of heat-inactivated S.aureus-rhodamine, and the bacterium is observed in the vacuoles of cellsoverexpressing Rab1b Q67L (lower panels), suggesting that these compart-ments have its fusogenic properties with the phagocytic pathway altered.(B) The upper panels show that in CHO cells overexpressing EGFP themajority of the Coxiella-vacuoles contain the endocytic marker dextran-rhodamine (red), whereas the absence of colocalization is observed in cellsoverexpressing Rab1b Q67L (lower panels), suggesting that the fusogeniccapacity with the endocytic pathway was altered. Bars, 20 �m. (C) Quantifi-cation of the percentage of colocalization of C. burnetii and heat-inactivatedStaphylococcus aureus-rhodamine or dextran-rhodamine. A total of 50 cellswere counted in each condition.

FIG. 8. Overexpression of a dominant-negative mutant of Sar1 pre-vents the recruitment of Rab1b Q67L to the C. burnetii vacuole mem-brane. CHO cells infected with C. burnetii for 24 h were cotransfectedwith pEGFP-Rab1b Q67L and pIRES-DsRed Sar1 T39N, pEGFP-Rab1b Q67L, and pIRES-DsRed Sar1 H79G or individually trans-fected with pEGFP, pIRES-DsRed Sar1 T39N, pIRES-DsRed Sar1H79G, or pEGFP-Rab1b Q67L as controls. At 24 h posttransfection,the cells were fixed and subjected to indirect immunofluorescence withspecific antibodies against C. burnetii (blue) and GM130 (red). Theimages were analyzed by confocal microscopy. (A) The insets in theleft panels show clearly the recruitment of Rab1b Q67L to the vacuolemembrane. In contrast, no recruitment of Rab1b Q67L to the CRV(see the inset) is observed under overexpression of Sar1 T39N (dasheddelineated cell). The middle panels show the different diameters of thevacuoles under Rab1b overexpression, coexpression of Rab1b Q67Land Sar1 T39N (dashed delineated cell), or untransfected cells (ar-rows). A disassembled Golgi apparatus is showing cells overexpressingSar1 T39N in contrast with an untransfected cell (arrowheads).(B) Quantification of the size of vacuoles at 48 h of infection from theexperiment presented in panel A. The data represent the means � theSEM of at least three independent experiments (P � 0.001). (C) Quan-tification of the vacuole size at 48 h postinfection from cells overex-pressing EGFP, EGFP-Rab1b Q67L, or DS-Red-Sar1 H69G or coex-pressing Rab1b Q67L and DsRed H69G. The data represent themeans � the SEM of at least three independent experiments (P �0.001). Bars, 15 �m.



.org/D

ownloaded from

http://iai.asm.org/

contribute to the development and expansion of the CRV (4,13, 16, 30). In the present study we present the first evidencethat the Coxiella-replicative niche also intercepts the early se-cretory pathway (see the model in Fig. 9). Our results indicatethat the CRVs are not only decorated by the small GTPaseRab1b and its GTPase-defective mutant but also that the for-mation of the CRV requires a functional Rab1b, a proteininvolved in ER-to-Golgi and intra-Golgi trafficking (27, 32).This close relationship with the secretory pathway seems tofacilitate expansion (see below) and contributes to the biogen-esis of the Coxiella-customized niche, which presents particularfeatures to facilitate bacterial growth.

The interaction between the secretory pathway and a patho-gen-containing compartment is not unique to Coxiella. Indeed,it is widely documented that the Legionella-containing vacuole(LCV), within 10 min of bacterial uptake recruits Rab1b andsubsequently interacts with vesicles derived from the early se-cretory pathway (8, 22, 42). However, in contrast to findingswith Legionella pneumophila, our results indicate that either inCHO cells or in Raw macrophages Rab1b is not recruited tothe Coxiella phagosome at early times after internalization(i.e., 1 h uptake) but when the Coxiella has developed largercompartments generated after 12 to 24 h infection. In the caseof Chlamydia trachomatis, the inclusion membrane is also dec-orated by Rab1b, as well as by Rab4 and Rab11 (7, 31),whereas neither Rab5 (early endosome marker) nor Rab7 (lateendosome marker) were recruited to the inclusion membrane.On the contrary, we have previously shown that both Rab5 andRab7 bind to the Coxiella phagosome at an early time afterinternalization (30), whereas no recruitment of Rab11 has

been observed at any time during infection (see Fig. S1A in thesupplemental material). Thus, all of these data indicate thateach microorganism recruits a subset of Rab proteins to spe-cifically interact with defined intracellular compartments, in atime window, in order to generate a self-tailored replicativeniche (for a review, see reference 5).

As mentioned above, Rab1b is a small GTPase that recruitsfactors necessary for the tethering and fusion of vesicles de-rived from the ER/intermediate compartment with targetmembranes (1, 25). We hypothesize that Rab1b is anchoredonto the CRV membrane to subsequently allow the tetheringof vesicles derived from the early secretory pathway. Thesevesicles may contribute with membranes to generate the spa-cious CRV or, alternatively, they may carry key fusion proteinssuch as SNAREs, i.e., transmembrane proteins required forspecific fusion processes (reviewed in reference 20). Indeed, inthe case of Legionella it has been shown that the ER-derivedvesicles transport the SNARE Sec22b to the LCV (22). Tothe best of our knowledge, no single SNARE molecule hasbeen yet identified in the Coxiella-containing vacuole. On-going work in our laboratory is currently designed to estab-lish some of the SNAREs molecules involved in Coxiellaintracellular trafficking.

We and others have demonstrated that the Coxiella-replica-tive compartment is highly fusogenic (4, 13, 16, 18, 30), andthese fusogenic properties are likely important for the gener-ation of the spacious CRV. We have observed that, in cellsoverexpressing the Rab1b mutant defective in GTPase activity(i.e., Rab1Q67L), the fusion capability of the CRVs with bothendocytic and phagocytic compartments was significantly im-paired. It is known that the dynamic association/dissociation ofa Rab protein is essential for an efficient transport through agiven pathway (38). As a compartment matures, the Rab com-position in specific domains changes, thus the dissociation of acertain Rab protein would allow the recruitment of anotherone, as well as its interacting proteins (e.g., effectors), leadingto new fusion and fission events favoring the maturation pro-cess. Therefore, it is likely that the irreversible association ofthe GTPase-deficient mutant of Rab1b to the CRV may ham-per the recruitment of a critical Rab or other key factorsrequired for fusion with endo/phagocytic compartments. Thesealterations in the fusion capability of the CRVs generated incells overexpressing the mutant Q67L would also explain theincreased vesicle number, albeit of smaller size, suggesting thathomotypic fusion between Coxiella vacuoles is also impeded.

There are some interesting similarities between the resultsobtained with Legionella and Coxiella regarding the relation-ship of both pathogens with the autophagic and secretory path-ways. It is well established that L. pneumophila exploits theautophagic pathway to generate an adequate replicative niche(10). At the ultrastructural level, the LCVs resemble nascentautophagosomes (36). In addition, the autophagic proteinsAtg7 and LC3 transiently associate with the pathogen-contain-ing vacuoles (3). Furthermore, amino acid depletion, whichstimulates autophagy, increases the association of the bacteriawith the ER and enhances bacterial growth. Besides, blockingthe early secretory pathway with Brefeldin A diminish thecolocalization of Atg7 and the LCV and prevents bacterialreplication (3), indicating a dynamic interaction between boththe secretory and autophagic pathways. Similarly, work from

FIG. 9. Model showing the interaction among C. burnetii and theendocytic, autophagic, and secretory pathways during the vacuole de-velopment process. Upon internalization, bacterium phagosomes in-teract with the autophagic pathway (the protein LC3 is present on thevacuole membrane). These Coxiella-containing phagosomes also inter-act with degradative organelles such as the lysosomes. These fusionevents contribute to generate a proper environment for the replicationprocess. The secretory pathway through Rab1b-labeled vesicles likelycontributes by supplying membrane to generate the spacious Coxiella-vacuole. The highly fusogenic properties of the Coxiella-phagosomeswould contribute not only to the formation of the large replicativeniche but also to the acquisition of key factors and nutrients to favorthe transformation of the bacteria into the replication-competent form(i.e., large cell variant).



.org/D

ownloaded from

http://iai.asm.org/

our laboratory indicates that the CRV is decorated by theautophagic protein LC3 (4). Physiological (i.e., starvation) orpharmacological (i.e., rapamycin treatment) induction ofautophagy significantly enhances C. burnetii replication andviability, as determined by an FFU assay (13), which indicatesthat autophagy favors the biogenesis of the Coxiella-replicativeniche. Our present results also indicate that the CRV recruitsRab1b and that adequate levels of this protein are critical forCoxiella replication since knockdown of this protein markedlyreduces bacterial growth. Previous evidences indicate that theoverexpression of Rab1b Q67L reverts the BFA effect (24). Inagreement with this result, we observed that under BFA treat-ment the recruitment of Rab1b Q67L to the vacuole mem-brane was not altered. In contrast, Rab1b Q67L was not re-cruited to the CRV in cells overexpressing neither Sar1 T39Nnor Sar1 H79G, indicating that vesicles produced by the Sar1/COPII system are necessary for transport of Rab1b to theCRV. This evidence agrees with the observation that in cellsoverexpressing the Sar1 H79G protein there is no tetheringbetween ER vesicles and the LCV in L. pneumophila-infectedcells (29). In summary, our results support the idea that theinteraction with ER-derived vesicles is critical for the genera-tion of the large CRV and sustain the hypothesis that the ERmight be the source of autophagosome membranes (11, 40).Indeed, recently published work support this hypothesis sincestudies by electron tomography have revealed that both ERand the phagophore/isolation membranes are interconnected(14, 47). Moreover, we have recently found that autophago-some formation depends on the small GTPase Rab1b andfunctional ER exit sites (50). Our present results indicate thatvesicles departing from ER exit sites intercept the Coxiella-replicative compartment. However, further studies are cer-tainly needed to fully understand the molecular interplay be-tween transport from the ER and the autophagic pathway inCoxiella-vacuole development.

ACKNOWLEDGMENTS

We are grateful to Hirofumi Kai and Akiko Niibori (Faculty ofPharmaceutical Sciences, Kumamoto University, Kumamoto, Japan)for providing the DrpIRES2DsRed2 bicistronic vectors encoding Sar1,Sar1 H79G, and Sar1 T39N. We thank Luis Mayorga and WalterBeron for critically reading of the manuscript. We also thank Alejan-dra Medero for technical assistance with tissue culture.

REFERENCES

1. Allan, B. B., B. D. Moyer, and W. E. Balch. 2000. Rab1 recruitment of p115into a cis-SNARE complex: programming budding COPII vesicles for fusion.Science 289:444–448.

2. Alvarez, C., R. Garcia-Mata, E. Brandon, and E. Sztul. 2003. COPI recruit-ment is modulated by a Rab1b-dependent mechanism. Mol. Biol. Cell 14:2116–2127.

3. Amer, A. O., and M. S. Swanson. 2005. Autophagy is an immediate macro-phage response to Legionella pneumophila. Cell Microbiol. 7:765–778.

4. Beron, W., M. G. Gutierrez, M. Rabinovitch, and M. I. Colombo. 2002.Coxiella burnetii localizes in a Rab7-labeled compartment with autophagiccharacteristics. Infect. Immun. 70:5816–5821.

5. Brumell, J. H., and M. A. Scidmore. 2007. Manipulation of Rab GTPasefunction by intracellular bacterial pathogens. Microbiol. Mol. Biol. Rev.71:636–652.

6. Colombo, M. I., M. G. Gutierrez, and P. S. Romano. 2006. The two faces ofautophagy: Coxiella and Mycobacterium. Autophagy 2:162–164.

7. Cortes, C., K. A. Rzomp, A. Tvinnereim, M. A. Scidmore, and B. Wizel. 2007.Chlamydia pneumoniae inclusion membrane protein Cpn0585 interacts withmultiple Rab GTPases. Infect. Immun. 75:5586–5596.

8. Derre, I., and R. R. Isberg. 2004. Legionella pneumophila replication vacuoleformation involves rapid recruitment of proteins of the early secretory sys-tem. Infect. Immun. 72:3048–3053.

9. Donaldson, J. G., D. Finazzi, and R. D. Klausner. 1992. Brefeldin A inhibitsGolgi membrane-catalyzed exchange of guanine nucleotide onto ARF pro-tein. Nature 360:350–352.

10. Dubuisson, J. F., and M. S. Swanson. 2006. Mouse infection by Legionella, amodel to analyze autophagy. Autophagy 2:179–182.

11. Dunn, W. A., Jr. 1990. Studies on the mechanisms of autophagy: formationof the autophagic vacuole. J. Cell Biol. 110:1923–1933.

12. Griffiths, G., M. Ericsson, J. Krijnse-Locker, T. Nilsson, B. Goud, H. D.Soling, B. L. Tang, S. H. Wong, and W. Hong. 1994. Localization of the Lys,Asp, Glu, Leu tetrapeptide receptor to the Golgi complex and the interme-diate compartment in mammalian cells. J. Cell Biol. 127:1557–1574.

13. Gutierrez, M. G., C. L. Vazquez, D. B. Munafo, F. C. Zoppino, W. Beron, M.Rabinovitch, and M. I. Colombo. 2005. Autophagy induction favours thegeneration and maturation of the Coxiella-replicative vacuoles. Cell Micro-biol. 7:981–993.

13a.Hackstadt, T., R. Messer, W. Cieplak, and M. G. Peacock. 1992. Evidence forthe proteolytic cleavage of the 120-kilodalton outer membrane protein ofrickettsiae: identification of an avirulent mutant deficient in processing.Infect. Immun. 60:159–165.

14. Hayashi-Nishino, M., N. Fujita, T. Noda, A. Yamaguchi, T. Yoshimori, andA. Yamamoto. 2009. A subdomain of the endoplasmic reticulum forms acradle for autophagosome formation. Nat. Cell Biol. 11:1433–1437.

15. Heinzen, R. A., T. Hackstadt, and J. E. Samuel. 1999. Developmental biologyof Coxiella burnetii. Trends Microbiol. 7:149–154.

16. Heinzen, R. A., M. A. Scidmore, D. D. Rockey, and T. Hackstadt. 1996.Differential interaction with endocytic and exocytic pathways distinguishparasitophorous vacuoles of Coxiella burnetii and Chlamydia trachomatis.Infect. Immun. 64:796–809.

17. Horwitz, M. A. 1983. The Legionnaires’ disease bacterium (Legionella pneu-mophila) inhibits phagosome-lysosome fusion in human monocytes. J. Exp.Med. 158:2108–2126.

18. Howe, D., J. Melnicakova, I. Barak, and R. A. Heinzen. 2003. Fusogenicity ofthe Coxiella burnetii parasitophorous vacuole. Ann. N. Y. Acad. Sci. 990:556–562.

19. Hutt, D. M., and W. E. Balch. 2008. Rab1b silencing using small interferingRNA for analysis of disease-specific function. Methods Enzymol. 438:1–10.

20. Jahn, R., and R. H. Scheller. 2006. SNAREs: engines for membrane fusion.Nat. Rev. Mol. Cell. Biol. 7:631–643.

21. Kagan, J. C., and C. R. Roy. 2002. Legionella phagosomes intercept vesiculartraffic from endoplasmic reticulum exit sites. Nat. Cell Biol. 4:945–954.

22. Kagan, J. C., M. P. Stein, M. Pypaert, and C. R. Roy. 2004. Legionella subvertthe functions of Rab1 and Sec22b to create a replicative organelle. J. Exp.Med. 199:1201–1211.

23. Kuge, O., C. Dascher, L. Orci, T. Rowe, M. Amherdt, H. Plutner, M. Ravaz-zola, G. Tanigawa, J. E. Rothman, and W. E. Balch. 1994. Sar1 promotesvesicle budding from the endoplasmic reticulum but not Golgi compart-ments. J. Cell Biol. 125:51–65.

24. Monetta, P., I. Slavin, N. Romero, and C. Alvarez. 2007. Rab1b interacts withGBF1 and modulates both ARF1 dynamics and COPI association. Mol. Biol.Cell 18:2400–2410.

25. Moyer, B. D., B. B. Allan, and W. E. Balch. 2001. Rab1 interaction with aGM130 effector complex regulates COPII vesicle cis–Golgi tethering. Traffic2:268–276.

26. Peyroche, A., B. Antonny, S. Robineau, J. Acker, J. Cherfils, and C. L.Jackson. 1999. Brefeldin A acts to stabilize an abortive ARF-GDP-Sec7domain protein complex: involvement of specific residues of the Sec7 do-main. Mol. Cell 3:275–285.

27. Plutner, H., A. D. Cox, S. Pind, R. Khosravi-Far, J. R. Bourne, R.Schwaninger, C. J. Der, and W. E. Balch. 1991. Rab1b regulates vesiculartransport between the endoplasmic reticulum and successive Golgi compart-ments. J. Cell Biol. 115:31–43.

28. Ren, M., G. Xu, J. Zeng, C. De Lemos-Chiarandini, M. Adesnik, and D. D.Sabatini. 1998. Hydrolysis of GTP on rab11 is required for the direct deliveryof transferrin from the pericentriolar recycling compartment to the cellsurface but not from sorting endosomes. Proc. Natl. Acad. Sci. U. S. A.95:6187–6192.

29. Robinson, C. G., and C. R. Roy. 2006. Attachment and fusion of endoplasmicreticulum with vacuoles containing Legionella pneumophila. Cell Microbiol.8:793–805.

30. Romano, P. S., M. G. Gutierrez, W. Beron, M. Rabinovitch, and M. I.Colombo. 2007. The autophagic pathway is actively modulated by phase IICoxiella burnetii to efficiently replicate in the host cell. Cell Microbiol. 9:891–909.

31. Rzomp, K. A., L. D. Scholtes, B. J. Briggs, G. R. Whittaker, and M. A.Scidmore. 2003. Rab GTPases are recruited to chlamydial inclusions in botha species-dependent and species-independent manner. Infect. Immun. 71:5855–5870.

32. Saraste, J., U. Lahtinen, and B. Goud. 1995. Localization of the smallGTP-binding protein rab1p to early compartments of the secretory pathway.J. Cell Sci. 108(Pt. 4):1541–1552.

33. Schwartz, S. L., C. Cao, O. Pylypenko, A. Rak, and A. Wandinger-Ness. 2007.Rab GTPases at a glance. J. Cell Sci. 120:3905–3910.



.org/D

ownloaded from

http://iai.asm.org/

34. Stenmark, H. 2009. Rab GTPases as coordinators of vesicle traffic. Nat. Rev.Mol. Cell. Biol. 10:513–525.

35. Stroud, W. J., S. Jiang, G. Jack, and B. Storrie. 2003. Persistence of Golgimatrix distribution exhibits the same dependence on Sar1p activity as a Golgiglycosyltransferase. Traffic 4:631–641.

36. Swanson, M. S., and R. R. Isberg. 1995. Association of Legionella pneumo-phila with the macrophage endoplasmic reticulum. Infect. Immun. 63:3609–3620.

37. Takai, Y., T. Sasaki, and T. Matozaki. 2001. Small GTP-binding proteins.Physiol. Rev. 81:153–208.

38. Tavitian, A. 1994. Cellular traffic and markers of subcellular compartments.Nouv. Rev. Fr. Hematol. 36(Suppl. 1):S29–S32.

39. Tisdale, E. J., J. R. Bourne, R. Khosravi-Far, C. J. Der, and W. E. Balch.1992. GTP-binding mutants of rab1 and rab2 are potent inhibitors of vesic-ular transport from the endoplasmic reticulum to the Golgi complex. J. CellBiol. 119:749–761.

40. Ueno, T., D. Muno, and E. Kominami. 1991. Membrane markers of endo-plasmic reticulum preserved in autophagic vacuolar membranes isolatedfrom leupeptin-administered rat liver. J. Biol. Chem. 266:18995–18999.

41. Ullrich, O., S. Reinsch, S. Urbe, M. Zerial, and R. G. Parton. 1996. Rab11regulates recycling through the pericentriolar recycling endosome. J. CellBiol. 135:913–924.

42. Urwyler, S., E. Brombacher, and H. Hilbi. 2009. Endosomal and secretorymarkers of the Legionella-containing vacuole. Commun. Integr. Biol. 2:107–109.

43. Veras, P. S., C. C. de, M. F. Moreau, V. Villiers, M. Thibon, D. Mattei, andM. Rabinovitch. 1994. Fusion between large phagocytic vesicles: targeting ofyeast and other particulates to phagolysosomes that shelter the bacterium

Coxiella burnetii or the protozoan Leishmania amazonensis in Chinese ham-ster ovary cells. J. Cell Sci. 107(Pt. 11):3065–3076.

44. Veras, P. S., C. Moulia, C. Dauguet, C. T. Tunis, M. Thibon, and M.Rabinovitch. 1995. Entry and survival of Leishmania amazonensis amasti-gotes within phagolysosome-like vacuoles that shelter Coxiella burnetii inChinese hamster ovary cells. Infect. Immun. 63:3502–3506.

45. Voth, D. E., and R. A. Heinzen. 2007. Lounging in a lysosome: the intracel-lular lifestyle of Coxiella burnetii. Cell Microbiol. 9:829–840.

46. Wilcke, M., L. Johannes, T. Galli, V. Mayau, B. Goud, and J. Salamero.2000. Rab11 regulates the compartmentalization of early endosomes re-quired for efficient transport from early endosomes to the trans-Golgi net-work. J. Cell Biol. 151:1207–1220.

47. Yla-Anttila, P., H. Vihinen, E. Jokitalo, and E. L. Eskelinen. 2009. 3Dtomography reveals connections between the phagophore and endoplasmicreticulum. Autophagy 5:1180–1185.

48. Yoshimura, S., A. Yamamoto, Y. Misumi, M. Sohda, F. A. Barr, G. Fujii, A.Shakoori, H. Ohno, K. Mihara, and N. Nakamura. 2004. Dynamics of Golgimatrix proteins after the blockage of ER to Golgi transport. J. Biochem.135:201–216.

48a.Zamboni, D. S., R. A. Mortara, and M. Rabinovitch. 2001. Infection of Verocells with Coxiella burnetii phase II: relative intracellular bacterial load anddistribution estimated by confocal laser scanning microscopy and morpho-metry. J. Microbiol. Methods 43:223–232.

49. Zerial, M., and H. McBride. 2001. Rab proteins as membrane organizers.Nat. Rev. Mol. Cell. Biol. 2:107–117.

50. Zoppino, F. C., R. D. Militello, I. Slavin, C. Alvarez, and M. I. Colombo.2010. Autophagosome formation depends on the small GTPase Rab1 andfunctional ER exit sites. Traffic 11:1246–1261.

Editor: R. P. Morrison



.org/D

ownloaded from

http://iai.asm.org/

the early secretory pathway contributes to the growth of ... · mids (1 mg ml1) using the...

Documents