ESX-1-mediated translocation to the cytosol controlsvirulence of mycobacteria

Diane Houben,1 Caroline Demangel,2

Jakko van Ingen,3 Jorge Perez,4 Lucy Baldeón,4

Abdallah M. Abdallah,1,4† Laxmee Caleechurn,2

Daria Bottai,2‡ Maaike van Zon,1 Karin de Punder,1

Tridia van der Laan,3 Arie Kant,5

Ruth Bossers-de Vries,5† Peter Willemsen,5†

Wilbert Bitter,4 Dick van Soolingen,3

Roland Brosch,2 Nicole van der Wel1 andPeter J. Peters1,6*1Division of Cell Biology II, Netherlands CancerInstitute–Antoni van Leeuwenhoek Hospital (NKI-AVL),1066 CX Amsterdam, the Netherlands.2Institut Pasteur Pathogénomique MycobactérienneIntégrée, 25 Rue du Docteur Roux, 75724 Paris,France.3National Tuberculosis Reference Laboratory, NationalInstitute of Public Health and the Environment, 3721 MABilthoven, the Netherlands.4Medical Microbiology and Infection Control, VUUniversity Medical Center, 1081 BT Amsterdam, theNetherlands.5Department of Bacteriology and TSE’s, CentralVeterinary Institute, 8203 AA Lelystad, the Netherlands.6Kavli Institute of Nanoscience, Delft University ofTechnology, 2628 CJ Delft, the Netherlands.

Summary

Mycobacterium species, including Mycobacteriumtuberculosis and Mycobacterium leprae, are amongthe most potent human bacterial pathogens. Thediscovery of cytosolic mycobacteria challenged theparadigm that these pathogens exclusively localizewithin the phagosome of host cells. As yet thebiological relevance of mycobacterial translocationto the cytosol remained unclear. In this currentstudy we used electron microscopy techniques toestablish a clear link between translocation and

mycobacterial virulence. Pathogenic, patient-derived mycobacteria species were found to trans-locate to the cytosol, while non-pathogenic speciesdid not. We were further able to link cytosolic trans-location with pathogenicity by introducing theESX-1 (type VII) secretion system into the non-virulent, exclusively phagolysosomal Mycobacte-rium bovis BCG. Furthermore, we show thattranslocation is dependent on the C-terminus of theearly-secreted antigen ESAT-6. The C-terminal trun-cation of ESAT-6 was shown to result in attenuationin mice, again linking translocation to virulence.Together, these data demonstrate the molecularmechanism facilitating translocation of mycobacte-ria. The ability to translocate from the phagolyso-some to the cytosol is with this study proven to bebiologically significant as it determines mycobac-terial virulence.

Introduction

Pathogenic bacteria have evolved elaborate mechanismsto manipulate the host cell for self-replication and propa-gation, while evading the host immune response. Myco-bacteria are no exception, and several pathogenicspecies cause severe diseases in humans, such as tuber-culosis with over 2 million deaths per year, and leprosy, aseriously disabling disease affecting 2–3 million people.Understanding the virulence mechanisms of Mycobacte-ria is essential for vaccine development and treatment ofinfections. The ability of pathogenic Mycobacteria toarrest phagosomal maturation and reside quietly withinphagosomal compartments in host macrophages wasthought to be essential to their virulence. Howeverrecently we reported that Mycobacterium tuberculosis isable to translocate from the phagolysosomal compart-ment into the cytosol of host cells (van der Wel et al.,2007), which completely shifted the paradigm of theknown virulence mechanisms of M. tuberculosis. Translo-cation has since been observed by others (Lee et al.,2008; Hagedorn et al., 2009), and has also previouslybeen observed in another pathogenic mycobacteriumspecies (Stamm et al., 2003; Gao et al., 2004). In light ofthis recent evidence, the argument that mycobacteria are

Received 24 December, 2011; revised 12 March, 2012; accepted 28March, 2012. *For correspondence. E-mail p.peters@nki.nl; Tel. (+31)20 512 2031; Fax (+31) 20 512 2029.Present addresses: †King Abdullah University of Science and Tech-nology Thuwal 23955-6900, Saudi Arabia; ‡Dipartimento di PatologiaSperimentale, Biotecnologie Mediche, Infettivologia ed Epidemiolo-gia, University of Pisa, 56126 Pisa, Italy.

Cellular Microbiology (2012) 14(8), 1287–1298 doi:10.1111/j.1462-5822.2012.01799.xFirst published online 8 May 2012

© 2012 Blackwell Publishing Ltd

cellular microbiology

restricted to the phagosome still remains, as can be seenfrom recent reviews (Barry et al., 2011; Blanchard andShastri, 2010; Ehrt and Schnappinger, 2009; Meena andRajni, 2010; Russell, 2011) and cell biological studies (deChastellier et al., 2009; Yuk et al., 2009; Brodin et al.,2010; Cardoso et al., 2010; Estorninho et al., 2010; Leeet al., 2010; Abramovitch et al., 2011; Seto et al., 2011).However, a number of these studies utilize non-pathogenic Mycobacterium avium (de Chastellier et al.,2009) or Mycobacterium bovis Bacille Calmette–Guérin(BCG) species (Cardoso et al., 2010; Lee et al., 2010) asa model for tuberculosis infection. Recent genomic analy-ses of several mycobacterial species have identifiedregions within the genome that are missing in non-pathogenic mycobacterial strains. For example, at least16 genomic regions of difference (RDs) present in thegenome of M. tuberculosis, are absent from M. bovisBCG (Behr et al., 1999; Gordon et al., 1999). One suchregion, RD1 (for region of differentiation 1) has beenidentified as a key region involved in M. tuberculosis andM. marinum virulence. It encodes a novel ESAT-6 secre-tion complex-1 (ESX-1 or type VII secretion system)(Abdallah et al., 2007) which enables secretion of myco-bacterial proteins early-secreted antigen 6 kDa (ESAT-6)and culture filtrate protein 10 kDa (CFP10) that are sug-gested to be involved in host membrane lysis and cell-to-cell spread (Guinn et al., 2004; de Jonge et al., 2007).Rendering the components of ESX-1 inactive abrogatesM. tuberculosis and M. marinum translocation to thecytosol (van der Wel et al., 2007). Could it be then that thedependence on these deleted regions (specificallyESX-1) explains the lack of cytosolic M. avium andM. bovis BCG in these studies (de Chastellier et al., 2009;Cardoso et al., 2010; Lee et al., 2010)? The question wethen ask, is there a link between cytosolic translocationand the ability of mycobacterial species to be pathogenic,and is translocation indeed a product of the secretedproteins encoded by ESX-1?

In this present study, we investigate whether there is alink between the presence of a functional ESX-1 secretionsystem and the ability of a wide range of Mycobacteriumspecies (known to be either pathogenic or non-pathogenic) to translocate into the cytosol. We demon-strate that virulence strongly correlates with the ability ofmycobacteria to translocate to the cytosol in humanphagocytic cells. Furthermore, we show that all of thepathogenic species possess an ESX-1 system thatsecretes ESAT-6. To test whether a functional ESX-1secretion system is necessary and sufficient for translo-cation, we reintroduced the extended RD1 gene cluster ofM. tuberculosis into M. bovis BCG (Pym et al., 2003) andshow that BCG with RD1, and thus a functional ESX-1secretion system, translocate to the cytosol and exhibit amore pathogenic phenotype. Finally we show that cytoso-

lic translocation is dependent on the last 12 amino acidsof ESAT-6 via an as yet undiscovered mechanism.Together, our data show that cytosolic translocation is anESAT-6-dependent process relevant only to pathogenicmycobacterial species. These findings provide importantinsights into mycobacterial virulence and demonstratethat translocation is key to mycobacterial pathogenesis.

Results and discussion

ESAT-6 secretion of selected mycobacteria

The ability of M. marinum, M. leprae and M. tuberculosisH37Rv to translocate from the phagolysosome to thecytosol has been previously described (Stamm et al.,2003; van der Wel et al., 2007) and is dependent on theESX-1-encoded secretion system. Here we hypothesizedthat virulence is correlated with ESX-1-mediated ESAT-6secretion. For our studies we selected several specieswith differing degrees of virulence (Table 1). Theseincluded five M. tuberculosis strains; the virulent labora-tory strain (H37Rv) and four patient isolates (two Beijinggenotypes and Dutch isolates, Harlingen and 1243),M. bovis, M. marinum (E11 and M), M. leprae, M. szulgai,M. avium, M. fortuitum, M. gilvum, M. kansasii type I andV, M. smegmatis and M. bovis BCG (Danish) (Table S1).To first test whether the presence of ESAT-6 is specific forpathogenic mycobacterial species we probed culture fil-trates of M. marinum, M. szulgai, M. kansasii type I and V,which encode ESX-1 and ESAT-6, and non-pathogenic(lacking ESX-1) M. fortuitum for ESAT-6 secretion(Fig. S1). All tested ESX-1- containing mycobacteriasecrete ESAT-6, although the amounts vary. The datacombined with data from literature (Table 1) reveal that allpathogenic species secrete ESAT-6. However, the pres-ence of a functional ESX-1 secretion system does notcorrelate with virulence.

Cytosolic translocation is a feature of pathogenicmycobacterial strains

To test whether mycobacterial cytosolic translocation cor-relates with a pathogenic phenotype, we examined thesubcellular localizations of mycobacteria (Table 1) usingelectron microscopy. Mycobacteria were cultured toOD600 0.5 (mid log phase) and used for infection of PMA-activated THP-1 human myeloid cell line as previouslydescribed (van der Wel et al., 2007). After 1–10 days(time was dependent on how long the cells survivedinfection with each mycobacterial species) the cells wereprocessed for cryo-immunogold electron microscopy. Inorder to obtain unbiased data the samples were con-cealed. For the quantification of the rate of translocation,only viable (adherent) cells were analysed, since the

1288 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

morphology of the non-adherent cells in these infectionexperiments is often affected (Fig. S2A and B). Eventhough immunogold labelling of lysosomal markers innon-viable, M. tuberculosis-infected cells is different thanin non-viable cells infected with M. bovis BCG (Fig. S2Cand D), the determination of the subcellular localizationwas carried out in viable cells only. We have previouslyshown that translocation precedes cell death (van derWel et al., 2007) and therefore the viable cell fraction ismost suitable for translocation studies. Images wereobtained and analysed in a double blind manner for sub-cellular localization for all species (Figs 1, S3 and S4)using two criteria: (i) the presence or absence of a pha-gosomal membrane and (ii) the presence or absence ofendosomal or lysosomal immunogold labelling (LAMP orCD63). Only when both the membrane and gold labellingwere absent, a bacterium was scored as cytosolic. As aninternal control, bacteria with phagosomal immunogoldlabelling and without a phagosomal membrane neverexceeded 2% (data not shown). The day on which thehighest percentage cytosolic bacteria were detected isrepresented in Fig. 2.

Pathogenic mycobacteria such as M. tuberculosis,M. leprae, M. bovis and M. marinum were found readily inthe cytosol. Most importantly, we identified that patientderived M. tuberculosis readily translocated into thecytosol (Fig. 2). In contrast, the opportunistic or non-pathogenic species M. avium, M. fortuitum, M. gilvum,M. kansasii type V, M. smegmatis and M. bovis BCG(Danish) remained phagolysosomal, while opportunisticM. szulgai and M. kansasii type I showed reduced trans-location (Figs 2 and S3). From these observations we

can conclude that cytosolic translocation is specific forpathogenic mycobacteria while it is reduced in two oppor-tunistic species tested. Other opportunistic and all non-pathogenic species remained phagosomal. Therefore wecan conclude that cytosolic localization is a feature ofpathogenic mycobacterial species. We would like to notethat part of the species which remained phagosomal inour studies, M. avium and M. smegmatis (opportunisticand non-pathogenic respectively), as well as M. bovisBCG are widely used by researchers to mimic M. tuber-culosis in laboratory settings.

Introduction of ESAT-6 secretion into M. bovis BCGenables cytosolic translocation

Previously we found that mutations that block CFP-10and ESAT-6 secretion abrogated translocation in M.tuberculosis (van der Wel et al., 2007). Cfp-10 andesat-6 genes belong to the RD1 region, which alsoencodes part of the ESX-1 secretion system. In thisstudy we showed that ESAT-6 is secreted by all patho-genic mycobacterial species (Table 1, Fig. S1), andthat pathogenic species translocate to the cytosol (Figs 1and 2); however, what still remains unknown is if thepresence of ESAT-6 is directly linked to cytosolic trans-location. To achieve this, we used a strain of M. bovisBCG which has a recombinant RD1 region reintroducedinto the bacterium (RD1-2F9 construct containinga 32 kb segment encoding Rv3861-Rv3885, hereincalled BCG::ESX-1, Fig. 3A) and is able to secreteESAT-6 (Brodin et al., 2005), to determine if reintroduc-tion of ESAT-6 secretion could induce translocation

Table 1. Characteristics of various mycobacterial species and their pathogenicity.

Mycobacterium strains Pathogenicity in humans ESX-1 genes ESAT-6 secretion Transloc in cells

M. tuberculosis H37Rv Pathogenic + +† +M. tuberculosis ancient Beijing Pathogenic + +* +M. tuberculosis Beijing Pathogenic + +* +M. tuberculosis Harlingen Pathogenic + +* +M. tuberculosis 1243 Pathogenic + +* +M. leprae Pathogenic + +** +M. bovis Pathogenic + + +M. marinum Pathogenic + +‡ +M. szulgai Opportunistic + +† +/-M. kansasii type I Opportunistic + + +/-M. kansasii type V Non-pathogenic + + -M. smegmatis Non-pathogenic + +*** -M. avium Opportunistic - -† -M. fortuitum Opportunistic - -† -M. bovis BCG Non-pathogenic - - -M. gilvum Non-pathogenic - -* -

The level of pathogenicity in humans for each strain is as described previously (Behr, 2008; Parwati et al., 2010). The presence of the ESX-1 genesis represented as ‘+’ and an incomplete complex or absence as ‘-’. Secretion of ESAT-6 was monitored by Western blot analysis comparing pelletand supernatant of in vitro cultures (Fig. S1). References indicate data shown by others. *Results not shown; **T cells of patients with leprosy reactstrongly to M. leprae ESAT-6, suggesting that M. leprae has an active ESX-1 secretion system (Geluk et al., 2002); ***M. smegmatis waspreviously shown to secrete ESAT-6 in some conditions (Converse and Cox, 2005), †(Sorensen et al., 1995), ‡(Abdallah et al., 2009). The abilityto translocate (Transloc in cells) was monitored in THP-1 cells and is represented as + when the percentage of bacteria in the cytosol was higherthan 20%, as +/- when the percentage was between 2% and 20% and as ‘-’ when the percentage was lower than 2% (see Figs 2 and S3).

Translocation is a virulence factor mycobacteria 1289

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

1290 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

in an otherwise phagolysosomal mycobacterium. PMA-activated THP-1 cells were infected with BCG::ESX-1 asdescribed above and subsequently fixed and processedfor cryo-immunogold electron microscopy to determine if,and at what rate, bacterial cytosolic translocation occurs.After 7 days of infection, we discovered in an unbiasedexperimental set-up that 16% of BCG::ESX-1 bacteriawere translocated to the cytosol and after 10 days thispercentage had increased to 46% (Fig. 3). BCG::ESX-1strain showed significantly increased virulence comparedwith the non-translocating BCG (Pym et al., 2002; Brodinet al., 2005). With these results we show that reintroduc-tion of ESAT-6 secretion apparatus encoded by the RD1region into normally phagosomal BCG provokes BCGcytosolic translocation and in turn increases the virulenceof the bacterium.

C-terminus ESAT-6 is required for M. tuberculosiscytosolic translocation

ESAT-6 and CFP-10 are known to be secreted into Myco-bacterium culture filtrate medium mediated by ESX-1-encoded secretion system and we have shown here thatESAT-6 is linked to cytosolic translocation. M. tuberculo-sis H37Rv EsxAD84-95 is a strain containing a deletionof the 12-amino-acid ESAT-6 C-terminus (Brodin et al.,2005). While this strain was shown to still secrete ESAT-6and CFP-10 efficiently, it was attenuated in a mouse infec-

tion model (Brodin et al., 2005). To determine whether thisattenuation is linked to its (in)ability to translocate, weinfected THP-1 cells as above with H37Rv EsxAD84-95strain and analysed electron microscopy images with thecriteria mentioned above in a double blind study. Asshown in Fig. 4, the H37Rv EsxAD84-95 mutant remainedin phagolysosomes, demonstrating that the C-terminusof the ESAT-6 protein is required for M. tuberculosistranslocation.

Both phagosomal-localized M. smegmatis and M.kansasii type V species have been found to secreteESAT-6 (Converse and Cox, 2005 and Fig. S1) so wehypothesized that the inability of these bacteria to trans-locate may be due to the lack of a specific C-terminusfound in pathogenic species. Performing sequenceanalysis of ESAT-6 amino acid sequences from severalspecies we fail to find a correlation between sequenceand ability to translocate (Fig. S5). It seems unlikely thatthe C-terminus of the ESAT-6 protein alone can explainthe inability to translocate and therefore must involveother unknown factors. We have yet to understand themechanism of ESAT-6 in cytosolic translocation and itremains an exciting area for future research. Still theresults of the C-terminal truncated H37Rv mutant dem-onstrate that this part of the M. tuberculosis ESAT-6protein is essential for its function in translocationand again demonstrates a link between virulence andtranslocation.

Fig. 1. Subcellular localization of different mycobacterial species in THP-1 cells immunogold labelled for CD63. THP-1 monocytic cell linewas infected with different pathogenic mycobacteria: M. tuberculosis 1243, M. bovis and opportunistic or non-pathogenic species M. szulgai,M. fortuitum, M. avium, M. kansasii type I and V and M. smegmatis, fixed and processed for immunogold labelling with CD63. An electronmicrograph example was taken at day 2 for M. tuberculosis 1243, day 7 M. bovis, day 3 M. szulgai, day 1 M. fortuitum, day 3 M. avium, day2 M. kansasii type I, day 3 M. kansasii type V and day 2 M. smegmatis. Asterisk, cytosolic bacterium; Encircled asterisk, phagosomalbacterium; M, mitochondrium; L, lysosome. Bar represents 400 nm.

Fig. 2. Translocation is specific for pathogenic mycobacteria. Percentage cytosolic bacteria in adherent, non-apoptotic THP-1 cells infectedwith different mycobacterial species (see Table 1 and Fig. S3). Infections were performed with moi 10 and kept as long as cells survived, oruntil day 10 in M. bovis BCG infections. The day (D) of infection on which the percentage of cytosolic bacteria was highest is indicated undereach bar. The pathogenic species are represented as red, opportunistic species as yellow and non-pathogenic as green.

Translocation is a virulence factor mycobacteria 1291

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

Cytosolic translocation results in transientubiquitin accumulation

We hypothesized that the phagosome must undergoultrastructural changes during translocation. Therefore,we attempted to search for intermediate stages of trans-location in cells infected with M. tuberculosis by probingfor ubiquitin, which tags proteins destined for proteaso-mal degradation. We observed that the area aroundcytosolic mycobacteria could be labelled with antibodiesagainst ubiquitin (Figs 5 and S6) on cryosections, whichwas prevented in the presence of excess purified ubiq-uitin, demonstrating specificity of the antibody. In con-trast, intact phagolysosomes containing M. tuberculosis,but also M. avium or M. bovis BCG, were less labelledfor ubiquitin (Fig. 5A). The percentage of cytosolic bac-teria immunogold labelled for ubiquitin increased to 30%of all bacteria at 96 h post infection. At this time of infec-tion, no longer all cytosolic bacteria were ubiquitinlabelled, demonstrating that the presence of ubiquitinsurrounding cytosolic bacteria is transient. We observedthat, although intermediate stages of translocation arevery rare, remainders of the phagosome can be tran-siently detected (Fig. S6B and D) and immunolabelledfor ubiquitin. This would suggest that the ubiquitinlabelled membranous structures could be degradingphagolysosomes. These structures need to be furtheranalysed and specifically the dynamics and role ofautophagic processes could be further investigated. Forour analysis of the translocation characteristics of myco-bacteria, ubiquitin immunogold labelling experimentsfurther confirm cytosolic localization.

Why has translocation gone so often unnoticed?

Our present data demonstrate the importance of translo-cation for the virulence of mycobacteria and reveal a partof the molecular mechanism underlying this subcellulartrafficking event. Our earlier report on the existence of

Fig. 3. RD1 knock-in is sufficient for translocation of BCG.A. Schematic representation of the RD1 deletion region naturallyoccurring in M. bovis BCG and the extended RD1 region naturallyoccurring in M. tuberculosis and reintroduced into M. bovis BCG(BCG::ESX-1) using the RD1-2F9 construct containing a 32 kbsegment encoding Rv3861-Rv3885.B. Percentage cytosolic bacteria at days 4, 7 and 10 post infectionin stimulated THP-1 cells. In the absence of RD1 (BCG::pYUBvector control) no bacteria were found in the cytosol, but in BCGwith ESX-1 knocked-in (BCG::ESX-1), translocation was detectedat 7 and 10 days.C. Representative electron micrograph of cytosolic BCG::ESX-1at day 10 post infection in a stimulated THP-1 cell. Immunogold(10 nm) labelling of LAMP-2 marks lysosomal structures (L). Barrepresents 300 nm; M, mitochondria; PM, plasma membrane.D. Representative electron micrograph of phagosomal M. bovisBCG at day 1 post infection, immunolabelled (10 nm gold) forCD63. Bar represents 200 nm; M, mitochondrium; L, lysosome.

1292 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

cytosolic mycobacteria was initially received with a certainscepticism, and since then many studies refer to thedogma of phagosomal localization of mycobacteria (Guti-errez et al., 2008; Jordao et al., 2008; Peyron et al., 2008;de Chastellier et al., 2009; Anand et al., 2010; Bonillaet al., 2010; Brodin et al., 2010; Caleffi et al., 2010;Cardoso et al., 2010; Estorninho et al., 2010; Ferrer et al.,2010; Sweet et al., 2010; Abramovitch et al., 2011; Mattoset al., 2011; Seto et al., 2011). First, it should be noted thatsome of the studies describe the early events of infection(Yuk et al., 2009; Bonilla et al., 2010; Brodin et al., 2010;Estorninho et al., 2010; Sweet et al., 2010; Seto et al.,2011). The early interactions of mycobacteria with thephagosome are without doubt crucial for the outcome ofthe infection, but will not allow for the detection of cytoso-lic bacteria, as translocation generally occurs later duringinfection, at time points that vary for the different myco-bacterial species. For M. tuberculosis this is generally 3–5days after infection (Fig. S3) (van der Wel et al., 2007;Lee et al., 2008). Second, immunofluorescence micros-copy will not reveal subpopulations of cytosolic bacteria.The absence of a (lysosomal) marker (Ferrer et al., 2010;Abramovitch et al., 2011) may not be recognized as evi-dence for the absence of the phagosomal membrane.Finally, as demonstrated here, the subcellular localizationof mycobacteria and their ability to translocate dependsupon the species, as M. avium and M. smegmatis forexample remain phagosomal (Gutierrez et al., 2008; deChastellier et al., 2009; Anand et al., 2010; Sweet et al.,2010). Results on the localization of these species shouldtherefore not be extrapolated to pathogenic species likeM. tuberculosis.

Other research groups in independent studies havenow confirmed the cytosolic localization of M. tuberculo-sis in human cells (Lee et al., 2008; Simeone et al.,2012), the amoeba Dictyostelium (Hagedorn et al., 2009)and of M. marinum in various in vitro systems and in vivoin mouse tail tissues (Lee et al., 2008; Hagedorn et al.,2009; Pandey et al., 2009; Carlsson et al., 2010). Inaddition, recently ESAT-6-dependent damaging effectson the phagosomal membranes were reported duringM. tuberculosis infections (Wong and Jacobs, 2011).Further, indirect evidence for the cytosolic localization ofM. tuberculosis comes from studies describing the myco-bacterial stimulation of the cytosolic Nod1, 2 pathway,which is known to respond to microbial products in thecytosol (Lee et al., 2008; Hagedorn et al., 2009; Pandeyet al., 2009; Carlsson et al., 2010). The translocationmodel is also supported by numerous reports ofmycobacterial-specific CD8 T-cell responses (Garceset al., 2010; Weerdenburg et al., 2010).

The first reports of mycobacterial cytosolic localizationare from 1984 (Leake et al., 1984; Myrvik et al., 1984);however, we are only now beginning to understand whythis subcellular event is so often overlooked. Transloca-tion, also often termed ‘escape’ is, however, well estab-lished as a virulence mechanism in several bacteria fromother genera, including Shigella flexneri, Listeria mono-cytogenes, Burkholderia pseudomallei, Neisseria menin-gitidis, Rickettsia prowazekii and Francisella tularensis(Ray et al., 2009). For many of these species, the bacte-rial factors that determine and facilitate translocationare already known. For L. monocytogenes, a faculta-tive intracellular pathogen that crosses the intestinal,

Fig. 4. C-terminus ESAT-6 crucial for translocation.A. Percentage cytosolic H37Rv (moi 2), H37Rv EsxAD84-95 (moi 10) or H37RvDRD1 (moi 10) at day 2, 4 or 6 post infection in stimulatedTHP-1 cells. The translocating M. tuberculosis strain, H37Rv, and non-translocating RD1 deletion mutant of M. tuberculosis H37Rv(H37RvDRD1) acted as positive and negative controls.B. Representative electron micrograph of phagolysosomal ESAT-6 mutant H37Rv EsxAD84-95 at day 2 post infection in THP-1 cells.Immunogold (10 nm) labelling of LAMP-2 marks the lysosomal structure (L). Bar represents 300 nm; M, mitochondria.

Translocation is a virulence factor mycobacteria 1293

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

Fig. 5. Ubiquitin on perturbed phagolysosomes.A. Representative electron micrograph of M. tuberculosis H37Rv at 3 days of infection in DCs, immunolabelled for ubiquitin with 5 nm goldcoupled to protein A and subsequently LAMP-1 with 10 nm gold. Asterisk represents cytosolic bacterium; Encircled asterisk, phagosomalbacterium; M, mitochondrium; L, lysosome; blue circles, ubiquitin label surrounding cytosolic M. tuberculosis; bar represents 100 nm.B. Quantification of ubiquitin labelling associated with clusters of M. tuberculosis. M. tuberculosis clusters were regarded as positive forubiquitin when 7 or more gold particles per bacterium were detected in the area extending out to 300 nm from the outer edges of the bacteria.Percentages of phagosomal bacteria ubiquitin labelled (+), cytosolic ubiquitin labelled (+) and cytosolic unlabelled (-) are plotted. Forsimplicity, the unlabelled phagosomal bacteria are left out of the graph.C. Primary human DCs infected with H37Rv for 3 days, immunogold labelled for ubiquitin using a ubiquitin antibody and protein A conjugatedto 5 nm gold and subsequently LAMP-1 with 10 nm gold. Labelling was present on small membranous structures. Asterisk represents cytosolicbacterium; M, mitochondrium; blue circles, ubiquitin label; bar represents 100 nm.

1294 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

materno-fetal and blood-brain barriers, lysis of the pha-gosomal membrane is mediated by the pore formingprotein listeriolysin O, a cholesterol-dependent cytolysinthat acts collaboratively with bacterial phospholipases todisrupt the membrane of the phagolysosome (Marquiset al., 1995; Smith et al., 1995; Schnupf and Portnoy,2007). M. tuberculosis encodes up to four phospholipaseC genes that are known to be important for growth ofmycobacteria in vivo (Raynaud et al., 2002). The role ofphospholipase C in translocation of mycobacteria ishowever still unknown.

Concluding remarks

In this study, we have established that mycobacteriarequire the ESX-1 secretion system and secreted ESAT-6for cytosolic translocation. We have now shown thatESAT-6 is secreted by all translocating species tested,and these therefore have a functional ESX-1 secretionsystem. We also show that specifically, the C-terminus ofESAT-6 is important for translocation, which explains theattenuation of the ESAT-6 C-terminal mutant in mice.Thus using several approaches, we have established aclear link between the ability of mycobacterial species totranslocate and their pathogenicity. This was verified bythe reintroduction of the M. tuberculosis ESX-1 genecomplex into attenuated, phagolysosomal M. bovis BCGand showed that this re-establishes translocation. Impor-tantly, this recombinant BCG::ESX-1 strain showed sig-nificantly increased virulence compared with the non-translocating BCG (Pym et al., 2002; Brodin et al., 2005).Collectively, the increased virulence of the BCG::ESX-1and the ability of all pathogenic species tested to trans-locate presents a crucial relationship. Even though trans-location was shown to be essential for virulence inmultiple genera, a crucial correlation within a subset ofspecies from one genus has not been described previ-ously. Thus here we show for the first time that translo-cation of mycobacteria is not only clinically relevant, it isa virulence factor. Clearly virulence factors are importantfor designing vaccines and thus we propose that novelmycobacterial vaccines should be functional in thecytosol. The vaccine currently used worldwide (M. bovisBCG), however, is restricted to the phagosome. Theresults described here indicate that translocation into thecytosol mediated by the ESX-1 system will improvevaccine efficacy, but also warns of the risk of developinga virulent vaccine (Pym et al., 2002; Brodin et al., 2005).Possibly by manipulating translocation and virulence, theefficacy of the BCG vaccine may increase and could beeffective for various mycobacterial infections. By estab-lishing that translocation is a virulence factor for myco-bacteria, the development of an improved vaccinebecomes feasible.

Experimental procedures

Cell culture

THP-1 cells were cultured in RPMI-1640 medium (Invitrogen)with 10% FCS and 100 units ml-1 penicillin/streptomycin at 37degrees and 5% CO2. Differentiation of the monocytic cell lineinto macrophages was stimulated by adding 10 ng ml-1 phorbol12-myristate 13-acetate (PMA) 1 day before infection. The nextday, cells were given fresh medium without PMA and pen/strep.Primary human DCs were infected with M. tuberculosis H37Rvand processed as described in van der Wel et al. (2007).

Bacterial cultures

Several pathogenic (Kiers et al., 1997; Tsolaki et al., 2005;Parwati et al., 2010), opportunistic (van Ingen et al., 2009a,b; vanIngen et al., 2008) and non-pathogenic mycobacterial specieswere isolated or obtained via several laboratories (Table S1). Allmycobacterial strains were grown to log-phase in 7H9 mediumsupplemented with 10% ADC enrichment and 0.05% Tween80.The bacteria were washed with RPMI-1640 medium, centrifugedtwice at 750 r.p.m. for 7 min to spin down aggregates andadded to the THP-1 cells at a multiplicity of infection (moi) of 10.Mutant strains BCG::pYUB, BCG::ESX-1 (Pym et al., 2002),H37RvDRD1 (gift from William R Jacobs Jr) and H37RvEsxAD84-95 (Brodin et al., 2005) were cultured in 7H9 with ADCenrichment, 0.05% Tween80 and hygromycin. These mutantswere sonicated, dispersed by a syringe and added to the cells atan moi of 10, except for M. tuberculosis H37Rv, for which moi 2was used, to slow the experiment to last until day 6. Cells wereincubated with bacteria for 1 h at 37°C (M. marinum andM. gilvum at 32°C), and then cells were washed 3¥ with RPMI-1640 medium to remove extracellular bacteria. Infection wasstopped at designated time points by adding paraformaldehydeor paraformaldehyde/glutaraldehyde (Peters et al., 2006). Cul-tures were continued until cells died or until day 14 for BCGinfections.

Electron microscopy

All samples were prepared for electron microscopy as describedbefore (Peters et al., 2006; van der Wel et al., 2007). For immu-nogold labelling, monoclonal LAMP-1 (Pharmingen H4A3) 1:150,monoclonal LAMP-2 (H4B4), monoclonal CD63 (SanquinM1544) 1:15, polyclonal ubiquitin (Sigma U5379) 1:300 or asequential labelling with first anti-ubiquitin and then anti-LAMP-1was used. To determine the specificity of the polyclonal anti-ubiquitin, pure ubiquitin (gift from F. El Oualid) was added indifferent concentrations. For monoclonal antibodies, a bridgingrabbit anti-mouse (DAKO) 1:200 was then added. For visualiza-tion of antibody binding, 10 nm protein A gold was used and forthe double labelling first 5 nm protein A gold and subsequently10 nm protein A gold were used. Samples were analysed bydouble-blind quantification using a CM10 electron microscope(FEI, Eindhoven, the Netherlands).

SDS-PAGE and immunoblot

Bacteria were washed and cultured in 7H9 medium supple-mented with 0.2% (w/v) dextrose and 0.05% Tween80. Pellets

Translocation is a virulence factor mycobacteria 1295

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

and supernatants were then separated from log-phase cultures.Proteins in the cell-free supernatants were precipitated with5% TCA (w/v), and the pellets were sonicated. Proteinswere separated by SDS-PAGE on 16% polyacrylamide gels.Immunoblots were stained with monoclonal ESAT-6 antibody(Staten Serum Institute, Hyb76-8) 1:500 and the presence ofperoxidase-conjugated secondary antibody was detected via4-chloronaphthol/3,3-diaminobenzidine staining.

Acknowledgements

We are grateful to William R Jacobs Jr (AECOM, New York),Astrid van der Sar, Ben J Appelmelk and Janneke Maaskant(VUmc, Amsterdam) for bacterial strains, Edith Houben (VUmc,Amsterdam) for help with Western Blot analysis, Nico Ong (NKI-AVL, Amsterdam) for photography and Sue Godsave, Alicia Lam-merts van Bueren (NKI-AVL, Amsterdam) and Michael Brenner(Harvard Medical School, Boston) for critical reading of the manu-script. We would like to thank David Russell (Cornell University)for critically evaluating the data and the manuscript, Farid ElOualid (NKI-AVL, Amsterdam) for providing purified ubiquitin andBen J Appelmelk (VUmc, Amsterdam), Jacques Neefjes (NKI-AVL, Amsterdam) and the members of the Peters’ lab for discus-sions. Work was supported by the Netherlands Leprosy ReliefFoundation (NLR), and by Aeras Global TB Vaccine Foundationvia a grant from the Netherlands Directorate-General of Devel-opment Cooperation, Dutch Ministry of Foreign Affairs. C.D., D.B.and R.B. were supported by the European Community’s SeventhFramework Programme ([FP7/2007–2013]) under grant agree-ment n°201762FP7. D.H. received travel grants from KNCVTuberculosis Foundation, Foundation ‘De Drie Lichten’ in TheNetherlands and the Ter Meulen Fund, Royal NetherlandsAcademy of Arts and Sciences, the Netherlands.

References

Abdallah, A.M., Gey van Pittius, N.C., Champion, P.A., Cox,J., Luirink, J., Vandenbroucke-Grauls, C.M., et al. (2007)Type VII secretion – mycobacteria show the way. Nat Rev5: 883–891.

Abdallah, A.M., Verboom, T., Weerdenburg, E.M., Gey vanPittius, N.C., Mahasha, P.W., Jimenez, C., et al. (2009)PPE and PE_PGRS proteins of Mycobacterium marinumare transported via the type VII secretion system ESX-5.Mol Microbiol 73: 329–340.

Abramovitch, R.B., Rohde, K.H., Hsu, F.F., and Russell, D.G.(2011) aprABC: a Mycobacterium tuberculosis complex-specific locus that modulates pH-driven adaptation tothe macrophage phagosome. Mol Microbiol 80: 678–694.

Anand, P.K., Anand, E., Bleck, C.K., Anes, E., and Griffiths,G. (2010) Exosomal Hsp70 induces a pro-inflammatoryresponse to foreign particles including mycobacteria. PLoSONE 5: e10136.

Barry, A.O., Mege, J.L., and Ghigo, E. (2011) Hijacked pha-gosomes and leukocyte activation: an intimate relationship.J Leukoc Biol 89: 373–382.

Behr, M.A. (2008) Mycobacterium du jour: what’s on tomor-row’s menu? Microbes Infect 10: 968–972.

Behr, M.A., Wilson, M.A., Gill, W.P., Salamon, H., Schoolnik,G.K., Rane, S., and Small, P.M. (1999) Comparative

genomics of BCG vaccines by whole-genome DNAmicroarray. Science 284: 1520–1523.

Blanchard, N., and Shastri, N. (2010) Topological journey ofparasite-derived antigens for presentation by MHC class Imolecules. Trends Immunol 31: 414–421.

Bonilla, D.L., Ly, L.H., Fan, Y.Y., Chapkin, R.S., andMcMurray, D.N. (2010) Incorporation of a dietary omega 3fatty acid impairs murine macrophage responses to Myco-bacterium tuberculosis. PLoS ONE 5: e10878.

Brodin, P., de Jonge, M.I., Majlessi, L., Leclerc, C., Nilges, M.,Cole, S.T., and Brosch, R. (2005) Functional analysis ofearly secreted antigenic target-6, the dominant T-cellantigen of Mycobacterium tuberculosis, reveals key resi-dues involved in secretion, complex formation, virulence,and immunogenicity. J Biol Chem 280: 33953–33959.

Brodin, P., Poquet, Y., Levillain, F., Peguillet, I., Larrouy-Maumus, G., Gilleron, M., et al. (2010) High contentphenotypic cell-based visual screen identifies Mycobacte-rium tuberculosis acyltrehalose-containing glycolipidsinvolved in phagosome remodeling. PLoS Pathog 6:e1001100.

Caleffi, K.R., Hirata, R.D., Hirata, M.H., Caleffi, E.R., Peixoto,P.R., Sposito, F.L., and Cardoso, R.F. (2010) Evaluation of85 A-C intergenic region PCR primers for detection ofMycobacterium leprae DNA in urine samples. Int J Derma-tol 49: 717–718.

Cardoso, C.M., Jordao, L., and Vieira, O.V. (2010) Rab10regulates phagosome maturation and its overexpressionrescues Mycobacterium-containing phagosomes matura-tion. Traffic 11: 221–235.

Carlsson, F., Kim, J., Dumitru, C., Barck, K.H., Carano, R.A.,Sun, M., et al. (2010) Host-detrimental role of Esx-1-mediated inflammasome activation in mycobacterial infec-tion. PLoS Pathog 6: e1000895.

de Chastellier, C., Forquet, F., Gordon, A., and Thilo, L.(2009) Mycobacterium requires an all-around closelyapposing phagosome membrane to maintain the matura-tion block and this apposition is re-established when itrescues itself from phagolysosomes. Cell Microbiol 11:1190–1207.

Converse, S.E., and Cox, J.S. (2005) A protein secretionpathway critical for Mycobacterium tuberculosis virulenceis conserved and functional in Mycobacterium smegmatis.J Bacteriol 187: 1238–1245.

Ehrt, S., and Schnappinger, D. (2009) Mycobacterial survivalstrategies in the phagosome: defence against hoststresses. Cell Microbiol 11: 1170–1178.

Estorninho, M., Smith, H., Thole, J., Harders-Westerveen, J.,Kierzek, A., Butler, R.E., et al. (2010) ClgR regulation ofchaperone and protease systems is essential for Mycobac-terium tuberculosis parasitism of the macrophage. Micro-biology 156: 3445–3455.

Ferrer, N.L., Gomez, A.B., Neyrolles, O., Gicquel, B., andMartin, C. (2010) Interactions of attenuated Mycobacteriumtuberculosis phoP mutant with human macrophages. PLoSONE 5: e12978.

Gao, L.Y., Guo, S., McLaughlin, B., Morisaki, H., Engel, J.N.,and Brown, E.J. (2004) A mycobacterial virulence genecluster extending RD1 is required for cytolysis, bacterialspreading and ESAT-6 secretion. Mol Microbiol 53: 1677–1693.

1296 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

Garces, A., Atmakuri, K., Chase, M.R., Woodworth, J.S.,Krastins, B., Rothchild, A.C., et al. (2010) EspA acts as acritical mediator of ESX1-dependent virulence in Mycobac-terium tuberculosis by affecting bacterial cell wall integrity.PLoS Pathog 6: e1000957.

Geluk, A., van Meijgaarden, K.E., Franken, K.L., Subronto,Y.W., Wieles, B., Arend, S.M., et al. (2002) Identificationand characterization of the ESAT-6 homologue of Myco-bacterium leprae and T-cell cross-reactivity with Mycobac-terium tuberculosis. Infect Immun 70: 2544–2548.

Gordon, S.V., Brosch, R., Billault, A., Garnier, T., Eiglmeier,K., and Cole, S.T. (1999) Identification of variable regionsin the genomes of tubercle bacilli using bacterial artificialchromosome arrays. Mol Microbiol 32: 643–655.

Guinn, K.M., Hickey, M.J., Mathur, S.K., Zakel, K.L., Grotzke,J.E., Lewinsohn, D.M., et al. (2004) Individual RD1-regiongenes are required for export of ESAT-6/CFP-10 and forvirulence of Mycobacterium tuberculosis. Mol Microbiol 51:359–370.

Gutierrez, M.G., Mishra, B.B., Jordao, L., Elliott, E., Anes, E.,and Griffiths, G. (2008) NF-kappa B activation controlsphagolysosome fusion-mediated killing of mycobacteria bymacrophages. J Immunol 181: 2651–2663.

Hagedorn, M., Rohde, K.H., Russell, D.G., and Soldati, T.(2009) Infection by tubercular mycobacteria is spread bynonlytic ejection from their amoeba hosts. Science 323:1729–1733.

van Ingen, J., Boeree, M.J., de Lange, W.C., de Haas, P.E.,Dekhuijzen, P.N., and van Soolingen, D. (2008) Clinicalrelevance of Mycobacterium szulgai in The Netherlands.Clin Infect Dis 46: 1200–1205.

van Ingen, J., Bendien, S.A., de Lange, W.C., Hoefsloot, W.,Dekhuijzen, P.N., Boeree, M.J., and van Soolingen, D.(2009a) Clinical relevance of non-tuberculous mycobacte-ria isolated in the Nijmegen-Arnhem region, The Nether-lands. Thorax 64: 502–506.

van Ingen, J., de Zwaan, R., Dekhuijzen, R., Boeree, M., andvan Soolingen, D. (2009b) Region of difference 1 in non-tuberculous Mycobacterium species adds a phylogeneticand taxonomical character. J Bacteriol 191: 5865–5867.

de Jonge, M.I., Pehau-Arnaudet, G., Fretz, M.M., Romain, F.,Bottai, D., Brodin, P., et al. (2007) ESAT-6 from Mycobac-terium tuberculosis dissociates from its putative chaperoneCFP-10 under acidic conditions and exhibits membrane-lysing activity. J Bacteriol 189: 6028–6034.

Jordao, L., Bleck, C.K., Mayorga, L., Griffiths, G., and Anes,E. (2008) On the killing of mycobacteria by macrophages.Cell Microbiol 10: 529–548.

Kiers, A., Drost, A.P., van Soolingen, D., and Veen, J. (1997)Use of DNA fingerprinting in international source casefinding during a large outbreak of tuberculosis in The Neth-erlands. Int J Tuberc Lung Dis 1: 239–245.

Leake, E.S., Myrvik, Q.N., and Wright, M.J. (1984) Phagoso-mal membranes of Mycobacterium bovis BCG-immunealveolar macrophages are resistant to disruption byMycobacterium tuberculosis H37Rv. Infect Immun 45:443–446.

Lee, B.Y., Clemens, D.L., and Horwitz, M.A. (2008) The meta-bolic activity of Mycobacterium tuberculosis, assessed byuse of a novel inducible GFP expression system, correlateswith its capacity to inhibit phagosomal maturation and

acidification in human macrophages. Mol Microbiol 68:1047–1060.

Lee, B.Y., Jethwaney, D., Schilling, B., Clemens, D.L.,Gibson, B.W., and Horwitz, M.A. (2010) The Mycobacte-rium bovis bacille Calmette-Guerin phagosome proteome.Mol Cell Proteomics 9: 32–53.

Marquis, H., Doshi, V., and Portnoy, D.A. (1995) The broad-range phospholipase C and a metalloprotease mediatelisteriolysin O-independent escape of Listeria monocytoge-nes from a primary vacuole in human epithelial cells. InfectImmun 63: 4531–4534.

Mattos, K.A., Lara, F.A., Oliveira, V.G., Rodrigues, L.S.,D’Avila, H., Melo, R.C., et al. (2011) Modulation of lipiddroplets by Mycobacterium leprae in Schwann cells:a putative mechanism for host lipid acquisition andbacterial survival in phagosomes. Cell Microbiol 13: 259–273.

Meena, L.S., and Rajni. (2010) Survival mechanisms ofpathogenic Mycobacterium tuberculosis H37Rv. FEBS J277: 2416–2427.

Myrvik, Q.N., Leake, E.S., and Wright, M.J. (1984) Disruptionof phagosomal membranes of normal alveolar macro-phages by the H37Rv strain of Mycobacterium tuberculo-sis. A correlate of virulence. Am Rev Respir Dis 129:322–328.

Pandey, A.K., Yang, Y., Jiang, Z., Fortune, S.M., Coulombe,F., Behr, M.A., et al. (2009) NOD2, RIP2 and IRF5 play acritical role in the type I interferon response to Mycobacte-rium tuberculosis. PLoS Pathog 5: e1000500.

Parwati, I., van Crevel, R., and van Soolingen, D. (2010)Possible underlying mechanisms for successful emer-gence of the Mycobacterium tuberculosis Beijing genotypestrains. Lancet Infect Dis 10: 103–111.

Peters, P.J., Bos, E., and Griekspoor, A. (2006) Cryo-immunogold electron microscopy; including 11 videos online. Curr Protoc Cell Biol 4: 7.1–4.7.19.

Peyron, P., Vaubourgeix, J., Poquet, Y., Levillain, F., Botanch,C., Bardou, F., et al. (2008) Foamy macrophages fromtuberculous patients’ granulomas constitute a nutrient-richreservoir for M. tuberculosis persistence. PLoS Pathog 4:e1000204.

Pym, A.S., Brodin, P., Brosch, R., Huerre, M., and Cole, S.T.(2002) Loss of RD1 contributed to the attenuation of thelive tuberculosis vaccines Mycobacterium bovis BCG andMycobacterium microti. Mol Microbiol 46: 709–717.

Pym, A.S., Brodin, P., Majlessi, L., Brosch, R., Demangel, C.,Williams, A., et al. (2003) Recombinant BCG exportingESAT-6 confers enhanced protection against tuberculosis.Nat Med 9: 533–539.

Ray, K., Marteyn, B., Sansonetti, P.J., and Tang, C.M. (2009)Life on the inside: the intracellular lifestyle of cytosolicbacteria. Nat Rev 7: 333–340.

Raynaud, C., Guilhot, C., Rauzier, J., Bordat, Y., Pelicic, V.,Manganelli, R., et al. (2002) Phospholipases C areinvolved in the virulence of Mycobacterium tuberculosis.Mol Microbiol 45: 203–217.

Russell, D.G. (2011) Mycobacterium tuberculosis and theintimate discourse of a chronic infection. Immunol Rev 240:252–268.

Schnupf, P., and Portnoy, D.A. (2007) Listeriolysin O: aphagosome-specific lysin. Microbes Infect 9: 1176–1187.

Translocation is a virulence factor mycobacteria 1297

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298

Seto, S., Tsujimura, K., and Koide, Y. (2011) Rab GTPasesregulating phagosome maturation are differentiallyrecruited to mycobacterial phagosomes. Traffic 12: 407–420.

Simeone, R., Bobard, A., Lippmann, J., Bitter, W., Majlessi,L., Brosch, R., and Enninga, J. (2012) Phagosomal ruptureby Mycobacterium tuberculosis results in toxicity and hostcell death. PLoS Pathog 8: e1002507.

Smith, G.A., Marquis, H., Jones, S., Johnston, N.C., Portnoy,D.A., and Goldfine, H. (1995) The two distinct phospholi-pases C of Listeria monocytogenes have overlapping rolesin escape from a vacuole and cell-to-cell spread. InfectImmun 63: 4231–4237.

Sorensen, A.L., Nagai, S., Houen, G., Andersen, P., andAndersen, A.B. (1995) Purification and characterization ofa low-molecular-mass T-cell antigen secreted by Mycobac-terium tuberculosis. Infect Immun 63: 1710–1717.

Stamm, L.M., Morisaki, J.H., Gao, L.Y., Jeng, R.L.,McDonald, K.L., Roth, R., et al. (2003) Mycobacteriummarinum escapes from phagosomes and is propelled byactin-based motility. J Exp Med 198: 1361–1368.

Sweet, L., Singh, P.P., Azad, A.K., Rajaram, M.V.,Schlesinger, L.S., and Schorey, J.S. (2010) Mannosereceptor-dependent delay in phagosome maturation byMycobacterium avium glycopeptidolipids. Infect Immun 78:518–526.

Tsolaki, A.G., Gagneux, S., Pym, A.S., Goguet de la Salmo-niere, Y.O., Kreiswirth, B.N., Van Soolingen, D., and Small,P.M. (2005) Genomic deletions classify the Beijing/Wstrains as a distinct genetic lineage of Mycobacteriumtuberculosis. J Clin Microbiol 43: 3185–3191.

Weerdenburg, E.M., Peters, P.J., and van der Wel, N.N.(2010) How do mycobacteria activate CD8+ T cells?Trends Microbiol 18: 1–10.

van der Wel, N., Hava, D., Houben, D., Fluitsma, D., van Zon,M., Pierson, J., et al. (2007) M. tuberculosis and M. lepraetranslocate from the phagolysosome to the cytosol inmyeloid cells. Cell 129: 1287–1298.

Wong, K.W., and Jacobs, W.R., Jr (2011) Critical role forNLRP3 in necrotic death triggered by Mycobacteriumtuberculosis. Cell Microbiol 2011: 1371–1384.

Yuk, J.M., Shin, D.M., Lee, H.M., Yang, C.S., Jin, H.S., Kim,K.K., et al. (2009) Vitamin D3 induces autophagy inhuman monocytes/macrophages via cathelicidin. Cell HostMicrobe 6: 231–243.

Supporting information

Additional Supporting Information may be found in the onlineversion of this article:

Fig. S1. ESAT-6 secretion of different mycobacteria. M. mari-num E11, M. szulgaiand M. kansasiishow expression and secre-tion of ESAT-6 into the supernatant fraction, detected with themonoclonal a-ESAT-6 antibody Hyb 76-8. M. fortuitum, which hasno RD1 region, is negative for ESAT-6.Fig. S2. Difference between morphology cell cultures withoutor with non-adherent, dying cells. After the infection period onlythe adherent (mostly viable) cells were fixed and processed

(A; M. tuberculosis H37Rv, 4 days) or all cells, adherent andnon-adherent were fixed, collected together and processed forelectron microscopy (B; M. tuberculosis patient derived, 5 days).In a high magnification an example of a non-viable cell infectedwith M. tuberculosis (H37Rv, day 3 of infection) and immunogoldlabelled for LAMP. In (D) a non-viable cell infected by M. bovisBCG for 9 days and immunogold labelled for CD63. Bars in (A)and (B) represents 5 mm and in C and D 200 nm.Fig. S3. Percentage cytosolic bacteria in THP-1 in time. Per-centage cytosolic bacteria in adherent, non-apoptotic THP-1 cellsinfected with different mycobacterial species (see Figs 1 and 2,Table 1 and Table S1). Infections were performed with moi 10and kept as long as cells survived, or until day 10 in M. bovisBCG infections. Then cells were fixed and processed for immu-nogold labelling with lysosomal marker (CD63 or LAMP-1). Bac-teria were scored as cytosolic when no immunogold label waspresent and no membranes were detected surrounding the bac-teria. All infections were performed without antibiotics, except forthe M. smegmatis infection in which Amikacin was used to killextracellular bacteria. The THP-1 cell survived the infection for 2days without and 4 days with antibiotic. The day (D) of infectionis indicated under each bar. The pathogenic species are repre-sented as red, opportunistic species as yellow and non-pathogenic as green.Fig. S4. Overview and details of THP-1 cell infected for 3 dayswith M. szulgai. The electron micrograph of the M. szulgai infec-tion presented in Fig. 1 is used to visualize phagosomal (A) andcytosolic (B) bacteria at a high magnification and at low magni-fication (C). At a high magnification the host cytosol, presence (A)or absence (B) of the phagosomal membrane, capsular layer,bacterial plasma membrane and bacterial cytosol can be dis-cerned. The low magnification (C) of the infected cell demon-strates that the CD63 labelling is specifically labelling thephagosomal membrane, and small vesicles. Bar represents100 nm; black boxes in (C) represents enlarged area A and B;asterisk, cytosolic bacterium; encircled asterisk, phagosomalbacterium; m represents mitochondria; v vesicles and red circlesrepresent CD63 labelling.Fig. S5. Amino acid sequence alignment of the esat-6 gene inRD1-positive mycobacteria. All Mycobacterium tuberculosis iso-lates used here (H37Rv, Beijing and ancient Beijing, Harlingen,and reactivated non-transmitting isolate 1243) have identicalsequences. GenBank accession numbers are FJ014499 (M. tu-berculosis), BX248347 (M. bovis), CP000854 (M. marinum M ),EU826486 (M. szulgai ), X90946 (M. leprae), EU888292 (M. kan-sasii type I), EU888297 (M. kansasii type V) and CP000480(M. smegmatis).Fig. S6. Limited colocalization ubiquitin and LAMP-1 on dis-rupted phagolysosomes. Primary human DCs infected withH37Rv for 96 h, immunogold labelled for ubiquitin and protein Aconjugated to 5 nm gold and subsequently labelled for LAMP-1,rabbit anti-mouse bridging and protein A conjugated to 10 nm.Bar represents 100 nm and blue circles ubiquitin (5 nm), redcircles LAMP-1 (10 nm) gold particles.Table S1. Strains codes, some characteristics and reference ofmycobacterial species used.

Please note: Wiley-Blackwell are not responsible for the contentor functionality of any supporting materials supplied by theauthors. Any queries (other than missing material) should bedirected to the corresponding author for the article.

1298 D. Houben et al.

© 2012 Blackwell Publishing Ltd, Cellular Microbiology, 14, 1287–1298