Hana Kovárová1

Petr Halada2

Petr Man3

Igor Golovliov4

Zuzana Krocová1

Josef Spacek5

Stanislava Porkertová1

Radka Necasová1

1Institute of Radiobiology andImmunology and ProteomeCentre for the Study ofIntracellular Parasitism, PurkyneMilitary Medical Academy,Hradec Králové, Czech Republic

2Institute of Microbiology,Academy of Sciences ofthe Czech Republic,Prague, Czech Republic

3Faculty of Science,Charles University,Prague, Czech Republic

4Umea University,Umea, Sweden

5Faculty of Medicine,Charles University,Prague, Czech Republic

Proteome study of Francisella tularensis live vaccinestrain-containing phagosome in Bcg/Nramp1congenic macrophages: Resistant allelecontributes to permissive environment andsusceptibility to infection

The phagocytosis of pathogens by macrophages classically initiates maturation of thephagosome that involves a dynamic exchange of phagosomal components with intra-cellular compartments of the endocytic pathway. The intracellular microorganismshave developed sophisticated mechanisms to sense environmental conditions andrespond to them by phenotypic alterations that ensure their adaptation, survival andproliferation inside the cell. They have learned also to utilise host cellular componentsto ensure own survival. Recent results suggest that the Bcg locus/Nramp1 gene (nat-ural resistance-associated macrophage protein 1) controls natural resistance to infec-tion by Francisella tularensis LVS (live vaccine strain) and its effect is opposite to thatobserved for other Bcg/Nramp1-controlled pathogens such as several mycobacterialspecies, Leischmania donovani, and Salmonella typhimurium. In the case of F. tularen-sis LVS infection, the mutant allele of the Bcg locus (Bcgs/Nramp1s) is associated withnatural resistance and, inversely, the wild type allele (Bcgr/Nramp1r) confers suscept-ibility. To determine whether differential allelic expression of the Bcg locus/Nramp1gene modifies the composition of F. tularensis LVS-containing phagosomes (FCP), wehave utilised an approach where we isolated FCP from infected Bcg congenic B10R(Bcgr/Nramp1r) and B10S (Bcgs/Nramp1s) macrophages of susceptible and resistantphenotype, respectively. Comparative proteomic analysis of the two phagosomalcompartments with subsequent mass spectrometric analysis allowed identification ofseveral proteins typical for FCP from B10R macrophages. They include a bacterialhypothetical 23 kDa protein, 60 kDa chaperonin GroEL, and host putative proteinsthat appeared to be mitochondrial ATP synthase �-chain and NADH-ubiquinone oxi-doreductase based on high cross-species homology. High abundance of the hypo-thetical 23 kDa protein correlates with the susceptible phenotype and, possibly, patho-genicity of F. tularensis LVS. The results demonstrate that F. tularensis LVS can exploition transport function of Bcg/Nramp1 to its own advantage.

Keywords: Innate immunity / Phagosome / Francisella tularensis / Bcg locus / Nramp1 / Two-dimensional gel electrophoresis / Mass spectrometry PRO 0147

1 Introduction

The phagocytosis of pathogens by macrophages classi-cally initiates maturation of the phagosome that involves adynamic exchange of phagosomal components withintracellular compartments of the endocytic pathway.Ultimately, the phagosome-lysosome fusion takes place

and delivery of lysosomal enzymes causes destruction ofpathogens [1]. This highly complex process is required tomount a successful host defence against microbes withthe ability to survive and multiply within phagocytic cells,so called intracellular microorganisms [2]. Commonlyused survival strategies for intracellular microorganismsinvolve inhibition of phagosome-lysosome fusion, escapefrom, or physical resistance to the toxic lysosomal envir-onment [3, 4]. The nature of phagosomal microenviron-ment can be manipulated by intracellular microorganismsand they have developed sophisticated mechanisms tosense environmental conditions and respond to it byphenotypic alterations that ensure their adaptation, sur-vival and proliferation inside the cell [5]. In addition, manyof intracellular bacteria have learned to utilise host cellu-lar components to ensure their own survival [6].

Correspondence: Dr. Hana Kovárová, Institute of Radiobiologyand Immunology, Purkyne Military Medical Academy, Trebesská1575, 500 01 Hradec Králové, Czech RepublicE-mail: kovarova@pmfhk.czFax: +420-49-551-3018

Abbreviations: BCA, bicinchoninic acid assay; FCP, F. tularensisLVS-containing phagosome; GFP, green fluorescence protein;LVS, live vaccine strain

Proteomics 2002, 2, 85–93 85

ª WILEY-VCH Verlag GmbH, 69451 Weinheim, 2002 1615-9853/02/0101–85 $17.50+.50/0

In spite of the complexity of host-pathogen interactions,innate immunity to pathogens can be in many instancesclassified simply in terms of susceptibility or resistance.Mouse Nramp1, originally known as Bcg/Ity/Lsh, is oneof the best studied genes contributing to the innate con-trol of infections caused by unrelated intracellular patho-gens such as Mycobacterium bovis BCG and severalother mycobacterial species, Salmonella typhimuriumand Leishmania donovani [7, 8]. Nramp1 is a member ofan ancient family of divalent cation transporters with thetypical structure of integral membrane proteins [9].Recent studies have demonstrated a role of the Nramp 1protein in transportation and distribution of iron and/ormanganese in the macrophage [10–12]. Nramp1 is local-ised in late endosomal and lysosomal compartments ofmacrophages and, after phagocytosis, targeted to pha-gosomal membrane [13, 14]. These findings indicate thatNramp1 can either directly influence the microenviron-ment of the host cell phagosome by its capability to fuseor converge with late endosomes/lysosomes or play arole in the regulation of fusion events and functions ofendocytic organelles, most probably due to the localchanges in the concentration of divalent ions.

Francisella tularensis, the best characterized member ofthe genus Francisella, is a facultative intracellular bacter-ium and etiological agent of the tularemia, a diseaseoccuring in a variety of animal hosts. The interaction ofinfected macrophages with Francisella-specific T lym-phocytes contributes significantly to the establishmentof protective immunity and ultimate clearance of thebacterium [15]. However, intracellular trafficking of Fran-cisella in macrophages, and the early innate phase ofthe host immune response remains largely unknownalthough it was shown that live attenuated variant of F.tularensis denoted LVS (live vaccine strain), virulent forvarious strains of mice, survive within membrane-boundphagocytic vesicules and lysosomes fail to fuse withphagosomes containing F. tularensis LVS [16]. Recentresults suggest that the Bcg locus, which is close oridentical to the Nramp1 gene, controls natural resistanceto infection by F. tularensis LVS and, interestingly, itseffect is opposite to that observed for other pathogenssuch as mycobacterial species, S. typhimurium and L.donovani. The Nramp1 gene is expressed in murineinbred strains in two allelic forms, wild type (Bcgr/Nramp1r) and mutant (Bcgs/Nramp1s); the latter has169Gly-to-169Asp substitution within predicted trans-membrane domain 4 which confers susceptibility toNramp1-controlled infections [17]. In the case of F. tular-ensis LVS infection, the mutant allele of the Bcg locus isassociated with natural resistance and, inversely, thewild type allele confers susceptibility [18]. Thus, theexpression of the mutant Bcgs/Nramp1s allele could

afford protection against F. tularensis LVS while the pres-ence of the wild type Bcgr/Nramp1r allele could mediateenvironment permissive to bacteria.

To determine whether differential allelic expression of theBcg locus modifies the composition of F. tularensis LVS-containing phagosomes (FCP), we have utilised anapproach where we isolated FCP from infected Bcg con-genic B10R (Bcgr/Nramp1r) and B10S (Bcgs/Nramp1s)macrophages. We assumed that comparative proteomicanalysis of the two phagosomal compartments mightallow identification of macrophage and bacterial proteinsthat contribute to a permissive phagosomal environmentfor bacterial growth in B10R macrophages or a hostileenvironment associated with degradation of the patho-gen in B10S cells. Here we describe differential expres-sion of several proteins associated with the susceptiblephenotype and, possibly, pathogenicity of F. tularensis.

2 Materials and methods

2.1 Materials

Immobiline DryStrips (IPG) pH 3–10 NL, 18 cm, Pharma-lytes pH 8–10.5 and pH 3–10 were purchased fromAmersham Pharmacia Biotech (Uppsala, Sweden);CHAPS, urea, DTTand acrylamide were from USB (Amer-sham Pharmacia Biotech); Tris base, Tris-HCl, agarose,iodoacetamide, thiourea, glycin, silver nitrate, bicinchon-inic acid (BCA) assay, p-nitrophenyl-N-acetyl-�-glucos-aminide, EDTA, EGTA, sucrose, tetracykline, HEPES, andTFA were from Sigma (St. Louis, MO, USA); piperazine-diacrylamide (PDA), ammonium persulfate (APS), SDSand TEMED were from Bio-Rad (Richmond, CA., USA);tributyl phosphine (TBP) was purchased from Fluka(Buchs, Switzerland); protease inhibitor complete tabletsand PVDF membrane were from Boehringer (Mannheim,Germany).

2.2 Experimental infection of macrophages

F. tularensis LVS (ATCC 29684, American Type CultureCollection, Manassas, VA, USA) harboring green fluores-cence protein (LVS-GFP) was kindly provided by AndersSjöstedt (Umea University, Umea, Sweden) [19–21] andcultured on McLeod agar plates supplemented withbovine haemoglobin, Iso VitaleX (Becton-Dickinson,Heidelberg, Germany) and tetracycline.

Macrophage cell lines B10R (Bcgr/Nramp1r, wild typeallele) and B10S (Bcgs/Nramp1s, mutated allele) werederived by immortalizing bone marrow macrophagesfrom B10.A.Bcgr and B10.A (Bcgs) mice congenic in theBcg locus. Studies on surface markers and functional

86 H. Kovárová et al. Proteomics 2002, 2, 85–93

activities of those macrophage lines confirmed that theyfaithfully reflect the morphological and functional proper-ties of native macrophages from M. bovis BCG-resistantand -susceptible mouse strains [22]. Cells were grown inDulbecco’s modified essential medium containing 4.5 g/Lglucose and glutamax I (L-alanyl-L-glutamine) supple-mented with 10% heat inactivated fetal calf serum(DMEM-FCS, GibcoBRL, Karlsruhe, Germany) at 37�Cand 5% CO2 for 48 h before infection. Cell suspensionswere collected, washed by PBS and adjusted to densityof 5�106 cells/mL in DMEM-FCS culture medium in poly-propylene tubes. Macrophages were infected at a multi-plicity of infection (MOI) of 100 for 2 h at 37�C and 5%CO2. Following this period, cells were washed three timeswith PBS, resuspended in DMEM-FCS medium and cul-tured for 48 h. Immediately after the internalization periodas well as after the incubation of 48 h, cell associatedbacteria were enumerated by lysing aliquots of cells inPBS containing 0.02% SDS and culturing serially dilutedlysates on McLeod plates.

2.3 Preparation of F. tularensis LVS phagosomalcompartment

Separation of FCP was performed as described by Chak-raborty [23] with slight modifications. Infected B10R andB10S macrophages (1�108) were lysed in homogeniza-tion buffer (25 mM sucrose, 20 mM HEPES, 0.5 mM EDTA,0.5 mM EGTA, protease inhibitor cocktail, pH 7.0) using aDounce homogenizer on ice. After repeated centrifuga-tion three times at 300�g for 10 min at 4�C, the postnuc-lear supernatant was pooled and centrifuged at 8000�gfor 10 min at 4�C to spin down mostly FCP. Sediment wasresuspended in gradient buffer (20 mM HEPES, 0.5 mM

EDTA, 0.5 mM EGTA, protease inhibitor cocktail, pH 7.0)and subjected to velocity sedimentation (1500�g, 1 h,4�C) through a 20/35/50% w/v sucrose step gradient.Fractions of 0.5 mL were collected from the top of the gra-dient and stored at –80�C until analysis.

2.4 Flow cytometry and analysisof �-hexosaminidase activityin gradient fractions

Flow cytometry was employed as a method to check thepresence of FCP containing vacuoles in the gradient frac-tions. Collected fractions were analysed using a EPICSXL flow cytometer (Coulter Corporation, Hielah, FL, USA)equipped with argon laser turned to 15 milliwatts of out-put at 488 nm. Fluorescence was measured using a530 nm pass filter. Lysosomal �-hexosaminidase activitywas detected using p-nitrophenyl-N-acetyl-�-glucos-aminide as described by Landegren [24].

2.5 Transmission electron microscopyof isolated F. tularensis phagosomalcompartment

Gradient fractions containing FCP as verified by FACSanalysis were fixed in phosphate-buffered fixative con-taining 2.5% glutaraldehyde, 2% paraformaldehyde and2 mM calcium chloride pH 7.35 for 1 h at 25�C, washedand embedded in 2% agar. Agar blocks containing pelletsof phagosomes were immersed in 2% osmium tetroxidefor 1 h, dehydrated in graded ethanol and embedded inepoxy resin. Ultrathin sections were cut and examinedby electron microscopy (Philips EM208; FEI, Eindhoven,The Netherlands).

2.6 2-DE

For 2-DE analysis, the proteins from pooled fractions con-taining FCP were precipitated by chloroform-methanol[25] and dissolved in 2-DE sample buffer (9 M urea, 3%CHAPS, 2% NP-40, 70 mM DTTand 2% carrier ampholytepH 8–10.5, trace of bromphenol blue) and centrifuged for 5min in 12 000 rpm in a microfuge (Joan A-14, Jouan, Sain-Herblain, France). Protein concentrations in the sampleswere measured using BCA assay [26]. Separations wereperformed as described by Hochstrasser [27]. Briefly, IPGstrips were rehydrated in buffer containing 2 M thiourea, 5 M

urea, 2% CHAPS, 2% SB 3–10 (N-decyl-N,N,-dimethyl-3-amonio-propanesulfonate), 2 mM TBP, 40 mM Tris, and0.5% carrier ampholyte pH 3–10 overnight. Samples equalto 50 �g and 150 �g of proteins for analytical run and massspectrometric identification, respectively, were loaded onIPG strips by adding into the sample cups positioned oncathodic edges of the strips. The isoelectric focusing withimmobilized sigmoid pH 3.5–10 gradient was carried out ina Multiphor apparatus (Amersham Pharmacia Biotech)with a 5000 V power supply (Serva, Heidelberg, Germany).The second dimension 9–16% polyacrylamide gradientgels were run using a Protean II xi 2D multi cell (Bio-Rad).Protein spots were visualized by sensitive ammoniacal sil-ver staining [28] or Coomassie Brilliant Blue R-250 (CBB).The gels were scanned using a laser densitometer (Per-sonal Densitometer, Molecular Dynamics, Sunnyvale, CA,USA, 4000�5000 pixels, 12 bits/pixel; stored on 16 bits)generating 16 megabyte images which were then trans-ferred to a SUN workstation for analysis with Melanie 3software (Bio-Rad). Proteins separated by 2-DE werequantitated in terms of their relative spot volumes. Corre-spondence analysis implemented in Melanie 3 was usedfor objective classification of samples. The isoelectricpoints and molecular weights were estimated using poly-peptide SDS-PAGE standards (Bio-Rad).

Proteomics 2002, 2, 85–93 Bcg/Nramp1modifies F. tularensis LVS-containing phagosome 87

2.7 Tryptic digestion and MALDI-MS

CBB stained protein spots were cut from the gel andwashed several times with 10 mM DDTand 50 mM 4-ethyl-morpholine acetate (pH 8.1) in 50% acetonitrile (ACN).The gel pieces were shrunk in ACN, washed with water,partly dried in a SpeedVac centrifuge and reconstitutedin a cleavage buffer containing 0.01% 2-mercaptoethanol,0.1 M 4-ethylmorpholine acetate, 1 mM CaCl2, 10% ACNand sequencing grade trypsin (50 ng/�L; Promega,Madison, WI, USA). Resulting peptides were extractedin 40% ACN/0.5% TFA after overnight digestion. Asaturated solution of �-cyano-4-hydroxycinnamic acid(Sigma, Steinheim, Germany) in aqueous 50% ACN/0.2% TFA was used as a MALDI matrix. Two �L of sampleand 2 �L of matrix solution were premixed in a tube, 0.5 �Lof the mixture was placed on the sample target andallowed to dry at the ambient temperature. Positive ionMALDI mass spectra were measured on a Bruker BIFLEXII reflectron time of flight mass spectrometer (Bruker-Franzen, Bremen, Germany) equipped with a multiprobesample inlet, a gridless delayed extraction ion source anda nitrogen laser (337 nm) (Laser Science, Cambridge, MA,USA). Ion acceleration voltage was 19 kV and the reflec-tron voltage was set to 20 kV. The spectrometer was cali-brated externally using the monoisotopic [M+H]+ ions ofpeptide standards angiotensin II and insulin (Sigma). Pro-tein spots were identified by searching for their peptidemass fingerprints in the nonredundant database NCBIusing the search program ProFound [29].

3 Results and discussion

We performed comparative 2-DE analysis of the proteincomposition of F. tularensis LVS-containing phagosomalcompartment isolated from infected B10R (Bcgr/Nramp1r)macrophages of F. tularensis LVS susceptible phenotypeand the “pseudophagosomal” compartment from in-fected Bcg congenic B10S (Bcgs/Nramp1s) macrophagesof F. tularensis LVS resistant phenotype. The “pseudo-phagosome” was pooled from the fractions that sedimentat the same position in the sucrose gradient as the FCPfrom B10R cells. We assumed that this comparisonenabled the elimination of possible nonphagosomal pro-teins that were contaminating the phagosomal fraction.Several proteins that were typically associated with thepermissive environment of B10R-FCP were selected andidentified. Since Bcg/Nramp1 congenic macrophagesrepresent a homogenous cell population, the differentialprotein expression in resistant and susceptible cellsupon infection reflects the interaction of F. tularensisLVS and the gene product(s) of the Bcg locus/Nramp1gene.

3.1 Characterization of sucrose gradientfractions

To analyse proteins of FCP, subcellular fractionation wasutilised. Macrophages were infected with live F. tularensisLVS for 2 h and, after removal of extracellular bacteria bywashing, further incubated for 48 h. The enumeration ofcell-associated bacteria was performed immediately after2 h internalization period and at the end of cultivation ofinfected cells for 48 h. As shown in Fig. 1, the numberof bacteria increased substantially in susceptible B10Rmacrophages compared to resistant B10S counterparts.This cultivation time period allowed isolation of sub-cellular organelles in sufficient quantities to permit bio-chemical analyses. Before gradient separation, the centri-fugation of the postnuclear fraction was performed thatallowed elimination of substantial amounts of the micro-somes/endoplasmic reticulum. Enrichment of FCP wasachieved by velocity sedimentation through the sucrosestep gradient. Identitication of FCP was facilitated usingF. tularensis LVS expressing green fluorescent protein(GFP). During gradient separation, the majority of GFP-labeled FCP present in susceptible B10R macrophages,but not in resistant B10S counterparts, were shiftedtoward interface 20/35% sucrose together with a portionof the total loaded protein (Fig. 2A and B). The phagoso-mal fraction appeared to be separated from lysosomesthat remain mostly on the interface sample/20% sucroseon the top of the column as assayed by the enzymaticactivity of the lysosomal marker �-hexosaminidase(Fig. 2A). Fractions 5–7 that displayed the highest GFP

Figure 1. Growth of F. tularensis LVS in congenic B10R(Bcgr/Nramp1r) and B10S (Bcgs/Nramp1s) macrophages.Macrophages were infected at a multiplicity of infection of100 for 2 h at 37�C and 5% CO2. Bacterial growth wasmonitored immediately after 2 h internalization (time ofinfection 0 h) and at the end of cultivation of infected cellsfor 48 h. Significant levels of differences between suscep-tible B10R macrophages and resistant B10S counterpartsat a given time point were determined for three independ-ent experiments by t-test and are indicated: *, p � 0.05.

88 H. Kovárová et al. Proteomics 2002, 2, 85–93

Figure 2. Separation of F. tularensis LVS-containing phagosomes by velocity sedimentation throughthe sucrose gradient. A, B10R and B10S macrophages were infected with F. tularensis LVS andphagosomal compartments were isolated 48 h postinfection. Typical profiles of vesicules are shownby fluorescence of GFP-labeled Francisella monitored by flow cytometry (solid line). Distribution oflysosomes was measured by �-hexosaminidase activity (broken line). The maximum of GFP fluores-cence corresponds to the fractions (fractions 5–7) enriched for F. tularensis LVS-containing phago-somes from B10R (B10R-FCP). Corresponding fractions in B10S contain only low fluorescence.B, Protein amount through the gradient. C, Transmission electron microscopy of B10R-FCP showingintact phagosomal membrane and partial contamination of B10R-FCP fraction.

fluorescence and were significantly enriched in FCP ofB10R macrophages (B10R-FCP) together with similarlyisolated “pseudophagosomes” from B10S cells (B10S-FCP) were used in further studies. Morphological analysisof B10R-FCP by electron microscopy showed that thephagosomal membranes were intact and indicated partialcontamination with other cellular components (Fig. 2C).

3.2 2-DE protein patterns ofF. tularensis-containing compartments

3.2.1 Objective classification of the gel samplesby correspondence analysis andidentification of host proteins

First, we performed multivariate correspondence analysisusing numeric taxonomy for computerized classificationof all analysed samples [30], e.g. two independently pre-pared B10R-FCP as well as B10S-FCP gels. Approxi-mately 20% of the total number of polypeptide spotswere successfully matched in all four gels and their valuesof relative volumes (% vol) served as the source for calcu-lations of correspondence analysis. Fig. 3C illustrates the

spatial distribution of gel samples. It is possible to recog-nise two groups of protein patterns: B10R-FCP is on theleft side, while B10S-FCP is on the right side along factor 1axis. The larger distance between these two groups indi-cates the dissimilarity in polypeptide abundance. Relativeabundance of “discriminant spots” selected by computeris demonstrated by histograms shown in Fig. 3B andthese spots are highlighted on gel images in Fig. 3A. Thesimplest explanation for the presence of spots commonlyshared between B10R-FCP and B10S-FCP is contam-ination of gradient fractions by other organelles. Indeed,two of these proteins, 189 and 808, were identified asmouse putative proteins and based on high cross-spe-cies homology, they appeared to be mitochondrial ATPsynthase �-chain and NADH-ubiquinone oxidoreductase,respectively (Fig. 3A and Table 1). In addition, we couldidentify calreticulin, a marker protein of endoplasmic retic-ulum, that was matched with reference mouse liver pro-tein map in the SWISS-2DPAGE database [31] (Fig. 2A).Nevertheless, the possibility that the protein spotsselected by correspondence analysis may be common,but differentially expressed in two different subcellularcompartments, cannot be excluded. These candidatespots may represent cellular proteins that can be more

Proteomics 2002, 2, 85–93 Bcg/Nramp1modifies F. tularensis LVS-containing phagosome 89

Figure 3. Correspondence analysis of gel images. A, The images of silver stained gels of F. tularensisLVS-containing phagosomes from B10R (B10R-FCP) and “pseudophagosomes” from B10S (B10S-FCP) macrophages with the positions of the most different proteins marked by arrows. Proteins arenumbered according to the reference gel. Asterisks indicate identified proteins specified in Table 1.Equal amounts of 50 �g of protein were loaded for 2-DE separations. B, The abundance of selectedproteins in two independently prepared B10R-FCP and B10S-FCP compartments. The values of rela-tive volumes were obtained using Melanie 3 software. C, The spatial distribution of 2-DE analyses ofB10R-FCP and B10S-FCP compartments in 2-D factor space.

either accumulated in or eliminated from the phagosomein the course of phagosomal trafficking. This view is sup-ported by recent observations of Garin et al. [32] demon-strating that four mitochondrial proteins including ATPsynthase �-chain were identified in latex bead-containingphagosomes and their presence appeared to be repre-sentative of more than a simple contamination. Further-more, proteome analysis of latex bead-containing phago-somes identified several molecules normally associatedwith the endoplasmic reticulum including calreticulin,and provided new insight into phagosome functions.Recruitment of endoplasmic reticulum to phagosomescan keep the particles within close compartment whileplasma membrane recycles back to the cell surface [32].Another differentially expressed protein detected in ourstudy, spot 585, gave acceptable peptide mass map butno positive identification was obtained indicating that it

may be an unknown protein. Unfortunately, due to thesmall quantity of protein, we did not obtain a sequencetag sufficient to search databases for a homologous pro-tein. Although many of the spots selected by correspon-dence analysis are more abundant in the B10R-FCP com-partment, spot 561 was dominant in B10S-FCP (Fig. 3Aand B).

3.2.2 Subtractive analysis of the phagosomalcompartments and identification ofbacterial proteins

Subtractive analysis for evaluation of qualitative differ-ences based on the criteron that the spot was present inB10R-FCP but absent in B10S-FCP gels was done. Asequal protein amounts of the lysates of gradient fractions

90 H. Kovárová et al. Proteomics 2002, 2, 85–93

Table 1. List of identified proteins

Spotno.

2-DEpI/mass(kDa)

Identificationmethod

No. ofmatchedpeptides

Sequencecoverage(%)

Protein name Organism NCBIaccessionno.

189 5.1/61 MALDI-MS 12 32 60 kDa Chaperonin (GroEL) F. tularensis 6225128190 5.0/61 IMMUNO-

BLOTTING– – 60 kDa Chaperonin (GroEL) F. tularensis –

840 5.7/22 MALDI-MS 12 63 Hypothetical 23 kDa protein F. tularensis 1842198189 5.1/61 MALDI-MS 14 41 Putativea) M. musculus 12845667808 5.5/27 MALDI-MS 10 51 Putativeb) M. musculus 12832533810 5.9/26 MALDI-MS 10 54 GFP [19–21]

a) highly homologous to ATP synthase, �-chainb) highly homologous to NADH-ubiquinone oxidoreductase

were loaded on the gels, the enrichment of certain proteinspots became obvious, although the intensities of a largenumber of protein spots were similar (Fig. 4). Due to thedifferential phenotypes that represent both fractions, thebacterial proteins, in addition to macrophage proteins,were expected to be revealed in total lysates of the sus-ceptible B10R-FCP compartment compared to the B10S-FCP fraction. The spots typical for B10R-FCP are high-lighted in Fig. 4.

To date we have identified several of the selected pro-teins (Table 1). Protein spot 840 (pI 5.70, mass 22 kDa)which showed extremely high expression in B10R-FCPwas identified as hypothetical 23 kDa protein of F.tularensis LVS. A slightly more basic form (pI 5.8) ofthis protein was identified by Golovliov [33]. This bac-terial protein was prominently induced inside cells ofthe macrophage-like line J774 and also when exposedextracellularly to hydrogen peroxide. Thus, the authors

Figure 4. Subtractive analysis of gel images. The images of silver stained gels of F. tularensis LVS-containing phagosomes from B10R (B10R-FCP) and “pseudophagosomes” from B10S (B10S-FCP)macrophages were evaluated by Melanie 3 software. The proteins specific for B10R-FCP compart-ment were selected and are indicated by their numbers according to the reference gel, arrowheads.They are comprised into A-D regions of the gel. Asterisks indicate identified proteins specified inTable 1.

Proteomics 2002, 2, 85–93 Bcg/Nramp1modifies F. tularensis LVS-containing phagosome 91

assumed the association between increased expres-sion of this protein and adaptation to the hostile envi-ronment inside the macrophage. Our observation rathersuggested a relation between high protein expressionand bacterial growth under permissive phagosomalenvironment of B10R macrophages. Although weassume that permissive conditions in B10R macro-phages are still likely to be stressful, another possibleexplanation is the presence of several charge-differentprotein forms that can be either differentially expressedin the virulent variant of Francisella compared to theattenuated vaccine strain [34] or may play a differentfunctional role under the influence of various combina-tions of stimuli in vacuolar microenvironment. Further-more, intracellular bacterium can successfully exploitnormal cellular functions to bypass host defense andmultiply inside the cell. We suppose that the transportfunction of Nramp1 allows Francisella to utilise ionsincluding essential nutrient iron and overcome hostimmune response of B10R macrophages. We haveobserved that the manipulation of intracellular ironaffects resistance/susceptibilty to Francisella conferredby Bcg/Nramp1. The addition of desferrioxamine, aniron chelator, to susceptible B10R macrophagesresulted in suppression of bacterial growth. In contrast,iron abrogated the capacity of B10S macrophages tocontrol the growth of F. tularensis LVS [unpublishedresults].

Another identified bacterial protein is 60 kDa chaperoninGroEL (Figs. 3 and 4, Table 1, spots 189 and 190). It wasdifficult to evaluate its contribution to the interaction

between F. tularensis LVS and B10R macrophages dueto its overlap with putative macrophage protein in spot189. The corresponding MALDI spectrum exhibited twocomplementary series of peptide peaks (Fig. 5). The firstset of peaks clearly identified GroEL, the second one wasunambiguously assigned to the putative mouse protein.The presence of GroEL in spot 190 would be expectedbecause this protein is an abundant bacterial protein[34]. Identity of the spot 810 with GFP was confirmed forone of the proteins specifically found in B10R-FCP only.Interestingly, high levels of the hypothetical 23 kDa pro-tein was comparable to GFP expressed by F. tularensisLVS.

4 Concluding remarks

To survive or coexist, bacteria as well as host cellsrespond, upon phagocytosis, to stimuli from the intracel-lular microenvironment by phenotypic modulation. Noneof the stimuli used in vitro can mimic complex intracellularconditions. The composition of bacterial phagosomespresents a delicate balance between the host and thepathogen. Proteome studies of these dynamic intracel-lular organelles offer a promising approach to understandhow this balance is regulated. The results presenteddemonstrate that differential allelic expression of hostBcg/Nramp1 modifies the composition of F. tularensisLVS-containing phagosomes and the resistant allele ofthis gene contributes to the permissiveness for Franci-sella inside B10R macrophages. The acquisition of an

Figure 5. Peptide mass map ofspot 189 containing two pro-teins: 60 kDa chaperonin GroEL(F. tularensis) and putativemouse protein homologous toATP synthase, �-chain. Thepeaks with asterisks correspondto tryptic fragments of GroEL,the peptide signals of the puta-tive protein are labeled with �signs.

92 H. Kovárová et al. Proteomics 2002, 2, 85–93

essential nutrient iron supported by ion transport functionof the Bcg/Nramp1 gene is a possible means thatdeserves further study.

This project was supported by the Grant Agency ofthe Czech Republic (# 310/99/1185), the Ministry ofEducation, Youth and Sports of the Czech Republic(# LN00A033), and The Wellcome Trust Travel Award,UK. We are grateful to A. Sjöstedt at Umea University forproviding F. tularensis LVS-GFP, helpful discussions andcritical reading of the manuscript. Macrophage cell linesB10R and B10S were kindly provided by D. Radzioch atMcGill University Health Centre. We thank E. Skamene atMcGill University Health Centre for constructive com-ments. The technical assistance of R. Krejcová, J. Záková,J. Michalicková, and A. Firychová is greatly appreciated.

Received April 4, 2001

5 References

[1] Desjardins, M., Trends Cell Biol. 1995, 5, 24–27.[2] Harding, C. J., Trends Cell Biol. 1995, 5, 105–109.[3] Frehel, C., de Chastellier, C., Lang, T., Rastogi, N., Infect.Immun. 1986, 52, 252–262.

[4] Joiner, K. A., J. Clin. Invest. 1997, 99, 1814–1817.[5] Mekalanos, J. J., J. Bacteriol. 1992, 174, 1–7.[6] Finlay, B. B., Cossart, P., Science 1997, 276, 718–125[7] Blackwell, J. M., Roach, T. I. A., Atkinson, S. E., Ajioka, J. W.,et al., Immunol. Lett. 1991, 30, 241–248.

[8] Vidal, S., Tremblay, M. L., Govoni, G., Gauthier, S., et al.,J. Exp. Med. 1995, 182, 655–666.

[9] Cellier, M., Prive, G., Belouchi, A., Kwan, T., et al., Proc. Natl.Acad. Sci. USA 1995, 92, 10089–10093.

[10] Baker, S. T., Barton, C. H., Biggs, T. E., J. Leukoc. Biol. 2000,67, 501–507.

[11] Jabado, N., Jankowski, A., Dougaparsad, S., Picard, V., etal., J. Exp. Med. 2000, 192, 1237–1247.

[12] Zwilling, B. S., Kuhn, D. E., Wikoff, L., Brown, D., Lafuse, W.,Infect. Immun. 1999, 67, 1386–1392.

[13] Grunheid, S., Pinner, E., Desjardins, M., Gros, P., J. Exp.Med. 1997, 185, 717–730.

[14] Searle, S., Bright, N. A., Roach, T. I., Atkinson, P. G., et al.,J. Cell Sci. 1998, 111, 2855–2866.

[15] Sjöstedt, A., Tärnvik, A., Sandström, G., FEMS Immunol.Med. Microbiol. 1996, 13, 181–184.

[16] Anthony, L. S. D., Burke, R., Nano, E., Infect. Immun. 1991,59, 3291–3296.

[17] Malo, D., Vogan, K., Vidal, S., Hu, J., et al., Genomics 1994,23, 51–61.

[18] Kovárová, H., Hernychová, L., Hajdúch, M., Sirová, M.,Macela, A., Infect. Immun. 2000, 68, 1480–1484.

[19] Norqvist, A., Kuoppa, K., Sandström, G., FEMS Immunol.Med. Microbiol. 1996, 13, 257–260.

[20] Ericsson, M., Golovliov, I., Sandström, G., Tärnvik, A., Sjö-stedt, A., Infect. Immun. 1997, 65, 1824–1829.

[21] Cormack, B. P., Valdivia, R. H., Falkow, S., Gene 1996, 173,333–338.

[22] Radzioch, D., Hudson, T., Boule, M., Barrera, L., et al., J.Leukoc. Biol. 1991, 50, 263–272.

[23] Chakraborty, P., Sturgill-Koszycki, S., Russell, D. G., Meth-ods Cell Biol. 1994, 45, 261–276.

[24] Landegren, U., J. Immunol. Methods 1984, 67, 379–388.[25] Wessel, D., Flugge, U. I., Anal. Biochem. 1984, 138, 141–

143.[26] Brown, R. E., Jarvis, K. L., Hyland, K. J., Anal. Biochem.

1989, 180, 136–139.[27] Hochstrasser, D. F., Frutiger, S., Paquet, N., Bairoch, A., et

al., Electrophoresis 1992, 13, 992–1001.[28] Hochstrasser, D. F., Merril, C. R., Appl. Theoet. Electrophor.

1988, 1, 35–40.[29] http://129.85.19.192/profound_bin/webProFound.exe[30] Manly, B. F. J., Multivariate Statistical Methods: A Primer,

Chapman and Hall, New York 1986, pp. 59–71.[31] http://www.expasy.ch/cgi-bin/map2/def?LIVER_MOUSE[32] Garin, J., Diez, R., Kieffer, S., Dermine, J.-F., et al., J. Cell

Biol. 2001, 152, 165–180.[33] Golovliov, I., Ericsson, M., Sandström, G., Tärnvik, A., Sjös-

tedt, A., Infect. Immun. 1997, 65, 2183–2189.[34] Hernychová, L., Stulík, J., Halada, P., Macela, A., et al., Pro-

teomics 2001, 1, 508–515.

Proteomics 2002, 2, 85–93 Bcg/Nramp1modifies F. tularensis LVS-containing phagosome 93