Studies on the mechanisms of maturation and membrane disassembly of vacuoles containing invasive bacteria

STUDIES ON THE MECHANISMS OF MATURATION AND MEMBRANE DISASSEMBLY OF VACUOLES CONTAINING INVASIVE BACTERIA JOSÉ CARLOS VIEIRA DOS SANTOS Tese de doutoramento em Biologia Básica e Aplicada 2015 JOSÉ CARLOS VIEIRA DOS SANTOS STUDIES ON THE MECHANISMS OF MATURATION AND MEMBRANE DISASSEMBLY OF VACUOLES CONTAINING INVASIVE BACTERIA Tese de Candidatura ao grau de Doutor em Biologia Básica e Aplicada submetida ao Instituto de Ciências Biomédicas Abel Salazar da Universidade do Porto. Orientador – Jost Enninga Categoria – Director de Investigação Afiliação – Institut Pasteur, Paris, França Co-orientador – Sandra Sousa Categoria – Investigador auxiliar Afiliação – Instituto de Biologia Molecular e Celular da Universidade do Porto, Porto, Portugal 2 3 “Ó mar salgado, quanto do teu sal São lágrimas de Portugal! Por te cruzarmos, quantas mães choraram, Quantos filhos em vão rezaram! Quantas noivas ficaram por casar Para que fosses nosso, ó mar! Valeu a pena? Tudo vale a pena Se a alma não é pequena. Quem quer passar além do Bojador Tem que passar além da dor. Deus ao mar o perigo e o abismo deu, Mas nele é que espelhou o céu.” – Fernando Pessoa; in Mensagem 4 5 Acknowledgments Serendipity often plays an important role in science. One of my most fortunate contacts with serendipity was meeting Jost for the first time. I am extremely grateful for that event and now, after making the balance of these last years, I could not be happier about choosing Jost as my PhD mentor. I had the chance to work in a fantastic and vibrant environment, for which Jost always was (and still is) the driving force. Our scientific project was difficult, bold, some people even told us we were a bit crazy, but Jost always trusted me and gave me the motivation to move forward and not to give up or give in. No words will ever suffice or convey how thankful I am. Anyway, I really want to thank Jost for teaching me so many important things. Things that I believe will help me during my scientific career; things that made me a better scientist; things that helped me in so many circumstances… Jost have always given me intellectual freedom to follow my scientific instincts, not to be afraid of failing, and to take the risk. All this with a perfect balance of proximity and independence, which I think was crucial to make these last years a personal and scientific success. In the end, it was fun, a lot of fun. I would also like to thank Jost for all the support in any kind of other situations: for helping me settling down in Paris, for all the personal advices, and for all the support in the preparation of my next scientific steps. I will always be grateful to Jost, who was an excellent mentor and a good friend. Over these last years at Pasteur, I met fantastic people in the DIHP lab. First I would like to acknowledge the ones that have already left. To the Portuguese crew, Cristina and Mónica: they helped me a lot when I arrived at the lab. It was great to have them around and to keep in mind that, as a Portuguese, we really love our country. I would also like to thank Ricardo: even if he was not from our lab, he was as part of the family (also because he is Cristina’s husband). It was great to be around him: a great friend always with a funny story, and with whom I learned a lot as well. Many thanks to Alex, whose help was precious when I needed to change apartment, and with whom was great to share conversations about football. I would like to thank Soudeh, with whom I worked closely together, and also Juliane, for all their important advices. It was great to have met and shared the lab with Marianne, Charlotte and Geneviève, and to have Andrew around, for his one-year sabbatical experience. 6 Many thanks to Anne-Marie and to Laurence for all the help taking care of the bureaucracies: without you, the lab would be a total chaos. I am also thankful to all the current members of the DIHP family. To Patricia, Célia and Laura for their feedback. To Camille and Virginie, for being so friendly and for sharing your knowledge with me. I would like to thank Laurent for all his support with everything: the lab would not be the same without his “MacGyverisms”. To Nora, many thanks. It was great to share many moments with her, to be desk neighbors (for the first years) and also lab bench neighbors. I would like to thank Jennifer for all the help during my project, for teaching me a lot of things and for all the linguistic help with English. She is a great friend with whom I really learned a lot. Last but not least, special thanks to Allon: a great friend, with whom I had the pleasure to share not only the office, but also cultural and scientific knowledge. I really learned a lot with Allon. I learned how to be a better scientist and also to appreciate the potential and the beauty of three-dimensional electron microscopy. I can say that meeting Allon changed my scientific career. All these people are fantastic, and my words are not enough. I would like to thank Paul Lazarow for sharing crucial knowledge about subcellular fractionation, and Genèvieve Milon and Javier Pizarro-Cerdá for the support during my last year of PhD. I want to thank Magalie, Mariette, Véronique and Julia for the amazing collaboration: I learned a lot from all of them, and proteomics can be really fun. Many thanks to Adeline, Perrine and Christine Schmitt for the help with electron microscopy. I would also like to thank to all the people that I met at Pasteur and to everyone that shared reagents with me. Nothing of this would be possible without the GABBA doctoral program, a very special setting for young scientists that gave me my PhD fellowship through the FCT. In particular, thanks to my co-supervisor Sandra Sousa, one of the persons that showed me for the first time the amazing world of cellular microbiology. And, of course, thanks to GABBA13 (Bruno, Clara, João, “Mega John”, Joana, Mariana, Sara, Susana and Olga) for all the great moments during my first year of PhD. Very importantly, I would like to thank my mother, my father and my brother, for all the support and good advices. It’s not easy to be away from them. Finally, I am extremely grateful to my girlfriend Mariana, with whom I shared amazing experiences during our time together in Paris. My PhD would not be possible without her, whose contribution goes far beyond what I can acknowledge here. 7 Table of contents i. List of abbreviations 9 ii. List of figures 12 iii. List of tables 13 iv. Abstract 14 v. Resumo 15 PART I – INTRODUCTION 19 1 Small GTPases as Master Regulators of Cellular Functions 21 A. The Ras superfamily of small GTPases: activity and regulation 21 B. Rho GTPases 25 Cdc42 27 Rac proteins 27 RhoA 28 C. Rab GTPases 28 2 Cellular Communication With the Outside Environment 31 A. Plasma membrane and the cytoskeleton as interface 31 Actin filaments 32 Microtubules 35 Intermediate filaments 36 Septins 36 B. Endocytosis and phagocytosis 38 Clathrin-mediated endocytosis 39 Clathrin-independent endocytosis 41 Macropinocytosis 42 Phagocytosis 43 3 Intracellular Organelle Communication 45 A. Regulators of vesicular membrane trafficking 45 Rab GTPases regulate vesicle budding and uncoating 45 Rab GTPases regulate vesicle motility 46 Rab GTPases regulate vesicle fusion 47 SNAREs 47 B. Endosome and phagosome maturation 49 The early endosome network 49 Formation and maturation of late endosomes 51 Endosome fusion with lysosomes 53 Phagosome maturation 56 14 iv. Abstract Eukaryotic cells receive diverse extracellular stimuli that lead to the generation of intracellular responses. To regulate this, cells have evolved ways to partition off different functions to various intracellular compartments, called organelles. This compartmentalization requires the existence of communication routes between organelles, which is mostly done via vesicular trafficking that transfer biological content and information in a precise and regulated manner. Invasive bacterial pathogens, such as Salmonella and Shigella, are able to manipulate many of these communication routes, in order to invade and survive inside eukaryotic epithelial cells. Within the host cell, Salmonella resides inside a membrane-bound vacuole, called Salmonella-containing vacuole (SCV), which undergoes several steps of maturation that can either result in the establishment of a bacterial replicative niche or in vacuolar membrane rupture. Similarly, Shigella is engulfed within a membrane-bound vacuole, but this is rapidly ruptured, with subsequent bacterial escape into the host cytosol. In both cases, the formation, maturation and rupture of the bacteria-containing vacuoles are crucial steps for bacterial survival. In order to obtain a better understanding of these steps, at a molecular level, we investigated the dynamic association of several host factors with the bacteriacontaining vacuoles, through the use of real time fluorescence microscopy. Our results show that several host small GTPases and kinases are dynamically recruited to the sites of Shigella entry and vacuolar rupture, and that these processes follow precise kinetics. Moreover, we isolated and determined the protein composition of bacteria-containing vacuoles, namely the SCV, which revealed hundreds of proteins associated with this compartment. By combining three-dimensional fluorescence and electron microscopies, we demonstrated that the SCV interacts with several host cell organelles, namely with endoplasmic reticulum-derived COPII vesicles and VAMP7-positive lysosome-like vesicles. The dynamic communication between the SCV and different host organelles modulates vacuolar maturation and rupture, affecting intracellular Salmonella localization and growth. Together, our results provide new insights into the mechanisms of maturation and membrane rupture of vacuoles containing invasive bacteria. 15 v. Resumo Células eucarióticas recebem diversos estímulos extracelulares que despoletam repostas intracelulares. Para regular estes processos, as células evoluíram de forma a dividir diferentes funções por vários compartimentos intracelulares, denominados organelos. Esta compartimentalização requer a existência de vias de comunicação entre organelos, o que é feito sobretudo através do tráfego de vesículas que transferem conteúdos biológicos de uma maneira precisa e regulada. Bactérias patogénicas intracelulares, como Salmonella e Shigella, têm a capacidade de manipular muitas destas vias de comunicação, de forma a invadirem e sobreviverem no interior de células epiteliais eucarióticas. Dentro da célula hospedeira, Salmonella reside no interior de um vacúolo membranar, designado SCV, que é sujeito a diversos passos de maturação que tanto podem resultar no estabelecimento de um nicho bacteriano replicativo como na ruptura da membrana do vacúolo. De forma semelhante, Shigella é internalizada num vacúolo membranar, mas este rapidamente sofre lise, com subsequente passagem da bactéria para o citosol do hospedeiro. Em ambos os casos, a formação, maturação e ruptura de vacúolos que contêm bactérias são passos cruciais para a sobrevivência bacteriana. De modo a melhor compreender estes passos, a um nível molecular, investigámos a presença de proteínas da célula hospedeira em vacúolos que contêm bactérias, através do uso de microscopia de fluorescência em tempo real. Os resultados aqui apresentados mostram que, durante infecção por Shigella, GTPases e proteínas cinases são recrutadas de forma dinâmica para os locais de entrada da bactéria e para locais de ruptura do vacúolo, e que estes processos seguem cinéticas precisas. Para além disto, isolámos e determinámos a composição proteica de vacúolos que contêm bactérias, nomeadamente do SCV, o que evidenciou a presença de centenas de proteínas neste compartimento. Através da combinação entre microscopias de fluorescência e electrónica tridimensionais, demonstrámos também que o SCV interage com vários organelos da célula hospedeira, particularmente com vesículas COPII derivadas do retículo endoplasmático e com vesículas lisossomais positivas para VAMP7. Esta comunicação dinâmica entre o SCV e diferentes organelos celulares controla a maturação e lise do vacúolo, 16 afectando a localização intracelular de Salmonella e o crescimento bacteriano. Na sua globalidade, estes resultados fornecem novas pistas acerca dos mecanismos de maturação e ruptura membranares de vacúolos que contêm bactérias intracelulares. 17 “… the cell is not just an inert playground for a few almighty masterminding molecules, but is a system, a hierarchically ordered system, of mutually interdependent species of molecules, molecular groupings, and supramolecular entities; and that life, through cell life, depends on the order of their interactions.” – Paul Alfred Weiss 18 ! 19 PART I INTRODUCTION 20 21 Small GTPases as Master Regulators of Cellular Functions ukaryotic cells, individually or as part of multicellular organisms, receive a multitude of extracellular stimuli and respond to their environment through the generation of intracellular responses. The extracellular stimuli alter the affected cell, both by inducing changes at the cell surface and by modifying the communication between intracellular organelles. Communication relies on the compartmentalization within eukaryotic cells and is dependent on several steps, such as budding of vesicles from donor membranes that are transported to specific acceptor membranes, followed by their docking and fusion, which leads to content mixing. These dynamic processes in the cell are regulated by several proteins, of which small GTPases play a crucial role in ensuring the correct crosstalk between the different cellular compartments. A. The Ras superfamily of small GTPases: activity and regulation The Ras superfamily of small GTPases is a group of low molecular weight (2030 kDa) proteins composed of more than 150 members with highly conserved structural and biochemical properties. Due to their capacity to hydrolyze guanosine-5’-triphosphate (GTP) to guanosine-5’-diphosphate (GDP), Ras GTPases can act as molecular switches that convert extracellular cues to intracellular signaling pathways, thus regulating numerous basic cellular processes, such as actin cytoskeleton regulation, adhesion, migration, phagocytosis, endocytosis, differentiation, polarization, morphogenesis and cell survival (Cherfils and Zeghouf, 2013; Goitre et al., 2014; Jaffe and Hall, 2005; Stenmark, 2009). Depending on the sequence and functional similarities, the Ras superfamily can be divided into five major families (Ras, Rho, Rab, Arf and Ran) and subfamilies (Table 1). Small GTPases are localized to distinct membrane microdomains, alternating cyclically between a GDP-bound “off” and a GTP-bound E Chapter!1! 22 “on” form, in response to different stimuli. Generally, GDP is tightly bound to the GTPases (Bos et al., 2007; Wittinghofer et al., 1997), therefore these proteins require the catalytic helping of guanine nucleotide exchange factors (GEFs), which facilitate GDP-to-GTP exchange (Cherfils and Zeghouf, 2013; Jaffe and Hall, 2005). Since GTP is highly abundant in the cytosol (~1 mM), it quickly binds as soon as GDP has been released. On the other hand, GTP-to-GDP conversion, and signaling termination, is driven by GTPase activating proteins (GAPs) given that GTP hydrolysis is very slow (Bos et al., 2007). Thus, the entire molecular switch is composed of a small GTPase, a GEF and GAP, altogether (Figure 1). Table 1 – List of human proteins of the Ras superfamily of small GTPases. Adapted from Goitre et al., 2014. Family Subfamily Function Members - Ras Cell proliferation, differentiation, survival, and apoptosis. Gene expression E-Ras; N-Ras; H-Ras; K-Ras; R-Ras; R-Ras-2; DiRas-1; Di-Ras-2; Di-Ras-3; NKIRas-1; NKIRas-2; Ras-D1; Ras-D2; Ras-L10A; Ras-L10B; Ras-L11A; Ras-L11B; Ras-L12, Rerg - Ral GTP-dependent exocytosis Ral-A; Ral-B Ras - Rap Cell adhesion Rap-1A; Rap-1B; Rap-2A; Rap-2B; Rap-2C - Rad Cell shape. Cell cycle checkpoint Rad; Gem; Kir; Rem-1; Rem-2 - Rheb mTOR pathway. Cell growth and cell-cycle progression Rheb; Rheb-L1 - Rit Neuronal differentiation and survival Rit-1; Rit-2; Rin; Ric Rho Cytoskeletal dynamics. Cell shape, polarity, adhesion and movement. Cell cycle progression. Gene expression. RhoA; RhoB; RhoBTB1; RhoBTB2; RhoBTB3; RhoC; RhoD; RhoF; RhoG; RhoH; RhoJ; RhoQ; RhoU; RhoV; Rnd1; Rnd2; Rnd3; Rac1; Rac2; Rac3; Cdc42 Rab Protein and membrane vesicle trafficking in the endocytic and secretory pathways. Rab1a; Rab1b; Rab2; Rab3a; Rab3b; Rab3c; Rab3d; Rab4a; Rab4b; Rab5a; Rab5b; Rab5c; Rab6a; Rab6b; Rab6c; Rab7a; Rab7b; Rab7L1; Rab8a; Rab8b; Rab9a; Rab9b; Rab10; Rab11a; Rab11b; Rab12; Rab13; Rab14; Rab15; Rab17; Rab18; Rab19; Rab20; Rab21; Rab22a; Rab23; Rab24; Rab25; Rab26; Rab27a; Rab27b; Rab28; Rab30; Rab31; Rab32; Rab33a; Rab33b; Rab34; Rab35, Rab36; Rab37 Arf Vesicle trafficking. Endocytosis and exocytosis. Arf1; Arf3; Arf4; Arf5; Arf6; Arl1; Arl2; Arl3; Arl4; Arl5; Arl5C; Arl6; Arl7; Arl8; Arl9; Arl10A; Arl10B; Arl10C; Arl11; Arl13A; Arl13B; Arl14; Arl15; Arl16; Arl17; TRIM23, Arl4D; ArfRP1; Arl13B Ran Nuclear transport. Mitotic spindle organization. Ran 23 Figure 1 – The small GTPase switch and its complex regulation. All small GTPases are activated by GDP/GTP exchange stimulated by GEFs (blue) and inactivated by GTP hydrolysis by GAPs (magenta). The individual GAP and GEF proteins acting in the regulatory switch (among several possible GAPs and GEFs) are depicted in dark colors. In addition, many GTPase families combine their GDP/GTP switch with a cytosol/membrane alternation, which is regulated by GDIs (beige). Targeting of the GTPase-GDI complex to specific membranes is mediated by interaction with a membrane-bound GDF (brown). Activation or termination signals act over a specific GEF or GAP, promoting conformational changes that result in their activation. Moreover, feedback loops (green broken arrow) from the active GTPase can control the GEF efficiency, and feed-forward signaling (blue broken arrow) emanating from the GEF contributes to the selection of the specific effector that will be recruited by the active GTPase. Feed-forward signaling from the selected effector (black broken arrow) can also provide an additional level of regulation over a specific GAP. Adapted from Cherfils and Zeghouf, 2013. Most small GTPases are modified at their C-terminus by the addition of a farnesyl or geranylgeranyl group, which act as lipid tails that anchor and localize these proteins to membranes. Guanine dissociation inhibitors (GDIs) play an important role in the regulation of GTPase activity and GTP-to-GDP conversion. GDIs remove certain small GTPases (mostly Rho and Rab proteins) from membranes and, by masking their lipid tails, establish cytosolic soluble complexes (Figure 1). In this way, GDIs prevent small GTPases from associating with membranes and thus inhibit interaction with regulators or effectors (Bos et al., GTP Effector GDP/GTP exchange GTP hydrolysis GDP GDI GAPGEF GDF GEFs Ac�va�onsignals Autoinhibi�on release GAPs Autoinhibi�on release Termina�onsignals Feed-forward signal Feedback loop Feed-forward signal Effectors 30 31 Cellular Communication With the Outside Environment he various environmental signals acting over eukaryotic cells induce the generation of selective cellular responses. These are governed by the activation of different proteins embedded in the plasma membrane, which induce dynamic remodeling of the cell cytoskeleton and generation of an intracellular signal transduction. A. Plasma membrane and the cytoskeleton as interface The plasma membrane of eukaryotic cells, together with all associated protein receptors, acts as an interface that separates the intracellular contents from its outside environment, and regulates what enters and exits the cell. Additionally, is also serves as a base of attachment for the cytoskeleton, thus helping in the maintenance of the cell shape. Despite etymological similarities with the body skeleton, the cytoskeleton is functionally very different, displaying a highly dynamic structure that is continuously reorganized in response to cellular requirements. Thus, the cytoskeleton plays major cellular functions such as spatial organization of the intracellular contents, physical and biochemical connection of the cell to the extracellular milieu and generation of coordinated forces that enable the cell to move and change shape. Most of these processes are accompanied by deformation of the cell membrane, such as the formation of filopodia and lamellipodia, endocytosis, phagocytosis, membrane trafficking events and intracellular movement of pathogenic bacteria. There are three main cytoskeletal building blocks (Figure 5): actin microfilaments, microtubules (MTs) and intermediate filaments. However, recently it has been proposed that septins constitute an additional component of the cytoskeleton (Chang and Goldman, 2004; Fletcher and Mullins, 2010; Mostowy and Cossart, 2012). T Chapter!2! 32 Figure 5 – Schematic representation of the main cellular cytoskeletal components: actin, microtubules, intermediate filaments and septins. Septins from different groups are depicted in different colors, and they interact through their GTP-binding domain (the G interface) and their N-terminal and Cterminal regions (the NC interface). Adapted from Mostowy and Cossart, 2012. Actin filaments Actin is one of the most abundant and highly conserved proteins in eukaryotic cells. This 42 kDa protein is able to bind ATP (adenosine triphosphate), and in its monomeric form is known as globular actin (G-actin), which can undergo cycles of self-polymerization into filamentous actin (F-actin), a process that is accompanied by ATP hydrolysis. Long polymerized chains of actin filaments are organized in a helix structure with a diameter of ~7 nm, and are highly dynamic branched networks. These can regulate a wide variety of cellular processes requiring generation of forces, such as phagocytosis, endocytosis, cell junction assembly, membrane ruffling and lamellipodia dynamics, and filopodia and stress fiber formation. F-actin filaments are polar, containing dynamic barbed plus ends (where monomers preferentially assemble), and less active pointed minus ends (where monomers preferentially disassemble). Moreover, filament turnover is 33 controlled by different actin-binding proteins, which can sequester or deliver actin monomers, or promote filament nucleation, capping, elongation, severing or depolymerization. Actin assembly is initiated by the generation of free plus ends that act as templates for polymerization. Nevertheless, spontaneous assembly of actin monomers into polymers (known as nucleation) is kinetically unfavorable and inefficient. Therefore, cells utilize several actin-binding proteins that directly nucleate actin filament formation (Figure 6), such as the Arp2/3 complex, formins and WASP homology 2 (WH2) domain-containing nucleators (Campellone and Welch, 2010; Mostowy and Cossart, 2012; Rottner et al., 2010; Rotty et al., 2012). Figure 6 – Schematic representation of different paths of actin polymerization. Adapted from Goley and Welch, 2006. The Arp2/3 complex is one the most important actin nucleators. This is a 220 kDa protein complex composed of seven stably associated polypeptides that include Arp2, Arp3 and five additional subunits, ArpC1-ArpC5. The Arp2/3 complex binds to the side of existing acting filaments and initiates the branching off and elongation of new filaments at a ~70º Y-branch angle (Figure 6). By itself,                  34 the Arp2/3 complex is not an efficient actin nucleator, but when bound to actin filaments its activity increases. Other factors contribute to increased nucleation and branching activity, such as Arp2 phosphorylation and the binding of nucleation-promoting factors (NPFs), such as the N-WASP factor (Campellone and Welch, 2010; Rotty et al., 2012). Signal transduction pathways initiated at the plasma membrane that include Cdc42 activation can result in activation of NWASP, through a conformational change, then leading to Arp2/3-mediated actin polymerization (Tomasevic et al., 2007). Interestingly, N-WASP function seems to be important for several processes such as plasma membrane ruffling and invagination, and endocytosis, a process in which F-actin facilitates membrane fission (Legg et al., 2007; Tsujita et al., 2006). Another group of NPFs are the WAVE proteins that, when activated by Rac GTPases, can activate the Arp2/3 complex and induce lamellipodia formation (Takenawa and Suetsugu, 2007). Other NPFs include the WASH (WASP and SCAR homologue), WHAMM (WASP homolog associated with actin, membranes and microtubules) and JMY (junctionmediating and regulatory protein) proteins, as well as cortactin, each exerting specific functions that extend the repertoire of Arp2/3 complex activation (Campellone and Welch, 2010; Rottner et al., 2010). Other actin nucleators, unlike Arp2/3, induce formation of unbranched actin filaments. One of the best characterized examples are the formins, such as DRF proteins. Formins bind to barbed ends and enable actin filament elongation by preventing other capping proteins from terminating this process (Figure 6). Additionally, formins help determining the length of F-actin filaments. More specifically, DRFs are activated by different Rho GTPases and participate in the formation of stress fibers, filopodia, lamellipodia and phagocytic cups (Campellone and Welch, 2010). Finally, a new class of actin nucleating proteins has been described: the WH2 domain-containing nucleators. Factors from this family include Spire, cordon-bleu and leiomodin proteins. These proteins induce the tethering of three or more G-actin monomers in either a single linear F-actin multimeric strand or a short trimer, thereby promoting actin nucleation at the barbed ends (Baum and Kunda, 2005; Qualmann and Kessels, 2009). Dissociation of Y-branches can be regulated by coronins and cofilin, which inhibit the Arp2/3 complex and promote actin depolymerization and recycling (Rotty et al., 2012; Sit and Manser, 2011). The tightly orchestrated equilibrium 35 between nucleation, branching and turnover allows the efficient formation of actin networks in the eukaryotic cell. Microtubules MTs are long, filamentous, tube-shaped protein polymers with a diameter of ~25 nm and that can be many micrometers long, composed of α-tubulin and βtubulin heterodimers. MTs are essential in all eukaryotic cells and play crucial roles in the development and maintenance of cell shape, in the transport of vesicles within the cell, in cell signaling, and in cell division and mitosis. Like actin filaments, MTs are polarized polymers, with a structurally asymmetrical molecular organization comprised of a plus end and a minus end. The polymerization of MTs is driven by a nucleation-elongation mechanism, is highly dynamic and is regulated both spatially and temporally (Jordan and Wilson, 2004; Mostowy and Cossart, 2012). These filaments are in a constant state of non-equilibrium dynamics that can occur via two different mechanisms. One kind of dynamic behavior is known as dynamic instability and involves the addition and loss of tubulin subunits at the plus end of MTs, which is driven mainly by GTP hydrolysis. GTP-bound tubulin subunits are incorporated into a MT plus end, creating a stable GTP cap that prevent filament depolymerization. However, GTP is rapidly hydrolyzed in the β-tubulin subunit, the MT becomes unstable and depolymerizes due to the adoption of a outward curved conformation, leading to further destabilization of the filament. When GDP is substituted by GTP in the disassembled tubulin subunits, the cycle can restart. Thus, MTs constantly switch between phases of growth (known as rescue) and shortening (known as catastrophe), which happens more rapidly and more extensively in the plus end than in the minus end. The other mechanism of MT dynamic behavior, known as treadmilling, occurs through the gain of tubulin at the plus end of MTs (growing) and loss of tubulin at the minus end of the filaments (shortening), together with an intrinsic flow of tubulin subunits (Ferreira et al., 2014; Jordan and Wilson, 2004). Among other factors, MT behavior is controlled through the binding of regulatory proteins that include MT-associated proteins (MAPs), such as the dynein and kinesis motor proteins, and dynactin-1. Many of these MAPs are able to recognize only the MT plus end, and are involved in the regulation of dynamic 36 instability parameters, therefore controlling MT activity (Akhmanova and Hoogenraad, 2005; Jordan and Wilson, 2004). Intermediate filaments Intermediate filaments are a group of around 70 proteins that, unlike actin and tubulin, do not form seeds to which individual subunits grow. Instead, intermediate filaments are built by the assembly of individual proteins that form a tetrameric subunit composed of two antiparallel coiled-coil dimers. Then, eight tetrameric subunits associate with each other laterally, forming a unit length filament (ULF), which then joins end-to-end with other ULFs to form short filaments, in a process that only takes a few seconds. Short filaments then grow into longer filaments through the longitudinal annealing with other ULFs (Herrmann et al., 2007; Mostowy and Cossart, 2012). Intermediate filaments have a diameter of ~ 11 nm and are not polarized. They are the least stiff of all the types of cytoskeletal polymers, so they are able to resist tensile forces much more effectively than compressive forces. Moreover, intermediate filaments can crosslink to actin filaments and MTs, through the action of plectin proteins, which regulates several cellular functions (Fletcher and Mullins, 2010). Within the cell, there is a system of cytoplasmic intermediate filaments that connects intercellular junctional complexes at the plasma membrane with the outer nuclear membrane, thus helping in the stabilization of the cell shape. One example is intermediate filament protein vimentin. Additionally, there is another intermediate filament system that is attached to the inner nuclear membrane, inside the nucleus, and that consists of lamins (Herrmann et al., 2007). Septins Septins comprise a family of highly conserved proteins in eukaryotes and were only recently recognized as a novel component of the cytoskeleton. Human cells have 13 septin genes that encode for over 30 protein isoforms (Hall et al., 2005; Mostowy and Cossart, 2012; Spiliotis, 2010). These are 30-65 kDa proteins structurally composed of a central core domain that can directly bind to phosphoinositides on the plasma membrane, a GTP-binding domain and 53 highly conserved amino acids of unknown function (Pan et al., 2007). All septins bind to GTP and most hydrolyze it to GDP, which appears to induce stable conformational 37 changes and therefore enable or prevent septin-septin interactions (Estey et al., 2011; Sirajuddin et al., 2007; 2009). In human cells, septins generally interact as hetero-oligomeric complexes of three different septin proteins, each of which is present in duplicate, therefore generating a hexameric structure. An example is the structure of the septin-2, -6 and -7 complex, which displays a linear non-polar hexamer with two copies of each septin symmetrically arranged in the order 7-6-22-6-7 (Mostowy and Cossart, 2012; Sirajuddin et al., 2007). Furthermore, recent work showed that both ends of this hexamer also bind septin-9, therefore forming an octamer (Kim et al., 2011). All septin complexes are able to form non-polar filaments as a result of end-to-end protein assembly, which results in structures that are more stable and less dynamic than other cytoskeletal elements such as actin microfilaments or MTs. Additionally, septin filaments can associate laterally, forming bundled filaments that ultimately tend to self-assemble into higher-ordered structures such as rings, both of which are considered to be the biological active forms of septins (Estey et al., 2011; Mostowy and Cossart, 2012). This plasticity results in a notable combinatorial diversity of septin complexes, which might provide them multiple roles that cannot be carried out by other cytoskeleton components. Septins have also been shown to interact with other components of the cytoskeleton such as actin, and seem to be involved in several cellular processes such as actin bundle formation and phagocytosis (Huang et al., 2008; Kinoshita et al., 2002). This suggests that septins could also associate with the Arp2/3 complex or with members of the WASP family to control actin polymerization (Mostowy and Cossart, 2012). Moreover, septins associate with MTs in different mammalian cells, where colocalization was observed with septin-2 and -9 (Nagata et al., 2003; Spiliotis et al., 2008). Therefore septins might contribute to MT-dependent movement of membranes, as well to the spatial coordination of cell motility and MT-actin crosstalk (Spiliotis, 2010). Functionally, septin filaments appear to act as macromolecular scaffolds that accumulate proteins and promote their interaction. Given that septins can interact with phospholipids on the plasma membrane (Bertin et al., 2010), they can regulate the distribution of membrane-bound proteins at this site, such as the receptor tyrosine kinase Met. Moreover, septins seem to have a role in clathrinmediated endocytosis (Mostowy and Cossart, 2009; Mostowy et al., 2011). It is 38 even possible that septin filaments might function as crosslinkers between the plasma membrane and actin (Saarikangas and Barral, 2011). Additionally, these septins scaffolds localize to sites of exocytosis, where they participate and regulate events of vesicle docking and fusion with the plasma membrane (Amin et al., 2008; Estey et al., 2011). The association of septin filaments with the plasma membrane is also involved in generating membrane diffusion barriers that compartmentalize membrane proteins to specific membrane domains. In such a way, septins can create barricades that isolate specific cellular domains thus preventing free diffusion of molecules in the membranes (Caudron and Barral, 2009; Estey et al., 2011). B. Endocytosis and phagocytosis Eukaryotic cells are able to take up macromolecules and particles from the surrounding medium through a process called endocytosis, coined for the first time in 1963, by Christian de Duve. In this process, signals at the plasma membrane trigger actin filaments to assemble locally, inducing membrane internalization and particle ingestion into an endocytic vesicle that pinches off into the cytoplasm. Given that the plasma membrane defines the boundary by which the cell communicates with the environment, its composition has to be tightly regulated by the cell in order to allow effective cellular response mechanisms. During endocytosis there is production of new intracellular membranous structures deriving from the plasma membrane. These structures contain membrane lipids, proteins and extracellular fluid. In this way, cells efficiently regulate the levels of surface receptors, as well as their interaction with the environment, which is crucial for the control of diverse intracellular cascades. Moreover, many pathogens exploit endocytosis to force their internalization into the host cell (described in more detail in chapter 4). Interestingly, there are different mechanisms and routes of endocytic uptake into the cell (Figure 7) (Doherty and McMahon, 2009). 39 Figure 7 – Schematic representation of different cellular endocytic pathways. Large particles are internalized by phagocytosis, whereas fluid internalization occurs by macropinocytosis. In both cases, there is the formation of considerably large intracellular vesicles. The majority of the endocytosed cargo is internalized to the early endosome, via clathrinor caveolin-coated vesicles that are derived from the plasma membrane. Cdc42-regulated endocytosis may lead to the internalization of cargoes that are delivered via clathrinand dynamin-independent carriers (CLICs) into a glycosyl phosphatidylinositol-anchored protein-enriched early endosomal compartment (GEEC), on the way to the early endosome. Adapted from Johannes et al., 2015 and Mayor and Pagano, 2007. Clathrin-mediated endocytosis Clathrin-mediated endocytosis is the most extensively studied and best understood process of endocytosis. As in any other endocytic pathway, it requires a budding structure from the plasma membrane, through a well-defined sequence of events. In mammalian cells, clathrin-mediated endocytosis involves the packaging of transmembrane receptors and their ligands into vesicles, called clathrin-coated vesicles (~80-100 nm in diameter), with the aid of adaptor proteins. This mechanism involves sequential steps of nucleation, cargo selection, coat assembly and invagination, and ultimately scission of the vesicle from the plasma membrane. The coat of the inward-budding clathrin-coated pit is formed by a triskelion shape composed of three clathrin heavy chains and three light chains, which confers structural support to the vesicle. Since clathrin does not bind directly to lipids or cargo, it requires additional accessory proteins and cargo adaptors that promote triskelion assembly, cargo sequestration and vesicle formation. One of the most abundant clathrin adaptors is the adaptor protein complex 2 (AP2), which acts as a bridge that links transmembrane cargo to the nucleating clathrin coat (Doherty and McMahon, 2009). Another important adaptor is the clathrin assembly                     46 vesicles. This is ensured by several factors such as the cargo itself, the lipid composition of a membrane and its curvature, and Rab GTPases. It was shown that Rab1 is involved in vesicle budding from the ER (Nuoffer et al., 1994) and that the late endosomal Rab9 facilitates sorting and recycle of mannose-6-phosphate receptors (M6PRs) from late endosomal recycling buds to the trans-Golgi network (TGN) (Carroll et al., 2001; Riederer et al., 1994). Rab5 was also described to have a role in cargo sequestration and vesicle formation, and it associates with early endosomes. This GTPase is a crucial factor for the assembly of clathrincoated pits at the plasma membrane and for clathrin-mediated endocytosis of transferrin receptors (McLauchlan et al., 1998). Thus it seems that at least some Rab GTPases influence vesicle budding from donor membranes, and they might also aid incorporation of cargo molecules into the nascent vesicles. Moreover, active Rab GTPase could act as a checkpoint that guarantees vesicle delivery to the correct target organelle (Stenmark, 2009). After their budding, vesicles are coated with molecular complexes that need to be shed to allow correct membrane engagement and fusion with the acceptor membrane. On clathrin-coated pits, for example, Rab5 activity helps in the uncoating of the adaptor protein complex AP2 from endocytic vesicles, which is important for proper membrane trafficking in epithelial cells (Semerdjieva et al., 2008). Rab GTPases regulate vesicle motility Actin filaments and MTs can function as tracks for molecular motors that carry vesicles between intracellular compartments. Rab proteins also regulate vesicle motility within the cell along the cytoskeletal filaments, which requires a high specificity in the binding of motor proteins to vesicles. In general, two main protein families associate with MTs to mediate trafficking: kinesins, which move intracellular material to the MT plus-end; and dyneins, which act as minus-enddirected MT motors (Chang and Goldman, 2004). It has been shown that active Rab6a (localized to the Golgi) associates with kinesin-1, kinesin-3 and kinesin-6, but also with the cytoplasmic dynein-1, thus regulating different trafficking steps in separate cellular contexts (Horgan and McCaffrey, 2011; Stenmark, 2009). Rab5 regulates both the attachment of early endosomes to, and the motility along, MTs (Nielsen et al., 1999). Rab11a regulates endosomal trafficking by associating with 47 kinesin-2, and can also control additional trafficking from sorting endosomes to recycling endosomes by binding to dynein-1. Rab27a and Rab27b associate with kinesin-1 to regulate axonal transport of certain vesicles. Moreover, Rab27a is able to mediate association of cargo vesicles to the actin motor myosin-Va, which is crucial to transport Rab27-positive vesicles to the cell periphery, in melanocytes (Stenmark, 2009). Finally, it was shown that the Rab7 effector Rab-interacting lysosomal protein (RILP) facilitates the dynein-mediated transport of late endosomes along MTs (Jordens et al., 2001). Rab GTPases regulate vesicle fusion Rab GTPases also cooperate with components of the vesicle docking and fusion machinery, such as the SNARE complexes. For example, the active GTPbound form of Rab5 (localized to early endosomes) is vital for homotypic early endosome fusion thus regulating endosome maturation (Stenmark et al., 1994). The abovementioned Rab27a also plays a role in vesicle fusion, since it controls docking of exocytic vesicles to the plasma membrane (Stenmark, 2009). SNAREs Membrane fusion is a fundamental process within eukaryotic cells. At a molecular level it is mostly regulated by SNARE proteins. In mammalian cells, at least 35 different SNAREs have been found, which act in the final step of docking of donor vesicles and their fusion with target compartments (Hong, 2005). Functionally, SNAREs were originally classified into v-SNAREs and t-SNAREs according to their donor vesicle (v) or target (t) membrane localization. Though, to avoid ambiguity in the case of homotypic membrane fusion, SNAREs have been reclassified as R-SNAREs (generally act as v-SNAREs) or Q-SNAREs (often act as t-SNAREs) (Chen and Scheller, 2001). Interaction between Rand Q-SNAREs leads to the formation of the trans-SNARE complex (or SNAREpin), in which one SNARE helix wraps around similar helixes on three other SNAREs (Figure 9A). In the center of the SNAREpin, one R-SNARE contributes with an arginine (R) residue, whereas three Q-SNAREs contribute with a glutamine (Q) residue each (Hong, 2005; Luzio et al., 2007). Many of the VAMP (vesicle-associated membrane protein) proteins belong to the group of R-SNAREs, whereas syntaxins and SNAPs (synaptosome-associated proteins) belong to the group of Q- 48 SNAREs. Most of these are found in specific subcellular compartments, indicating that they are selectively involved in particular intracellular trafficking steps (Chen and Scheller, 2001). For example syntaxin-1, syntaxin-2, syntaxin-4 and SNAP25 are found at the plasma membrane. Syntaxin-5 and VAMP4 localize to the Golgi, VAMP8 localizes to early and late endosomes, and VAMP7 localized to late endosomes and lysosomes (Jahn and Scheller, 2006). Figure 9 – Schematic representation of SNARE-mediated vesicle membrane fusion. Adapted from Jahn and Scheller, 2006. Mechanistically, it is not energetically favorable to fuse two membranes in an aqueous environment. The formation of the trans-SNARE complex exerts a mechanical force on membranes, which direct two membranes towards each other and creates membrane curvature and tension, directly causing fusion. After the formation of a tight trans-SNARE complex, apposing membranes start hemifusing, followed by distal leaflet membrane breakdown and opening of a fusion pore (Figure 9B). Subsequently, the fusion pore expands, resulting in content mixing and membrane relaxation (Chen and Scheller, 2001; Jahn and Scheller, 2006).               $" 6* 1           " #  '              49 B. Endosome and phagosome maturation The early endosome network Endocytosis and endosome maturation have been extensively studied, mostly by using mammalian tissue culture cells. Endosomes play a major role in controlling the reutilization or degradation of membrane components, thus regulating basic cellular processes, such as nutrient uptake, immunity, signaling, adhesion, and membrane turnover. Extracellular material that is endocytosed by several pathways (as described on chapter 2B) is delivered to a common organelle, the early endosome. From there, many cell membrane receptors and other proteins are recycled back to the plasma membrane (via recycling endosomes), whereas other molecules are directed towards the TGN or the lysosomes for degradation (Figure 10). In this way, the early endosomes functions as a key sorting station in the cell, ensuring that components that need to be reutilized are separated from the ones that need to be degraded (Huotari and Helenius, 2011; Scott et al., 2014). Figure 10 – The endosomal/lysosomal network. The early endosomes accumulate cargo and support recycling back to the plasma membrane, through fast and slow pathways, or retrograde transport to the TGN. The early endosome can also mature, by successive steps that result in the conversion into maturing endosomes and late endosomes. At the same time, the endosomes move to the perinuclear space along microtubules (MT). Endosomes undergo homotypic fusion reactions and grow in size, with the formation of intraluminal vesicles. Late endosomes fuse with lysosomes, generating a transient acidic hybrid organelle, the endolysosome, in which active degradative reactions take place. The endolysosome is then converted into a dense lysosome. Adapted from Huotari and Helenius, 2011. 50 The early endosome is the first endocytic membrane-bound vesicle to accept incoming cargo. This organelle is a highly dynamic structure, composed of regions of thin tubular extensions (~60 nm) and large vesicles (~ 400 nm) that have membrane invaginations and present a multi-vesicular appearance (Figure 11) (Gruenberg, 2001; Sachse et al., 2002). It is thought that the morphologically distinct early endosome sub-domains are important for different functions: the tubular membranes might be involved in recycling pathways, whereas the multi-vesicular sub-domain may be involved in sorting of cargo to the degradative pathways (Jovic et al., 2010). Typically, within minutes (~10 minutes) after endocytosis and early endosome formation, there is the separation of the different cargo molecules. The majority of plasma membrane proteins and lipids are actively and accurately recycled back to the cell surface for additional rounds of internalization (Figure 10). Many receptors are uncoupled from their ligands at the early endosome mildly acidic pH (pH ~ 6.2) and then returned to the plasma membrane either through a direct and rapid recycling pathway (fast recycling route) or via the recycling endosomes (slow recycling route) (Goldstein et al., 1985; Huotari and Helenius, 2011; Scott et al., 2014). The fast recycling route (t1/2 = 2-5 minutes) is regulated by the small GTPases Rab4 and Rab35. The slow recycling route (t1/2 = 15-30 minutes) involves the transport of cargo proteins in vesicles, along MTs, to the perinuclear endocytic recycling compartment (ERC), which is localized near the microtubule-organizing center (MTOC) and Golgi complex in many of the tissue culture cells. Subsequently, recycling endosomes detach from the ERC and are transported to the plasma membrane. This pathway is coordinated by Rab11, which associates with recycling endosomes and the ERC (Grant and Donaldson, 2009; Jovic et al., 2010). In addition to its role as a divergent point between the recycling and degradative pathways, the early endosome is also the intersection point between Figure 11 – The early endosome ultrastructure. Early endosomes (EE) display inward budding invaginations (arrows with an asterisk) and recycling membrane tubules (arrowheads). Additionally, the early endosome on the right side of the figure contains intraluminal vesicles (ILVs, shown with the + signal). Data from chinese hamster ovary cells is depicted (Sachse et al. 2002). endosomal vacuoles increased with 17% after wortmannin treatment, these data indicate a reduction of coated membrane area. We conclude that the formation and/or maintenance of the endosomal bilayered coat depends on PtdIns 3-kinase activity. Bilayered Coats Are Not Associated with Actin or Tubulin Cytoskeleton The movement and intracellular positioning of endocytic organelles depends on association with microtubules (Matteoni and Kreis, 1987; Habermann et al., 2001) and actin filaments (van Deurs et al., 1995). We therefore sought to determine whether the bilayered coat on endosomal vacuoles might function as possible anchor site for the cytoskeleton. Labeling for ! -tubulin revealed long linear patterns of gold, reflecting the longitudinal sectioning of microtubules (Figure 4, A and B). Although we frequently observed the association of LEs and lysosomes with microtubules (Figure 4A; our unpublished data), we found no linear tracks or single clusters of tubulin in close vicinity to the coated areas of EE vacuoles (Figure 4B). In addition, the microtubuledepolymerizing agent nocodazole had no influence on the presence of the bilayered coats (our unpublished data). Actin filaments are required for transport toward lysosomes (van Deurs et al., 1995). In addition, in Xenopus egg extracts, actin nucleation on endosomal vacuoles was demonstrated in vitro (Taunton et al., 2000). Labeling with an antibody against actin revealed actin present in the cytosol, especially in the region under the plasma membrane and in cellular protrusions (Figure 4C). In addition, label was sometimes observed near the limiting membrane of endosomal vacuoles, but not in association with the coated areas of their limiting membrane (Figure 4D). Incubation of cells with latrunculin A, which causes disassembly of actin filaments, had no effect on the occurrence of the bilayered coats (our unpublished data). In summary, these data do not yield any indication that the coated membranes provide the site of interaction of EEs with the cytoskeleton. EGFR and GHR but Not TfR Are Concentrated in Bilayered Coats of Early Endosomes To investigate a possible involvement of the bilayered coated areas in protein sorting within EEs, we next established the steady-state distribution of a recycling receptor, the TfR, and two receptors that are degraded upon ligand binding, the EGFR and GHR. As shown in many other cell types, the majority of intracellular TfR in wtGHR cells was found in REs, and only a minor amount at the limiting membrane of EE vacuoles (Figure 5A). Of this relatively low amount of label associated with the EE vacuole, 11.1% was localized on internal vesicles, which most likely represents the small percentage of TfR that is targeted to lysosomes (Omary and Trowbridge, 1981). Similar distribution patterns were found in the other cell lines used in this study. At the limiting membrane of the EE vacuoles, we found !10% of the TfR label in bilayered coated areas (Table 2). Almost similar values were obtained when the Tf–TfR complex was labeled with anti-Tf (Table 2), indicating that TfR antigenicity was not masked by the presence of the coat. Also in HeLa cells, only a very small percentage of TfR at the limiting membrane of EEs was localized in the coated areas (Table 3). Figure 1. Bilayered coats on endosomal vacuoles contain clathrin heavy and light chain. (A) Chinese hamster ovary cell. Clathrin heavy chain (10-nm gold) is found on the flat, coated region of an EE. The clathrin coat consists of a narrow electron-dense layer opposed to the endosomal limiting membrane (arrows) and a fuzzier layer facing the cytoplasm (arrowheads). Note that the recycling tubule emerges opposite to the coated region (small arrowheads). (B) Also in wtGHR cells, clathrin heavy chain (10-nm gold) is found in the coated areas (arrows). Note that no label is found on the inward budding vesicles (small arrows). As in A the emerging tubule has no continuity with the coated region (small arrowheads). (C) Example of an early endosome in wtGHR cells, labeled for clathrin light chain (10-nm gold) in the coated areas (arrows). PM, plasma membrane. Bars, 200 nm. M. Sachse et al. Molecular Biology of the Cell1318 * * endosomal vacuoles increased with 17% after wortmannin treatment, these data indicate a reduction of coated membrane area. We conclude that the formation and/or maintenance of the endosomal bilayered coat depends on PtdIns 3-kinase activity. Bilayered Coats Are Not Associated with Actin or Tubulin Cytoskeleton The movement and intracellular positioning of endocytic organelles depends on association with microtubules (Matteoni and Kreis, 1987; Habermann et al., 2001) and actin filaments (van Deurs et al., 1995). We therefore sought to determine whether the bilayered coat on endosomal vacuoles might function as possible anchor site for the cytoskeleton. Labeling for ! -tubulin revealed long linear patterns of gold, reflecting the longitudinal sectioning of microtubules (Figure 4, A and B). Although we frequently observed the association of LEs and lysosomes with microtubules (Figure 4A; our unpublished data), we found no linear tracks or single clusters of tubulin in close vicinity to the coated areas of EE vacuoles (Figure 4B). In addition, the microtubuledepolymerizing agent nocodazole had no influence on the presence of the bilayered coats (our unpublished data). Actin filaments are required for transport toward lysosomes (van Deurs et al., 1995). In addition, in Xenopus egg extracts, actin nucleation on endosomal vacuoles was demonstrated in vitro (Taunton et al., 2000). Labeling with an antibody against actin revealed actin present in the cytosol, especially in the region under the plasma membrane and in cellular protrusions (Figure 4C). In addition, label was sometimes observed near the limiting membrane of endosomal vacuoles, but not in association with the coated areas of their limiting membrane (Figure 4D). Incubation of cells with latrunculin A, which causes disassembly of actin filaments, had no effect on the occurrence of the bilayered coats (our unpublished data). In summary, these data do not yield any indication that the coated membranes provide the site of interaction of EEs with the cytoskeleton. EGFR and GHR but Not TfR Are Concentrated in Bilayered Coats of Early Endosomes To investigate a possible involvement of the bilayered coated areas in protein sorting within EEs, we next established the steady-state distribution of a recycling receptor, the TfR, and two receptors that are degraded upon ligand binding, the EGFR and GHR. As shown in many other cell types, the majority of intracellular TfR in wtGHR cells was found in REs, and only a minor amount at the limiting membrane of EE vacuoles (Figure 5A). Of this relatively low amount of label associated with the EE vacuole, 11.1% was localized on internal vesicles, which most likely represents the small percentage of TfR that is targeted to lysosomes (Omary and Trowbridge, 1981). Similar distribution patterns were found in the other cell lines used in this study. At the limiting membrane of the EE vacuoles, we found !10% of the TfR label in bilayered coated areas (Table 2). Almost similar values were obtained when the Tf–TfR complex was labeled with anti-Tf (Table 2), indicating that TfR antigenicity was not masked by the presence of the coat. Also in HeLa cells, only a very small percentage of TfR at the limiting membrane of EEs was localized in the coated areas (Table 3). Figure 1. Bilayered coats on endosomal vacuoles contain clathrin heavy and light chain. (A) Chinese hamster ovary cell. Clathrin heavy chain (10-nm gold) is found on the flat, coated region of an EE. The clathrin coat consists of a narrow electron-dense layer opposed to the endosomal limiting membrane (arrows) and a fuzzier layer facing the cytoplasm (arrowheads). Note that the recycling tubule emerges opposite to the coated region (small arrowheads). (B) Also in wtGHR cells, clathrin heavy chain (10-nm gold) is found in the coated areas (arrows). Note that no label is found on the inward budding vesicles (small arrows). As in A the emerging tubule has no continuity with the coated region (small arrowheads). (C) Example of an early endosome in wtGHR cells, labeled for clathrin light chain (10-nm gold) in the coated areas (arrows). PM, plasma membrane. Bars, 200 nm. M. Sachse et al. Molecular Biology of the Cell1318 200 nm + + 51 the endocytic and biosynthetic routes. Some lipids and proteins, such as the cation-independent mannose-6-phosphate receptor (CI-M6PR), are trafficked to the TGN (Figure 10), a process called retrograde transport (Bonifacino and Rojas, 2006; Scott et al., 2014). This trafficking pathway is mediated by the retromer complex, which consists of a core of at least three cargo selection proteins (Vps26, Vps29 and Vps35) and members of the sorting nexin (SNX) family (SNX1, SNX2, and possibly SNX5 and SNX6) (Seaman, 2012). Formation and maturation of late endosomes The early endosome is also the starting point in the degradative endolysosomal pathway. Since the majority of the endocytosed cargo is actively recycled back to the plasma membrane, the transport to lysosomes can be considered as a side pathway that is limited to a relatively small fraction of the internalized components. Furthermore, cargo that is directed to the degradative pathway needs to undergo stringent selective selection (Figure 10). Different proteins associate with the early endosome membrane thus orchestrating its maturation and function. Rab5 is a key component of the early endosome and, together with its effector hVPS34, a PI(3)K, regulates generation of phosphatidylinositol-3-phosphate (PtdIns(3)P), which is the most abundant phosphoinositide on this organelle membrane. Generation of PtdIns(3)P recruits other Rab5 effectors that bind to this phosphoinositide via a FYVE domain, such as early endosomal antigen-1 (EEA-1) or hepatocyte growth factor-regulated tyrosine kinase substrate (Hrs), which in turn mediate homotypic fusion of early endosomes (Huotari and Helenius, 2011; Jovic et al., 2010). Additionally, several SNAREs associate with the early endosome, mediating events of membrane fusion during its maturation. Rab5 is also a crucial protein in endosome maturation, as it is the main regulator of the conversion to late endosomes at later stages. As described before, early endosomes have a complex structure with tubular and vacuolar domains (Figure 11), and the latter are important for gradual maturation into late endosomes. The presence of the Hrs protein on the early endosome allows recruitment of the endosomal sorting complex required for transport (ESCRT) -I, through interaction with its Tsg101 subunit. Subsequently, ESCRT-II and ESCRT-III complexes are recruited to the early endosome, leading 52 to membrane invagination and formation of intraluminal vesicles (ILVs). Furthermore, homotypic fusion leads to the accumulation of additional ILVs and formation of multivesicular bodies (MVBs), which are enriched in PtdIns(3)P and PtdIns(3,5)P2 (Jovic et al., 2010; Scott et al., 2014; Wollert and Hurley, 2010). Once formed, MVBs quickly acidify to a pH ~ 5.5, and move along MT towards the cell center for further maturation into late endosomes (Figure 10). The switch from early-to-late endosomes involves the conversion from Rab5 to Rab7, and this change in membrane identity ensures that different functions in the endosomal pathway remain spatially, temporally and functionally separated (Huotari and Helenius, 2011). Initial Rab5 activation, promoted by the GEF Rabex-5 and Rabaptin-5, establishes a feedback loop in which Rab5-GTP promotes further Rab5 binding, thus defining the identity of the early endosome (Horiuchi et al., 1997; Lippé et al., 2001). At later stages of endosome maturation, Rab5 recruits and activates Rab7, which in turn inactivates Rab5 and promotes its dissociation through a negative feedback loop (Del Conte-Zerial et al., 2008). In this way, there is the formation of the late endosome, an hybrid compartment that is typically round or oval and has a diameter of 250-1000 nm. Rab7-GTP then recruits several downstream effectors, including RILP (a protein that connects late endosomes to dynein motors), components of the homotypic fusion and protein sorting (HOPS) complex (function as a tether of late endosome fusion), and machinery for heterotypic vesicle fusion (Zhang et al., 2009). This leads to the formation of an organelle containing multiple ILVs. One of the crucial roles of the Rab5/Rab7 conversion is to exchange the fusion machinery on the endosomal membrane, guarantying that late endosomes can only fuse with other late endosomes, lysosomes, and possibly autophagosomes (Huotari and Helenius, 2011). Importantly, late endosomes also interact with vesicles deriving from the TGN, thereby acquiring protein markers such as the lysosomal associated membrane protein-1 (Lamp-1), M6PR and lysosomal hydrolase precursors (Scott et al., 2014). Endosomal motility is crucial for its maturation. Is has been suggested that the distance to the nucleus is a key parameter that influences intracellular position, number, size, and cargo contents of endosomes during endosomal maturation (Collinet et al., 2010). Movement of endosomes depends on both dynein and kinesin motors, which provide opposing forces that move endosomes in opposite 53 directions. Additionally, these motor proteins play a role in fusion of endosomes with each other during different steps of maturation (Driskell et al., 2007; Soppina et al., 2009). Whereas kinesins are implicated in the movement of early endosomes and late endosomes, dynein-dependent transport seems to be the major factor involved in the motility of late endosomes. This compartment binds dynein either directly or through the adaptor protein dynactin, which is dependent on Rab7-mediated recruitment of RILP (Huotari and Helenius, 2011). Moreover, members of the annexin family of proteins, such as annexins A2 and A8, link the actin cytoskeleton the endosome membranes, which might regulate early endosome fission and contribute to late endosome biogenesis (Goebeler et al., 2008; Morel et al., 2009). Endosome fusion with lysosomes Late endosomes that are formed continue to undergo maturation mechanisms and eventually fuse with lysosomes. Lysosomes are acidic membrane-bound organelles (luminal pH can drop to 4.5) containing proton-pumping vacuolar ATPases (vATPases) and acid hydrolases (Mellman et al., 1986). They were discovered 60 years ago by Christian de Duve, as a result of studies using subcellular fractionation (DE DUVE et al., 1955), and were then visualized by electron microscopy (NOVIKOFF et al., 1956), which showed that lysosomes appear heterogeneous in size and morphology (Figure 12). More recently, time-lapse confocal microscopy studies have contributed to the understanding of how lysosomes dynamically interact with endosomes. Electron microscopy data, together with cell-free content-mixing assays, provided evidence that late endosomes and MVBs fuse directly with lysosomes (reviewed in (Luzio et al., 2007)). The majority of these fusions events happen in the perinuclear region, since late endosomes and lysosomes are concentrated near the MTOC. The usage of correlative live-cell and electron microscopy has allowed the understanding of the successive steps during endosome-lysosome abc Dominant-negative mutant A protein encoded by a mutated gene that prevents the function of the wild-type protein in cells in which both the mutant and wild-type proteins are expressed at the same time. involved in the fusion of late endosomes with lysosomes (FIG. 3). In common with other fusion events in the secretory and endocytic pathways, the fusion of late endosomes and lysosomes requires the presence of N-ethylmaleimide sensitive factor (NSF), soluble NSF attachment proteins (SNAPs) and a small GTPase of the Rab family, probably RAB7 (REF. 48). Similar to other fusion events, the process can be considered as having three sequential steps: tethering, the formation of a trans-SNARE (SNAP receptor) complex that bridges across the two organelles and membrane fusion. Tethering. A prerequisite to organelle fusion is organelle tethering, whereby two organelles form links between each other that extend over distances of >25 nm from a given membrane surface. The physical existence of tethers between late endocytic organelles was implied by morphological observations of fine striations between adjacent late endosomes and lysosomes in cultured cells46,47,49, as well as by the ability to show endosome– lysosome interactions in cell-free systems50 (FIG. 2c). The composition of the tethers has not been established but the mammalian homotypic fusion and vacuole protein sorting (HOPS) complex, which is recruited by RAB7, is a good candidate (FIG. 3a). Overexpression of the mammalian HOPS complex components VPS18 and VPS39 caused clustering of late endosomes and lysosomes51,52 and depletion of VPS18 resulted in organelle dispersion52. However, it is clear that these components are not specific for heterotypic late endosome–lysosome fusion and that they also function in homotypic endosome fusion even in the early part of the endocytic pathway53. Overexpression of RAB7 and some RAB7 effectors can also cause clustering of late endocytic organelles54,55, and dominant-negative mutants of RAB7 cause dispersion54. It should be noted that tethering is a separate process from that which causes the accumulation of late endosomes and lysosomes in the juxtanuclear region around the microtubule-organizing centre, although this might also increase the efficiency of delivery of endocytosed macromolecules to lysosomes. Juxtanuclear accumulation reflects the balance of long-range bidirectional movement on microtubules and short-range movement on actin filaments. Such movement is mediated by motor proteins and proteins that are required for the optimal attachment of these motors to late endocytic organelles, including the RAB7 effector RAB7-interacting lysosomal protein (RILP)56 and BLOC3 (REF. 57). Trans-SNARE complex formation. Following tethering, a trans-SNARE complex must form in which the ~16-turn helix of one SNARE wraps around similar helices on three other SNAREs to form a parallel four-helix bundle called a SNAREpin, which is essential for membrane fusion58. The centre of the four-helix bundle contains an ionic layer comprising an arginine (R) and three glutamine (Q) residues, each contributed by a different SNARE. These residues are termed R-SNARE and Qa-, Qband QcSNAREs, respectively59,60. A functional trans-SNARE complex must contain one helix of each type61. Antibody-mediated function-blocking experiments in cell-free systems have provided the most compelling evidence that the same Qa, Qb and Qc SNAREs — syntaxin-7, VTI1B (VPS10 tail interactor1B) and syntaxin-8, respectively — are required both for homotypic late endosome fusions and heterotypic late endosome–lysosome fusions62–64. What distinguishes the two fusion events is the R-SNARE, which is vesicleassociated membrane protein-8 (VAMP8) for homotypic late endosome fusion, and VAMP7 (also known as Figure 2 | Electron microscopy of endosome–lysosome fusion. a | Dense-core lysosomes in normal rat kidney (NRK) cells were loaded with colloidal gold conjugated with bovine serum albumin for 4 h followed by a 24 h chase. The lysosomes (dark grey) can be compared with a less-dense late endosome in the centre of the image. b | An electrondense lysosome (arrowhead) in an NRK cell is captured in the process of fusing directly with an electron-lucent endosome by correlative live-cell imaging and electron microscopy. The image shown is from a 50 nm serial section immediately adjacent to that shown in Bright et al. (REF. 12). c | Immunogold electron microscopy of dense lysosomes from a rat liver preparation (labelled with a cathepsin D lysosomal marker; 15 nm gold). Lysosomes were isolated following an in vitro content-mixing assay48. The image shows that multiple lysosomes can form robust attachments (outer arrows) with an endosome (central arrowhead). The endosome was loaded with asialofetuin–avidin for 6 min and was subsequently immuno-labelled with 10 nm colloidal gold. The scale bar in part a is 500 nm; parts b and c are the same scale as part a. REVIEWS NATURE REVIEWS | MOLECULAR CELL BIOLOGY VOLUME 8 | AUGUST 2007 | 625 Figure 12 – Electron microscopy of late endosomes and lysosomes. Dense-core lysosomes (dark grey) are structurally different from a less-dense late endosome seen in the center of the image. Scale bar is 500 nm (Luzio et al. 2007). 54 fusion (Bright et al., 2005). There is initial physical contact between the organelles, that then transiently fuse (kissing events) or undergo permanent fusion. Additionally, live-cell imaging studies have helped to establish the protein machinery involved in the fusion of late endosomes with lysosomes, which requires small GTPases, SNAREs, among others (Luzio et al., 2007). Figure 13 – Schematic representation of homotypic late endosome fusion and heterotypic late endosome-lysosome fusion. A. Different cytosolic proteins, such as Rab7, NSF (N-ethylmaleimide sensitive factor) and SNAPs (soluble NSF attachments proteins), tether endosomes with lysosomes or endosomes with endosomes. B. Two different trans-SNARE complexes induce either homotypic late endosome fusion (VAMP8-dependent) or heterotypic late endosome-lysosome fusion (VAMP7-dependent). C. Vesicle fusion is dependent on the release of lumenal Ca2+ (only shown for heterotypic fusion). Reformation of dense-core lysosomes from the endolysosome hybrid organelle requires the loss and retrieval of proteins. Adapted from Luzio et al., 2007. Heterotypic late endosome-lysosome fusion starts with organelle tethering involving the formation of links between the two structures (Figure 13A). For this process, the HOPS complex is thought to be an important player, possibly in conjunction with Rab7 (Luzio et al., 2010). Subsequently, fusion depends on a core protein machinery that includes cytosolic factors and the formation of a SNAREpin composed of an R-SNARE and three Q-SNAREs (one each of Qa-, Qband QcSNAREs) (Luzio et al., 2007; Weber et al., 1998). For heterotypic late endosome-lysosome fusion, the Q-SNAREs are syntaxin-7, Vti1b and syntaxin-8 Lumenal Ca2+ Lumenal Ca2+ Calmodulin Hybrid organelle Late endosome Late endosome L y sosome Retrieval a Tethering and docking b SNARE assembly c Fusion Homotypic fusion Heterotypic fusion RAB7 NSF SNAP Condensation VAMP8 VAMP7 Syntaxin-7 VTI1B Syntaxin-8 SNAREs: Tethers Liposome An artificial lipid vesicle that encloses an aqueous interior. Isopycnic ultracentrifugation Centrifugation of samples (organelles or macromolecules) in a density gradient until an equilibrium is reached, such that the density of the sample is the same as that part of the density gradient in which it equilibrates. Retromer complex A complex of cytoplasmic proteins that are required for some retrograde membranetrafficking pathways that deliver cargo from endosomes to the trans-Golgi network. tetanus-neurotoxin-insensitive VAMP (TI-VAMP) or synaptobrevin-like-1 (SYBL1)) for heterotypic late endosome–lysosome fusions64 (FIG. 3b). As described below and in BOX 2, VAMP7 is found in numerous combinatorial SNARE complexes, several of which involve lysosomes. VAMP7 is an unusual R-SNARE because it has a relatively long (~110 amino acid) N-terminal extension that might function as a regulatory domain. This so-called longin domain is required for the delivery of VAMP7 to late endocytic compartments as a result of binding to AP3 (REF. 65). Membrane fusion. It has not been proven whether transSNARE complex formation on its own is sufficient to result in phospholipid bilayer fusion between endosome and lysosome membranes. Trans-SNARE complexes can cause fusion of liposomes but kinetic differences with biological fusion reactions have been reported61. In the case of heterotypic fusion of endosomes with lysosomes, there is evidence from cell-free contentmixing assays that membrane fusion is dependent on Ca2+ and calmodulin13. Ca2+ is released from the lumen of the fusing organelles late in the mechanistic pathway of fusion (FIG. 3c). Hybrid organelles. The immediate product of direct and complete fusion between a late endosome and a lysosome is a hybrid organelle that contains a full complement of lysosomal enzymes but still contains some MPRs. This hybrid organelle is the site of degradation of endocytosed macromolecules. The demonstration of direct fusion between the two organelles is consistent with the idea that lysosomes are, fundamentally, storage granules for mature lysosomal enzymes, and that they periodically fuse with late endosomes to form a compartment, sometimes referred to as a ‘cell stomach’, in which degradation occurs66. It is noticeable that lysosomes morphologically resemble regulated secretory granules66, and their content might well be less aqueous than the cytoplasm or the lumen of the endosome. It is also known that lysosomes behave as dense organelles following isopycnic ultracentrifugation on a variety of density gradients used for subcellular fractionation. Hybrid organelles have an intermediate density between those of lysosomes and late endosomes48. Reformation of lysosomes from hybrids. The direct and complete fusion of late endosomes with lysosomes would consume both organelles if no recovery process occurred. Therefore, lysosome reformation from hybrid organelles is necessary and requires content condensation and a membrane-retrieval process to remove endosomal membrane proteins and recycle SNAREs. Lysosomes can be reformed from hybrid organelles in a cell-free system, in which content condensation requires a proton-pumping ATPase and lumenal Ca2+ (REF. 13). In a study of asialoglycoprotein endocytosis and degradation in rat hepatocytes, it was also suggested that phosphoino sitide-3-kinase activity is required for the reformation of dense lysosomes from hybrid organelles67. In live-cell experiments following endosome–lysosome fusion, small vesicular tubular structures have been observed budding off hybrid organelles, consistent with a lysosome reformation process12. VAMP7-positive vesicles were also observed budding from terminal endocytic compartments in a live-cell study of organelles containing the Niemann–Pick C1 protein68. Overall, the lysosome reformation process will be one of maturation and, by definition, it is only at the point at which no MPRs are detectable in the organelles that they can be called lysosomes. One candidate for the machinery that mediates membrane retrieval from the hybrid organelles is the retromer complex. This complex was first described in yeast as a complex of Vps5, Vps17, Vps26, Vps29 and Vps35. Depletion of the Vps26 orthologue in mammalian cells leads to a phenotype in which there is some swelling and vacuolarization of lysosomal compartments15,69. Figure 3 | Schematic models of heterotypic late endosome–lysosome fusion and homotypic late endosome fusion. a | The small GTPase RAB7, possibly in conjunction with the mammalian homotypic fusion and vacuole protein sorting (HOPS) complex, is thought to tether endosomes and lysosomes (or endosomes with endosomes). The fusion of late endosomes and lysosomes requires N-ethylmaleimide sensitive factor (NSF) and soluble NSF attachment proteins (SNAPs). b | trans-SNARE (SNAP receptor) complex formation requires syntaxin-7, VTI1B (Vps10 tail interactor-1B) and syntaxin-8 in both homotypic late endosome fusions and heterotypic late endosome–lysosome fusions. Whereas vesicle-associated membrane protein-8 (VAMP8) is required for homotypic late endosome fusion, VAMP7 is needed for heterotypic late endosome– lysosome fusions. Two different combinatorial trans-SNARE complexes are shown. c | The release of lumenal Ca2+ (shown only for heterotypic fusion) leads to phospholipid bilayer fusion. Reformation of lysosomes from hybrid organelles requires the loss of mannose-6-phosphate receptors, SNARE retrieval and condensation of lumenal content to produce dense-core lysosomes. It should be noted that all five of the SNAREs shown have been observed on both late endosomes and lysosomes. Release of lumenal Ca2+, which is necessary for membrane fusion between endosomes and lysosomes, is probably promoted by trans-SNARE complex formation, as has been described for homotypic vacuole fusion in yeast133. REVIEWS 626 | AUGUST 2007 | VOLUME 8 www.nature.com/reviews/molcellbio A.#Tethering#and#docking# B.#SNARE#assembly# C.#Fusion# 55 (Qa-, Qband QcSNAREs, respectively), whereas the R-SNARE is VAMP7. Interestingly, the Q-SNAREs required for homotypic late endosome fusion are the same, however the R-SNARE is VAMP8 instead of VAMP7 (Figure 13B) (Luzio et al., 2009; Pryor et al., 2004). VAMP7 is a protein of particular interest since it is known to be present in numerous combinatorial SNARE complexes (Table 2), many of which involve lysosomes. Its presence on the lysosome membrane might help the definition of the lysosome and regulate its fusion with late endosomes. Furthermore, VAMP7 is involved in lysosome fusion with the plasma membrane, where it forms a transSNARE complex with syntaxin-4 and SNAP23 (Luzio et al., 2009; Rao et al., 2004). Table 2 – VAMP7 can associate with several Q-SNAREs and form different combinatorial trans-SNARE complexes. Adapted from Luzio et al., 2007. Qa-SNARE Qb-SNARE Qc-SNARE Qb/c-SNARE Function Syntaxin-7 Vti1B Syntaxin-8 Late endosome-lysosome fusion Syntaxin-7 Vti1 Syntaxin-8 Macropinosome fusion Syntaxin-4 SNAP23 Lysosome-plasma membrane fusion Syntaxin-3 SNAP23 Vesicle-apical plasma membrane fusion Syntaxin-1 SNAP25 Vesicle-neurite plasma membrane fusion The final stages of late endosome-lysosome fusion are also dependent on calmodulin and Ca2+ (Figure 13C), which is released from the lumen of the organelles and allows phospholipid bilayer fusion (Pryor et al., 2000). After complete heterotypic fusion, the formation of a transient hybrid organelle called endolysosome takes place, which can be referred to as a “cell stomach” where degradation occurs. Here, hydrolase precursors are proteolytically cleaved into active forms, thus contributing to the degradative environment of the endolysosome. Finally, endolysosomes undergo content condensation and membrane-retrieval processes to remove endosomal membrane proteins and recycle SNAREs. In this way there is reformation of classical lysosomes (Figure 10), which are Lamp-1-positive but M6PR-negative spherical dense storage organelles (Luzio et al., 2010). 62 recent data in yeast show that regions of the Golgi can approach and contact the ERES. This, together with the collapse of COPII coats, enables cargo capture into the Golgi, thus ensuring efficient and targeted cargo transport from the ERES to the Golgi (Kurokawa et al., 2014). Figure 17 – The ER-Golgi interface in eukaryotic cells. COPII-coated vesicles that bud from the ERES are released into the cytoplasm and travel along MTs to the ERGIC. In parallel, there is depolymerization of the Sec13-Sec31 cage. Vesicles eventually tether with the ERGIC, in a Sec23 and TRAPPI-dependent manner. COPI mediates cargo transport from the ERGIC to toward the Golgi, as well as recycling back to the ER membrane (not shown in this scheme). Adapted from Brandizzi and Barlowe, 2013. The ERGIC can also form COPI-coated vesicles that are generated from anterograde carriers as they move towards the Golgi. Though, the most biological relevant role of COPI-coated vesicles is to regulate retrograde transport from the Golgi to back the ER (Szul and Sztul, 2011). The COPI complex consists of a heptameric (α, β, β’, γ, δ, ε, ζ) complex, also called coatomer. Two main subcomplexes then form the coatomer: the γ-COP-δ-COP-ζ-COP-β-COP tetrameric complex, which constitutes the inner layer core; and the α-COP-β’- COP-ε-COP trimeric complex, which forms the outer layer of the COPI coat. COPI is recruited to membranes by activated Arf1 GTPases. Once activated by ARFGEFs, Arf1 recruits preassembled COPI coats to the Golgi membrane, which then facilitates transport out of the Golgi and ERGIC (Brandizzi and Barlowe, 2013; Jackson, 2014; Szul and Sztul, 2011). Interestingly, several studies also show that inhibition of the retrograde transport route leads to the collapse of anterograde trafficking (Lippincott-Schwartz et al., 1989; Niu et al., 2005; Richter ? ER Golgi proteins has also been observed upon photobleaching of Golgi in untreated cells24, again supporting the model that ER export can occur toward motile Golgi. However, it has not yet been established whether there are differences in ER–Golgi transport kinetics between cells with disrupted versus intact actin, or between fast or slow moving Golgi stacks. Different forms for unique functions A conserved core machinery transports biosynthetic cargo forward in the early secretory pathway and is balanced with retrieval routes that maintain the ER and Golgi compartments. In eukaryotic cells, COPII assembly produces transport intermediates from the ER, which then rely on RAB GTPase-dependent tethering factors and the SNARE-dependent membrane fusion machinery for delivery of cargo to the Golgi. Similarly, retrograde transport depends on the COPI machinery to produce retrograde-directed vesicles that also require conserved tethering factors and SNARE proteins for fusion with ER membranes. In spite of this conservation, diversity in eukaryotic cell types and cellular functions generate tremendous variety in the organization of this ER–Golgi interface. Each trafficking step between ER and Golgi compartments provides multiple opportunities for regulation that could influence vesicle and organelle size and shape. Indeed, there are now clear examples in which the expression levels of these core components and their covalent modification (for example, phosphorylation or ubiquitylation) alter the morphology of COPII carriers. Moreover, the additional control provided by accessory factors such as TANGO1 and cTAGE5 and cytoskeletal components in certain cell types also affects ER–Golgi organization. Finally, inherent flexibility in both the COPII35 and COPI107 coats that allows cargos of different sizes to be accommodated can influence the morphology of ER–Golgi transport intermediates in different cell types. The influence of cargo. The striking differences in the organization of ER–Golgi interfaces across species that are revealed by in vivo morphological analyses probably reflect cellular specialization related to the types of cargo that must be transported as well as the overall cell structure and function. If we consider the possibility that distinct morphologies observed reflect functional adaptation of the core transport machinery, a molecular understanding of these differences should provide important insights into how trafficking and compartmental organization are integrated within the early secretory pathway. For example, the existence of an ERGIC in certain cell types and not in others remains mysterious. As discussed above, P. pastoris yeast cells56 and vacuolated plant cells106 lack an ERGIC but instead contain compact ER–Golgi units (300 nm distance between the two organelles) that are thought to be firmly connected through a tethering matrix. However, such a compact arrangement may not provide the space that would be needed for the assembly of large cargo into dissociating transport carriers without generating membrane connections between the ER and Golgi compartments. Indeed, there are no known large Figure 3 |The ER–Golgi interface and ERES have a distinct organization in mammals and plants. a | In mammalian cells, ER exit sites (ERES) are orientated towards a juxtaposed endoplasmic reticulum (ER)–Golgi intermediate compartment (ERGIC). Coat protein complex II (COPII)-coated vesicles originate within cup-shaped ER subdomains, which are associated with the plus end of microtubules. Upon fission of vesicles from the ERES, the SEC13–SEC31 cage is depolymerized, but the SEC23–SEC24 coat is partially retained. Vesicles reach the ERGIC in a microtubule-independent man ne r w he r e t h ey ar e t e th er e d t h r o u g h t h e i n te r acti on b et w ee n S EC 23 an d t h e TR A PPI (t ransport protein particle I) tethering complex. COPI mediates forward protein transport from the ERGIC towards the Golgi as well as recycling back to the ER membrane (the latter is not shown).b| In plant cells, ERES and Golgi are closely associated, possibly through a mat ri x (i n d i cate d i n g r ey) t h at h o l d s t h e E R an d t h e G o l g i tog eth er. T he e xi ste nc e o f C O PI I ve sicles in plants is still debated (denoted by the question mark); vesicle-like structures have been seen89,97,99,106 rarely in electron microscopy analyses of high-pressure frozen Arabidopsisthaliana tissues, although it was unclear whether they were undergoing budding or fusion. Unlike mammalian cells, plant cell ER–Golgi transport does not rely on the microtubule cytoskeleton. REVIEWS 388 | JUNE 2013 | VOLUME 14 www.nature.com/reviews/molcellbio © 2 0 13 M ac mi l l an P ubl i s her s Li mi t ed. A l l ri ght s res er ved Cytoplasm ER Golgi ERGIC COPI TRAPPI Sec13-Sec31 Sec23-Sec24 MT COPII vesicles 63 et al., 2007; Sciaky et al., 1997). Thus, it is likely that the integrity of the anterograde pathway depends on the homeostasis of the retrograde pathway, which guarantees not only the retrieval of resident proteins that escape the ER but also facilitates the recycling of lipids back to the ER. 64 65 Invasive Bacteria – Master Hijackers of Cellular Processes athogenic bacteria can cause diseases that are devastating for humans and other animals. During their long lasting coexistence and coevolution with their hosts, many of these bacteria have developed diverse strategies to survive in the host, either replicating outside the cells of the infected organism or within the host-cell cytoplasm. In this way, these so called intracellular bacterial pathogens avoid the harsh extracellular environment and many immune defense mechanisms, such as circulating antibodies or complement-induced destruction. Furthermore, professional phagocytic cells, such as macrophages, engulf pathogenic bacteria (such as Mycobacterium tuberculosis and Legionella pneumophila). The pathogens have evolved ways to impair phagosome maturation and fusion with lysosomes, thus avoiding their degradation. Other bacteria use an arsenal of virulence factors (or effectors) that subvert the host cell cytoskeleton and endocytic machineries. In this way, these so called invasive bacteria (such as Salmonella enterica, Shigella flexneri or Listeria monocytogenes) trigger their own uptake into non-phagocytic cells, such as epithelial cells, and are internalized into an endocytic membrane-bound vacuole (Figure 18). These bacteria also modify their internalization vacuole, either to survive within it (intravacuolar lifestyle) or to promote its rupture and escape to the cytosol (cytosolic lifestyle) (Cossart and Helenius, 2014; Fredlund and Enninga, 2014; Ham et al., 2011; Kumar and Valdivia, 2009). In both cases, this requires a fine tuned manipulation of the host cell vesicular trafficking by the bacteria, which will be addressed in detail in this chapter. Over the last decades, the combination of microbiology and cell biology into a new discipline, cellular microbiology (Cossart et al., 1996), has offered important insights into bacterial pathogenesis, transmission and dissemination. Moreover, the study of how bacteria hijack their host cells has also been crucial for the understanding of fundamental aspects of cell biology. P Chapter!4! 66 Figure 18 – Cellular invasion routes of intracellular bacterial pathogens and their intracellular fate. Some intracellular bacteria are phagocytosed by macrophages and then reside inside a modified phagosome. Whereas some bacteria can invade epithelial cells through an endocytosis-like “zipper” mechanism, other trigger massive membrane ruffling of the host cell surface, which is also accompanied by macropinocytosis (“trigger” mechanism). In the case of Bartonella henselae, specialized invasion structures called invasomes can be formed. Once inside the host cell, different bacteria have different fates: whereas some reside and replicate within a membrane-bound vacuole, others lyse this compartment and then replicate in the cytosol. Adapted from Cossart and Helenius, 2014. A. Exploiting host cell mechanisms for bacterial entry During bacterial-induced internalization into host cells, the bacterium plays the major and active role during the interplay between the two organisms. The other important player during this process is the host cell plasma membrane and underlying cytoskeleton, whose plasticity is exploited by the pathogen. Some invasive bacteria, such as Listeria monocytogenes, express surface proteins that bind to eukaryotic surface receptors, causing receptor clustering. In this way, signaling events are triggered that culminate in bacterial entry into epithelial cells, through a “zippering” process that involves relatively small cytoskeletal rearrangements and plasma membrane extensions for bacterial engulfment into an endocytic vacuole (Figure 19A).                         67 Figure 19 – Invasive bacteria hijack the actin cytoskeleton to promote invasion into epithelial cells. A. Listeria binds to the cell surface, via interactions between surface bacterial proteins and host cell receptors, and enter through a “zipper” mechanism, which involves minor actin rearrangements. B. Bacteria such as Shigella and Salmonella inject effector proteins into the host cell cytosol, via a type 3 secretion system (T3SS), leading to massive actin rearrangements and bacterial internalization via a “triggering” mechanism. On the lower panels, the perturbations on the host cell surface are shown, via scanning electron microscopy [left A and B (Cossart and Sansonetti, 2004)] or fluorescence microscopy [right A (Bierne et al., 2005) and B (Ehsani et al., 2012)]. Adapted from Haglund and Welch, 2011. Other invasive bacteria, such as Shigella flexneri and Salmonella enterica, use dedicated needle-like secretion systems, called type III secretion systems (T3SSs). The T3SSs are essential virulence determinants of diverse Gramnegative bacteria and are used to inject virulence proteins into eukaryotic host target cells. T3SSs are complex syringe-like macromolecular machines (Figure 20) assembled in a hierarchical manner. They consist of the structural components of the export machinery itself (called injectisome), secreted proteins (including pore-forming translocators and effectors), chaperones and cytoplasmic regulators. The injectisome is made up of a basal body that spans both bacterial membranes, and an extracellular needle complex that protrudes from the bacterial surface. Both the basal body and the needle complex are hollow, containing a channel that acts as a tube for protein secretion. The needle complex terminates in a tip structure made of several copies of the tip protein. Upon S. flexneri or S. enterica contact with a host cell membrane, the T3SS exports two different categories of proteins: the translocators that form a pore in the target membrane, and the effectors that are translocated through the pore into the host cell (Deane et al.,     68 2010; Enninga and Rosenshine, 2009; Parsot, 2009). In this way there is the injection of a cocktail of bacterial effector molecules into the host cell cytosol. Here, the injected bacterial proteins can interact with and manipulate diverse host molecules. One of the first cellular consequences of T3SS effector injection is the manipulation of the host cell machinery that regulates the actin cytoskeleton dynamics, “triggering” massive cytoskeleton changes that cause the formation of macropinocytosis-like membrane ruffles and result in bacterial entry (Figure 19B) (Cossart and Helenius, 2014; Cossart and Sansonetti, 2004; Haglund and Welch, 2011; Rottner et al., 2005). Figure 20 – Overview of the components of the type III secretion system. Left and middle panels depict, respectively, side and cut surface views representations of 3D reconstructions of needle complexes from S. flexneri and S. enterica, based on cryo-electron microscopic data. The right panel shows a schematic representation of the T3SS with all its components. Cytoplasmic components are shown in yellow, export apparatus components in red, base and needle components in blue, and needle tip and translocator proteins in green. The unified Sct names (secretion and cellular translocation) are shown. OM, bacterial outer membrane; IM, bacterial inner membrane; MS, membrane and supramembrane. Adapted from Diepold and Wagner, 2014. Even though actin polymerization is of central importance in both processes, invasive bacteria also exploit the other components of the host cell cytoskeleton. Over the last decades this has helped the scientific community to understand the dynamics of the eukaryotic cytoskeleton regulation. refer to dedicated reviews for details on structure (Moraes et al., 2008; Hodgkinson et al., 2009; Spreter et al., 2009; Erhardt et al., 2010; Izor! eet al., 2011; Schraidt & Marlovits, 2011; Chatterjee et al., 2013) and function (Cornelis, 2006; Gal! an & Wolf-Watz, 2006; B€ uttner, 2012). Needle The extracellular needle is generated by helical polymerization of a small hairpin protein (SctF) (Cordes et al., 2003; Deane et al., 2006; Fujii et al., 2012; Loquet et al., 2012). An oligomer of hydrophilic translocator proteins forms the needle tip (Mueller et al., 2005; Johnson et al., 2006; Broz et al., 2007), which is proposed to act as a scaffold for the assembly of hydrophobic translocator proteins that permeate the host cell membrane (Tardy, H # akansson et al., 1996; Blocker et al., 1999; Neyt & Cornelis, 1999; Goure et al., 2004; Picking et al., 2005; Montagner et al., 2011); reviewed in (Matte€ ıet al., 2011). In attaching and effacing animal pathogens such as E. coli pathotypes EPEC and EHEC, and Citrobacter rhodentium, in Bordetella, and in plant pathogens, the tip is replaced by a protein forming a pilus or filament (Jin & He, 2001; Chen & Frankel, 2005; Medhekar et al., 2009; Bergstrom et al., 2012). The needle is presumably anchored to the base by the alpha-helical inner rod protein SctI (Marlovits et al., 2004; Wood et al., 2008; Zhong et al., 2012). Membrane rings/the base The base consists of two membrane-spanning ring structures (Fig. 1). The outer membrane (OM) ring, which extends deeply into the periplasm, consists of 12–15 copies of a protein from the secretin family (SctC) (Koster et al., 1997; Kubori et al., 1998; Blocker et al., 2001; Spreter et al., 2009; Schraidt & Marlovits, 2011; Bergeron et al., 2013). Secretins were found to require dedicated pilot proteins for proper functioning. Pilotins form a structurally dissimilar group of OM lipoproteins that are thought to passage secretins piggyback through the periplasm by use of the Lol system. At the OM, pilotins are believed to facilitate the insertion, oligomerization, and assembly of their cargo (Koster et al., 1997; Crago & Koronakis, 1998; Daefler & Russel, 1998; OM IM Host membrane Needle tip C ring protein (SctQ) Accessory protein (SctK) Minor export apparatus proteins (SctRST) Major export apparatus protein (SctV) Switch protein (SctU) Translocon Stator (SctL) Stalk (SctO) ATPase (SctN) Secretin (SctC) Inner rod (SctI) Outer MS ring protein (SctD) Inner MS ring protein (SctJ) Outer rings Neck Inner rings Effectors Needle filament (SctF) Fig. 1. Overview of the injectisome and its components. The left and middle panels show surface representations of 3D reconstructions of NCs based on cryo-electromicroscopic data (left panel side view, middle panel cut view). Images were kindly provided by Thomas Marlovits. The right panel shows a drawing of the type III secretion holo-complex indicating all its components. Cytoplasmic components are shaded in yellow, export apparatus components in red, base and needle components in blue, and needle tip and translocator proteins in green. The drawing is based on our current structural knowledge of the complex but includes also presumed localizations for components that have so far not been unambiguously localized (e.g. the accessory protein and the inner membrane assembly of the export apparatus). FEMS Microbiol Rev 38 (2014) 802–822ª2014 Federation of European Microbiological Societies. Published by John Wiley & Sons Ltd. All rights reserved 804 A. Diepold & S. Wagner 69 Listeria monocytogenes entry by receptor signaling Listeria monocytogenes is a food-borne Gram-positive bacterial pathogen that preferentially infects immunocompromised hosts, eliciting a severe and often lethal disease, called listeriosis. In human hosts, clinical manifestations can range from gastroenteritis to severe systemic disease, such as meningitis, encephalitis, sepsis, and fetal infections that can result in abortion (Swaminathan and GernerSmidt, 2007). Classically, L. monocytogenes has been described to enter into epithelial cells through a “zipper” mechanism (Cossart and Sansonetti, 2004). This bacterium is able to bind to host cell surface receptors, through dedicated bacterial surface proteins called internalins. So far, the only two L. monocytogenes internalins that have been directly implicated in bacterial invasion are internalins A and B (InlA and InlB, respectively) (Pizarro-Cerdá et al., 2012). InlA interacts with E-cadherin, an adhesion molecule involved in the formation of adherens junctions in some epithelial barriers, and this promotes bacterial invasion in specific cellular subpopulations. InlB interacts with the hepatocyte growth factor receptor Met, promoting L. monocytogenes internalization into a broad range of cell types from epithelial origin. In this way, E-cadherin and Met are phosphorylated and ubiquitinated (via the ubiquitin ligases Hakai or Cbl, respectively), which leads to subsequent recruitment of the molecular machinery for clathrin-mediated endocytosis (Dab2, clathrin, dynamin, HIP1R, myosin-VI). In both cases, this coordinates local recruitment and activation of kinases and small GTPases that lead to Arp2/3 complex-dependent actin rearrangements and bacterial uptake (Figure 21). Given that myosin-VI has the ability to move towards the minus end of actin filaments, it is thought that it might pull the bacteria to the interior of the host cell. Ultimately, the actin remodeling is down-regulated by the recruitment of proteins such as cofilin or OCRL (reviewed in (Cossart and Helenius, 2014; Cossart and Roy, 2010; Pizarro-Cerdá et al., 2012)). Interestingly, is was also shown that some members of the septin family (septin-2 and -9) are recruited to the entry site of L. monocytogenes, contributing to the anchorage of Met to the actin cytoskeleton and to bacterial entry in epithelial cells (Mostowy et al., 2009; 2011). Together, this highlights that L. monocytogenes entry into epithelial cells is a complex process that requires several host molecular machineries. 70 Figure 21 – InlAand InlB-dependent L. monocytogenes invasion pathways. Binding of InlA to E-cadherin (A) or InlB to Met (B) lead to post-translational modifications of the receptors, recruitment of clathrin-mediated endocytosis machinery and initial actin reorganization. Subsequent receptor signaling results in further actin polymerization and bacterial internalization (Pizarro-Cerdá et al., 2012). Shigella flexneri entry by effector secretion Shigella spp. are Gram-negative, rod-shaped, facultative anaerobic, nonmotile pathogenic bacteria that cause bacillary dysentery (shigellosis) in humans. Worldwide, shigellosis constitutes a significant public health burden, mostly in developing countries, where it is thought that about 160 million cases occur Listeria Listeria Clustering Arp2/3 Caveolin Caveolin Src InlA Lipid raft Hakai p120 Arp2/3 Src The InlB pathway The InlA pathway A Actin α-Catenin α-Catenin p120 β-Catenin β-Catenin E-cadherin Hip1R MyoVI Clathrin Dab2 Listeria Listeria Ub Ub Ub Ub P P P P Ub Ub Ub P P P P P P P Listeria Listeria ListeriaListeria InIB InIB GAGs Met Cbl gC1q-R Met Soluble InIB Arp2/3 Actin Cholesterol Clathrin Hip1R Dab2 MyoVI Arp2/3 WAVE/N-WASP LIM-K Rac/Cdc42 Gab1 PI3K Cofilin Cofilin Cortactin Dynamin Actin Hip1R MyoVI Clathrin Dab2 Lipid raft PIP3 B Cortactin Dynamin Figure 2. Signaling cascades activated via the InlAand InlB-invasion pathways. Interaction of InlA and/or InlB with their respective host-cell surface receptors E-cadherin and Met induces ubiquitination of the receptors by the ubiquitin ligases Hakai in the case of E-cadherin or Cbl in the case of Met and subsequent recruitment of the clathrin endocytosis machinery (Dab2, clathrin, dynamin, Hip1R, MyoVI), which provides an initial platform for actin cytoskeleton polymerization. (See following page for legend.) Listeria monocytogenes Entry in Mammalian Epithelial Cells Cite this article as Cold Spring Harb Perspect Med 2012;2:a010009 3 www.perspectivesinmedicine.org on July 1, 2015 - Published by Cold Spring Harbor Laboratory Press http://perspectivesinmedicine.cshlp.org/Downloaded from 71 annually, resulting in over 1 million deaths especially among children under the age of 5 years (Kotloff et al., 1999). This disease is transmitted by the fecal-oral route, usually by the ingestion of contaminated water or food, and is characterized by severe inflammatory destruction of the colonic epithelium. Currently, there is no vaccine available to prevent shigellosis, and the development of antibiotic resistance to multiple drugs is a growing problem (Phalipon and Sansonetti, 2007). The pathogenicity of these bacteria is mainly studied using the species Shigella flexneri. S. flexneri pathogenicity is mainly related to a large virulence plasmid that encodes the mxi-spa pathogenicity island: this genetic element encodes proteins composing a T3SS and the effector proteins that are subsequently injected into the host cell (Parsot, 2009). Together, the T3SS and the translocated bacterial effectors are the main elements of bacterial invasion, survival and evasion of the host immune system. The S. flexneri T3SS is assembled during bacterial growth at 37ºC, however it is not active as the translocators and some effector are kept in the bacterial cytoplasm. However, within seconds upon bacterial contact with host cell membranes, the T3SS is activated and the translocators IpaB and IpaC are secreted through the needle complex to form a translocation pore (translocon) in the host plasma membrane, through which a cocktail of about 25-30 bacterial effectors can be injected into the host cytoplasm (Enninga et al., 2005; Parsot, 2009). Only recently the cellular mechanisms eliciting the initial contact between S. flexneri and the host cell have been described. It was shown that the bacteria establish contacts and are captured by filopodial-like structures that emanate from the host cell, in a T3SS-dependent manner. These structures then retract, bringing the bacteria into contact with the host cell membrane, where invasion occurs (Romero et al., 2011). Then, S. flexneri adhesion to the host cell surface is facilitated by the interaction between the bacterial IpaB and CD44 receptor, which is present on the epithelial cell membrane in lipid raft-enriched areas, thus stimulating site-specific invasion (Skoudy et al., 2000). After contact with the host cell there is the formation of a translocon complex, which delivers a subset of effectors through the needle complex, both to the host cell membrane and into the host cytoplasm. These effectors include IpaA, IpaB, IpaC, IpgB1, IpgB2, IpgD and 78 Recent data also suggest the existence of a novel membrane rufflingindependent mechanism for Salmonella entry, which operates independently of Arp2/3 complex. It was shown that this entry mechanism relies on myosin-II contractility at invasion sites, through the activation of RhoA/Rho kinase signaling pathway, in a SopB-dependent manner (Hänisch et al., 2010; 2011). Additionally, it was recently shown that Salmonella can invade epithelial cells in a T3SSindependent manner, which might be important under certain conditions. At least two non-fimbrial outer membrane proteins (Omps), Rck and PagN, can mediate bacterial “zipper”-like invasion of non-phagocytic cells. Rck-mediated entry is dependent on Arp2/3 complex, Rac1 and Cdc42, which induce actin rearrangements that result in membrane ruffling and bacterial internalization (Mijouin et al., 2012; Rosselin et al., 2010). PagN binds to cell surface heparin sulfate proteoglycans, mediating Salmonella adhesion and invasion (Lambert and Smith, 2009). It is likely that these alternate invasion mechanisms, when used in combination with T3SS-1-mediated entry, might help provide cell and host specificity. B. Bacteria exploit the host intracellular environment to survive After internalization, invasive bacterial pathogens are surrounded by a host cell membrane and are enclosed within a membrane-bound vacuole, which is thought to derive mostly from the plasma membrane. Since the host cell has efficient ways to destroy invading bacteria, the latter have evolved diverse strategies to survive and proliferate within the host. Many intracellular bacterial pathogens modulate their vacuoles by hijacking the host cell endomembrane system: they can either rupture the vacuole to reach the host cell cytoplasm (cytosolic lifestyle), or build a replicative niche inside the vacuole (intravacuolar lifestyle) (Figure 18). Each lifestyle has important consequences for both the infected cell and the invading pathogen, since the different niches have distinct physiological environments. By rupturing their vacuole, cytosolic bacterial pathogens, such as L. monocytogenes, S. flexneri, and Francisella tularensis, have direct access to the cytoplasm, which is a rich source of nutrients. 79 Nevertheless, they face other challenges, notably the cytoplasmic innate immune system that can induce inflammatory responses and autophagy-mediated bacterial degradation (Fredlund and Enninga, 2014; Ray et al., 2009). In contrast, intravacuolar bacteria, such as Salmonella, L. pneumophila, M. tuberculosis and Chlamydia trachomatis, must actively avoid vacuolar maturation into a degradative compartment, as well as subvert host cellular trafficking to obtain nutrients. However, the formation of a compartmentalized and unique pathogen-containing organelle allows them to avoid most of the host cell defense mechanisms (Cossart and Helenius, 2014; Creasey and Isberg, 2014; Kumar and Valdivia, 2009). Even though each bacterial pathogen usually adopts a specific intracellular lifestyle, recent studies have described that some intravacuolar bacteria, namely Salmonella and M. tuberculosis, can also access the host cytoplasm (Fredlund and Enninga, 2014; Malik-Kale et al., 2011; Simeone et al., 2012). Listeria and Shigella rupture the vacuole and grow in the host cytoplasm Escape from the bacteria-containing vacuole to the host cytoplasm is a crucial step in the life cycle of cytosolic invasive bacteria (Figure 24). Most of these pathogens are found free in the cytoplasm within 30 minutes after entry into the host cell, which means that vacuolar rupture occurs very rapidly. For many years it has been thought that this process was exclusively driven by bacterial effectors, but it is now clear that vacuolar rupture depends on the intricate interplay between host and bacterial factors (Fredlund and Enninga, 2014; Ray et al., 2009; 2010). Figure 24 – The intracellular lifestyle of cytosolic bacteria, such as L. monocytogenes and S. flexneri (Ray et al., 2009). Bacteria Entry Internalization in primary vacuole Secretion of bacterial escape proteins Disruption of vacuole membrane Bacteria free in the cytosol Actin polymerization Cell–cell spread Internalization in secondary vacuole Secretion of bacterial escape proteins Disruption of double-membrane vacuole Bacteria free in the cytosol Cell cytosol Replication Replication Cell cytosol Plasma membrane Plasma membrane Nature Reviews | Microbiology Cell protrusion Zipper mechanism A mechanism used by bacteria, such as the genera Yersinia or Listeria, to enter cells. Bacteria contact and adhere to the eukaryotic cell through the binding of a bacterial surface protein to a eukaryotic surface receptor, often a transmembrane cell-adhesion protein. Modest membrane extensions and cytoskeletal rearrangements engulf the bacterium in an entry vacuole. Vacuole A single-membrane organelle within the cell cytosol that encloses a fluid-filled compartment. Phagolysosome A membrane-enclosed organelle formed by the fusion of a lysosome, which is an organelle containing hydrolytic enzymes, and a phagosome, which is a membranous vacuole formed around a particle. Lysosome A membrane-bound organelle that contains hydrolytic enzymes. Microbicidal An activity that is lethal for microorganisms. infected cells from damage24,25. Furthermore, LLO is optimally active only at the acidic conditions (pH 5.5) of the vacuole5,11,20,25. As another level of control, LLO is activated by a host factor, γ-interferon-inducible lysosomal thiol reductase, only within the vacuole26; mice lacking this enzyme are resistant to L. monocytogenes infection owing to delayed escape of the pathogen from the vacuole26,27. By contrast, the mechanisms that S. flexneri uses to escape the vacuole are not fully understood. The type III secretion system (T3SS) effector IpaB is required for vacuolar escape in macrophages and, similarly to LLO, displays haemolytic activity15,28. The T3SS functions as a molecular syringe through which bacteria can inject proteins directly into eukaryotic cells. IpaB assembles with IpaC into a pore complex that binds cholesterol and inserts into cell membranes during invasion, whereas another effector, IpaD, enhances the efficiency of insertion of the complex29,30. Therefore, the IpaB–IpaC pore complex may be involved in disruption of the vacuolar membrane and subsequent bacterial escape. However, the function of IpaB and other effectors during vacuolar escape in epithelial cells cannot be confirmed, as they are also required for invasion28. It is also possible that other bacterial factors are involved in lysis. S. flexneri possesses a large virulence plasmid (LVP) that encodes all the genes required for cell invasion3,31,32. It has been suggested that some genes on the LVP that are not essential for cell entry are involved in vacuole lysis33,34. For example, the role of the LVP protein IpaH7.8 in vacuole lysis has yet to be clarified: ipaH7.8 has been reported to facilitate escape from vacuoles in both murine and human macrophages33, whereas other data suggest that ipaH7.8 mutants escape efficiently from vacuoles in epithelial cells34. There is little information about how Rickettsia spp. and B. pseudomallei escape from the vacuole. BPSS1539, a protein of unknown function, does not have a role during invasion of epithelial cells by B. pseudomallei, but seems to facilitate bacterial escape35. Rickettsia spp. produce haemolysin C and phospholipases, which, at least for Rickettsia prowazekii, seem to play a part in escape from Table 1 | Mechanisms of escape from the vacuole by cytosolic bacteria Shigella flexneri Listeria monocytogenes Burkholderia pseudomallei Francisella tularensis Rickettsia spp. Bacterial gene required IpaB28; Mxi–Spa T3SS4 LLO15–17 and type C phospholipases18 Unknown; BPSS1539 mutants are trapped in a vacuole35; Bsa T3SS121 IglC and MglA42; FTT1103 (REF. 43) Phospholipases36–39 and haemolysin C36 Host factors required Unknown GILT activates LLO by a thiol reductase mechanism26 Unknown Unknown Unknown Vacuole conditions required Unknown Acidic vacuole (pH 5.5)5,20 Unknown Acidic vacuole10 Unknown Kinetics of escape 15–30 minutes9 17 minutes5,7 Unknown 30–60 minutes10 12 minutes for Rickettsia coronii12 Actin-based motility Yes47 (using IcsA) Yes47 (using ActA) Yes47 (using BimA) No Yes47 (using RickA) GILT, γ-interferon-inducible lysosomal thiol reductase; LLO, listeriolysin O; T3SS, type III secretion system. Figure 1 | The intracellular lifestyle of cytosolic pathogens. During entry into the host cell, bacteria are engulfed in a primary vacuole. Once inside the vacuole, bacteria secrete proteins that facilitate escape from the vacuole by disrupting the vacuolar membrane. Bacteria replicate once free in the cytosol. With the exception of Francisella tularensis, all cytosolic bacteria polymerize actin at the bacterial pole and are therefore capable of intracellular and intercellular motility. During cell-to-cell spread, bacteria are enclosed in a secondary double-membrane vacuole. Bacteria secrete proteins that disrupt both membranes, allowing the bacteria to escape into the cytosol of an adjacent cell. The bacteria then replicate and continue their intercellular spread, disseminating the infection. REVIEW S 334 | MAY 2009 | VOLUME 7 www.nature.com/reviews/micro REVIEW S 80 In the case of L. monocytogenes, vacuolar rupture occurs 15 minutes upon internalization, and its mechanism has been studied in some detail. It has been proposed to be mediated mostly by the bacterial effectors listeriolysin O (LLO) and type C phosphatases. LLO is a secreted effector that oligomerizes and inserts into cholesterol-enriched domains of the vacuolar membrane, forming pores that destabilize and rupture the vacuole. In this way, L. monocytogenes avoids fusion with lysosomes. LLO production is tightly regulated, and this enzyme is optimally active only when the vacuole is slightly acidified (pH 5.5) (Beauregard et al., 1997; Hamon et al., 2012). LLO activity is also regulated by host factors, such as the γinterferon-inducible lysosomal thiol reductase (GILT) (Singh et al., 2008). Moreover, L. monocytogenes secretes phosphatidylinositol-phospholipase C (PIPLC) and phosphatidylcholine-phospholipase C (PC-PLC), which are required for vacuolar escape into the cytosol, and disruption of the double-membrane vacuole during cell-to-cell spread, respectively (Marquis and Hager, 2000; Smith et al., 1995). After internalization, S. flexneri also escapes rapidly from the vacuole into the host cytoplasm. Several recent studies that allow tracking of vacuolar rupture in real time, by fluorescence microcopy, revealed that this process occurs in less than 10 minutes after bacterial entry (Ehsani et al., 2012; Paz et al., 2010; Ray et al., 2010). Over the last years, novel insights have been made into the rupture of the S. flexneri containing vacuole, which is changing the previous paradigm of a bacterial-induced mechanism. It was initially proposed that T3SS effectors could play a role in vacuolar rupture, within macrophages. In particular, it was proposed that the T3SS translocon proteins, IpaB and IpaC, could form a pore complex in cholesterol-rich membrane domains of the vacuole, causing its disruption and subsequent bacterial escape to the cytosol (Hayward et al., 2005; High et al., 1992). However, the role of these bacterial effectors during vacuolar escape in epithelial cells was never confirmed, since they are also required for invasion (High et al., 1992; Ogawa et al., 2008). After vacuolar rupture and bacterial escape to the cytoplasm, there is the formation of a pool of smaller vesicles derived from vacuolar membrane remnants, whose associated proteins are polyubiquitinated. Then, the autophagy marker microtubule-associated protein light chain 3 (LC3) and the adaptor p62 are recruited, as well as inflammasome components and caspase-1, and the membrane remnants are targeted to autophagic degradation 81 (Dupont et al., 2009; Ehsani et al., 2012). This suggests that the location of vacuolar rupture constitutes a site for the establishment of a signaling platform within the host cell. Recently, a high-content small interfering RNA (siRNA) study showed that several host proteins induce S. flexneri vacuolar membrane rupture (Mellouk et al., 2014), specifically Rab11, a component of the host ERC. It was shown that the conversion of PtdIns(4,5)P2 to PtdIns(5)P, by the bacterial effector IpgD, results in the recruitment of Rab11-positive vesicles to the invasion site and around the bacteria-containing vacuole, before its rupture. In this way, S. flexneri promotes vacuolar rupture through the modulation of host vesicles. Figure 25 – Intracellular actin comet tails induced by L. monocytogenes (top) and S. flexneri (bottom). The left panels show electron microscopy data (Gouin et al., 2005), and the right panels show fluorescently labeled actin comet tails (Egile et al., 1999; Gouin et al., 2004). Once in the host cytoplasm, L. monocytogenes and S. flexneri subvert the host cytoskeleton machinery by expressing effectors that induce actin polymerization at the surface of one bacterial pole. In this way, bacteria induce the formation of a so-called actin comet tail that propels bacteria within the host cytoplasm, moving them intraand intercellularly. Intracellular bacterial motility can reach speeds of 10-87 µm/minute, for L. monocytogenes, or 3-26 µm/minute, for S. flexneri (Ray et al., 2009; Stevens et al., 2006). Structurally, actin comet tails formed by these two bacterial pathogens are composed of short, highly branched crosslinked actin filaments that can leave a long trail behind one of the bacterial poles (Figure 25). In the case of L. monocytogenes, the effector ActA is expressed at one of the bacterial poles, and it directly interacts with and activates the Arp2/3 complex (Figure 26). Thus, ActA mimics the activity of N-WASP, 38 Host–microbe interactions: bacteria Figure 2 Shigella Vaccinia Rickettsia Electron microscopy of actin tails analyzed after myosin S1 decoration. For Shigella, Listeria and Rickettsia, see [41]; reprinted with permission. Copyright 1999 The company of biologists Ltd. For vaccinia see [85]; rep rinted with permiss ion, C op yright 1995 Nature publis hing group. Strikingly, Rickettsia tails are made of long and unbranched filaments. It is not the case for Listeria, Shigella tails and vaccinia pedestals which display the characteristic Y branched structures. Figure 3 38 Host–microbe interactions: bacteria Figure 2 Shigella Vaccinia Rickettsia Electron microscopy of actin tails analyzed after myosin S1 decoration. For Shigella, Listeria and Rickettsia, see [41]; reprinted with permission. Copyright 1999 The company of biologists Ltd. For vaccinia see [85]; rep rinted with permiss ion, C op yright 1995 Nature publis hing group. Strikingly, Rickettsia tails are made of long and unbranched filaments. It is not the case for Listeria, Shigella tails and vaccinia pedestals which display the characteristic Y branched structures. Figure 3 L.monocytogenes S.flexneri 38 Host–microbe interactions: bacteria Figure 2 Shigella Vaccinia Rickettsia Electron microscopy of actin tails analyzed after myosin S1 decoration. For Shigella, Listeria and Rickettsia, see [41]; reprinted with permission. Copyright 1999 The company of biologists Ltd. For vaccinia see [85]; reprinted with permission, C opyright 1995 Nature publishing group. Strikingly, Rickettsia tails are made of long and unbranched filaments. It is not the case for Listeria, Shigella tails and vaccinia pedestals which display the characteristic Y branched structures. Figure 3 ons: bacteria Vaccinia ils analyzed after myosin S1 decoration. For Shigella, Listeria and Rickettsia, see [41]; reprinted with permission. of biologists Ltd. For vaccinia see [85]; reprinted with permission, Copyright 1995 Nature publishing group. ade of long and unbranched filaments. It is not the case for Listeria, Shigella tails and vaccinia pedestals Ybranchedstructures. l. 38 Host–microbe interactions: bacteria Figure 2 Shigella Vaccinia Rickettsia Electron microscopy of actin tails analyzed after myosin S1 decoration. For Shigella, Listeria and Rickettsia, see [41]; reprinted with permission. Copyright 1999 The company of biologists Ltd. For vaccinia see [85]; reprinted with permission, C opyright 1995 Nature publishing group. Strikingly, Rickettsia tails are made of long and unbranched filaments. It is not the case for Listeria, Shigella tails and vaccinia pedestals which display the characteristic Y branched structures. Figure 3 mplex in which the equilibrium Kdof omolar range. The activation reached er than with Cdc42. Interestingly, IcsA fect the polymerization of actin alone, ns as low as 0.1 mM, it activated the some extent in the absence of N-WASP ig. 4 e), which indicates that IcsA interomplex. In the presence of N-WASP, true triangular complex is formed beteins at the surface of Shigella, enhanche edifice. of N-WASP Binds G-actin and mbly in a Profilin-like Fashion SP was described as a filament depolysevering protein (Miki et al., 1996, ., 1998). These conclusions do not apthe present work. The interaction of actin was examined in steady state and In the presence of 5 mM VCA, a modest s concentration plots from 0.12 to 0.21 mM tin was observed, while a 3.5-fold larger sh unassembled actin) was expected, using 0.8 mM. These results are reminiscent of filin in actin assembly (Pantaloni and C they demonstrate that when barbed ends centration of ATP–G-actin at steady sta VCA, suggesting that VCA–actin compl actin, participates in barbed end assembl the experiment shown in Fig. 5 c, the amo bled actin, [A]0, at steady state is 0.21 mM utes into free G-actin, [A], and VCA [VA]. The concentrations of free G-actin are linked by the law of mass action: ( [VA]) / [VA] 5Kd); and by the sum: [A] resulting in a quadratic equation that has [A] 50.03 mM, [VA] 50.18 mM. The free actin, thus, appears fourfold lower th value that is measured in the absence of V Figure 3. tion of N-W complex i fected wit actin tails cells were munolabe anti-N-WA tibodies ( anti-Arp3 bodies (C phalloidin B represe optical sec represent a tion close Texas redantibody w both immu calization bacterium cated by a 10 mm. 82 functioning as a NPF (Gouin et al., 2005; Stevens et al., 2006). It has been proposed, from in vitro reconstitution experiments, that L. monocytogenes propulsion in the host cytoplasm might require an initial Arp2/3 complexdependent nucleation step, followed by an Arp2/3 complex-independent tailelongation phase that necessitates fascin and results in the formation of F-actin bundles (Brieher et al., 2004). As for S. flexneri, it induces the formation of actin comet tails by expressing the outer membrane protein IcsA (also called VirG) at one pole. IcsA directly interacts with and recruits a host cell NPF (N-WASP), which in turn activates the Arp2/3 complex to form an actin comet tail (Figure 26) (Stevens et al., 2006). Figure 26 – Mechanisms of intracellular bacterial actin-based motility, induced by L. monocytogenes and S. flexneri. Listeria ActA directly interacts with Arp2/3 complex, through the central and acidic (CA) regions. In this way there is Arp2/3 activation. Shigella IcsA recruits N-WASP to the bacterial surface, through a glycinerich domain, which in turn activates the Arp2/3 complex. Adapted from Stevens et al., 2006. Bacterial escape from the endocytic vacuole to the cytoplasm allows them to escape degradation within the endolysosomal pathway. Nevertheless, the host cytoplasm also contains a range of innate immune defense mechanisms, such as antimicrobial peptides and the Nod-like receptors (NLRs). An additional component of the host defense against cytosolic pathogens is autophagy, a degradative pathway by which cytosolic contents, organelles and pathogens are delivered to lysosomes as part of cellular homeostasis (He and Klionsky, 2009). Though, cytosolic bacteria have developed strategies to interact and modify the autophagic pathway, in order to promote their survival. Once in the host cytoplasm, L. monocytogenes avoid autophagic degradation by expressing the phospholipases PI-PLC and PL-PLC, by a mechanism that remains to be © 2006 Nature Publishing Group Actin L. monocytogenes ActA R. conorii RickA S. flexneri IcsA B. pseudomallei BimA Proline-rich motif WH2 (W) domain (binds actin monomers) CA domain (binds/ activates Arp2/3) Glycine-rich repeats Arp2/3 N-WASP Src homology-3 (SH3) domain A small module of ~50 aminoacid residues found in proteins that interact with proline-rich motifs. Focal adhesion sites Adhesions by which cells attach to the underlying substrate. Many structural, cytoskeletal and signalling proteins are concentrated in these structures. Stress fibres Contractile filaments involved in the maintenance of cell shape that support the motile apparatus of the cell. At least one end is anchored by a focal adhesion site. Central and acidic (CA) domain Forms part of a larger domain called a WCA domain. This region of WASP-family proteins stimulates the Arp2/3 complex. WASP homology-2 (WH2) domain WASP homology-2 domain motifs are composed of approximately 35 amino acids and are conserved in cellular proteins that recruit actin monomers. it is thought that ActA functions as an NPF, directly activating the Arp2/3 complex by mimicking the activity of WASP. In infected cells, immunocytochemistry has shown that the ActA protein is asymmetrically distributed on the bacterial surface, and is localized at the site of actintail formation9. The mature ActA protein is expressed on the bacterial surface as a protein comprising 610 amino acids with three functional domains (FIG. 2). The N-terminal domain, composed of the first 233 aminoacid residues, contains all the necessary residues for motility in cultured cells and cell-free extracts28,29 and can stimulate the activity of the Arp2/3 complex in vitro17. The central domain (amino acids 234–394), which contains proline/glutamic-acid-rich repeats reminiscent of Src homology-3 (SH3) domains, is not essential for motility, but deletion of this region shortens actin tails in infected cells and decreases the velocity of movement in Xenopus oocyte extracts28. Indeed, the number of proline-rich repeats in this region has been correlated with the velocity of movement, with each contributing to ~2.5 µm per minute30. The hydrophobic C-terminal portion of ActA probably constitutes a membrane anchor18, which tethers ActA to the bacterial surface after it secretion. Several host-cell proteins have been localized to the actin tails of L. monocytogenes in infected cells. Their importance in motility has been extensively reviewed elsewhere and is summarized in Supplementary information S1 (table). The proline-rich region and C-terminal portion of ActA shows significant sequence similarity with zyxin, an actin-binding protein that is associated with focal adhesion sites and stress fibres31. Interestingly, although ActA mimics WASP-family proteins in its activation of Arp2/3-dependent actin assembly in host cells, ActA and zyxin also harbour similar Arp2/3-independent actin-polymerization activities in vitro32. From in vitro reconstitution experiments, it has been proposed that Listeria propulsion in host cells might involve an initial Arp2/3-dependent nucleation step followed by an Arp2/3-independent tail-elongation phase that requires the actin-filament-bundling protein fascin33. Rickettsia species. Rickettsia are Gram-negative obligate intracellular bacteria that cause arthropod-borne diseases of humans, including typhus (Rickettsia prowazekii and Rickettsia typhi) and spotted-fevers (Rickettsia conorii, Rickettsia rickettsii, Rickettsia sibirica and Rickettsia montana). Rickettsia species of the spotted-fever group show actin-based motility in the cytosol and nucleus of host cells at 5–8 µm per minute and are capable of intercellular spread34–37. Comparison of the complete genome sequence of R. conorii with that of R. prowazekii, which cannot use actin-based motility, identified a 2-kb R. conorii-specific region that encodes a predicted protein of 517 amino acids known as RickA38,39. RickA contains a central proline-rich domain and a C-terminal WH2 domain followed by a region with homology to the central and acidic (CA) domains of WASP-family proteins, including an amphipathic helix predicted to bind the Arp2/3 complex39 (FIG. 2). The RickA protein lacks predicted signal sequences for secretion or obvious hydrophobic domains that might serve as a membrane anchor. Separate analysis of the R. rickettsii genome for WASP-like proteins identified a 494-amino-acid RickA homologue39. RickA homologues are also found in the genomes of R. sibirica and R. montana and share the same general organization, although there are differences in the number and sequence of proline-rich repeats and C-terminal WASP homology-2 (WH2) domains. RickA proteins from R. conorii, R. rickettsii and R. sibirica have a single WH2 domain similar to WASP, whereas R. montana RickA Figure 2 | Mechanisms of intracellular bacterial actin polymerization. Schematic diagram depicting the proposed mechanisms by which intracellular pathogens recruit and activate actin-related protein-2/3 (Arp2/3)-dependent actin polymerization at the bacterial surface. Listeria monocytogenes ActA directly interacts with, and activates, the Arp2/3 complex. The Wiskott–Aldrich syndrome protein (WASP) homology-2 (WH2) domain binds to actin monomers, and the central and acidic (CA) regions bind to, and activate, Arp2/3. Shigella flexneri IcsA functions as a Cdc42 mimic, recruiting neural (N-)WASP to the bacterial surface through a glycine-rich region, which in turn recruits and activates the Arp2/3 complex. Rickettsia conorii RickA interacts with, and directly activates, the Arp2/3 complex in a manner similar to that of ActA; however, Rickettsia-induced actin tails lack the Y-branches seen in Listeria-induced tails, and other bacterial or cellular factors might organize the actin filaments into unbranched arrays. The structure of Burkholderia pseudomalleiinduced actin tails and the molecular mechanism by which they form are unclear. The schematic reflects the tertiary structure of the proteins. REVIEWS 94 | FEBRUARY 2006 | VOLUME 4 www.nature.com/reviews/micro © 2006 Nature Publishing Group Actin L. monocytogenes ActA R. conorii RickA S. flexneri IcsA B. pseudomallei BimA Proline-rich motif WH2 (W) domain (binds actin monomers) CA domain (binds/ activates Arp2/3) Glycine-rich repeats Arp2/3 N-WASP Src homology-3 (SH3) domain A small module of ~50 aminoacid residues found in proteins that interact with proline-rich motifs. Focal adhesion sites Adhesions by which cells attach to the underlying substrate. Many structural, cytoskeletal and signalling proteins are concentrated in these structures. Stress fibres Contractile filaments involved in the maintenance of cell shape that support the motile apparatus of the cell. At least one end is anchored by a focal adhesion site. Central and acidic (CA) domain Forms part of a larger domain called a WCA domain. This region of WASP-family proteins stimulates the Arp2/3 complex. WASP homology-2 (WH2) domain WASP homology-2 domain motifs are composed of approximately 35 amino acids and are conserved in cellular proteins that recruit actin monomers. it is thought that ActA functions as an NPF, directly activating the Arp2/3 complex by mimicking the activity of WASP. In infected cells, immunocytochemistry has shown that the ActA protein is asymmetrically distributed on the bacterial surface, and is localized at the site of actintail formation9. The mature ActA protein is expressed on the bacterial surface as a protein comprising 610 amino acids with three functional domains (FIG. 2). The N-terminal domain, composed of the first 233 aminoacid residues, contains all the necessary residues for motility in cultured cells and cell-free extracts28,29 and can stimulate the activity of the Arp2/3 complex in vitro17. The central domain (amino acids 234–394), which contains proline/glutamic-acid-rich repeats reminiscent of Src homology-3 (SH3) domains, is not essential for motility, but deletion of this region shortens actin tails in infected cells and decreases the velocity of movement in Xenopus oocyte extracts28. Indeed, the number of proline-rich repeats in this region has been correlated with the velocity of movement, with each contributing to ~2.5 µm per minute30. The hydrophobic C-terminal portion of ActA probably constitutes a membrane anchor18, which tethers ActA to the bacterial surface after it secretion. Several host-cell proteins have been localized to the actin tails of L. monocytogenes in infected cells. Their importance in motility has been extensively reviewed elsewhere and is summarized in Supplementary information S1 (table). The proline-rich region and C-terminal portion of ActA shows significant sequence similarity with zyxin, an actin-binding protein that is associated with focal adhesion sites and stress fibres31. Interestingly, although ActA mimics WASP-family proteins in its activation of Arp2/3-dependent actin assembly in host cells, ActA and zyxin also harbour similar Arp2/3-independent actin-polymerization activities in vitro32. From in vitro reconstitution experiments, it has been proposed that Listeria propulsion in host cells might involve an initial Arp2/3-dependent nucleation step followed by an Arp2/3-independent tail-elongation phase that requires the actin-filament-bundling protein fascin33. Rickettsia species. Rickettsia are Gram-negative obligate intracellular bacteria that cause arthropod-borne diseases of humans, including typhus (Rickettsia prowazekii and Rickettsia typhi) and spotted-fevers (Rickettsia conorii, Rickettsia rickettsii, Rickettsia sibirica and Rickettsia montana). Rickettsia species of the spotted-fever group show actin-based motility in the cytosol and nucleus of host cells at 5–8 µm per minute and are capable of intercellular spread34–37. Comparison of the complete genome sequence of R. conorii with that of R. prowazekii, which cannot use actin-based motility, identified a 2-kb R. conorii-specific region that encodes a predicted protein of 517 amino acids known as RickA38,39. RickA contains a central proline-rich domain and a C-terminal WH2 domain followed by a region with homology to the central and acidic (CA) domains of WASP-family proteins, including an amphipathic helix predicted to bind the Arp2/3 complex39 (FIG. 2). The RickA protein lacks predicted signal sequences for secretion or obvious hydrophobic domains that might serve as a membrane anchor. Separate analysis of the R. rickettsii genome for WASP-like proteins identified a 494-amino-acid RickA homologue39. RickA homologues are also found in the genomes of R. sibirica and R. montana and share the same general organization, although there are differences in the number and sequence of proline-rich repeats and C-terminal WASP homology-2 (WH2) domains. RickA proteins from R. conorii, R. rickettsii and R. sibirica have a single WH2 domain similar to WASP, whereas R. montana RickA Figure 2 | Mechanisms of intracellular bacterial actin polymerization. Schematic diagram depicting the proposed mechanisms by which intracellular pathogens recruit and activate actin-related protein-2/3 (Arp2/3)-dependent actin polymerization at the bacterial surface. Listeria monocytogenes ActA directly interacts with, and activates, the Arp2/3 complex. The Wiskott–Aldrich syndrome protein (WASP) homology-2 (WH2) domain binds to actin monomers, and the central and acidic (CA) regions bind to, and activate, Arp2/3. Shigella flexneri IcsA functions as a Cdc42 mimic, recruiting neural (N-)WASP to the bacterial surface through a glycine-rich region, which in turn recruits and activates the Arp2/3 complex. Rickettsia conorii RickA interacts with, and directly activates, the Arp2/3 complex in a manner similar to that of ActA; however, Rickettsia-induced actin tails lack the Y-branches seen in Listeria-induced tails, and other bacterial or cellular factors might organize the actin filaments into unbranched arrays. The structure of Burkholderia pseudomalleiinduced actin tails and the molecular mechanism by which they form are unclear. The schematic reflects the tertiary structure of the proteins. REVIEWS 94 | FEBRUARY 2006 | VOLUME 4 www.nature.com/reviews/micro 83 addressed (Birmingham et al., 2007; Py et al., 2007). In the case of S. flexneri, the IcsA effector, which is required for actin-based motility, also induces autophagy in the host cytoplasm, by binding the autophagic protein Atg5 (Ogawa et al., 2005). Moreover, septins are recruited to the IcsA-mediated actin comet tail, forming cage-like structures that prevent intracellular bacterial motility, entrapping and targeting S. flexneri to autophagic degradation (Mostowy and Cossart, 2011; Mostowy et al., 2010). To evade autophagy, S. flexneri secretes the T3SS effector IcsB, which competitively binds to IcsA and prevent Atg5-dependent recognition and septin cage entrapment (Mostowy et al., 2010; Ogawa et al., 2005). Legionella builds a replicative vacuole inside host cells Legionella pneumophila, the causative agent of the potentially fatal pneumonia Legionnaires’ disease, is an accidental human pathogen that replicates intracellularly within environmental protozoa and alveolar macrophages (Hilbi et al., 2011; Isberg et al., 2009). After being phagocytosed, L. pneumophila secretes around 300 effectors into the host cell, via the Dot/Icm type IV secretion system (T4SS), some of which drive maturation of the Legionella-containing vacuole (LCV). Different cellular microbiology studies and proteomic analysis of purified LCVs revealed that this unique compartment intercepts the secretory vesicle trafficking pathway, interacts with the ER and communicates with the endosomal pathway without fusing with lysosomes (Hilbi and Haas, 2012; Hoffmann et al., 2014; Hubber and Roy, 2010; Isberg et al., 2009; Urwyler et al., 2009). Maintenance of the LVC is Dot/Icm-dependent, because non-virulent Δdot/icm mutants escape to the cytosol of infected cells soon after invasion (Molmeret et al., 2007). By recruiting vesicles derived from the ER, the LCV matures into a specialized ER-like compartment that supports bacterial growth (Figure 27). L. pneumophila has the ability to subvert the functions of the GTPase Rab1 (a ER-toGolgi traffic regulator) and of the v-SNARE Sec22b, in order to facilitate the transport and fusion of ER-derived vesicle with the LCV (Kagan et al., 2004). Additional evidence indicating that vesicles that exit the ER fuse with the LCV includes the presence of ER-derived proteins on the vacuole, such as glucose-6phosphatase and protein disulphide isomerase, as detected by electron microscopy (Robinson and Roy, 2006). 84 Figure 27 – The L. pneumophila intracellular lifestyle. A. After uptake into amoebae or macrophages, the LCV escapes transport to the endolysosomal pathway and it interacts with mitochondria and ER-derived vesicles. Eventually, the LCV matures into an ER-like compartment, covered with ribosomes, where the bacteria replicate. B. Several Dot/Icm T4SS effectors associate with the LCV and recruit host proteins involved in vesicle trafficking, such as Sec22b, Rab1 and Arf1. The host proteins Atg7 and Atg8 also associate with the LCV, possibly as a result of autophagic membrane recruitment to the vacuole. Adapted from Isberg et al., 2009. The LCV is also decorated with host small GTPases involved the regulation of secretory or endosomal trafficking (Arf1, Rab4, Rab5, Rab7, Rab8 and Rab11), whose activity can be regulated by bacterial effectors (Hoffmann et al., 2014; Kagan and Roy, 2002; Urwyler et al., 2009). Additionally, L. pneumophila utilizes PtdIns(4)P and PtdIns(3)P on the LCV membrane to anchor bacterial effectors to the vacuole (Hilbi and Haas, 2012; Ragaz et al., 2008), and interferes with the !! ! ! ! A B 85 retrograde and lysosomal pathways. The Dot/Icm T4SS effector RidL, which localizes to the LCV membrane, binds the subunit Vps29 of the retromer complex, inhibiting retrograde endosome-Golgi trafficking and promoting intracellular bacterial replication (Finsel et al., 2013), and the effector SidK, which inhibits vATPase, prevents LCV acidification (Xu et al., 2010). Since L. pneumophila uses a T4SS to inject effectors across host membranes, it is likely that it could also cause LCV membrane damage and compromise its integrity. Even though it has been described that this bacterium escapes into the host cytoplasm late in infection, this still needs further investigation (Kumar and Valdivia, 2009; Molmeret et al., 2004). Interestingly, it was recently shown that the bacterial effector SdhA is required to maintain the integrity of the LCV membrane in macrophages, which allows intracellular bacterial replication (Creasey and Isberg, 2012). Salmonella life inside a vacuole – nesting to grow After internalization into the host cell, Salmonella reside within the Salmonellacontaining vacuole (SCV). This is a unique and modified membrane-bound compartment that, through the action of both T3SS-1 (SPI-1 effectors) and T3SS-2 (SPI-2 effectors), enables intravacuolar bacterial survival and replication. The SCV is subjected to different stages of maturation (early, maturing, and late), each associated with the presence of different host cell factors on the vacuole (MalikKale et al., 2011). It has been shown that, immediately after formation, the early SCV [< 30 minutes post-invasion (p.i.)] shares some similarities with early endosomes and undergoes rapid membrane remodeling (Figure 28). This is mostly driven by the activities of the T3SS-1 effector SopB (a homologue of the S. flexneri effector IpgD), which is translocated during bacterial entry and then persists in the host cell for several hours (Drecktrah et al., 2005; Kubori and Galán, 2003). SopB is located on the cytosolic side of the SCV, where it recruits the small GTPase Rab5 to the SCV membrane. This in turn recruits the Rab5-interacting protein Vps34, a PI(3)K that phosphorylates PtdIns into PtdIns(3)P on the SCV membrane, which is necessary for the recruitment of EEA-1 and VAMP8 (Dai et al., 2007; Mallo et al., 2008; Scott et al., 2002). It has been suggested that the presence of PtdIns(3)P on the SCV stimulates fusion with other vesicles containing PtdIns(3)P, which might 86 bring nutrients required for bacterial growth (Hernandez et al., 2004). SopD is another T3SS-1 effector that may act cooperatively with SopB during the initial steps of SCV biogenesis (Bakowski et al., 2007). Figure 28 – Biogenesis and maturation of the Salmonella-containing vacuole (SCV), and the bimodal lifestyle of intracellular Salmonella, in epithelial cells. Invasive Salmonella use T3SS1 to inject effectors into the host cell and promote actin-mediated ruffling and bacterial internalization into a modified vacuole, the SCV. The SCV then undergoes successive maturation steps, with extensive remodeling of its membrane. T3SS-2 effectors are required for SCV maturation and for its positioning in a perinuclear location, close to the MTOC. In the figure, the bacterial effectors SopB, SifA and PipB2, are depicted in dark blue. Some SCVs do not undergo maturation, and instead are ruptured, either releasing bacteria into the host cytosol or being targeted by the autophagy system. In the late SCV there is Salmonella replication and the formation of dynamic membrane tubules. For reasons of simplicity, only the Lamp-1-enriched Salmonellainduced filaments (SIFs) are represented. Cytosolic bacteria that have escaped autophagy start to hyper-replicate, and become activated for T3SS-1 and flagella. See text for details.            !"  #$% &' () *()     *()&" +*,-&" &"      #$%  ,. /), 0 0,, 1  ,. ,2 *(3 4( 4(3 5 5 5 5 5 5 5 " " " " " " " " "  ( * 87 Moreover, within 15 minutes p.i., there is the recruitment of SNX1 to the SCV, in a SopB-dependent manner, which promotes exclusion of CI-M6PR from the SCV (Bujny et al., 2008). Since the CI-M6PR is used to deliver soluble lysosomal enzymes to lysosomes and it is generally excluded from the SCV at later timepoints, this was first seen as evidence for the lack of lysosomal fusion with the SCV (Garcia-del Portillo and Finlay, 1995). Nevertheless, this has been controversial (Oh et al., 1996) and recent data clearly indicate that the SCV dynamically interacts with the host cell endolysosomal system (Drecktrah et al., 2007). Therefore, an alternative explanation for the exclusion of the CI-M6PR from the SCV is that it is still recruited to the early SCV but is then efficiently removed in a SNX1-dependent manner (Bujny et al., 2008). This implies that Salmonella might not avoid interactions between the SCV and lysosomes but, alternatively, can control host trafficking pathways to remove unwanted factors. SopB also mediates recruitment of SNX3 to the early SCV, which is important for vacuole maturation (Braun et al., 2010). Therefore, it is likely that SopB-mediated PtdIns modulation on the SCV membrane is crucial for SCV formation and maturation. After the formation of the SCV, the T3SS-2 is induced and there is secretion of SPI-2 effectors across the SCV membrane. Upregulation and delivery of SPI-2 T3SS effectors across the SCV membrane into the host cytoplasm is a precisely controlled process. It involves the ordered assembly of the secretion apparatus on the cell surface, followed by the secretion of effectors that form de needle complex and the translocon pore (formed by SseB, SseC and SseD) (Chakravortty et al., 2005). Activation of genes encoding the secretion apparatus is mediated by twocomponent regulatory systems, in response to the slightly acidic pH (pH of around 5) and poor nutritional status of the SCV lumen. After assembly of the translocon pore on the SCV membrane, the pH of the host cytosol is sensed by an unknown component of the T3SS-2. Subsequently, a SPI-2-encoded regulatory complex in the bacteria dissociates, derepressing the translocation of proteins, and about 30 different bacterial effectors are injected across the SCV membrane into the host cytoplasm (Figueira and Holden, 2012; van der Heijden and Finlay, 2012), which leads to vacuole maturation. The intermediate or maturing SCV (between 30 minutes and 5 hours p.i.) then undergoes extensive membrane remodeling. Rab5 is depleted from the SCV, and this acquires late endosome/lysosomal markers, such as Rab7, Lamp-1 and vATPase (Figure 28). However, the maturing SCV is 94 Subcellular fractionation allows the separation of organelles according to their physical and biological properties. The most usual strategy consists of mechanically disrupting or homogenizing the cell, in a detergentfree isotonic buffer solution, in order to release all the organelles and other cellular constituents as a free suspension of intact individual components. Then, several techniques exploiting the physical and biological parameters of the different compartments can be used to isolate organelles and membranes. Among those, equilibrium ultracentrifugation in a density gradient is considered to be the most effective. In this way, it is possible to separate all the different organelles on the basis of their distinct densities: after reaching equilibrium in the gradient, a specific organelle will be positioned in an area (called “fraction”) where its density matches the density of the surrounding medium. In this way, several organelles can be purified to obtain nearly pure fractions (Graham and Rickwood, 1997). Even though high enrichments can be achieved, many endosomal compartments have similar physical properties. Thus, a particular fraction of interest might contain some degree of contamination from a different organelle. Despite this limitation, subcellular fractionation proved to be crucial for the identification and characterization of several organelles, as exemplified by the groundbreaking work performed by George Palade, Christian de Duve and Albert Claude, during the 1950s and 1960s (Bergeron et al., 2010). Other techniques used to separate subcellular fractions, alone together with centrifugation, include immuno-isolation, affinity electrophoresis, free-flow electrophoresis or direct alteration of the physical properties of an organelle. 1998 2000 2002 2004 2006 2008 2010 2012 2014 0 200 400 600 800 1000 1200 Year Number of publications (organelle OR membrane OR subcellular) AND proteomics Figure 29 – Number of publications per year in the field of organelle proteomics. The PubMed query is shown on top of the figure. 95 B. Mass spectrometry-based proteomics The principles of MS-based proteomics MS is a way to accurately identify the type and amount of a molecule, by measuring its mass-to-charge ratio (m/z). Given that mass analysis uses electromagnetic fields in a vacuum, molecules first have to be subjected to electrospray ionization. In this technique, molecules such as peptides are dissolved in liquid that passes through a needle at high electrical potential. The applied voltage causes the liquid to disperse into small, highly charged droplets, which evaporate and transfer the molecules into the gas phase in an ionized form. After electrospray ionization, electrically charged molecules are transferred into the vacuum of a mass spectrometer and their m/z ratio is determined by their trajectories in an electric field. Popular mass analyzers include, for example, quadrupole-time of flight (TOF) instruments, where the m/z is determined by the time ions need to travel through an electric field to arrive at a detector. Alternatively, ions can be captured in a linear ion trap, where they can be accumulated and manipulated for further analysis. This can be used in combination with an Orbitrap, a type of ion trap mass analyzer where ions oscillate along and around a central spindle-shaped electrode, and that enables extremely accurate m/z measurements and a high mass spectrometric resolution. The high resolution allows the mass spectrometer to distinguish hundreds of thousands of different peptides from each other, which is crucial for their accurate identification and quantification (Walther and Mann, 2010). Nevertheless, it can be difficult to measure the mass of entire proteins (“top-down proteomics”), because the m/z differences of distinct proteins with similar compositions are small. Therefore, for most experiments it is measured the m/z ratio of peptides derived from the entire proteins, after enzymatic cleavage (“bottom-up proteomics”). As the peptides are electrosprayed, the mass spectrometry yields a MS-spectrum of m/z ratios. Then the acquisition software selects a preset number of peptides in the mass spectra and isolates each one of them, to fragment them in the mass spectrometer and to measure the mass spectra of the fragments, which allows peptide sequence identification. This is called tandem mass spectrometry (MS/MS). The data are then scanned through an amino acid sequence database that calculates and 96 predicts a sequence for each peptide, which enables protein identification. The most popular commercially available peptide search engines are Mascot and Sequest (Cox and Mann, 2011; Walther and Mann, 2010). In experiments where it is necessary to determine the protein composition of an organelle or a protein complex, it is crucial to accurately detect specific proteins in complex mixtures. In addition, increasing layers of complexity are generated when proteins are digested to peptides. Thus, several approaches can be used to reduce complexity. Initial techniques involved the separation of soluble proteins by two-dimensional gel electrophoresis (2DGE), followed by protein excision from the gel, digestion with a protease, and analysis by MS techniques, such as matrixassisted laser desorption/ionization-time of flight (MALDI-TOF). However, 2DGE has significant limitations for the separation of transmembrane proteins, and the extraction of hundreds of spots from the gel is extremely time-consuming (Molloy et al., 1998; Yates et al., 2005). Therefore, the most widely method is to separate the proteins by one-dimensional SDS-PAGE (sodium dodecyl sulfatepolyacrylamide gel electrophoresis), which effectively solubilizes membrane proteins, due to the detergent SDS. Proteins can then be subjected to in-gel digestion by a protease, and the remaining peptides are extracted from the gel. Alternatively, it is possible to digest proteins in-solution, avoiding the tedious steps of gel separation and extraction (Walther and Mann, 2010). In these two last examples (SDS-PAGE proteins separation followed by in-gel digestion, or protein digestion in-solution) it is then essential to resolve and detect peptides with high accuracy. To this end, peptides can be separated according to their hydrophobicity, by very low flow high-performance liquid chromatography (HPLC) that is linked directly to a tandem mass spectrometer through electronspary ionization (this technique is called LC-MS/MS, for liquid chromatography-tandem mass spectrometry). This allows extremely high efficiency, yielding peptide spectra that are sufficiently well resolved for a comprehensive detection of hundreds of proteins present in a complex sample (Walther and Mann, 2010; Yates et al., 2005). Novel LC-MS/MS-based proteomics methodologies generate gigabytes of high resolution data per day and per mass spectrometer, which need to be analyzed by computational tools. For this, a computational proteomics workflow has been developed, based on the use of the MaxQuant algorithms for the analysis of large 97 MS datasets. MaxQuant is equipped with its own search engine, called Andromeda, and enables peak detection in the raw data with a high peptide identification rate, as well as peptide quantification (Cox and Mann, 2008; Cox et al., 2011). Quantitative proteomics Usually it is more important to determine how protein levels change between different conditions, than it is to know just whether a protein is present or not. In this regard, quantification is of central importance in MS-based proteomics, and it can determine the absolute amount of each of the proteins in a mixture or their relative change between two or more conditions. There are two main approaches to turn MS quantitative: stable isotope-labeling or label-free methods (Cox and Mann, 2011; Ong and Mann, 2005). One of the labeling approaches that has recently gained popularity is called iTRAQ (isobaric tag for relative and absolute quantitation), where digested peptides are chemically modified with isobaric tags. iTRAQ uses up to eight isobaric tags (even though usually only four tags are used in simultaneous) that react with primary amine groups of peptides. During MS analysis, the tags are further fragmented into low mass reporter ions with a tag-specific mass, and then the relative peak intensities of the different reporter ions are used to derive the relative abundance of the corresponding peptides and proteins (Ong and Mann, 2005; Sadowski et al., 2006; Walther and Mann, 2010). Alternatively, peptides can be metabolically labeled, as in the SILAC (stable isotope labeling by amino acids in cell culture) approach. SILAC procedure consists of growing two cell populations in media. One of the cell populations is fed with normal (called “light”) amino acids, whereas the other population is fed with stable (nonradioactive) “heavy” 13Cand 14N-labeled forms of lysine and/or arginine. When cells grow in this “heavy” medium, cells will integrate the “heavy” amino acids into all proteins in the course of several cell doublings. In this way, after protein digestion and MS analysis, the “heavy” labeled proteome is distinguishable from the “light” control proteome by a characteristic mass shift, and the relative intensity of the peaks reflects the relative abundance of the proteins in the mixture (Ong and Mann, 2005; Walther and Mann, 2010). 98 A label-free quantitative MS methodology is becoming increasingly used. Here, the basic principle is to align and compare the signals of the same eluting peptides from separate LC-MS/MS runs, which allows calculating differences in peak intensities of the same peptides detected in each run. This approach is usually less accurate than the isotope labeling techniques. Nevertheless, in combination with sophisticated algorithms, label-free quantification can be a robust alternative, mostly because it can be applied to any cell type without the need of introducing isotopes (Cox and Mann, 2011; Luber et al., 2010; Walther and Mann, 2010). Insights from organelle proteomics The analysis of organelles using proteomics methods is an active field of research, and significant progress is being made in order to define the proteomes of different organelles. For this, subcellular fractionation approaches can be combined with MS-based proteomics to detect and measure proteins in purified organelles (Yates et al., 2005). The mitochondrial proteome was determined after organelle purification by sucrose gradient fractionation, followed by SDS-PAGE protein separation and LC-MS/MS, which identified hundreds of factors associated with this organelle (Pflieger et al., 2002; Taylor et al., 2002; 2003b). Similarly, proteins associated with the Golgi (Wu et al., 2000), clathrin-coated vesicles (Blondeau et al., 2004; Wasiak et al., 2002), the ER lumen (Knoblach et al., 2003), the ERGIC (Breuza et al., 2004) or peroxisomes (Marelli et al., 2004) were determined after subcellular fractionation and MS. As described in the previous chapters of this thesis, there are complex mechanisms of intracellular communication and contact sites between organelles. This complexity makes it hard to evaluate the biological significance of proteins that are usually associated with one organelle but are detected in the proteome of a different organelle. Additionally, different organelles can have similar physical properties, which means that during biochemical subcellular fractionation they can co-fractionate together, and thus some proteins can be associated with multiple organelle fractions. This shows that not all proteins in a fraction are bona fide constituents of the organelle of interest, but might be instead a result of subcellular fractionation artifacts or contaminants. Nevertheless, it is not possible to exclude that these proteins could also be of biological significant. One of the simplest 99 solutions to the difficulty in obtaining pure organelles fractions is to perform “subtractive proteomics”, in which the inventory of proteins that are found in a control state or fraction are subtracted from an experimental state or organelleenriched fraction (Yates et al., 2005). The remaining proteins only detected in the target fraction are thus considered to be enriched in components of the organelle of interest. It is also possible to combine this with quantitative proteomics, where the relative abundance of proteins present in a target and similar control fractions is compared. A quantitative subtractive proteomics has the main advantage of discriminating between true protein members of the organelle and background hits. It is however important that the two fractions under comparison share as much as possible of technical variability to avoid systematic biases that would invalidate the entire approach (Gatto et al., 2010). C. The proteome of bacteria-containing vacuoles: the Holy Grail More than twenty years ago, scientists started making important contributions to the understanding of the protein composition of phagosomes (Desjardins et al., 1994). Using phagocytes as cellular model, phagosomes containing inert, lowdensity latex beads (latex bead-containing phagosomes, or LBPs) were first isolated by subcellular fractionation. Given that LBPs have a lower density than most other organelles, they float after ultracentrifugation and thus can be isolated with a very high degree of purity. Over the last two decades, several subsequent studies have been performed, which, with the development of more sensitive mass spectrometers, identified hundreds of proteins associated with LBPs (CampbellValois et al., 2012; Garin et al., 2001; Rogers and Foster, 2007; Stuart et al., 2007; Trost et al., 2009). These studies, together with many others, helped to understand phagosome biogenesis and maturation, as well as to identify novel sources of membrane during the process (Li et al., 2010; Rogers and Foster, 2008). Phagocytosed latex beads are an excellent tool to understand normal phagosome maturation, but some of the most interesting aspects of phagosome maturation involve the ability of intracellular bacterial pathogens to manipulate the 100 normal maturation process, as described in chapter 4B. However, in the field of host-pathogen interactions, isolation and protein identification of bacteriacontaining phagosomes is considered a “Holy Grail” in phagosome proteomics, and only a few studies have been successfully performed. One example is the determination of LCV protein composition. Since L. pneumophila, through its T4SS, injects and selectively anchors the effector SidC into the LCV membrane, it is possible to isolate this compartment by using immuno-magnetic separation followed by density gradient centrifugation (Hoffmann et al., 2013; Urwyler et al., 2009). The proteomics analysis of isolated LCVs revealed more than 670 (in amoeba) or 1150 (in macrophages) host proteins (Hoffmann et al., 2014). Many of these factors are Rab GTPases, which corroborates that LCVs communicates with different cellular signaling and vesicle trafficking pathways. It was also possible to isolate enriched fractions containing Mycobacterium bovis bacillus CalmetteGuérin (BCG)-containing phagosomes, through subcellular fractionation methodologies, and determine their proteome composition (Lee et al., 2010). This revealed the presence of 447 host proteins on the BCG-containing phagosomes. Finally, recent work established a novel protocol for the isolation of Salmonellamodified membranes (Vorwerk et al., 2015), which include all host cell membranes modified by activities of intracellular Salmonella (such as the late SCV, SIFs, SISTs and other unknown membrane structures). Using fractionation followed by affinity immuno-precipitation and subtractive LC-MS/MS analysis, the authors identified 247 host proteins uniquely associated with Salmonella-modified membranes. The analysis revealed that these membranes are enriched in proteins derived from the TGN, recycling endosomes, ER and other host organelles. 101 “Ceux qui préfèrent leurs principes à leur bonheur. Ils refusent d'être heureux en dehors des conditions qu'auparavant ils ont fixées à leur bonheur. S'ils le sont, par surprise, les voilà désemparés - malheureux d'être privés de leur malheur.” – Albert Camus; in Carnets III: Mars 1951-Décembre 1959 102 103 PART II GOALS OF THE PROJECT 110 111 Manuscript 1 – “Hierarchies of host factor dynamics at the entry site of Shigella flexneri during host cell invasion” Soudeh Ehsani1, José Carlos Santos1,2, Cristina D. Rodrigues1, Ricardo Henriques3, Laurent Audry1, Christophe Zimmer3, Philippe Sansonetti4,5, Guy Tran Van Nhieu6,7, Jost Enninga1 1. Unit of Dynamics of Host-Pathogen Interactions, Institut Pasteur, Paris, France 2. Graduate Program in Areas of Basic and Applied Biology (GABBA), Universidade do Porto, Portugal 3. Group “Imagerie et Modélisation”, Institut Pasteur, Paris, France; CNRS URA 2582 4. Unit “Pathogénie Microbienne Moléculaire”, Institut Pasteur, Paris, France 5. INSERM unit 789, Paris, France 6. Interdisciplinary research group “Intercellular communication of microbial infection”, College de France, Paris, France 7. INSERM unit 1050, Paris, France Contribution to this manuscript I contributed extensively for this work. Together with Soudeh Ehsani, we performed most of the experiments presented in the manuscript, except the ones presented in Figure 4 and Figure S2 (performed by Cristina Rodrigues and Ricardo Henriques), and Figure 6 (performed solely by Soudeh Ehsani). Data presented in Figures 3 and 5 was performed solely by myself. Together with Soudeh Ehsani and Jost Enninga, we assembled all figures, except Figures 4 and Figure S2. I also contributed to text editing and manuscript revision. State of publication Published in Infection & Immunity, in July 2012. 112 Hierarchies of Host Factor Dynamics at the Entry Site of Shigella flexneri during Host Cell Invasion Soudeh Ehsani, a José Carlos Santos, a,b Cristina D. Rodrigues, a Ricardo Henriques, c Laurent Audry, a Christophe Zimmer, c Philippe Sansonetti, d,e Guy Tran Van Nhieu, f,g and Jost Enninga a Group Dynamics of Host-Pathogen Interactions, Institut Pasteur, Paris, France a ; Doctoral Program in Areas of Basic and Applied Biology (GABBA), Universidade do Porto, Porto, Portugal b ; Group Imagerie et Modélisation, Institut Pasteur, Paris, France c ; Unit Pathogénie Microbienne Moléculaire, Institut Pasteur, Paris, France d ; INSERM Unit 789, Paris, France e ; Interdisciplinary Research Group Intercellular Communication of Microbial Infection, College de France, Paris, France f ; and INSERM Unit 1050, Paris, France g Shigella flexneri, the causative agent of bacillary dysentery, induces massive cytoskeletal rearrangement, resulting in its entry into nonphagocytic epithelial cells. The bacterium-engulfing membrane ruffles are formed by polymerizing actin, a process activated through injected bacterial effectors that target host small GTPases and tyrosine kinases. Once inside the host cell, S. flexneri escapes from the endocytic vacuole within minutes to move intraand intercellularly. We quantified the fluorescence signals from fluorescently tagged host factors that are recruited to the site of pathogen entry and vacuolar escape. Quantitative time lapse fluorescence imaging revealed simultaneous recruitment of polymerizing actin, small GTPases of the Rho family, and tyrosine kinases. In contrast, we found that actin surrounding the vacuole containing bacteria dispersed first from the disassembling membranes, whereas other host factors remained colocalized with the membrane remnants. Furthermore, we found that the disassembly of the membrane remnants took place rapidly, within minutes after bacterial release into the cytoplasm. Superresolution visualization of galectin 3 through photoactivated localization microscopy characterized these remnants as small, specular, patchy structures between 30 and 300 nm in diameter. Using our experimental setup to track the time course of infection, we identified the S. flexneri effector IpgB1 as an accelerator of the infection pace, specifically targeting the entry step, but not vacuolar progression or escape. Together, our studies show that bacterial entry into host cells follows precise kinetics and that this time course can be targeted by the pathogen. Invasive pathogens such as Shigella flexneri,Salmonella enterica, or Listeria monocytogenes are capable of subverting host factors to induce their uptake into typically nonphagocytic epithelial and/or endothelial cells (32). This is achieved via bacterial constituents or adhesive molecules present on the pathogen surface, the secretion of soluble bacterial factors, or the translocation of effectors into the host cell through specialized molecular injection devices. Independent of the mode of interaction, the internalization process is rapid for all studied pathogens, requiring only a few minutes and featuring a complex and coordinated interplay between host and bacterial factors. After ingestion of spoiled food or water, as few as 10 to 100 bacteria are sufficient to cause an infection resulting in mucosal ulceration and bloody diarrhea, qualifying Shigella as a potent enteroinvasive pathogen (20). Upon contact of S. flexneri with an epithelial cell, the injection of effectors through the type III secretion system (T3SS) leads to the formation of a signaling platform consisting of the bacterial translocon complex constituents IpaB and IpaC and the targeting of host factors through injected effectors (25,37). Together, these induce a complex rearrangement of the cortical cytoskeletal components, resulting in the formation of lamelipodia that engulf the pathogen and lead to its uptake. These events are coordinated by bacterial effectors, e.g., IpgB1 and IpgB2, IpgD, or IpaC, and host factors, mainly the small GTPases of the Rho family, Rac, Cdc42, and kinases, such as Abl and Src (4, 7,38). Apart from GTPases and kinases, other signaling molecules have been implicated in the entry process of S. flexneri, namely, inositol signaling, which is targeted by IpgD (7). Examplarily, the bacterial effector IpgB1 mimics RhoG at the host plasma membrane and interferes with the ELMO/Dock180 pathway (12,27). Furthermore, the homologous effector IpgB2, together with IpgB1, orchestrates bacterial entry through their GEF activities, which have been reported for both of them in vitro (11,18). The activation of the GTPases induces members of the WASP family verprolin-homologous protein family (WAVE) that in turn activate the actin-nucleating Arp2-Arp3 complex. Additionally, it has been suggested that the C terminus of IpaC is involved in the activation of the kinase Src (24). In turn, Src and another tyrosine kinase, Abl, play a role in the bacterial entry process via the phosphorylation of CrkII at the plasma membrane, which leads to the recruitment of phosphorylated cortactin to the S. flexneri entry site (3,5). Upon internalization, S. flexneri is surrounded by an endocytic vacuole that is subsequently ruptured, thereby releasing the pathogen into the host cellular cytoplasm (6,35). We have recently shown that the rupture happens within minutes after uptake and can be spotted using fluorescently labeled galectin 3 as a marker for the disassembled membranes (30,34). It is believed that the membrane remnants are then processed into smaller vesReceived 29 December 2011 Returned for modification 22 January 2012 Accepted 12 April 2012 Published ahead of print 23 April 2012 Editor: J. B. Bliska Address correspondence to Jost Enninga, [email protected]. Supplemental material for this article may be found at http://iai.asm.org/. Copyright © 2012, American Society for Microbiology. All Rights Reserved. doi:10.1128/IAI.06391-11 2548 iai.asm.org Infection and Immunity p. 2548–2557 July 2012 Volume 80 Number 7 on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 113 icles, and it has been shown that they are targeted to the autophagy machinery by the tethering of autophagy markers to the site of ruptured membranes, which in turn leads to the induction of signaling pathways (8). Further, it has been found that autophagyassociated signaling is also triggered from S. flexneri surrounded by septin structures termed septin cages (22,23). The dynamic recruitment of host factors to the forming vacuole and to the membrane remnants upon vacuolar disruption is still poorly understood. So far, only a few studies have tracked bacterial entry into living host cells in real time compared to investigations that used endpoint assays (2,24,29,34). Therefore, we aimed at obtaining a more precise picture of the temporal events surrounding the entry of S. flexneri. HeLa cells were transfected with a set of host factors, and the time course of their tethering to the site of bacterial entry and the disassembly of the endocytic vacuole was monitored. This allowed us to delineate the functional hierarchies of host factor recruitment during the entry process, which involves different families of signaling molecules, namely, kinases or GTPases. We observed the simultaneous recruitment of the small GTPases Rac, RhoA, and Cdc42 and of the kinases Src and Abl to the site of bacterial entry, which contrasted with a specific sequence of events for their dispersal during the process of vacuolar rupture. Further, we show that the vacuolar membranes disassemble rapidly upon release of the pathogen into the cytoplasm. Finally, we revealed that the bacterial effector IpgB1 is responsible for the rapid entry of bacteria into the host and propose that it acts as a pacemaker of infection. MATERIALS AND METHODS Cell culture and infection assays. All cell culture reagents were purchased from Invitrogen unless otherwise stated. Human epithelial HeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum (FBS), 50 !g/ml penicillin, 50 !g/ml streptomycin, and 2 mM L-glutamine at 37°C, 5% CO 2 . All live-cell fluorescence microscopy was performed in EM buffer (120 mM NaCl, 7 mM KCl, 1.8 mM CaCl 2 , 0.8 mM MgCl 2 , 5 mM glucose, 25 mM HEPES, pH 7.3). Overnight bacterial cultures were inoculated at a 1/100 dilution in tryptic casein soy broth (TCSB) with the appropriate antibiotic if required and grown to an optical density at 600 nm (OD 600 ) of "0.3. Before infection, the bacteria were washed with PBS and coated with poly-L-lysine (Sigma Corp.) at a final concentration of 10 !g/ml to facilitate bacterial adhesion to cells. After 10 min of incubation at room temperature, the bacteria were washed 2 times with PBS, resuspended in the medium of the cells that were prepared for infection, and used immediately. Bacterial strains. The following S. flexneri strains were used: S. flexneri M90T (wild type), S. flexneri M90T expressing dsRed, BS176 (no virulence plasmid), an ipgB1 strain (nonpolar mutant of the effector IpgB1), and the complemented ipgB1/pHA61 strain (these strains were described previously [11,36,19]). All bacterial strains were grown in TCSB at 37°C. The growth medium was supplemented with kanamycin (100 !g/ml) or ampicillin (50 !g/ml), depending on the resistance of the strain used. Plasmids and transfection. For the expression of actin-mOrange, the mOrange coding sequence was inserted into the actin-EGFP-C3 plasmid by excising the enhanced green fluorescent protein (EGFP) sequence at the NheI/XhoI sites and inserting the mOrange sequence (the primers used were as follows: 5=, AGAGCTGCTAGCATGGTGAGCAAGGGCGA GGA, and 3=, AGAGTCCTCGAGATCTGAGTCCGGACTTCTACAGCT CGTCCATGC). The construct was verified by sequencing. For the expression of galectin 3-tandem Eos fluorescent protein (tdEos), the tdEos coding sequence was inserted in the peGFP-N1 plasmid (BD Biosciences Clontech) by excising the EGFP at the BamHI/NotI sites and inserting the tdEos sequence (the primers used were as follows: 5=, AGCTGGATCCAT CCACCGGTCGCCACCATG, and 3=, AGTCGCGGCCGCTCTAGAGT CGCGGCCGCTTA) and by inserting the galectin 3 coding sequence at the KpnI/BamHI sites (the primers used were as follows: 5=, ATGCGGT ACCCGCCACCATGGCAGACAATTTTTCGCTC, and 3=, GCATGGAT CCGTATCATGGTATATGAAGCACT). The construct was verified by sequencing. The other plasmids have been described previously, as follows: pEGFP-galectin 3 (30), pEGFP-actin (24), pOrange-galectin 3 (34), pOrange-RhoA (34), pEYFP-RhoA (17), pEYFP-Rac1 (17), pOrangeRac1 (34), pEGFP-Abl (5), and pEGFP-Src (24). For transfection of samples that were processed for indirect immunofluorescence, HeLa cells were plated on 12-well plates containing glass coverslips at a density of 2 #10 4 cells per coverslip (diameter, 12 mm; thickness, 0.13 to 0.17 mm) 24 h before transfection. For transfection of samples that were processed for live-cell imaging, HeLa cells were seeded into glass bottom dishes 35 mm in diameter at a density of 2 #10 5 cells per dish (Mattek) or 96-well glass bottom plates (Nunc) at a density of 3 #10 4 cells per plate 24 h before transfection. Microscopy and image analysis. Bacterial invasion was measured in real time on a Leica DM or Nikon inverted microscope equipped with a heated stage, using a 40#N-Plan Objective for simultaneous phase-contrast imaging (Leica or Nikon), and fluorescence imaging was performed with excitation at 465 to 500 nm (fluorescein isothiocyanate [FITC]) and 532 to 554 nm (rhodamine), and emission was detected with 516to 556-nm (FITC) and 573to 613-nm (rhodamine) filters. Images were captured using a Cascade 512B camera or a CoolSnap2 camera (Roeper Scientific). Images were acquired in the two fluorescent channels and in trans every 30 s or 90 s. Time lapse series were analyzed with the freeware program ImageJ (http://rsb.info.nih.gov/ij/) and further processed using Excel (Microsoft). Superresolution imaging. Glass coverslips (diameter, 18 mm; thickness, 1) were cleaned using acetone (high-grade pure) treatment for 1 h, followed by overnight treatment with a potassium hydroxide solution at 0.1 M and extensive cleaning with autoclaved ultrapure water. HeLa cells were maintained in DMEM F-12 medium (without phenol red) supplemented with 4% FBS and 1% penicillin/streptomycin. Fortyeight hours before bacterial infection, HeLa cells were seeded on the treated coverslips and transfected with 1.5 !g of the plasmid galectin 3-tdEos. For the infection assay, S. flexneri M90T bacterial cultures were prepared as previously described. Galectin 3-tdEos-transfected cells were incubated for 45 min with S. flexneri M90T at a multiplicity of infection of 100. Afterward, samples were fixed with 2% paraformaldehyde for 30 min at 37°C, washed, and incubated with 0.5 !l of fluorescent beads (2 mM 0.1-!m TetraSpeck microspheres, fluorescent blue/green/orange/dark red; Invitrogen) for 30 min. For imaging, coverslips were mounted in PBS. Photoactivated localization microscopy (PALM) imaging was performed on a home-made setup, previously described in detail (1,15), based on a Nikon Ti-E eclipse microscope system. For tdEos activation and imaging, the system uses a solid-state laser with an emission wavelength at 488 nm (Spectra Physics, Japan) and diode lasers with emission wavelengths at 405 nm and 561 nm (Spectra Physics, Japan). Observations were performed with a 100#oil immersion objective (numerical aperture [NA], 1.49) and detected by an electron-multiplying charge-coupled device (EM-CCD) camera (Ixon DV887ECS-BV; Andor, Belfast, Northern Ireland). Imaging was performed with a final magnification of #150, corresponding to a pixel size of 107 nm. Sequences of 50,000 to 100,000 wide-field fluorescence images compatible with PALM were acquired with a 50-ms exposure time. During acquisition of PALM image sequences, the 561-nm laser power was kept constant for readout of the tdEos activated state, and the 405-nm laser was pulsed at decreasing frequencies adjusted manually to maintain a stable number of active fluorophores per frame. The imaging parameters were set using the !Manager freeware (http://www.micro-manager.org/), and laser control was achieved with custom software (15). After acquisition, each sequence of raw diffraction-limited images was processed with ImageJ and QuickPALM (15). The QuickPALM software Host Factor Dynamics during Shigella Entry July 2012 Volume 80 Number 7 iai.asm.org 2549 on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 114 computed the positions of individual molecules and reconstructed superresolution images with an initial arbitrary pixel size of 10 nm by superimposing Gaussian spots with a full-width-Hald-maximum (FWHM) of 30 nm centered on these positions. The imaged fluorescent beads were used as fiducial markers for drift correction. To estimate the average resolution, 23 small clusters of galectin 3-tdEos were aligned by their center of mass, and the FWHM of the superimposed clusters was calculated, yielding an estimated resolution of 28 nm. RESULTS Numerous studies have described the recruitment of the host actin cytoskeleton, regulatory factors like small GTPases of the Rho family, and tyrosine kinases to the entry site of enteroinvasive bacteria, such as Shigella or Salmonella (9,31). The studies discovered a growing family of host proteins involved in the entry process. We aimed at studying the spatiotemporal hierarchies between some representative host protein family members at the bacterial entry site throughout the successive steps of internalization during cell challenge with S. flexneri wild type and the ipgB1 mutant strain. Host proteins involved in cytoskeletal rearrangements are recruited simultaneously to the S. flexneri entry site. Figure 1A (for more detail, see Movie S1 in the supplemental material) displays a series of time lapse images of actin-GFP-transfected HeLa cells challenged with dsRed-expressing S. flexneri. These images confirm the well-documented massive actin rearrangements upon host cellular contact with the bacterium. We used such image series to establish a quantification procedure with the open-source software ImageJ to determine the time points when the accumulation of a host factor became detectable at the site of bacterial entry and when it reached its maximum at the entry site before the disassembly of the individual entry focus. These quantifications were performed in order to decipher the order of host factor recruitment at the bacterial entry site. As a control, we compared the bacterial entry kinetics in transfected cells with those of nontransfected cells and found no significant differences. We then went on to challenge HeLa cells with S. flexneri cotransfected with fluorescently tagged actin and members of the small GTPase family. Figure 1B (for more detail, see Movie S2 in the supplemental mateFIG 1 Sequential recruitment and dispersal of host molecules involved in the entry of S. flexneri into epithelial cells. (A) HeLa cells transfected with actinEGFP were challenged with S. flexneri expressing dsRed and monitored by time lapse fluorescence microscopy. The recruitment of actin to the entry site highlights the successive internalization steps, bacterial contact, focus formation, collapse of the entry focus, and intracytoplasmic bacteria moving along actin tails. (B) Simultaneous tracking of host cellular actin and the small GTPase RhoA at the entry site shows that they are both recruited to the bacterial entry site at the same time. (C) Simultaneous tracking of the host cellular kinase Abl and the GTPase RhoA at the entry site shows that they are both recruited at the same time. Abl is tethered to the apical edges of the forming foci, whereas RhoA spreads diffusely through the focus and is enriched around the vacuole containing bacteria. Representative data from 5 to 10 independent experiments are shown. FIG 2 Quantification of the sequence of recruitment of bulk actin, small GTPases, and kinases to the entry site of S. flexneri. The images were quantified by thresholding from the image sequences displayed in Fig. 1 (see Materials and Methods for details). Actin, small GTPases, and kinases are rapidly recruited to the pathogen entry site (between 200 and 400 s after contact), and these factors reach their peak at the entry foci between 425 and 725 s after bacterial contact. The standard deviations (SD) (error bars) indicate high variance of the events temporally, making it impossible to delineate clear hierarchies of recruitment. 1, delay of host factor recruitment to the pathogen entry site upon bacterial contact; 2, time window between host factor recruitment and the time point with the peak of the individual foci before their disassembly. Data from 5 to 10 independent experiments are shown. Ehsani et al. 2550 iai.asm.org Infection and Immunity on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 115 rial) shows that actin and RhoA are simultaneously recruited at the entry site upon challenge. Similar results were obtained with other small GTPases, such as Rac1 or Cdc42 (data not shown). Tyrosine kinase recruitment was exemplarily studied using Abl (Fig. 1C; for more detail, see Movie S3 in the supplemental material) and Src (24) (see Fig. S1 in the supplemental material) fused to EGFP. Cotransfection of Abl and RhoA showed that both were recruited to the bacterial entry site at the same time points; however, their localizations differed upon recruitment. Abl was located at the distal tips of the forming membrane ruffles, whereas the GTPases, such as RhoA, were spread throughout the entry focus or around the entering bacterium, as was previously shown for Src (24,25). This was expected, since distinct functions during the entry process have previously been attributed to the different recruited host factors (25,28). Next, we quantified the fluorescence intensities at the S. flexneri entry sites, setting stringent thresholds in the ImageJ software. The time point of the start of host factor recruitment was determined as the time point when fluorescence values were at least 50% above the background levels before bacterial contact. Then, we determined the time span between this time point and the time point of bacterial contact that could be identified in the TRANS channel (1 in Fig. 2). Second, we measured the mean fluorescence intensity in the area of bacterial entry and determined the time point when it reached its maximum. We then subtracted the time point of bacterial contact from the time point when a maximum mean fluorescence intensity was present within the entry focus. This was used to determine the time it took to reach maximal host factor recruitment, coordinating the cytoskeletal rearrangements at the entry site (2 in Fig. 2). Figure 2 depicts the summary of these quantifications for the recruitment of actin, Rac1, RhoA, and Src. Strikingly, we found that all these factors are recruited simultaneously at the bacterial entry site upon bacterial contact, and the large standard deviations show pronounced variability at the single-cell level (Fig. 2, columns numbered 1). Further, recruitment of the analyzed host factors appeared rapidly upon contact, within 200 to 400 s. Similar results were obtained when analyzing the time it took to reach maximal recruitment of host factors to the entry foci. Here, we found again that the maxima were reached within similar time intervals for all measured host factors and that the standard deviations were pronounced for all analyzed scenarios, highlighting large variability in the kinetics of the entry process. We also realized that there is some variability in the diminishing of the recruited host factor signals at the entry site, e.g., the massive early actin recruitment disappeared before the RhoA sigFIG 3 Rapid assembly and disassembly of galectin 3 around the ruptured membranes after vacuolar escape of S. flexneri. (A) Galectin 3 flags the membrane remnants upon vacuolar rupture of the pathogen. The remnants (indicated by the arrow) are typically present only for a few minutes upon membrane rupture and cannot be distinguished from background signal (arrowheads) at later time points. (B) In some instances, galectin 3 remains associated with the membrane remnants (arrows) for hours after escape from the vacuole. Representative data from at least 10 independent experiments are shown. Host Factor Dynamics during Shigella Entry July 2012 Volume 80 Number 7 iai.asm.org 2551 on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 116 nal in Fig. 1B. However, we have not been able to establish a reliable algorithm to quantify this phenomenon. Disassembly of the ruptured vacuolar membrane remnants is highly dynamic. We went on to investigate the subsequent step of S. flexneri invasion, bacterial escape from the endocytic vacuole. In particular, we were interested in tracking the fate of the disassembling membranes that constituted the endocytic vacuole containing bacteria using the recently described marker galectin 3 (30). It has been shown that fluorescently tagged galectin 3 is recruited to the bacterial entry site within seconds after vacuolar rupture and that it targets the disassembling membranes surrounding the bacterium (30,34). By cotransfecting HeLa cells with actin-EGFP and galectin 3-mOrange, we confirmed this rapid recruitment of galectin 3 to the entering bacteria, which highlights the ruptured membranes as “bacterial ghost-like” structures (Fig. 3A and B, top; see Movie S1 in the supplemental material for more details). Strikingly, our time lapse experiments demonstrated that the galectin 3 vacuolar-membrane-wrapping signal was very short-lived, disappearing within 5 to 15 min after bacterial escape from the vacuole (Fig. 3A, arrow). At later time points, the galectin 3 signal was similar to background signals in cells that were not invaded by S. flexneri (compare the specific signal, indicated by an arrow, and the nonspecific background signals, indicated by arrowheads, in Fig. 3A). In a few cases, galectin 3 highlighted the bacterial ghost-like structures for more than 30 min without disassembling into smaller membrane vesicles. However, it turned out to be difficult to quantify the extent of this phenomenon, due to the heterogeneity of the measured signal (Fig. 3B, arrows in the bottom row). Together, these findings show the rapid processing of the disassembling membranes at the bacterial entry site that are subsequently targeted to autophagy (8). To analyze the membrane disassembly in more detail, we performed PALM, which achieves a 10-fold increase in resolution over the classical optical limit in microscopy (roughly 30 nm versus 300 nm) (1,16). PALM imaging of the photoactivatable tdEos fused to galectin 3 was used to obtain superresolution insight into the cellular localization of the protein upon the rupture of the vacuole containing bacteria. HeLa cells were transfected with galectin 3-tdEos and infected with S. flexneri M90T. We selected seven different individual fixed cells featuring visible galectin 3 staining accumulating around the ruptured membranes of the invading bacteria and superresolved them through PALM imaging. Two representative samples are shown in Fig. 4A to D (first example) and E to H (second example). PALM allowed us to discern small patchy accumulations of galectin 3 in the vicinity of the bacteria (Fig. 4C and G). Importantly, no structures with a lumen (resembling vesicles) could be observed. Further analyses of the cluster sizes of these small accumulations of galectin 3 permitted us to observe a distribution of sizes ranging from 30 nm (the estimated resolution of PALM images) to 300 nm in diameter (see Fig. S2 in the supplemental material). These structures are beyond the resolution limit of standard wide-field epifluorescence microscopy and thus cannot be accurately discerned by standard imaging methods (Fig. 4B and F). Larger clusters of galectin 3 with diameters between 100 and 200 nm are found predominantly in the vicinity of the disassembling bacterial vacuole, but not in the rest of the host cellular cytoplasm (see Fig. S2 in the supplemental material), indicating the vacuolar degradation surrounding the bacteria. We then investigated the localization of the simultaneously recruited host factors (Fig. 1 and 2) at the site of the vacuole containing bacteria around the time of vacuolar escape. In contrast to their recruitment to the entry site, we found that the disassembly process showed a higher level of organization with regard to the temporal sequence of events. We found that the accumulated actin surrounding the bacteria within vacuoles diminished before the recruitment of galectin 3 (Fig. 5A and C; for more detail, see Movie S5 in the supplemental material). Interestingly, the galectin FIG 4 High-resolution analysis of disassembling vacuoles by superresolution microscopy. (A to H) HeLa cells transfected with galectin 3-tdEos fluorescent protein (tdEosFP) were infected with S. flexneri M90T. PALM shows galectin 3-tdEos being organized in heterogeneous patchy clusters in the bacterial enveloping vacuoles. The images correspond to two different acquisitions representative of a total of 7 cells from 2 independent experiments. (A and E) Merges of the bright-field (gray), wide-field epifluorescence (green), and superresolution PALM (red) images. (B to D and F to H) Zoomed images of the bacterial region. (B and F) Wide-field epifluorescence. (C and G) PALM imaging. (D and H) Merges of wide-field epifluorescence and PALM imaging. Scale bars: A and E, 1,000 nm; B to D and F to H, 500 nm. Ehsani et al. 2552 iai.asm.org Infection and Immunity on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 117 3 (arrowheads) and actin (arrows) signals around the disassembling vacuoles were mutually exclusive. Furthermore, actin could not be readily identified around smaller vesicular structures potentially derived from disassembling vacuoles at later time points after the escape of S. flexneri to the cytoplasm. In contrast, the small GTPase Rac1 was also recruited to the vacuole containing bacteria; however, it remained located around the membrane remnants upon vacuolar rupture, colocalizing with galectin 3 (Fig. 5B and D; for more detail, see Movie S6 in the supplemental material). The tyrosine kinase Src (24) (see Fig. S1 in the supplemental material) is also recruited to the Shigella-containing vacuole, but this event was not seen with the kinase Abl and the small GTPase RhoA, which were both recruited rather diffusely or at the distal ends of the forming entry foci (Fig. 1). Together, these findings show that, despite simultaneous recruitment of the investigated host factors to the entering bacteria, the sequence of events during vacuolar disassembly appear to be temporally well organized. IpgB1 accelerates the pace of invasion by S. flexneri.It has been reported that the S. flexneri T3SS effector IpgB1 mimics small GTPases to promote their entry into epithelial host cells (11,12, 27). Using low multiplicities of infection (MOIs) (between 1 and 5) and following bacterial internalization at successive time points by gentamicin protection assay, we found that the ipgB1 strain showed reduced entry. However, increasing the load of challenging bacteria or increasing the time periods of infection impeded the readout of this endpoint assay (data not shown). Therefore, we chose to track the dynamics of the time course of internalization and the recruitment of host factors to the bacterial entry site using the ipgB1 mutant with the aim of measuring the effects on the entry kinetics. Performing time lapse microscopy on HeLa cells transfected with the host factors shown in Fig. 1 and 2, we found that the ipgB1 mutant was able to enter host cells; however, the time periods of host factor recruitment to the bacterial contact site and entry were massively reduced (Fig. 6A and C; see Fig. S3 and Movie S7 in the supplemental material) compared to the wildtype bacteria (Fig. 6B and C; see Fig. S3 in the supplemental material). The wild-type phenotype was restored when the ipgB1 complemented strain was used (see Fig. S3 in the supplemental material). We also tested whether IpgB1 affected the subsequent step of vacuolar maturation and escape of the bacteria into the cytoplasm. To do this, we performed time lapse microscopy on HeLa cells coexpressing fluorescent actin and galectin 3, challenged with either the wild-type, ipgB1, or complemented strain (Fig. 7). Again, we found that the entry of the ipgB1 strain was delayed (Fig. 7A; FIG 5 Hierarchies of host factor dispersal from the disassembling vacuoles. (A) HeLa cells were cotransfected with galectin 3-mOrange (arrow, red) and actin-EGFP (arrowhead, green) before being challenged with S. flexneri. Time lapse microscopy shows that actin disperses from the vacuole containing bacteria before its disassembly. Staining of the two factors is mutually exclusive. (B) Enlargement of an image from panel A. The arrowhead shows the actin signal (green) surrounding one moiety of a bacterium, and the arrow shows the other moiety with galectin 3 accumulation (red). (C) HeLa cells were cotransfected with galectin 3-mOrange (arrow, red) and Rac1-EGFP (arrowhead, green) before being challenged with S. flexneri. Rac1 surrounds the entering bacteria and remains associated with the galectin 3-positive membrane remnants after vacuolar rupture. (D) Enlargement of an image from panel C. The arrowhead and arrow point to the overlapping signal (orange) of Rac and galectin 3. Representative data from 7 independent experiments are shown. Host Factor Dynamics during Shigella Entry July 2012 Volume 80 Number 7 iai.asm.org 2553 on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 118 see Movie S8 in the supplemental material). Interestingly, we identified some actin accumulation around the bacteria in contact with the host plasma membrane (Fig. 7A, middle; for more detail see Fig. S4 in the supplemental material), hinting at the involvement of multiple pathways for actin cytoskeletal rearrangement. Quantifying the time course of vacuolar escape, we were not able to measure significant differences between the ipgB1, the wildtype, and the complemented strains (Fig. 7B). In conclusion, the data presented in Fig. 6 and 7suggest that IpgB1 represents a bacterial T3SS effector that accelerates the pace of S. flexneri entry into host cells as a means to boost bacterial infectivity but that it does not impact the subsequent steps. DISCUSSION Using time lapse microscopy, we demonstrated the simultaneous recruitment of host small GTPases and kinases to the site of S. flexneri entry into epithelial cells. Even though it is evident that bacterial invasion has to be highly organized, the functional hierarchies between the involved host factors have not been identified with precision (2,7). So far, hierarchies of small GTPases have been described only in the case of Salmonella infection (28,29). Simultaneous host factor recruitment (Fig. 1 and 2) may call into question the importance of strict hierarchies during cellular invasion or may highlight the fact that the entry process is following not only one pathway, but multiple pathways, as previously proposed for S. enterica (13,14). However, it also highlights the limitations of the performance of our microscopes with regard to both spatial and temporal resolution. Another fact that has to be considered is the activation of the signaling molecules involved; for example, GTPases can switch between GDPand GTP-bound states, and kinases can be activated via phosphorylation. Such activities have already been taken into account by some studies on host-pathogen interactions, for example, in the case of Yersinia, using functionalized fluorescence resonance energy transfer (FRET) probes for the GTP state of small GTPases (39). Nevertheless, these studies require a high level of experimental sophistication, impeding its broad use throughout the scientific community. Considering the manifold and seemingly contradictory functions of the injected bacterial effectors on the host, such as actin polymerization via IpaC and its depolymerization via IpaA, we suggest that future studies will be required to reveal the precise recruitment of host factors and their dispersal (24). This will be facilitated once the precise enzymatic functions of the injected effectors have been revealed. It is possible to track the step of vacuolar escape of S. flexneri with more precision. Recently, we have shown that this event takes place rapidly upon internalization (26,34). Vacuolar escape has also been found for other cytoplasmic bacteria, for example, Listeria and Rickettsia (35). Strikingly, in this study, we show the rapid disassembly of the membrane remnants (Fig. 3) that have been reported to be coated with autophagy markers upon vacuolar rupture (8). This shows that the signaling events leading to the “digestion” of the membrane remnants have to be very rapid, efficient, and transient. PALM microscopy revealed the heterogeneous, patchy nature of the membrane remnants (Fig. 4). First, we found that the disassembled membranes were not hollow, highlighting their multilayered or micellar organization. Second, their resolvable size distribution spanned 30 to 300 nm (see Fig. S1 in the supplemental material). Based on these observations, we suggest that the membrane remnants are eventually disassembled or recycled, not only by a single mechanism, but by multiple mechFIG 6 The S. flexneri ipgB1 strain enters HeLa cells, but at a reduced rate. (A) Time lapse microscopy of cells cotransfected with actin and Rac1 and challenged with the S. flexneri wild-type strain. Arrows show bacterial entry sites. (B) Cells transfected similarly to those in panel A and challenged with an S. flexneri ipgB1 mutant. Arrows show bacterial entry sites. (C) Quantification of the entry dynamics of the S. flexneri wild-type (WT) and ipgB1 strains was performed using the time-lapse image series. Data from 5 independent experiments are shown. The error bars indicate SD. Ehsani et al. 2554 iai.asm.org Infection and Immunity on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 119 anisms. Further, it is likely that the heterogeneous vesicular remnants result in the induction of multiple signaling pathways. Taking advantage of our single-cell-based assays, we also found that some bacteria remained vacuole bound in vesicles coated with actin. It would be interesting to investigate the fate of these vacuole-bound bacteria, for example, if they are also targeted to the autophagy pathway. So far, it has been reported that cytoplasmic bacterial pathogens can be trapped within septin cages that send specific signals to trigger autophagy (22,23). In addition to these signaling events, we suggest that future studies investigate whether septins can also be recruited to S. flexneri trapped within actincoated vacuoles and whether the signals emanating from the trapped vacuolar bacteria are different from the signals emanating from cytoplasmic bacteria trapped in septin cages. Where S. flexneri succeeded in escaping from the endocytic vacuoles, we noted that the dispersal of the host factors surrounding the cytoplasmic escaping bacteria appeared to be more organized than the recruitment during the initial steps of entry (Fig. 5). Since the membrane remnants appear to be targeted to the autophagy machinery, it will be important to investigate whether the sequential recruitment and dispersal of host factors around these remnants impacts the signaling cascades emanating from them. Challenging host cells with a number of bacterial mutants, for example, with the translocon component IpaB or IpaC, results in very little or no invasion of the host cells (10). In contrast, mutant strains for other effectors show attenuated and subtle invasion phenotypes. One of these is the GEF IpgB1 that interferes with the ELMO/Dock180 pathway at the host plasma membrane upstream of the small GTPases (12,27). An S. flexneri mutant strain for IpgB1 and IpgB2 was as invasive as the wild-type strain, indicating FIG 7 Effects of Shigella IpgB1 on entry and vacuolar disassembly. (A) Time lapse microscopy of HeLa cells cotransfected with actin-EGFP and galectin 3-mOrange challenged with the wild-type strain, with the ipgB1strain, or with the ipgB1 strain complemented with IpgB1. The ipgB1 mutant escaped efficiently from the endocytic vacuole. (B) Quantification of the time intervals between entry focus formation and escape of S. flexneri wild-type, S. flexneri ipgB1, and complemented strains from vacuoles. The wild-type bacteria escaped from the vacuoles at the same rate as the ipgB1 mutant. Representative data from 4 independent experiments are shown. The error bars indicate SD. Host Factor Dynamics during Shigella Entry July 2012 Volume 80 Number 7 iai.asm.org 2555 on June 18, 2012 by INSTITUT PASTEUR-Médiathèquehttp://iai.asm.org/Downloaded from 126 Supplemental Movie Legends Movie S1. Time-lapse movie corresponding to Fig. 1A: Infection of actin-GFPtransfected cells with Shigella flexneri expressing the dsRed protein. Movie S2. Time-lapse movie corresponding to Fig. 1B: HeLa cells cotransfected with actin-eGFP and RhoA-mOrange challenged with Shigella flexneri. Movie S3. Time-lapse movie corresponding to Fig. 1C: HeLa cells cotransfected with RhoA-mOrange and Abl-eGFP challenged with Shigella flexneri. Movie S4. Time-lapse movie corresponding to Fig. 3: HeLa cells cotransfected with actin-eGFP and galectin-3–mOrange challenged with Shigella flexneri. Movie S5. Time-lapse movie corresponding to Fig. 5A: HeLa cells cotransfected with galectin-3–mOrange and actin-eGFP challenged with Shigella flexneri. Movie S6. Time-lapse movie corresponding to Fig. 5B: HeLa cells cotransfected with Rac1-citrin and galectin-3–mOrange challenged with Shigella flexneri. Movie S7. Time-lapse movie corresponding to Fig. 6A: HeLa cells cotransfected with Rac1-mOrange and actin-eGFP challenged with the ipgB1 mutant. Movie S8. Time-lapse movie corresponding to Fig. 7A: HeLa cells cotransfected with actin-eGFP and galectin-3–mOrange challenged with the ipgB1 mutant. Movies S1-S8 are shown on the Electronic Annex, on the attached CD. 127 Manuscript 2 – “The COPII complex and lysosomal VAMP7 determine intracellular Salmonella localization and growth” José Carlos Santos1,2, Magalie Duchateau3, Jennifer Fredlund1, Allon Weiner1, Adeline Mallet4, Christine Schmitt4, Mariette Matondo3, Véronique Hourdel3, Julia Chamot-Rooke3,5 and Jost Enninga1 1. Unit of Dynamics of Host-Pathogen Interactions, Institut Pasteur, Paris, France 2. Graduate Program in Areas of Basic and Applied Biology (GABBA), Universidade of Porto, Portugal 3. Structural Mass Spectrometry and Proteomics Unit, Institut Pasteur, Paris, France 4. Plate-forme Microscopie Ultrastructurale, Institut Pasteur, Paris, France 5. CNRS UMR3528, Paris, France Contribution to this manuscript For this manuscript, I contributed to all parts of the work. For nearly one and a half years I developed and optimized the protocol for the biochemical purification of SCVs, in parallel with the assay for assessing SCV integrity. In collaboration with Magalie Duchateau, from the Structural Mass Spectrometry and Proteomics Unit of the Institut Pasteur, we optimized the best conditions to analyze the SCV enriched fraction by LC-MS/MS. The quantitative analysis of the SCV proteome was performed by Mariette Matondo and Véronique Hourdel. Nonetheless, I also extensively discussed the methodologies for data analysis with them, Julia Chamot-Rooke and Jost Enninga. I performed the evaluation of hyper-replicating Salmonella together with Jennifer Fredlund. For the CLEM, I imaged the samples by fluorescence spinning disk confocal microscopy, Christine Schmitt prepared the Epon blocks and Adeline Mallet performed the FIB/SEM acquisition. All these steps were performed with crucial help from Allon Weiner, who also helped me and taught me how to do the data correlation, analysis and segmentation with the Amira software. I prepared all the figures presented in the manuscript and I wrote the entire manuscript, with the proofreading of Jennifer Fredlund and Jost Enninga. 128 State of publication Accepted for publication in Cellular Microbiology, in June 2015. 129 The COPII complex and lysosomal VAMP7 determine intracellular Salmonella localization and growth José Carlos Santos,1,2 Magalie Duchateau,3 Jennifer Fredlund,1Allon Weiner,1Adeline Mallet,4 Christine Schmitt,4Mariette Matondo,3 Véronique Hourdel,3Julia Chamot-Rooke3,5 and Jost Enninga1* 1Unit of Dynamics of Host-Pathogen Interactions,Institut Pasteur,Paris, France. 2Graduate Program in Areas of Basic and Applied Biology (GABBA),University of Porto,Porto, Portugal. 3Structural Mass Spectrometry and Proteomics Unit, Institut Pasteur,Paris, France. 4Plate-forme Microscopie Ultrastructurale,Institut Pasteur,Paris, France. 5CNRS UMR3528,Paris, France. Summary Salmonella invades epithelial cells and survives within a membrane-bound compartment, the Salmonella-containing vacuole (SCV). We isolated and determined the host protein composition of the SCV at 30 min and 3 h of infection to identify and characterize novel regulators of intracellular bacterial localization and growth. Quantitation of the SCV protein content revealed 392 host proteins specifically enriched at SCVs, out of which 173 associated exclusively with early SCVs, 124 with maturing SCV and 95 proteins during both timepoints. Vacuole interactions with endoplasmic reticulum-derived coat protein complex II vesicles modulate early steps of SCV maturation, promoting SCV rupture and bacterial hyper-replication within the host cytosol. On the other hand, SCV interactions with VAMP7-positive lysosome-like vesicles promote Salmonella-induced filament formation and bacterial growth within the late SCV. Our results reveal that the dynamic communication between the SCV and distinct host organelles affects both intracellular Salmonella localization and growth at successive steps of host cell invasion. Introduction Salmonella enterica serovar Typhimurium (Salmonella)isa Gram-negative enteric pathogen that can cause acute gastroenteritis in humans after ingestion of contaminated food or water. Salmonellosis is one of the most common sources of food-borne disease in humans and is a major public health and economic burden worldwide (Majowicz et al., 2010; Agbor and McCormick, 2011). A key aspect of Salmonella virulence is its ability to invade and survive within nonphagocytic intestinal epithelial cells, processes driven by two type III secretion systems (T3SS1 and T3SS2) that together inject more than 30 effector proteins into the host cell (Haraga et al., 2008; Figueira and Holden, 2012). Host cell invasion is mostly mediated by T3SS1-injected effectors, triggering fast and massive rearrangements of the actin cytoskeleton (Zhou et al., 2001; Scott et al., 2005; Pateland Galán, 2006) followedby the formation of plasma membrane (PM) ruffles and Salmonella engulfment into the Salmonella-containing vacuole (SCV). The SCV is a unique, modified membrane-bound compartment that enables bacterial survival and replication. The early SCV [<30 min post-invasion (p.i.)] has been shown to share some similarities with early endosomes, namely an association with phosphatidylinositol 3-phosphate, Rab5, Vps34, early endosomal antigen-1 (EEA-1) and sorting nexins -1 and -3 (Hernandez et al., 2004; Bujny et al., 2008; Mallo et al., 2008; Steele-Mortimer, 2008; Braun et al., 2010). The maturing SCV [between 30 min and 5 h p.i.] undergoes extensive membrane remodelling and acquires late endosome/lysosomal markers, such as Rab7, lysosomal-associated membrane-associated protein-1 (Lamp-1) and vacuolar ATPase (vATPase) (reviewed in Steele-Mortimer, 2008; Schroeder et al., 2011). This is accompanied by movement of the SCV along microtubules (MTs) to a juxtanuclear position adjacent to the MT-organizing centre (Harrison et al., 2004). Maintaining the SCV in the perinuclear region is thought to be important for promoting bacterial replication, which is initiated 3–4 h p.i. (Ramsden et al., 2007a,b; Bakowski et al., 2008). The last stages of SCV maturation (>5 h p.i.), mostly mediated by T3SS2 effectors, are characterized by concomitant intravacuolar bacterial replication and formation of Lamp-1-enriched membrane tubules, named Salmonella-induced filaments (SIFs), that extend from the SCV along MTs (Drecktrah et al., 2008). SIFs are highly dynamic structures and can spread throughout the entire Received 15 April, 2015; revised 22 May, 2015; accepted 12 June, 2015. *For correspondence. E-mail [email protected]; Tel. (+33) 1 44 38 94 13; Fax (+33) 1 40 61 35 83. Cellular Microbiology (2015) doi:10.1111/cmi.12475 © 2015 John Wiley & Sons Ltd cellular microbiology 130 cell to form a complex network. They are also enriched in vATPase, Rab7 and cholesterol (Brumell et al., 2001). In addition to SIFs, other tubular networks emanate from the late SCV, such as Salmonella-induced SCAMP3 tubules (Mota et al., 2009) and Lamp-1-negative tubules (Schroeder et al., 2010). However, the biological role of all these Salmonella-induced tubules remains largely unknown (Schroeder et al., 2011). Membrane damage during the early or maturing stage has also been described forsome SCVs, which givesbacteriaaccess to the hostcell cytosol. This can allow for bacterial detection and degradation by autophagy mechanisms (Jo et al., 2013). Nevertheless, recent data show that in approximately 9% of infected epithelial cells, these cytoplasmic bacteria can replicate at much faster rates than those within an SCV (doubling time of ∼20 min); this is termed hyper-replication (Knodler et al., 2010; Malik-Kale et al., 2012). Thus, intracellular Salmonella growth can be different depending on its localization within the host cell. Proteomics has been used to reveal the protein composition of phagosomes (Desjardins et al., 1994; Gagnon et al., 2002; Stuart et al., 2007; Rogers and Foster, 2008). Other proteomics studies have reported the protein composition of bacteria-containing vacuoles, such as Legionella pneumophila (Shevchuk et al., 2009; Urwyler et al., 2009; Hoffmann et al., 2014) or Mycobacterium bovis BCG (Lee et al., 2010), and have given important insights into the mechanisms of intracellular bacterial survival. In this study, we applied a quantitative proteomics approach to identify novel host factors associated with the SCV at different stages of its maturation. We reproducibly isolated SCVs and determined their protein composition. Using functional and correlative ultrastructural approaches, we then characterized and showed that two specific SCV–protein interactions affect intracellular Salmonella growth. We demonstrate that early interactions between the SCV, the endoplasmic reticulum (ER) and the coat protein complex II (COPII) complex promote cytoplasmic Salmonella localization and hyper-replication. At late stages, bacterial growth is regulated by interactions between intact SCVs and VAMP7-positive lysosome-like vesicles, which also determine SIF formation. Results Purification of SCVs from infected epithelial host cells We developed a fractionation methodology to obtain a subcellular fraction highly enriched in intact SCVs that could be used to determine the SCV proteome. Vacuoles were isolated at two time-points of Salmonella infection representing two stages of SCV maturation: 30 min, corresponding to the early SCV; and 3 h, corresponding to the maturing SCV. Later stages of SCV maturation were not isolated, as bacterial replication and Salmonellainduced tubule formation interfere with our purification procedures. In order to obtain intact SCVs with sufficient purity, we performed careful cell homogenization and centrifugal separation in density gradients. SCV integrity was quantified in the post-nuclear supernatant (PNS) after cell homogenization, by a novel ELISA-based assay (Fig. 1; see Experimental procedures for details). In this assay, non-vacuolarized Salmonella adhere to an immobilized antibody and are then quantified by a secondary, biotinylated antibody, whereas vacuolarized bacteria do not adhere and are thus not counted (Fig. 1A, left panel). Total bacteria in each sample were also determined using this method by first subjecting samples to osmotic shock in order to rupture the SCVs and free all vacuolarized bacteria. In order to evaluate the robustness of our assay and to quantify the number of non-vacuolarized Salmonella in the PNS, a standard curve with known amounts of bacteria was generated and statistically validated (Fig. S1A and B). In the non-infected control PNS, to which a known amount of bacteria was added, the percentage of non-vacuolarized and total, osmotically shocked, Salmonella was the same (Fig. 1A, right panel). At the two time-points of infection, non-vacuolarized Salmonella accounted for only 16% of the total number of bacteria in the PNS, indicating that the vast majority of SCVs are intact after cell homogenization. Fig. 1. Isolation of highly enriched and intact SCVs from infected host cells. A. Salmonella-infected HeLa cells were mechanically homogenized in an isotonic buffer and the SCV integrity in the PNS was tested by ELISA. Non-infected PNS supplemented with a known amount of added bacteria was used as control. A small aliquot of sample was subjected to a sandwich ELISA, either at isotonic conditions (non-treated) or after osmotic shock. Data are represented as the mean ±SEM from three independent and representative experiments. B. The different PNSs (non-infected control, 30 min and 3 h infection) were fractionated by ultracentrifugation in a density gradient and the number of Salmonella in each fraction (F1 to F12) was determined by counting CFUs (black bars). Simultaneously, the density of each fraction (expressed in g cm−3) was measured in a refractometer and plotted together in the same graph (white dots). Data represent the mean ±SEM from five independent experiments. C. Aliquots of fractions F6 to F8 were subjected to the ELISA method described in (A). Data show the mean ±SEM from three independent and representative experiments. D. All fractions F1 to F12 and the PNS were tested using markers for the following compartments: early endosomes (EEA-1 and Rab5), late endosomes and lysosomes (Lamp-1), Golgi (GM130), peroxisomes (catalase), mitochondria (TOM22) and ER (calreticulin). The organelle distribution in the density gradients was the same in the non-infected control (left panel), 30 min (right panel) and 3 h (Fig. S2). Salmonella distribution, detected with a specific anti-Salmonella LPS antibody, was the same in the 30 min (right panel) and 3 h (Fig. S2) fractionations. All P-values were determined using the Student’s t-test. 2J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 131 SCV interaction with the ER and lysosomes 3 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 132 We performed five independent infection experiments, followed by ultracentrifugation, with highly reproducible linear density gradients (Fig. S1C). This separated SCVs from remaining subcellular organelles. To determine the SCV position within the gradient, we measured the number of colony forming units (CFUs) of total Salmonella in each fraction. As a control, we added Salmonella to a non-infected PNS before fractionation and found that the majority of non-vacuolarized bacteria (Fig. 1B, left panel) accumulated in fractions F7 and F8 (1.11 and 1.12 g cm−3 respectively). In contrast, the fractionation of the infected PNS, both at 30 min and 3 h (Fig. 1B, middle and right panels), led to a 500-fold enrichment of total bacteria in fraction F6 (1.10 g cm−3). We next evaluated SCV integrity in fractions F6 to F8 by ELISA. The number of nonvacuolarized bacteria in fractions F6 and F7 was significantly lower than the total Salmonella as determined by osmotic shock (Fig. 1C, middle and right panels). This indicates that the bacteria in these two fractions were mostly within intact SCVs. In contrast, fraction F8 contained solely non-vacuolarized Salmonella, as no increase was seen after osmotic shock. Thus, in fractions F6 and F7, we successfully isolated intact SCVs and separated them from non-vacuolarized bacteria. We assessed the separation efficiency of different subcellular organelles by Western blot (Fig. 1D). As expected, the Golgi apparatus and early endosomes showed a low density (between 1.03 and 1.09 g cm−3), whereas peroxisomes had a high density (between 1.12 and 1.15 g cm−3) and did not overlap with fractions F6 and F7. Late endosomes/lysosomes accumulated mostly in F5 (1.09 g cm−3) but also showed partial overlap with the SCV-enriched fraction F6. The ER displayed a broad distribution within the gradient, spanning from F5 to F8 (1.09 to 1.12 g cm−3) and overlapped with the SCV fractions. This distribution can be explained by the complex structure of the ER and its heterogeneous physical properties. Importantly, the tested conditions led to the accumulation of mitochondria only in fractions F7 and F8 (1.11– 1.12 g cm−3), not overlapping with the SCV fraction, F6. For all tested antibodies, equal results were obtained for the 30 min (Fig. 1D, right panel) and 3 h fractionation (Fig. S2). We successfully isolated intact and highly enriched SCVs in fraction F6, which we then used to determine the protein composition of the SCV. The dynamic quantitative host protein composition of the SCV To determine the repertoire of host proteins enriched at the SCV at the two time-points, we used a label-free quantitative mass spectrometry (MS)-based proteomics approach. We compared the relative protein abundance in the SCV fraction (30 min or 3 h) with a non-infected control fraction, prepared in parallel. MS data obtained from five independent experiments were analysed by MaxQuant, both for protein identification and quantification (Cox and Mann, 2008; Luber et al., 2010). As protein abundance from contaminating organelles should not differ between the tested conditions (Rao et al., 2009), factors that are enriched in the SCV can be identified through their positive fold-change ratio compared with the control. Analysis of the entire MS data set identified 2522 host proteins, 392 of which (∼15%) showed a statistically significant fold-change increase at one or both of the two time-points of infection (Fig. 2A and Fig. S3A and B and Tables S1 and S2). In detail, 173 proteins were enriched solely at the 30 min SCV (red), 124 uniquely at the 3 h SCV (green) and 95 at both time-points (orange) (Fig. 2A). These three specific subsets of host factors (only 30 min SCV; only 3 h SCV; 30 min +3 h SCV; complete list is in Table S3) reflect how SCV protein composition is altered during its maturation.An overview of some selected proteins dynamically associated with the SCV can be found in Table S4. The identified proteins were grouped according to their putative subcellular localization or biological function by gene ontology analysis. We detected a significant enrichment of ER-, Golgiand vesicle-derived proteins, all of which decreased with time. Also, lysosome-derived proteins were exclusively increased in the 30 min SCV (Fig. 2B). We observed significant enrichment of proteins involved in ER to Golgi vesicle-mediated transport at the 30 min SCV or at both time-points (Fig. S3C). Together, these data highlight that SCV protein composition dynamically varies with time and that specific biological processes are implicated at each step of SCV maturation. Several host proteins identified by our quantitative approach were previously described to be associated with the SCV (Rathman et al., 1997; Steele-Mortimer et al., 1999; Harrison et al., 2004; Boucrot et al., 2005; Smith et al., 2007; Thurston et al., 2012) or involved in Salmonella infection (Criss and Casanova, 2003; Hänisch et al., 2010; Jolly et al., 2014), confirming the potential of our work (Fig. 2C and Table S4). Importantly, we also identified several novel host proteins associated with the SCV (Tables S3 and S4). Among them, we were particularly interested in ER-derived proteins, all the factors from the coat protein complex II (COPII) machinery and lysosomederived proteins (Fig. 2C). Early SCV interactions with ER-derived COPII complex promote vacuolar rupture and cytoplasmic Salmonella growth through hyper-replication A key finding in our quantitative proteomic analysis was that approximately 20% of the proteins enriched in early SCVs were derived from the ER (Fig. 2B and Table S3 for 4J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 133 ER Golgi Vesicle Lysosome Cytoskeleton 0 5 10 15 20 25 Host proteins identified (%) Only 30 min SCV ND **** **** **** ** **** ** *** *** Both 30 min + 3 h SCV Only 3 h SCV * ND *** * ** **** AB C 173 124 95 30 min SCV 3 h SCV Number of host proteins enriched in the SCV + (enriched relative to control) – (depleted relative to control) n.d. ( no diference relative to control ) ER-derived proteins Early SCV Maturing SCV Calnexin + + Re!cu l on-4 +n. d . VAP-A + – VAP - B + – COPII-complex proteins Early SCV Maturing SCV Sec13 +n.d. S ec 23 + + Sec24 +n.d. Sec31 +n.d. Lysosome-derived proteins Early SCV Maturing SCV Cathepsin B + – Cathepsin Z + – CI Man-6-P receptor +n.d. Dipep!dyl pep!dase 2 + – Known SCV associated proteins Early SCV Maturing SCV Dynac!nn.d. + D y nein ++ Galec!n-3 + – Kinesin light chain 1 n.d. + Kinesin light chain 2 n.d. + Kinesin-1 heavy chain + + Lysosomal acid phosphatase + – Rab-4a +n.d. Rab-7a +n.d. vATPase +n. d . Proteins known to be involved in Salmonella infec!on Early SCV Maturing SCV Arp2/3 complex +n.d. Ac!n+n.d. Annexin A2 +n.d. F il am i n++ Tropomyosin-4 +n.d. Tubulin n.d. + Fig. 2. The SCV protein composition dynamically changes during vacuole maturation. Protein data from fraction F6 of each condition (control, 30 min and 3 h) was analysed with label-free quantitation algorithms (MaxQuant software). Relative host protein abundances were compared between the control and any two infection time-points. Proteins enriched at a specific time-point were considered as SCV constituents. A positive fold-change was considered when the abundance ratio 30 min/control or 3 h/control was >1.3 (log2fold-change >0.387) with a P-value <0.05. A. Thirty minute SCV-enriched proteins were compared with 3 h SCV-enriched proteins and a Venn diagram was built. B. The host proteins enriched in the SCV were grouped according to their subcellular localization. For each term, the analysis was performed for the proteins enriched uniquely either at the 30 min SCV (only 30 min SCV, red bars) or at the 3 h SCV (only 3 h SCV, green bars) or for the proteins enriched at both time-points (both 30 min +3 h SCV, orange bars). The graphs show the percentage of host proteins enriched at the SCV, relative to the total number of proteins identified in each condition. Statistics were performed by determining the P-values (EASE score, as mentioned in the Experimental procedures) and show the robustness of gene-term enrichment for each condition. ND, non-enriched factors in comparison with the control. C. Selected proteins enriched in the early (30 min) or maturing (3 h) SCV are shown with a ‘+’ symbol. SCV interaction with the ER and lysosomes 5 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 134 details). Figure 3A depicts some of these host factors, such as calnexin, which was previously described to be associated with phagosomes (Gagnon et al., 2002) and with the Legionella-containing vacuole (LCV) (Lu and Clarke, 2005; Ragaz et al., 2008). Hence, we investigated its subcellular localization during Salmonella infection by immunofluorescence microscopy. Fast and transient calnexin accumulation at the invasion site was observed, which peaked at 10 min p.i. (Fig. 3B, upper panel; quantification in Fig. 3C), and by 30 min p.i., we observed calnexin accumulation around some of the SCVs (Fig. 3B, magnifications). Reticulon-4 immunostaining was used to observe tubular ER during bacterial infection. Unlike calnexin, reticulon-4 did not consistently accumulate at the Salmonella entry site (data not shown); however, we observed that the early SCV seemed to be wrapped by ER filaments (Fig. 3D, insets for details). Additionally, Western blot analysis of the isolated subcellular fractions confirmed enrichment of reticulon-4 in the 30 min fraction F6 (Fig. S4A). The precise nature of the SCV–ER interactions could not be determined by light microscopy due to its resolution limit. Therefore, we applied an emerging technique, termed correlative-focused ion beam/scanning electron tomography (C-FIB/SEM). Fluorescence microscopy and large-volume ultrastructural tomography are combined in a single three-dimensional (3D) data set, allowing for precise identification of molecules of interest within the ultrastructural volume. This technique was previously used in our laboratory to characterize the host cell environment around the Shigella-containing vacuole (Mellouk et al., 2014). In short, cells were infected with Salmonella for 30 min and fixed. The ER was labelled with a reticulon-4 antibody followed by indirect immunofluorescence, DNA (bacteria and host cell nuclei) was stained and cells were imaged by confocal microscopy followed by FIB/SEM tomography at the exact same location (Fig. 3E, see Experimental procedures for details). The two data sets were then correlated and combined into a single data set presented here. DAPI and reticulon-4 fluorescent signals were segmented by thresholding (left middle and lower panels) and the corresponding FIB/SEM data (upper middle and right panels) show the segmentation of bacteria (blue) and of the SCV lumen (yellow). After superimposing the ER fluorescent signal with the FIB/SEM data, we could segment the ER ultrastructure (red). Strikingly, in all data sets (n=4) the SCVs were surrounded by ER (Fig. 3E, lower right panel). Moreover, in all data sets we could observe membrane interactions between the ER and SCVs (Fig. 3F and Fig. S4B show different data sets; in Video Clip S1, ER contacts with multiple SCVs are also observed) that resembled membrane contact sites (MCS) (Orci et al., 2009; Eden et al., 2010; Stefan et al., 2013). Therefore, we conclude that there is membrane contact between the early SCV and the ER. Interestingly, we found that the early SCV proteome was enriched in all constituents of the COPII complex, namely Sec13, Sec23, Sec24 and Sec31 [proteins form a complex leading to vesicle budding from the ER membrane and the transport of cargo to the Golgi or the cell surface (Haucke, 2003; Sato, 2004; Lord et al., 2013)]. Using immunofluorescence microscopy, we observed the accumulation of COPII complex around a small percentage of intracellular Salmonella (8–15%), as seen by Sec13-positive bacteria (Fig. 4B and C and Fig. S4C). This confirmed our proteomic data and also suggested a potential link between the ER and some of the early SCVs, possibly via COPII-coated vesicles. To examine if COPII affects the intracellular Salmonella lifestyle, we measured bacterial growth within epithelial cells using gentamicin assays (Elsinghorst, 1994), after inhibiting COPII function via Sec13 siRNA treatment. Salmonella entry into HeLa cells, measured 1 h p.i., was not affected after Sec13 knockdown (data not shown). Strikingly, Sec13-depletion impaired intracellular bacterial growth from 3 h p.i. onwards, as compared with the control (Fig. 4D). These data show that the COPII complex is crucial for bacterial growth within the host cell. Then, we tested if there was a functional link between the Fig. 3. The SCV contacts with the host cell ER. A. Log2fold-change in ER-derived protein abundance in the SCV-enriched fraction compared with the non-infected control from five independent MS/MS experiments. Relative protein abundances were considered different when the log2fold-change >0.387 (red line) or <−0.387 (green line), with a P-value <0.05. Statistics are relative to the non-infected control for each protein. B–D. HeLa cells were infected with Salmonella (B) or Salmonella-dsRed (D), fixed at different time-points and immunostained for calnexin or reticulon-4 respectively. F-actin was stained with phalloidin and DNA with DAPI (cyan). Representative confocal microscopy images are shown. In (B), arrows indicate the site of bacterial entry, for which calnexin accumulation at the different time-points was quantified (C) from three independent experiments. Scale bars represent 10 μm. E and F. Fluorescence confocal microscopy was followed by large-volume FIB/SEM of the same Salmonella-infected cell. (E) Confocal image of an infected cell is shown in the upper left corner, together with the region that was imaged by FIB/SEM (white box). Corresponding zoomed 3D views are shown for the native and segmented fluorescent signals (middle and lower left panel, respectively), together with the volume of FIB/SEM acquisition. A 3D view of the FIB/SEM acquisition is depicted in the upper middle panel. Using Amira software, the bacteria (blue), the SCVs lumen (yellow) and the ER (red) were segmented (upper right panel). The FIB/SEM segmentation was correlated with reticulon-4 fluorescent signal, showing overlap between the ER segmentation and the fluorescence (lower right panel). (F) A xy-view FIB/SEM section shows a site of contact between the SCV and the ER membrane (upper panel, white box). A corresponding partial 3D segmentation showing the bacteria (blue), the SCV lumen (yellow) and the ER (red) is presented in the lower panel. 6J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 135 COPII complex and bacterial hyper-replication within the cytosol, which requires SCV rupture (Knodler et al., 2010; 2014; Malik-Kale et al., 2012). By 6 and 9 h p.i., we found COPII localizing to bacterial-shaped structures in 100% of the cells with hyper-replicating Salmonella (Fig. S4D and E), although only surrounding some of the bacteria. These structures were specific to bacterial hyper-replication, as they were absent in infected cells not containing FIB section (xy)FIB segmentation 2 1 12 Reticulon-4 F-ActinMerge Salmonella-dsRed tesnInim03 VAP-A VAP-B Reticulon-4 SERCA2 Calnexin Sec20 Sec22b ERGIC-32 -1.0 -0.5 0.0 0.5 1.0 1.5 ER-derived proteins log2Fold-change (relative to control) 30 min SCV 3hSCV ** * **** * **** *** * * **** *** *** **** * BA 5 min 10 min 30 min 020406080100 Calnexin recruitment to actin entry foci (%) 10 min DAPI Calnexin F-Actin Merge 30 min CD E 400 nm FIB/SEM Confocal microscopy Segmentation 3D view Segmentation Bacteria SCVs ER F 2 1 12 2 1 12 SCV interaction with the ER and lysosomes 7 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 142 Lysosomal interaction with the SCV has remained a matter of debate. Initial studies pointed to the SCV avoiding fusion with late endosomes and lysosomes, as hydrolases and the M6PR were reported to be absent from the late SCV (Buchmeier and Heffron, 1991; Garcia-del Portillo and Finlay, 1995; Rathman et al., 1997; Hashim et al., 2000). This is, however, controversial (Oh et al., 1996) and has been recently questioned, as extensive dynamic interactions between the lysosomal system and the SCV were observed during vacuole maturation (Drecktrah et al., 2007). Intriguingly, our proteomics data show that the early SCV is enriched in lysosomal proteins, pointing to a model where the early SCV remains accessible to incoming lysosomal content that does not take place at later time-points of vacuole maturation. Such direct interactions were visualized by C-FIB/SEM, in which we observed fusion events of VAMP7/Lamp-1positive vesicles with the early SCV. Thus, fusion between 3 hours 6 hours 9 hours 0.00 3.00 6.00 9.00 12.00 15.00 Time p.i. Number of bacteria/cell scramble siRNA VAMP7 siRNA ns ns **** A Scramble VAMP7 0 10 20 30 40 50 60 70 % infected cells with SIFs siRNA: **** B C DAPIVAMP7-RFP Lamp-1-GFP Merge Lysosomal-like vesicles (VAMP7) ER COPII SCV rupture Maturing SCV SIF formation Salmonella hyper-replication Early SCV Late SCV D VAMP7 Actin scramble siRNA VAMP7 siRNA 14 J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 143 lysosome-like vesicles and the early SCV could be part of a normal mechanism of vacuole maturation. When the vacuole reaches the stage of the maturing SCV, it contains fewer lysosomal proteins, such as VAMP7. One possible explanation is that Salmonella could control recycling pathways in the host cell to remove unwanted proteins from the SCV in order to avoid degradation, as previously suggested (Bujny et al., 2008). Moreover, interaction with the ER might modulate SCV maturation, as the acquisition of ER-derived membranes to phagosomes was shown to alter normal phagosomal maturation in dendritic cells (Cebrian et al., 2011). Interestingly, we show here that interactions between SCVs and VAMP7positive vesicles do not play a role in the establishment of the early and maturing SCV. Surprisingly, at later time-points (>5 h p.i.), we verified that VAMP7-positive but LysoTracker-negative vesicles are recruited to the late SCV and also to SIFs. Our data also show that VAMP7 is important for SIF formation and Salmonella replication exclusively within the late SCV. Due to the role of VAMP7 on vesicle fusion, we hypothesize that this cellular mechanism could provide membrane for SIF elongation and bacterial replication. Despite the late SCV being again enriched in lysosomal proteins such as VAMP7 and Lamp-1, it is LysoTracker negative. This confirms the hypothesis that the early SCV interacts with lysosomes, and that Salmonella actively reduces the lysosomal content/activity of its replicative niche during vacuole maturation (McGourty et al., 2012). Moreover, as SIFs display reduced acidity, they might be involved in diluting the lysosomal proteins that are delivered to the SCV (Schroeder et al., 2011). In the future, it will be interesting to identify bacterial effectors that control the successive cycles of interactions between the SCV and the lysosomal system. It could also be informative to compare our data with SCV remodelling in macrophages, as SCVlysosome fusions have been reported in those cells (Oh et al., 1996), or with intestinal epithelia, where the mechanisms of SCV maturation are largely unknown. In summary, quantitative proteomic analysis of the SCV combined with cell biology techniques revealed that this unique and specialized organelle interacts with several host cell compartments. We demonstrate that Salmonella growth within epithelial cells is regulated by the interactions between the SCV and either the ER or the lysosomal system, in distinct ways. ER-derived COPIIvesicle activity promotes SCV rupture and Salmonella hyper-replication within the cytosol, while successive interactions between the SCV and VAMP7-positive vesicles regulate bacterial growth within the SCV, by promoting formation of SIFs. Experimental procedures Bacterial strains The following Salmonella strains were used: SL1344 (wild type), SL1344 pM965 (Salmonella-GFP) and SL1344 expressing dsRed (Salmonella-dsRed). Bacteria were grown in lysogeny broth (LB) medium supplemented with 0.3 M NaCl at 37°C in an orbital shaker. LB was supplemented with streptomycin (50 μg ml−1) and, when appropriate, with ampicillin (50 μg ml−1). Plasmids, siRNAs and cell transfection HeLa cells were plated either on 12-well plated containing glass coverslips (1 ×105cells per well) or into 96-well glass bottom plates (Greiner) (7 ×103cells per well) 24 h before plasmid transfection. Cells were then transfected with one or two expression plasmids using the X-tremeGENE 9 DNA Transfection Reagent (Roche) for 24 or 48 h, according to the manufacturer’s instructions. The pRFP-VAMP7 plasmid was kindly provided by Thierry Galli (Institut Jacques Monod, Université Paris 7). The plasmids encoding the CFP-tagged Sar1, Sar1[T39N] and Sar1[H79G] were a kind gift from Franck Perez (Institut Curie). All siRNAs SMARTpool were obtained from Dharmacon: VAMP7 (6845), Sec13 (6396), non-specific non-targeting pool. Cells were reversed transfected for 72 h with Lipofectamine RNAiMAX (Life Technologies) reagent according to the manufacturer’s instructions. Protein knockdown efficiency was assessed by Western blot, by lysing cells with RIPA buffer at 4°C. Equal protein Fig. 7. VAMP7 recruitment to the late SCV is important for bacterial replication within the vacuole and SIF formation. A. HeLa cells were treated with scramble or VAMP7 siRNA and infected with Salmonella. Gentamicin was added to kill all extracellular bacteria. Then, cells were lysed at the indicated time-points of infection and the number of viable intracellular bacteria was counted by CFUs. The graph shows the mean ±SEM number of bacteria normalized to the number of cells counted for each siRNA after trypsinization of a non-infected well. P-values were determined using the two-way analysis of variance test for multiple comparisons from three independent experiments performed in triplicate. VAMP7-knockdown level was evaluated by Western blot as described in the Experimental procedures. B. Cells were double transfected with VAMP7-RFP and Lamp-1-GFP and infected with Salmonella for 16 h. Cells were imaged by confocal microscopy without fixation, in order to preserve SIFs structure. C. Cells were treated with scramble or VAMP7 siRNA and, 24 h later, transfected with Lamp-1-mCherry. After infection with Salmonella-GFP for 16 h, the number of infected cells containing SIFs was counted by confocal microscopy. Results presented in the graph show the mean ±SEM percentage of infected cells with SIFs from three independent experiments performed in triplicate. Thirty infected cells per condition were counted. P-values were determined using the Student’s t-test. D. Model for VAMP7 and COPII contribution to SCV maturation and Salmonella growth in epithelial cells. The early SCV is enriched and fuses with acidic lysosome-like vesicles (red with purple fill). The lysosomal content, including VAMP7, is then temporarily depleted from the maturing SCV. At later stages, VAMP7-positive vesicles with decreased acidity (red) associate again with the late SCV, promoting SIF formation and Salmonella replication within the vacuole. There is also contact between the early SCV and the host cell ER. COPII complex (green) assembly and accumulation around the early SCV promote vacuolar rupture, bacterial escape into the host cytosol and hyper-replication. Scale bars represent 10 μm. SCV interaction with the ER and lysosomes 15 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 144 amounts were separated in a SDS-PAGE gel, transferred to a nitrocellulose membrane and immunoblotted with Sec13 or VAMP7 specific antibodies. An actin antibody was used as loading control. Cell culture and infection assays All cell culture reagents were purchased from Invitrogen unless otherwise stated. Human epithelial HeLa cells (clone CCL-2 from the American Type Culture Collection (ATCC)) were cultured in Dulbecco’s modified eagle’s medium supplemented with 10% (v/v) fetal bovine serum (FBS) at 37°C, 5% CO2. All live-cell fluorescence microscopy and infection assays were performed in EM buffer (120 mM NaCl, 7 mM KCl, 1.8 mM CaCl2, 0.8 mM MgCl2, 5 mM glucose, 25 mM HEPES, pH 7.3). For invasion experiments, overnight bacterial cultures were subcultured 1/20 and grown until late exponential/early stationary phase. Before infection, bacteria were gently washed with PBS and resuspended in EM buffer. Except for live-cell microscopy, bacteria were added to the cells at a multiplicity of infection (MOI) of 100, incubated for 5 min at room temperature and then at 37°C for 10 min, so that a synchronized infection could be followed. Non-internalized bacteria were washed three times with warm EM buffer and incubated up to 30 min at 37°C. Extracellular bacteria were killed by adding EM containing 50 μg ml−1gentamicin for 1 h. The concentration of gentamicin was then decreased to 10 μg ml−1for the remainder of the experiment and 10% FBS was added to the medium. At the desired time-points, the cells were either processed for fractionation, enumeration of intracellular bacteria or fixed for immunofluorescence analysis. To count the number of intracellular bacteria, infectedcells were gently washed with PBS andlysed with ice-cold distilled water containing 0.2% Triton X-100 for 5 min. Bacteria were then serially diluted and plated onto LB agar. Cell fractionation and isolation of the SCV For the isolation of the SCV, approximately 6 ×107HeLa cells were used in T225 flasks. Cells were infected as described before and, at the selected time-points (30 min or 3 h), they were extensively washed with ice-cold homogenization buffer (HB: 250 mM sucrose, 0.5 mM ethylene glycol tetraacetic acid (EGTA), 20 mM HEPES-KOH pH 7.4, supplemented with complete protease inhibitors and 5 μg ml−1cytochalasin-D), detached and then homogenized in HB with a Dounce homogenizer. All steps were performed at 4°C. Between 30 and 40 strokes were performed, until more than 80% free nuclei were visible. Nuclei and intact cells were removed by performing three sequential centrifugations, at 100 g (1000 rpm) for 5 min each, in order to obtain the PNS. In parallel, a non-infected control was prepared. In order to separate all the subcellular organelles, the PNS was loaded on top of a 10–25% (1.06–1.15 g cm−3) linear OptiPrep (Sigma) gradient with a 50% (1.22 g cm−3) cushion, in a 14 ×89 mm ultracentrifuge tube (Beckman), and then centrifuged at 210000 g (35000 rpm) for 4 h at 4°C, in a SW-41 swinging bucket rotor, with low acceleration and slow brake. Approximately 5 ×107bacteria were added to the non-infected control PNS. After ultracentrifugation, 1.0 ml fractions were collected from the top to the bottom of the gradient and each fraction was analysed by measuring its refractive index, the number of bacteria by CFUs and the organelle separation efficiency by Western blot. Equal amount of proteins in each fraction were analysed by Western blot. The following primary antibodies were used: mouse anti-EEA-1 (1:2500, BD Biosciences), mouse antiRab5 (1:2000, BD Biosciences), rabbit anti-Lamp-1 (1:2000, Abcam), mouse anti-GM130 (1:1000, BD Biosciences), rabbit anti-Catalase (1:2500, Abcam), mouse anti-Calreticulin (1:2000, Abcam), mouse anti-TOM22 (1:2000, Sigma), rabbit anti-VAMP7 (1:2000, Pierce), rabbit anti-Salmonella Lipopolysaccharide (LPS) (1:20000, Abcam) and mouse anti-reticulon-4 (1:2000, Thermo-Scientific). The secondary antibodies were diluted 1:10000 (anti-mouse-HRP and anti-rabbit-HRP, Amersham). ELISA for quantification of intact SCVs The bottom of the wells of an ELISA plate (Nunc) was coated with a polyclonal rabbit anti-Salmonella antibody (Abcam) in PBS and incubated overnight at 4°C. Blocking was carried out by incubating the wells with a 2% BSA solution at room temperature for 90 min. Samples from the PNS or from F6 to F8 were added to the wells and incubated for 1 h at room temperature. As a control, an aliquot from the same sample was subjected to osmotic shock by incubation with distilled water. Standard curves were generated by adding known amounts of bacteria in 1:2 dilution series. Bacteria were then detected by incubation with the same antibody, biotinylated in 2% BSA. Signal was quantified at 450 nm after sequential incubation with streptavidin peroxidase (Sigma) and o-Phenylenediamine dihydrochloride (Sigma). Sample preparation for MS, data processing and analysis Each experiment was carried out in five biological replicates. After density centrifugation, 150 μl of fraction F6, which was highly enriched is intact SCVs, was subjected to methanol/ chloroform protein precipitation (Wessel and Flügge, 1984). Proteins were then separated in by one-dimensional SDS-PAGE in order to eliminate OptiPrep contaminations, each lane was cut into 10 slices and in-gel tryptic digestion was performed as described previously (Wilm et al., 1996). Peptides were finally extracted in 50 mM NH4HCO3/acetonitrile/formic acid (42.5/42.5/ 5), dried down and reconstituted in H2O/acetonitrile/formic acid (98/2/0.1) before LC-MS/MS analysis. Tryptic digests were analysed by nanoLC-MS/MS using an Ultimate 3000 RSLC system (Dionex, Thermo-Scientific, Waltham, MA, USA) coupled to the nanoelectrospray ion source of a Q-Exactive mass spectrometer (Thermo-Scientific, Bremen, Germany). One microgram of each digest was loaded on a C-18 μ-precolumn (C-18 PepMap100, 5μm, 100 Å, Dionex, Thermo-Scientific, Waltham, MA, USA) at a flow rate of 30 μl min−1of solvent A and the separation was performed using an in-house packed 15 cm nano-HPLC column (75 μm inner diameter) with C-18 resin (3 μm particles, 100 Å pore size, Reprosil C-18, Dr. Maisch GmbH). Peptides were separated at a flow rate of 300 nl min−1using a gradient of 2% to 55% solvent B for 30 min, followed by a 10 min washing step at 100% solvent B and a reconditioning step at 2% B for 20 min. Solvent A was H2O/acetonitrile/formic acid (98/2/0.1) and solvent B was H2O/acetonitrile/formic acid (20/80/0.08). NanoLC-MS/MS experiments were conducted in data-dependent acquisition mode. A resolution of 70 000 (at m/z 400) was used for MS scans. The 10 most intense ions were selected for HCD frag16 J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 145 mentation and fragments were analysed in the orbitrap. A dynamic exclusion window of 30 s was used. Raw files were processed with MaxQuant software (version 1.3.0.5) (Cox and Mann, 2008; Luber et al., 2010). Protein identification was carried out using Andromeda (Cox et al., 2011) against a concatenated database including Salmonella strain SL1344 proteins (tax 216597 – 4657 proteins) and human proteins (tax 9606 – 20233 proteins). Trypsin was chosen as specific enzyme with a maximum number of two miscleavages. Possible modifications included carbamidomethylation (Cys, fixed), oxidation (Met, variable) and N-terminal acetylation (variable). Mass tolerance for MS was set to 20 ppm for the first search then 6 ppm for the main search and 10 ppm was used for MS/MS. The ‘match between run’ option was selected with a maximal retention time window of 2 min. Five amino acids were required as minimum peptide length. A false discovery rate of 1% was used for the identification. Salmonella, reverse and contaminant proteins were excluded and only proteins identified with a minimum of two peptides were considered. Statistical relative protein quantification was performed with the peptide intensities extracted from the ‘peptides.txt’ MaxQuant output file using MSstats R package (Choi et al., 2014). MSstats enables protein significance analysis between different conditions (here control, 30 min and 3 h p.i.) and statistical protein quantification from label-free LC-MS experiments. The gene ontology analysis was performed with DAVID/EASE tools (Huang et al., 2009a,b) (http:// david.abcc.ncifcrf.gov/). Microscopy and image analysis Time-lapse microscopy of living cells was performed at 37°C in a PerkinElmer UltraView spinning disk confocal microscope with a 40 ×/1.3 NA oil objective. Every 10, 20 or 30 min, a stack of 15 z-planes (500 nm step size) was acquired sequentially. In 4% paraformaldehyde (PFA)-fixed samples, the following antibodies were used after cell permeabilization with 0.5% saponin for 10 min: mouse anti-Sec13 (1:100, Abnova), mouse antireticulon-4 (1:300, Thermo Scientific), rabbit anti-calnexin (1:100, Stressgen), rabbit anti-Lamp-1 (1:500, Abcam), Alexa Fluor 488conjugated anti-mouse or anti-rabbit (1:200, Life Technologies) and Cy5-conjugated anti-rabbit (1:200, Life Technologies). Z-stacks of 300 nm step size were acquired with a 60 ×/1.2 NA water objective. The following excitation lasers were used: 405, 488, 561 or 640 nm. Fluorescence emission was detected with 445 (W60), 525 (W50), 615 (W70) or 705 (W90) nm filters respectively. All images were further analysed with ImageJ and FiJi softwares. All confocal microscopy derived images shown correspond to maximum 3D projections. Correlative-focused ion beam/scanning electron large-volume tomography HeLa cells grown on MatTek dishes with a finder grid were fixed with 0.1% glutaraldheyde (GA) and 4% PFA for 15 min. After confocal microscopy imaging using a 60 ×/1.3 NA water objective, positions of interest were marked at 10 ×magnifications. Then cells were fixed overnight with 2.5% GA in 0.1 M HEPES pH 7.2 and post fixed in 1% osmium/1.5% potassium ferrocyanide in 0.1M HEPES for 1 h. Samples were treated for 30 min with 1% tannic acid and 1 h with 1% osmium tetroxide, rinsed in water and dehydrated in an ethanol series of 25%, 50%, 75%, 90% and 100% (5 min each). Cells were embedded in hard Epon resin. After resin polymerization, Epon blocks were mounted on a SEM stub and coated with a 25 nm layer of gold/palladium. Samples were placed in an Auriga FIB/SEM system (Zeiss) and the site of interest previously visualized by light microscopy was relocated using the imprint of the gridded dish in the Epon. Site was prepared and data acquired using ATLAS 3D software (Zeiss) using the backscatter detector at 2 kV with pixel sizes of 5 nm (Fig. 2E and F) or 10 nm (Fig. 5 and Fig. S4B). Slice thickness was 10 nm in all acquisitions. Stack alignment was performed with ImageJ and 3D visualization, data correlation, manual segmentation and video clips with Amira (FEI). Quantification of cells containing hyper-replicating bacteria Identification of cells containing hyper-replicating bacteria was performed using fluorescence microscopy. HeLa cells were infected with dsRed-expressing Salmonella and fixed at 6 h p.i., stained with DRAQ5 to enable computer segmentation and imaged on a Perkin Elmer, Opera spinning disk microscope. Images were analysed using Columbus software (Perkin Elmer) to quantify the red intensity (representing bacteria) per cell and defining intensity thresholds for uninfected, infected and hyperreplicating cells. The percentage of hyper-replication was defined as the number of cells containing hyper-replicating bacteria divided by the total number of infected cells. Statistical analysis Statistical analyses were performed in GraphPad Prism software v6, or MSstats R package for MS data. Significance was referred as *, **, *** and **** for P-values <0.05, <0.01, <0.001 and <0.0001 respectively. Acknowledgements We would like to thank all the members of the Dynamics of host-pathogen interaction (DIHP) unit, Paul Lazarow and Randy Schekman for productive discussions. We thank T. Gally and F. Perez for fluorescent constructs. This work was supported by a PhD fellowship from the Portuguese Fundação para a Ciência e a Tecnologia (SFRH/BD/51006/2010) to JCS; a Pasteur Foundation Fellowship to JF; and fellowships from the PasteurWeizmann Council, EMBO and the Fondation pour la Recherche Médicale to AW. JE is member of the LabEx consortium IBEID and is supported by the Institut Pasteur CARNOT-MIE and the PTR (grant number PTR-460) programmes. JE also acknowledges support of an European Research Council (ERC) starting grant for this work (Rupteffects, Nr. 261166). The platform microscopie ultrastructurale was supported by France Bioimaging. Author contributions JCS, JC-R and JE designed the study. JCS performed all experiments with the help of JF and AW, except mass SCV interaction with the ER and lysosomes 17 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 146 spectrometry performed by MD. AM and CS performed electron microscopy sample preparation and data acquisition. Data were analysed by JCS, MD, JF, AW, MM and VH. JCS and JE wrote the manuscript. Conflict of interest The authors declare that they have no conflict of interest. References Agbor, T.A., and McCormick, B.A. (2011) Salmonella effectors: important players modulating host cell function during infection. Cell Microbiol 13: 1858–1869. Akimana, C., Al-Khodor, S., and Abu Kwaik, Y. (2010) Host factors required for modulation of phagosome biogenesis and proliferation of Francisella tularensis within the cytosol. PLoS ONE 5: e11025. Bakowski, M.A., Braun, V., and Brumell, J.H. (2008) Salmonella-containing vacuoles: directing traffic and nesting to grow. Traffic 9: 2022–2031. Boucrot, E., Henry, T., Borg, J.-P., Gorvel, J.-P., and Méresse, S. (2005) The intracellular fate of Salmonella depends on the recruitment of kinesin. Science 308: 1174– 1178. Braun, V., Wong, A., Landekic, M., Hong, W.J., Grinstein, S., and Brumell, J.H. (2010) Sorting nexin 3 (SNX3) is a component of a tubular endosomal network induced by Salmonella and involved in maturation of the Salmonella-containing vacuole. Cell Microbiol 12: 1352– 1367. Brumell, J.H., Tang, P., Mills, S.D., and Finlay, B.B. (2001) Characterization of Salmonella-induced filaments (Sifs) reveals a delayed interaction between Salmonellacontaining vacuoles and late endocytic compartments. Traffic 2: 643–653. Buchmeier, N.A., and Heffron, F. (1991) Inhibition of macrophage phagosome-lysosome fusion by Salmonella typhimurium.Infect Immun 59: 2232–2238. Budnik, A., and Stephens, D.J. (2009) ER exit sites – localization and control of COPII vesicle formation. FEBS Lett 583: 3796–3803. Bujny, M.V., Ewels, P.A., Humphrey, S., Attar, N., Jepson, M.A., and Cullen, P.J. (2008) Sorting nexin-1 defines an early phase of Salmonella-containing vacuole-remodeling during Salmonella infection. J Cell Sci 121: 2027–2036. Campbell-Valois, F.-X., Trost, M., Chemali, M., Dill, B.D., Laplante, A., Duclos, S., et al. (2012) Quantitative proteomics reveals that only a subset of the endoplasmic reticulum contributes to the phagosome. Mol Cell Proteomics 11: M111.016378. Campoy, E.M., Mansilla, M.E., and Colombo, M.I. (2013) Endocytic SNAREs are involved in optimal Coxiella burnetii vacuole development. Cell Microbiol 15: 922–941. Cebrian, I., Visentin, G., Blanchard, N., Jouve, M., Bobard, A., Moita, C., et al. (2011) Sec22b regulates phagosomal maturation and antigen crosspresentation by dendritic cells. Cell 147: 1355–1368. Choi, M., Chang, C.-Y., Clough, T., Broudy, D., Killeen, T., MacLean, B., and Vitek, O. (2014) MSstats: an R package for statistical analysis of quantitative mass spectrometrybased proteomic experiments. Bioinformatics 30: 2524– 2526. Cox, J., and Mann, M. (2008) MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat Biotechnol 26: 1367–1372. Cox, J., Neuhauser, N., Michalski, A., Scheltema, R.A., Olsen, J.V., and Mann, M. (2011) Andromeda: a peptide search engine integrated into the MaxQuant environment. J Proteome Res 10: 1794–1805. Criss, A.K., and Casanova, J.E. (2003) Coordinate regulation of Salmonella enterica serovar Typhimurium invasion of epithelial cells by the Arp2/3 complex and Rho GTPases. Infect Immun 71: 2885–2891. Davis, M.J., Gregorka, B., Gestwicki, J.E., and Swanson, J.A. (2012) Inducible renitence limits Listeria monocytogenes escape from vacuoles in macrophages. J Immunol 189: 4488–4495. Delevoye, C., Nilges, M., Dehoux, P., Paumet, F., Perrinet, S., Dautry-Varsat, A., and Subtil, A. (2008) SNARE protein mimicry by an intracellular bacterium. PLoS Pathog 4: e1000022. Desjardins, M., Huber, L.A., Parton, R.G., and Griffiths, G. (1994) Biogenesis of phagolysosomes proceeds through a sequential series of interactions with the endocytic apparatus. J Cell Biol 124: 677–688. Drecktrah, D., Knodler, L.A., Howe, D., and Steele-Mortimer, O. (2007) Salmonella trafficking is defined by continuous dynamic interactions with the endolysosomal system. Traffic 8: 212–225. Drecktrah, D., Levine-Wilkinson, S., Dam, T., Winfree, S., Knodler, L.A., Schroer, T.A., and Steele-Mortimer, O. (2008) Dynamic behavior of Salmonella-induced membrane tubules in epithelial cells. Traffic 9: 2117– 2129. Eden, E.R., White, I.J., Tsapara, A., and Futter, C.E. (2010) Membrane contacts between endosomes and ER provide sites for PTP1B-epidermal growth factor receptor interaction. Nat Cell Biol 12: 267–272. Ehsani, S., Santos, J.C., Rodrigues, C.D., Henriques, R., Audry, L., Zimmer, C., et al. (2012) Hierarchies of host factor dynamics at the entry site of Shigella flexneri during host cell invasion. Infect Immun 80: 2548–2557. Elsinghorst, E.A. (1994) Measurement of invasion by gentamicin resistance. Methods Enzymol 236: 405– 420. English, A.R., and Voeltz, G.K. (2013) Endoplasmic reticulum structure and interconnections with other organelles. Cold Spring Harb Perspect Biol 5: a013227. Fajardo, M., Schleicher, M., Noegel, A., Bozzaro, S., Killinger, S., Heuner, K., et al. (2004) Calnexin, calreticulin and cytoskeleton-associated proteins modulate uptake and growth of Legionella pneumophila in Dictyostelium discoideum.Microbiology 150: 2825–2835. Figueira, R., and Holden, D.W. (2012) Functions of the Salmonella pathogenicity island 2 (SPI-2) type III secretion system effectors. Microbiology 158: 1147–1161. Friedman, J.R., and Voeltz, G.K. (2011) The ER in 3D: a multifunctional dynamic membrane network. Trends Cell Biol 21: 709–717. 18 J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 147 Gagnon, E., Duclos, S., Rondeau, C., Chevet, E., Cameron, P.H., Steele-Mortimer, O., et al. (2002) Endoplasmic reticulum-mediated phagocytosis is a mechanism of entry into macrophages. Cell 110: 119–131. Garcia-del Portillo, F., and Finlay, B.B. (1995) Targeting of Salmonella typhimurium to vesicles containing lysosomal membrane glycoproteins bypasses compartments with mannose 6-phosphate receptors. J Cell Biol 129: 81–97. Guermonprez, P., Saveanu, L., Kleijmeer, M., Davoust, J., Van Endert, P., and Amigorena, S. (2003) ER-phagosome fusion defines an MHC class I cross-presentation compartment in dendritic cells. Nature 425: 397–402. Haraga, A., Ohlson, M.B., and Miller, S.I. (2008) Salmonellae interplay with host cells. Nat Rev Microbiol 6: 53–66. Harrison, R.E., Brumell, J.H., Khandani, A., Bucci, C., Scott, C.C., Jiang, X., et al. (2004) Salmonella impairs RILP recruitment to Rab7 during maturation of invasion vacuoles. Mol Biol Cell 15: 3146–3154. Hashim, S., Mukherjee, K., Raje, M., Basu, S.K., and Mukhopadhyay, A. (2000) Live Salmonella modulate expression of Rab proteins to persist in a specialized compartment and escape transport to lysosomes. J Biol Chem 275: 16281–16288. Haucke, V. (2003) Vesicle budding: a coat for the COPs. Trends Cell Biol 13: 59–60. Hänisch, J., Ehinger, J., Ladwein, M., Rohde, M., Derivery, E., Bosse, T., et al. (2010) Molecular dissection of Salmonella-induced membrane ruffling versus invasion. Cell Microbiol 12: 84–98. Hernandez, L.D., Hueffer, K., Wenk, M.R., and Galán, J.E. (2004) Salmonella modulates vesicular traffic by altering phosphoinositide metabolism. Science 304: 1805– 1807. Hoffmann, C., Finsel, I., Otto, A., Pfaffinger, G., Rothmeier, E., Hecker, M., et al. (2014) Functional analysis of novel Rab GTPases identified in the proteome of purified Legionella-containing vacuoles from macrophages. Cell Microbiol 16: 1034–1052. Houde, M., Bertholet, S., Gagnon, E., Brunet, S., Goyette, G., Laplante, A., et al. (2003) Phagosomes are competent organelles for antigen cross-presentation. Nature 425: 402–406. Huang, D.W., Sherman, B.T., and Lempicki, R.A. (2009a) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4: 44–57. Huang, D.W., Sherman, B.T., and Lempicki, R.A. (2009b) Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37: 1–13. Hughes, H., and Stephens, D.J. (2008) Assembly, organization, and function of the COPII coat. Histochem Cell Biol 129: 129–151. Jo, E.-K., Yuk, J.-M., Shin, D.-M., and Sasakawa, C. (2013) Roles of autophagy in elimination of intracellular bacterial pathogens. Front Immunol 4: 97. Jolly, C., Winfree, S., Hansen, B., and Steele-Mortimer, O. (2014) The Annexin A2/p11 complex is required for efficient invasion of Salmonella Typhimurium in epithelial cells. Cell Microbiol 16: 64–77. Kagan, J.C., and Roy, C.R. (2002) Legionella phagosomes intercept vesicular traffic from endoplasmic reticulum exit sites. Nat Cell Biol 4: 945–954. Knodler, L.A., Vallance, B.A., Celli, J., Winfree, S., Hansen, B., Montero, M., and Steele-Mortimer, O. (2010) Dissemination of invasive Salmonella via bacterial-induced extrusion of mucosal epithelia. Proc Natl Acad Sci USA 107: 17733–17738. Knodler, L.A., Nair, V., and Steele-Mortimer, O. (2014) Quantitative assessment of cytosolic Salmonella in epithelial cells. PLoS ONE 9: e84681. Kurokawa, K., Okamoto, M., and Nakano, A. (2014) Contact of cis-Golgi with ER exit sites executes cargo capture and delivery from the ER. Nat Commun 5: 3653. Lee, B.-Y., Jethwaney, D., Schilling, B., Clemens, D.L., Gibson, B.W., and Horwitz, M.A. (2010) The Mycobacterium bovis bacille Calmette-Guerin phagosome proteome. Mol Cell Proteomics 9: 32–53. Lord, C., Ferro-Novick, S., and Miller, E.A. (2013) The highly conserved COPII coat complex sorts cargo from the endoplasmic reticulum and targets it to the golgi. Cold Spring Harb Perspect Biol 5: a013367. Lu, H., and Clarke, M. (2005) Dynamic properties of Legionella-containing phagosomes in Dictyostelium amoebae.Cell Microbiol 7: 995–1007. Luber, C.A., Cox, J., Lauterbach, H., Fancke, B., Selbach, M., Tschopp, J., et al. (2010) Quantitative proteomics reveals subset-specific viral recognition in dendritic cells. Immunity 32: 279–289. Luzio, J.P., Pryor, P.R., and Bright, N.A. (2007) Lysosomes: fusion and function. Nat Rev Mol Cell Biol 8: 622–632. Luzio, J.P., Gray, S.R., and Bright, N.A. (2010) Endosomelysosome fusion. Biochem Soc Trans 38: 1413–1416. McGourty, K., Thurston, T.L., Matthews, S.A., Pinaud, L., Mota, L.J., and Holden, D.W. (2012) Salmonella inhibits retrograde trafficking of mannose-6-phosphate receptors and lysosome function. Science 338: 963–967. Majowicz, S.E., Musto, J., Scallan, E., Angulo, F.J., Kirk, M., O’Brien, S.J., et al. (2010) The global burden of nontyphoidal Salmonella gastroenteritis. Clin Infect Dis 50: 882–889. Malik-Kale, P., Winfree, S., and Steele-Mortimer, O. (2012) The bimodal lifestyle of intracellular Salmonella in epithelial cells: replication in the cytosol obscures defects in vacuolar replication. PLoS ONE 7: e38732. Mallo, G.V., Espina, M., Smith, A.C., Terebiznik, M.R., Alemán, A., Finlay, B.B., et al. (2008) SopB promotes phosphatidylinositol 3-phosphate formation on Salmonella vacuoles by recruiting Rab5 and Vps34. J Cell Biol 182: 741–752. Mellouk, N., Weiner, A., Aulner, N., Schmitt, C., Elbaum, M., Shorte, S.L., et al. (2014) Shigella subverts the host recycling compartment to rupture its vacuole. Cell Host Microbe 16: 517–530. Mota, L.J., Ramsden, A.E., Liu, M., Castle, J.D., and Holden, D.W. (2009) SCAMP3 is a component of the Salmonellainduced tubular network and reveals an interaction between bacterial effectors and post-Golgi trafficking. Cell Microbiol 11: 1236–1253. Müller-Taubenberger, A., Lupas, A.N., Li, H., Ecke, M., Simmeth, E., and Gerisch, G. (2001) Calreticulin and SCV interaction with the ER and lysosomes 19 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 148 calnexin in the endoplasmic reticulum are important for phagocytosis. EMBO J 20: 6772–6782. Oh, Y.K., Alpuche-Aranda, C., Berthiaume, E., Jinks, T., Miller, S.I., and Swanson, J.A. (1996) Rapid and complete fusion of macrophage lysosomes with phagosomes containing Salmonella typhimurium.Infect Immun 64: 3877– 3883. Orci, L., Ravazzola, M., Le Coadic, M., Shen, W.-W., Demaurex, N., and Cosson, P. (2009) From the Cover: STIM1-induced precortical and cortical subdomains of the endoplasmic reticulum. Proc Natl Acad Sci USA 106: 19358–19362. Patel, J.C., and Galán, J.E. (2006) Differential activation and function of Rho GTPases during Salmonella-host cell interactions. J Cell Biol 175: 453–463. Paz, I., Sachse, M., Dupont, N., Mounier, J., Cederfur, C., Enninga, J., et al. (2010) Galectin-3, a marker for vacuole lysis by invasive pathogens. Cell Microbiol 12: 530–544. Ragaz, C., Pietsch, H., Urwyler, S., Tiaden, A., Weber, S.S., and Hilbi, H. (2008) The Legionella pneumophila phosphatidylinositol-4 phosphate-binding type IV substrate SidC recruits endoplasmic reticulum vesicles to a replication-permissive vacuole. Cell Microbiol 10: 2416– 2433. Ramsden, A.E., Holden, D.W., and Mota, L.J. (2007a) Membrane dynamics and spatial distribution of Salmonellacontaining vacuoles. Trends Microbiol 15: 516–524. Ramsden, A.E., Mota, L.J., Münter, S., Shorte, S.L., and Holden, D.W. (2007b) The SPI-2 type III secretion system restricts motility of Salmonella-containing vacuoles. Cell Microbiol 9: 2517–2529. Rao, P.K., Singh, C.R., Jagannath, C., and Li, Q. (2009) A systems biology approach to study the phagosomal proteome modulated by mycobacterial infections. Int J Clin Exp Med 2: 233–247. Rathman, M., Barker, L.P., and Falkow, S. (1997) The unique trafficking pattern of Salmonella typhimurium-containing phagosomes in murine macrophages is independent of the mechanism of bacterial entry. Infect Immun 65: 1475– 1485. Robinson, C.G., and Roy, C.R. (2006) Attachment and fusion of endoplasmic reticulum with vacuoles containing Legionella pneumophila.Cell Microbiol 8: 793–805. Rocha, N., Kuijl, C., van der Kant, R., Janssen, L., Houben, D., Janssen, H., et al. (2009) Cholesterol sensor ORP1L contacts the ER protein VAP to control Rab7-RILP-p150 Glued and late endosome positioning. J Cell Biol 185: 1209–1225. Rogers, L.D., and Foster, L.J. (2007) The dynamic phagosomal proteome and the contribution of the endoplasmic reticulum. Proc Natl Acad Sci USA 104: 18520–18525. Rogers, L.D., and Foster, L.J. (2008) Contributions of proteomics to understanding phagosome maturation. Cell Microbiol 10: 1405–1412. Rowland, A.A., Chitwood, P.J., Phillips, M.J., and Voeltz, G.K. (2014) ER contact sites define the position and timing of endosome fission. Cell 159: 1027–1041. Sato, K. (2004) COPII coat assembly and selective export from the endoplasmic reticulum. J Biochem 136: 755–760. Schindler, A.J., and Schekman, R. (2009) In vitro reconstitution of ER-stress induced ATF6 transport in COPII vesicles. Proc Natl Acad Sci USA 106: 17775–17780. Schroeder, N., Henry, T., de Chastellier, C., Zhao, W., Guilhon, A.-A., Gorvel, J.-P., and Méresse, S. (2010) The virulence protein SopD2 regulates membrane dynamics of Salmonella-containing vacuoles. PLoS Pathog 6: e1001002. Schroeder, N., Mota, L.J., and Méresse, S. (2011) Salmonella-induced tubular networks. Trends Microbiol 19: 268–277. Scott, C.C., Dobson, W., Botelho, R.J., Coady-Osberg, N., Chavrier, P., Knecht, D.A., et al. (2005) Phosphatidylinositol-4,5-bisphosphate hydrolysis directs actin remodeling during phagocytosis. J Cell Biol 169: 139–149. Shevchuk, O., Batzilla, C., Hägele, S., Kusch, H., Engelmann, S., Hecker, M., et al. (2009) Proteomic analysis of Legionella-containing phagosomes isolated from Dictyostelium.Int J Med Microbiol 299: 489–508. Smith, A.C., Heo, W.D., Braun, V., Jiang, X., Macrae, C., Casanova, J.E., et al. (2007) A network of Rab GTPases controls phagosome maturation and is modulated by Salmonella enterica serovar Typhimurium. J Cell Biol 176: 263–268. Steele-Mortimer, O. (2008) The Salmonella-containing vacuole: moving with the times. Curr Opin Microbiol 11: 38–45. Steele-Mortimer, O., Méresse, S., Gorvel, J.P., Toh, B.H., and Finlay, B.B. (1999) Biogenesis of Salmonella typhimuriumcontaining vacuoles in epithelial cells involves interactions with the early endocytic pathway. Cell Microbiol 1: 33–49. Stefan, C.J., Manford, A.G., and Emr, S.D. (2013) ER-PM connections: sites of information transfer and interorganelle communication. Curr Opin Cell Biol 25: 434–442. Stuart, L.M., Boulais, J., Charriere, G.M., Hennessy, E.J., Brunet, S., Jutras, I., et al. (2007) A systems biology analysis of the Drosophila phagosome. Nature 445: 95–101. Thurston, T.L., Wandel, M.P., von Muhlinen, N., Foeglein, A., and Randow, F. (2012) Galectin 8 targets damaged vesicles for autophagy to defend cells against bacterial invasion. Nature 482: 414–418. Urwyler, S., Nyfeler, Y., Ragaz, C., Lee, H., Mueller, L.N., Aebersold, R., and Hilbi, H. (2009) Proteome analysis of Legionella vacuoles purified by magnetic immunoseparation reveals secretory and endosomal GTPases. Traffic 10: 76–87. Walther, T.C., and Mann, M. (2010) Mass spectrometrybased proteomics in cell biology. J Cell Biol 190: 491–500. Ward, T.H., Polishchuk, R.S., Caplan, S., Hirschberg, K., and Lippincott-Schwartz, J. (2001) Maintenance of Golgi structure and function depends on the integrity of ER export. J Cell Biol 155: 557–570. Wessel, D., and Flügge, U.I. (1984) A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Anal Biochem 138: 141–143. Wilm, M., Shevchenko, A., Houthaeve, T., Breit, S., Schweigerer, L., Fotsis, T., and Mann, M. (1996) Femtomole sequencing of proteins from polyacrylamide gels by nano-electrospray mass spectrometry. Nature 379: 466–469. 20 J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 149 Zhou, D., Chen, L.M., Hernandez, L., Shears, S.B., and Galán, J.E. (2001) A Salmonella inositol polyphosphatase acts in conjunction with other bacterial effectors to promote host cell actin cytoskeleton rearrangements and bacterial internalization. Mol Microbiol 39: 248–259. Supporting information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Fig. S1. (Related to Fig. 1). Validation of the ELISA assay and density gradients quality. A–B. The bottom of the wells of an ELISA plate was coated with an anti-Salmonella antibody and known amounts of bacteria were added in 1:2 dilution series. Salmonella were then detected by incubation with a biotinylated anti-Salmonella antibody and streptavidin peroxidase. Signal was quantified at 450 nm, correlated to number of bacteria per ml and plotted as linear-linear (A) or log-log (B). C. After ultracentrifugation of the different PNSs, 12 fractions were collected (F1–F12) from the top to the bottom of the gradients. The refractive index was measured in each fraction, using a refractometer, and then the density (in g cm−3) was calculated. The graph shows the mean ±SEM from 15 different gradients from five independent experiments. Fig. S2. (Related to Fig. 1). Subcellular organelle distribution in the 3 h infection fractionation. All fractions F1 to F12 and the PNS were tested using markers for the following compartments: early endosomes (EEA-1 and Rab5), late endosomes and lysosomes (Lamp-1), Golgi (GM130), peroxisomes (catalase), mitochondria (TOM22) and ER (calreticulin). Salmonella distribution was assessed with a specific anti-Salmonella LPS antibody. Fig. S3. (Related to Fig. 2). Relative protein quantitative differences between the different protein subsets and gene ontology analysis. A–B. Volcano plots of relative protein abundance differences between the control fraction and the 30 min SCV fraction (A) or the 3 h SCV fraction (B), as a function of statistical significance. Relative protein abundances were considered different when the log2fold-change >0.387 or <−0.387, with a P-value <0.05. Proteins with no statistically significant abundance difference between conditions are represented in gray. Proteins enriched at the 30 min SCV are depicted in red (A), whereas proteins enriched at the 3 h SCV are shown in green (B). In both graphs, factors enriched in the control non-infected fraction are shown in blue. Values correspond to the average fold-changes of five independent experiments. C. The host proteins enriched in the SCV were subjected to gene ontology analysis and grouped according to their biological process. For each term, the analysis was performed for the proteins enriched uniquely either at the 30 min SCV (only 30 min SCV, red bars) or at the 3 h SCV (only 3 h SCV, green bars) or for the proteins enriched at both time-points (both 30 min +3 h SCV, orange bars). The graphs show the percentage of host proteins enriched at the SCV, relative to the total number of proteins identified in each condition. Statistics were performed by determining the P-values (EASE score) and show the robustness of gene-term enrichment for each condition. ND, non-enriched factors in comparison with the control. Fig. S4. (Related to Figs 3 and 4). The SCV contacts with the host cell ER and the COPII complex accumulates around intracellular Salmonella. A. Equal amounts of protein from subcellular fractions F3 to F7 were immunoblotted, using markers for reticulon-4 or VAMP7. B. Cells infected with Salmonella for 30 min were fixed and imaged by fluorescence confocal microscopy followed by FIB/ SEM. Several yz-views are shown, corresponding to different acquisition planes, and depict sites of membrane contact between the SCV and the ER (arrowheads). We show here a different data set from the one shown on Fig. 3E–F. (b, bacteria; s, SCV lumen). C. Cells were infected with Salmonella-dsRed, fixed at different time-points and immunostained for Sec13 (green). DNA was stained with DAPI (cyan). Representative confocal microscopy images show the association of Sec13 around intracellular bacteria. Arrows point to Sec13-positive bacteria. D. HeLa cells were infected with Salmonella-dsRed, fixed at different time-points and immunostained for Sec13. DNA was stained with DAPI (cyan). Images show representative infected cells with or without hyper-replicating Salmonella after confocal microscopy analysis. Arrows point to Sec13-positive structures resembling bacterial-shaped membrane remnants. E. The percentage of infected cells containing Sec13-positive bacterial-shaped membrane remnants was determined by counting 50 cells with or without hyper-replicating bacteria, from triplicate wells. Scale bars correspond to 10 μm. Fig. S5. (Related to Video Clips S2–S4). VAMP7-positive lysosome-like vesicles are recruited to the early SCV, which then gets depleted in acidic content. HeLa cells were transfected with VAMP7-RFP for 48 h, treated with LysoTracker deep red at 50 nM for 30 min and infected with GFP-expressing Salmonella at a MOI of 30. Infected cells were then imaged by time-lapse confocal microscopy. After 30 min cells were washed and 50 μg ml−1gentamicin was added for 1 h. The concentration of gentamicin was then decreased to 10 μg ml−1for the remainder of the experiment and 10% FBS was added to the medium. Arrows point to intracellular bacteria. Scale bar corresponds to 10 μm. Representative data from five independent experiments are shown. Fig. S6. (Related to Fig. 7 and Video Clip S6). VAMP7 does not affect Salmonella hyper-replication but is recruited to SIFs. A. HeLa cells were treated with scramble or VAMP7 siRNA for 72 h and infected with dsRed-expressing Salmonella for 6 h. Cells were then fixed, counterstained with DRAQ5 and imaged by fluorescence microscopy. Statistics were determined using the Student’s t-test. ns, not significant. Data from three independent experiments are shown. B. Cells were transfected with VAMP7-RFP and Lamp-1-GFP and infected with Salmonella. Gentamicin was added to kill extracellular bacteria and cells were then imaged by time-lapse confocal microscopy every 10 min. Arrows indicate SIF tubules. Representative data from three independent experiments are shown. Table S1. Relative protein abundances at the 30 min SCV compared with the non-infected control for the complete list of host proteins identified by proteomics. Table S2. Relative protein abundances at the 3 h SCV compared with the non-infected control for the complete list of host proteins identified by proteomics. SCV interaction with the ER and lysosomes 21 © 2015 John Wiley & Sons Ltd, Cellular Microbiology 150 Table S3. Complete list of host proteins enriched in the SCV at 30 min, 3 h or at both time-points. Table S4. Selected host proteins enriched in the SCV identified by proteomics. Video Clip S1. (Related to Fig. 3E and F – 38 MB; 1 min and 11 s). Ultrastructural characterization of ER interactions with the early SCV. Successive FIB/SEM sections are shown in the xy-view and show sites of membrane contact between SCVs and the ER (arrowheads). The bacteria (blue), the SCV lumen (yellow) and the ER (red) were segmented using Amira software. Video Clip S2. (Related to Fig. S5A – 771 KB; 20 s). HeLa cells were transfected with VAMP7-RFP (red) and lysosomes were stained with LysoTracker deep red (purple). Cells were then infected with Salmonella-GFP (green) and imaged by confocal microscopy. VAMP7and LysoTracker-positive vesicles accumulate around the early SCV, in the first 2 h of infection. Then the SCV loses both markers and by 5 h p.i. gets again enriched in VAMP7 but not LysoTracker. Scale bar corresponds to 10 μm. Video Clip S3. (Related to Fig. S5B – 1.3 MB; 30 s). HeLa cells were transfected with VAMP7-RFP (red) and lysosomes were stained with LysoTracker deep red (purple). Cells were then infected with Salmonella-GFP (green) and imaged by confocal microscopy. VAMP7and LysoTracker-positive vesicles accumulate around the early SCV, in the first 2 h of infection. Then the SCV loses both markers and by 5 h p.i. gets again enriched in VAMP7 but not LysoTracker. By 6 h p.i., bacterial replication is observed, together with VAMP7-positive tubules emanating from the SCV. Scale bar corresponds to 10 μm. Video Clip S4. (Related to Fig. S5C – 1.3 MB; 18 s). HeLa cells were transfected with VAMP7-RFP (red) and lysosomes were stained with LysoTracker deep red (purple). Cells were then infected with Salmonella-GFP (green) and imaged by confocal microscopy. VAMP7and LysoTracker-positive vesicles accumulate around the early SCV, in the first 2.5 h of infection. Then the SCV loses LysoTracker and by 6 h p.i., bacterial replication is observed, together with VAMP7-positive tubules emanating from the SCV. Scale bar corresponds to 10 μm. Video Clip S5. (Related to Fig. 6 – 49 MB; 1 min and 23 s). Ultrastructural characterization of the interactions between early SCVs and VAMP7/Lamp-1-positive vesicles. Fluorescence confocal microscopy of DAPI (blue), VAMP7 (red) and Lamp-1 (green) was followed by 3D FIB/SEM at the same cell. Segmentation of the bacteria (blue), vesicles and SCVs lumen (yellow) show that some VAMP7/Lamp-1-poisitive vesicles interact and fuse with the SCVs. Fluorescence staining of the bacteria was used as correlating fiducials. Video Clip S6. (Related to Fig. S6B – 1.3 MB; 27 s). HeLa cells were transfected with VAMP7-RFP and Lamp-1-GFP and infected with Salmonella. Lamp-1and VAMP7-positive SIFs emanate from the SCV 5 h p.i. Scale bar corresponds to 10 μm. 22 J. C. Santos et al. © 2015 John Wiley & Sons Ltd, Cellular Microbiology 151 Supplemental information from Manuscript 2 Inventory of Supplemental information Fig. S1 (related to Fig. 1) Fig. S2 (related to Fig. 1) Fig. S3 (related to Fig. 2) Fig. S4 (related to Fig. 3 and 4) Fig. S5 (related to Movies S2-S4) Fig. S6 (related to Fig. 7 and Movie S6) Movie S1 (related to Fig. 3E-3F) Movie S2 (related to Fig. S5A) Movie S3 (related to Fig. S5B) Movie S4 (related to Fig. S5C) Movie S5 (related to Fig. 6) Movie S6 (related to Fig. S6B) Table S1 Table S2 Table S3 Table S4