Bacterial Invasion: The Paradigms of Enteroinvasive Pathogens
Pascale Cossart, et al.
Science 304, 242 (2004);
DOI: 10.1126/science.1090124
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.
Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.
The following resources related to this article are available online at
www.sciencemag.org (this infomation is current as of March 7, 2012 ):
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/304/5668/242.full.html
A list of selected additional articles on the Science Web sites related to this article can be
found at:
http://www.sciencemag.org/content/304/5668/242.full.html#related
This article has been cited by 332 article(s) on the ISI Web of Science
This article has been cited by 100 articles hosted by HighWire Press; see:
http://www.sciencemag.org/content/304/5668/242.full.html#related-urls
This article appears in the following subject collections:
Microbiology
http://www.sciencemag.org/cgi/collection/microbio
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2004 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
107. T. C. Pierson, R. W. Doms, Curr. Top. Microbiol. White, Annu. Rev. Cell Dev. Biol. 12, 627 (1996). grants from the Swiss National Science Foundation,
Immunol. 281, 1 (2003). the Swiss Federal Institute of Technology (SEP), and
110. We thank J. Heuser, J. Kartenbeck, and N. Pante for
108. F. Reggiori, H. R. Pelham, Nature Cell Biol. 4, 117 (2002). the European Union (Euro Gene Drug) (A.H.) and by
electron micrographs, andA. Tagawa andL. Ellgaard
109. L. D. Hernandez, L. R. Hoffman, T. G. Wolfsberg, J. M. the Human Frontiers Science Program (A.E.S.).
for critical reading of the manuscript. Supported by
REVI EW
Bacterial Invasion: The Paradigms of
Enteroinvasive Pathogens
Pascale Cossart1* and Philippe J. Sansonetti2*
Invasive bacteria actively induce their own uptake by phagocytosis in normally of a macropinocytic pocket, loosely bound to
nonphagocytic cells and then either establish a protected niche within which they the bacterial body.
survive and replicate, or disseminate from cell to cell by means of an actin-based
The Zipper Mechanism of Entry
motility process. The mechanisms underlying bacterial entry, phagosome matura-
tion, and dissemination reveal common strategies as well as unique tactics evolved Yersinia pseudotuberculosis and Listeria
by individual species to establish infection. monocytogenes both harness transmembrane
cell-adhesion proteins as receptors for entry
To establish and maintain a successful infec- How the cell senses the bacterial intruders into mammalian cells (Figs. 1A and 2A).
tion, microbial pathogens have evolved a va- and adjusts its transcription and translation Entry can be divided into three successive
riety of strategies to invade the host, avoid or programs to its new life with a parasite is steps: (i) Contact and adherence. This step is
resist the innate immune response, damage an important issue. Apoptosis and anti- independent of the actin cytoskeleton and
the cells, and multiply in specific and nor- apoptosis, as well as cell cycle and inflam- involves only the bacterial ligand and its
mally sterile regions. Based on their capacity mation-related signaling pathways, are repro- receptor. It leads to receptor clustering. (ii)
to deal with these critical issues, bacteria can grammed after infection to help the cell Phagocytic cup formation. This step is trig-
be grouped in different categories. Here we to survive the stress induced by the infection. gered by the transient signals occurring
review the so-called invasive bacteria, i.e., The success of an infection depends on the after formation of the first ligand-receptor
bacteria that are able to induce their own messages that the two players the bacterium complexes and propagating around the in-
phagocytosis into cells that are normally and the cell send to each other. At each step of vading microbe. These signals induce actin
nonphagocytic. We focus on the tactics used the infectious process, the bacterium exploits the polymerization and membrane extension.
by enteroinvasive bacteria to trigger their host cell machinery to its own profit. (iii) Phagocytic cup closure and retraction,
uptake by epithelial cells and discuss their and actin depolymerization.
Entry Mechanisms
intracellular life-styles. The mechanisms of The Yersinia outer-membrane protein in-
entry and life-styles of other intracellular patho- To enter nonphagocytic cells such as intesti- vasin binds to integrin receptors that have the
gens have been reviewed elsewhere (1 4). nal epithelial cells, some microbial pathogens 1 chain and are normally implicated in ad-
During phagocytosis by phagocytes, express a surface protein able to bind eukary- herence of cells to the extracellular matrix
bacteria play a passive role. In contrast, otic surface receptors often involved in cell- (6). Invasin does not possess the RGD motif
during bacterial-induced phagocytosis, the matrix or cell-cell adherence. Expression of present in fibronectin, but both proteins inter-
bac- this protein leads to the formation of a vacu- act with integrins by a structurally similar
terium is the key and active player in the ole that engulfs the bacterium through a  zip- domain. Invasin has a higher affinity for in-
complex interplay between the invading pering process in which relatively modest tegrins and can oligomerize, inducing inte-
microbe and the host cell (5). Another im- cytoskeletal rearrangements and membrane grin clustering and efficient downstream sig-
portant component is the cytoskeleton, extensions occur in response to engagement naling. The cytoplasmic tail of the 1 chain,
whose plasticity is critical and optimally of the receptor. The initial interactions be- which normally interacts with the cytoskele-
exploited. After internalization, some bac- tween the bacterial protein and its receptor ton in focal complexes of adhesion plaques,
teria remain in a vacuole, in which they trigger a cascade of signals, including protein is critical for entry, but surprisingly, alter-
replicate. They prevent the normal matura- phosphorylations and/or recruitment of adap- ations of this domain that impair interaction
tion and trafficking of the phagosome and tors and effectors, and activation of cytoskel- with the cytoskeleton increase internaliza-
impair its normal bacteriolytic activities. eton components that culminate in phagocyt- tion. Thus, a lower affinity of the integrin for
Other bacteria escape from the vacuole and ic cup closure and bacterial internalization. the cytoskeleton could allow higher mobility
replicate in the cytosol. In some cases, they Other pathogens have devised mechanisms to of the receptors in the membrane.
also move and disseminate by means of an bind a protein that can itself act as a bridge Activation of integrins leads to tyrosine-
actin-based motility process. between the bacterium and a transmembrane phosphorylation events required for entry.
receptor, which then mediates the entry pro- The tyrosine kinase FAK (focal adhesion ki-
1
Unit� des Interactions Bact�ries-Cellules, INSERM cess. Finally, pathogens can also bypass the nase) is the most attractive candidate for
2
Unit� 604, Unit� de Pathog�nie Microbienne Mo-
first step of adhesion and interact directly transmitting a signal from clustered integrins
leculaire, INSERM Unit� 389, D�partement de Biologie
with the cellular machinery that regulates the to the cytoskeleton, because the 1-chain cy-
Cellulaire et Infection, Institut Pasteur, 28 Rue du
actin cytoskeleton dynamics by injecting ef- toplasmic domain binds to FAK, and domi-
Docteur Roux, Paris 75015, France.
fectors through a dedicated secretory system. nant-inhibitory mutations in FAK strongly
*To whom correspondence should be addressed. E-
The effector molecules cause massive cy- impair invasin-mediated uptake (7). Src,
mail: [e-mail] (P.C.); [e-mail]
(P.J.S.) toskeletal changes that trigger the formation phosphoinositide 3-kinase (PI 3-kinase), and
242 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org
S
PECIAL
S
ECTION
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
Rac are also involved in invasin-mediated junctions of polarized epithelial cells. The LRR LRR family of proteins and is only loosely
uptake. Why there is a requirement for phos- domain surrounds the first ectodomain of E- attached by its C-terminal repeats to the
phoinositide 3-kinase is unknown. Efficient cadherin (13). This weak-affinity interaction bacterial surface, where it interacts with
entry involves a Rac1-Arp2/3 pathway which cannot take place if proline-16 is changed into lipotechoic acids. Soluble InlB can reasso-
may involve N-WASP (8 10). The local glutamic acid, as in murine E cadherin (14). ciate with the bacterial surface of an InlB
concentration of phosphatidylinositol 4,5- Formation and maintenance of adherens junc- mutant and promote entry.
bisphosphate [(PIP2, PI(4,5)P2] is critical for tions require the integrity of the E-cadherin cy- InlB interacts with three cellular ligands (12,
entry, and Arf6 may play a role in activa- toplasmic domain that binds catenins ( , , and 18). The most relevant one is Met, a transmem-
tion of phosphoinositol-4-phosphate-5-kinase p120 catenins), which interact with the cell actin brane receptor tyrosine kinase that upon interac-
(PIP5 kinase) and control of cytoskeleton re- cytoskeleton (15). Similarly, entry of Listeria tion with its normal ligand, the hepatocyte
arrangements and membrane traffic involved into cells requires the terminal 35 amino acids of growth factor (HGF), dimerizes and elicits phos-
in closure of the phagocytic cup (11). E-cadherin. The latter binds to -catenin, which phorylation on two critical residues that act as
Several surface proteins contribute to entry recruits -catenin, which in turn interacts with docking sites to recruit signaling and adaptor
of L. monocytogenes into nonphagocytic cells in actin. Actin polymerization during internalin- molecules (20). Met binding to the concave sur-
vitro (12). The best-characterized protein, in- mediated entry is Rac dependent and mediated face of the InlB LRRs also leads to its transient
ternalin (InlA), is a surface protein that is co- by Arp2/3, but how Arp2/3 is activated is un- phosphorylation and to the recruitment and phos-
valently anchored to the cell wall and belongs to known (16). Entry also requires an unconven- phorylation of the adaptor proteins Cbl, Gab1,
a large family of leucine-rich repeat (LRR) pro- tional myosin, myosinVIIa, and its ligand veza- and Shc, and activation of PI 3-kinase with the
teins. As for invasin, coating of latex beads with tin (17). These two proteins probably play a role generation of PIP3 at the plasma membrane (21).
internalin promotes their entry, thus facilitating in the dynamics of the phagocytic cup. How the Optimal activity of Met requires the presence of
dissection of the specific pathway. Entry of Lis- tension generated by the myosin motor is cou- glycosaminoglycans (GAGs) on the cell surface,
teria into cells involves interaction between the pled to actin polymerization required for entry probably promoting oligomerization of the
LRR region of internalin and the first ectodo- has not been established. growth factor and/or its protection from extra-
main of human E-cadherin, a transmembrane The second well-characterized L. mono- cellular proteases. GAGs also increase Listeria
glycoprotein normally involved in homophilic cytogenes invasion protein is InlB (12, 18, InlB-dependent entry into the target cell. Heparin
E-cadherin E-cadherin interactions at adherens 19). This surface protein belongs to the can detach InlB from the bacterial surface, rein-
Fig. 1. Mechanisms used by bacteria to enter cells. (A) The zipper mechanism used by Yersinia and Listeria. (B) The trigger mechanism used by
Salmonella and Shigella.
www.sciencemag.org SCIENCE VOL 304 9 APRIL 2004 243
S
P E C I A L
S
E C T I ON
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
with lipid rafts (27),
allows entry of extra-
cellular calcium and
stimulates entry (28).
Even in the absence
of LLO, both interna-
lin- and InlB-mediat-
ed entry are depen-
dent on the presence
of raft microdomains,
suggesting that for en-
try, Listeria take ad-
vantage of raft mi-
crodomains, which
are known to be en-
riched in receptors
and signaling mole-
cules. Interestingly,
cholesterol depletion
does not affect the in-
ternalin- and InlB-
mediated pathways at
the same step of the
entry process (29).
The Trigger
Mechanism of
Entry
Both Shigella and
Salmonella use this
mechanism to enter
the cell (Fig. 1B and
Fig. 2. The zipper and the trigger mechanisms. (A): Zipper mechanism. From left to right: x-ray structure of internalin interacting
Fig. 2B). Contact be-
with E-cadherin [reprinted from (13) with permission from Elsevier]; scanning electron micrograph of Listeria entering into Caco2
tween bacteria and
cells; immunofluorescence images of Listeria entering into Vero cells (red: Met; green, actin; and blue: bacteria). (B) Trigger
cells is mediated by
mechanism. From left to right: Reconstitution of the TTSS; scanning electron micrograph of Shigella entering into cells;
immunofluorescence images of Shigella entering into Caco 2 cells (red: cortactin; green: actin; and blue: bacteria). (C) InlB-mediated the type III secretory
ruffling. Control cells and cells ruffling upon incubation with soluble InlB (green: actin). (D) Shigella entering into Src dominant-
system (TTSS) (Fig.
negative cells (red: cortactin; green: actin; and blue: bacteria). Src-dependent tyrosine phosphorylation of cortactin is essential to
1). The TTSS allows
trigger massive extension of actin filaments at a distance from the entry focus; thus, cells expressing a Src dominant-negative
direct activation of
construct form inefficient entry foci with limited actin polymerization tightly around the entry vacuole.
components of the cy-
toskeleton by delivery
forcing the hypothesis that InlB may act as a new free ends for polymerization by severing actin of dedicated bacterial effectors. In Salmonella, the
soluble protein. Thus, InlB mimics HGF, the filaments. In the initial steps of cell entry, cofilin TTSS is encoded by a chromosomal patho-
normal Met ligand, and similarly to growth fac- activity is modulated by LIM kinase. Then progres- genicity island (SPI-1) and in Shigella by a
tors, soluble InlB induces actin-rich membrane sive accumulation of cofilin on filaments favors plasmid-located pathogenicity island (PAI).
ruffles (Fig. 2C). filament disassembly and retraction of the phago- These PAIs encode the structural components
InlB also interacts with gC1qR/p32, a cytic cup. Thus, the InlB-Met interactions probably of the TTSS and some of their dedicated
ubiquitous protein first identified as the elicit both a Rac-WAVE-ARP2/3 and a Rac-PAK- effectors. Two of these components (i.e.,
receptor for the globular part of the comple- LIM-kinase-cofilin cascade. It is still unknown SipB/C in Salmonella, IpaB/C in Shigella)
ment component C1q (22). However, the sub- how Rac is activated downstream of Met. The role form a pore, or translocator, that delivers the
cellular location and function of gC1qR re- of PI 3-kinase is also unknown. The working hy- effectors into the cell cytoplasm, creating a
main controversial, and its role in cell entry pothesis is that, as in phagocytosis, PI 3-kinase continuum between the bacterial and eukary-
remains to be clarified. facilitates cup closure, probably by recruiting mem- otic cytoplasms (30, 31).
Contact between Met and InlB, present on brane vesicles and actin regulators. It may also The interaction of bacteria with their epithe-
the bacterium or released from its surface, ini- induce sustained activation of Rac. lial cell target occurs in four successive stages:
tiates actin nucleation and polymerization via the InlB is thus a strong signaling protein that by 1) A pre-interaction stage. At 37�C, the ef-
small guanosine triphosphatase (GTPase) Rac, itself acts as an invasin but may also potentiate fector molecules stored in the bacterial cyto-
WAVE, and the Arp2/3 complex (19, 23). Actin other bacterial factors involved in Listeria entry plasm are associated with dedicated chaperones,
filament elongation, which provides the driving and tissue tropism, such as internalin. Other pro- whose major role is to avoid premature associa-
force for membrane extension around the bacte- teins such as the autolysins Ami, Auto, and ActA tion of the effector molecules and their proteo-
rium, involves VASP, which may act as an contribute to Listeria adherence and entry (24). lytic degradation (32). In exponentially growing
anticapping protein at the barbed ends. Cofilin also In addition, listeriolysin O (LLO), a pore-form- bacteria, the TTSSs are properly assembled, but
participates in this process. This protein increases ing, cholesterol-dependent cytolysin involved the secretion of effector proteins is repressed
actin turnover by triggering actin depolymerization mainly in escape from the internalization vacu- until the bacterium establishes contact with its
at pointed ends of actin filaments and by creating ole (25, 26) and that, like other toxins, interacts cell target.
244 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org
S
PECIAL
S
ECTION
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
2) An interaction stage. This stage encom- dylinositol phosphatase (39), stimulates actin re- strategies (4) aimed at surviving in a hostile and
passes complex events leading to the formation arrangements and mediates bacterial entry, changing environment characterized by poor nu-
of a signaling platform. A recognition event is whereas SipA binds and stabilizes actin fila- trient content, progressive decrease of the pH,
likely to take place at the tip of the TTSS, ments (40). Shigella has evolved a similar pro- and delivery of antibacterial peptides and lyso-
activating the secretory process via a retroactive cess of boosting cytoskeletal rearrangements, al- somal enzymes as late endosomes mature to
signaling, possibly involving an adenosine though through different molecular mechanisms. lysosomes. In macrophages, these conditions are
triphosphatase in the TTSS basal body (33). In The C-terminal domain of IpaC is central to the even more drastic and exacerbated by the deliv-
Shigella, the high-affinity binding of IpaB to activation of Cdc42 and Rac-1, which is quickly ery of reactive oxygen and nitrogen intermedi-
CD44 the hyaluronic acid receptor that is followed by activation of the tyrosine kinase ates. Two major strategies can be recognized,
strongly expressed on the basolateral membrane c-src upon contact with IpaC (41), recruitment of although a given species may use a combination
of intestinal epithelial cells and on the surface of cortactin to the membrane upon its c-src medi- of both: (i) Bacteria may adapt to and eventually
many other cell types, including cells of myeloid ated tyrosine phosphorylation, and further mas- resist these hostile conditions, thus developing a
lineage may be a key step in achieving tran- sive actin polymerization in the vicinity of the state of metabolic adaptation to the stress im-
sient adherence to the cell surface, activation of original actin cup via the Arp2/3 complex (42) posed by these conditions; (ii) alternatively, bac-
the secretory machinery, and insertion of the (Fig. 2C). This process is amplified by IpgD, a teria may alter the biogenesis and dynamics of
IpaB/C translocon into the eukaryotic cell mem- Shigella homolog of SopB/SigD. IpgD expresses their vacuolar compartment, thus creating for
brane. Consistent with the association of CD44 a phospatidylinositol phosphatase activity that themselves a less hostile niche that is permissive
with cholesterol and sphingolipid-rich mem- hydrolyzes PI(4,5)P2 into PI(5)P [phosphatidyl- for their survival and growth. Salmonella repre-
brane rafts, this step of the interaction is depen- inositol 5-phosphate], thus disconnecting the ac- sents a paradigm of the complex combination of
dent on intact rafts (34). Cholesterol extraction tin subcortical cytoskeleton from the membrane these two survival and growth strategies (Fig. 3).
disrupts binding to and entry into epithelial cells, and favoring actin dynamics at the entry site After a few hours of invasion, bacteria reside in
and IpaB and CD44 segregate in these rafts. (43). The Abl family of tyrosine kinases is also an atypical acidic compartment called the SCV
Similarly, in Salmonella, the protein components involved in Shigella entry through phosphoryl- (Salmonella containing vacuole), which is nei-
of the SipB/C translocon also segregate in rafts. ation of the adaptor molecule Crk (44). ther a late nor an early endosome (48). How
The initial interaction may take place in these 4) Actin depolymerization and closing of bacteria redirect the fate of this compartment
membrane subdomains because (i) the targeted the macropinocytic pocket. This final stage is away from the normal phagosomal pathway in-
receptor is enriched in rafts; (ii) the lipid com- similar in Shigella and Salmonella, despite volves transient acquisition of rab5, PI3-kinase,
position of rafts is optimal for the formation of important differences between the effectors EEA1, and finally rab7 (49). In addition, merg-
the pore and translocon, in a way similar to the involved and the molecular mechanisms ex- ing of the SCV with the endoplasmic reticulum
cholesterol dependence of several hemolysins ploited. In the case of Salmonella, SptP, a appears to contribute to early SCV maturation
(27); and (iii) these domains are enriched in TTSS-secreted protein, has two activities: (i) (50) and membranes of the trans-Golgi network
signaling molecules such as tyrosine kinases of a tyrosine-phosphatase activity that regulates surround the SCV at late times of infection (51),
the src family. activity of the mitogen-activated protein ki- suggesting interactions with both the endocytic
3) The formation of a macropinocytic pocket. nase (MAPK) induced by entry; and (ii) a and the biosynthetic pathway. Numerous bacte-
This stage involves localized but massive rear- GAP (GTPase-activating protein) activity on rial genes are required for survival and replica-
rangements of the cell surface, characterized by Cdc42 and Rac that antagonizes the activity tion. A key role is played by the SPI2 effector
the formation of intricate filopodial and lamelli- of SopE, thus leading to shrinking of the SifA a protein required for the formation of
podial structures that appear similar in Salmonel- entry focus by blocking further actin poly- Sifs, filaments enriched in lysosomal glycopro-
la and Shigella. Rearrangements of the actin merization (45). It may seem strange that teins (Lgps), and extensions of the SCV, in
cytoskeleton largely account for the formation of proteins of opposite functions are injected epithelial cells (52). The function of SifA may be
the entry focus. At the early stage of Shigella simultaneously into the target cell. Recent to mediate the recruitment of vesicles and in-
entry, VirA, a plasmid-encoded protein secreted evidence indicates that, despite equivalent crease the SCV membrane surface area to ac-
through the TTSS, induces local destabilization amounts delivered by the TTSS, SopE is commodate replicating bacterial cells.
of the microtubules that results in their depoly- rapidly degraded through a proteasome-
Life in the Cytosol and Actin-Based
merization (35). The latter affects the early dependent pathway, whereas SptP is more
Intra- and Intercellular Motility
events of actin rearrangement through the deac- stable (46). In the case of Shigella, IpaA, a
tivation of RhoA, leading to Rac1 activation and TTSS-secreted protein, binds the N-terminal Some intracellular pathogens able to induce
formation of Rac1-IRSp53-WAVE2 complex head domain of vinculin, a key protein in the their own phagocytosis into epithelial cells
that recruits Arp2/3. IpaC in Shigella (36) and formation of cell-adherence plaques, and in- escape from the internalization vacuole, rep-
SipC in Salmonella (37) initiate actin nucle- duces actin depolymerization (47). licate in the cytosol, and move by recruiting
ation through their C-terminal domain, which and polymerizing actin (53) (Fig. 4). Actin
Intracellular Life-Styles
is exposed to the cytoplasm of the eukaryotic polymerization at one pole of the bacterium
cell, via the IpaB/C or SipB/C pore. The mech- After internalization, bacteria remain in a provides the energy for movement and en-
anism of initial actin nucleation, however, re- vacuole or escape to the cytosol, where they ables the bacteria to reach the plasma mem-
mains uncertain. SipC can nucleate actin alone replicate. Some intracytosolic bacteria may brane, where they form protrusions that are
in vitro (37), but IpaC requires activation of also move by a process of polarized actin endocytosed by neigboring cells, allowing
Cdc42 and Rac 1 (36). polymerization that takes place at one pole of the formation of a two-membrane vacuole,
Massive extension of the actin filaments that the bacterium and provides the force for bac- cell to cell spread, and tissue dissemination.
form entry foci seems to respond to different terial locomotion inside the cytosol and to- For Listeria, escape from the vacuole is me-
mechanisms in Salmonella and Shigella. In Sal- ward neighboring cells. diated by a pore-forming toxin called listerioly-
monella, the translocated SopE proteins (SopE1 sin O (LLO), a potent signaling molecule that
The Vacuole as an Intracellular
and SopE2) act as exchange factors for the activates nuclear factor B (NF- B) and a vari-
Replication Compartment
Cdc42 and Rac-1 GTPases, thus massively ety of other pathways (25). Intracytosolic repli-
boosting the initial nucleation event (38). More- Bacteria that replicate inside the internalization cation requires expression of a sugar-uptake
over, SopB/SigD, a TTSS-secreted phosphati- vacuole have developed an impressive array of system, which is absent in the nonpathogenic
www.sciencemag.org SCIENCE VOL 304 9 APRIL 2004 245
S
P E C I A L
S
E C T I ON
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
that manipulate cell apoptotic processes. Three
major pathways have so far been identified: (i)
Intracellular Shigella and Salmonella, respec-
tively, secrete IpaB and SipB through their
TTSS. These two proteins activate the pro-apop-
totic cysteine protease caspase-1, which causes
apoptotic death of infected macrophages while
also initiating an inflammatory response through
processing or maturation of two potent pro-in-
flammatory cytokines, interleukin-1 (IL-1 )
and IL-18 (62, 63). (ii) Yersinia translocate plas-
mid-encoded Yop proteins, one of which, YopP/
YopJ, binds to and neutralizes the activity of a
MAPK kinase, thereby blocking the activation of
NF- B, an essential system supporting cell sur-
vival (64). (iii) The third pathway, although not
yet clearly described in enteroinvasive bacteria,
is worth mentioning. Upon interaction of Neis-
seria gonorrhoeae with epithelial cells, the se-
creted protein PorB causes Ca2 fluxes that ac-
tivate caspases, and consequently cell apoptosis
(65). PorB creates mitochondrial pores, thus in-
ducing apoptosis through the release of cyto-
Fig. 3. Intracellular life-styles. Schematic representation of the Salmonella-containing vacuole (see
chrome c. Finally, epithelial cells infected by
text). Listeria and Shigella lyse the vacuole and move in the cytosol by an actin-based motility
process mediated by ActA or IcsA/VirG, which interact with Arp2/3 or N-WASP and Arp2/3,
Shigella undergo activation of their connexin-
respectively. EE: early endosome; LE: late endosome; Ly: lysosome; ER: endoplasmic reticulum
constituted hexameric hemichannels. The infect-
ed cells release ATP, which acts as a paracrine
species L. innocua (25). Actin recruitment by vitro and is recruited on the McRettsial sur- mediator activating Ca2 fluxes in neighboring
Listeria and polymerization are triggered by the face was unexpected, providing a new tool cells, thus increasing their competence for bac-
surface protein ActA, which recruits and acti- to address Arp2/3 regulation. terial invasion and cell-to-cell spread (66).
vates the seven-protein Arp2/3 complex, hence
Cell Responses to Intracellular Pathogens
generating a dendritic network of branched actin
filaments (54). Modulation and control of actin- In addition to the transient posttranslational
based movements involve several other proteins: modifications occurring upon entry, intracel-
(i) cofilin; (ii) capping protein, which caps the lular bacteria induce drastic changes in the
barbed ends of actin filaments; (iii) profilin, pattern of transcription and translation of in-
which binds to monomeric actin and, in complex fected cells. This is particularly true for in-
with actin, to actin-filament barbed ends, hence testinal epithelial cells that, upon invasion by
providing actin monomers to growing barbed Salmonella or Shigella, behave as sentinels
ends; (iv) -actinin, which cross-links actin fil- by inducing a transcriptional program whose
aments; and (v) VASP, which binds to ActA and major function is to up-regulate innate im-
F-actin and modulates branch density and move- mune defense mechanisms (58). This pro-
ment. Shigella, after escaping from the vacuole gram occurs largely in response to the induc-
upon the action of IpaB, expresses on its surface tion of NF- B that regulates a large portion
an outer-membrane protein called IcsA/VirG. of the pro-inflammatory genes. The pro-
This protein, which is unrelated to ActA, recruits inflammatory program of epithelial cells in
the cellular protein called N-WASP (55, 56). contrast to the outside-in signaling pathway
Cellular N-WASP is functionally and structural- that Toll-like receptors mediate in phagocytic
ly related to bacterial ActA and can recruit and cells, in the presence of bacterial PAMPs
activate the Arp2/3 complex, highlighting how (pathogen-associated molecular patterns)
bacteria may either mimic or recruit mammalian appears to be mediated by an intracellular
proteins to harness eukaryotic pathways (5). sensing system involving cytosolic proteins
Even though Rickettsia is not an enteroinva- of the Nod family (59). Nod1 is prevalent in
sive microorganism, it is worth mentioning that intestinal epithelial cells and shows specific
after its escape into the cytoplasm, it forms actin recognition for muropeptides originating
tails made of long, unbranched actin fila- from the peptidoglycan of Gram-negative mi-
ments, which differ from those generated by croorganisms (60, 61). Another cytosolic pro-
ActA or IcsA/N-WASP (Fig. 4). Similar to tein, Nod2, recognizes peptidoglycans from
proteins of the WASP family, the bacterial any bacterial species, essentially because it is
surface protein involved, RickA (57), is able to recognize muramyl-dipeptide, a struc-
composed of three regions, with a central ture common to all peptidoglycans.
proline-rich region and a C-terminal part Through their capacity to regulate gene tran-
Fig. 4. Actin-based motility of Listeria, Rickettsia,
that recruits Arp2/3. Because Arp2/3 gener- scription and by other pathways, intracellular
and Shigella. Electron micrographs of actin tails
ates a network of branched actin filaments, bacteria can take over the fate of their host cell.
labeled with fragment S1 of myosin (69) [reprint-
the discovery that RickA activates Arp2/3 in Among the most striking paradigms are bacteria ed with permission from Journal of Cell Science].
246 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org
S
PECIAL
S
ECTION
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
Fig. 5. The invasive strategies of enteroinvasive pathogens. Intestinal is an alternative to the Spi1-dependent apoptotic killing of macrophages
epithelial cells (IECs) maintain a physical barrier against commensal flora, similar to that caused by Shigella. Having crossed the epithelial barrier
although specializedsites such as the follicle-associatedepithelium (FAE) and circumvented the threat of phagocytosis, the bacterial species
allow constant sampling of the luminal flora through M cells. Translo- considered here proceed along different pathways. L. monocytogenes
cated bacteria thus exposed to macrophages, dendritic cells (DCs), and B disseminate systemically, possibly inside circulating monocytes and DCs.
lymphocytes are captured, killed, processed, and presented to the im- Yersinia may invade IECs through their basolateral pole, a process medi-
mune system. Invasive pathogens take advantage of this route to cross ated by invasin; they also cause local and mesenteric abscesses in local
the epithelial barrier. Once translocated, bacteria must survive attack by and loco-regional lymphoid structures. Shigella proceeds to TTSS/Ipa-
macrophages. The four bacterial species consideredhave solvedthis issue dependent entry into epithelial cells followed by escape into the cyto-
differently: L. monocytogenes are phagocytosed but escape into the plasm, intracellular motility, andcell-to-cell spread, thus establishing the
cytoplasm, and thus avoid being killed in lysosomal compartments. infectious process at the mucosal level, without extensive systemic
Yersinia adopt an antiphagocytic strategy by intracellular injection of dissemination. Salmonella may, like Shigella, enter IECs through their
YopE, H, and T that inactivate the actin cytoskeleton. In addition, they basolateral pole in a T TSS/Sop-dependent manner. Alternative routes of
adopt an anti-inflammatory strategy, with YopP/J blocking tumor necro- invasion involve IECs directly, away from the FAE. In particular, invasion
sis factor production, which prevents further local recruitment of by L. monocytogenes is mediatedby internalin (InlA) andpossibly InlB. In
predators such as monocytes and polymorphonuclear leukocytes. Alter- addition, Salmonella are able to dislocate the brush border cytoskeleton
natively, phagocytosed Yersinia may cause YopP/J-dependent apoptosis and cause an apical entry ruffle. Shigella and Yersinia seem unable to
of their host cell. Shigella not only cause apoptosis of macrophages and disrupt the epithelial barrier from a luminal position unless massive
monocytes, thus ensuring their own survival, but also trigger early inocula are used. A third process of translocation may involve DCs
mucosal inflammation through the release of mature IL-1 and IL-18, crawling between IECs or sending pseudopods to capture luminal bacte-
which disrupts epithelial impermeability and facilitates bacterial spread ria and retract in a subepithelial position. Salmonella are able to trans-
at a distance. Finally, Salmonella remodel their phagosomes, thus avoid- locate in this way, possibly followedby systemic diffusion of Salmonella-
ing its transition to a lysosome and creating an intracellular niche loaded DCs. It is not yet clear whether this type of translocation occurs
that allows their efficient replication; this Spi2-dependent process in the other invasive species.
www.sciencemag.org SCIENCE VOL 304 9 APRIL 2004 247
S
P E C I A L
S
E C T I ON
Downloaded from www.sciencemag.org on March 7, 2012
C E L L UL AR I NVAS I ONS
Bacterial Invasion: In Vivo Veritas References and Notes
36. G. Tran Van Nhieu, E. Caron, A. Hall, P. J. Sansonetti,
1. C. R. Roy, L. G. Tilney, J. Cell Biol. 158, 415 (2002).
EMBO J. 18, 3249 (1999).
2. K. A. Fields, T. Hackstadt, Annu. Rev. Cell Dev. Biol.
37. R. D. Hayward, V. Koronakis, EMBO J. 18, 4926 (1999).
A major issue is to validate, in vivo, the molec-
18, 221 (2002).
38. W. D. Hardt, L. M. Chen, K. E. Schuebel, X. R. Bustelo,
ular and cellular events analyzed in vitro. If one
3. C. R. Hauck, T. F. Meyer, Curr. Opin. Microbiol. 6, 43
J. E. Galan, Cell 93, 815 (1998).
(2003).
focuses on invasion of the intestinal barrier, it is
39. F. A. Norris, M. P. Wilson, T. S. Wallis, E. E. Galyov,
4. C. C. Scott, R. J. Botelho, S. Grinstein, J. Membr. Biol.
P. W. Majerus, Proc. Natl. Acad. Sci. U.S.A. 95, 14057
clear that L. monocytogenes, Shigella, Salmonel-
193, 137 (2003).
(1998).
la, and Yersinia, despite their shared capacity to
5. B. B. Finlay, P. Cossart, Science 276, 718 (1997).
40. D. Zhou, M. S. Mooseker, J. E. Galan, Proc. Natl. Acad.
invade epithelial cells in vitro, differ with regard 6. R. R. Isberg, P. Barnes, J. Cell Sci. 114, 21 (2001).
Sci. U.S.A. 96, 10176 (1999).
7. M. A. Alrutz, R. R. Isberg, Proc. Natl. Acad. Sci. U.S.A.
41. J. Mounier, P. J. Sansonetti, G. Tran Van Nhieu, in
to (i) the capacity to disrupt, invade, and even-
95, 13658 (1998).
preparation.
tually cause the inflammatory destruction of the
8. M. A. Alrutz et al., Mol. Microbiol. 42, 689 (2001).
42. L. BougnŁres et al., in preparation.
epithelium; and (ii) the possibility of proceeding
9. K. McGee, M. Zettl, M. Way, M. Fallman, FEBS Lett.
43. K. Niebuhr et al., EMBO J. 21, 5069 (2002).
509, 59 (2001).
44. E. A. Burton, R. Plattner, A. M. Pendergast, EMBO J.
to systemic dissemination and possibly coloni-
10. K. McGee, P. Holmfeldt, M. Fallman, FEBS Lett. 533,
22, 5471 (2003).
zation of organs at a distance.
35 (2003).
45. C. E. Stebbins, J. E. Galan, Mol. Cell 6, 1449 (2000).
A major handicap to studying the respective
11. K. W. Wong, R. R. Isberg, J. Exp. Med. 198, 603 (2003).
46. T. Kubori, J. E. Galan, Cell 115, 333 (2003).
12. P. Cossart, J. Pizarro-Cerda, M. Lecuit, Trends Cell Biol.
invasive phenotypes in vivo has been the lack 47. R. Bourdet-Sicard et al., EMBO J. 18, 5853 (1999).
13, 23 (2003).
48. D. W. Holden, Traffic 3, 161 (2002).
of a mouse model simulating the intestinal and
13. W. D. Schubert et al., Cell 111, 825 (2002).
49. S. Meresse, O. Steele-Mortimer, B. B. Finlay, J. P.
systemic diseases observed in humans (67).
14. M. Lecuit et al., EMBO J. 18, 3956 (1999).
Gorvel, EMBO J. 18, 4394 (1999).
This was particularly the case for L. monocyto- 15. M. Lecuit et al., Proc. Natl. Acad. Sci. U.S.A. 97,
50. O. Steele-Mortimer, S. Meresse, J. P. Gorvel, B. H. Toh,
10008 (2000).
B. B. Finlay, Cell. Microbiol. 1, 33 (1999).
genes, until a transgenic mouse line expressing
16. S. Sousa, M. Lecuit, P. Cossart, in preparation.
51. S. P. Salcedo, D. W. Holden, EMBO J. 22, 5003 (2003).
the human E-cadherin receptor of internalin
17. S. Sousa et al., J. Cell Sci., in press.
52. F. Garcia-del Portillo, M. B. Zwick, K. Y. Leung, B. B.
became available, thus unlocking the transintes- 18. H. Bierne, P. Cossart, J. Cell Sci. 115, 3357 (2002).
Finlay, Proc. Natl. Acad. Sci. U.S.A. 90, 10544 (1993).
19. H. Bierne et al., in preparation.
53. F. Frischknecht, M. Way, Trends Cell Biol. 11, 30 (2001).
tinal route for this pathogen, i.e., via invasion of
20. Y. Shen, M. Naujokas, M. Park, K. Ireton, Cell 103, 501
54. M. D. Welch et al., Science 281, 105 (1998).
enterocytes (68). A relevant animal model has
(2000).
55. T. Suzuki, H. Miki, T. Takenawa, C. Sasakawa, EMBO J.
yet to be found for Shigella because, unlike
21. K. Ireton et al., Science 274, 780 (1996).
17, 2767 (1998).
56. C. Egile et al., J. Cell Biol. 146, 1319 (1999).
infected humans, mice do not undergo exten- 22. L. Braun, B. Ghebrehiwet, P. Cossart, EMBO J. 19,
1458 (2000).
57. E. Gouin et al., Nature 427, 457 (2004).
sive invasion and inflammatory destruction
23. H. Bierne et al., J. Cell Biol. 155, 101 (2001).
58. M. F. Kagnoff, in Microbial Pathogenesis and the
of their rectal and colonic mucosae. Despite
24. O. Dussurget, J. Pizarro-Cerda, P. Cossart, Annu. Rev. Intestinal Epithelial Cell, G. A. Hecht, Ed. (American
Microbiol., in press.
these limitations, a picture is emerging (Fig. Society for Microbiology, Washington, DC, 2003).
25. J. A. Vazquez-Boland et al., Clin. Microbiol. Rev. 14, 59. S. E. Girardin et al., EMBO Rep. 2, 736 (2001).
5) concerning the various strategies used by
584 (2001). 60. S. E. Girardin et al., Science 300, 1584 (2003).
these pathogens.
26. A. L. Decatur, D. A. Portnoy, Science 290, 992 (2000). 61. M. Chamaillard et al., Nature Immunol. 4, 702 (2003).
In conclusion, although current work aims 27. F. G. Van der Goot, T. Harder, Semin. Immunol. 13, 62. H. Hilbi et al., J. Biol. Chem. 273, 32895 (1998).
89 (2001). 63. D. Hersh et al., Proc. Natl. Acad. Sci. U.S.A. 96, 2396
to elucidate the in vivo relevance of the now
28. S. Dramsi, P. Cossart, Infect. Immun. 71, 3614 (2003). (1999).
well-understood mechanisms used by inva-
29. S. Seveau, S. Giroux, H. Bierne, P. Cossart, in 64. K. Orth et al., Science 285, 1920 (1999).
sive bacteria in vitro, future efforts should
preparation. 65. A. Muller et al., EMBO J. 18, 339 (1999).
30. J. E. Galan, Annu. Rev. Cell Dev. Biol. 17, 53 (2001). 66. G. Tran Van Nhieu et al., Nature Cell Biol. 5, 720 (2003).
focus on understanding both bacterial and
31. P. J. Sansonetti, FEMS Microbiol. Rev. 25, 3 (2001). 67. M. Lecuit, P. Cossart, Trends Mol. Med. 8, 537 (2002).
host cell transcription and translation pro-
32. C. Parsot, C. Hamiaux, A. L. Page, Curr. Opin. Micro- 68. M. Lecuit et al., Science 292, 1722 (2001).
grams during infection, in various cells and
biol. 6, 7 (2003). 69. E. Gouin et al., J. Cell Sci. 112, 1697 (1999).
33. A. Blocker, K. Komoriya, S. Aizawa, Proc. Natl. Acad. 70. We thank E. Veiga, E. Gouin, and O. Dussurget for
tissues. This information should provide
Sci. U.S.A. 100, 3027 (2003). help with manuscript preparation and H. Bierne, L.
vital clues in the ongoing battle against
34. F. Lafont, G. Tran Van Nhieu, K. Hanada, P. San- BougnŁres, and G. Tran van Nhieu for generously pro-
bacterial disease and for elaborating new
sonetti, F. G. van der Goot, EMBO J. 21, 4449 (2002). viding unpublished figures. The authors are internation-
therapeutic strategies. 35. S. Yoshida et al., EMBO J. 21, 2923 (2002). al Scholars of the Howard Hughes Medical Institute.
REVI EW
Intracellular Parasite Invasion Strategies
L. D. Sibley
Intracellular parasites use various strategies to invade cells and to subvert cellular detect and destroy foreign objects. Overcom-
signaling pathways and, thus, to gain a foothold against host defenses. Efficient cell ing these defenses and breaching the final
entry, ability to exploit intracellular niches, and persistence make these parasites barrier imposed by the cell membrane is a
treacherous pathogens. Most intracellular parasites gain entry via host-mediated formidable challenge. By entering into the
processes, but apicomplexans use a system of adhesion-based motility called  glid- confines of a host cell, the parasite assures
ing to actively penetrate host cells. Actin polymerization dependent motility itself of both a ready source of nutrients and
facilitates parasite migration across cellular barriers, enables dissemination within a potential means to avoid immune clearance.
tissues, and powers invasion of host cells. Efficient invasion has brought widespread Parasites that practice this life-style have typ-
success to this group, which includes Toxoplasma, Plasmodium, and Cryptosporidium. ically given up the capacity for extracellular
growth, which leaves them vulnerable if en-
Parasites exist in virtually every conceivable many of which are responsible for lethal and try is impeded. Defining how parasites gain
niche, but none is so specialized as that of the debilitating diseases in animals and humans. entry into their host cells is thus important for
obligate intracellular parasite, which must Our defenses present an array of barriers to rational design of improved therapies. Para-
gain entry into the cells of its host to survive. infection, including skin, mucosa, connective sites are among the earliest branching eu-
Most intracellular parasites are protozoans, tissue, and an active surveillance system to karyotes (1); their study expands our knowl-
248 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org
S
PECIAL
S
ECTION
Downloaded from www.sciencemag.org on March 7, 2012