Listeria monocytogenes transiently altersmitochondrial dynamics during infectionFabrizia Stavrua,b,1, Frédéric Bouillaudc,d,e, Anna Sartorif, Daniel Ricquierc,d,e, and Pascale Cossarta,b,1

aUnité des Interactions Bactéries-Cellules, Institut Pasteur; U604, Institut National de la Santé et de la Recherche Médicale and bUSC2020, Institut Nationalde la Recherche Agronomique, 75015 Paris, France; cInstitut Cochin, dInstitut National de la Santé et de la Recherche Médicale U1016, and eCentreNational de la Recherche Scientifique Unité Mixte de Recherche 8104, Université Paris Descartes, 75014 Paris, France; and fPlate-Forme de MicroscopieUltrastructurale, Imagopole, Institut Pasteur, Unité de Recherche Associée 2185; 75015 Paris, France

Contributed by Pascale Cossart, January 20, 2011 (sent for review August 30, 2010)

Mitochondria are essential and highly dynamic organelles, con-stantly undergoing fusion and fission. We analyzed mitochondrialdynamics during infection with the human bacterial pathogenListeria monocytogenes and show that this infection profoundlyalters mitochondrial dynamics by causing transient mitochondrialnetwork fragmentation. Mitochondrial fragmentation is specific topathogenic Listeria monocytogenes, and it is not observed withthe nonpathogenic Listeria innocua species or several other intra-cellular pathogens. Strikingly, the efficiency of Listeria infection isaffected in cells where either mitochondrial fusion or fission hasbeen altered by siRNA treatment, highlighting the relevance ofmitochondrial dynamics for Listeria infection. We identified thesecreted pore-forming toxin listeriolysin O as the bacterial factormainly responsible for mitochondrial network disruption and mi-tochondrial function modulation. Together, our results suggestthat the transient shutdown of mitochondrial function and dy-namics represents a strategy used by Listeria at the onset of in-fection to interfere with cellular physiology.

bacterial infection | calcium influx | bioenergetics

Mitochondria are essential organelles providing most cellu-lar ATP as well as biosynthetic intermediates. They have

emerged as important integrators of several signaling cascades(1). Fusion and fission of mitochondria occurs constantly, regu-lating their size and subcellular distribution and reflecting theirfunctional state (1, 2). Disturbance of mitochondrial dynamicsleads to pathological conditions exacerbated in tissues with highmetabolic demand such as neuronal or muscle tissues (3). Al-though it often is unclear whether defects in mitochondrial dy-namics are the cause or effect of the disease, mitochondria havenow emerged as drug targets for several pathologies (4, 5).At the single-cell level, long-term impairment of either fusion

or fission can cause respiratory defects (6, 7). Key players in thefusion process include the outer mitochondrial membraneGTPases mitofusin 1 and 2 (Mfn1/2) (8, 9), and the inner mem-brane GTPase optic atrophy 1 (Opa1) (10, 11), but the moleculardetails of the fusion mechanism remain obscure (12, 13). Mito-chondrial fission critically depends on the GTPase dynamin-related protein 1 (Drp1) (14), which oligomerizes on the outermitochondrial membrane to constrict mitochondria at divisionsites (15). Mitochondrial fission also occurs during apoptosis andmay be required for apoptosis progression under specific circum-stances (16), although it does not induce apoptosis per se (17–19).Several pathogens including both viruses and bacteria directly

or indirectly target mitochondria to interfere with the host ap-optotic machinery (20–24). Depending on the pathogen andhost-cell type, this interference can inhibit cell death to preservethe pathogen’s replication niche or induce cell death to promoteinfection spreading (25). Mitochondria also function as signalingplatforms in the innate immune response, and this function hasbeen linked recently to mitochondrial dynamics during viral in-fection (26, 27).To investigate the interrelation between host-cell mitochon-

drial dynamics and bacterial infection, we focused on Listeriamonocytogenes, a facultative intracellular bacterium causing the

food-borne disease listeriosis. While listeriosis is a public healthissue, L. monocytogenes has been instrumental in elucidatingfundamental cell biological questions, e.g., actin polymerizationprinciples (reviewed in refs. 28 and 29).After cell invasion, L. monocytogenes uses the pore-forming

toxin listeriolysin O (LLO) to escape from the phagosome. Al-though LLO function had been characterized first in Listeria es-cape from the phagosome under acidic conditions (30), severalstudies now indicate that this crucial virulence factor also displaysactivity at neutral pH and acts on cells before bacterial entry (31,32). Indeed, low levels of LLO secreted before bacterial entry aresufficient to activate prosurvival signaling cascades such as theNFκB and MAPK pathways (33, 34), transcriptionally reprogramhost cells (35), and trigger global deSUMOylation (36).Here we report that L. monocytogenes causes dramatic alter-

ations of mitochondrial dynamics via LLO. Strikingly, mitochon-drial fragmentation induced by Listeria infection is a transientphenomenon, indicating that mitochondria are not terminallydamaged. We propose that modulation of mitochondrial dy-namics and function is a strategy used by pathogenic Listeria atthe onset of infection to slow down mitochondrial activity andmitochondria-dependent processes.

ResultsInfection with Pathogenic Listeria Induces Mitochondrial Fragmentation.We first investigated the effects of L. monocytogenes infection onmitochondria of epithelial cells by using confocal laser scanningmicroscopy. Fig. 1A shows L. monocytogenes infection of HeLacells [1 h, multiplicity of infection (MOI) of 50] inducing strongand rapid mitochondrial network fragmentation. Importantly, thisnetwork fragmentation is specific to pathogenic L. monocytogenes,because it is not observed with the closely related nonpathogenicspecies Listeria innocua, even when cells are infected with L.innocua overexpressing the L. monocytogenes invasin Internalin B(InlB) to enter cells [L. innocua(InlB), Fig. 1A]. This findingindicates that mitochondrial fragmentation is not a consequenceof stress imposed by the engulfment of bacteria. Furthermore, L.monocytogenes-induced mitochondrial fragmentation is not re-stricted to HeLa cells, because it also occurs in human placental(Jeg3) and green monkey kidney (Vero) cells (Fig. S1A).To assess whether other invasive pathogens also cause mito-

chondrial fragmentation, we infected cells with Salmonella enter-ica serovar typhimurium, Escherichia coli(Inv) as a model forYersinia pseudotubercolosis (37), enteropathogenic E. coli (EPEC),and Shigella flexneri. Importantly, infection with either the extra-cellular pathogen EPEC or with invasive pathogens that remain

Author contributions: F.S. and P.C. designed research; F.S. and A.S. performed research;F.B. and D.R. contributed new reagents/analytic tools; F.S., F.B., and P.C. analyzed data;and F.S., F.B., and P.C. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.1To whom correspondence may be addressed. E-mail: pcossart@pasteur.fr or fstavru@pasteur.fr.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1100126108/-/DCSupplemental.

3612–3617 | PNAS | March 1, 2011 | vol. 108 | no. 9 www.pnas.org/cgi/doi/10.1073/pnas.1100126108

confined to a phagocytic vacuole [i.e., E. coli(Inv) and Salmonellaenterica serovar typhimurium] did not appear to affect mitochon-dria (Fig. 1B), even at an MOI of 100 and up to 3 h of infection,supporting the notion that L. monocytogenes-induced mitochon-drial fragmentation is not a general stress response to bacterialinfection. Moreover, mitochondrial fragmentation is not causedby the presence of bacteria in the cytosol, because infection withShigella flexneri, which, like Listeria, escapes from the vacuole andpolymerizes actin to move intra- and intercellularly, did not causemitochondrial fragmentation (Fig. 1B).Correlative light/transmission electron microscopy (TEM)

showed that fragmented mitochondria in Listeria-infected cellshad a disorganized ultrastructure with remodeled cristae com-pared with mitochondria in noninfected cells (Fig. S1B).

Impairment of Mitochondrial Dynamics Affects Infection by Listeriamonocytogenes. Having established that L. monocytogenes infec-tion induces mitochondrial fragmentation, we asked whether in-hibiting fusion or fission by siRNA would affect early Listeriainfection stages. Cells depleted of the fusion proteins Mfn1 andMfn2 (resulting in fissioned mitochondria) or of the fission pro-tein Drp1 (resulting in hyperfused mitochondria) were infectedwith L. monocytogenes. Infection was strongly impaired in cellswith fissioned mitochondria, and the strongest inhibition wasseen with codepletion of both mitofusins (Fig. 2 A–C and Fig. S2A and B). Infection was more efficient in cells with hyperfusedmitochondria (Fig. 2D and Fig. S2C). This result suggests that, forefficient infection, Listeria requires the ability to induce mito-chondrial fission, because cells with mitochondria that already arefissioned at the onset of infection provide an unfavorable envi-ronment for infection. Silencing of proteins regulating mitochon-drial dynamics may also affect later stages of infection or haveindirect effects on mitochondrial function and infection.

Listeriolysin O Is Sufficient to Cause Mitochondrial Fragmentation.Because mitochondria appeared to fragment at early stages ofinfection, we tested whether a secreted effector of L. mono-cytogenes could cause mitochondrial fragmentation. We first useda noninvasive mutant of L. monocytogenes lacking InlB and foundthat this mutant was still able to induce mitochondrial fragmen-tation (Fig. 3A). We then asked whether the best-characterizedsecreted effector of L. monocytogenes, the pore-forming toxinLLO, could induce mitochondrial fragmentation. Strikingly, mi-tochondrial fragmentation was abolished when an LLO-deletionmutant (L.monocytogenesΔhly) was used (Fig. 3A), indicating thatLLO is required for fragmentation of host-cell mitochondria.Bacteria whose hemolysin (hly) gene carries a point mutationdisrupting pore formation (38) did not affect mitochondrial mor-phology (Fig. 3B), revealing that fragmentation depends on thepore-forming ability of LLO.LLO appears necessary and sufficient to induce mitochondrial

fragmentation, because addition of the purified toxin at nano-molar concentrations [3–6 nM, considered noncytotoxic and notcausing lactate dehydrogenase (LDH) release (35)] recapitulatedthe mitochondrial phenotype observed upon infection (Fig. 3C).Mitochondrial fragmentation was found to occur in a fast, all-or-nothing manner (i.e., within less than 10 min) upon addition ofrecombinant LLO (Movie S1). Increasing incubation time or LLOconcentration resulted in an increased number of cells displayingfragmented mitochondria (Fig. 3D).

Fig. 1. Mitochondrial fragmentation upon infection with Listeria mono-cytogenes. (A) HeLa cells infected with L. monocytogenes (1 h; MOI of 50)display strongly fragmented mitochondria, whereas cells infected withL. innocua expressing InlB to enter cells do not. Mitochondria were stainedwith MitoTracker Deep Red (red). L. innocua or L. monocytogenes werestained with polyclonal antibodies R6 and R11 (green). Arrows indicate in-fected cells, and insets show 2× enlargements. (B) Infection (1 h, MOI of 50)with E.coli(Inv) as a model for Yersinia infection, Shigella flexneri (M09T),Salmonella enterica serovar typhimurium, and enteropathogenic E. coli (EPEC)do not cause mitochondrial fragmentation. Cell nuclei and bacteria werestained with DAPI (blue). Salmonella additionally expresses GFP (green). Mi-tochondria were stained as in A. Arrows indicate infected cells, and insetsshow 2× enlargements.

Fig. 2. Early Listeria infection is affected by impaired mitochondrial dy-namics. (A) HeLa cells treatedwith two different siRNAs (#A and #B) targetingthe mitochondrial fusion protein Mfn1 were infected with L. monocytogenes(MOI of 50) in a gentamicin survival assay. The relative number of intracellularbacteria was determined by cfu count at 3 h postinfection. L. monocytogenesinfection is significantly impaired in Mfn1-knockdown cells (P < 0.001, one-tailed Student’s t test). Each experiment was performed in triplicate, anda representative experiment with SDs is shown. At least three independentexperiments were performed for each condition. (B) Infection of Mfn2-knockdown cells was performed as in A. Mfn2 knockdown also impairs Lis-teria infection, although a lesser extent. (C) Down-regulation of both Mfn1and Mfn2 has a cumulative effect in impairing Listeria infection, suggestingthat these proteins play nonredundant roles in infection. (D) L. mono-cytogenes infection is promoted in Drp1-knockdown cells. Experiment andstatistical analysis were performed as in A–C. **P < 0.005, ***P < 0.001 byone-tailed Student’s t test.

Stavru et al. PNAS | March 1, 2011 | vol. 108 | no. 9 | 3613

CELL

BIOLO

GY

Mitochondrial fragmentation could result from enhanced fis-sion or decreased fusion (39). We thus analyzed whether LLOtreatment would affect total levels of key mitochondrial dynamicsmediators, i.e., Mfn1, Mfn2, and Drp1. Total levels of these pro-teins did not decrease in cells treated with 3–6 nM recombinantLLO (Fig. 3E). Because LLO forms ion-permeable pores inmembranes (30), we hypothesized that LLO might modulate mi-tochondrial dynamics by inducing ion flux. We tested whetherblocking K+ efflux or Ca2+ influx would prevent LLO-inducedmitochondrial fragmentation, given that changes in these ions areknown to affect mitochondrial morphology (40, 41). Blocking K+

efflux from LLO-treated cells by incubation in high extracellularK+ concentrations (135 mM) had no effect (Fig. 4). In contrast,interfering with LLO-induced Ca2+ influx by performing LLOtreatment in Ca2+-free medium strongly prevented mitochondrialfragmentation (Fig. 4). This result suggested that Ca2+ influx(rather than K+ efflux) through LLO pores is the signal inducingmitochondrial fragmentation.

LLO Treatment Causes Mitochondrial Membrane Potential Loss anda Drop in Respiration and Cellular ATP. To investigate the func-tionality of mitochondria in LLO-treated cells, we measured themitochondrial membrane potential ΔΨ by assessing release oftheΔΨ-dependent mitochondrial dye tetramethylrhodamine ethylester (TMRE) and found that LLO treatment caused a significantdrop inΔΨ compared with untreated cells (Fig. 5A). At the single-cell level, time-lapse microscopy of mitochondria loaded with theTMRE derivative tetramethylrhodamine methyl ester (TMRM)indicated that ΔΨ loss was concomitant with or immediately pre-ceded mitochondrial fragmentation (Fig. 5B).LLO treatment caused a drop in respiratory activity as mea-

sured with an oxygen consumption chamber (Fig. S3). Respirationresumed when the mitochondrial respiratory substrate succinatewas added, indicating that no major damage occurred to mito-chondria, because the respiratory chain (complex II–complex IV)appeared functional. Furthermore, the resumption of respiration

demonstrates that LLO does not form pores in the mitochondrialinner membrane, a finding that is supported further by immuno-fluorescence analysis, showing plasma membrane rather thanmitochondrial localization of LLO (Fig. S4). In contrast, LLOappears to permeabilize the otherwise succinate-impermeableplasma membrane, contributing to mitochondrial depolarizationby allowing leakage of bioenergetic substrates out of the cell.Infection also affected cellular ATP levels in an LLO-

dependent manner: Infection with wild-type L. monocytogenes(1 h, MOI of 50) induced a 50% decrease in intracellular ATPlevels, but this decrease was not observed with L. innocua(InlB)or theL.monocytogenes Δhlymutant (Fig. 5C). Recombinant LLO(6 nM, 10 min) was sufficient to cause an even more pronounceddecrease in intracellular ATP levels (Fig. 5C). A similar decreasewas obtained by a combination of chemically induced mitochon-drial uncoupling with carbonyl cyanide m-chlorophenylhydrazone(CCCP) and Ca2+ influx (A23187), although uncoupling or Ca2+

influx separately did not lead to a significant difference (Fig. 5D).These data strongly suggest that L. monocytogenes infection notonly affects mitochondrial dynamics but also interferes with cel-lular bioenergetics and mitochondrial function.

LLO-Induced Mitochondrial Fragmentation Does Not Correlate withClassical Apoptosis. Mitochondrial fragmentation has been de-scribed in apoptotic cells, prompting us to analyze apoptosismarkers such as cytochrome c release. Cytochrome c was not re-leased from fragmented mitochondria upon infection or LLOtreatment (Fig. 3A and Fig. S5A). In contrast, cytochrome c re-lease was observed in the positive control, i.e., staurosporine-treated cells (Fig. S5A). Interestingly, the observed mitochondrialfragmentation did not depend on mitochondrial transition pore(mTP) opening, because it was not blocked by the mTP inhibitorcyclosporin A. To test further for apoptosis in cells displayinginfection-induced mitochondrial fragmentation, we analyzed ac-tivated B-cell lymphoma 2 (Bcl2)-associated X (Bax) protein byimmunostaining. The proapoptotic Bcl2 protein family member

Fig. 3. LLO induces mitochondrial fragmentation with-out affecting total levels of key mitochondrial dynamicsproteins. (A) HeLa cells infected (1 h, MOI of 50) witha noninvasive L. monocytogenes mutant (ΔinlB) displayfragmented mitochondria, indicating bacterial entry isnot required. In contrast, no fragmentation occurs withinfection by an LLO-deficient mutant (Δhly). Bacteriawere stained with anti L. monocytogenes (R11, green),mitochondria were stained with anti-cytochrome c (red),and DNA was stained with DAPI (blue). Insets show 2×enlargements of mitochondria. Arrows point to infectedcells. (B) L. monocytogenes carrying a point mutation inthe hly gene inactivating its pore-forming ability (W492A)do not cause mitochondrial fragmentation. Mitochondria(red) and bacteria (green) were stained as in A. (C) HeLacell treatment with recombinant LLO (10 min, 6 nM) wassufficient to induce mitochondrial fragmentation. Mito-chondria (red) were stained as in A. (D) HeLa cells weretreated with different concentrations of LLO for 10 min orwith 6 nM LLO for different periods of time. Morpho-metric analysis indicated that LLO-induced mitochondrialfragmentation is concentration and time dependent atthe cell-population level. The percentage of fragmenta-tion was determined by counting at least 100 cells perdata point; data from at least two independent experi-ments were pooled, and P values were calculated usingone-tailed Student’s t test (***P < 0.005, *P < 0.25). (E)Cells treated for 10 min with the indicated amounts ofLLO were analyzed by Western blot for Drp1, Mfn1, orMfn2, showing that LLO treatment does not affect theirtotal levels.

3614 | www.pnas.org/cgi/doi/10.1073/pnas.1100126108 Stavru et al.

Bax is activated by several apoptotic signals and translocates tomitochondria, where it forms pores and participates in fission (42,43). We could not detect mitochondrial recruitment of activatedBax upon infection or LLO treatment (Fig. S5B).Together, these experiments suggest that neither L. mono-

cytogenes infection nor treatment with sublytic LLO concen-trations causes classical apoptosis in HeLa cells.

Infection-Induced Mitochondrial Fragmentation Is Transient. An im-portant question was whether mitochondria that fragment be-cause of LLO produced at initial stages of Listeria infectionrecover their original shape and reform an interconnected net-work. To this end, we followed recovery by infecting cells for 1 hand then observing mitochondrial morphology at different timepoints after bacteria removal. The proportion of cells with frag-mented mitochondria decreased steadily (Fig. 6A). Time-lapseimaging indicated that in most cases tubular mitochondrial mor-phology is recovered within the first few hours (Fig. 6B). In linewith these results, we found that intracellular ATP levels re-covered by 4 h postinfection (Fig. S6), suggesting that infectedcells are not terminally damaged.

DiscussionWe show here that infection with pathogenic L. monocytogenescauses transient mitochondrial network fragmentation and iden-tify the secreted bacterial toxin LLO as the main factor affectingmitochondrial morphology and function at early time points ofinfection. LLO induces Ca2+-dependent mitochondrial frag-mentation, accompanied by a decrease in themitochondrial mem-brane potential ΔΨ and in respiration. As ΔΨ and respirationconcomitantly decrease upon LLO treatment, ATP regenerationproceeds inefficiently, contributing to a decrease in intracellularATP levels. This decrease suggests a transient metabolic “slow-down” of host cells, favoring early stages of infection by interferingwith the capacity of the host cell to respond to this event. Suchstrategy could be common to several bacteria secreting pore-forming toxins, because we found that different recombinantpore-forming toxins had effects comparable to LLO (Fig. S7).LLO pore formation has been studied at the biophysical and

physiological level and induces Ca2+ influx (32, 44). Interestingly,several pathogens manipulate host-cell physiology by inducingCa2+fluxes (45). Importantly, this Ca2+ influx enhances Listeriaentry into cells (31), suggesting that early LLO action is a crucialstep in epithelial cell infection. Ca2+ influx probably representsa first bioenergetic insult to the cell, inducing mitochondrialfragmentation and depolarization as well as blocking mitochon-drial movement (Movie S1), although such damage is reversible.While neither uncoupling nor Ca2+ ionophore treatment signifi-cantly reduce intracellular ATP levels, such an ATP decrease is

reproduced partially by the synergistic action of an uncoupler and aCa2+ ionophore but these drugs do not cause the dramatic mito-chondrial fragmentation observed upon LLO treatment (Fig. S8).In the case of Listeria infection, the bioenergetic crisis is probablyaggravated by leakage of small molecules, including respiratorysubstrates or glycolysis intermediates through LLO pores.LLO-inducedΔΨ decrease also may reflect a host-cell response

to prevent mitochondrial Ca2+ accumulation to cytotoxic levels,because mitochondrial Ca2+ uptake is ΔΨ dependent (46). In-deed, several markers of apoptosis were absent in cells with LLO-or infection-induced mitochondrial fragmentation. Our data areconsistent with the notion that epithelial cells recover from theattack of pore-forming toxins (including LLO) at sublytic con-centrations, regain membrane integrity, and resume the cell cycle(32, 47, 48). The molecular mechanisms underlying LLO recoveryare currently unclear: LLO does not colocalize with endocyticmarkers at early time points (Fig. S4). Infected cells appear torestore their mitochondrial network both morphologically (Fig. 6)and functionally, because intracellular ATP levels recover 4 hpostinfection (Fig. S6). Consistent with the view that cytosolicLLO is inactivated rapidly (30), mitochondrial fragmentation isinduced only by extracellular LLO, i.e., early during Listeria in-fection; at late time points of infection intracellular bacteria do notaffect mitochondrial morphology, even though they produce LLOto escape from the vacuole (Fig. S9). Together, these data indicatethat Listeria affects host-cell mitochondria only transiently. Per-manent impairment of mitochondrial function and dynamicswould harm host cells and therefore would be counterselected for,because it would eliminate the bacterial replication niche.Our data suggest that active induction of mitochondrial frag-

mentation early during infection is critical for infection. Indeed,infection is impaired in cells with previously fissioned mito-chondria and is enhanced in cells with hyperfused mitochondria.Treatments that impair mitochondrial respiration cause fragmen-tation (49), and, conversely, mitochondrial fragmentation has beenshown to limit the spread of incoming Ca2+ across the mitochon-drial network (50). Accordingly, fragmentation appears to be anappropriate response to avoid propagation of the consequences ofCa2+ influx to the entire mitochondrial network. A prefissionedstate would limit the extent of damage caused by LLO action andallow the mitochondrial network to restore cellular bioenergeticsmore efficiently. Consequently, invading Listeria would have lesstime to take advantage from the transient bioenergetic “slow-down” it induces. The opposite effect would occur in Drp1-knockdown cells with a hyperfused mitochondrial network.In conclusion, we propose a scenario in which the normal

bioenergetic state of the cell represents a barrier to Listeria in-vasion. Consequently, an LLO-induced transient metabolic re-programming of the cell would promote efficient infection. In

Fig. 4. LLO-induced mitochondrial fragmentation ismediated by calcium influx, not by potassium efflux. Cellswere treated for 10 min with 6 nM LLO in the presence orabsence of extracellular calcium. Mitochondria (stainedwith anti-cytochrome c) showed strongly inhibited frag-mentation in the absence of calcium, whereas blockingpotassium efflux by incubation in 135 mM KCl did notprevent LLO-induced mitochondrial fragmentation.

Stavru et al. PNAS | March 1, 2011 | vol. 108 | no. 9 | 3615

CELL

BIOLO

GY

agreement with this notion, mitochondrial dysfunction has beenlinked to increased susceptibility to bacterial infection (51, 52).Our work shows that mitochondrial dynamics plays a role in in-fection with the human pathogen L. monocytogenes. Whether inthis case specific signaling cascades are activated downstream ofthe induced mitochondrial fragmentation and dysfunction is cur-rently unknown, but mitochondrial dynamics and infection doappear to influence each other mutually, because Listeria in-fection specifically leads to transient disruption of mitochondrialmorphology and function, and prior disruption of mitochondrialdynamics by siRNA affects Listeria infection efficiency. Inter-estingly, the cytomegalovirus protein vMia induces mitochondrialfragmentation and thereby prevents innate immune signalingdownstream of mitochondrial antiviral signaling protein (MAVS)(26). Listeria may act similarly, explaining MAVS-independentactivation of the innate immune response (53). Given that bio-energetics affect mitochondrial shape, and vice-versa (2, 49), it islikely that mitochondrial localization of innate immunity com-ponents serves to coordinate innate immune responses with cel-lular energy levels via sensing of mitochondrial morphology andfunction. How innate immunity components and the fission/

Fig. 5. LLO causes a decrease in the mitochondrial membrane potentialΔΨ concomitant with fragmentation and loss of intracellular ATP. (A) Themitochondrial membrane potential ΔΨ of HeLa cells was measured ina plate-based TMRE-release assay. Cells were treated with the indicateddrugs and loaded with TMRE in a ΔΨ-dependent manner. MitochondrialTMRE then was released by overnight incubation at 4 °C and measured at585 nm, reflecting drug-induced changes in ΔΨ. Mitochondria wereuncoupled with 10 μM CCCP/0.5 μg/mL oligomycin (control for ΔΨ de-crease), whereas oligomycin alone induced higher ΔΨ by blocking the F1FOATPase (hyperpolarization control). LLO treatment caused a decrease inΔΨ. Normalized average values from two independent experiments areshown, and P values were calculated (**P < 0.01, *P < 0.25, one-tailedStudent’s t test). (B) Time-lapse analysis of HeLa cells stained with Mito-Tracker 488 (green) and the potential-sensitive dye TMRM (red) shows thatΔΨ decreases shortly before mitochondrial fragmentation induced by ad-dition of 6 nM LLO becomes apparent. Fragmented mitochondria loseTMRM and retain only the potential-insensitive MitoTracker 488. (C ) In-tracellular ATP levels were measured 1 h after infection with L. innocuaexpressing InlB, L. monocytogenes Δhly, or wild-type L. monocytogenes.Wild-type L. monocytogenes causes the strongest drop in intracellular ATPlevels; the effect is more pronounced when cells are incubated withrecombinant LLO (10 min, 6 nM). Cell permeabilization with detergent (0.1%Triton X-100, 10 min) causes complete intracellular ATP release. Experimentswere performed three times in duplicate, and mean values (normalized tountreated cells) are shown. ***P < 0.0005, **P < 0.005, *P < 0.05, one-tailedStudent’s t test. (D) HeLa cells were treated for 10 min with 6 nM LLO or 0.1%Triton X-100, for 30 min with the mitochondrial uncouplers CCCP (100 μM),carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) (1 μM), andvalinomycin (100 nM), or with the Ca2+ ionophore A23187 (1 μM). In-tracellular ATP levels do not decrease significantly upon Ca2+ influx oruncoupling but do decrease significantly with the combination of bothtreatments. Statistical analysis was performed as in C. n.s., nonsignificantvalues.

Fig. 6. Mitochondria fragmented upon L. monocytogenes infection canrecover tubular morphology. (A) HeLa cells infected with L. monocytogenes(1 h, MOI of 50) were fixed directly or were washed and further incubatedfor 1.5 h, 3 h, or overnight (i.e., 2.5 h, 4 h, and overnight after inoculation) inthe presence of 80 μg/mL gentamicin to kill extracellular bacteria rapidly andto slow intracellular bacterial replication. The number of cells displayingfragmented mitochondria decreases steadily with time, indicating mito-chondrial network recovery. The percentage of cells with fragmented mi-tochondria was determined by counting >100 cells per time point in twoindependent experiments. (B) HeLa cells were loaded with 100 nM Mito-Tracker 488, treated as in A, and imaged at the indicated time points. Arrowsindicate cells with fragmented mitochondria that recover tubular morphol-ogy starting at 2.5 h postinfection.

3616 | www.pnas.org/cgi/doi/10.1073/pnas.1100126108 Stavru et al.

fusion machinery sense changes in the bioenergetic status of mi-tochondria is still unknown. The strong and rapid action of LLOdescribed in our work provides an additional model system toaddress this question.

Materials and MethodsReagents, cell lines, and bacterial strains used in this study, detailed exper-imental protocols, and nonstandard abbreviations are provided in SI Mate-rials and Methods.

ACKNOWLEDGMENTS. We acknowledge the expert help of C. Schmitt andG. Pehau-Arnaudet of the Plate-Forme de Microscopie Ultrastructurale andthe continuous support of the Plate-Forme d’Imagerie Dynamique staff atthe Institut Pasteur. We thank E. Veiga and L. Dortet for pioneering experi-ments, K. Rogers and M. Lecuit for stimulating discussions and laboratorymembers for critical reading of the manuscript. This work received supportfrom the European Research Council (Advanced Grant 233348), the Fonda-tion Le Roch Les Mousquetaires, and the Fondation Jeantet. P.C. is a HowardHughes Medical Institute International Fellow. F.S. was funded by a RouxFellowship (Institut Pasteur) and currently is a European Molecular BiologyOrganization Long-Term Fellow.

1. Soubannier V, McBride HM (2009) Positioning mitochondrial plasticity within cellularsignaling cascades. Biochim Biophys Acta 1793:154–170.

2. Sauvanet C, Duvezin-Caubet S, di Rago JP, Rojo M (2009) Energetic requirements andbioenergetic modulation of mitochondrial morphology and dynamics. Semin Cell DevBiol 21:558–565.

3. Detmer SA, Chan DC (2007) Functions and dysfunctions of mitochondrial dynamics.Nat Rev Mol Cell Biol 8:870–879.

4. Aliev G, et al. (2009) Brain mitochondria as a primary target in the development oftreatment strategies for Alzheimer disease. Int J Biochem Cell Biol 41:1989–2004.

5. Fulda S, Galluzzi L, Kroemer G (2010) Targeting mitochondria for cancer therapy. NatRev Drug Discov 9:447–464.

6. Chen H, Chan DC (2005) Emerging functions of mammalian mitochondrial fusion andfission. Hum Mol Genet 14(Spec No. 2):R283–289.

7. Parone PA, et al. (2008) Preventing mitochondrial fission impairs mitochondrialfunction and leads to loss of mitochondrial DNA. PLoS One 3:e3257.

8. Rojo M, Legros F, Chateau D, Lombès A (2002) Membrane topology and mitochondrialtargeting of mitofusins, ubiquitous mammalian homologs of the transmembraneGTPase Fzo. J Cell Sci 115:1663–1674.

9. Santel A, Fuller MT (2001) Control of mitochondrial morphology by a humanmitofusin. J Cell Sci 114:867–874.

10. Alexander C, et al. (2000) OPA1, encoding a dynamin-related GTPase, is mutated inautosomal dominant optic atrophy linked to chromosome 3q28. Nat Genet 26:211–215.

11. Delettre C, et al. (2000) Nuclear gene OPA1, encoding a mitochondrial dynamin-related protein, is mutated in dominant optic atrophy. Nat Genet 26:207–210.

12. Chen H, et al. (2003) Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrialfusion and are essential for embryonic development. J Cell Biol 160:189–200.

13. Eura Y, Ishihara N, Yokota S, Mihara K (2003) Two mitofusin proteins, mammalianhomologues of FZO, with distinct functions are both required for mitochondrialfusion. J Biochem 134:333–344.

14. Smirnova E, Shurland DL, Ryazantsev SN, van der Bliek AM (1998) A human dynamin-related protein controls the distribution of mitochondria. J Cell Biol 143:351–358.

15. Smirnova E, Griparic L, Shurland DL, van der Bliek AM (2001) Dynamin-related proteinDrp1 is required for mitochondrial division in mammalian cells.Mol Biol Cell 12:2245–2256.

16. Frank S, et al. (2001) The role of dynamin-related protein 1, a mediator of mitochondrialfission, in apoptosis. Dev Cell 1:515–525.

17. Estaquier J, Arnoult D (2007) Inhibiting Drp1-mediated mitochondrial fissionselectively prevents the release of cytochrome c during apoptosis. Cell Death Differ14:1086–1094.

18. Cassidy-Stone A, et al. (2008) Chemical inhibition of the mitochondrial divisiondynamin reveals its role in Bax/Bak-dependent mitochondrial outer membranepermeabilization. Dev Cell 14:193–204.

19. Parone PA, et al. (2006) Inhibiting the mitochondrial fission machinery does notprevent Bax/Bak-dependent apoptosis. Mol Cell Biol 26:7397–7408.

20. Ashktorab H, et al. (2004) Bax translocation and mitochondrial fragmentationinduced by Helicobacter pylori. Gut 53:805–813.

21. Faherty CS, Maurelli AT (2008) Staying alive: Bacterial inhibition of apoptosis duringinfection. Trends Microbiol 16:173–180.

22. Kepp O, Rajalingam K, Kimmig S, Rudel T (2007) Bak and Bax are non-redundantduring infection- and DNA damage-induced apoptosis. EMBO J 26:825–834.

23. Blanke SR (2005) Micro-managing the executioner: Pathogen targeting of mitochondria.Trends Microbiol 13:64–71.

24. Boya P, Roques B, Kroemer G (2001) New EMBO members’ review: Viral and bacterialproteins regulating apoptosis at the mitochondrial level. EMBO J 20:4325–4331.

25. Arnoult D, Carneiro L, Tattoli I, Girardin SE (2009) The role of mitochondria in cellulardefense against microbial infection. Semin Immunol 21:223–232.

26. Castanier C, Garcin D, Vazquez A, Arnoult D (2010) Mitochondrial dynamics regulatethe RIG-I-like receptor antiviral pathway. EMBO Rep 11:133–138.

27. Yasukawa K, et al. (2009) Mitofusin 2 inhibits mitochondrial antiviral signaling. SciSignal 2(ra47):1–11.

28. Cossart P, Toledo-Arana A (2008) Listeria monocytogenes, a unique model in infectionbiology: An overview. Microbes Infect 10:1041–1050.

29. Hamon MA, Cossart P (2008) Histone modifications and chromatin remodeling duringbacterial infections. Cell Host Microbe 4:100–109.

30. Schnupf P, Portnoy DA (2007) Listeriolysin O: A phagosome-specific lysin. MicrobesInfect 9:1176–1187.

31. Dramsi S, Cossart P (2003) Listeriolysin O-mediated calcium influx potentiates entry ofListeria monocytogenes into the human Hep-2 epithelial cell line. Infect Immun 71:3614–3618.

32. Gekara NO, et al. (2007) The multiple mechanisms of Ca2+ signalling by listeriolysin O,the cholesterol-dependent cytolysin of Listeria monocytogenes. Cell Microbiol 9:2008–2021.

33. Tang P, Rosenshine I, Cossart P, Finlay BB (1996) Listeriolysin O activates mitogen-activated protein kinase in eucaryotic cells. Infect Immun 64:2359–2361.

34. Kayal S, et al. (1999) Listeriolysin O-dependent activation of endothelial cells duringinfection with Listeria monocytogenes: Activation of NF-kappa B and upregulation ofadhesion molecules and chemokines. Mol Microbiol 31:1709–1722.

35. HamonMA, et al. (2007) Histone modifications induced by a family of bacterial toxins.Proc Natl Acad Sci USA 104:13467–13472.

36. Ribet D, et al. (2010) Listeria monocytogenes impairs SUMOylation for efficientinfection. Nature 464:1192–1195.

37. Isberg RR, Falkow S (1985) A single genetic locus encoded byYersinia pseudotuberculosispermits invasion of cultured animal cells by Escherichia coli K-12. Nature 317:262–264.

38. Michel E, Reich KA, Favier R, Berche P, Cossart P (1990) Attenuated mutants of theintracellular bacterium Listeria monocytogenes obtained by single amino acidsubstitutions in listeriolysin O. Mol Microbiol 4:2167–2178.

39. Chan DC (2006) Mitochondrial fusion and fission in mammals. Annu Rev Cell Dev Biol22:79–99.

40. Frieden M, et al. (2004) Ca(2+) homeostasis during mitochondrial fragmentation andperinuclear clustering induced by hFis1. J Biol Chem 279:22704–22714.

41. Han XJ, et al. (2008) CaM kinase I alpha-induced phosphorylation of Drp1 regulatesmitochondrial morphology. J Cell Biol 182:573–585.

42. Karbowski M, et al. (2002) Spatial and temporal association of Bax with mitochondrialfission sites, Drp1, and Mfn2 during apoptosis. J Cell Biol 159:931–938.

43. Autret A, Martin SJ (2009) Emerging role for members of the Bcl-2 family inmitochondrial morphogenesis. Mol Cell 36:355–363.

44. Meixenberger K, et al. (2010) Listeria monocytogenes-infected human peripheralblood mononuclear cells produce IL-1beta, depending on listeriolysin O and NLRP3.J Immunol 184:922–930.

45. TranVan Nhieu G, Clair C, Grompone G, Sansonetti P (2004) Calcium signalling duringcell interactions with bacterial pathogens. Biol Cell 96:93–101.

46. Dong Z, Saikumar P, Weinberg JM, Venkatachalam MA (2006) Calcium in cell injuryand death. Annu Rev Pathol 1:405–434.

47. Bischofberger M, Gonzalez MR, van der Goot FG (2009) Membrane injury by pore-forming proteins. Curr Opin Cell Biol 21:589–595.

48. Babiychuk EB, Monastyrskaya K, Potez S, Draeger A (2009) Intracellular Ca(2+)operates a switch between repair and lysis of streptolysin O-perforated cells. CellDeath Differ 16:1126–1134.

49. De Vos KJ, Allan VJ, Grierson AJ, Sheetz MP (2005) Mitochondrial function and actinregulate dynamin-related protein 1-dependent mitochondrial fission. Curr Biol 15:678–683.

50. Szabadkai G, et al. (2004) Drp-1-dependent division of the mitochondrial networkblocks intraorganellar Ca2+ waves and protects against Ca2+-mediated apoptosis.Mol Cell 16:59–68.

51. Ayres JS, Freitag N, Schneider DS (2008) Identification of Drosophila mutants alteringdefense of and endurance to Listeria monocytogenes infection. Genetics 178:1807–1815.

52. Francione L, et al. (2009) Legionella pneumophila multiplication is enhanced bychronic AMPK signalling in mitochondrially diseased Dictyostelium cells. Dis ModelMech 2:479–489.

53. Soulat D, Bauch A, Stockinger S, Superti-Furga G, Decker T (2006) Cytoplasmic Listeriamonocytogenes stimulates IFN-beta synthesis without requiring the adapter proteinMAVS. FEBS Lett 580:2341–2346.

Stavru et al. PNAS | March 1, 2011 | vol. 108 | no. 9 | 3617

CELL

BIOLO

GY