Université Pierre Et Marie Curie

Home , Curie (lunar crater)

Université Pierre et Marie Curie

Ecole doctorale Complexité du Vivant Unité de Biologie des Interactions Cellulaires / URA CNRS 2582

The intracellular pathogen Chlamydia trachomatis targets proteins of the ESCRT machinery

Par François VROMMAN

Thèse de doctorat de Microbiologie Cellulaire

Dirigée par Agathe SUBTIL

Présentée et soutenue publiquement le 10 juin 2014

Devant un jury composé de :

Pr. Vincent MARECHAL (Professeur UPMC) Président Dr. Reynaldo CARABEO (Directeur de laboratoire) Rapporteur Dr. Guy TRAN VAN NHIEU (Directeur de laboratoire) Rapporteur Dr. Nolwenn JOUVENET (Chargée de recherche) Examinatrice Dr. Agathe SUBTIL (Directrice de laboratoire) Directrice de thèse

ACKNOWLEDGMENTS

Je veux tout d’abord remercier la Pr. Alice Dautry qui a accepté de m’accueillir dans son laboratoire. Je la remercie pour ses remarques concernant mon travail malgré ses obligations de Directrice de l’Institut Pasteur, ainsi que pour ses conseils avisés notamment en communication. Je veux particulièrement remercier Agathe Subtil, ma directrice de thèse pour son implication dans ma thèse sur tous les points. Je suis reconnaissant de l’autonomie qu’elle a su me donner au cours de ces quatre années passées ici. Sa présence pour la préparation de communication et son travail de relecture, jusqu’à l’écriture de ce manuscrit, m’ont permis d’apprendre de mes erreurs et de progresser. Enfin je tiens à la remercier pour son soutien dans mes démarches extra-professionnelles (comme StaPa), dans le recrutement de stagiaires de l’université de Metz, ou encore pour les différentes conférences internationales auxquelles j’ai pu participer. Pour tout ça Agathe, MERCI. Merci à Stéphanie Perrinet pour son aide et son expertise technique, et surtout sa bonne humeur. En grande professionnelle, elle a su prendre sur elle même pour ne pas assommer ce « petit stagiaire » que j’étais en arrivant qui la bombardait de questions. Un grand merci aux membres du labo qui font vivre les couloirs vides du quatrième étage de Calmette. Natalie Sauvonnet pour ses conseils précieux, Marc et Goran pour leur aide lors de la dernière année et les petites discussions de fin de journée. Merci à Valérie, Lena, Mathilde, Natalia, Sandrine et Flora pour leur bonne humeur de tous les jours. Je tiens à remercier Alexandre Dufour pour son amitié d’abord, mais aussi pour son aide dans mon travail. Sa disponibilité pour les petits réglages de dernière minute sur ICY m’a toujours été d’un grand secours. Egalement je tiens à le remercier pour son soutien pendant mon année de « StaPa president » qui aurait été nettement plus difficile à boucler sans son expérience et ses conseils. Un grand merci à Chloé, la blonde ! Pour son amitié, les petits chablis du lundi, les lunch du vendredi et les brunch du dimanche… Sans oublier sa petite touche comptabilité qu’elle même ne suspectait pas avant de devenir trésorière de StaPa. Je tiens finalement à remercier toute ma famille pour leur soutien durant toutes ses années. Je ne serai jamais arrivé jusque là sans eux.

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 1! TABLE OF FIGURES ...... 4! ABBREVIATIONS ...... 5! ABSTRACT ...... 6! RESUME ...... 7! INTRODUCTION ...... 17! I.! Chlamydia trachomatis, a human pathogen ...... 18! 1.! Pathology & epidemiology: which diseases and where? ...... 18! a.! Trachoma ...... 18! b.! Urogenital infections ...... 19! c.! Lymphogranuloma venereum ...... 20! d.! Diagnosis and treatments ...... 20! 2.! A bit of history: once upon a time...... 21! 3.! Chlamydia, a bacterium apart from others ...... 22! 4.! The Chlamydia special: a biphasic developmental cycle ...... 24! a.! Characteristics of EBs and RBs ...... 25! b.! The very first step: the bacterial adhesion ...... 25! c.! Knock-knock: The Chlamydia entry ...... 26! d.! Home sweet home: Setting up the inclusion ...... 27! e.! Multiplication and development: the reticulate bodies party ...... 29! f.! The persistent form: a party forever? ...... 30! g.! Parasitism of the host cell by Chlamydia ...... 31! h.! The end of the show: a double way out ticket ...... 32! 5.! Chlamydia’s swiss knife: the secretion systems ...... 33! a.! The T2SS: the general secretory pathway (GSP) ...... 33! b.! The T5SS: the twin-arginine translocation (tat) pathway ...... 34! c.! The T3SS: the chlamydial “injectisome” ...... 34! 6.! The challenge Chlamydia: from darkness to light ...... 36! a.! A long time in the darkness… ...... 36!

1 b.! … and then there was light: welcome to a new Chlamydia world...... 37! II.! Interactions between Chlamydia and host membranes ...... 38! 1.! The first Chlamydia interaction: entering the cell ...... 38! 2.! Chlamydia fattens up: delivery of lipids to the inclusion ...... 40! 3.! A sociable guest: multiple interactions between Chlamydia and host organelles ...... 42! a.! Chlamydia attracts most if not all the host organelles ...... 42! b.! The Inc proteins: key components of the Chlamydia-host cell interaction 43! c.! Manipulation of the Rab proteins, key trafficking regulators ...... 44! 4.! Chlamydia trachomatis infection impairs cytokinesis ...... 46! 5.! The defensive answers of the host ...... 46! 6.! Chlamydia and ESCRT ...... 47! III.! Biology of the ESCRT system ...... 48! 1.! Overview of a machinery with many complexes ...... 48! 2.! One system, many functions ...... 50! a.! MVB biogenesis ...... 50! b.! Abscission during cytokinesis ...... 53! c.! Virus budding ...... 53! d.! Exosomes secretion ...... 54! e.! Autophagy ...... 55! 3.! Focus on Hrs ...... 56! a.! From Hrs to the ESCRT-0 complex ...... 56! b.! Dissection of Hrs: domains, interactions, regulations ...... 56! c.! Hrs, a multi-functions protein ...... 58! 4.! Focus on Tsg101: a central protein for the ESCRT system ...... 59! 5.! ESCRT system, where else? ...... 59! MATERIALS & METHODS ...... 61! I.! Chlamydia strains ...... 62! II.! Cell culture, transfection and chemicals ...... 62! III.! Cloning procedures ...... 63! IV.! Production and purification of recombinant protein ...... 64! V.! Yeast-two-hybrid (Y2H) assays ...... 64! VI.! Immunofluorescence and western blotting analysis ...... 65! VII.! Quantification of bacterial entry ...... 66!

2 VIII.! Flow cytometry on infected cells and EBs...... 67! IX.! Immunoprecipitation (IP) ...... 68! RESULTS ...... 69! I.! Article 1: Identification of a family of effectors secreted by the type III secretion system that are conserved in pathogenic chlamydiae ...... 70! II.! Article 2: Monitoring of Chlamydia trachomatis developmental cycle using GFP-expressing bacteria, microscopy and flow cytometry ...... 81! III.! The DUF582 protein CT619 targets the ESCRT proteins Hrs and Tsg101 during Chlamydia trachomatis infection ...... 106! 1.! CT619 interacts with Hrs and Tsg101 ...... 106! a.! CT619 is expressed late in the cycle ...... 106! b.! Two-hybrid screen in yeast designates Hrs as a candidate partner for the DUF582 ...... 108! c.! Validation of the interaction between the DUF582 proteins and Hrs ...... 109! d.! CT619 interacts with Tsg101 ...... 112! 2.! Hrs and Tsg101 are not required for chlamydial growth in vitro ...... 114! a.! Hrs and Tsg101 levels decrease during infection ...... 114! b.! Perturbation of the ESCRT system does not affect chlamydial development ...... 117! c.! Uptake of LC3 positive compartments in the inclusion is Tsg101 dependent ...... 119! DISCUSSION ...... 120! BIBLIOGRAPHY ...... 128! ANNEXE ...... 148!

3 TABLE OF FIGURES Figure 1: Map of the 53 known countries with endemic trachoma (WHO)...... 19! Figure 2: Phylogeny of the Chlamydia genus ...... 23! Figure 3: Developmental cycle of C. trachomatis...... 24! Figure 4: Exit of Chlamydia by two different ways, (A) lysis or (B) extrusion...... 33! Figure 5: The assembly of the T3SS of Chlamydia...... 35! Figure 6: Chlamydia entering the cell ...... 39! Figure 7: Schematic of the different pathway that Chlamydia uses to get host lipids and nutrients...... 42! Figure 8: Overview of different Rab proteins and their specific localization in membranes in close relationship with phosphoinositides...... 44! Figure 9: Table of the ESCRT subunits and associated proteins...... 49! Figure 10: The different implication of ESCRT in the cell...... 50! Figure 11: Shematic organization of the human ESCRT-0...... 51! Figure 12: Scheme resuming the ILV formation via the ESCRT pathway ...... 52! Figure 13: Scheme of the abscission process mediated by ESCRT-I and –III proteins...... 53! Figure 14: Scheme of the HIV budding...... 54! Figure 15: Scheme of microvesicles (MVs) and exosome release...... 55! Figure 16: Schematic representation of Hrs domain...... 56! Figure 17: ESCRT along evolution...... 60! Table 1: Sequences of the different siRNA used in this study ...... 62! Table 2: Primers used to amplify each DUF582 recombinant protein used in this study...... 63! Table 3: List of primers used to amplify each constructs used in the directed yeast assay...... 65! Figure 18: CT619 and CT712 are made late in the Chlamydia developmental cycle...... 108! Figure 19: The Hrs protein interacts with the DUF582 domains of CT619 and CT621...... 109! Figure 20: Interations between the members of the DUF582 family and Hrs...... 110! Figure 21: The ESCRT-I protein Tsg101 interacts with CT619...... 113! Figure 22: Localization of Hrs during the infectious cycle...... 114! Figure 23: Chlamydia infection modulates Hrs and Tsg101 levels in a CPAF independent manner...... 116! Figure 24: Modulation the ESCRT system does not affect Chlamydia infection...... 118! Figure 25: LC3 positive structures are imported in the inclusion lumen in a Tsg101-dependent manner...... 119

4 ABBREVIATIONS

DUF: Domain of Unknown Function

EB: Elementary body

EE: Early endosome

RB: Reticulate body

GFP: Green fluorescence protein

Hpi: Hours post-infection

Hrs: hepatocyte growth factor-regulated tyrosine substrate

LGV: Lymphogranuloma venereum

L2: Chlamydia trachomatis LGV 2

MOI: Multiplicity of infection

MTOC: Microtubule-organizing center

MVB: Multivesicular bodies

PDI: Protein disulfide isomerase

PI3P: Phosphatidylinositol-3-phosphate

PIP3: Phosphatidylinositol-3,4,5 triphosphate

Tsg101: Tumor susceptibility gene 101

WHO: World Heatlth Organization

Y2H: Yeast-two-hybrid

5 ABSTRACT Chlamydia trachomatis is an obligate intracellular human pathogen. This Gram-negative bacterium is the first infectious cause of blindness and the most common cause of sexually transmitted diseases of bacterial origin. Like all chlamydiae, the developmental cycle of C. trachomatis takes place in a membrane bound compartment called the inclusion. The elementary body (EB) adheres and enters the host-cell. Once in the inclusion, EBs differentiate into reticulate bodies (RBs) that multiply several times until they differentiate back to EBs. Bacteria exit the host cell through two pathways: cell burst or inclusion extrusion. Using a strain of C. trachomatis serovar LGV expressing a fluorescent protein we developed novel microscopy and flow cytometry based methods to quantify several steps of this developmental cycle. These methods will facilitate future studies aimed at testing anti-bacterial compounds or various culture conditions. Chlamydiae depend on their host to complete their developmental cycle. They interfere with many cellular processes, in particular via the secretion of bacterial proteins through a type 3 secretion (T3S) system. We identified a family of proteins that possess T3S signals. They share a domain of about 400 amino acids, designated as DUF582, which is only found in pathogenic chlamydiae. C. trachomatis possesses five DUF582 proteins. We obtained specific antibodies against these five proteins and showed that they are expressed at the mid and late phases of infection by C. trachomatis serovar L2. We provide direct evidence for the secretion in the host cell of three out of the five proteins. A yeast-two-hybrid (Y2H) screen revealed that the protein Hrs is a common interactor for the DUF582. The interaction was confirmed by co-immunoprecipitation experiments in co-transfected cells. In addition, Y2H and co- immunoprecipitation experiments demonstrated that the N-terminal domain of one of the DUF582 proteins, CT619, also interacted with the protein Tsg101. Hrs and Tsg101 are both implicated in a well conserved machinery of the eukaryotic cell called the ESCRT machinery, which is involved in several cellular processes requiring membrane fission. We demonstrated that infection leads to a decrease in Hrs and Tsg101 levels in the late phase of the developmental cycle. Using RNA interference we showed that Hrs and Tsg101 are dispensable for bacterial entry and growth. This last result indicates that DUF582 proteins actually prevent Hrs and/or Tsg101 driven processes. Alternatively, the bacteria might highjack the ESCRT machinery but redundant mechanisms would explain the absence of phenotype on bacterial development observed in the silencing experiments. In light of our results and the known functions or Hrs and Tsg101 we favor three hypotheses as to the possible role of the DUF582 proteins in infection. They might participate to the escape of the nascent inclusion from the default pathway towards degradation, permit the uptake of host material at the inclusion membrane through an ESCRT driven mechanisms and/or allow bacterial exit from the cells via an ESCRT-driven extrusion mechanism.

6 RESUME

I. INTRODUCTION

Chlamydia trachomatis est une bactérie intracellulaire obligatoire apparentée aux bactéries Gram négatives. Ce pathogène de l’Homme est la première cause infectieuse de cécité ainsi que de maladies sexuellement transmissible d’origine bactérienne. Elle infecte principalement des cellules épithéliales, de la conjonctive de l’œil ou des muqueuses génitales. La bactérie a un cycle de développement qui fait intervenir deux formes. Sous sa forme « corps élémentaire » (CE), la bactérie adhère à la cellule hôte puis entre, formant une vacuole appelée inclusion. Le CE se différencie alors en « corps réticulé » (CR), forme métaboliquement active de la bactérie permettant sa réplication au sein de l’inclusion. Le retour à la forme CE prend place peu avant la fin du cycle infectieux. De l’entrée à la fin de son développement, la bactérie interagit constamment avec l’hôte, notamment en utilisant une « seringue moléculaire », le système de sécrétion de type III (SST3). Les effecteurs du SST3 participent à la manipulation de la cellule hôte, contribuant directement au développement de la bactérie. L’étude de la bactérie Chlamydia a beaucoup évolué ces dernières années. La manipulation génétique, encore impossible deux ans auparavant, est devenue accessible même si elle reste à l’heure actuelle encore laborieuse. Ceci a conduit à l’obtention de Chlamydia exprimant des protéines fluorescentes qui seront utilisées au cours de cette étude pour développer de nouvelles techniques d’étude de la bactérie. En utilisant le SST3 de Shigella, nous avons identifié plusieurs protéines de Chlamydia candidates à être des substrats du SST3. L’une d’entre elles possédait un motif de fonction inconnue (le DUF582) que l’on a retrouvé dans quatre autres protéines de Chlamydia trachomatis. Ce motif étant présent uniquement chez les Chlamydia et plus particulièrement dans les espèces pathogènes, nous avons décidé d’étudier la(es) fonction(s) de cette famille de protéines.

7 II. RESULTATS

1. Article 1 : Identification of a family of effectors secreted by the type III secretion system that are conserved in pathogenic chlamydiae 1 1 Sandra Muschiol, Gaelle Boncompain , François Vromman , Pierre Dehoux, Staffan Normark,

Birgitta Henriques-Normark, and Agathe Subtil. Inf. & Immun. (2011) 79 571

La détection d’effecteur du SST3 de Chlamydia, réalisée avec le SST3 hétérologue de Shigella flexneri, a démontré que la protéine CT712 de C. trachomatis ainsi que ses homologues chez C. caviae et C. pneumoniae possèdent un signal de ST3 à leur extrémité amino-terminale (Subtil et al, 2005). Dans cette étude, nous nous sommes concentrés sur la famille de protéine qui inclut CT712. Les protéines de cette famille partagent un domaine commun de 400 acides-aminés, le DUF582, qui est présent dans toutes les Chlamydia pathogènes et dans aucun autre organisme. Chaque espèce comporte 4 à 5 protéines (5 pour C. trachomatis), composée du domaine conservé DUF582 ainsi que d’un domaine N-terminal variable. En utilisant le système de sécrétion hétérologue de Shigella flexneri, nous avons montré que tous les membres de cette famille de protéine possèdent un signal de ST3 à leur extrémité N-terminale. Le DUF582 est prédit pour être majoritairement composé d’hélices alpha et possède un domaine « coiled-coil » en son centre. Les analyses bioinformatiques ont montré que la conservation du domaine est supérieure entre les orthologues qu’entre les paralogues pour les quatre groupes définit par le domaine N-terminal. Chez C. trachomatis, le pourcentage d’identité entre deux protéines DUF582 ne dépasse pas 39%. De plus, les domaines N- terminaux ne montrent aucune similarité. CT712 est uniquement composé du domaine DUF582. Enfin, aucune similarité n’a été trouvée entre les protéines DUF582 et toutes les protéines référencées dans les bases de données. Nous avons obtenu des anticorps de lapin dirigés contre trois des cinq protéines DUF582 de C. trachomatis (CT620, CT621 et CT711). Nous avons observé que ces protéines sont exprimées à partir du milieu du cycle infectieux et qu’elles sont présentes dans les formes CE. Egalement, nous avons démontré que CT620 et CT621 sont sécrétées dans le lumen de l’inclusion ainsi que dans le cytoplasme de la cellule hôte après trente heures d’infection.

1 Les auteurs ont contribué également à ce travail.

8 2. Article 2: Monitoring of Chlamydia trachomatis developmental cycle using GFP-expressing bacteria, microscopy and flow cytometry

François Vromman, Marc Laverrière, Stéphanie Perrinet, Alexandre Dufour and Agathe

Subtil. Plos One (2014), in press.

Une méthode d’obtention de Chlamydia fluorescentes a été publiée fin 2011. Nous avons choisi d’utiliser ce nouvel outil pour suivre l’infection plus facilement et plus quantitativement. Ainsi, nous avons développé des méthodes, basées sur la microscopie et de cytométrie en flux, permettant de suivre tout le cycle de développement de la bactérie. Nous avons choisi de travailler avec la souche L2 transformée avec un plasmide exprimant la GFP sous contrôle du promoteur du gène incD obtenu de I. Derré (Agaisse & Derré, 2013). Cette souche montre, par qPCR, une courbe de croissance similaire à la souche parentale. Nous montrons également que la GFP est détectée dans les CE. L’attachement et la multiplication des bactéries ont été mesurés en utilisant la cytométrie en flux. Cette méthode fournit des données statistiquement valables de façon simple, rapide et peu coûteuse. Nous avons montré que la quantité de GFP présente dans les CE était suffisante pour détecter leur attachement à partir d’une MOI de un. L’étape d’entrée a été quantifiée par microscopie couplée à une analyse d’image automatique. Pour la première fois pour C. trachomatis, nous avons quantifié précisément la cinétique d’entrée des bactéries. En utilisant le logiciel ICY développé à l’Institut Pasteur, nous avons mesuré que 50 % des bactéries sont internalisées après 10 min d’incubation à 37°C. La cytométrie en flux est aussi une méthode simple et statistiquement valable pour quantifier le taux d’infection ainsi que la charge bactérienne, plus particulièrement pour les populations cellulaires non-homogènes. Enfin, nous avons utilisé la cytométrie en flux pour compter directement les CE en utilisant un cytomètre qui détecte les particules de petite taille jusqu’à 0,2 µm. Cet important résultat offre la possibilité de compter directement les préparations bactériennes ou les lysats cellulaires frais.

9 3. La protéine DUF582 CT619 cible les protéines Hrs et Tsg101 au cours de l’infection à Chlamydia trachomatis François Vromman, Stéphanie Perrinet et Agathe Subtil. Manuscrit en préparation.

a. CT619 interagit avec Hrs et Tsg101

Pour obtenir des anticorps reconnaissant CT619 et CT712, deux protéines DUF582, nous avons produit les protéines recombinantes GST-!81CT619!625 et GST-CT712 exprimée chez E. coli et nous les avons purifiées pour l’immunisation de lapins. Les anticorps obtenus se sont révélés spécifiques des protéines respectives pour l’analyse par immunofluorescence de cellules transfectées avec la protéine d’intérêt fusionnée avec une étiquette GFP. Toutefois, ces anticorps utilisés en immunofluorescence ne permettent pas de visualiser de structure particulière dans les cellules infectées, à part les bactéries elles-mêmes (dans le cas des anticorps contre CT619, même les bactéries ne sont pas détectées par le sérum). Une recherche de protéines de l’hôte susceptibles d’interagir avec les protéines DUF582 a été effectuée dans un modèle de levure en utilisant la technique du double-hybride. En utilisant les domaines DUF582 de CT619 et de CT621 comme appât, nous avons obtenu une proie commune sur quelques millions de candidats testés, la protéine Hrs. Cette protéine fait partie du complexe ESCRT-0, lui-même appartenant au système ESCRT (Endosomal Sorting Complex Required for Transport) impliqué dans le tri des protéines au sein de la cellule eucaryote. En se fixant aux protéines ubiquitinées, Hrs initie le tri des protéines vers les lysosomes. Plus particulièrement, le système ESCRT est responsable de la formation des vésicules intra-luminales présentes dans les MVB (MultiVesicular Bodies). L’un des marqueurs les plus étudiés de cette voie est le récepteur à l’EGF. En co-transfectant une construction myc-Hrs et les protéines DUF582 fusionnées avec une étiquette GFP, nous avons montré par immunofluorescence que ces protéines co- localisaient. Des expériences d’immunoprécipitation de myc-Hrs ou de Hrs endogène ont permis de co-précipiter les protéines DUF582 transfectées. Nous avons poursuivi l’approche par la technique du double-hybride pour préciser le domaine d’interaction des protéines à DUF582 avec Hrs. L’interaction semble particulièrement portée par le domaine coiled-coil présent dans le domaine DUF582. CT619 montre une interaction beaucoup plus forte avec Hrs que CT621

10 Le crible double hybride avait également mis en évidence une interaction de la partie N-terminale de la protéine CT619 avec la protéine Tsg101. Cette dernière fait partie du complexe ESCRT-I, interagissant directement avec ESCRT-0 au niveau de Hrs. Nous avons confirmé cette interaction par des expériences d’immunoprécipitation et précisé que l’interaction se fait spécifiquement entre la partie N-terminale de CT619 et le domaine « coiled-coil » de Tsg101. L’interaction entre ces deux protéines a également été observée au cours de l’infection. Pour résumer, la protéine CT619 semble interagir avec deux protéines importantes du système ESCRT : Hrs et Tsg101. Il est important de souligner ici que le knock-down de l’une de ces deux protéines bloque la fonctionnalité du système ESCRT.

b. Hrs et Tsg101 ne sont pas requis pour la croissance de Chlamydia trachomatis in vitro

Ayant mis en évidence une interaction entre Hrs et le domaine DUF582, nous avons étudié la localisation de Hrs au cours de l’infection par immunofluorescence. Hrs est souvent située proche de l’inclusion mais n’est pas spécialement enrichie à la membrane de l’inclusion proprement dite. Nous avons également analysé le niveau de la protéine Hrs, mesuré par western blot, au cours de l’infection par Chlamydia trachomatis L2. Cette mesure a révélé que la quantité de Hrs diminuait à partir de vingt-quatre heures d’infection jusqu’à pratiquement disparaître en fin de cycle. L’ajoût de chloramphénicol après seize heures d’infection bloque la diminution de Hrs, montrant le rôle de la bactérie dans ce processus. Cette diminution n’est pas due à un clivage opéré par CPAF, une protéase de Chlamydia, car l’expression de cette dernière dans des lignées stable n’entraîne aucune diminution des niveaux de Hrs. De plus, nous avons démontré que la protéine Tsg101 décroît également au cours de l’infection. Afin de comprendre l’avantage pour la bactérie de cibler Hrs et Tsg101, nous avons utilisé la technique de l’ARN interférence pour inhiber l’expression de ces protéines. Nous avons démontré, en utilisant les techniques développées dans l’article 2, que la déplétion de Hrs ou de Tsg101 n’affecte pas l’entrée de la bactérie, leur développement, ou encore leur pouvoir infectieux. Pour confirmer ce résultat, nous avons étudié la protéine VPS4, protéine indispensable au désassemblage du système ESCRT. Le dominant négatif de cette protéine est connu pour abolir également les fonctions du système ESCRT. En transfectant la version sauvage ou le

11 dominant négatif de cette protéine suivi de l’infection des cellules HeLa, nous avons démontré que le blocage du système ESCRT n’affecte pas le cycle infectieux de Chlamydia.

III. DISCUSSION

1. Les protéines DUF582 sont des effecteurs du SST3 de C. trachomatis

Les protéines à domaine DU582 sont retrouvées uniquement chez les chlamydiae, et parmi elles, seulement chez les chlamydiae pathogènes et non chez les souches environnementales. Par espèce, on compte 4 à 5 protéines qui partagent le domaine de fonction inconnue DUF582 et dont la conservation est plus forte entre les orthologues qu’entre les paralogues. La plupart des protéines DUF582 possèdent également un domaine N-terminal variable et spécifique des chlamydiae. CT712 et les autres membres de ce groupe ne sont composés que du domaine DUF582. En utilisant des anticorps spécifiques contre des protéines recombinantes, nous avons démontré que chaque membre de la famille de protéine de C. trachomatis est exprimé à partir du milieu du cycle infectieux et qu’elles sont présentent dans les CE purifiés. Ces données sont confirmées par les données de transcriptomique publiées. Nous avons montré que ces protéines possèdent un signal de ST3 sur leur extrémité N- terminale en utilisant le système hétérologue de Shigella flexneri. Ceci est confirmé par la visualisation, par microscopie, de CT620 et CT621 dans le cytoplasme de la cellule hôte.

2. Le domaine DUF582 interagit avec Hrs

Le crible double hybride réalisé sur les deux domaines DUF582 de CT619 et CT621ont révélé Hrs comme étant le seul candidat commun d’interaction sur tout le protéome humain testé. Nous avons confirmé l’interaction par des expériences d’immunoprécipitation ainsi que par des expériences de double-hybride. Ces dernières ont montré que l’interaction avait lieu préférentiellement avec le domaine DUF582 mais que les domaines N-terminaux pouvaient également interagir bien que plus faiblement.

12 3. CT619 interagit avec Tsg101

Le crible double hybride a révélé également la protéine Tsg101 comme interagissant avec le domaine N-terminal de CT619 et non avec CT621. Nous avons démontré que l’interaction se fait au niveau du domaine « coiled-coil » de Tsg101 en utilisant la technique d’immunoprécipitation. Cette même technique nous a permis de montrer l’interaction à des temps tardifs de l’infection. La propriété de CT619 d’interagir avec Hrs et Tsg101 est intéressante car ces deux protéines interagissent ensemble au sein du système ESCRT. Ces données indiquent que CT619 pourrait potentiellement interagir avec Hrs et Tsg101 simultanément.

4. Hrs et Tsg101 ne sont pas nécessaire au développement de C. trachomatis

Nous avons montré que les niveaux de Hrs et de Tsg101 diminuent à partir de 20 heures d’infection. Cette diminution n’est pas due à l’activité protéolytique de la protéase de Chlamydia CPAF. Cette diminution est corrélée à la cinétique d’expression des protéines DUF582 au cours de l’infection par C. trachomatis. Toutefois, l’expression par transfection des protéines DUF582 n’affecte pas le niveau de Hrs et Tsg101. Aussi, il n’est pas clair que la disparition de Hrs et Tsg101 observée durant l’infection résulte directement de leur interaction avec les protéines DUF582. Il faut également noté que les expériences mesurant la quantité de Hrs ou Tsg101 ont été effectuées à une MOI de 5 pour atteindre 100% d’infection. Aussi, avec une MOI de 1, condition normale d’infection, la disparition de Hrs ou Tsg101 intervient vraisemblablement à la toute fin du cycle infectieux. Il n’y a donc pas évident de comprendre le bénéfice de cette disparition, s’il y en a un, pour la bactérie. L’absence d’impact du traitement par ARN interférence contre Hrs ou Tsg101 sur le cycle infectieux peut s’expliquer par deux hypothèses : (i) les bactéries cherchent elles-mêmes à inactiver Hrs et/ou Tsg101, et dans ce cas aucun phénotype n’est à attendre de l’interférence qui ne fait qu’amplifier un phénomène naturel. (ii) Les processus dépendant de Hrs et/ou de Tsg101 ont lieu pendant le développement de la bactérie, mais des mécanismes redondants fonctionnent également et empêchent l’observation de phénotype. De plus, il est possible que

13 ces protéines Hrs et /ou Tsg101 soient impliquées dans la dissémination de la bactérie, phénomène qui n’a pas encore été étudié.

5. Trois scénarios possibles pour les résultats obtenus a. Scénario 1 : Les protéines DUF582 interagissent avec Hrs pour échapper à la dégradation lysosomale.

Il a été décrit que Chlamydia entre via différents récepteurs tel que le PDGFR ou l’EGFR. Ces deux récepteurs sont tous les deux dégradés via la voie endo-lysosomale, qui est initiée par la reconnaissance des récepteurs activés poly-ubiquitinés par Hrs. Aussi, l’absence de marqueurs des endosomes précoces tels que EEA1 même 5 min après le début de l’entrée suggère l’échappement rapide de la bactérie de cette voie lysosomale. Il a été montré que la bactérie a besoin de synthèse de novo pour échapper à cette voie de dégradation même si il faut tout de même plusieurs heures à la bactérie pour se retrouver dans les lysosomes. Comme nous détectons les protéines DUF582 dans les CEs purifiés, il est possible que la sécrétion de CT619 et sa capacité à interagir avec Hrs et Tsg101 soit suffisante pour bloquer le système ESCRT et donc éviter ainsi d’être dirigé vers les lysosomes.

b. Scénario 2 : Les protéines DUF582 interagissent avec Hrs pour acquérir du matériel de l’hôte

Bien que plusieurs constituants de la cellule hôte ait été observés dans le lumen de l’inclusion, les mécanismes impliqués dans leur import restent largement inconnus. Nous n’avons pas réussi à observer l’import de CD63, LBPA ou encore de gouttelettes lipidiques au sein de l’inclusion alors que leur présence dans l’inclusion a été décrite. Nos résultats préliminaires montrent l’import de LC3 au sein de l’inclusion, et ce de façon dépendante de Tsg101. Ce résultat suggère que les protéines DUF582 pourraient recruter le système ESCRT à la membrane de l’inclusion pour favoriser l’import de différents constituants de l’hôte, de manière analogue à la formation des vésicules internes dans les MVBs.

14 c. Scénario 3 : CT619 interagit avec Tsg101 pour la sortie de la bactérie

Il est possible que la bactérie ait besoin des protéines DUF582 pour sa sortie. Cette éventualité n’a pas été évaluée dans cette étude. L’un des moyens de sortie de la bactérie se fait par extrusion. Ce processus a été décrit pour impliquer des protéines actrices de la cytokinèse. Ainsi, leur déplétion entraîne une diminution du nombre d’extrusions sans affecter le reste du cycle de la bactérie. Ce processus de sortie de la bactérie est très semblable au mécanisme d’abscission au cours de la cytokinèse, dans lequel Tsg101 est déterminant. L’interaction de CT619 avec Tsg101 pourrait recruter Tsg101 afin d’effectuer l’extrusion.

IV. CONCLUSION

Nos résultats sont en accord avec au moins trois fonctions possibles des protéines DUF582 pendant l’infection. Ces fonctions ne s’excluent pas mutuellement, au contraire il est probable que les différentes protéines de la famille soient impliquées dans des mécanismes distincts. Les trois hypothèses décrites ci-dessus sont actuellement mis à l’épreuve, et plus particulièrement en utilisant les récents outils développés pour exprimer des effecteurs de Chlamydia fusionnés avec une étiquette, ainsi que l’obtention de mutants de délétion. Cette dernière stratégie ne fonctionnera que sur les gènes non-essentiels, et nous ne savons pas si ce sera le cas pour les protéines que nous étudions. Pour conclure, nous avons montré que les protéines DUF582 sont des effecteurs de type III des Chlamydia pathogènes, que leur domaine commun cible Hrs, et que probablement la bactérie manipule un processus ESCRT- dépendant. En considérant l’évolution du parasitisme intracellulaire des Chlamydia, il n’est pas surprenant qu’elles aient acquis des outils sophistiqués pour interagir avec cette machinerie ancestrale ESCRT, impliquée dans de nombreuse fonctions essentielles de leur hôte eucaryote.

INTRODUCTION

17 I. Chlamydia trachomatis, a human pathogen

1. Pathology & epidemiology: which diseases and where?

Chlamydia is an obligate intracellular bacterium, which cause a number of diseases in humans and animals. About ten different species have been described so far, which differ by their host range and the clinical manifestations of infection. Four species have been isolated in humans: C. abortus, C. psittaci, C. pneumoniae and C. trachomatis. The first two species are animal pathogens (mostly found in cattle and birds, respectively) and transmission to humans is only observed in populations exposed to the animal reservoir. In contrast, while C. pneumoniae is also commonly found in many animal hosts (including marsupials, amphibians and reptiles), it is also widespread in the human population, with efficient human-to-human transmission by aerosols. It causes acute respiratory diseases and has been associated to a number of chronic infections. This thesis will mainly focus on Chlamydia trachomatis, for which humans are the only host. It is responsible for two different kinds of infections: ocular infections, also called trachoma, and infections of the genital tract (Collier, Balows, & Sussman, 1998).

a. Trachoma

Trachoma is a neglected tropical disease that remains the leading cause of blindness by infection in the world. It is responsible for visual impairment of about 2.2 million people, including 1.2 million with an irreversible blindness (Pascolini & Mariotti, 2012). It mainly affects children, but the blindness cases are seen in adults between thirty to forty years old. Human-to-human transmission is frequent especially between children. Flies are also able to spread trachoma (Emerson et al., 2004). The large majority of clinical isolates of ocular C. trachomatis infection belong to the serovars A, B, Ba and C (Pedersen, Herrmann, & MÃ ller, 2009). The bacteria develop in the conjunctiva of the eye, causing conjunctival inflammation. Even in the presence of inflammation, the Chlamydia infection remains mostly asymptomatic. Without treatment, chronic inflammation can lead to inward turning of the eyelid with subsequent scarring of the conjunctiva (triachiasis). Abrasion of the cornea ultimately leads to corneal opacification, low vision and finally blindness. The World Health Organization (WHO) estimated in 2013 that trachoma is endemic in at least 53 countries (Figure 1).

18 Endemic trachoma is prevalent in large parts of Africa, Asia, Australia and the Middle East. Conditions favorable to its transmission are the lack of hygiene, difficulties for water access, over-population, and poverty in general. The WHO is currently running a program called Global Alliance for the Elimination of Trachoma by 2020 (GET 2020), which aims to eliminate blinding trachoma by 2020 (WHO, 2013).

Figure 1: Map of the 53 known countries with endemic trachoma (WHO).

b. Urogenital infections

C. trachomatis also infects the genital tracts of men and women. It is the primary cause of sexually transmitted infection of bacterial origin, with over 105.7 million cases worldwide in 2008 (World Health Organization, 2012). In the USA, the Centers for Disease Control (CDC) have estimated that the total number of cases in 2012 was approximately 2.86 million (http://www.cdc.gov). The main serotypes recovered from genital infections are serotypes D to K (Pedersen et al., 2009). Infection can remain limited to lower parts of the genital tract or propagate to upper parts. The majority of urogenital infections by Chlamydia is asymptomatic or clinically sub-acute, however acute presentations of cervicitis, endometritis and salpingitis for women, and uretritis or epididymitis for men, are not uncommon. If untreated, infection can lead to ectopic pregnancy and infertility (Borriello, Murray, & Funke, 2005). Also, Chlamydia has more recently been linked to cervical cancer

19 (Paavonen, 2012). Infections are acquired during sexual intercourse, but it can also be passed from a mother to her baby during delivery often resulting in a conjunctivitis.

c. Lymphogranuloma venereum

Lymphogranuloma venereum (LGV) is caused by the L1, L2 and L3 serovars of C. trachomatis. In contrast to other serovars, LGV is an invasive biovar, spreading to the lymphatic and subepithelial tissues. The infection reaches the lymphatic channels after entry through skin lesions. It starts with a lymphangitis before migrating to regional lymph nodes causing lymphadenitis and sometimes to suppurative necrosis. Since 2003, LGV proctitis cases have emerged in Europe, North America and Australia in a population of men who have sex with men. A majority of the patients were HIV-infected. It is important to note here that the ulcerative nature of LGV could facilitate both acquisition and transmission of HIV and other STDs "#$%&'()*+,$-./!.&!$/01!23435.

d. Diagnosis and treatments

Diagnoses are adapted to the different diseases caused by Chlamydia trachomatis. Concerning trachoma, an initial physical examination of the eye with a magnifying glass by a doctor or qualified paramedic allows the grading of the disease from one to five "678/.9:%;1! <$=;:(1! >:(.;1! ?.;&1! @! 6$8/:%1! 4ABC5. Confirmation that Chlamydia infection is at the origin of the disease can be obtained by nucleic acid amplification tests (NAATs) - although C. trachomatis detection is sometimes impossible (Tang & Bavoil, 2012). Treatments range from surgery (to reorient the eyelid) to antibiotic treatment. Tetracycline or azithromycin administrated locally for a minimum of six weeks is recommended (Hu et al., 2010). Efforts to improve the hygienic conditions, for instance by limiting the proximity between cattle/fly companions and households, are currently the most efficient strategy to decrease the incidence of conjunctival Chlamydia infection. Detection of Chlamydia in urogenital tracts is easily performed via NAATs on urine samples or patient-collected vaginal swabs. The typical treatment for serovars D to K is a single dose of azithromycin or seven days of doxycycline. Observational studies showed that azithromycin efficiency is about 92% (Batteiger et al., 2010). Real-time multiplex polymerase chain reaction (PCR) allows the discrimination between LGV and non-LGV (C.-Y. Chen et al., 2006). Treatment for LGV is doxycycline

20 twice-daily for three to four weeks. Differences in length and strength of the treatment in genital Chlamydia infection show how the precise identification of the serotypes is important. Up to now, the best strategy to limit the spreading of Chlamydia genital infections remains prevention. Use of sexual protection and detection of potential STDs are necessary to prevent and cure the mainly asymptomatic Chlamydia infections. Developing a vaccine remains the most appealing strategy to fight against Chlamydia infections. Difficulties due to the biology of the bacterium make the search for a protective vaccine particularly challenging. Several strategies focus on finding the best chlamydial antigen to generate a strong immune response. Outer-membrane proteins and well-studied secreted proteins are the main candidates. It is generally accepted that the best strategy is to induce both local neutralizing antibodies to prevent infection by the extracellular microbe and a cell-mediated immune response to target the intracellular infection (Hafner, Wilson, & Timms, 2014).

2. A bit of history: once upon a time...

Chlamydia infection is one of the oldest documented human infectious diseases. It may have originated in Central Asia based on the observations that the rate of trachoma was higher in people with Central Asian ancestry (Taboriski, 1952). In 1990, an Australian archeologist named Stephen Webb published what could be the first evidence of trachoma on an Australian aborigine dated to around 8000 years before Christ (B.C.) (Webb, 1990). Some Chinese reports from 2600 B.C. described therapies against trachoma. Other descriptions of the disease itself and related treatments were found in different civilizations: The Egyptians in their Ebers papyrus (1500 B.C.) as well as in Ancient Greece and Rome (1200 B.C. to 400 A.D.). It is in this period that the term “trachoma” (from a Greek word meaning “roughness”) was first used by the Greek doctor Dioscoride. Spreading of trachoma over the years is largely linked to wars. Crusades in the medieval age and the Napoleonic campaign in Egypt contributed to the spread of trachoma throughout Europe (Schlosser, 2013). For a long time trachoma was an infection with no causal agent identified. It was the pioneering work of Halberstaeter and Von Prowasek in 1907, that first showed through Giemsa staining evidence of microcolonies (inclusions) observed in conjunctival epithelial cells. Visualization of transmission to baboons confirmed that it was a pathogenic agent (Freney, Renaud, Hansen, & Bollet, 1994). Around that time, it was also shown that newborns

21 could develop trachoma because of a genital infection present in parents (Borriello et al., 2005). In 1933, the epidemiologist Bedson studied the outbreak of atypical pneumonia in Hamburg. He showed that it originated from parrots and that the infection could spread to other animals. He next qualified it as the psittacosis “virus” and during the following two decades the infectious agent was thought to be a virus (Bedson, 1936). Descriptions of cases and different serovars accumulated during the 20th century. In vitro cultures of Chlamydia were initially performed in eggs (Tang, Chang, & Huang, 1957) but, rapidly, growth of the “trachoma agent” in cell culture was introduced (Gordon, Quan, & Trimmer, 1960). Observations of the “trachoma virus” by electron microscopy suggested a bacteria-like behavior (Litwin, 1962) and lead, with many other observations, to the confirmation of Chlamydia being a bacterium in 1966 (Moulder, 1966). Over the years, the understanding of Chlamydia biology has increased slowly. The application of novel techniques from the first sequencing of its genome (Stephens et al., 1998) to the recent ability to genetically manipulate it (Y. Wang et al., 2011) have opened many novel avenues of research.

3. Chlamydia, a bacterium apart from others

After the difficulty to recognize Chlamydia as a bacterium came that of its classification among bacteria. Evolutionary studies seem to date the first Chlamydiales around 700 million years ago (mya). Concerning C. trachomatis, the species appeared around 6 mya at the same time as the “human/chimpanzee” divergence (Nunes & Gomes, 2014; Steiper & Young, 2006). This obligate intracellular pathogen has a genome of about one million base pairs supplemented or not by a seven to eight thousand base pairs plasmid. Chlamydia has a rather small genome (with around 1000 genes) compared to other human bacterial pathogens such as E. coli, which has around 4000 genes (Subtil, Collingro, & Horn, 2014). In terms of phylogeny, C. trachomatis belongs to the Chlamydiaceae family and has been recently reunited into the single genus “Chlamydia” (Stephens, Myers, Eppinger, & Bavoil, 2009). Nowadays, the classification of Chlamydia is based on whole genome sequences. Inside the C. trachomatis species, the phylogeny divergence between serovars mirrors their tissue tropism (Figure 2) (Nunes & Gomes, 2014).

Figure 2: Phylogeny of the Chlamydia genus (Nunes & Gomes, 2014). A) Phylogenic link between Chlamydia species and their respective main host and capacity to infect human. B) Phylogeny of C. trachomatis grouped by tissue tropism.

23 4. The Chlamydia special: a biphasic developmental cycle

Chlamydia is a Gram-negative, obligate intracellular bacterium with a very peculiar and characteristic development cycle alternating between two distinct forms of the bacterium (Abdelrahman & Belland, 2005). The first description of this cycle was done in 1932 by Bedson and Bland (Bedson & Bland, 1932). After entering the host cell, the elementary body (EB) differentiates into a reticulate body (RB), which multiplies within a growing parasitophorous vacuole termed the inclusion. In the mid-late phase of the cycle, the RBs differentiate into EBs before the exit of the bacteria after 48 to 72 hours of infection (Figure 3).

Figure 3: Developmental cycle of C. trachomatis. Adapted from Abdelrhaman & Belland. (Abdelrahman & Belland, 2005). EBs attach via reversible and irreversible interactions and induce their entry secreting effectors like TARP. EBs are subject to a primary differentiation into RBs, which start to divide and multiply. Along the cycle, the bacteria interact with the host via effector secretion. Bacteria differentiate a second time back to the condense EB state. Under stress condition generated by different inducers, Chlamydia grow abnormaly and turn into a persistent state generating aberrant RBs. At the end of the cycle, bacteria exit in two different ways: host-cell lysis or extrusion. TTSS: type three secretion system, IF: intermediate form.

24 a. Characteristics of EBs and RBs

EBs are the infectious forms of Chlamydia. They are roughly spherical, with a diameter of about 0.3 "m, and electron dense (Matsumoto, 1981). For a long period, it was believed that EBs were metabolically inert and described as “spore-like” forms of Chlamydia. Two recent papers have demonstrated that EBs show some reduced metabolic activity (Omsland, Sixt, Horn, & Hackstadt, 2014; Sixt et al., 2013). On the other hand, the reticulate bodies (RBs) are larger than EBs with a diameter of 1 "m. The RBs are intracellular and not adapted to the environment outside the inclusion, in which they multiply by binary fission. The characteristic inner and outer membranes of Gram negative bacteria are seen by electron microscopy (Abdelrahman & Belland, 2005). For a long time, the absence of detectable peptidoglycan, together with the sensitivity of Chlamydia to penicillin, and the presence in its genome of all the genes required for peptidoglycan synthesis constituted the “chlamydial anomaly”. However, the presence of peptidoglycan has recently been demonstrated thanks to new technologies (Liechti et al., 2013; Pilhofer et al., 2013).

b. The very first step: the bacterial adhesion

The first step of chlamydial invasion is its adhesion to a host cell. This process occurs in two steps. Primarily, the bacterium interacts with the host in a reversible manner via specific heparan sulfate glycosaminoglycans (HS-GAG) produced by the bacteria (Zhang & Stephens, 1992). In this paper, authors have shown an inhibition of Chlamydia adhesion by incubating the bacteria with heparin-sulfate receptors, heparin, or heparin sulfate lyase prior to infection. The effect of heparin sulfate lyase treatment was reversed by addition of exogenous heparin sulfate indicating the need of those molecules for Chlamydia adhesion. Other chlamydial adhesins are described to play a role in the attachment of the bacterium such as the major outer membrane protein (MOMP), the proteins of the outer membrane complex (OMC) OmcA and OmcB, or even the polymorphic outer membrane proteins (Pmps) (Tang & Bavoil, 2012). Carabeo & Hackstadt in 2001 described a secondary binding step coming rather from the host cell, which is irreversible (Carabeo & Hackstadt, 2001). They observed that washing the cells with heparin after 30 min of incubation in a normal medium (without heparin) is not enough to abolish the bacterial adhesion. Thus, the authors also demonstrated that the

25 hallmark of this secondary adhesion is temperature-dependent. Also, the Stephens’ lab described that the protein disulfide isomerase (PDI) is implicated not only in the attachment of Chlamydia (Conant & Stephens, 2007) but also in the entry of the bacterium (Abromaitis & Stephens, 2009). It remains so far the only host eukaryotic protein identified to be necessary for different Chlamydia species adhesion. Finally, it has been shown recently that sulfation is required for Chlamydia attachment (Rosmarin et al., 2012). Authors highlighted the fact that separate knockout of three host genes implicated in sulfation is sufficient to impair the bacterium binding.

c. Knock-knock: The Chlamydia entry

Entry of bacterial pathogens can occur in two non-mutually exclusive mechanisms: through the activation of host receptors or via the translocation of bacterial activators of endocytosis (Pizarro-Cerdá & Cossart, 2006). Both processes occur in a Chlamydia entry mechanism that is yet not fully understood. The host cell plays an active role in the uptake of the bacterium. Entry may engage lipid microdomains in the plasma membrane. Indeed, markers of lipid rafts localizes with the bacterium even after five hours of infection by C. trachomatis (Jutras, Abrami, & Dautry- Varsat, 2003). The involvement of the actin cytoskeleton in the entry process is very well established. Inhibitors of actin polymerization such as cytochalasin D completely inhibit invasion (Boleti, Benmerah, Ojcius, Cerf-Bensussan, & Dautry-Varsat, 1999; Coombes & Mahony, 2002; Ward & Murray, 1984). Early work has shown that small GTPases of the Rho (Carabeo, Grieshaber, Hasenkrug, Dooley, & Hackstadt, 2004; Subtil, 2004) and Arf (Balañá et al., 2005) families were activated during entry. In addition, an active mechanism induced by Chlamydia is at work. The bacterium uses the type three-secretion system (T3SS) to translocate effectors triggering its active entry. The most described secreted protein at the entry site is CT456, the translocated actin- recruiting phosphoprotein (TARP) (Clifton et al., 2004). This protein has been shown to be responsible for actin nucleation and recruitment and to be part of different phosphorylation events at the entry site of Chlamydia. This will be discussed more in detail in the second chapter. Even if the mechanism of entry is not completely solved, it seems clear that Chlamydia enters via a receptor, possibly clathrin for C. trachomatis (Hybiske & Stephens, 2007a). Engel’s lab showed that PDGFR! receptor and more importantly the FGF2 receptor

26 are two key receptors needed for the binding and the entry of the bacterium (Elwell, Ceesay, Kim, Kalman, & Engel, 2008; Kim, Jiang, Elwell, & Engel, 2011). With all the knowledge available on those first steps, it is still not known how the bacterium escapes from early endo- lysosomal trafficking. Known makers of the early endosomes like EEA1 or Rab5 are not colocalized with the EBs or the inclusion (Rzomp, Scholtes, Briggs, Whittaker, & Scidmore, 2003; Scidmore, Fischer, & Hackstadt, 2003). Nevertheless, if EEA1 staining has been investigated before one hour of infection, it is not the case for Rab5 for which the localization has not been studied before 18 hours post infection (hpi). A rapid response of the bacterium might be the cause of this escape since a very short and transient staining of PIP3 is observed in the first minutes of entry by Lane and colleagues (Lane, Mutchler, Khodor, Grieshaber, & Carabeo, 2008). The precise mechanism still remains to be investigated. Finally, it is very likely that the receptors and signaling pathways implicated show differences between species. For instance, the EGF receptor might serve C. pneumoniae but not C. trachomatis entry (Mölleken, Becker, & Hegemann, 2013). TARP itself shows high dissimilarities between species.

d. Home sweet home: Setting up the inclusion

Immediately following entry, EBs remain in a tight membrane-bound vesicle termed “inclusion”. Protected by this vacuole, the EB differentiates into a RB in a window of time comprised between 4 hours and 8 hours post infection (Shaw et al., 2000). Relaxation of the EB’s crossed-linked membranes by reduction of disulfide bridges in its outer membrane may result from the activity of the host PDI protein (Abromaitis & Stephens, 2009). One other important transformation takes place at the bacterial nucleoid. EB DNA is highly compacted with two histone H1 homologues: Hc1 and Hc2 (Barry, Brickman, & Hackstadt, 1993). Those two proteins, tightly attached to DNA, are surprisingly not regulated by degradation or transcription. In fact, it is the translation of the genes hctA and hctB, coding respectively Hc1 and Hc2 that is impaired. Using E. coli and the lethal phenotype observed by hctA expression, Grieshaber et al. identified two genes of Chlamydia able to rescue E. coli. The first protein, IspE, seems to be indirectly implicated in the production of small metabolites necessary to disrupt the DNA-Hc1 binding (N. A. Grieshaber, Fischer, Mead, Dooley, & Hackstadt, 2004). The other gene codes for a small regulatory RNA named IhtA (inhibitor of hctA translation). This small, non-coding RNA inhibits HctA translation and does not affect transcription or the mRNA stability (N. A. Grieshaber,

27 Grieshaber, Fischer, & Hackstadt, 2005). In a third paper, the same authors demonstrated that Hc2 regulation is mediated via IspE and not via IhtA (N. A. Grieshaber, Sager, Dooley, Hayes, & Hackstadt, 2006). To better understand the beginning of the development cycle, different teams have studied the transcriptional profiles of Chlamydia. The paper of Shaw and colleagues described three temporal classes of genes wherein the early translocated genes correspond to the EB-to- RB differentiation (Shaw et al., 2000). While Belland et al. confirmed these results in 2003, they added the detection of new transcripts as early as one-hour post infection (Belland et al., 2003). Two different groups of genes belong to the early class: “immediate early” for 29 genes and “early” for another 200 genes. Chlamydia immediately expresses so many genes at the beginning of its cycle because it needs to “build up” its niche. This parasitophorous vacuole is considered as “non-fusogenic” with the lysosomal pathway because it does not resemble other known eukaryotic intracellular compartments. Already after one hour of infection, markers of plasma membrane or early endosomes are not seen on the nascent inclusion (Scidmore et al., 2003). This requires bacterial activity since when bacterial protein synthesis is inhibited, the inclusion eventually fuses with lysosomes, although that event is surprisingly delayed. Lane and colleagues investigated the recruitment of several host proteins within minutes after infection. They showed that phosphatidylinositol 3,4,5- triphosphate (PIP3) is rapidly and transiently recruited to the inclusion membrane; the PIP3 binding protein Vav2 is also recruited at the entry site (Lane et al., 2008). One chlamydial protein, CT147, has been proposed to participate in the regulation of the interaction between the inclusion and early endosomes. CT147 shows some similarities with the early endosome antigen 1 (EEA1), a marker of early endosomes. Like EEA1, it has a zinc-finger domain for PI3P binding, but lacks the Rab5 interacting domain (Belland et al., 2003). This particular Rab protein is known to regulate the fusion of EE with each other or with other vesicles (Stenmark, 2009). In their study, Belland et al. show that CT147 is detected around the inclusion from 8 hours post-infection and not before, even though its transcription is detected from the first hour of infection. Within two hours after entry, the bacteria are trafficked toward the microtubule- organizing center (MTOC) (Clausen, Christiansen, Holst, & Birkelund, 1997). This phenomenon is dependent on the host cell dynein and on bacterial protein synthesis (S. S. Grieshaber, Grieshaber, & Hackstadt, 2003). Interestingly, the overexpression of the p50 dynamitin, which in other systems is sufficient to block dynein/cargo interaction, had no

28 effect on Chlamydia transport. One hypothesis is that a bacterial protein present on the inclusion membrane links it to dynein to travel to the MTOC. The presence of bacterial proteins on the inclusion is well documented. It has been estimated that 7% to 10% of Chlamydia proteome is represented by inclusion proteins (Inc) (Dehoux, Flores, Dauga, Zhong, & Subtil, 2011). This family of proteins shares the particular characteristic of having a large hydrophobic domain of sixty residues (Bannantine, Griffiths, Viratyosin, Brown, & Rockey, 2000). Among all the Inc proteins, some are transcribed between the entry moment and the first two hours of infection (Scidmore-Carlson, Shaw, Dooley, Fischer, & Hackstadt, 1999). For example, incD, incE, incF, and incG, which belong to the same operon, are transcribed during this time. The four proteins encoded by these genes are observed at the membrane of inclusions by immunofluorescence. Moreover, IncG is detectable by 2 hpi. Because Inc proteins are exposed to the host cytoplasm, they likely participate in the interactions between the bacteria and the host.

e. Multiplication and development: the reticulate bodies party

It is considered that each EB has fully differentiated into RB after 8 hpi (Abdelrahman & Belland, 2005). Three phases of gene transcription have been described: early, mid-cycle and late (Shaw et al., 2000). Chlamydia retains the ability to regulate gene expression transcriptionally and post-translationally. Regulation of transcription in Chlamydia remains poorly understood, however, since only a few transcription factors are known (Tang & Bavoil, 2012). Moreover, the recent description of regulating non-coding RNA (ncRNA) gives an additional level of regulation for chlamydial gene expression (Abdelrahman, Rose, & Belland, 2011). As already mentioned above, the reticulate bodies are metabolically active and are the replicative form of the bacterium. The replication of the bacteria follows first an exponential phase with a doubling time of around 2 hours before slowing down and reaching a plateau at the end of the cycle (Shaw et al., 2000). The division mechanism per se of Chlamydia remains poorly understood. The bacteria lack certain essential genes of bacterial division such as FtsZ (Ouellette, Karimova, Subtil, & Ladant, 2012; Stephens et al., 1998). Intriguingly they express a protein associated with the rod-shape of bacillus bacteria, MreB, when Chlamydia clearly has a coccoid shape (Ouellette et al., 2012). The work of Ouellette and colleagues brought the first clues to this surprising Chlamydia characteristic. Using specific inhibitors of Pbps and MreB, they showed that the penicillin binding proteins Pbp2, Pbp3, and MreB are

29 required for Chlamydia division. Also, the interaction of MreB with another division protein FtsK led to the proposition that MreB acts as a substitute for FtsZ. Thus, Ouellette and colleagues brought the first piece of evidence for the existence of a very unusual division mechanism in Chlamydia.

The late phase in the developmental cycle starts around 20 hpi, when RBs start to differentiate into EBs in an asynchronous manner (Shaw et al., 2000). To date, it is not known what triggers the RB-to-EB differentiation. One possibility would be a quorum sensing system even though there is no evidence of such a system or its sensors in the annotated Chlamydia genome. Still, many genes in its genome remain hypothetical and the existence of quorum sensing in a community inside the inclusion microenvironment is an attractive hypothesis. Another hypothesis of the RB-to-EB transition is that the release of the RBs from the inclusion membrane to the inclusion lumen triggers the differentiation to the EB form. The paper of Wilson and collaborators supports this theory by following the movement of both forms in live imaging (Wilson, Whittum-Hudson, Timms, & Bavoil, 2009). However, this model would only be feasible in C. trachomatis because other species of Chlamydia do not require attachment to the inclusion membrane (e.g. C. pneumoniae). The physical process of differentiation itself is better understood. The transcription of the genes hctA and hctB coding for the histone-like proteins Hc1 and Hc2 is concomitant with the timing of the differentiation (Belland et al., 2003). Moreover the same study shows the concomittant expression of gyrases, which might act on the DNA topology when it needs to be compacted. RB-to-EB transition also involves reassembly of the chlamydial outer membrane complex (COMC) including OmcA and OmcB. Disulfide cross-links are made in the COMC most likely by isomerases also expressed late in the cycle (Tang & Bavoil, 2012). The disulfide bonding of components of the type III secretion apparatus follows the same pattern, with a reduction in EB-to-RB differentiation and an oxidation in RB-to-EB differentiation (Betts- Hampikian & Fields, 2011).

f. The persistent form: a party forever?

The possibility of persistent C. trachomatis infection arose from the observation of seemingly chronic genital infections. Persistence is defined as a long-term association between Chlamydia and the host in which the bacteria are viable but cannot propagate. As discussed, the same patient can show several episodes of chlamydial infections in his lifetime,

30 which can result from reinfection but also from persistence of the organism after unresolved infection. Early studies of persistence revealed that, in vitro, Chlamydia became abnormal in the presence of IFN-" (Shemer & Sarov, 1985), a state later described by Wyrick as: “morphologically enlarged, aberrant, non-dividing, viable but non cultivable”. This “phenotype” can actually be observed following a variety of treatments: iron, amino acids, or nutrient starvation; exposure to penicillin or IFN-"; or even during host cell maturation. Transcriptomics studies showed that the gene expression profiles differ with the inducer, suggesting that, although morphologically identical, the bacteria are not in an equivalent state. The main question in the field remains to understand how these in vitro systems reflect clinical situations (Wyrick, 2010). The persistent stage might allow the bacteria to survive in a transiently unfavorable environment. This is observed in vitro after adding back nutrients in the culture medium for example or in vivo in the case of lack of treatment observance for patients infected by Chlamydia (Hogan, Mathews, Mukhopadhyay, Summersgill, & Timms, 2004).

g. Parasitism of the host cell by Chlamydia

Chlamydia has a surprisingly low number of genes (about 900 for C. trachomatis) and comparison of the pathogenic strains with their environmental relatives show that, if both have lost several biosynthetic pathways, the pathogenic strain has further reduced its genome (Omsland et al., 2014). As a consequence, Chlamydia relies heavily on the host cell for many metabolites such as amino acids and nucleotides (Collingro et al., 2011; Horn et al., 2004; Stephens et al., 1998). The main metabolites gained from the host will briefly be exposed here. The second chapter will give some details about the pathways involved. To begin with, it has been described that Chlamydia takes up several host lipids including sphingolipids (Hackstadt, Rockey, Heinzen, & Scidmore, 1996; Hackstadt, Scidmore, & Rockey, 1995; Moore, Fischer, Mead, & Hackstadt, 2008), cholesterol (Carabeo, Mead, & Hackstadt, 2003), and glycerophospholipds (Wylie, Hatch, & McClarty, 1997) mostly for building its own membrane. Chlamydia is able to import ATP from its host and to hydrolyze it as an energy source (T. P. Hatch, Al-Hossainy, & Silverman, 1982). Nevertheless, it has been shown that the bacterium has the capacity to produce its own ATP from glucose metabolism and possibly even oxidative phosphorylation (Iliffe-Lee & McClarty, 1999). This theory fits with the work of Schwöppe and colleagues which demonstrated the possible uptake of glucose-6-phosphate

31 from the host (Schwoppe, Winkler, & Neuhaus, 2002). Moreover, genome sequencing of C. trachomatis showed the presence of an intact pentose phosphate pathway as well as all the enzymes necessary for glycogen storage and degradation (Stephens et al., 1998). In addition, ribonucleoside triphosphates are derived from the host mainly to synthesize bacterial RNA (T. P. Hatch, 1975).

h. The end of the show: a double way out ticket

At the end of the cycle, the inclusion contains several hundreds of EBs ready to infect new cells (Wyrick, 2000). In vitro, lysis of the infected cell is most commonly observed at the end of the cycle. However, early electron microscopy experiments imaging infected McCoy cells showed evidence of parts of inclusions released outside of the cell as a cell-free inclusion (la Maza & Peterson, 1982). Moreover, authors pictured some intact cells with a “crater”, interpreted as a scar left by the exit of Chlamydia inclusion without cell destruction. Years later, the group of Stephens studied more specifically this phenomenon that has been called extrusion (Hybiske & Stephens, 2007b). Using a stable GFP expressing HeLa cell line, Hybiske followed by live microscopy the evolution of the chlamydial inclusion seen as a growing black hole inside the cell. This approach allowed the authors to distinguish the two different ways of exit, lysis or extrusion, and to establish that both ways can be employed by Chlamydia (Figure 4). The lysis event starts with the rupture of the inclusion membrane followed by the bursting of the cell within a maximum of 15 min in a protease dependent manner. In contrast, extrusion is slower (3 hours) and requires actin polymerization, N-WASP, myosin II, and Rho-GTPase.

Figure 4: Exit of Chlamydia by two different ways, (A) lysis or (B) extrusion. (Hybiske & Stephens, 2007a). A) Lysis happens in two steps with first the rupture of the inclusion membrane followed by that of the host cell membrane. B) Extrusion is a process, which is close to cytokinesis and involves the actin cytoskeleton and Rho GTPases.

5. Chlamydia’s swiss knife: the secretion systems

As with other intracellular pathogens, Chlamydia interacts with its host mainly through different secretion systems. Three secretion systems (SS) are present in Chlamydia: type II (T2SS), type III (T3SS), and type V (T5SS). T2SS or T5SS might in some instance be followed with the formation of outer membrane vesicles (OMV) to export specific cargo out of the bacteria (Tang & Bavoil, 2012). Regardless, T3SS is likely the major pathway used by Chlamydia to secrete proteins.

a. The T2SS: the general secretory pathway (GSP)

This secretion pathway is also known as the GSP pathway or the Sec pathway. It is a very conserved system found in all living organisms (Saier, 2006). Chlamydia possesses 5 components out of 15 in E. coli (Peabody, 2003). It is likely that other uncharacterized proteins are present in the genome to replace the missing components. The secretion of some chlamydial proteins such as the chlamydial protease-like activity factor (CPAF) appear to depend on T2SS (D. Chen et al., 2010). The T2SS does not accomplish the translocation of

33 effectors into the host cytoplasm and an additional system is needed. Small vesicles inside the inclusion lumen might do the job (Giles, Whittimore, LaRue, Raulston, & Wyrick, 2006). These vesicles are most probably OMV, well reviewed by Ellis and Kuehn (Ellis & Kuehn, 2010). Thus further work on this pathway (including OMV) is necessary to understand its role in the context of chlamydial growth and infectivity.

b. The T5SS: the twin-arginine translocation (tat) pathway

Also called the autotransporter pathway, this process is dependent on the Sec pathway for secretion of effectors to the periplasm and then a cleavage releases a part of the effector. A signal sequence drives the protein from the bacterial cytoplasm across the inner membrane to the periplasm and is removed from the protein by a signal peptidase. The beta-barrel of the cleaved protein is then inserted into the external membrane exposing the N-terminal part of the protein to the outside of the bacterium. Other cleavage could free part of the protein by cutting the beta-barrel. One of the best examples in the Chlamydia field is PmpD (Kiselev, Skinner, & Lampe, 2009; Kiselev, Stamm, Yates, & Lampe, 2007; Wehrl, Brinkmann, Jungblut, Meyer, & Szczepek, 2004). Authors showed that PmpD can be separated in three parts: a signal sequence, a passenger domain, and a beta barrel. The first part is cleaved while being secreted by the Sec pathway, delivering the protein to the periplasm. The beta-barrel is inserted in the outer membrane with the passenger domain exposed outside. Afterwards, cleavage of the passenger domain may free functional domains, which may be further secreted into the host cell through OMVs.

c. The T3SS: the chlamydial “injectisome”

This system is the masterpiece of Chlamydia’s arsenal and has attracted many studies. It is present on both EBs and RBs. Proteins of the system are present in purified EBs even before contact with the host cell (Fields, Mead, Dooley, & Hackstadt, 2003). Upon binding to the plasma membrane, the so-called injectisome might rapidly be activated through its association with sphingolipids, cholesterol, or unknown ligands (Jamison & Hackstadt, 2008). The secretion is active within the first minute of bacterial binding since TARP phosphorylation has been detected as early as 5 min after infection. It is not known whether genes are rapidly transcribed de novo (the first time point in the transcriptomics study comes

34 90 min after infection (Belland et al., 2003; Fields et al., 2003)), but it is clear that several proteins are pre-packaged in the EBs for secretion in the early steps. The first Inc proteins are detected 2 hours after infection and are secreted throughout the developmental cycle, attesting to the presence of active secretion in the RB state as well. The secretion apparatus is composed of more than ten proteins (Figure 5). Instead of having all the genes coding for this system on a pathogenicity island or on a plasmid, Chlamydia has them dispersed throughout the genome (Hefty & Stephens, 2007; Subtil, Blocker, & Dautry-Varsat, 2000). The assembly of the system happens sequentially. It starts with the insertion of the CdsC protein in the outer membrane followed by its interaction with CdsD, which triggers CdsJ, and both constitute the base of the apparatus in the inner membrane. Other proteins will associate in the inner membrane before the formation of the “needle” and the tip, respectively composed of CdsF and CT584, from the outer membrane to the host membrane or the inclusion membrane (Betts-Hampikian & Fields, 2010). Several chaperones have also been identified in the Chlamydia genomes. Their interaction with different components of the secretion apparatus have been described (Brinkworth et al., 2011; Spaeth, Chen, & Valdivia, 2009)

Figure 5: The assembly of the T3SS of Chlamydia (Betts-Hampikian & Fields, 2010).

35 Secretion of effectors is accomplished by recognition of a signal sequence contained in the N-terminal thirty amino acids of the protein. Several laboratories have worked on the identification of the substrates of the T3S, called effector proteins. Since no genetic tools were available at the time, heterologous secretion systems have been used, based on the universality of the secretion signals across all bacteria tested so far (Dean, 2011). Yersinia (Fields & Hackstadt, 2000), Salmonella (Ho & Starnbach, 2005), and Shigella (Subtil, Parsot, & Dautry-Varsat, 2001) have been used to identify more than thirty candidate effector proteins (not including the Inc proteins, which are also T3SS substrates).

6. The challenge Chlamydia: from darkness to light

This section reviews the specific challenges met by researchers in their investigation of Chlamydia biology.

a. A long time in the darkness…

Ever since the initial observation of inclusions in 1907, Chlamydia has not been a “cooperative” bug. It took nearly 60 years to categorize it as a bacterium. Because of its biphasic developmental cycle, Chlamydia is not easily genetically manipulated. The EBs survive many experimental conditions, but their cross-linked cell wall make it difficult to introduce DNA or other molecules. Even if this step is achieved, the compact DNA and the absence of replication make all common techniques of bacterial genetic manipulation ineffective. On the other hand, while RBs would be more amenable to genetic manipulation, other difficulties arise. First of all, RBs are inside inclusions, meaning that two eukaryotic and two prokaryotic membranes separate its DNA from the experimenter. If manipulated outside the cell, it is very fragile and rapidly killed. In any case, the strongest obstacle to a strategy based on extracellular manipulation of the RB stage is its inability to initiate an infectious cycle, making selection of mutants impossible. Finally, the last disadvantage of this organism is that it is not cultivable on axenic media, requiring tissue culture for growth. Selection of mutants (for instance after chemical mutagenesis) is therefore a very slow and tedious process.

36 b. … and then there was light: welcome to a new Chlamydia world.

For the Chlamydia world, the big (r)evolution happened in 2011. Various examples mentioned above show the crucial need of a method to knock out the gene of interest to find out its function. A first improvement came from two different strategies to isolate isogenic mutant strains. These strategies are rather time-consuming and labor-intensive but succeeded at isolating the first mutant of a gene of interest (Kari et al., 2011) or for a phenotype of interest (incriminated genes were then identified by whole genome sequencing) (Nguyen & Valdivia, 2012).

Meanwhile, the group of Clarke described the first stable transformation of C. trachomatis, based on classical CaCl2-based treatment to make the EB competent for DNA uptake (Y. Wang et al., 2011). They took advantage of the existence, in all C. trachomatis isolates, of a plasmid of around 7.5 kb, in which they introduced a penicillin resistance gene to serve as a selection marker. As a proof of concept, they obtained green fluorescent chlamydiae, which were able to develop normally in the presence of penicillin. This approach opened many possibilities in the field and several developments based on this method have already been published (Agaisse & Derré, 2013; Wickstrum, Sammons, Restivo, & Hefty, 2013). Recently, the transformation protocol has been used to insert a plasmid with a tagged version of the protein of interest under an inducible promoter (Bauler & Hackstadt, 2014). The authors showed that addition of a flag tag at the C-terminus of the IncD protein did not impair its secretion by the T3SS, opening the possibility to localize T3S substrates by immunofluorescence. Until now, this required obtaining good antibodies, a clear bottleneck in the field. Another long-awaited development was the possibility to perform targeted gene deletion. Johnson and Fisher used the “TargeTron™” system from Sigma to insert an intron in a selected place into the genome (C. M. Johnson & Fisher, 2013). As a proof of principle, they knocked out IncA, reproducing a well-established phenotype of fusion-incompetent inclusions (Hackstadt, Scidmore-Carlson, Shaw, & Fischer, 1999). Notably, the use of this attractive technology will only allow the knock out of non-essential genes. Moreover, since Chlamydia has already a very restricted genome, it is very likely that only few genes will be consider as non essential.

37 The possibility to use fluorescent bacteria is particularly attractive for immunofluorescence studies, in particular live imaging (Agaisse & Derré, 2013). We reasoned that, if the fluorescent signal is strong enough, it might also allow the monitoring of the bacterial developmental cycle by microscopy and by flow cytometry, in a faster and more quantitative manner than previously achieved. Part of this doctoral work addressed this question (Article 2).

II. Interactions between Chlamydia and host membranes

The first membrane encountered by Chlamydia is the plasma membrane. Interactions that occur at this stage are of utmost importance for the bacterium, since its growth depends on its ability to enter the cell. Once inside the host cell, interactions between the bacteria and host membranes remain very important: accesses to lipids, which are incorporated into the bacterial membranes, and to many nutrients depend on the bacteria modifying the inclusion membrane. Because the bacteria remain in the inclusion throughout their developmental cycle, some interactions are mediated through components inserted in the inclusion membrane, which is designed by the bacteria for that purpose. In this chapter, the different levels of interactions between Chlamydia and host plasma membrane or other intracellular compartments are reviewed. Exploiting host lipids is not restricted to intracellular bacteria. This dissertation includes one article that reviews the different reasons why bacteria target eukaryotic lipids and the strategies employed (Article 4 in appendix).

1. The first Chlamydia interaction: entering the cell

After attachment to a susceptible host cell, Chlamydia is dependent on an active mechanism driven by the bacterium to enter the cell (Figure 6). The role of the protein CT456, known as TARP, has been well studied but other effectors are emerging as well. TARP was the first type III secreted effector involved in entry, identified by the Hackstadt group (Clifton et al., 2004). Authors showed that the protein was not exposed at the bacterium surface and was rapidly translocated via the T3SS into the cytoplasm where it acts. They observed that the entry site is phosphorylated within 30 seconds after the beginning of

38 the infection. Transfected TARP was also phosphorylated and induced intense actin polymerization, suggesting that the protein triggers the actin polymerization by itself. Further investigations demonstrated that TARP nucleates actin assembly (Jewett, Fischer, Mead, & Hackstadt, 2006) and that the phosphorylation is not necessary for actin recruitment (Clifton et al., 2005). Also, the dissection of the protein showed that actin recruitment could be achieved with the C-terminal domain of TARP while a domain composed of 50 amino acids tandem repeats triggered the phosphorylation. An other molecular study of the C. trachomatis TARP highlighted special domains important in the binding to F-actin and the bundling of actin filaments (Jiwani et al., 2013). Phosphorylation of the protein is attributed to Src family tyrosine kinases triggered by the presence of many Src- like consensus targets on TARP (Jewett, Dooley, Mead, & Hackstadt, 2008). A chaperone protein, Slc1, facilitates TARP translocation (Brinkworth et al., 2011). Very recently, Slc1 was shown to engage multiple early chlamydial effectors, including the newly translocated early phosphoprotein, TepP (Y.-S. Chen et al., 2014). Translocation of TARP and TepP contributes to the activation of signaling cascades in the host cell, with the recruitment of various kinases at the site of entry such as Rac1 or the phosphatidylinositol 3- kinase (PI3K) (Carabeo et al., 2004; Lane et al., 2008). Also, the phosphorylation of different residues of TARP allows the protein to bind specifically to two different guanine nucleotide exchange factors (GEF), Sos1 and Vav2. This binding triggers Rac GTPase activation followed by WAVE complex assembly, which stimulates Arp2/3 complex for a synergistic action with TARP on actin polymerization (Lane et al., 2008).

Figure 6: Chlamydia entering the cell (Carabeo, 2011). Chlamydia acts on both the WAVE signaling pathway and on actin nucleation to trigger its uptake. Chlamydia T3 secreted effectors, such as TARP or CT694, are key components for the bacterium entry.

39 2. Chlamydia fattens up: delivery of lipids to the inclusion

C. trachomatis is poorly equipped with genes implicated in lipid anabolism. Enzymes implicated in phosphatidylserine and phosphatidylethanolamine synthesis are present but others are missing (Stephens et al., 1998). Trafficking of lipids from the host to the bacteria likely has been observed over the years (Carabeo et al., 2003; Hackstadt et al., 1995; G. M. Hatch & McClarty, 1998; Wylie et al., 1997). However, even if the bacteria acquire most of its lipids from the host (Figure 7), the mechanism of uptake and the origin of the host lipids remain unclear (Elwell & Engel, 2012). Import of sphingolipids and cholesterol has attracted the most attention. Hackstadt and colleagues followed the fluorescent marker C6-NBD-Cer, which traffics to the Golgi apparatus where it serves as a precursor for sphingomyelin synthesis (Hackstadt et al., 1995). The marker was trafficked in an active way to the inclusion membrane and sphingomyelin incorporation into Chlamydia cell walls was detected. Furthermore, sphingomyelin incorporation starts one hour after the entry, when the bacterium is not in the Golgi apparatus region. The origin of the sphingomyelin is not the plasma membrane but the Golgi apparatus, suggesting a specific re-routing for this metabolite (Hackstadt et al., 1996). From the same group, Scidmore showed later that the exocytosis of glycoproteins out of the Golgi occurred normally (Scidmore, Fischer, & Hackstadt, 1996). Chlamydia also affects the Golgi-to-plasma membrane trafficking during the specific sphingomyelin vesicles uptake (Moore et al., 2008). Thus, the uptake of sphingomyelin in the inclusion is a selective process. Two small GTPases of the Rab family, Rab6 and Rab11, were implicated in this process based on the observation that knock down of one or the other abrogated sphingolipids delivery (Rejman Lipinski et al., 2009). Recent work showed that, at least for C. trachomatis L2, the vesicle-associated membrane protein 4 (Vamp4) as well as syntaxin 6 are implicated in sphingomyelin acquisition (Kabeiseman, Cichos, Hackstadt, Lucas, & Moore, 2013). These two species recruit Vamp4 at the inclusion membrane, and knocking down Vamp4 blocked sphingomyelin uptake by the bacteria. Surprisingly, C. trachomatis serovar D does not recruit Vamp4, and, in this strain, sphingomyelin acquisition proceeds normally in Vamp4 depleted cells, suggesting that sphingomyelin acquisition proceeds through species-specific mechanisms. Concerning cholesterol, it is likely that it follows similar pathways as sphingomyelin (Carabeo et al., 2003). Multivesicular bodies (MVB) have also been described as a possible source for sphingolipids and cholesterol (Beatty, 2006; 2008). Carabeo and colleagues

40 pointed out that the cholesterol source could be either LDL or de novo synthesis. Indeed, it was shown that the high density lipoprotein (HDL) biogenesis machinery, including ABCA1 and CLA 1, is recruited around the inclusion membrane (Cox, Naher, Abdelrahman, & Belland, 2012). Moreover, the extracellular lipid acceptor ApoA-1 is present inside the inclusion with some phosphatidylcholine. The knock down of ABCA1 blocked C. trachomatis growth. Finally, the authors showed that drugs inhibiting the transport activity of ABCA1 or CLA 1 negatively affect the bacterial development. An alternative strategy for the bacteria to obtain lipids from its host is to target lipid droplets (LDs). The Valdivia group identified three chlamydial proteins (Lda proteins - Lipid droplets associated proteins) translocated into the host cell cytoplasm where they associate with lipid droplets (Kumar, Cocchiaro, & Valdivia, 2006). In collaboration with the Hackstadt lab they later described the translocation of LDs into the inclusion lumen (Cocchiaro, Kumar, Fischer, Hackstadt, & Valdivia, 2008). Recent work by our laboratory shows that another small organelle of the host, the peroxisome, is translocated into the inclusion lumen. Because peroxisomes and LD are known to associate, they might be ingested simultaneously. Mass spectrometry analysis also revealed, in the bacteria, the presence of particular phospholipids called plasmalogens, whose synthesis occurs in part in peroxisomes. (Boncompain et al., 2014). Finally, lipid uptake might also follow non-vesicular routes thanks to the protein CERT, a lipid transfer protein (Derré, Swiss, & Agaisse, 2011; Elwell et al., 2011). CERT transfers ceramide from the endoplasmic reticulum to the Golgi apparatus. CERT is recruited to the membrane of the inclusion by the IncD protein and might transfer ceramide directly from the adjacent endoplamic reticulum into the inclusion membrane. Early studies showed that Chlamydia tolerates variations in their lipid composition, since changing lipid homeostasis of the host changed bacterial lipid composition without impact on growth or infectivity (G. M. Hatch & McClarty, 1998). It has now become clear that Chlamydia obtains lipids from multiple sources, and this redundancy protects the bacteria from possible changes in lipid fluxes in the host. A consequence of this is that, in vivo, the bacteria may have variable lipid composition depending on the lipid homeostasis in the infected tissue.

Figure 7: Schematic of the different pathway that Chlamydia uses to get host lipids and nutrients (Saka & Valdivia, 2010). Chlamydia acquires lipids by targeting different organelles: Golgi apparatus (a), MVBs (d), lipid droplets (e); and by using different pathways: Rab proteins (f), Vamp proteins (g).

3. A sociable guest: multiple interactions between Chlamydia and host organelles

a. Chlamydia attracts most if not all the host organelles

The bacterium uses the cytoskeleton to move from the entry foci to the MTOC region of the cell. Van Ooij and coworkers were the first to look by immunofluorescence of Golgi markers (van Ooij, Apodaca, & Engel, 1997). Staining the trans-golgi network (TGN) protein TGN38, they noticed that the TGN was losing integrity in infected cells. Later, Heuer and colleagues followed several Golgi markers (GM130, golgin-84 or giantin) to arrive to the same conclusion: the Golgi apparatus is fragmented by Chlamydia (Heuer et al., 2009). Inducing fragmentation using siRNA against different components of the apparatus favored Chlamydia growth, suggesting that Golgi fragmentation is important for Chlamydia development.

42 The work of Beatty provided some evidences for the action of Chlamydia on the MVB, describing MVB specific components (CD63, MLN64 and LBPA) inside the inclusion ((Beatty, 2006; 2008). Though, regarding the weakness of these stainings, it is possible that the respective antibodies cross reacted with the bacteria. Indeed, the work of Ouellette and Carabeo invalidated the uptake of CD63. They followed by live microscopy LC3-GFP and showed that it does not enter in the inclusion (Ouellette & Carabeo, 2010). The recruitment of the endoplasmic reticulum (ER) at a very close proximity of the chlamydial vesicle has been recently reported (Agaisse & Derré, 2014; Derré et al., 2011; Dumoux, Clare, Saibil, & Hayward, 2012). Markers of early endosome (transferrin) as well as late endosomes (mannose-6- phosphate receptor) have been described around the inclusion at 4 hpi but also at 20 hpi (van Ooij et al., 1997). Recycling endosomes are also intimately associated with the inclusion (Ouellette & Carabeo, 2010). Ouellette and Carabeo showed that the recycling of transferrin is an important pathway for Chlamydia and that its disruption by specific inhibitors affects bacterial growth, probably by affecting the ability of the organism to acquire iron Interestingly, it is to note that C. trachomatis, as well as C. pneumoniae, share the particularity not to be surrounded by mitochondria while many other chlamydial species are. (Matsumoto, Bessho, Uehira, & Suda, 1991).

b. The Inc proteins: key components of the Chlamydia-host cell interaction

Among all the Chlamydia proteins, the family of Inc proteins lie at the forefront of the host-pathogen interactions, since they are located on the membrane that makes the interface between the bacteria and the cell. This family is specific to the Chlamydia phylum. Bioinformatic data suggest that each Chlamydia proteome is composed of 7% to 10% of Inc proteins and it was experimentally shown that they are secreted via the T3SS. Their particularity is to have at least one bi-lobal hydrophobic domain made of two transmembrane helices separated by thirty amino acids (Dehoux et al., 2011). The hydrophobic domain anchors the Inc proteins to the inclusion membrane facing the host cytoplasm. One of the most studied member of the family is IncA. This protein has been shown to have a SNARE-like motif (Delevoye et al., 2008). Also, microinjection of anti-IncA antibodies and natural IncA deficient strain (Suchland, Rockey, Bannantine, & Stamm, 2000) contributed to show that IncA is responsible for the homotypic fusion of the inclusion membrane (Hackstadt et al., 1999). This result was confirmed by the first targeted intron

43 mediated mutant in Chlamydia, targeting IncA, which resulted in a multi-inclusion phenotype (C. M. Johnson & Fisher, 2013). Many other examples already mentioned in this manuscript showed the implication of Inc proteins in uptake of nutrients and the recruitment of host-proteins including regulators of transporter proteins.

c. Manipulation of the Rab proteins, key trafficking regulators

Rab proteins are small GTPase associated to membrane of organelles or at the cell surface (Figure 8). They have a central role in membrane trafficking in general, contributing to the specificity of the fusion between two compartments.

Figure 8: Overview of different Rab proteins and their specific localization in membranes. The distribution of Rab proteins is linked to the distribution of specific phosphoinositides. Taken from Jean and Kiger, 2012 (Jean & Kiger, 2012).

Recruitment of membrane trafficking mediators such as Rab1, Rab4, Rab6, Rab10, Rab11 and Rab14 around the inclusion has been described from 10 hpi (Capmany & Damiani, 2010; Rzomp et al., 2003; Rzomp, Moorhead, & Scidmore, 2006). Rab GTPases are key

44 regulators in host cell trafficking: they contribute to defining the identity and the function of the different compartments in the cell (Pfeffer, 2013). Thus, manipulation or recruitment of Rab proteins by the bacteria represent a powerful strategy to access many cellular functions. If most of those Rabs are described to localize around the inclusion, only few were investigated in more detail. The study of Rab11 reveals recruitment as early as one hour post infection. Later in the cycle, the IncG protein associates with Rab11 around the inclusion. Also, this phenomenon is microtubule independent, as nocodazole or dimethyl-sulfoxyde treatments do not affect the Rab11 localization at 8 hpi or 18 hpi (Rzomp et al., 2003). On the other hand, Rab14 is recruited only from 10 hpi and remains stable around the inclusion until the end of the developmental cycle. Its recruitment depends on chlamydial activity and is crucial for bacterial development since siRNA against Rab14 decreased bacterial multiplication and infectivity. It is probably likely due to a role of Rab14 in lipid acquisition (Capmany & Damiani, 2010). More recently, recruitment of Rab11 and Rab14 has been linked to the protein FIP2, a member of the Rab11-family of interacting proteins. It has been shown that FIP2 possesses a Rab-binding domain allowing it to interact with both Rab11 and Rab14. Leiva and colleagues observed the transitory recruitment of FIP2 from 2 hpi to 18 hpi (Leiva, Capmany, & Damiani, 2013). Moreover, they showed that FIP2 is specifically targeted among all the Rab11-family of interacting proteins, in a bacterial driven process. Finally, they demonstrated that Rab11 is first recruited around the inclusion, then recruits FIP2, which in turn recruits Rab14. Authors also showed the co-localization of those Rabs with IncG, putting this Inc protein as a key member in this mechanism. Similarly the protein OCRL1, a protein linked to the Golgi complex and able to bind several Rab GTPase, is recruited to the inclusion (A. M. Moorhead, Jung, Smirnov, Kaufer, & Scidmore, 2010). Its presence around the chlamydial vacuole is detected at 2 hpi and is mediated by its Rab- binding domain. As well as FIP2, this characteristic fits with the recruitment of a Rab protein first before having other Rab-adapters recruited around the inclusion. The authors also showed that OCRL1 knock down affects Chlamydia growth, as observed for knock down of many proteins linked to the Golgi apparatus function. Altogether, this defines recruitment of different host components in a specific order for the acquisition of important resources (lipids) to the bacteria by the bacteria. Moreover, another protein Rab4 has also been observed at early time points (2 hpi) to co-localize with the inclusion (Rzomp et al., 2006). Authors demonstrated that the Inc protein CT229 interacts with Rab4, and more precisely with the activated form GTP-Rab4.

45 4. Chlamydia trachomatis infection impairs cytokinesis

Besides the influence on host cell compartments, Chlamydia subvert a number of cellular function. Because of its potential relevance for our findings we describe briefly the effect of C. trachomatis infection on cell division. Multi-nucleated infected cells have been observed since the first description of the Chlamydia development cycle. This characteristic of Chlamydia-infected cells is induced by the bacterium since antibiotic addition decreases the number of cells with several nuclei (Greene & Zhong, 2003). Also, several links between Chlamydia and proteins implicated in cytokinesis have been described in the past years. The mitotic protein cyclin B1 has been shown to be cleaved during infection (Balsara, Misaghi, Lafave, & Starnbach, 2006; H. M. Brown, Knowlton, & Grieshaber, 2012). However, this cleavage is probably an artifact due to a lack of proteases inhibitors in the lysis buffer used in that study (A. L. Chen, Johnson, Lee, Sütterlin, & Tang, 2012). However, Brown and colleagues provided other clues about the cytokinesis deficiency. They observed that Chlamydia-infected cells are able to form midbody, one of the last steps before separation of the two daughter cells. Thus, the authors concluded that Chlamydia somehow blocks the abscission during the cell cytokinesis resulting in multinucleated infected cells. A separate study by the group of Sütterlin showed that C. trachomatis infection led to supernumerary centrosomes, likely due to a dysregulation of the host centrosome duplication pathway (K. A. Johnson, Tang, & Sütterlin, 2009). It has been suggested to explain the association of Chlamydia genital infection and possible cervical cancers.

5. The defensive answers of the host

Parasitizing the host-cell is not achieved without consequence and the infected cell reacts to attempt fight against this invasion (Bastidas, Elwell, Engel, & Valdivia, 2013). For example, the bacteria are detected through their LPS via TLR4 and through their Hsp60 via TLR2 and TLR4 (Bulut et al., 2009; Heine, Muller-Loennies, Brade, Lindner, & Brade, 2003). Among other bacterial components detected by the immune system is chlamydial cyclic-di- AMP, recognized by the stimulator of interferon genes (STING), triggering IFN responses (Barker et al., 2013). Moreover, the activation of the innate immune response is observed through the production of diverse chemokines and cytokines in a host-cell and Chlamydia species dependent manner (Bastidas et al., 2013). One strategy to block the NF#B mediated

46 transcriptional response of the host has been described. C. trachomatis secrete a deubiquitinating enzyme ChlaDub1 (Misaghi et al., 2006), which might inhibit the ubiquitination and degradation of I#B$, which in turn would suppress NF#B activation (Le Negrate et al., 2008). The ultimate host-defense for a eukaryotic cell against such an intracellular pathogen remains programmed cell death, also known as apoptosis. Chlamydia actively blocks apoptosis during infection. The CPAF protease was implicated in different cases including the degradation of the pro-apoptotic BH3 only proteins Bad, Puma and Bim (Pirbhai, Dong, Zhong, Pan, & Zhong, 2006) before Puma and Bim degradation was attributed to experimental artefacts (A. L. Chen et al., 2012). Stabilization of the inhibitor of apoptosis protein 2 (IAP2) (Rajalingam et al., 2006) or the anti-apoptotic Mcl-1 up-regulation (Sharma et al., 2011) were also implicated. More recently, the MAP/ERK anti-apoptotic signaling pathway was linked to the up-regulation of Bag-1 (Kun, Xiang-lin, Ming, & Qi, 2013). There, the authors showed that Bag-1 depletion removes the anti-apoptotic effect of Chlamydia while its over-expression in non-infected cells leads to an anti-apoptotic phenotype. Considering the importance for Chlamydia to keep the host cell alive until the completion of its developmental cycle, it is not surprising that redundant pathways are at work to inhibit apoptosis.

6. Chlamydia and ESCRT

The endosomal sorting complex required for transport (ESCRT) is a very important machinery implicated in various mechanisms such as MVB or exosome formation, cytokinesis or even virus budding. So far, no link between Chlamydia and the ESCRT machinery has been reported. Our results show that Chlamydia targets the ESCRT system, which will be introduced in the next chapter.

47 III. Biology of the ESCRT system

1. Overview of a machinery with many complexes

In 2001, the Emr lab characterized in yeast a three protein complex of 350 kDa, which recognized ubiquitinated cargos in MVBs and whose function was needed for their sorting into MVBs intraluminal vesicles (Katzmann, Babst, & Emr, 2001; McCullough et al., 2006). The very first description of an endosomal sorting complex required for transport (ESCRT) was made, called ESCRT-I. A year later, the same lab described one after the other the ESCRT II and ESCRT III complexes, composed of respectively three and four proteins, and implicated in ubiquitinated cargo delivery to the vacuole (lysosome-like compartment in yeast) (Babst, Katzmann, Estepa-Sabal, Meerloo, & Emr, 2002a; Babst, Katzmann, Snyder, Wendland, & Emr, 2002b; Henne, Buchkovich, & Emr, 2011). Another five years later, a fourth complex (ESCRT-0) was discovered, composed this time of only two proteins (Prag et al., 2007; Scott et al., 2005). The proteins composing the ESCRT system have been categorized as “class E” vacuolar protein sorting (vps), based on their common phenotype when knocked down, e.g. a failure in cargo delivery to the vacuole (Henne et al., 2011; Raymond, Howald-Stevenson, Vater, & Stevens, 1992). Since the first description in 2001, about 750 papers have been published in PubMed with “ESCRT” in the title or in the abstract, representing a very dense bibliography for a single machinery. The ESCRT machinery is no longer restricted to a unique function since it is now described to be involved in at least five processes in eukaryotic cells: MVB biogenesis, cell abscission, virus budding, exosome secretion and autophagy (Henne et al., 2011; Hurley & Hanson, 2010). Altogether, the ESCRT system is composed of more than 20 proteins including some associated proteins (Figure 9). It proceeds through a succession of assembly and disassembly of complexes. The aim of this chapter is to give an overview of these different activities. We will include a focus on two specific members of the ESCRT machinery, Hrs and Tsg101, because our results show that they are specific targets of chlamydial infection.

Figure 9: Table of the ESCRT subunits and associated proteins. (Hurley & Hanson, 2010). ESCRT proteins are grouped by complex of proteins: ESCRT-0, -I, -II, and –III. ESCRT associated proteins such as VPS4, Vta1 and Bro1 are also reported.

49 2. One system, many functions

The different components of the ESCRT machinery, from ESCRT-0 to ESCRT-III, are all implicated in the sorting of ubiquitinated cargo into MVBs, which is the process to which they were initially associated. In contrast, we will see that only some of the complexes are necessary to fulfill the other ESCRT-dependent function (Figure 10).

Figure 10: The different implication of ESCRT in the cell. (McCullough, Colf, & Sundquist, 2013). ESCRT machinery acts in MVB formation (ESCRT-0 to ESCRT-III), abscission (ESCRT-I and Alix), in exosome formation (ESCRT-0 and –I) and in virus budding (ESCRT- I and –III).

a. MVB biogenesis

MVBs were named after the observation, by electron microscopy, of large endocytic vesicles containing an accumulation of smaller vesicles in their lumen. In the chronology of the endo-lysosomal pathway, MVBs are situated between the early endosomes (EE) and the late endosomes (LE). Indeed, a MVB is a matured EE.

50 Proteins internalized from the plasma membrane are directed to the EE from which they are either recycled to the plasma membrane, or targeted to the lysosomes for degradation. The signal for degradation is made through ubiquitination of the internalized cytoplasmic tail of the membrane protein. The genesis of the intraluminal vesicles (ILV) inside the EE is made through different steps. EE are enriched in phosphatidylinositol-3-phosphate (PI3P) and Hrs binds to this lipid thanks to its FYVE motif, facilitating the targeting to EE (Agromayor & Martin-Serrano, 2013; Hurley, 2010). Hrs and STAM, composing the ESCRT-0 complex, recognize the ubiquitinated membrane proteins (Figure 11). They bind to ubiquitin via two different kinds of domains: the ubiquitin interacting motifs (UIM) and the VHS (Vps27, Hrs, STAM) motif. Binding to ubiquitinated proteins leads to their clustering.

Figure 11: Shematic organization of the human ESCRT-0. Adapted from Hurley, 2010 (Hurley, 2010). Hrs (purple) and STAM (green) bind to poly-ubiquitinated membrane proteins (red squares). UiM: ubiquitin-interacting motif, DUiM: double UiM, FYVE: zinc- finger domain, VHS: domain present in Vps27, Hrs and STAM, PSAP: motif of interaction with Tsg101, EE: Early endosome.

The second step is the invagination of the membrane, which is mediated by both ESCRT-I and -II. ESCRT-I recruitment is mediated by an interaction between Hrs and Tsg101. In fact, this interaction via the PSAP motif of Hrs defines this protein as a key regulator in MVB biogenesis (Bache, 2003; Hurley & Hanson, 2010). The ESCRT-I components Tsg101 and hMvb12 are also able to bind ubiquitinated cargoes (McCullough et al., 2013; Votteler & Sundquist, 2013). With Vps28 and Vps37, they form a four protein complex. ESCRT-I recruits ESCRT-II (Babst, Katzmann, Snyder, Wendland, & Emr, 2002b; Pornillos et al., 2003) and the two complexes create and/or act in the stability of the vesicle neck (Weiss et al., 2010; Wollert & Hurley, 2010). Loss of either ESCRT-I or ESCRT-II impairs the formation of MVBs (Jouvenet, 2012; Votteler & Sundquist, 2013).

51 The ESCRT-III complex is the central membrane scission machinery, responsible for the closure of the vesicle neck (Hurley, 2010; Votteler & Sundquist, 2013). It is activated by an interaction with components of the ESCRT-II complex. The recruitment of the ALIX protein, responsible for the removal of the complex from the membrane, marks the end of the ILV formation. However, ESCRT-III recruitment via ALIX is dispensable for the lysosomal targeting of cargoes (Bissig & Gruenberg, 2014; Raposo & Stoorvogel, 2013). Even if it has been extensively studied, the exact mechanism of ESCRT-III complex mediated scission is not completely understood (Colombo et al., 2013; McCullough et al., 2013). Before the closure of the ILV, the cargo is de-ubiquitinated by two proteins, AMSH (associated molecular with SH3 domain of STAM) and UBPY. The first one binds to ESCRT-0 and ESCRT-III (McCullough et al., 2006; Raposo & Stoorvogel, 2013) and recognizes only Lys-63-linked polyubiquitin chains while UBPY deubiquitinates both Lys-48 and Lys-63 polyubiquitinated chains (Filimonenko et al., 2007; Henne et al., 2011; J.-A. Lee, Beigneux, Ahmad, Young, & Gao, 2007; Rusten et al., 2007; Tamai et al., 2007). The very end of the ILV formation happens when the ATPase VPS4 binds to ESCRT- III to remove it in an energy dependent manner (Rusten & Stenmark, 2009; Scott et al., 2005). The VPS4 protein activity is modulated by the Vta1 protein (Henne et al., 2011; Rusten, Vaccari, & Stenmark, 2012). The five steps of ILV budding are recapitulated in figure 12.

Figure 12: Schematic view of ILV formation via the ESCRT pathway (Henne et al., 2011)

52 b. Abscission during cytokinesis

Cytokinesis is the final step of cell division. It ends up with abscission, which designates the finale cut in the mid body to make definitive the separation between two daughter cells (Agromayor & Martin-Serrano, 2013; Komada, Masaki, Yamamoto, & Kitamura, 1997) (Figure 13). This process uses only some of the ESCRT proteins: the ESCRT-I and ESCRT-III complexes, and the associated proteins ALIX and VPS4 (Asao et al., 1997; Jouvenet, 2012). The centrosomal protein 55 (Cep55) is localized at the midbody during cytokinesis. It initiates the recruitment of the ESCRT components necessary for the abscission by binding to ALIX and Tsg101. This first recruitment triggers the recruitment of ESCRT-III components like the charged multivesicular bodies proteins (CHMPs) including CHMP4B. To finish, the VPS4 ATPase favors the disassembly of the ESCRT-III, which ends the abscission process possibly by inducing a constriction around the midbody (Agromayor & Martin-Serrano, 2013; Kanazawa et al., 2003).

Figure 13: Scheme of the abscission process mediated by ESCRT-I and –III proteins. (Agromayor & Martin-Serrano, 2013). Interaction of Cep55 with Tsg101 and ALIX triggers the recruitment of ESCRT-III, which might induce constriction around the midbody before its removal via VPS4 action.

c. Virus budding

The role of ESCRT in virus budding from the plasma membrane is very analogous to its role in the abscission process. The same proteins, ESCRT-I, ESCRT-III, ALIX, Nedd4, and VPS4, are implicated (Babst, 2005; Hurley & Hanson, 2010). Up to date, almost forty

53 different viruses were shown to use ESCRT components to bud from their host cell (Nikko & André, 2007; Votteler & Sundquist, 2013). Viruses use diverse ways to hijack the needed ESCRT components. Only the case of the HIV will be described here as an example (Figure 14). The HIV interacts directly with the ubiquitin E2 variant (UEV) domain of Tsg101 via the protein Gag by a P(T/S)AP sequence, which is also found in the ESCRT-0 component as well as in their specific adaptors (Bache, 2003; Pornillos et al., 2003). It has been shown that K-63 polyubiquitin chains present on the Gag protein might favor the budding of the virus (Hirano et al., 2006; Weiss et al., 2010). Ubiquitination of the Gag protein via the E3 ubiquitin ligase Nedd4L seems to be part of the budding process as well (Hayakawa, 2003; Komada et al., 1997; Votteler & Sundquist, 2013). This ubiquitination could favor the interaction with Tsg101 and ALIX via two distinct motifs. In turn, activation of ESCRT-I and ALIX implies the recruitment of ESCRT-III components for the scission of the virion outside of the cell.

Figure 14: Scheme of the HIV budding. (Votteler & Sundquist, 2013).

d. Exosomes secretion

Exosomes are extracellular vesicles (EV) originating from the fusion of MVBs with the plasma membrane (Figure 15) (Mizuno, Kawahata, Okamoto, Kitamura, & Komada,

54 2004; Raposo & Stoorvogel, 2013). The lipids, RNA, or cytosolic proteins originally contained with the ILVs and that are released in the extracellular space play important roles in intercellular communication. The Raposo lab recently evaluated the implication of ESCRT in exosomes biogenesis and secretion (Colombo et al., 2013; Lu, Hope, Brasch, Reinhard, & Cohen, 2003). They followed different marker of exosomes such as the major histocompatibility complex II (MHC-II), CD63 and HSC70 after depletion of several ESCRT proteins. Their results indicate that the individual depletion of different ESCRT proteins affect the exosome biogenesis: Hrs, STAM, Tsg101, CHMP4C, Alix, and VPS4. If the ESCRT-0 components (Hrs and STAM), as well as the Tsg101 (ESCRT-I) depletion mediated by ShRNA decreased the overall exosome secretion, the depletion of VPS4 increased it. Also authors showed that Alix depletion triggers an increase of MHC-II markers detected. Thus, ESCRT proteins are implicated in the exosome biogenesis and secretion.

Figure 15: Scheme of microvesicles (MVs) and exosome release. (Raposo & Stoorvogel, 2013). MVE: multivesicular endosome.

e. Autophagy

Macroautophagy and ESCRT have been linked because the loss of function of ESCRT triggers an accumulation of autophagosomes (Filimonenko et al., 2007; J.-A. Lee et al., 2007; Rusten et al., 2007; Stern et al., 2007; Tamai et al., 2007) Indeed, many ESCRT proteins seem to affect autophagosome biosynthesis since their depletion triggers a changing of the organelle phenotype (Rusten & Stenmark, 2009).

55 The neck closure in the phagophores is similar to the neck observed on ILV in formation. The implication of the ESCRT machinery in the autophagy process remains unclear (Rusten et al., 2012). One likely hypothesis is that ESCRT is required for the fusion of autophagosomes with endolysosomes.

3. Focus on Hrs a. From Hrs to the ESCRT-0 complex

The first description of the hepatocyte growth factor-regulated tyrosine substrate (Hrs) dates back to 1995, when Komada and coworkers described a protein which migrate at 115 kDa and is phosphorylated after growth factor receptor activation (Komada & Kitamura, 1995). It was followed with the localization of Hrs at the membrane of EE (Komada et al., 1997), and the identification of the signal-transducing adaptor molecule (STAM) as its partner (Asao et al., 1997). Some years later, the implication of the complex in an endosomal protein sorting pathway was shown (Kanazawa et al., 2003), and the discovery of an interaction between Tsg101 and Hrs highlighted the importance of Hrs in receptor downregulation (Lu et al., 2003). Identification of the ESCRT-II and ESCRT-III complexes brought the link between the Hrs-STAM complex and the ESCRT machinery (Babst, 2005) and the name ESCRT-0 was proposed two years later (Nikko & André, 2007).

b. Dissection of Hrs: domains, interactions, regulations

Hrs is a 777 amino-acid protein with a theoretical molecular weight of 86 kDa. Different domains have been identified in this protein (Figure 16).

Figure 16: Schematic representation of Hrs domain (Bache, 2003). VHS (Vps27, Hrs, STAM), UIM (Ubiquitin interaction motif), Pro (Proline-rich domain), CC (Coiled coil), Pro/Gln (Proline/Glutamine-rich region) and CBD (Clathrin-binding domain). The PSAP motif allows the interaction with Tsg101.

The VHS domain and the double UIM (DUIM) (Hirano et al., 2006) allow binding to ubiquitinated cargoes. The FYVE motif is known for its affinity for PI3P and is thought to target Hrs to PI3P-enriched EE (Hayakawa, 2003; Komada et al., 1997). A clathrin binding domain (CBD) is also present at the C-terminal extremity of Hrs, allowing for the formation of clathrin coats when the sorting is initiated. Hrs has two main partners in the ESCRT pathway. It binds to STAM via its coiled coil domain (Asao et al., 1997). This interaction stabilizes STAM and is required for its localization to the EE membrane (Mizuno et al., 2004). Regarding the interaction between Hrs and Tsg101, the domains involved are still debated. Two different groups used a two-hybrid strategy in yeast to define the interaction between these two proteins. According to the first study, the interaction takes place between the UEV domain of Tsg101 and the four amino- acid motif PSAP of Hrs (Lu et al., 2003). The next studies pointed to two other sites: the PTAP motif of Tsg101 would interact with both, separated or not, coiled coil domain and the proline/glutamine acid-rich domain (Bouamr et al., 2007; Pornillos et al., 2003). If the partners of Hrs are well described, the regulation of this key ESCRT component was poorly understood until recently. It needs to be phosphorylated and ubiquitinated to contribute to EGF receptor degradation (Stern et al., 2007). These modifications are dependent on the activity of the ubiquitin-ligase c-Cbl (Katzmann et al., 2001; Visser Smit et al., 2009). Very recently, a study in drosophila showed that Hrs needs to be de-ubiquitinated for the complete sorting of the signaling proteins (Babst, Katzmann, Estepa-Sabal, Meerloo, & Emr, 2002a; Babst, Katzmann, Snyder, Wendland, & Emr, 2002b; Zhang, Du, Lei, Liu, & Zhu, 2014). In the absence of deubiquitinase activity, ubiquitinated Hrs is degraded in lysosomes, instead of the proteasome, and enlarged endosomes containing accumulating signaling proteins are formed. Post-translational modifications have an effect on Hrs localization. Indeed, the paper of Gasparrini and colleagues showed that ubiquitinated Hrs is mainly cytosolic while the phosphorylated one is associated to membranes (Gasparrini et al., 2012; Prag et al., 2007).

57 c. Hrs, a multi-functions protein

Besides regulating the MVB biogenesis by recruiting ESCRT components at the EE, Hrs has been implicated in other functions. Hrs has been shown to inhibit the homotypic fusion of EE by interfering with the endosomal SNARE complex (Raymond et al., 1992; Sun, Yan, Vida, & Bean, 2003). SNARE proteins are essential elements of membrane fusion. By binding to several SNARE proteins, the coiled coil domain of Hrs blocks the homotypic fusion of endosomes. As part of the ESCRT machinery, Hrs participates to the sorting of proteins toward degradation. Nevertheless, several publications showed that the ESCRT-0 protein is probably also implicated in the recycling of membrane proteins to the plasma membrane. Indeed, Hrs is a subunit of the complex CART (cytoskeleton-associated recycling or transport) composed of actinin-4, BERT and myosin V (Hurley & Hanson, 2010; Yan et al., 2005). A large region containing the PSAP motif is responsible for the incorporation of Hrs in the complex. Disruption of the complex impaired transferrin recycling. An other group showed the implication of Hrs in the rapid recycling of adrenergic receptors to the plasma membrane (Hanyaloglu, McCullagh, & Zastrow, 2005; Hurley & Hanson, 2010). Overexpression or depletion of Hrs abrogated the recycling of adrenergic receptors. This phenotype is ESCRT independent since Tsg101 or VPS4 knock down have no effect. The VHS domain of Hrs is required for this activity. Interaction of Hrs with sortin nexin 1 (SNX1) suggested that Hrs might also be implicated in retrograde transport (Chin, Raynor, Wei, Chen, & Li, 2001; McCullough et al., 2013). Indeed, Hrs has been shown to colocalize with the RME-8 retromer component, supporting its implication in the transport from endosomes to the trans-Golgi network (TGN) (Hurley, 2010; Popoff et al., 2009). Last but not least, Hrs has been implicated in endosomal trafficking of cholesterol (X. Du, Kazim, Brown, & Yang, 2012; Hurley, 2010). Hrs depletion blocks cholesterol exit from endosomes, while depletion of other ESCRT proteins has no such effect. This suggests that cholesterol accumulation in endosomes does not result from a blockage in the ESCRT process as a whole. Thus, Hrs is a multifunctional protein, which is implicated in many different trafficking pathways. It holds a central place in the regulation of intracellular traffic.

58 4. Focus on Tsg101: a central protein for the ESCRT system

If Hrs seems to be implicated in a variety of cell trafficking pathways, which do not always include other ESCRT complexes, Tsg101 is a central regulator of the ESCRT system itself. As discussed above, the ESCRT-I complex is required in all ESCRT-driven processes. Moreover, the four constitutive components of ESCRT-I have been shown to be mutually dependent for stability (Bache, 2003; Yan et al., 2005). If Tsg101 amount exceeds that of its partners the Tsg101-associated ligase polyubiquitinates the C-terminal region of Tsg101 which gets degraded by the proteasome (McCullough et al., 2013; McDonald & Martin- Serrano, 2007)

5. ESCRT system, where else?

Until recently, it was believed that only eukaryotes and viruses use the ESCRT system. The discovery of an ESCRT-III-like complex in archaea challenged this view (Babst, Katzmann, Snyder, Wendland, & Emr, 2002b; Ettema & Bernander, 2009). The ESCRT system is indeed conserved among archaea and eukaryotes. A recent paper on archaea division shows, using electron-microscopy, that archeal ESCRT-III proteins form a belt where the division takes place (Dobro et al., 2013; Wollert & Hurley, 2010). The fact that these organisms use ESCRT to divide suggests that the first function of the ESCRT system in evolution was in cell division, and was carried out by ESCRT-III proteins alone. Phylogenic analyzes support this hypothesis, with a later apparition of the ESCRT I-II complexes, followed with ESCRT-0 (Field, Sali, & Rout, 2011; Jouvenet, 2012). Thus, it is quite remarkable that the constituent of the ESCRT machinery have been “invented backwards” (ESCRT-III to -0) relative to the sequence of events that takes place during MVB biogenesis (ESCRT-0 to –III) (Figure 17).

Figure 17: ESCRT along evolution. Adapted from Field and colleagues, 2011 (Field et al., 2011). ESCRT complexes are highlighted in yellow. FECA and LECA are respectively the first and last eukaryotic common ancestors.

Last but not least, although bacteria do not possess ESCRT-like proteins, they also exploit this machinery. A RNAi screen performed in drosophila revealed that ESCRT proteins among others are modulating the Mycobacterium phagosome (Bissig & Gruenberg, 2014; Philips, Porto, Wang, Rubin, & Perrimon, 2008). Further investigation allowed the authors to show that ESCRT is restricting the proliferation of the bacteria inside the host. Depletion of key ESCRT protein such as Tsg101 allows non-pathogenenic Mycobacterium to proliferate. More recently, the lab of Philips demonstrated that Mycobacterium tuberculosis is targeting ESCRT to impair the trafficking of the host via a secreted protein. Authors used a high throughput yeast-two-hybrid platform to screen for interaction between the bacterium secretome and host proteins and identify the mycobacterial protein EsxH as an Hrs interactor. Depletion of Hrs, Tsg101, or Rab7 increase infection, indicating that the bacteria target the ESCRT system to escape lysosomal degradation (McCullough et al., 2013; Mehra et al., 2013). Mycobacterium is so far the only example in which a bacterium interferes with ESCRT-driven processes of its host.

MATERIALS & METHODS

61 I. Chlamydia strains

One clonal population from the strain C. trachomatis L2 (strain 434/Bu – ATCC VR- 902B) was plaque isolated before transformation with SW2::GFP (Y. Wang et al., 2011), or with p2TK2-SW2 IncDProm-RSGFP-IncDTerm (Agaisse & Derré, 2013), as previously described (Y. Wang et al., 2011). Bacteria were propagated as previously described (Boleti et al., 1999). EBs were purified on a density gradient (Scidmore, 2005).

II. Cell culture, transfection and chemicals

HeLa and HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium with Glutamax (DMEM, Invitrogen) supplemented with 10% fetal calf serum (FCS). Cells were transfected with the indicated plasmids 24 h after seeding using JetPrime transfection kit (Polyplus transfection). Depending on the experiment, cells were infected or transfected first. Unless otherwise indicated, a MOI of 1 was used. SiRNA (Dharmacon) were mixed with Lipofectamine RNAi max reagents (Invitrogen) and cells before seeding. The different siRNA sequences used are described in the table 1. Chloramphenicol (stock at 35 mg/mL in ethanol) was purchased from Sigma and stored at -20°C.

Table 1: Sequences of the different siRNA used in this study

!"#$%& !'()'*+'& ,-*./-0& !""!!!#$#!"!$"#$!##% 1/2&345& "##$$$#$#$"!$"$$!!"% 1/2&365& "#""!###$"!$$"!##$#% 728494&345& $$#"!$!!$!$!$"!$$!#% 728494&365&& "##"!#"$$"#""!!"#!#%

62 III. Cloning procedures

The genes coding the five DUF582 proteins (CT619, CT620, CT621, CT711, CT712) were amplified from C. trachomatis D/UW-3/CX genomic DNA by PCR with Phusion high- fidelity DNA polymerase (Finnzyme) according to the manufacturer’s instructions and cloned into a pEGFP-derived destination vector providing a N-terminal green fluorescent protein (GFP) tag using the Gateway technology. The same open reading frames (ORFs) have been similarly cloned into two other vectors with pCIneo as a backbone: one without any tag, and the other one with a Flag tag positioned on the N-terminal part. Truncated versions of those genes have been amplified by PCR and inserted with the same set of vectors. A list of the different constructs obtained along this project is detailed in table 2. To overcome the low expression of some of these bacterial genes in mammalian expression systems we used gene synthesis with codon optimization for ct619, ct621, and ct712 (Genscript).

Table 2: List of primers used to amplify each DUF582 recombinant protein used in this study.

,-*2./)+.& :0;2<"=& 7;8& :/"<'/2& 234% 555556756678866556656585655775866855655656776756656% !&'$()'*+),-% ./01('-% "1(% 9:;% 5655567866687785875558758888755757565587558% .$<=>3?% =3%856?% 234% 55555675667886655665658565577586678878857858885558888566775578% $()'*% .0"2@$'% "2@% 9:;% 5655567866687785885588565685558887767867% .$<=>3?% =3%856?% 234% 55555675667886655665658565577586678878857858885558888566775578% $()'*+A&&% .0"2@$'% "2@% 9:;% 5655567866687785875878557888886558757868588787% .$<=>3?% =3%856?% $()'*3.8% BCD8E:BFG:H%6:D:%I":DB74F.8J% .0"2@$'% "2@% .$<=>3?% =3%856?% 234% 555556756678866556656585655775866688785865575858885585555885578786% $(),K% .0"2@$'% "2@% 9:;% 56555678666878857855775677568888788688% .$<=>3?% =3%856?% 234% 5555567566788665566565856557758665776858875876857575566% $(),'% .0"2@$'% "2@% 9:;% 56555678666878857858788556565885767678558% 234% 56555678666878857858788556565885767678558% +L&L$(),'% .0"2@$'% "2@% 9:;% 5555567566788665566565856557758666888656855855788868888657% .$<=>3?% =3%856?% $(),'3.8% BCD8E:BFG:H%6:D:%I":DB74F.8J% .0"2@$'% "2@% .$<=>3?% =3%856?% 234% 55555675667886655665658565577586675585755778575877588878885% $(M''% .0"2@$'% "2@% 9:;% 5655567866687785885888555878576658755788567% .$<=>3?% =3%856?% 234% 55555675667886655665658565577586665557758776588775% $(M',% .0"2@$'% "2@?% 5655567866687885785678565567755868878% ./01('-% "1(% 9:;%

$(M',3.8% .0"2@$'% "2@% BCD8E:BFG:H%6:D:%I":DB74F.8J%

63 Other constructs were obtained from various sources: myc-Hrs (F. Gesbert, Université Paris Sud, Villejuif, France), all the different Ha-Tsg101 constructs (E.O Freed, NIH Bethesda, USA), VPS4 constructs (J. M. Serrano, King’s College London School of Medecine, London, UK).

IV. Production and purification of recombinant protein

ORFs coding for ct619 and ct712 were amplified by PCR in order to obtain the full gene of ct712 and the region of ct619 truncated of the first 81 and last 625 codons, and were cloned using the Gateway system into the pDEST15 (Invitrogen) destination vector, providing a GST tag at the N-terminus. Expression of the recombinant proteins was made after transformation of BL21 E. coli strain. BL21 bacteria transformed with the GST- !81CT619!625 construct were cultured in LB media supplemented with ampicillin at 37 °C until the optical density reached 0.6 before addition of isopropyl #-D-1-thiogalactopyranoside (IPTG) for expression induction. The GST-CT712 construct has been expressed in BL21 cultured in the same medium in microfermentors at the plateform of protein production at the Institut Pasteur Paris. Protein purification was performed on column using glutathione-sepharose beads (GE Healthcare) following the manufacturer indications. Purity of each fraction has been checked on coomassie gel and the purest samples were sent to AgroBio (La Ferté, France) for immunization of rabbits.

V. Yeast-two-hybrid (Y2H) assays

The Y2H screens were performed by Hybrigenics (Paris, France). Four baits were used: each ORF coding for CT619 and CT621 were separated in two parts: amino-acids 1 to 490 and 480 to 877 for CT619; amino-acids 1 to 441 and 442 to 832 for CT621. For each screen, 7.107 interactions between each bait and a human placenta library were tested. Interactions were further tested using the Matchmaker kit (Clontech) following the manufacturer’s guidelines. The DUF582 protein constructs were cloned into the yeast vector pGBKT7 carrying the GAL4 DNA binding domain while the Hrs constructs were cloned into the pGADT7 vector carrying the GAL4 activation domain. The constructs used in this yeast

64 assay are listed in the table 3. The yeast strain AH109 was co-transformed with both vectors and plated on double dropout medium (DDO; SD/-Leu/-Trp) for 48 h at 30 °C. Single colonies were then cultured over-night at 30 °C in a YPD medium and from this culture a serial dilution of the same number of yeast was plated on three selective media DDO, TDO (SD/-His/-Leu/-Trp) and QDO (SD/-Ade/-His/-Leu/-Trp). Results were analyzed 48 h later.

Table 3: List of primers used to amplify each constructs used in the directed yeast assay.

,-*2./)+.& :0;2<"=& 7;8& :/"<'/2& "#PL%QFDHFD6% 234% 58676658777865865655857568658875555% $()'*$8:4% ."NO(M% H3R5FD% 9:;% 586778675688558856568555888776786785% "#PL%QFDHFD6% 234% 5867665877868788788578588855588885667% $()'*%=8:4% ."NO(M% H3R5FD% 9:;% 5867786756875878557888886558757868588787% "#PL%QFDHFD6% 234% 58676658777865865655857568658875555% $()'*$8:4+'*'% ."NO(M% H3R5FD% 9:;% 586778675688557555585568888787778778% "#PL%QFDHFD6% 234% 58676658777865865655857568658875555% $()'*$8:4+&*% ."NO(M% H3R5FD% 9:;% 586778675688556885878858588858778868758558868% "#PL%QFDHFD6% 234% 586766587786557768588758768575% $(),'=8:4% ."NO(M% H3R5FD% 9:;% 5867786756875755577578855768887865% "#PL%QFDHFD6% 234% 58676658776855855788868888657667% $(),'$8:4% ."NO(M% H3R5FD% 9:;% 586778675678587885565658857676785587% "#PL%QFDHFD6% 234% %5867665877585586757556788788885756588% $(M''$8:4% ."NO(M% H3R5FD% 9:;% %58677867568858885558785766587557885675% "#PL%QFDHFD6% 234% %586766587758675655685675677657857% $(M',% ."NO(M% H3R5FD% 9:;% %586778675678567856556775586887858585758858%

"#PL% 234% 5866587758586666765667567% S91% ."#/(M% 578F;58F3D% H3R5FD% 6587787656875687655865558656786% 9:;% "#PL% 234% 58665877585775576866866565577% +A)&S91+'A&% ."#/(M% 578F;58F3D% H3R5FD% 658778765687586786668575868566757% 9:;%

VI. Immunofluorescence and western blotting analysis

For immunofluorescence (IF), HeLa cells fixed for 30 min at room temperature in PFA 4% were permeabilized for 15 min in PBS 1X, BSA 1 mg/mL, and saponin 0.05% (IF buffer). For anti-Hrs staining, cells were fixed for 30 min on ice in PFA 4%. Coverslips were

65 incubated with primary antibodies in IF buffer for one hour at room temperature, followed with three washes in the same solution. Primary antibodies used for IF were: Anti-Hrs A-5 (Enzo Life Science), anti-Myc (kind gift from A. Israel, Institut Pasteur), and anti-LC3 4E12 (MBL). Secondary antibodies coupled with a fluorophore were incubated for one hour together with 0.5 µg/ml Hoechst 33342 (Molecular Probes) in IF buffer. Coverslips were then mounted in Mowiol buffer and analysed using an Axio observer Z1 microscope equipped with an ApoTome module (Zeiss, Germany) and a 63$ Apochromat lens. Pictures were taken with a Coolsnap HQ camera (Photometrics, Tucson, AZ) using the software Axiovision. For western blot (WB) analysis, cells were lysed in a 1% sodium dodecyl sulfate (SDS), 8 M urea, 150 mM NaCl, and 30 mM Tris-HCl pH 8. Samples normalized to protein content were analyzed by sodium dodecyl-sulfate poly-acrylamide gel electrophoresis (SDS- PAGE). Proteins were transferred on polyvinylidene difluoride (PVDF) membranes and incubated for one hour in 0.1% tween in PBS 1X supplemented with 5% milk. Primary and secondary antibodies were both sequentially incubated in 0.1% tween in PBS 1X on the membranes for one hour separated by washes. Primary antibodies used for WB were: Anti- Hrs (Bethyl laboratories), anti-Tsg101 4A10 (GeneTex), anti-GFP (Santa-Cruz), anti- vimentin (kind gift from P. Cossart, Institut Pasteur), anti-actin (Sigma), and anti-HA-11 (Covance). For qualitative analysis, secondary antibodies are coupled with horse-radish peroxidase and the WB are revealed by chemiluminescence (KODAK). For quantitative analysis, secondary antibodies are coupled with alkaline-phosphatase and revealed with a Typhoon system (GE Healthcare). Quantification of bands intensity was made with the multi- gauge software (Fujifilm). Levels of Hrs and Tsg101 were normalized on actin levels and presented as percentage of control.

VII. Quantification of bacterial entry

Bacteria were stained with a mouse anti-MOMP-LPS (Argene #11-114) antibody followed with Cy5-conjugated secondary antibodies. Bacteria were permeabilized for 15 min in 0.3% Triton X-100 in PBS 1mg/ml BSA prior to DNA staining for 30 min using mounted in a Mowiol solution. Entry experiments were performed on cells seeded the day before on coverslips (40,000 cells/well) in 24-well plates. To disrupt bacterial aggregates, EBs purified on a density gradient were briefly sonicated prior to infection. Cells were incubated at 4 °C for 15

66 min in DMEM 10% FCS before adding the bacteria (MOI=10) for another 30 min at 4 °C. Medium was replaced by medium prewarmed at 37 °C, and plates were transferred to the 37 °C incubator for the indicated times before fixation in ice-cold fixation buffer for 30 min. Extracellular bacteria were stained with a mouse anti-MOMP-LPS (Argene #11-114) antibody followed with Cy5-conjugated secondary antibodies. Pictures of fields with 5-10 cells were acquired using the same microscope described above. A minimum of 10 fields was analyzed per condition. We designed an automatic, ready-to-use analysis protocol for the Icy software (de Chaumont et al., 2012) to perform the quantification on entire image folders without manual intervention (The Chlamentry protocol will be made publicly available on the Icy website upon publication). First, a wavelet-based detection module was used to detect all objects in the green and red channels. Then, an object-based colocalization module was used to visualize and quantify the colocalization between the two detection sets. Two detections were considered colocalized under a distance threshold of 4 pixels (i.e. 400 nm) between their center of mass, accounting for the chromatic aberration of the imaging setup. Finally, the protocol produced a comprehensive result sheet containing the number and location of detected objects in each channel, the number of colocalized detections (i.e. number of extracellular bacteria), and a final script calculated the ratio of [green - colocalized detection] to [green detection] (i.e. ratio of internalized bacteria). Of note, to avoid bacterial aggregates in the inoculums, bacteria were gently sonicated before use. This procedure can lead to the appearance of red-and-not-green dots. These red dots are also not visible in the blue channel (DNA), and presumably correspond to bacterial wall debris. We used conditions where such events represented less than 10% of the total red staining. In addition, these objects are not scored by the software since they are not green, and therefore do not affect the measured efficiency of entry.

VIII. Flow cytometry on infected cells and EBs.

Cells were siRNA treated or not as described above while seeding in 6-well plates (400,000 cells/well) the day before infection. Cells were infected for one hour with L2incDGFP EBs purified on a density gradient at the indicated MOI before changing the medium and returning the plate to the incubator. At the indicated times, cells were washed with PBS and gently detached using 0.5 mM EDTA in PBS. Samples were fixed in PFA 2% in PBS and stored over-night at 4 °C. Flow cytometry analysis was performed with a FACS Gallios

67 (Beckton Coulter) using the FL-1 (detecting fluorescence emission between 505 and 545 nm), the FSC (relative cell size) and the side scatter detectors (cell granulometry or internal complexity) on 1/10 of the sample diluted in PBS. A minimum total of 10,000 gated events were collected for each sample. Data were analyzed using the Kaluza 1.2 software (Beckman Coulter). For analysis of transfected cells, after fixation the cells were centrifuged at 1500 xg, washed in PBS, centrifuged again, and incubated for 1 h with home-made rabbit anti-myc antibodies in 1 mg/ml BSA, 0.05% saponin in PBS. The cells were washed and incubated for one hour in the same buffer with anti-rabbit antibodies conjugated to Cy5. The cells were washed, resuspended in PBS and analyzed by flow cytometry in the FL-1 (green) and FL-4 (far-red) channels.

IX. Immunoprecipitation (IP)

Cells were lysed on ice for 45 min in ice cold 150 mM NaCl, 10 mM NaF, 1 mM EDTA, 1mM EGTA, 0.5 triton % 50 mM Tris pH 7.5 (IP buffer) supplemented with 5 mM PMSF, 1 mM vanadate and 1:100 of proteases inhibitor cocktail (P8340, Sigma). Cells were removed with a scraper, homogenized and centrifuged at 16000 g for 20 min at 4 °C. For anti- myc IP, the supernatant was incubated at 4°C for 1-2 hours with anti-myc 9E10 antibody (Santa-Cruz) followed by 1.5 hours incubation with protein G coupled with sepharose-beads. For anti-HA IP, the supernatant was incubated for 2 hours with the anti-HA antibody coupled with beads (Sigma) at 4 °C. Immunoprecipitates were then washed at 4°C a minimum of 5 times with 1 mL of the IP buffer with gentle centrifugations of maximum 400 g. Beads were then resuspended in a Laemmli buffer supplemented with 1% ß-mercapto-ethanol.

RESULTS

69 I. Article 1: Identification of a family of effectors secreted by the type III secretion system that are conserved in pathogenic chlamydiae

§ § Sandra Muschiol, Gaelle Boncompain , François Vromman , Pierre Dehoux, Staffan

Normark, Birgitta Henriques-Normark, and Agathe Subtil. Inf. & Immun. (2011) 79 571

A screen to identify new chlamydial effectors, using the heterologous T3SS of Shigella flexneri had shown that the C. trachomatis protein CT712 and its orthologs in C. caviae and C. pneumoniae possessed an amino-terminal T3S signal (Subtil et al., 2005). In this study, we focused on a family of proteins that includes CT712. Members of the family share a common domain of about 400 amino acids, the DUF582, which is only found in pathogenic Chlamydia and in no other sequenced organism. Each pathogenic species has 4 to 5 proteins (5 for C. trachomatis), composed of the conserved DUF582 domain in C- terminal and a variable N-terminal domain. We showed, using the same heterologous system of secretion in S. flexneri, that different members of each cluster have a T3S signal at their N- terminal extremity. The DUF582 is predicted to present mainly $-helices. It contains a conserved coiled coil domain in its center. Bioinformatics analysis showed that the conservation of the DUF582 is higher between orthologs than between paralogs for the 4 clusters defined by the variable N-terminal domains. With proteins of the family in C. trachomatis, the identity score between two DUF582 is not higher than 39%. In addition, the N-terminal domains do not show any similarities. CT712 contains only the DUF582, with no additional N-terminal domain. Finally neither the DUF582 nor the variable N-terminal domains show sequence similarities with proteins of other organisms than chlamydiae. We obtained rabbit polyclonal antibodies against three out the five C. trachomatis DUF582 proteins (CT620, CT621 and CT711). We observed that these proteins were expressed at the mid-late phase of infection and that they were also present in purified EBs. Immunofluorescence microscopy demonstrated the secretion of CT620 and CT621 in the lumen of the inclusion and in the host cytoplasm around 30 hpi. Finally, we showed that CT620 and CT711 are detected in the nuclear fraction of infected cells, confirming that they are secreted outside the bacteria.

Identification of a Family of Effectors Secreted by the Type III Secretion System That Are Conserved in Pathogenic Chlamydiaeᰔ Sandra Muschiol,1,2,3,4 Gaelle Boncompain,1,2§ François Vromman,1,2§ Pierre Dehoux,5 Staffan Normark,4 Birgitta Henriques-Normark,3,4 and Agathe Subtil1,2* Institut Pasteur, Unite´de Biologie des Interactions Cellulaires, Paris, France1; CNRS URA 2582, Paris, France2; Swedish Institute for Infectious Disease Control, SE-171 82 Solna, Sweden3; Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, Sweden4; and Institut Pasteur, Plate-Forme Inte´gration et Analyse Geńomique, Paris, France5

Received 29 July 2010/Returned for modiﬁcation 31 August 2010/Accepted 3 November 2010

Chlamydiae are Gram-negative, obligate intracellular pathogens that replicate within a membrane-bounded compartment termed an inclusion. Throughout their development, they actively modify the eukaryotic environment. The type III secretion (TTS) system is the main process by which the bacteria translocate effector proteins into the inclusion membrane and the host cell cytoplasm. Here we describe a family of type III secreted effectors that are present in all pathogenic chlamydiae and absent in the environment-related species. It is deﬁned by a common domain of unknown function, DUF582, that is present in four or ﬁve proteins in each Chlamydiaceae species. We show that the amino-terminal extremity of DUF582 proteins functions as a TTS signal. DUF582 proteins from C. trachomatis CT620, CT621, and CT711 are expressed at the middle and late phases of the infectious cycle. Immunolocalization further revealed that CT620 and CT621 are secreted into the host cell cytoplasm, as well as within the lumen of the inclusion, where they do not associate with bacterial markers. Finally, we show that DUF582 proteins are present in nuclei of infected cells, suggesting that members of the DUF582 family of effector proteins may target nuclear cell functions. The expansion of this family of proteins in pathogenic chlamydiae and their conservation among the different species suggest that they play important roles in the infectious cycle.

Chlamydiae are Gram-negative bacteria that constitute a mydiae remain within a membrane-bounded compartment distinct phylum. Long considered to comprise exclusively the termed an inclusion (23). This localization restricts the inter- family of Chlamydiaceae, which are obligate intracellular actions between the host and the bacteria. However, chlamyd- pathogens of vertebrates, chlamydiae now include a number of iae have acquired the ability to secrete a number of proteins species described as symbionts of free-living amoebae and into the host cell, including the inclusion membrane, presum- other eukaryotic hosts (16). The species that are pathogenic for ably to create an environment favorable for survival and rep- humans include C. trachomatis, an agent of chronic genital and lication (reviewed in references 4 and 33). Most of these pro- ocular infection, and C. pneumoniae, a prevalent cause of res- teins, often called effectors, are secreted by a type III secretion piratory infections that is also possibly involved in atheroscle- (TTS) mechanism, which is also found in many Gram-negative rosis (30). Other species, which infect primarily animals, also pathogenic bacteria (11, 17, 18). Little is known about chla- have zoonotic potential. mydial effectors and how they manipulate host cellular pro- All Chlamydia species share a unique, biphasic developmen- cesses (4, 33). In Chlamydia, efforts to identify effectors and tal cycle, which involves two distinct morphological and func- their functions have been hampered by the absence of tools to tional forms of the bacteria: the extracellular and invasive genetically modify the bacteria and by their obligate intracel- elementary body (EB) and the intracellular and replicative lular lifestyle. The observation that TTS-dependent proteins of reticulate body (RB) (16, 23). Infection starts with the attach- one bacterium can be secreted by a heterologous TTS appa- ment of an EB to a host cell. Upon bacterial internalization, ratus of another bacterium opened the possibility of screening EBs gradually convert into RBs, which divide several times for chlamydial effectors. Using the heterologous TTS systems before differentiating back to the EB form at the end of the of Yersinia, Shigella, and Salmonella, numerous putative chla- cycle. At 2 to 3 days after the initiation of infection, EBs are mydial effector proteins have been identiﬁed (10, 14, 31). TTS released in the extracellular space, ready to initiate a new cycle. is active both during the intracellular multiplication phase of Importantly, throughout their developmental cycle, chla- the cycle, as illustrated by the large family of proteins translocated into the inclusion membrane, the Inc proteins (27, 32), and during the entry step. Detection of TARP (trans- * Corresponding author. Mailing address: Unite´de Biologie des located actin-recruiting phosphoprotein) in the host cyto- Interactions Cellulaires, Institut Pasteur, 25 rue du Dr Roux, 75015 plasm immediately after infection provided the ﬁrst evi- Paris, France. Phone: 33 1 40 61 30 49. Fax: 33 1 40 61 32 38. E-mail: [email protected]. dence that EBs are also capable of performing TTS across § These authors contributed equally to this work. the plasma membrane (7). ᰔ Published ahead of print on 15 November 2010. In a directed screen to identify new chlamydial effector pro-

571 572 MUSCHIOL ET AL. INFECT.IMMUN. teins we previously showed that the C. pneumoniae protein ␤-D-thiogalactoside (IPTG) (Sigma). Bacterial culture pellets, resuspended in CPn0853 and its homologs in C. trachomatis and C. caviae buffer containing 5 mM imidazole, 300 mM sodium chloride, and 20 mM Tris- (CT712 and CCA00914, respectively), possessed an amino- HCl, were lysed with a French press. Lysates were spun to separate soluble and insoluble material. 6ϫHis-tagged fusion proteins were purified from the soluble terminal signal recognized for TTS in Shigella flexneri (31). In fraction by affinity chromatography using Ni-nitrilotriacetic acid (NTA) His Bind this study, we show that these proteins belong to a large family resin (Novagen, EMD Chemicals, Inc., Gibbstown, NJ) according to the manu- of chlamydial proteins that share a domain of unknown func- facturer’s guidelines. Purified proteins were used to immunize New Zealand tion, referred to as DUF582. All pathogenic Chlamydia species White rabbits for production of polyclonal antisera (Agro-Bio, La Ferte´Saint- Aubin, France). sequenced so far possess four or five proteins belonging to the Immunodetection. For immunoblot analyses, samples of HeLa cells grown in family. We provide evidence that members of this family are six-well plates were infected with C. trachomatis L2 for the indicated time, TTS substrates, which translocate into the host cell. washed twice in PBS, and detached in PBS supplemented with 0.5 mM EDTA. Where indicated, 100 ␮g/ml chloramphenicol dissolved in ethanol (34 mg/ml) was added 90 min prior to cell lysis; ethanol alone was added to the control cells. MATERIALS AND METHODS The cells were lysed in 1% sodium dodecyl sulfate (SDS), 6 M urea, 150 mM Cell culture and bacterial culture conditions. HeLa 229 cells (American Type NaCl, and 30 mM Tris-HCl. The protein content of each lysate was quantified Culture Collection) were cultured in Dulbecco’s modified Eagle’s medium with using the bicinchoninic acid (BCA) protein assay kit (Thermo) according to the Glutamax (Invitrogen Life Technologies) supplemented with 10% (vol/vol) fetal manufacturer’s instructions. An equal amount of total protein was loaded on 10 bovine serum (Invitrogen). Chlamydia trachomatis L2 strain 434 (ATCC) was or 8% acrylamide gels, resolved by SDS-PAGE, and transferred to polyvinyli- propagated in HeLa cells as previously described (6). For infection, semiconflu- dene difluoride (PVDF) membranes. After incubation with primary antibodies ent monolayers of HeLa cells were inoculated with Chlamydia at a multiplicity of to CT620, CT621, CT711, major outer membrane protein (MOMP) (mouse infection (MOI) of 0.5 to 1 for1hat37°C. Infected cells were washed with monoclonal antibody MyBioSource no. 310190), or actin (mouse monoclonal phosphate-buffered saline (PBS) and incubated in fresh medium for the indi- antibody clone AC-15; Sigma catalog no. 5441), membranes were probed with cated times. Shigella flexneri mxiD and ipaB strains were grown as described horseradish peroxidase-conjugated secondary antibodies and visualized with previously (32). Escherichia coli strains DH5␣ and BL21 were cultivated at 37°C Amersham ECL Plus (GE Healthcare UK Limited). on Luria-Bertani (LB) plates or in LB liquid cultures. C. trachomatis L2 EBs were purified on density gradients as described previ- Bioinformatics. Protein sequences containing a DUF582 domain were ob- ously (29), lysed in the lysis buffer, and run on SDS-PAGE in parallel with the tained by a HMMER search (9) using the Pfam DUF582 model (12). The whole-cell lysates. analysis was performed on eight Chlamydia genomes (C. trachomatis D/UW-3/ Immunofluorescence assay and microscopy. HeLa cells grown on glass cover- CX, C. muridarum, C. abortus S26/3, C. caviae GPIC, C. felis Fe/C-56, C. pneu- slips were infected with C. trachomatis L2 and fixed with 4% paraformaldehyde moniae CWL029, Protochlamydia amoebophila UWE25, and Waddlia chon- (PFA) at 31 h postinfection (p.i.). Cells were permeabilized in PBS containing drophila), resulting in a set of 26 sequences (no DUF582 domain was detected in 0.05% saponin (Sigma) and 1 mg/ml bovine serum albumin (BSA) for indirect the two environmental chlamydiae). Clustering of DUF582 protein families was immunofluorescence. Proteins were detected with antibodies against CT620 or performed by BLAST analysis after masking regions matching Pfam CT621. Bacteria were labeled with mouse monoclonal anti-Hsp60 (Affinity DUF582HMMs. DUF582 domain sequences were aligned using the multiple- BioReagents) or anti-MOMP (Argene no. 11-111) or with polyclonal rabbit sequence alignment program Promals (24). The plot of the sequence conserva- antibodies to CT260 (a kind gift of R. Valdivia, Duke University). The respective tion of the multiple alignment was obtained by EMBOSS-Plotcon (26) with antigens were visualized with Fluorolink Cy5-labeled goat anti-mouse antibody default parameters. Coiled-coil domains were detected by submitting the multi- (Amersham) and Alexa 488-conjugated goat anti-rabbit antibody (Molecular ple alignment to Pcoil at the MPI Bioinformatics facilities (http://toolkit Probes). DNA was stained with 0.5 ␮g/ml Hoechst 33342 (Molecular Probes) in .tuebingen.mpg.de) (5). the mounting medium. In antibody competition experiments, anti-CT620 or Heterologous TTS assay. The Shigella flexneri-based TTS assay was performed anti-CT621 antibodies were preincubated for 30 min with 4 ␮g purified His- as previously described (32). Chimeras comprising the 5Ј parts of different chla- CT620 or His-CT621 protein, respectively, or the irrelevant C. caviae protein mydial genes upstream of the gene coding for the calmodulin-dependent ade- CCA00037-His (31) in buffer used for immunolabelings. Images were acquired nylate cyclase of Bordetella pertussis were constructed by PCR as described using an ApoTome microscope (Zeiss) equipped with a 63ϫ objective and a previously (32). The constructs include about 30 nucleotides upstream from the Roper Scientific Coolsnap HQ camera, permitting optical sections of 0.7 ␮m. proposed translation start sites and the first 24 to 30 codons of the chlamydial Nuclear isolation. HeLa cells (in two 10-cm dishes) infected for 40 h were washed genes, using the following primers: CT619 (30 codons), AGTCAAGCTTGTAA once with PBS, pelleted, and resuspended in 0.4 ml buffer A (10 mM HEPES [pH TAGTTTTGTTTTTATGTCTTCTTACTATTT and TCTCTAGAAAAATCT 7.9], 10 mM KCl, 0.1 mM EDTA, and 1 mM EGTA plus protease inhibitor cocktail GAGCCAGGATTGG; CT621 (28 codons), AGTCAAGCTTGTAAATATTA [Sigma, P8340]) for 10 min before addition of 0.2% NP-40 and four passages TTGGGATAGGTTCGC and AGTCTCTAGATGGTTTGGCAATCTTCT through a 26-gauge syringe. Nuclei were pelleted at 800 ϫ g for 5 min, while the TTG; CPn0726 (24 codons), GTCAAGCTTGTAACTGTTCTTATCTAAGCA supernatant was saved as the cytoplasmic fraction. Nuclei were washed two times GACATTGA and AGTCTCTAGATGCAAAAGAATGCATTGAAGAC; and with 1 ml buffer A, resuspended in radioimmunoprecipitation assay (RIPA) buffer CPn0852 (26 codons), AGTCAAGCTTCTGATAAATACCGTGACGCTAC (50 mM Tris [pH 7.4], 150 mM NaCl, 2 mM EDTA, 1% NP-40, 0.5% Na-deoxy- and AGTCTCTAGATGACGTATCAATCTTACCACCAG. The chimeric con- cholate, and 0.1% SDS plus protease inhibitor cocktail) for 30 min, and centrifuged structs were transformed in the S. flexneri strains SF401 and SF620, which are at 16,000 ϫ g for 10 min to obtain the nuclear soluble fraction in the supernatant. derivatives of M90T in which the mxiD and ipaB genes, respectively, have been Protein concentrations were determined using the Bradford assay (Bio-Rad protein inactivated (1, 22). Secretion in liquid cultures was assayed as described previ- assay), and equal quantities of the cytosolic and nuclear fractions (5 ␮g) were loaded ously (32). Monoclonal antibodies against the adenylate cyclase (N. Guiso, In- for analysis by Western blotting. In addition to the antibodies already described, stitut Pasteur) were used to detect the chimera, polyclonal antibodies against the antibodies to poly(ADP ribose) polymerase (PARP), EF-Tu, and Rab-GDP disso- Shigella type III effector IpaD (22) were used to control that TTS was not ciation inhibitor ␤ (GDI␤) were used. Monoclonal anti-PARP antibody was pur- impaired by transformation of the various constructs, and antibodies against the chased from Trevigen, monoclonal antibody against chlamydial EF-Tu was a kind cyclic AMP (cAMP) receptor protein (CRP) (A. Ullmann, Institut Pasteur) were gift from Y.-X. Zhang (Boston), and polyclonal rabbit antibody against GDI␤ was a used to control for bacterial lysis during fractionation. kind gift from B. Goud (Institut Curie, France). Cloning, production of recombinant protein, and antibody production. The Ectopically expressed GFP fusion proteins. The open reading frames coding open reading frames coding for the hypothetical proteins CT711, CT620, and for the hypothetical proteins CT711, CT620, and CT621 from the C. trachomatis CT621 from the C. trachomatis genome were amplified from C. trachomatis L2 genome were amplified from C. trachomatis L2 DNA by PCR with Phusion DNA by PCR with Phusion high-fidelity DNA polymerase (Finnzyme, Espoo, high-fidelity DNA polymerase (Finnzyme, Espoo, Finland) according to the Finland) according to the manufacturer’s instructions and cloned into pET manufacturer’s instructions and cloned into a plasmid pEGFP-derived destina- expression vector pET28 or pET30, providing an N-terminal six-histidine tag tion vector providing a N-terminal green fluorescent protein (GFP) tag using the (Novagen, EMD Chemicals, Inc., Gibbstown, NJ). Gateway technology. HeLa cells were transiently transfected with these plasmids For protein expression, constructs were transformed into E. coli BL21. The C. for 24 h using Fugene reagent (Roche), and the cells were fixed, briefly perme- trachomatis proteins were expressed as 6ϫHis-tagged N-terminal fusion proteins, abilized with 0.05% saponin, mounted in Mowiol supplemented with 0.5 ␮g/ml and expression was induced in logarithmically growing cultures with isopropyl- Hoechst, and observed with an ApoTome microscope as described above. VOL. 79, 2011 DUF582 PROTEINS ARE TTS EFFECTORS IN CHLAMYDIACEAE 573

TABLE 1. Proteins with DUF582 domains in different Chlamydia species

Chlamydia Protein (aa) in cluster: species 1234 C. trachomatis CT712 (390) CT711 (767) CT620 (838), CT621 (832) CT619 (877) C. pneumoniae CPn0853 (389) CPn0852 (766) CPn0726 (831) CPn0727 (872) C. caviae CCA00914 (388) CCA00915 (747) CCA00017 (838) CCA00016 (870) C. muridarum TC0085 (390) TC0084 (769) TC0910 (828), TC0911 (831) TC0909 (875) C. felis CF0100 (388) CF0099 (746) CF0989 (849) CF0990 (864) C. abortus CAB882 (388) CAB833 (746) CAB017 (885) CAB016 (866)

RESULTS DUF582 proteins, with the exceptions of C. trachomatis and C. muridarum, which have five DUF582 proteins (Table 1). Mul- Identification of the chlamydial DUF582 family of proteins. tiple-alignment analysis of all DUF582 domains showed that We have previously reported that the C. trachomatis protein they have 17 to 88% sequence identity. Most DUF582 domains CT712 and its homologs in C. pneumoniae and C. caviae are putative TTS substrates because they were secreted by the are associated with a second domain located in the N-terminal heterologous secretion system of S. flexneri (31). CT712 is parts of the proteins. These domains were analyzed for their predicted to encode a 390-amino-acid protein, consisting es- sequence similarity relationships and found to define four dif- sentially of a domain of unknown function, annotated as ferent clusters, with each cluster containing one (or two, for DUF582 in the PFAM database. Bioinformatics analysis CT620/CT621 and TC0910/TC0911) proteins from each ge- showed that this domain is also found at the carboxyl-terminal nome (Table 1 and Fig. 1A and B). Interestingly, DUF582 extremities of four other C. trachomatis proteins, CT711, domains are more conserved between orthologs (i.e., within CT619, CT620, and CT621. Moreover, DUF582 proteins are the same cluster) than between paralogs (i.e., within the same present in all pathogenic chlamydiae sequenced so far, and species). For example, DUF582 proteins of cluster 1 show 45 only in these bacteria. Each genome codes for four different to 88% identity, while identity between DUF582 domains

FIG. 1. Conservation of the DUF582 proteins in Chlamydiaceae. (A) Schematic overview of the genes encoding DUF582-containing proteins in the C. trachomatis, C. pneumoniae, and C. caviae genomes. (B) Schematic representation of the Chlamydia DUF582 proteins. The proteins are grouped in four clusters based on sequence similarity. In all of them, DUF582 is located at the carboxy terminus. Proteins belonging to clusters 2, 3, and 4 contain additional N-terminal domains. The gray-level code corresponds to the DUF582 genes depicted in panel A. (C) Alignment of cluster 1 DUF582 proteins in six chlamydial genomes: C. trachomatis (CT), C. pneumoniae (CPn), C. muridarum (TC), C. caviae (CCA), C. felis (CF), and C. abortus (CAB). Identical residues are shown in red. Amino acids with high similarity are highlighted in green. Asterisks delimit the segment predicted to adopt a coiled-coil conformation. (D) Sequence conservation in the multiple alignment of all DUF582 sequences. 574 MUSCHIOL ET AL. INFECT.IMMUN.

were all detected in the supernatant of the transformed ipaB strain, indicating that they were secreted by the bacteria. Im- portantly, CRP, a cytosolic Shigella protein, was detected only in the pellet fractions, showing that detection of the chimera in the supernatant did not result from bacterial lysis. An endogenous TTS substrate of Shigella, IpaD, was also detected in the supernatant, indicating that expression of the chimera did not prevent TTS of Shigella effectors. In contrast, the chimeras FIG. 2. DUF582 proteins CT621, Cpn0852, CPn0726, and CT619 were not detected in the supernatant of transformed mxiD have TTS signals. The amino-terminal segments of the indicated pro- cultures, which are deficient for type III secretion. This result teins (24 to 30 amino acids, depending on the constructs) were fused to implies that secretion of the four chimeras observed in the ipaB the Cya reporter protein and expressed in a Shigella flexneri ipaD (constitutive TTS) or mxiD (defective TTS) strain. Exponential-phase background was type III dependent. In conclusion, this exper- cultures expressing the reporter fusion protein were fractionated into iment demonstrates that the amino-terminal regions of all the supernatants (S) and pellets (P). Samples were resolved by SDS- DUF582 family members tested contain TTS signals recog- PAGE, transferred to a PVDF membrane, and probed with anti-Cya nized by S. flexneri. This result strongly supports the hypothesis (to detect chlamydial fusion proteins), anti-IpaD (Shigella secreted protein), or anti-CRP (Shigella nonsecreted protein) antibodies. All that all the members of the DUF582 family are Chlamydia TTS results shown are representative of at least two separate experiments. effectors. C. trachomatis DUF582 proteins are expressed in the middle and late phases of the infectious cycle. For the remainder of our within the C. trachomatis species ranges between 18% (CT711- study, we concentrated on the C. trachomatis DUF582 pro- CT712) and 39% (CT620-CT621). teins: CT712 (cluster 1; 390 amino acids [aa]; expected molec- Within one cluster, the amino-terminal domains show be- ular mass, 44 kDa), CT711 (cluster 2; 767 aa; expected molec- tween 25 and 75% identity within their primary sequence, ular mass, 86 kDa), CT620 (cluster 3; 838 aa; expected while no similarity was found when the amino-terminal do- molecular mass, 93kDa), CT621 (cluster 3; 832 aa; expected mains of proteins from different clusters were compared. Da- molecular mass, 93 kDa), and CT619 (cluster 4; 877 aa; ex- tabase iterative searches using the N-terminal multiple align- pected molecular mass, 97 kDa). ments retrieved no homolog sequences, meaning that these All proteins were expressed as His-tagged recombinant pro- N-terminal domains are also Chlamydia specific. teins in E. coli. CT620, CT621, and CT711 were purified by Finally, DUF582 proteins were analyzed with structure pre- affinity chromatography and used to immunize rabbits to ob- diction tools. DUF582 is predicted to be mainly ␣-helical (80% tain antiserum. CT619 was expressed at a very low level, and as estimated by the Gor IV secondary-structure prediction CT712 was insoluble; both were excluded from further analy- method [13] and confirmed in the Promals alignment) and to sis. The antisera obtained for CT620, CT621, and CT711 were contain a segment adopting a coiled-coil conformation (Fig. tested for their specificity on HeLa cell lysates infected or not 1C). When all DUF582 sequences are compared, the coiled- for 30 h with C. trachomatis L2 (Fig. 3A). Antibodies against coil domain and the carboxy-terminal part stand as the best CT621 reacted with a one protein product of the expected conserved segments of this domain of unknown function (Fig. molecular mass which was present only in the infected samples. 1D). The amino-terminal parts of the proteins of clusters 2 to In infected-cell lysates, antibodies against CT620 reacted 4 are also predicted to be rich in alpha helices, possibly adopt- mainly with a 83-kDa product and only faintly with a 93-kDa ing coiled-coil structures. product, which is the expected molecular mass of this protein. DUF582-containing proteins are TTS substrates. We have In addition, the serum reacted with a higher-molecular-mass previously described a secretion assay in Shigella that allowed product that was also present in noninfected cells and which us to identify potential chlamydial TTS substrates (31). Three we therefore considered nonspecific. Strikingly, the same proteins of cluster 1 of the DUF582 protein family were shown pattern of migration was observed with antibodies against to possess TTS signals (CT712, CPn0853, and CCA00914). To CT711: the full-length protein was hardly visible in total cell investigate whether other members of the family possessed extracts, and the antibodies reacted mainly with a lower- TTS signals, we tested proteins representative of each cluster molecular-mass product that could represent a truncation of in the Shigella secretion assay. Because the CT620/Cya and 10 kDa (Fig. 3A). CT711/Cya chimeras were not well expressed (data not This similarity in patterns, using two different antibodies, shown), we constructed chimeras with the C. pneumoniae gene suggests that the two proteins undergo similar processing. We of the same cluster. Fusion constructs were made between the suspected that the small amount of full-length CT620 or N-terminal part of the respective chlamydial gene and a re- CT711 detected by Western blotting was due to a rapid pro- porter molecule, the calmodulin-dependent adenylate cyclase cessing into the truncated forms. To test this hypothesis, we of Bordetella pertussis (Cya). Each construct was transformed blocked bacterial protein synthesis by incubating infected cells into the TTS-competent Shigella ipaB strain and into the mxiD for 90 min in 100 ␮g/ml chloramphenicol before cell lysis. strain, which is deficient for TTS (1, 22). Shigella organisms Following this treatment, the amounts of full-length CT620 expressing the chlamydial constructs were grown to exponen- and CT711 detected in cell lysates strongly decreased com- tial phase and harvested. Cultures were fractionated into cell pared to those in lysates of untreated cells, demonstrating that culture supernatant and cell pellet and analyzed by Western these two forms have a short half-life (Fig. 3B). In comparison, blotting (Fig. 2). CPn0852/Cya (cluster 2), CT621/Cya and inhibition of protein synthesis for 90 min only slightly de- CPn0726/Cya (both from cluster 3), and CT619/Cya (cluster 4) creased the amounts of CT621 and of the truncated forms of VOL. 79, 2011 DUF582 PROTEINS ARE TTS EFFECTORS IN CHLAMYDIACEAE 575

FIG. 3. Expression of the C. trachomatis DUF582 proteins CT711, CT620, and CT621 during chlamydial infection. (A) Polyclonal antiserum obtained after immunization of rabbits with recombinant CT711, CT620, and CT621 was tested for its speciﬁcity on whole-cell lysates. HeLa cells were mock infected (Ϫ) or infected with C. trachomatis L2 (ϩ) for 30 h. Total cell lysates were resolved in polyacrylamide gels and probed by immunoblotting with antibodies to CT711, CT620, and CT621. (B) HeLa cells were infected or not with C. trachomatis L2 for 40 h at 37°C, prior to addition of 100 ␮g/ml chloramphenicol (Cm) in the indicated sample. Ninety minutes later, the cells were collected and whole-cell lysates were probed with the indicated antibodies. Density gradient-puriﬁed EBs were lysed and run in parallel with the whole-cell lysates. Arrowheads point to the expected molecular size for each DUF582 protein. (C) HeLa cell lysates infected with C. trachomatis L2 for the indicated time were run on SDS-PAGE and transferred to membranes. Membranes were probed with antibodies to CT711, CT621, CT620, and MOMP. Antibodies against actin were used as a loading control.

CT620 and CT711, showing that these products are stable. This below the detection threshold with this technique, and was difference in stability between the full-length and truncated therefore excluded from our immunolocalization studies. forms can account for their respective abundances in total cell HeLa cell monolayers were infected with C. trachomatis for lysates. Importantly, bacterial lysates made from EBs purified 31 h, fixed, and labeled for CT620 or CT621. Both proteins on a density gradient showed the same double-band migration were observed in association with the inclusion and with the profile for CT620 and CT711 as previously observed on whole cytoplasm of infected HeLa cells (Fig. 4A). To control for the infected-cell lysates, and in the same proportion of full-length specificity of these stainings, we verified that they disappeared versus truncated protein products, suggesting that processing when the sera were used in the presence of an excess of the occurs in the bacteria (Fig. 3B). However, it is also possible purified proteins against which they were raised but not in the that the shorter forms of CT620 and CT711 result from spu- presence of an irrelevant His-tagged purified protein, rious postlysis degradation of these proteins rather than spe- CCA00037-His (Fig. 4A). When we performed a time course cific processing. of infection, the earliest detectable signal was obtained at 18 to Although these three proteins share a DUF582 domain, we 20 h p.i., and at these early time points the labeling was re- did not expect the sera to cross-react with the different mem- stricted to the inclusion (data not shown). Cells with cytoplas- bers of the family because the domain is not very conserved mic staining for CT620 and CT621 were frequently observed at between paralogs. Indeed, we did not observe cross-reaction 31 h p.i., and by later time points, these proteins were found in when testing our sera against the different purified DUF582 the cytoplasm of almost all infected cells. proteins (data not shown). Altogether, we concluded that our In addition to their cytosolic localization, CT620 and CT621 sera were specific for the proteins against which they were were detected in the inclusion. The inclusion staining for raised and could be used for detection by immunoblotting. CT620 and CT621 did not fully overlap with the bacterial Next, we analyzed when the C. trachomatis DUF582 con- marker Hsp60 and was found in areas devoid of bacteria (Fig. taining proteins are expressed during the course of infection. 4A). In comparison, rabbit serum against the bacterial chap- We found that all three proteins were expressed by 24 h postin- erone CT260 showed a perfect colocalization with Hsp60 (Fig. fection, with a sharp increase between the 18-h and 24-h time 4B). Similar results were observed after methanol fixation points, which corresponds to an increase in bacterial load ob- (data not shown), indicating that it is not an artifact of fixation. served with the antibody against MOMP (Fig. 3C). By microar- As an additional control, we used anti-MOMP antibodies to ray analysis, transcription of the genes coding for C. trachoma- stain the outer membrane of bacteria. Again, a large propor- tis DUF582 proteins was first detected at 8 h postinfection tion of the luminal signal for CT620 or CT621 did not colo- (p.i.), when RB-to-EB conversion is completed (3). These data calize with MOMP (Fig. 4C). These observations indicate that are consistent with our finding that the DUF582 proteins are in addition to being secreted in the host cytoplasm, a portion of detected coincidently with an increase in bacterial number and CT620 and CT621 is secreted from the bacteria within the remain present throughout the remainder of the cycle. lumen of the inclusion. Localization of C. trachomatis DUF582-containing proteins DUF582 proteins are detected in the nuclei of infected cells. in the host cell cytoplasm and in the lumen of the chlamydial During the course of this study, CT621 was reported to be inclusion. We next analyzed the expression and localization of detected in the nuclei of infected cells (15). Our anti-CT620 DUF582-containing proteins by indirect immunofluorescence and anti-CT621 sera also detected proteins within the nuclei of microscopy. Our antibody for CT711 failed to detect a specific infected cells (Fig. 5A). Importantly, for both markers, nuclear signal, probably because the protein is expressed at a level staining was observed in about 80% (n ϭ 20 for each antibody) 576 MUSCHIOL ET AL. INFECT.IMMUN.

FIG. 4. The C. trachomatis DUF582 proteins CT620 and CT621 are detected in the host cytoplasm during chlamydial infection. HeLa cells were infected with C. trachomatis L2 and ﬁxed 31 h later. (A) C. trachomatis DUF582 proteins were immunolabeled with anti-CT620 and anti-CT621 as indicated, and bacteria were detected with a mouse anti-Hsp60 antibody. Coverslips were subsequently incubated with Alexa 488-conjugated anti-rabbit antibody to detect the DUF582 proteins (green) and with Cy5-conjugated anti-mouse antibody for Hsp60 (red). Host cell nuclei and bacterial DNA were stained with Hoechst stain (blue). For antibody competition, anti-CT620 and anti-CT621 were incubated in the presence of an excess of His-CT620, His-CT621, or CCA00037-His (an irrelevant His-tagged protein), as indicated. Arrows point to cells in which cytoplasmic staining for CT620 and CT621 is observed. Note that in addition to the cytoplasmic staining, CT620 and CT621 also localize to the lumen of the chlamydial inclusion with only little colocalization with the bacterial protein Hsp60. (B) Infected cells were labeled with antibodies to the chlamydial chaperones CT260 (green) and Hsp60 (red), which colocalize. (C) C. trachomatis DUF582 proteins were immunolabeled with anti-CT620 and anti-CT621 as indicated, and bacteria were detected with a mouse anti-MOMP antibody. Coverslips were subsequently incubated with Alexa 488-conjugated anti-rabbit antibody to detect the DUF582 proteins (green) and with Cy5-conjugated anti-mouse antibody for MOMP (red). Host cell nuclei and bacterial DNA were stained with Hoechst stain (blue). Images were acquired using an ApoTome ﬂuorescence microscope. Bars, 5 ␮m.

of the infected cells in which the proteins were also observed in nuclear fractions with dense bacteria. In addition to a cytoplas- the cytoplasm. In noninfected cells, or in cells in which the mic distribution, CT620 and CT711 were found to be enriched DUF582 proteins were observed only in the inclusion, no nu- in the nuclear fraction of infected cells compared to the three clear staining was visible with either anti-CT620 or anti-CT621 bacterial markers (Fig. 5B). Interestingly, although there was antibodies, indicating that the staining is specific. To confirm less full-length than truncated CT620 in the nucleus, it was still the nuclear localization of DUF582 proteins by another ap- significantly enriched in this fraction relative to its total expres- proach, we subjected infected cells to a subcellular fraction- sion level. The low level of expression of full-length CT711 ation protocol to separate the cytoplasmic material (including does not allow us to ascertain its presence in the nuclear the bacteria) from the nuclei. To control for the purity of our fraction, while the truncated form was clearly visible. In con- fractions, we probed the membrane for GDI␤ (found only in trast, CT621 was not clearly enriched in the nuclear fraction the cytoplasmic fractions) and PARP (found only in the nu- compared to other nonsecreted chlamydial proteins. This re- clear fractions). In addition, antibodies against three different sult is in slight contradiction with the detection of CT621 in the nonsecreted chlamydial proteins (Hsp60, EF-Tu, and MOMP) nuclei of infected cells by microscopy. It could be accounted were used to determine the level of contamination of the for by the observation that although CT621 is more abundant VOL. 79, 2011 DUF582 PROTEINS ARE TTS EFFECTORS IN CHLAMYDIACEAE 577

FIG. 5. DUF582 proteins are detected in the host nucleus during infection. (A) HeLa cells were infected with C. trachomatis L2 and fixed 31 h later. C. trachomatis DUF582 proteins were immunolabeled with anti-CT620 and anti-CT621 as indicated, and bacteria were detected with a mouse anti-MOMP antibody. Coverslips were subsequently incubated with Alexa 488-conjugated anti-rabbit antibody to detect the DUF582 proteins (green) and with Cy5-conjugated anti-mouse antibody for MOMP (red). Host cell nuclei and bacterial DNA were stained with Hoechst stain (blue). Both CT620 and CT621 were observed in the cell nuclei. (B) HeLa cells were infected with C. trachomatis for 40 h. Nuclei were isolated as described in Materials and Methods, and both cytosolic and nuclear fractions were analyzed by Western blotting. The quality of the fractionation was controlled using GDI␤ as a cytosolic marker and PARP as a nuclear marker. In addition, three nonsecreted chlamydial proteins, EF-Tu, Hsp60, and MOMP, were used to assess the degree of contamination of the nuclear fraction with bacteria. Note that the proportion of CT621 detected in the nuclear fraction is similar to what is observed for the nonsecreted proteins. In contrast, both CT620 and CT711 (arrowheads) are enriched in the nuclear fractions, above the contamination level. Blots are representative of three separate experiments. (C) HeLa cells were transfected for 24 h with the indicated GFP fusion proteins before fixation. Nuclei were stained with Hoechst stain (blue). While GFP-CT620 and GFP-CT711 are observed in the nuclei of all transfected cells, nuclear staining was not detected in all GFP-CT621-transfected cells. Bars, 5 ␮m. than CT620 and CT711 (as assessed by immunofluorescence teins are present in the nuclear fractions, they distribute pri- and Western blotting), the nuclear staining for CT621 was not marily in the cytoplasmic fractions, in agreement with our stronger than that for CT620 (Fig. 5A). Therefore, the pro- microscopy observations. portion of nuclear CT621 is probably lower than that of CT620 Finally, to determine whether DUF582 proteins were able to and just below the detection level by the subcellular fraction- translocate to the cell nucleus even in the absence of infection, ation technique. It is important to note that equal protein we expressed N-terminal GFP fusion proteins by transfection amounts were loaded in each lane. This represents about a in HeLa cells. GFP-CT620, GFP-CT621, and GFP-CT711 5-fold concentration of the nuclear fractions relative to the were all detected at least to some extent in the nuclei of cytoplasmic fractions. Therefore, even though DUF582 pro- transfected cells, with differences between constructs (Fig. 5C). 578 MUSCHIOL ET AL. INFECT.IMMUN.

Nuclear staining was easily detected for GFP-CT620, while it tionation experiments analyzed by Western blotting, CT711 was visible in only some of the GFP-CT621-transfected cells. was observed in the nuclear fraction of infected cells, bringing GFP-CT711 was very poorly expressed and could be seen in evidence that this protein is also secreted out of the inclusion the nuclei of transfected cells. Altogether, GFP fusion proteins during infection. CT620 was also detected in the nuclear frac- reproduced what was observed in infection in terms of level of tion and observed in the nucleus by microscopy techniques. expression (with GFP-CT621 being better expressed than While we were able to confirm earlier CT621 detection in the GFP-CT620 and GFP-CT711 being very poorly expressed) and nuclei of infected cells (15) by microscopy, only a small pro- in terms of nuclear localization (with GFP-CT620 and GFP- portion of the total pool of CT621 is nuclear, making it difficult CT711 being more enriched in the nucleus than GFP-CT621). to detect over bacterial contamination by subcellular fractionation techniques. DISCUSSION The detection of all three DUF582 proteins tested in the nuclei of infected cells is intriguing. These proteins are too Chlamydiae, like other Gram-negative pathogens, use a TTS large to diffuse passively into the nucleus, suggesting that active system to translocate effector proteins into their host to mod- import occurs, although no obvious nuclear localization signal ulate cellular functions. In this study, we have identified a new was identified in their sequence. Interestingly the proteins family of chlamydial effector proteins, called the DUF582 pro- were also observed in the nucleus when ectopically expressed teins, that are secreted by a TTS mechanism during infection. by transfection, indicating that other bacterial factors are not This conclusion is based on (i) the presence of a TTS signal in needed for their translocation in the nucleus in infection. The the seven DUF582 proteins from three different species that only other nuclear effector of Chlamydia identified so far, we tested (this report and reference 31), (ii) the demonstration NUE, is found predominantly in the nucleus (25). This is not of the cytoplasmic localization of two DUF582 proteins in the case for the DUF582 proteins, which are more abundant in infected cells, and (iii) the detection of three DUF582 proteins the cytosol than in the nucleus and might shuttle between these in the nuclei of infected cells. The family consists of 26 mem- locations. NUE has histone methyltransferase activity, suggest- bers in the six annotated genomes that we analyzed. ing that chlamydial proteins may target host gene expression We have previously shown that the DUF582 proteins (25). Whether DUF582 proteins function in the cytoplasm CPn0853, CT712, and CCA00914 are recognized by the heter- and/or in the nucleus during infection requires further inves- ologous TTS system of Shigella flexneri (31). These proteins tigation. share a domain of unknown function, DUF582, and we hy- In addition to their cytoplasmic localization, CT620 and pothesized that other DUF582 proteins might be recognized CT621 were also observed in the lumen of the inclusion. Pro- for secretion by the TTS system. Indeed, we showed here that teins stored in the bacteria prior to secretion do not account four other members of the family possessed a TTS signal rec- for the totality of the luminal staining, because it does not fully ognized by the Shigella TTS system. Altogether, we now have overlap with the distribution of two bacterial markers, Hsp60 demonstrated the presence of a TTS signal in at least one and MOMP. More likely, most of the luminal signal corre- member of each of the four different DUF582 subfamilies that sponds to proteins secreted out of the bacteria inside the in- can be distinguished based on sequence analysis, suggesting clusion lumen. This observation suggests that TTS can occur that the whole family serves as substrates for TTS. So far, no independently of the formation of a translocation pore across well-conserved sequence motif has been identified in TTS sub- the inclusion membrane. In Shigella, it has been reported that strates. However, two independent research groups presented there is always 4 to 5% secretion into the extracellular medium a sequence-based computational approach to predict TTS sig- even in the absence of host cells (21). In addition, it now nals (2, 28). Both groups applied this approach to chlamydial appears that several membranous compartments from the host genomes and identified a number of putative TTS effectors. might make their way to the inside of the inclusion (reference Interestingly, the C. trachomatis DUF582 proteins CT620 and 8 and our own observations), which could provide the neces- CT711 were among the top 10% of predicted chlamydial ef- sary trigger for TTS within the inclusion. The luminal staining fectors by one of the predictors (28). However, CT712, CT619, pattern for CT620 and CT621 appears to be more punctate and CT621 were not predicted to be TTS substrates, showing than the diffuse staining pattern we observed in the cytosol. the limitations of this approach. Of the seven DUF582 proteins One could speculate that once DUF582 proteins reach the for which we demonstrated the presence of a TTS signal, the lumen of the inclusion they tend to aggregate because they are second predictor (2) successfully classified CCA00914, not in the right cytosolic environment. This would be consis- CPn0726, CPn0852, and CT621 as potential TTS substrates but tent with Shigella proteins that also aggregate if they are se- not CPn0853, CT619, or CT712, indicating again that the ex- creted by “leakage” into the culture medium (21). Aggregation perimental approach using heterologous secretion remains to would concentrate the signal and might explain why when date the best method to identify potential TTS effectors. looking at CT620 and CT621 distribution over time, we always Most soluble TTS effectors studied so far are produced in observed the luminal staining, while cytoplasmic staining was small amounts, hampering their detection by immunofluores- visible only at later times of infection. Whether luminal cence. CT621 and, to a lesser extent, CT620 were abundantly DUF582 proteins have activity or represent leakage remains to expressed, allowing for detection by Western blotting and im- be determined. Two other chlamydial proteins of unknown munofluorescence. Both proteins were observed in the cyto- function, Pls1 and Pls2, were shown to localize to globular plasm of infected cells, demonstrating that they are secreted structures within the inclusion lumen and at the inclusion during infection. CT711 expression was lower, and the protein membrane (19). was not detected by immunofluorescence. In subcellular frac- CT620 and CT711 share an intriguing migration profile. We VOL. 79, 2011 DUF582 PROTEINS ARE TTS EFFECTORS IN CHLAMYDIACEAE 579 have shown that both proteins are present in whole-cell lysates REFERENCES as a high-molecular-mass product, which corresponds to the 1. Allaoui, A., P. J. Sansonetti, and C. Parsot. 1993. MxiD, an outer membrane expected size of the full-length protein, and a truncated prod- protein necessary for the secretion of the Shigella flexneri lpa invasins. Mol. Microbiol. 7:59–68. uct with a molecular mass 10 kDa lower. The high-molecular- 2. Arnold, R., S. Brandmaier, F. Kleine, P. Tischler, E. Heinz, S. Behrens, A. mass product is unstable, disappearing within 90 min of Niinikoski, H. W. Mewes, M. Horn, and T. Rattei. 2009. Sequence-based prediction of type III secreted proteins. PLoS Pathog. 5:e1000376. inhibition of protein synthesis, while the truncated product 3. Belland, R. J., G. M. Zhong, D. D. Crane, D. Hogan, D. Sturdevant, J. is stable. Both forms were detected in density gradient- Sharma, W. L. Beatty, and H. D. Caldwell. 2003. Genomic transcriptional purified EBs in the same proportions as in the whole-cell profiling of the developmental cycle of Chlamydia trachomatis. Proc. Natl. Acad. Sci. U. S. A. 100:8478–8483. lysates, suggesting that cleavage occurs before translocation 4. Betts, H. J., K. Wolf, and K. A. Fields. 2009. Effector protein modulation of out of the bacteria. These observations suggest that the host cells: examples in the Chlamydia spp. arsenal. Curr. Opin. Microbiol. 12:81–87. full-length CT620 and CT711 are rapidly processed into 5. Biegert, A., C. Mayer, M. Remmert, J. Soding, and A. N. Lupas. 2006. The truncated forms. However, we cannot at this stage rule out MPI Bioinformatics Toolkit for protein sequence analysis. Nucleic Acids the hypothesis that the shorter forms result from postlysis Res. 34:W335–W339. 6. Boleti, H., A. Benmerah, D. Ojcius, N. Cerf-Bensussan, and A. Dautry- degradation of the proteins. Varsat. 1999. Chlamydia infection of epithelial cells expressing dynamin and DUF582 proteins are present in all pathogenic chlamyd- Eps15 mutants: clathrin-independent entry into cells and dynamin-dependent productive growth. J. Cell Sci. 112:1487–1496. iae sequenced so far and are absent from the two environ- 7. Clifton, D. R., K. A. Fields, N. A. Grieshaber, C. A. Dooley, E. R. Fischer, mental chlamydiae sequenced, Protochlamydia amoebophila D. J. Mead, R. A. Carabeo, and T. Hackstadt. 2004. A chlamydial type III Waddlia chondrophila translocated protein is tyrosine-phosphorylated at the site of entry and as- and (this study and reference 15). sociated with recruitment of actin. Proc. Natl. Acad. Sci. U. S. A. 101:10166– Their conservation in pathogenic bacteria, together with our 10171. finding that they probably all represent TTS effectors, sug- 8. Cocchiaro, J. L., Y. Kumar, E. R. Fischer, T. Hackstadt, and R. H. Valdivia. 2008. Cytoplasmic lipid droplets are translocated into the lumen of the gests that they may likely play an important role in patho- Chlamydia trachomatis parasitophorous vacuole. Proc. Natl. Acad. Sci. genesis. CT620, CT621, and CT711 were first detected only U. S. A. 105:9379–9384. about 18 to 24 h after infection, which corresponds to the 9. Eddy, S. R. 1998. Profile hidden Markov models. Bioinformatics 14:755–763. 10. Fields, K. A., and T. Hackstadt. 2000. Evidence for the secretion of Chla- middle to late phase of the C. trachomatis serovar L2 devel- mydia trachomatis CopN by a type III secretion mechanism. Mol. Microbiol. opmental cycle. More broadly, in transcriptomic analyses, 38:1048–1060. 11. Fields, K. A., D. J. Mead, C. A. Dooley, and T. Hackstadt. 2003. Chlamydia all genes coding for DUF582 proteins came up as midcycle/ trachomatis type III secretion: evidence for a functional apparatus during tardy genes (3, 20), suggesting that our results can be ex- early-cycle development. Mol. Microbiol. 48:671–683. 12. Finn, R. D., J. Tate, J. Mistry, P. C. Coggill, S. J. Sammut, H. R. Hotz, G. trapolated to all DUF582 proteins. Sequence analysis failed Ceric, K. Forslund, S. R. Eddy, E. L. Sonnhammer, and A. Bateman. 2008. to provide clues to the potential function of these proteins, The Pfam protein families database. Nucleic Acids Res. 36:D281–D288. which show no similarity with known proteins. Importantly, 13. Garnier, J., J. F. Gibrat, and B. Robson. 1996. GOR method for predicting protein secondary structure from amino acid sequence. Methods Enzymol. while they share the DUF582 domain, they diverge largely 266:540–553. regarding the N-terminal part. This domain is absent from 14. Ho, T. D., and M. N. Starnbach. 2005. The Salmonella enterica serovar Typhimurium-encoded type III secretion systems can translocate Chla- members of cluster 1, or very different in sequence in mem- mydia trachomatis proteins into the cytosol of host cells. Infect. Immun. bers from clusters 2, 3, and 4, suggesting that each cluster 73:905–911. might represent addition of a different module to the same 15. Hobolt-Pedersen, A.-S., G. Christiansen, E. Timmerman, K. Gevaert, and S. Birkelund. 2009. Identification of Chlamydia trachomatis CT621, a protein functional DUF582 domain. Their expression profile sug- delivered through the type III secretion system to the host cell cytoplasm and gests that they could be involved in the middle and/or late nucleus. FEMS Immunol. Med. Microbiol. 57:46–58. 16. Horn, M. 2008. Chlamydiae as symbionts in eukaryotes. Annu. Rev. Micro- steps of development, including exit from the host cell. The biol. 62:113–131. CT620 and CT621 distributions suggest that these DUF582 17. Hsia, R.-C., Y. Pannekoek, E. Ingerowski, and P. Bavoil. 1997. Type III proteins, and maybe other members of the family, are abun- secretion genes identify a putative virulence locus of Chlamydia. Mol. Mi- crobiol. 25:351–359. dantly present in the host cytoplasm during infection. This 18. Hueck, C. 1998. Type III protein secretion systems in bacterial pathogens of property, together with the observation that they are found in animals and plants. Microbiol. Mol. Biol. Rev. 62:379–433. 19. Jorgensen, I., and R. H. Valdivia. 2008. Pmp-like proteins Pls1 and Pls2 are all pathogenic chlamydiae and not in other bacteria, makes secreted into the lumen of the Chlamydia trachomatis inclusion. Infect. them interesting candidates for major histocompatibility com- Immun. 76:3940–3950. plex (MHC) class I presentation of bacterial antigens. Our 20. Maurer, A. P., A. Mehlitz, H. J. Mollenkopf, and T. F. Meyer. 2007. Gene expression profiles of Chlamydophila pneumoniae during the develop- present efforts are concentrating on the identification of inter- mental cycle and iron depletion-mediated persistence. PLoS Pathog. acting proteins for several members of the family, which might 3:752–769. 21. Meńard, R., P. Sansonetti, and C. Parsot. 1994. The secretion of the Shigella give clues to the putative function(s) of the conserved DUF582 flexneri Ipa invasins is activated by epithelial cells and controlled by IpaB domain. and IpaD. EMBO J. 13:5293–5302. 22. Meńard, R., P. J. Sansonetti, and C. Parsot. 1993. Nonpolar mutagenesis of the ipa genes defines IpaB, IpaC, and IpaD as effectors of Shigella flexneri entry into epithelial cells. J. Bacteriol. 175:5899–5906. ACKNOWLEDGMENTS 23. Moulder, J. W. 1991. Interaction of chlamydiae and host cells in vitro. Microbiol. Rev. 55:143–190. We thank Vale´rieMalarde´for excellent technical help and Scot 24. Pei, J., B. H. Kim, M. Tang, and N. V. Grishin. 2007. PROMALS web server Ouellette for donation of density gradient-purified EBs and critical for accurate multiple protein sequence alignments. Nucleic Acids Res. 35: reading of the manuscript. W649–W652. 25. Pennini, M. E., S. p. Perrinet, A. Dautry-Varsat, and A. Subtil. 2010. Histone This work was supported by the European Marie Curie program methylation by NUE, a novel nuclear effector of the intracellular pathogen European Initiative for Basic Research in Microbiology and Infectious Chlamydia trachomatis. PLoS Pathog. 6:e1000995. Diseases and by the Agence Nationale pour la Recherche (ANR-06- 26. Rice, P., I. Longden, and A. Bleasby. 2000. EMBOSS: the European Molec- JCJC-0105). ular Biology Open Software Suite. Trends Genet. 16:276–277. 580 MUSCHIOL ET AL. INFECT.IMMUN.

27. Rockey, D. D., M. A. Scidmore, J. P. Bannantine, and W. J. Brown. 2002. 31. Subtil, A., C. Delevoye, M. E. Balanã´, L. Tastevin, S. Perrinet, and A. Proteins in the chlamydial inclusion membrane. Microbes Infect. 4:333–340. Dautry-Varsat. 2005. A directed screen for chlamydial proteins secreted by 28. Samudrala, R., F. Heffron, and J. E. McDermott. 2009. Accurate prediction a type III mechanism identifies a translocated protein and numerous other of secreted substrates and identification of a conserved putative secretion new candidates. Mol. Microbiol. 56:1636–1647. signal for type III secretion systems. PLoS Pathog. 5:e1000375. 32. Subtil, A., C. Parsot, and A. Dautry-Varsat. 2001. Secretion of predicted Inc 29. Scidmore, M. A. 2005. Cultivation and laboratory maintenance of Chlamydia proteins of Chlamydia pneumoniae by a heterologous type III machinery. trachomatis. Curr. Protoc. Microbiol. 11:A11.11–A11.25. Mol. Microbiol. 39:792–800. 30. Stephens, R. S. (ed.) 1999. Chlamydia. Intracellular biology, pathogenesis, 33. Valdivia, R. H. 2008. Chlamydia effector proteins and new insights into and immunity. American Society for Microbiology, Washington, DC. chlamydial cellular microbiology. Curr. Opin. Microbiol. 11:53–59.

Editor: J. B. Bliska II. Article 2: Monitoring of Chlamydia trachomatis developmental cycle using GFP-expressing bacteria, microscopy and flow cytometry

François Vromman, Marc Laverrière, Stéphanie Perrinet, Alexandre Dufour and Agathe

Subtil. Plos One (2014), in press.

A method to obtain fluorescent chlamydiae was published at the end of 2011. We reasoned that, based on this tool, new methods could be developed to follow the progression of infection faster and more quantitatively than before. Therefore, we developed microscopy and flow cytometry methods to monitor the entire developmental cycle. We chose to work with the L2 strain transformed with a plasmid expressing the GFP under the promoter of the incD gene obtained from I. Derré (Agaisse & Derré, 2013). This strain showed similar growth curves as the parental strain, as measured by qPCR. We also showed that the GFP signal is detectable in single EBs. The binding and the multiplication of the bacteria were measured using flow cytometry. This method yielded statistically relevant results in an easy, cheap and fast manner. We showed that the GFP amount present in EBs was suitable to detect the binding of the bacteria from a multiplicity of infection (MOI) of one. The entry step was quantified by microscopy coupled to automatic image analysis. For the first time for Chlamydia trachomatis, we quantified precisely the kinetics of entry of single bacteria. Using the ICY image analysis software developed at the Institut Pasteur, we measured that fifty percent of the bacteria have entered after 10 min of incubation at 37 °C. Flow cytometry is also an easy and a statistically relevant procedure to quantify the infection rate and bacterial load, especially in a non-homogeneous cell population. Finally, we used flow cytometry to directly count EBs using a cytometer that detects particles of small size, down to 0.2 "m. This is an important result because with this methods EBs can be counted directly in bacterial preparation or in crude cell lysates.

81 Quantitative monitoring of the Chlamydia trachomatis developmental cycle using GFP-expressing bacteria, microscopy and flow cytometry

François Vromman1,2,3, Marc Laverrière1,2, Stéphanie Perrinet1,2, Alexandre Dufour2,4 and Agathe Subtil1,2#

1 Institut Pasteur, Unité de Biologie des Interactions Cellulaires, 25 rue du Dr Roux, 75015 Paris, France 2 CNRS URA 2582, Paris, France 3 Université Pierre et Marie Curie, Cellule Pasteur UPMC, Paris, France 4 Institut Pasteur, Unité d’Analyse d’images biologiques, Paris, France

#Corresponding authors: Unité de Biologie des Interactions Cellulaires, Institut Pasteur 25 rue du Dr Roux, 75015 Paris, France Tel: +33 1 40 61 30 49 Fax: + 33 1 40 61 32 38 E-mail: [email protected]

Running title: Quantitative measurement of Chlamydia infection

82 Abstract

Chlamydiae are obligate intracellular bacteria. These pathogens develop inside host cells through a biphasic cycle alternating between two morphologically distinct forms, the infectious elementary body and the replicative reticulate body. Recently, C. trachomatis strains stably expressing fluorescent proteins were obtained. The fluorochromes are expressed during the intracellular growth of the microbe, allowing bacterial visualization by fluorescence microscopy. Whether they are also present in the infectious form, the elementary body, to a detectable level has not been studied. Here, we show that a C. trachomatis strain transformed with a plasmid expressing the green fluorescent protein (GFP) accumulates sufficient quantities of the probe in elementary bodies for detection by microscopy and flow cytometry. Adhesion of single bacteria was detected. The precise kinetics of bacterial entry were determined by microscopy using automated procedures. We show that during the intracellular replication phase, GFP is a convenient read-out for bacterial growth with several advantages over current methods. In particular, infection rates within a non-homogenous cell population are easily quantified. Finally, in spite of their small size, individual elementary bodies are detected by flow cytometers, allowing for direct enumeration of a bacterial preparation. In conclusion, GFP-expressing chlamydiae are suitable to monitor, in a quantitative manner, progression throughout the developmental cycle. This will facilitate the identification of the developmental steps targeted by anti-chlamydial drugs or host factors.

83 INTRODUCTION

Chlamydiae are obligate intracellular bacteria that grow in very diverse eukaryotic hosts, including humans. Chlamydia trachomatis is the most prevalent sexually transmitted bacterial pathogen (more than 2.5 million infections annually in the United States as estimated by the Center for Disease Control in 2012) and can lead to severe pathologies including infertility, ectopic pregnancy, and pelvic inflammatory disease. Conjunctival inflammation as a result of C. trachomatis infection is the leading cause of blindness by an infectious agent, with about 8 million people irreversibly visually impaired by trachoma and an estimated 84 million cases in need of treatment (World Health Organization 2011) [1]. Chlamydiae develop in a biphasic cycle, which is a landmark of the order [2]. The infectious forms of the bacteria, called elementary bodies (EBs), are characterized by a small size (around 0.3 µm), a rigid cell wall, densely packed DNA, and reduced metabolic activity. Upon entry into a host cell, typically an epithelial cell, EBs convert to reticulate bodies (RBs). RBs are larger (1-2 µm), metabolically active, and multiply within a membrane-bound vacuole called the inclusion. After several rounds of division, RBs convert back to the EB form, still within the inclusion, before ultimately exiting the host cell. Completion of the whole cycle takes two or more days depending on the species. The initial steps of infection (adhesion, entry, conversion, and division) are asynchronous, ultimately leading to a population of infected cells with inclusions of variable sizes. These inclusions typically contain a mixture of EBs and RBs at later times during infection. The obligate intracellular growth of chlamydiae, and the absence of genetic manipulation tools, have limited the development of tools to measure the progression of the infectious cycle through its different steps in vitro. Adhesion can be quantified by flow cytometry using antibodies against the major component of the outer membrane (MOMP) [3] or using bacteria labeled with a fluorescent dye [4]. To measure bacterial entry, the method of choice is microscopy, using a two-step permeabilization protocol to distinguish intracellular from extracellular bacteria [5,6]. This method is time consuming and requires intensive work to be precise. Intracellular growth is usually assessed by quantitative PCR, measuring genome number [7], or by microscopy, measuring inclusion number and size. Alterations in the infectious cycle affect the number of infectious particles produced, and the “infectious progeny” is enumerated through re-infection assays and quantification of the inclusion forming units (IFUs) by microcopy. This requires fixation of the samples followed by manual or automatic counting of the inclusions by microscopy after inclusion staining with anti-

84 bacterial antibodies [8,9]. Various methods for staining and directly counting EBs under the microscope have also been described, yet all are rather tedious and rarely employed [10]. Recently, following the pioneering work by the Clarke lab, C. trachomatis strains stably expressing fluorescent proteins were obtained in various laboratories [11,12,13]. Bacteria expressing the green fluorescent protein (GFP) are of particular interest because all fluorescent microscopes and flow cytometers are equipped with the lasers and filters required to detect this fluorochrome. Here, we show that the fluorescent signal of GFP-expressing chlamydiae allows monitoring of the progression through the developmental cycle in a quantitative manner. This has several advantages over current methods. Thousands of events can be rapidly analyzed by flow cytometry, generating highly quantitative data even on rare events. Calibration of the instrument allows detection of individual fluorescent EBs and thus direct counting of the particles. The fluorescence level in individual EBs is also sufficient for detection by microscopy, and we show that automated tools allow for rapid quantification of bacterial entry into cells. Using these quantitative tools, the action of anti-chlamydial compounds on different steps of the chlamydial developmental cycle can easily be assessed.

MATERIAL AND METHODS C. trachomatis strains. One clonal population from C. trachomatis L2 strain 434 (ATCC) was plaque isolated before transformation with SW2::GFP [11] or p2TK2-SW2 IncDProm-RSGFP-IncDTerm [13] as has been previously described [11]. EBs were purified on a density gradient as described [14].

Cell culture, transfection and chemicals. HeLa cells were cultured in Dulbecco’s modified Eagle’s medium with Glutamax (DMEM, Invitrogen) supplemented with 10% (v/v) fetal calf serum (FCS). Construction of the histidine-tagged IncA constructs was described elsewhere [15]. Cells were transfected 24 hrs after seeding using JetPrime transfection kit (Polyplus transfection) and infected 24 hrs later. Tetracycline (12.5 mg/ml stock in ethanol) and cytochalasin D (5 mg/ml stock in DMSO) were purchased from Sigma and stored at - 20 °C.

Quantitative-PCR analysis. Cells were infected at a multiplicity of infection (MOI) between 0.1 and 0.2 for each of the three strains tested. At the indicated time points, cells

85 were gently detached using 0.5 mM EDTA in PBS, pelleted by centrifugation at 400xg, resuspended in PBS, and stored at -20 °C. DNA was extracted using the DNeasy blood & tissue kit (Qiagen) following the manufacturer instructions, with cell lysis at 65 °C for better bacterial lysis. Total DNA concentrations were measured and normalized to 50 ng/µl using a Nanodrop spectrophotometer (Thermo Scientific). Note that the contribution of bacterial DNA to total DNA is negligible compared to host DNA in these experimental conditions. Primers targeting the ompA gene (GGTTTCGGCGGAGATCCT and AGTAACCAACACGCATGCTGAT) were used at 10 µM with the Quantitect SYBRgreen PCR kit (Qiagen) following the manufacturer’s instructions. Q-PCR reactions were run in an LC480 Lightcycler thermocycler (Roche), starting with an activation phase of 15 min at 95 °C, followed with 40 cycles with an annealing temperature of 54 °C. Data were normalized to standard curves of purified L2 genomic DNA amplified in parallel with the experimental samples.

Observation of fluorescent EBs. L2 or L2incDGFP EBs were centrifuged at 400xg on poly-L-lysine treated coverslips and fixed for 30 min in 4% paraformaldehyde (PFA) 120 mM sucrose in phosphate buffered saline solution, PBS (fixation buffer). Bacteria were stained with a mouse anti-MOMP-LPS (Argene #11-114) antibody followed with Cy5-conjugated secondary antibodies. Bacteria were permeabilized for 15 min in 0.3% (v/v) Triton X-100 in PBS 1mg/ml BSA prior to DNA staining for 30 min using 0.5 µg/ml Hoechst 33342 (Molecular Probes) in PBS with 1mg/ml BSA. Coverslips were then mounted in a Mowiol solution.

Quantification of entry with a semi-automated procedure. Entry experiments were performed on cells seeded the day before on coverslips (40,000 cells/well) in 24-well plates. Prior to infection, cells were incubated at 37 °C for 30 min in culture medium supplemented or not with 1 µg/ml cytochalasin D (Sigma) or DMSO and were maintained in this buffer until fixation. To disrupt bacterial aggregates, EBs purified on a density gradient [14] were briefly sonicated prior to infection. Cells were incubated at 4 °C for 15 min in DMEM 10% FCS before adding the bacteria (MOI=10) for another 30 min at 4 °C. Medium was replaced by medium prewarmed at 37 °C, and plates were transferred to the 37 °C incubator for the indicated times before fixation in ice-cold fixation buffer for 30 min. Extracellular bacteria were stained with a mouse anti-MOMP-LPS as described above. Pictures of fields with 5-10 cells were acquired using an Axio observer Z1 microscope equipped with an ApoTome

86 module (Zeiss, Germany) and a 63$ Apochromat lens. Pictures were taken with a Coolsnap HQ camera (Photometrics, Tucson, AZ) using the software Axiovision. A minimum of 80 bacteria was analyzed per condition. We designed an automatic, ready-to-use analysis protocol for the Icy software [16] to perform the quantification on entire image folders without manual intervention (The Chlamentry protocol will be made publicly available on the Icy website upon publication). First, a wavelet-based detection module [17] was used to detect all objects in the green and red channels. Then, an object-based colocalization module was used to visualize and quantify the colocalization between the two detection sets. Two detections were considered colocalized under a distance threshold of 4 pixels (i.e. 400 nm) between their center of mass, accounting for the chromatic aberration of the imaging setup. Finally, the protocol produced a comprehensive result sheet containing the number and location of detected objects in each channel, the number of colocalized detections (i.e. number of extracellular bacteria), and a final script calculated the ratio of [green - colocalized detection] to [green detection] (i.e. ratio of internalized bacteria). Of note, the light sonication procedure preceeding infection can lead to the appearance of red-and-not-green dots. These red dots are also usually not visible in the blue channel (DNA), and presumably correspond to bacterial wall debris. We used conditions where such events represented less than 10% of the total red staining. In addition, these objects are not scored by the software since they are not green, and therefore do not affect the measured efficiency of entry.

Flow cytometry on infected cells. Cells were seeded in 6-well plates (400,000 cells/well) the day before infection. Cells were infected for one hour with L2incDGFP EBs purified on a density gradient at the indicated MOI before changing the medium and returning the plate to the incubator. At the indicated times, cells were washed with PBS and gently detached using 0.5 mM EDTA in PBS. Samples were fixed in PFA 2% in PBS and stored over-night at 4 °C. For adhesion experiments, cells were incubated for 4 hrs at 4 °C with L2incDGFP EBs at the indicated MOI before being washed, detached and fixed as described above. Flow cytometry analysis was performed with a FACS Gallios (Beckton Coulter) using the FL-1 (detecting fluorescence emission between 505 and 545 nm), the FSC (relative cell size) and the side scatter detectors (cell granulometry or internal complexity) on 1/10 of the sample diluted in PBS. The FACS Gallios parameter FSC collection angle N (Narrow FSC angles 1 – 8°) was used, triggering on the FSC channel during acquisition. A minimum total of 10,000 gated events were collected for each sample. Data were analyzed using the Kaluza

87 1.2 software (Beckman Coulter) and FlowJo. Calibration beads were purchased from Spherotech. For analysis of transfected cells, after fixation the cells were centrifuged at 1500xg, washed in PBS, centrifuged again, and incubated for 1 hr with home-made rabbit anti- histidine tag antibodies in 1 mg/ml BSA, 0.05% (w/v) saponin in PBS. The cells were washed and incubated for one hour in the same buffer with anti-rabbit antibodies conjugated to Cy5. The cells were washed, resuspended in PBS and analyzed by flow cytometry in the FL-1 (green) and FL-4 (far-red) channels.

Flow cytometry on EBs. EBs purified on a density gradient were serially diluted and fixed in 2% PFA in PBS. Flow cytometry analysis was performed with a FACS Gallios using the FL-1, the FSC, and the side scatter detectors on 1/10 of the sample diluted in PBS. The W2 (Enhanced Wide angle) was used for the FSC parameter collection, triggering on the FL- 1 channel during the acquisition, 100,000 events were acquired. Serial dilutions of EBs freshly prepared (no freezing step) were used, on one hand to quantify particle concentration by flow cytometry, and, on the other hand to infect fresh HeLa cells and determine the titer expressed as IFU/ml by serial dilutions, as has been described [14].

RESULTS GFP is detected both in RBs and EBs GFP-expressing C. trachomatis will be suitable to monitor infection if they show similar growth rate as the parental strain and if the fluorescent protein is expressed at a level sufficient for detection. We compared the growth kinetics of the parental C. trachomatis L2 strain with two different strains obtained by transformation with a plasmid expressing GFP in one case under the control of a promoter derived from Neisseria meningitidis [11] and in the second case from the promoter of the incD gene of C. trachomatis [13]. In the rest of the manuscript, these three strains are designated L2, L2nmGFP and L2incDGFP, respectively. Bacterial growth was measured by Q-PCR of chlamydial genome content over time between 17 and 23 hrs of infection, which are within the exponential growth phase [7]. Similar growth rates were observed (Fig. 1A), demonstrating that expression of the exogenous gene has no impact on bacterial multiplication. We measured a doubling time of about 3 hrs, a value similar to the doubling time reported for L2 [7].

88 Fluorescent RBs have been shown to be amenable to microscopy, including live [11,13], but whether GFP is also present in EBs to a detectable level has not been studied. Purified EBs from the L2incDGFP strain were attached to a coverslip using poly-L-lysine, and the bacteria were fixed using PFA and stained with anti-MOMP antibodies followed with secondary antibodies coupled to a red fluorescent dye. Green particles were observed on the coverslip incubated with purified L2incDGFP EBs and not with purified L2 EBs (Fig. 1B). These particles correspond to bacteria since they are co-stained with anti-MOMP antibodies and with DNA labelling. Thus, the level of GFP present in EBs is sufficient for detection by microscopy. Fluorescence of L2incDGFP EBs was stronger than that of L2nmGFP EBs, a difference that was also observed when the fluorescence of whole inclusions was compared (not shown). This higher expression does not impact growth as infection with L2incDGFP resulted in similar infectious progeny as with the parental strain [13]. We therefore chose to use the L2incDGFP strain to ask if these bacteria were suitable to monitor the different steps of the developmental cycle using flow cytometry.

Quantitative measure of the adhesion step by flow cytometry We examined whether bacterial adhesion could be measured by flow cytometry. We incubated cells with different amounts of purified L2incDGFP EBs for 4 hrs at 4 °C. Cells were then fixed and analyzed by flow cytometry. Even at a low MOI of 1.11, a small shift of the cell-associated green fluorescence was observed (Fig. 2A). The difference in mean fluorescence increased in a linear manner with bacterial concentration up to the highest MOI tested (MOI=30) (Fig. 2B). Thus, the GFP signal can be used as a convenient read-out to quantify EB adhesion by flow cytometry. The possibility of analyzing a large number of events make this approach quantitative even at low MOI.

Fluorescence of infectious particles is sufficient for tracking entry by automated microscopy tools We next asked whether the GFP signal could be used to monitor bacterial entry into cells. For future applications, we anticipated that co-localization with host proteins, or other information on the spatial organization of the entry step, would be sought. We therefore used a microscopy-based approach, which, in contrast to flow cytometry, gives access to spatial information. Cells were incubated with bacteria at 4 °C to allow bacterial attachment but not entry, and then transferred to a 37 °C incubator for variable times. The cells were then fixed

89 with paraformaldehyde, and extracellular bacteria were stained with anti-MOMP antibodies followed with Cy5-conjugated secondary antibodies. For each time points, 10 pictures in randomly selected fields were acquired in the green and the far-red channels and analyzed using the ICY software, as detailed in the Methods. Preliminary experiments done with cells infected for 15 min showed that the same proportion of bacteria was scored intracellular when using this automated procedure or when counting manually. Furthermore, when the bacteria were stained in two steps as previously described [6], similar internalization levels were observed as with the one step procedure described here, validating the new method. We next used this procedure to follow C. trachomatis L2 entry over the first hour of infection. As a control we included cells treated with cytochalasin D, a known inhibitor of chlamydial entry [18,19,20], or the solvent DMSO. Around 50 percent of the bacteria were internalized within the first 10 min of incubation at 37 °C and reached a plateau with 75% intracellular bacteria within 30 min (Fig. 3). To the best of our knowledge these kinetics represent the first precise description of the rate of entry of C. trachomatis. Our data fit with the reported efficiency of internalization of L2 in HeLa cells within 30 min [5]. As expected, depolymerization of the actin cytoskeleton using cytochalasin D completely abrogated entry. Interestingly, it also reduced bacterial adhesion at 4 °C by about one third (data not shown).

Quantitative measurement of antichlamydial activities Bacterial entry is followed by differentiation into the RB form, and intracellular multiplication. We followed the intracellular growth of L2incDGFP using flow cytometry. Green bacteria were observed at all time points by microscopy and detected by flow cytometry. At early time points, the GFP signal is too low to fully separate the infected cells from the non-infected ones (Fig. S1). This limitation is due to the fact that each infected cell only contains one to a few bacteria, whose GFP signal is within the range of the autofluorescence level of the host cell. The population of infected cells was fully discernible from non-infected cells in the green fluorescence channel 18 hrs post infection (Fig. 4A). As expected, within one infection round, the percentage of infected cells remained stable between 18 hrs and 30 hrs of infection while the mean fluorescence increased as the bacteria divided. We verified that the percentage of infected cells measured by flow cytometry matched the percentage of infection measured by visual examination of the population by microscopy, which is currently the method of choice for quantifying IFUs (Fig. 4B).

90 The mean intensity of the whole population (including infected and non infected cells) can be used as a practical read-out of the activity of anti-bacterial drugs. To illustrate this application we measured the mean fluorescence in infected cells exposed to increasing concentrations of tetracycline, a potent inhibitor of bacterial growth. We determined the IC50 for this antibiotic to be 40 ng/ml (Fig. 4C). One strong advantage of the single cell analysis lies in the possibility to measure differences in infection rates within non-homogenous populations. For instance, it can show in one simple step whether the overexpression of a particular protein affects the bacterial developmental cycle. As a proof of principle we evaluated the effect of ectopic expression of the inclusion protein IncA on bacterial growth. Cells were transfected with a plasmid expressing His-tagged IncA, either full-length or with a 75 amino acid truncation at its N- terminus. Twenty-four hours later, cells were infected with L2incDGFP strain, and 24 hrs later they were fixed and processed for analysis by flow cytometry, with detection of the transfected population in the red channel and of the infected population in the green channel (Fig. 4D). We observed a three-fold reduction in the infection rate of cells expressing full- length IncA-His compared to non-transfected cells from the same well. This result is in agreement with our previous observation, based on microscopy data, that ectopic IncA expression [21], but not that of the deletion mutant [15], ultimately results in cell death of infected cells. This was not a non-specific consequence of transfection since cells expressing %75IncA-His were actually slightly better infected than the non-transfected cells, a trend that we repeatedly observed with transfection of other negative controls (data not shown).

Quantification of EBs production Having determined that the GFP signal emitted by L2incDGFP EBs is detected by microscopy, we asked whether this fluorescence could be used to detect individual EBs using flow cytometry. New generation flow cytometers have lowered the size limit for detection of particles and are suitable for the detection of small particles including bacteria. However, the small size of EBs (around 0.3 µm) makes them particularly challenging to detect by this method. Using uninfected cells to set up the background of fluorescence, we detected fluorescent particles in a sample containing purified L2incDGFP EBs that were absent from a preparation of purified L2 EBs and from non-infected cells. When gating on this fluorescent population, and using calibration beads, we observed that the particle size ranged between 0.22 and 0.45 µm, which is the expected size for EBs (Fig. 5A). The distribution of the

91 fluorescence values gives a sharp peak, indicating that the fluorescence of individual EBs was homogenous (Fig. 5B). By calibrating on the fluid speed, we estimated the concentration of serial dilutions of one fresh bacterial preparation with a good approximation over 4 logs, down to 103 EBs/ml (Fig. 5C). The same preparation, determined by flow cytometry to contain 17.7±2x106 EBs/ml, was used to serially infect cells. We found that the sample contained 16.9±4 x106 IFU/ml, validating the direct particle enumeration. This result also shows that the particles detected by flow cytometry were infectious.

DISCUSSION The L2incDGFP strains proved suitable to track the chlamydial developmental cycle by flow cytometry: (i) its growth is identical to that of the parental non-fluorescent strain ([13] and Fig. 1A), (ii) the fluorescence signal can be used as a read-out of bacterial multiplication, (iii) the fluorescent signal was detected in the EBs and allowed for direct quantification of the extracellular and intracellular bacteria. For the same bacterial load, the fluorescent signal was about three-fold stronger in cells infected with L2incDGFP than with L2nmGFP at the mid-phase of the cycle (not shown). It likely reflects the fact that the GFP signal expressed by SW2::GFP plasmid is hindered by the fusion of the GFP with chloramphenicol acetyl transferase [11]. It might also reflect a difference in strength of the two promoters or in their timing of activity during the cycle. The incD promoter is expressed very early and during the whole developmental cycle [22], which might explain why we could detect bacterial growth even at early time points. Use of other promoters might delay the detection of the fluorescent marker. However, it is not known if the distinction between early and late promoters holds true when the genes are expressed from the plasmid.

Traditionally, effects of antichlamydial compounds on infection are assessed by measuring genome numbers by quantitative PCR or by measuring the resulting IFU through reinfection of fresh cells. These methods are costly, labor intensive, and often only one time point is selected for the analysis. Analysis of infection via flow cytometry is cheaper, faster, more quantitative, especially when analyzing rare events, and gives additional information. Measure of the GFP signal gives immediate access to the percentage of infected cells and to the mean intensity of fluorescence in the infected population. These data are very useful to determine the inhibition properties of antimicrobial compounds. Unlike measures of IFUs or

92 genome numbers, they distinguish between antimicrobial activities that prevent establishment of the infection (fewer infected cells, normal rate of accumulation of GFP in the infected population) from those that prevent bacterial growth (same percentage of infected cells, slower accumulation of GFP). In addition, these parameters can be correlated to a third one such as the expression of a protein in a non-homogenous population of transfected cells, as illustrated with the inhibition of bacterial development in cells expressing full-length IncA. Altogether, these GFP-expressing strains, coupled to high throughput flow cytometry, offer a powerful tool in a wide range of studies, in particular rapid screening of anti-chlamydial compounds. For instance we found that the IC50 for tetracycline is 40 ng/ml. This finding is in agreement with a previous report [23]. In addition, we showed that GFP-loaded EBs allow the measurement of bacterial adhesion to cells by flow cytometry, simplifying an alternative procedure using bacteria labeled with a fluorescent dye [4].

Until now the most direct method to measure EB entry was a two-step procedure, in which external and internal bacteria were successively stained before and after permeabilization of the sample [5,6]. This method, followed with manual counting of internalized bacteria is tedious. Consequently, EB internalization rates are mostly absent from the literature, with, to the best of our knowledge, a single report on the kinetics of C. pisttaci entry [24]. In addition, using FITC-coupled bacteria, we had noticed that some of the internalized bacteria were not accessible to antibodies, even after saponin permeabilization, possibly because actin polymerization around nascent inclusions limits antibody access. The green fluorescence of the EBs reduces the time needed to process the sample (one single labeling step), and eliminates underscoring of intracellular bacteria due to limited antibody access. Also, the green fluorescence is strong enough to allow automated detection by the ICY software, which drastically reduces the analysis time. Provided that enough images are collected, the procedure can be applied to very low MOI. It will facilitate future discovery of anti-chlamydial compounds that target the internalization step per se.

Finally, we show here that, in spite of their small size, GFP-expressing EBs can be enumerated by flow cytometers designed for the detection of small particles. This is a huge advantage over current methods of titrating bacterial preparations, which require reinfecting fresh cells to measure IFUs. Using calibrating beads, we determined that the EB diameter was between 0.22 and 0.45 µm, which fits well with observations at the ultrastructural level. We could also observe fluorescent particles between 0.88 and 1.33 µm, the expected size range

93 for RBs (data not shown). However, even when the bacteria were collected at mid-cycle, either in PBS or in the conventional sucrose phosphate glutamate buffer used for chlamydial preparation, we recovered a minority of these larger particles compared to EBs, suggesting that the majority of RBs is lost by lysis during sample preparation.

C. trachomatis L2 was the first Chlamydia strain engineered to express GFP. Recent work has shown that the same plasmid can be stably inserted in other C. trachomatis strains, as well as in C. pneumoniae [25,26]. It is very likely that the quantitative methods we developed here are also applicable to other chlamydial species. The use of GFP-expressing Chlamydia will greatly improve the quantitative assessment of the progression of these different chlamydial species throughout their developmental cycles and will aid in the identification of compounds that affect specific steps of the cycle.

ACKNOWLEDGMENTS We thank Dr. S. Ouellette and Dr. A. Dautry for critical reading of the manuscript, Dr. I. Clarke and Dr. I. Derré for providing the SW2::GFP and p2TK2-SW2 IncDProm-RSGFP- IncDTerm plasmids, respectively, Clémence Taisne for generating stably transformed bacteria, P-H Commère and Dr. M. Nguyen-de Bernon (Plateforme de Cytométrie, Institut Pasteur) for help in flow cytometry. FV is funded by the Ministère de l'Education Nationale, de la Recherche et de la Technologie and by the Fondation pour la Recherche Médicale. ML was funded by a DIM-Malinf (Ile-de-France) post-doctoral fellowship. This work was supported by an ERC Starting Grant to AS (NUChLEAR N°282046), the ANR (Ménage à trois, ANR- 12-BSV2-0009-02), the Institut Pasteur and the Centre National de la Recherche Scientifique.

REFERENCES 1. Batteiger BE (2012) Chlamydia infection and epidemiology. In: Tan M, Bavoil PM, editors. Intracellular pathogens I; Chlamydiales. Washington,DC: ASM press. 2. AbdelRahman YM, Belland RJ (2005) The chlamydial developmental cycle. FEMS Microbiol Rev 29: 949-959. 3. Levitt D, Zable B, Bard J (1986) Binding, ingestion, and growth of Chlamydia trachomatis (L2 serovar) analyzed by flow cytometry. Cytometry 7: 378-383.

94 4. Molleken K, Schmidt E, Hegemann JH (2010) Members of the Pmp protein family of Chlamydia pneumoniae mediate adhesion to human cells via short repetitive peptide motifs. Mol Microbiol 78: 1004-1017. 5. Carabeo RA, Grieshaber SS, Fischer E, Hackstadt T (2002) Chlamydia trachomatis induces remodeling of the actin cytoskeleton during attachment and entry into HeLa cells. Infect Immun 70: 3793-3803. 6. Subtil A, Wyplosz B, Balañá ME, Dautry-Varsat A (2004) Analysis of Chlamydia caviae entry sites and involvement of Cdc42 and Rac activity. J Cell Sci 117: 3923-3933. 7. Shaw EI, Dooley CA, Fischer ER, Scidmore MA, Fields KA, et al. (2000) Three temporal classes of gene expression during the Chlamydia trachomatis developmental cycle. Molecular Microbiology [ print] 37: 913-925. 8. Osaka I, Hills JM, Kieweg SL, Shinogle HE, Moore DS, et al. (2012) An automated image- based method for rapid analysis of Chlamydia infection as a tool for screening antichlamydial agents. Antimicrob Agents Chemother 56: 4184-4188. 9. Bogdanov A, Endresz V, Urban S, Lantos I, Deak J, et al. (2013) Application of DNA chip scanning technology for the automatic detection of Chlamydia trachomatis and Chlamydia pneumoniae inclusions. Antimicrob Agents Chemother. 10. Gerloff RK, Ritter DB, Watson RO (1971) Counting Chlamydial Particles by Negative Staining with Congo Red. J Infect Dis 123: 429-432. 11. Wang Y, Kahane S, Cutcliffe LT, Skilton RJ, Lambden PR, et al. (2011) Development of a transformation system for Chlamydia trachomatis: restoration of glycogen biosynthesis by acquisition of a plasmid shuttle vector. PLoS Pathog 7: e1002258. 12. Wickstrum J, Sammons LR, Restivo KN, Hefty PS (2013) Conditional Gene Expression in Chlamydia trachomatis Using the Tet System. PLoS One 8: e76743. 13. Agaisse H, Derre I (2013) A C. trachomatis cloning vector and the generation of C. trachomatis strains expressing fluorescent proteins under the control of a C. trachomatis promoter. PLoS One 8: e57090. 14. Scidmore MA (2005) Cultivation and laboratory maintenance of Chlamydia trachomatis. Curr Protocols Microbiol: 11A11.11-11A11.25. 15. Delevoye C, Nilges M, Dehoux P, Paumet F, Perrinet S, et al. (2008) SNARE protein mimicry by an intracellular bacterium. PLoS Pathog 4: e1000022. 16. de Chaumont F, Dallongeville S, Chenouard N, Herve N, Pop S, et al. (2012) Icy: an open bioimage informatics platform for extended reproducible research. Nat Methods 9: 690-696.

95 17. Olivo Marin JC (2002) Extraction of spots in biological images using multiscale products. Pattern Recognition 35: 1989-2016. 18. Boleti H, Benmerah A, Ojcius D, Cerf-Bensussan N, Dautry-Varsat A (1999) Chlamydia infection of epithelial cells expressing dynamin and Eps15 mutants: clathrin- independent entry into cells and dynamin-dependent productive growth. J Cell Sci 112: 1487-1496. 19. Coombes BK, Mahony JB (2002) Identification of MEK- and phosphoinositide 3-kinase- dependent signalling as essential events during Chlamydia pneumoniae invasion of HEp2 cells. Cellular Microbiology 4: 447-460. 20. Ward ME, Murray A (1984) Control mechanisms governing the infectivity of Chlamydia trachomatis for HeLa cells: mechanisms of endocytosis. J Gen Microbiol 130: 1765- 1780. 21. Delevoye C, Nilges M, Dautry-Varsat A, Subtil A (2004) Conservation of the biochemical properties of IncA from Chlamydia trachomatis and Chlamydia caviae: oligomerization of IncA mediates interaction between facing membranes. J Biol Chem 279: 46896-46906. 22. Belland RJ, Zhong GM, Crane DD, Hogan D, Sturdevant D, et al. (2003) Genomic transcriptional profiling of the developmental cycle of Chlamydia trachomatis. Proc Natl Acad Sci U S A 100: 8478-8483. 23. Walsh M, Kappus EW, Quinn TC (1987) In vitro evaluation of CP-62,993, erythromycin, clindamycin, and tetracycline against Chlamydia trachomatis. Antimicrob Agents Chemother 31: 811-812. 24. Hodinka RL, Wyrick PB (1986) Ultrastructural-Study of Mode of Entry of Chlamydia- Psittaci into L-929 Cells. Infect Immun 54: 855-863. 25. Ding H, Gong S, Tian Y, Yang Z, Brunham R, et al. (2013) Transformation of sexually transmitted infection-causing serovars of chlamydia trachomatis using Blasticidin for selection. PLoS One 8: e80534. 26. Gerard HC, Mishra MK, Mao G, Wang S, Hali M, et al. (2013) Dendrimer-enabled DNA delivery and transformation of Chlamydia pneumoniae. Nanomedicine 9: 996-1008.

96 FIGURE LEGENDS

Figure 1: GFP-expressing C. trachomatis serovar L2 strains are suitable to track RBs and EBs. (A) Genomic DNA accumulation over time for the parental strain and the two GFP-expressing strains, L2nmGFP and L2incDGFP, with GFP expression controlled by a neisserial (nm) promoter or the promoter for the chlamydial incD gene, respectively. Bacterial DNA was quantified by Q-PCR on the ompA gene. An experiment representative of three is shown. In this experiment, the MOI for the parental strain was about 2 to 3 times less than that for the two other strains, accounting for the lower bacterial DNA amount at all time points. Nevertheless, the growth curves are similar for the three strains (B) Coverslips were coated with L2 (top) or L2incDGFP (bottom) EBs. After fixation, the bacterial envelope was stained with an anti-MOMP/LPS antibody followed with Cy5 coupled secondary antibodies, and bacterial DNA was labeled with Hoechst. The scale bar represents 5 µm.

Figure 2: Quantification of C. trachomatis adhesion by flow cytometry. Cells were incubated for 4 hrs at 4 °C with the indicated MOI of L2incDGFP purified EBs, before fixation and analysis in the green channel by flow cytometry. (A) Histograms of fluorescence in the green (FL-1) channel. In each panel, the fluorescence of non-infected cells is reported for comparison (dashed line). The horizontal bar delimits the fluorescence above background level and the percentage of cells (out of total cells) reaching these fluorescence values is indicated above. NI = non-infected. (B) The mean fluorescence of non-infected cells was subtracted from the mean fluorescence of infected cells and the resulting fluorescence value was plotted against the MOI. The inset shows an enlargement of the values obtained at low MOI. This experiment is representative of two.

Figure 3: Quantification of C. trachomatis internalization by automated microscopy. Cells were pre-incubated at 37 °C for 30 min in culture medium alone (untreated) or supplemented with 1 µg/ml cytochalasin D or solvent (DMSO). Cells were then transfered to 4 °C and incubated with bacteria (MOI= 10) for 30 min. At time zero, the plates were transferred to 37 °C and incubated for the indicated times. The cells were fixed, extracellular bacteria were labeled with anti-MOMP antibodies followed with Cy5-coupled secondary antibodies, and DNA was labeled with Hoechst. (A) Representative fields of the untreated control in the blue (first column), green (second column) and far-red (third column) channels with the merged pictures shown on the right. Prior to transfer to 37 °C (top) all bacteria

97 (green) are extracellular (red), while after 10 min at 37 °C half of the bacteria are internalized (arrowheads). The scale bar represents 5 µm. (B) Kinetics of bacterial entry. Images were processed with the ICY software as described in the methods section. Each time point represents averages on more than 80 bacteria from 10 different fields. One experiment representative of three is shown.

Figure 4: Quantitative measurement of the C. trachomatis growth cycle using flow cytometry. (A) Kinetics of chlamydial growth. Cells were infected at a MOI = 0.5 and fixed at the indicated times. Samples were analyzed by flow cytometry as described in the Methods section, and histograms of fluorescence in the green channel (FL-1) are shown. For each time point, 10,000 cells were analyzed. The horizontal bar delimits the fluorescence above background level (= infected cells). The percentage of cells included in this gate, and their mean fluorescence, are indicated. NI = non-infected. (B) Comparison of the determination of infection rates by flow cytometry and microsocopy. Cells were infected at MOI=0.2 in a 36 mm dish with one coverslip in the dish. Twenty-four hours later, cells on the coverslips were fixed and processed for microscopy toma nually count the percentage of infected cells (top left panel, cellular DNA appears in blue and inclusions are green, bar=20 µm). The rest of the dish was used for determining the infection rate by flow cytometry (top right panel). The average of five measurements with each method is shown. A t-test showed no statistical difference between the two methods. (C) Determination of tetracycline IC-50 by flow cytometry. Cells were infected (MOI=0.3) for one hour prior to tetracycline addition and incubated for 24 hrs before analysis by flow cytometry as described above. The histogram shows the mean fluorescence in the infected population relative to the value measured in infected cells treated with ethanol only, the error bars show the standard deviation in the triplicate of this experiment. The experiment has been reproduced three times. (D) Analysis of infection rates in a non-homogenous population. Cells were transfected for 24 hrs with the indicated construct prior to infection with L2incDGFP. At 24 hrs post-infection, cells were detached and fixed, and the His tag was stained in red as described in the Methods section. Transfected cells were positive in the red channel (FL-4), infected cells in the green channel (FL-1). A dot-plot analysis of the two parameters is shown for cells transfected with IncA-His (left) or %75IncA-His (right). For each condition, the percentage of infected cells in the transfected and non-transfected populations is reported in the histogram. The experiment

98 shown is representative of three independent assays with error bars showing the standard deviation. In panels B, C and D results of t-tests are reported.

Figure 5: Quantification of EBs by flow cytometry. (A) Dot-plot of the size and granulometry of fluorescent particles in an EB preparation. The threshold of background fluorescence in the green channel was determined using non-infected cells broken with glass beads and fixed in 2% (w/v) PFA. A purified EB preparation fixed in 2% PFA was then acquired, and parameters in the FSC and SSC channels of events above the background green fluorescence threshold are shown. The squares delimit the perimeter of detection of calibration beads of the indicated size. The size of the fluorescent particles detected ranged between 0.22 and 0.45 µm. (B) Fluorescence of the L2incDGFP EBs. 68.2 percent (µ ±1 &) of the particles have a fluorescence between 40 and 105, thus reflecting a maximum of 2.5-fold variation in the fluorescence of individual EBs. (C) A fresh EB preparation was serially diluted and fixed in 2% PFA. For each dilution, the concentration of particles in the sample was calculated by flow cytometry based on the fluid speed (ml.min-1) and event detection (event.min-1).

Figure S1: Analysis of the early times of C. trachomatis L2 development using flow cytometry. Cells were infected at a MOI = 0.3 and fixed at the indicated times. Samples were analyzed by flow cytometry as described in the Methods section, and histograms of fluorescence in the green channel (FL-1) are shown. For each time point, 10,000 cells were analyzed. The horizontal bar delimits the fluorescence above background level. The percentage of cells included in this gate, and their mean fluorescence, are indicated. NI = non- infected. For each time point, one coverslip infected in the same conditions was fixed and permeabilized to stain the DNA with Hoechst 33342. DNA appears in blue and GFP- expressing bacteria (arrowheads) in green, bar=10 µm.

99 Figure 1

100 Figure 2

101 Figure 3

102 Figure 4

103 Figure 5

104 Figure S1

105 III. The DUF582 protein CT619 targets the ESCRT proteins Hrs and Tsg101 during Chlamydia trachomatis infection François Vromman, Stéphanie Perrinet and Agathe Subtil. Manuscript in preparation.

1. CT619 interacts with Hrs and Tsg101 a. CT619 is expressed late in the cycle

When this work was initiated, the group had obtained specific antibodies against three out of the five DUF582 proteins, and antibodies against CT712 and CT619 were missing. We decided to produce and purify recombinant GST-tagged CT619 and CT712 to use as immunogens in rabbits. CT619 hydrophobic domains were excluded to make the protein soluble, resulting in the following recombinant protein: GST-!81CT619!625. The GST- CT712 recombinant protein was not well expressed in the standard conditions tested (37°C or 16°C), so we produced it in microfermenters. Both recombinant proteins were expressed in the E. coli strain BL21 at 37 °C in LB medium, and purified using glutathione-sepharose beads. Purification efficiency was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of the different fractions followed with coomassie staining (Figure 18A). The best fractions of each recombinant protein were pooled and injected into rabbits for antibodies production (AgroBio, La Ferté, France). Recognition of their respective target by these antibodies was tested on HeLa cells transfected with GFP-tagged proteins (Figure 18B). Immunofluorescence (IF) analysis showed that each antibody reacted with the protein it was raised against, and not with the other DUF582 protein GFP-CT621. The antibodies were further tested by western blot (WB), on infected and non-infected cellular lysates (Figure 18C). Each antibody recognized a protein of the expected size. Other bands were also observed, which were absent from non infected cells and presumably correspond to cross-reacting bacterial proteins. Notably, the antibody against CT619 recognizes two protein species: a upper band whose size corresponds to the expected size for the full-length protein and a lower band about 10 kDa below. A similar migration pattern had been observed for CT620 and CT711 (Article 1). Both CT619 and CT712 were detected 24 hpi and in increasing quantities thereafter, like the other DUF582 proteins (Article 1), indicating that they are mid to late proteins of C. trachomatis.

106

107 Figure 18: CT619 and CT712 are made late in the Chlamydia developmental cycle. A) Purification of recombinant proteins. GST-!81CT619!625 and GST-CT712 were produced in BL21 cultured at 37 °C and purified on a sepharose beads coupled to glutathione. Eluted fractions were analyzed by SDS-PAGE followed with protein coloration with Coomassie B) Antibodies against CT619 and CT712 recognize their cognate target by immunofluorescence. Cells were transfected with GFP-tagged CT619 and CT712 for 24 h and fixed in PFA 4%. Fixed cells were stained using the antibodies directed against CT619 and CT712, followed with Cy5-coupled anti-rabbit secondary antibodies. C) Anti-CT619 and –CT712 antibodies recognize the respective endogenous protein by WB and by IF. HeLa cells were infected with C. trachomatis L2 strain and lysed in 8M urea buffer at the indicated times. Samples were normalized to identical protein content, and analyzed by SDS-PAGE followed with transfer on PVDF membrane and immunoblotting with anti-CT619 and anti-CT712 antibodies (left). Cells grown on coverslips were fixed 42 h after infection in 4% PFA and stained with the indicated antibodies followed with Alexa488-coupled secondary antibodies (green), and Hoechst to label the DNA (blue). IB: Immunoblot, NI: Non-infected. Bar scale = 5 "m.

To determine the localization of the proteins during infection, HeLa cells were infected for 42 h and fixed in PFA 4%. Staining by both antibodies showed a dotty localization in the inclusion lumen or at its surface (Figure 17C). Staining with anti-CT619 was very weak, hardly above background levels observed with the preimmune serum (not shown).

b. Two-hybrid screen in yeast designates Hrs as a candidate partner for the DUF582

One strategy to find the function of a protein is to identify proteins with which it interacts. Since we showed that the DUF582 proteins are secreted by the T3SS of Chlamydia, we assumed that it might bind host proteins. We decided to use two-hybrid in yeast to find proteins that might interact with the DUF582 proteins. Bioinformatics showed that members of the DUF582 family have a variable N- terminus part. This domain is very different for each of the five proteins except for CT620 and CT621, which likely result from more recent gene duplication. The DUF582 itself is only moderately conserved since the amino acid identity ranges between 18% (CT711-CT712) to 39% (CT620-CT621) (Article 1). To determine if we could identify a common function for the DUF582 domain, we chose as baits the two domains of CT619 and CT621, which share 19.6% of identity. The N-terminal parts of these two proteins were also used as baits. We screened a bank expressing human placental cDNA. One single protein, Hrs, came as a common prey with both DUF582 domains used as baits (Figure 19). The interaction for CT619 was scored with the very high confidence (1st rank /6) while the one for CT621 was

108 scored with a low confidence (4th/6). In both cases, the minimal region of Hrs that interacted with the DUF582 corresponded to a central region of Hrs. They overlapped between the amino acids 372 and 480, thus including part of the proline-rich and the coiled coil domains of Hrs. Interestingly, Hrs was also a hit in the screen performed with the N-terminal part of CT621, although with a very low confidence score (5th/6). Each of these screens raised a high number of candidate interactors, more than any other screen we performed with other chlamydial proteins. Among those hits was Tsg101, which interacted with the N-terminal domain of CT619.

Figure 19: The Hrs protein interacts with the DUF582 domains of CT619 and CT621. Schematic view of the domains in Hrs. FYVE: zinc finger domain, UIM: Ubiquitin Interating motif, CC: Coiled coil domain, Pro/Gln: proline/glycine rich region, CBD: clathrin binding domain. The minimal regions of interaction with CT619 and CT621 are indicated below in yellow, and the overlap between these two regions is mapped in red.

c. Validation of the interaction between the DUF582 proteins and Hrs

We next studied the localization of GFP-tagged DUF582 proteins and a myc-tagged Hrs expressed by transfection in HeLa cells (Figure 20A). The overexpression of Hrs results in the formation of abberant endosomal compartments with a massive recruitment of clathrin around the endosomes (Bache, 2003; Urbé et al., 2003), which appeared as rings in the cell cytoplasm. Expression of the GFP-tagged versions of the DUF582 proteins gave a rather diffuse staining in the host cytoplasm (Figure 18B). Interestingly, the GFP staining changed completely and co-localized with the myc-Hrs rings or appeared to be trapped in those circular structures (Figure 20A). The high degree of co-localization between these two constructs supports the hypothesis that they interact. To address this question further, we immunoprecipitated myc-Hrs and looked for co- immunoprecipitation of GFP-tagged proteins. Because the efficiency of transfection in HeLa

109

Figure 20: Interations between the members of the DUF582 family and Hrs. A) IF analysis. HeLa cells were transfected with myc-Hrs and the indicated GFP-tagged DUF582 proteins for 24 h and fixed in PFA 4%. Cells were stained with a mouse anti-myc antibody followed with Cy5 coupled anti-mouse antibodies. Co-transfected HeLa cells show a strong co-localisation of the overexpressed GFP-tagged DUF582 proteins and myc-Hrs. B) GFP-DUF582 proteins co-immunoprecipitate with myc-Hrs. Hek293T cells were co-transfected with myc-Hrs and the indicated constructs for 24 h, lysed and immunoprecipitated with an anti-myc antibody. Samples were run on SDS-PAGE gel, transferred on PVDF membranes and analysed with an anti-GFP antibody. An aliquot from each lysate was loaded on a separate gel to compare the levels of expression of each protein (Input, bottom panels) IB: Immunoblot. C) Analysis of the interactions between different domains of the DUF582 proteins and Hrs by two-hybrid in yeast. Top panels show the schematic view of the constructs used in the assay. The middle panel shows an example of the assay with no interaction (-) and a strong interaction (+++). The table summaries the interactions measured. D) Analysis of the interaction site within the DUF582 of CT619. Deletion of the coiled coiled domain in the DUF582 and C-terminal did not abolish the interaction with Hrs.

110 cells was quite low, we used HEK293T cells (Figure 20B). No interaction was detectable between the GFP protein alone and myc-Hrs. In contrast, all the DUF582 proteins tested co- immunoprecipitated with myc-Hrs except for GFP-CT619. The absence of detection of CT619 in the precipitated fraction could be due to the lower rate of transfection of this construct compared to the other constructs used. However, when the DUF582 of CT619 was used, the interaction with myc-Hrs was clearly visible, even though its expression was similarly low. Thus, it is possible that the conformation of GFP-CT619 full-length is not compatible with Hrs binding, for unknown reasons.

To test other DUF582 proteins for interactions with Hrs and delimit the interaction domains we used the two-hybrid strategy further. In this assay, the DUF582 domains of CT619, CT621 CT711 and CT712 of C. trachomatis L2 have been tested. In addition we tested the N-terminal domains of CT619 and CT621. These constructs were cloned as baits (Figure 20C). The tested preys were the full length Hrs and a shorter version (%368Hrs%138) corresponding to the minimal domain interacting with CT619 and CT621 (Figure 19). Yeast were co-transformed with each plasmid and plated on selective medium for isolation of co-transformed cells. Two days after, the selected colonies were grown overnight in a liquid medium allowing counting of single cells the day after. Different dilutions of the same quantity of co-transformed yeast were plated on different selective media from low to high stringency. The results were read at 48 h post- plating (Figure 20C). Interaction with high confidence was detected for all of the DUF582 domains tested (CT619, CT711, CT712) except for the one of CT621. The DUF582 of proteins CT711 and CT712 showed a stronger interaction with Hrs than the one of CT619. The N-terminal parts of CT619 and CT621 both showed an interaction with Hrs, although weaker than the DUF582 domain. For all the observed interactions, the smaller version of Hrs displayed stronger interactions than the full length. To delimit further the interaction between CT619 and Hrs, we decided to test truncated versions of the CT619 DUF582 domain (Figure 20D). We generated a construct CT619%89, deleting the last 89 amino acids that correspond to the most conserved region in the DUF582 (Article 1). We also designed a construct truncated for the last 191 amino acids (CT619%191), which no longer contains the coiled coil region common to all DUF582.

111 Surprisingly, the CT619%191 construct is the one showing the strongest interaction with the small Hrs. All three CT619 constructs also interacted with the full length Hrs.

d. CT619 interacts with Tsg101

Interestingly, in the initial yeast-two-hybrid screens, the protein Tsg101 interacted with a high score of confidence (1st/6) with the N-terminal part of CT619 and with none of the other baits (Figure 21A). The minimal domain of interaction corresponds to a coiled-coil domain of Tsg101. To verify this interaction, we performed immunoprecipitation experiments using HA- tagged Tsg101 constructs of various lengths (Figure 21B). HEK293T cells were co- transfected with HA-tagged Tsg101 constructs and GFP-tagged CT619, or CT621 as control. Immunoprecipitation was performed with anti-HA antibodies coupled to agarose beads. Inputs and immunoprecipitated fractions were analyzed by WB with anti-HA and anti-GFP antibodies. CT619 co-immunoprecipitated with the two constructs of Tsg101 that contained the coiled-coil domain, and not with others (Figure 21C). The interaction of Tsg101 with CT619 is taking place between the coiled-coil of Tsg101 and the N-terminal part of CT619 (Figure 21D). In contrast, CT621 does not interact with any of the Tsg101 constructs. We next investigated the interaction during infection. HeLa cells were transfected with the HA-Tsg 3’ construct, infected for 35 hours and immunoprecipitated using anti-HA antibodies coupled with beads. The experiment reveals a co-immunoprecipitation of the endogenous CT619 with the HA-Tsg 3’ construct, confirming the interaction (Figure 21E). Altogether, these data demonstrate a specific interaction of the ESCRT-I protein Tsg101 with the DUF582 protein CT619.

112

Figure 21: The ESCRT-I protein Tsg101 interacts with CT619. A) Schematic view of the domains in Tsg101. The minimal region of interaction with CT619 is mapped in red. B) Schematic view of the HA-tagged Tsg101 constructs used. C & D) Analysis of the interaction by co-immunoprecipitation in transfected cells. HEK293T cells were co-transfected with Ha- tagged Tsg101 constructs and the indicated GFP tagged DUF582 proteins constructs for 24 h, lysed and immunoprecipitated with anti-HA coupled with agarose beads. Samples were run on SDS-PAGE gel, transferred on PVDF membranes and analyzed with an anti-GFP antibody (top) and anti-Ha (bottom). Both inputs and immunoprecipitated fractions are shown. E) Analysis of the interaction by co-immunoprecipitation in infected cells. HeLa cells were transfected with Ha-Tsg 3’and infected by LGV2 strains (MOI = 5) for 18 hpi or 38 hpi, before performing anti-HA immunoprecipitation as described above. Right panels: samples were run on SDS-PAGE gel, transferred on PVDF membranes and analyzed with anti-CT619 antibodies (top) and anti-HA antibodies (bottom) to verify the quality of the immunoprecipitation. Both inputs and immunoprecipitated fractions are shown. IB: immunoblot, IP: immunoprecipitation.

113 2. Hrs and Tsg101 are not required for chlamydial growth in vitro

a. Hrs and Tsg101 levels decrease during infection

We next analyzed the localization of Hrs during infection with C. trachomatis. HeLa cells were infected with C. trachomatis L2 (MOI= 0.3) for different times, fixed and stained with the indicated antibodies and analyzed on a microscope equipped with an Apotome module (Figure 22). Antibodies against the inclusion protein CT529 were used to stain the inclusion membrane. Hrs was often observed in close proximity with this membrane but did not appear to be enriched there.

Figure 22: Localization of Hrs during the infectious cycle. HeLa cells were infected by LGV2 strain at MOI of 0.3 for the indicated time before fixation on ice in PFA 4%. Cells were immunostained with rabbit antibodies against CT529 (inclusion protein) and mouse antibodies against Hrs, which were respectively targeted with antibodies coupled with Alexa488 and Cy5. DNA was stained with Hoechst 33342.

114 Next, we studied whether infection had an effect on the level of Hrs. We infected HeLa cells at a MOI of 5 to reach 100% of infected cells. At each time point, infected and uninfected cells were lysed and whole cell lysates were analysed by SDS-PAGE followed with immunoblotting using antibodies specific for Hrs. The experiment revealed that the level of Hrs was stable for the first 20 hours of infection and diminished afterwards, up to a 80% decrease 45 hpi (Figure 23A). To test whether this phenotype required bacterial activity, we repeated the experiment adding chloramphenicol 16 hpi, a time point preceding the start of Hrs decline. This antibiotic blocks bacterial protein synthesis and thus bacterial multiplication. Chloramphenicol strongly reduced the decrease of Hrs levels, indicating that bacterial factors are required for this effect (Figure 23B). The remaining drop in Hrs levels in the presence of chloramphenicol might be due to bacterial factors already present at the time of drug addition. Tsg101 levels also decreased during infection although to a lesser extent than Hrs (Figure 23C). One bacterial factor that has been incriminated in the past for the degradation of several host proteins, including vimentin (Paschen et al., 2008), is the protease CPAF. To test if CPAF was responsible for Hrs disappearance we used the cell line T-Rex 293T stably transfected with the protein CPAF under an inducible promoter. This cell line has been kindly given by Georg Häcker (Freiburg University). CPAF expression was induced by tetracycline treatment for different times, the cells were lysed in 8M urea buffer to block any CPAF activity during processing of the samples, and levels of Hrs, and vimentin as a positive control, were analyzed by WB (Figure 23D). As expected, vimentin was increasingly processed into short fragments over time. In contrast, Hrs levels remained stable showing that this protein is insensitive to CPAF proteolytic activity. In conclusion, Hrs and Tsg101 are two important proteins of the ESCRT machinery, whose levels decrease in cells infected by C. trachomatis after the mid-phase of infection, in a bacterial dependent and CPAF independent manner.

115

Figure 23: Chlamydia infection modulates Hrs and Tsg101 levels in a CPAF independent manner. For all the panels, HeLa cells were infected or not (NI) at a MOI of 5 for the indicated times before lysis in 8M urea buffer. Equal amount of proteins for each time point were analyzed by SDS-PAGE followed with transfer on a membrane and immunoblotting. A) Hrs levels in infected relative to non-infected cells. One representative experiment is shown, and the histogram shows the quantification on two independent experiments. For each time point the amount of Hrs and actin were quantified on a Typhoon, Hrs was normalized to equal actin amount, and compared to its level in the non-infected sample. B) Chloramphenicol was added or not 16 hpi and the samples were analyzed as described above. C) Same experiment as in B, with parallel analysis of the amount of Hrs and Tsg101 D) 293T cells were treated with tetracycline for the indicated time before cell lysis in 8M urea buffer and analysis by SDS-PAGE, transfer on a membrane and immunoblotting using antibodies against Hrs, and vimentin as control.

116 b. Perturbation of the ESCRT system does not affect chlamydial development

We next asked whether experimentally modulating the level of Hrs and Tsg101 could affect chlamydial development. We investigated first the effect of a depletion in Hrs or in Tsg101 on infection. Each protein was depleted using two different siRNA (Figure 24A). Immunoblotting of Hrs or Tsg101 revealed a depletion of at least 85% of each protein. HeLa cells were transfected with siRNA for 24 hours then infected with the fluorescent L2IncDGFP strain described in Article 2. Analysis of bacterial entry (MOI=10) was performed using microscopy coupled with the image analysis application ChlamEntry (Figure 24B). No difference in the efficiency of entry was observed in the cells depleted for Hrs or Tsg101 compared to control cells. Effect of Hrs or Tsg101 depletion on bacterial growth was next analyzed by flow cytometry (Figure 24C). Cells were infected at MOI=1 and bacterial load was measured 24 h later. No differences were observed between the different conditions. Also the infection rates were identical for all samples, confirming that depleting Hrs or Tsg101 has no impact on bacterial entry. Finally we measured the growth rate during the exponential phase of infection and observed no difference between the different conditions. In conclusion Hrs and Tsg101 are not required for bacterial growth in vitro. Although we observed a strong (>85 %) depletion of Hrs and Tsg101, it is possible that residual proteins are sufficient for chlamydial need, if any. We decided to use a dominant negative form of VPS4, which prevents the disassembly of the ESCRT machinery at the end of the process (Figure 24D). The construct is described as the most potent method to inhibit ESCRT-mediated processes. By looking at infection in cells transfected with this construct we wanted to rule out the possibility that Tsg101 or Hrs depletion were insufficient to abolish ESCRT-mediated processes, and to address whether a functional ESCRT machinery is required for chlamydial growth or not. HeLa cells were transfected with myc-tagged VPS4 constructs, wild type (WT) or dominant negative (DN, mutation K173 to Q), and infected four hours later for 24 h before fixation and analysis by flow cytometry. If anything, a slight increase in the percentage of infection was observed in the cells expressing myc-VPS4-DN compared to cells expressing the WT. Bacterial loads were similar in both condition. Thus functional ESCRT is not required for bacterial growth.

117

Figure 24: Modulation the ESCRT system does not affect Chlamydia infection. A) Hrs and Tsg101 depletion by siRNA measured by WB. HeLa cells were transfected for 24 h with two different siRNA for each target, lysed and the quantity of Hrs and Tsg101 in the samples was analyzed by immunoblotting (IB). Actin levels are shown as loading controls. B) Hrs or Tsg101 depletion does not impair Chlamydia entry. Twenty-four hours after siRNA treatment, HeLa cells were infected with L2IncDGFP for one hour and fixed with PFA 4%. Immunofluorescence staining of the bacteria outside the cells was performed using anti- MOMP antibody without permeabilising the cell. Pictures were analysed using the ICY software plugin Chlamentry (>70 bacteria analyzed per condition). Yellow harrow heads indicate entered bacteria. Bar scale = 5 "m. C) Hrs or Tsg101 depletion does not impact bacterial growth nor the formation of infectious particles. Twenty-four hourts after siRNA treatment, HeLa cells were infected by L2IncDGFP at a MOI of 1. Twenty-four hours later, one

118 half of the cells was fixed in PFA 2% for analysis of the infection rate by flow cytometry (black bars). The rest of the cells was gently lysed and lysates were used to re-infect fresh HeLa cells. Re-infection rates were measured by flow cytometry one day later. The bottom plot shows a measurement of the growth rate of L2IncDGFP in the indicated conditions, between 18 and 24 hpi. D) Impairment of the ESCRT system using VPS4 dominant negative (DN) does not affect Chlamydia development. HeLa cells were transfected with the WT or the DN form of VPS4 with a myc-tag, then infected with L2IncDGFP for 24 h. Cells were fixed in PFA 2%, permeabilized and stained with anti-myc antibody followed with Cy5-coupled secondary antibodies, and were analysed by flow cytometry (Cy5 = FL4 channel, GFP = FL1 channel). The average of three measurements is shown, T-Test show no significant difference in the infection rates between transfected and non-transfected samples.

c. Uptake of LC3 positive compartments in the inclusion is Tsg101 dependent

Several host compartments such as LDs, peroxisomes, Rab14 positive compartments were described inside the inclusion lumen. We hypothesize that their entry might involve the ESCRT machinery, in a process analogous to the formation of intraluminal vesicles in MVBs. We therefore looked for eukaryotic markers whose entry in the inclusion lumen might depend on the ESCRT machinery. Surprisingly, we observed that LC3 positive structures are visible inside most inclusions in control cells, but only in few inclusions in cells depleted of Tsg101 (Figure 25). These results are still preliminary and additional experiments will be required to confirm that LC3 positive structures enter the inclusion in an ESCRT-driven pathway. The nature of these structures also remains to be characterized.

Figure 25: LC3 positive structures are imported in the inclusion lumen in a Tsg101-dependent manner. HeLa cells were transfected for 24 h with siRNA control (SiCTRL) or with siRNA against Tsg101 then infected for 24 h before fixation in PFA 2%. The cells were permeabilzed and stained with mouse anti-LC3 and rabbit anti-CT529 antibodies, followed with Alexa488- coupled anti mouse and Cy3-coupled anti-rabbit secondary antibodies. DNA was stained with 0.5 µg/ml Hoechst 33342 (blue). Bar scale = 5 "m. Inclusions were manually examined for the presence or absence of LC3 staining in the lumen in three separate fieds (n> 25 cells analyzed).

119

DISCUSSION

120 DUF582 proteins are type III effectors of C. trachomatis Proteins with a DUF582 are only found in chlamydiae, and among them, only in pathogenic chlamydiae, not in their environmental relatives. Each pathogenic Chlamydia species counts four to five proteins sharing this domain of unknown function, which is more conserved between orthologs than between paralogs (Article 1). Most DUF582 domains are associated with a second domain located in the N-terminal parts of the proteins, which define three sub-families of DUF582 proteins. CT712 and its orthologs are only composed of the DUF582 and make a fourth DUF582 proteins subfamily. The N-terminal domains are specific for each sub-family and are not found in other bacteria. It seems likely that the DUF582 domain first appeared in the common ancestor to all pathogenic chlamydiae, and was duplicated several times before the divergence of the pathogenic species during their adaptation to different hosts. It is likely that these duplication events reflect different functions acquired by each sub-family of DUF582 proteins. In addition, it is also possible that the DUF582 domains themselves have conserved a common target. Based on this last hypothesis we undertook the functional study of the DUF582 proteins of C. trachomatis. Using specific antibodies against recombinant DUF582 proteins, we demonstrated that each member of the family is expressed at the mid and late phases of infection. These data are consistent with transcriptional data: mRNA for the corresponding genes are first detected at 8 hpi and increase in quantities until the end of the cycle (Belland et al., 2003). This expression pattern suggests that the proteins might accumulate in EBs, that start forming at the mid phase of the cycle. Indeed, CT619 and CT711 were detected in purified EBs in a proteomic study (Saka et al., 2011). This is confirmed by our data showing that all the DUF582 proteins are present in purified EBs. We demonstrated using the heterologous TTSS of Shigella flexneri that all the DUF582 proteins possess a signal at their N-terminal extremity that is recognized by type three secretion machineries. CT620 and CT621 were indeed observed in the inclusion lumen and host cytoplasm by immunofluorescence, confirming that they are type three effectors of C. trachomatis (Article 1). Other members of the family showed dotty staining inside the inclusion. We verified by immunofluorescence, that these sera recognized the protein against which they had been raised. However, their expression level is relatively low (assessed by western blot), and we cannot exclude that the signal we detect by immunofluorescence results from cross-reaction of the antibodies with other bacterial proteins. In spite of our efforts to purify these sera, we could still observe contaminating cross-reacting species by western blot. Thus, we cannot conclude for the moment as to the timing at which other DUF582 proteins,

121 and CT619 in particular, are secreted, and where they localize. We observed that transfected GFP-CT619 appears to be enriched at the plasma membrane, suggesting that it might associate with the inclusion membrane. With the recent advance in C. trachomatis manipulation we are developing epitope-tagged version of all DUF582 proteins, which should allow us to answer these important questions left open.

The DUF582 binds Hrs One classical strategy to find the function of a secreted protein is to screen for potential interactors among the human proteome. We performed a yeast-two-hybrid screen using both C-terminal (DUF582 domain) and N-terminal parts of CT619 and CT621 as baits, and a human placenta cDNA library as preys. Only one protein, Hrs, interacted with the two DUF582 domains tested. Still using the two-hybrid technology we confirmed the interaction between Hrs and the DUF582 domain for CT619, CT711 and CT712, and not for CT621. We were surprised not to detect an interaction between the DUF582 domain of CT621 and Hrs, because it is an exception among DUF582 domains and because it had been detected in the initial screen and by co-immunoprecipitation. However, in the screen the interaction had a low score, which might explain the discrepancy between the screen and the targeted two- hybrid assay. The targeted two-hybrid assay also showed an interaction of the N-terminal domains of CT619 and CT621 with Hrs. Only the N-terminal domain of CT621 had pointed to Hrs in the initial screen, and with a very weak confidence score. This last piece of data matches with our observation that Hrs interacted more strongly with the DUF582 domains than with any N- terminal domain. Also, the N-terminal domains of DUF582 proteins colocalized with myc- Hrs only very partially in HeLa cells in co-expression experiments (data not shown). In addition we did not confirm the interactions with the N-terminal domains through another method like co-immunoprecipitaion. Thus we think that the interactions detected between Hrs and the N-terminal domains of CT619 and CT621 need to be regarded cautiously. We found that removing the coiled coil region present in the DUF582 domain of CT619 did not abrogate the interaction with Hrs, indicating that this interaction does not rely only on coiled coil association. Surprisingly, GFP-CT619 did not co-immunoprecipitate with myc-Hrs while its DUF582 domain expressed alone did, and while the two proteins colocalized by immunofluorescence. There is no trivial explanation for these discrepancies. Expression of the DUF582 alone might expose the domain of interaction with Hrs better than in the full-length protein.

122 If we restrict ourselves to the functional study of CT712, all our results are consistent. Expressed with a GFP tag or without tag (not shown) in HeLa cells CT712 showed a high level of co-localization with Hrs. The interaction with Hrs was observed by two independent techniques, two hybrid in yeast and co-immunoprecipitation in mammalian cells. CT712, composed of the DUF582 alone, might represent the founding member of the family, that gave rise to multiple proteins by addition of N-terminal domains of varied forms. Thus our results are consistent with the DUF582 domain being an Hrs interacting domain.

CT619 interacts also with Tsg101 The initial two-hybrid screens indicated that the N-terminal part of CT619 interacted strongly with Tsg101 while that of CT621 did not. The minimum region for the interaction corresponds to the proline rich and the coiled coil domains of Tsg101. This was confirmed by co-immunoprecipitation experiments using different constructs of Tsg101, in transfected cells as well as in infected cells, with endogenous CT619. Analysis of different experiments revealed that the interaction takes place between the N-terminal part of CT619 and the coiled coil domain of Tsg101. Interaction between Tsg101 and CT619 in infected cells was observed in cells infected for 38 h. However, it might occur earlier without being detected detected due to limiting amounts of CT619 expressed. The ability for CT619 to interact with Hrs and Tsg101 via two distinct domains (respectively the N-terminal and the DUF582 domains) means that CT619 could potentially bind both proteins simultaneously, but this was not directly investigated.

Hrs and Tsg101 are dispensable for C. trachomatis development The levels of Hrs and Tsg101 decreased after 20 h of infection. We have shown this decrease to be dependent on Chlamydia activity but independent on CPAF proteolytic activity. The decrease in Hrs and Tsg101 levels correlates with the timing of expression of the DUF582 proteins. However, expression of the DUF582 proteins in mammalian cells did not affect Hrs nor Tsg101 levels (data not shown). Thus, there is no evidence that the decrease in Hrs and Tsg101 is linked in any way to the secretion of the DUF582 proteins. It is also useful to note that to monitor the effect of infection on Hrs and Tsg101 we worked at MOI=5, to reach 100% infection. With a MOI=1, which are the usual working conditions, disappearance of Hrs and Tsg101 would likely be delayed to a time that corresponds to the very end of the infectious cycle. It is thus not clear how their disappearance could benefit bacteria already largely engaged in RB to EB conversion. Interestingly, it was recently shown that many

123 proteins are stabilized or destabilized by infection (Olive et al., 2014). The mechanisms involved are not known. In conclusion, it is not clear at this point whether the decreases in Hrs and Tsg101 levels are specific, or if they reflect a more global change in protein homeostasis in the host. In infected cells Hrs sometimes localized in close proximity with the inclusion membrane, but did not appear to be enriched there. Hrs positive structures were slightly bigger in infected cells, which could be interpreted as aggregates or clusters of Hrs. However, co-staining with CT619 or CT712 did not show any co-localization (data not shown) so it might not be related to DUF582 protein secretion.

To explore what might be the benefit for Chlamydia to interact with ESCRT proteins we asked whether the depletion of these targets, using siRNA, might impact bacterial development. We adapted and improved existing techniques of microscopy coupled to image analysis and flow cytometry in order to quantify several steps of C. trachomatis developmental cycle. The advantages of these methods are discussed in the Article 2. Hrs and Tsg101 depletion had no effect on Chlamydia entry, development or infectious progeny. Experiments with the dominant negative form of VPS4 confirmed that a functional ESCRT machinery is not required for C. trachomatis LGV development in vitro. We used this strain because it is very efficient at infecting cells in vitro, and because fluorescent bacteria are available. The drawback is that this strain is less sensitive to suboptimal growth conditions than other serovars. However, preliminary experiments using serovar D indicate that Hrs or Tsg101 depletion does not affect growth of this serovar either. Since Hrs and Tsg101 levels decrease after 20 h of infection we tested whether overexpression of these proteins 24 h after infection had an effect on bacterial growth. Bacterial load measured at 48 h post infection was identical in transfected and non-transfected cells, suggesting that Hrs or Tsg101 overexpression does not impair late bacterial growth (data not shown).

Two main hypotheses could explain the fact that Hrs or Tsg101 depletion have no impact on C. trachomatis development, in spite of being targets of the DUF582 proteins: (i) the function of the DUF582 proteins is to prevent Hrs and/or Tsg101 mediated processes. In that case, depleting the proteins only mimics what the bacteria normally do and no phenotype is to be expected. If overexpression of Hrs and/or Tsg101 could in theory counteract the

124 DUF582 activity, it is difficult to obtain the right timing and balance of overexpression in these multi-subunits systems, and we might have failed to do so. (ii) Hrs and/or Tsg101 mediated processes (possibly ESCRT mediated processes) take place during bacterial development, but redundant mechanisms are also at work so that preventing their activity has no phenotype in vitro. One alternative for this hypothesis is that we did not look at the right phenotype. Indeed, it is possible that ESCRT driven processes are not required for bacterial growth but for dissemination for instance. In light of our results and the known functions or Hrs and Tsg101 we discuss below what the function of the DUF582 proteins might be.

Scenario 1: DUF582 proteins interact with Hrs to escape lysosomal degradation It is established that at least some of these receptors that chlamydiae use to bind to host cells are targeted to degradation after endocytosis. For instance C. trachomatis and C. pneumoniae enter the cell via PGDF receptor and the EGF receptor, respectively (Elwell et al., 2008; Mölleken et al., 2013). Both receptors are degraded via the endo-lysosomal pathway (Er, Mendoza, Mackey, Rameh, & Blenis, 2013), which is initiated by the Hrs recognition of the poly-ubiquitinated activated receptor. Absence of EEA1 staining as early as 5 min after the entry suggests a fast escape from this pathway (Scidmore et al., 2003). Bacterial escape from the degradation pathway requires de novo protein synthesis since chloramphenicol treated EBs eventually acquire lysosomal markers, but only very slowly (Scidmore et al., 2003). In addition, we were able to visualize endogenous Hrs in close contact with invasive bacteria 30 min after infection (data not shown). This was only observed in a minority of events, suggesting that Hrs recruitment to the nascent inclusion might only be transient. Collectively these data suggest that bacterial factors pre-loaded in the infectious EBs might be sufficient to prevent early escape from the lysosomal pathway. All the DUF582 proteins are detected in purified EBs, and could play a role at this step. In particular, CT619’s ability to associate with membranes might retain it to the nascent inclusion membrane, reaching a local concentration sufficient to bind Hrs and Tsg101 and prevent the recruitment of the ESCRT machinery. The silencing experiment would only reproduce what occurs normally, explaining the absence of phenotype on infection.

Scenario 2: DUF582 proteins interact with Hrs to acquire material from the host If Chlamydia imports many host constituents inside the inclusion, the mechanisms involved are largely unknown. Videomicroscopy by the Valdivia laboratory looking at lipid droplet import in the inclusion is suggesting of a vesicle mediated transport with inward

125 invagination of the inclusion membrane topologically similar to the ESCRT dependent ILV formation in MVBs (Cocchiaro et al., 2008). Importantly, there are clear redundant pathways for nutrient acquisition by the chlamydial inclusion. Thus, blocking one ESCRT-driven mechanism would not necessarily impact bacterial growth, making this hypothesis compatible with our results. We tested this hypothesis directly by comparing the distribution of some host markers in control cells and cells depleted for Hrs or Tsg101. Lipid droplet import was not observed in our hands, we can thus not conclude as to the involvement of the ESCRT machinery in that process. Similarly, we did not observe CD63 nor LBPA import in the inclusion lumen, in contrast to reported results (Beatty, 2006; 2008). Finally, peroxisome import (Boncompain et al., 2014) is difficult to quantify but appears to proceed normally in Hrs or Tsg101 depleted cells. Our preliminary data indicate that LC3 positive structures reach the inclusion lumen in a Tsg101 dependent manner (LC3 staining was still observed in Hrs-depleted cells, but this could be due to a weaker effect of the depletion on the ESCRT driven process. In our hands, Hrs depletion had less impact than Tsg101 depletion on EGF receptor traffic for instance). LC3 positive structures were not autophagosomes since they were also observed in inclusions formed in ATG5-/- cells (data not shown). This staining remains thus to be confirmed (it is to note that a previous report did not describe LC3 staining in the inclusion lumen (Al-Younes et al., 2011), and if so, the nature of the structures will be determined.

Scenario 3: CT619 interacts with Tsg101 for bacterial exit through extrusion As discussed earlier, it is possible that we failed to detect a requirement for the ESCRT machinery in chlamydial development because we did not look at the right phenotype. We have measured bacterial growth and infectious progeny, but we did not study the exit process itself. Chlamydiae exit cells through two pathways, one being extrusion (Hybiske & Stephens, 2007a). This exit pathway seems to have many similarities with cytokinesis, which requires Tsg101 activity. Tsg101 is recruited by CEP55 before abscission to trigger ESCRT- III recruitment (Agromayor & Martin-Serrano, 2013). Proteins implicated in cytokinesis (i.e. MLCK and myosin II) (Deschamps, Echard, & Niedergang, 2013) have been found to be determinant for Chlamydia extrusion and their depletion decrease the number of extrusion events. In addition, the extrusion event ends by an abscission-like process as illustrated in the video n°6 recorded by Hybiske and colleagues (Hybiske & Stephens, 2007a). Thus, secretion of CT619 late in the cycle could recruit Tsg101 to engage exit through the extrusion pathway.

126 The decrease in Tsg101 levels does not fit well with this hypothesis. However we never observed total disappearance of Tsg101 during infection. Thus, a small quantity of Tsg101, locally recruited by CT619, might be enough to trigger extrusion.

Conclusion Our results are consistent with at least three possible functions for the DUF582 proteins in infection. These functions are not mutually exclusive, on the contrary it is likely that the different proteins of the family are implicated in distinct mechanisms. The three hypotheses described above are currently being tested, in particular using the very recently developed tools to express tagged chlamydial effectors (Bauler & Hackstadt, 2014), (Agaisse & Derré, 2014) and to obtain null mutants (C. M. Johnson & Fisher, 2013). Importantly this last strategy will only succeed for non-essential genes, we do not know if it will be the case for our genes of interest. Altogether, we have shown that the DUF582 proteins are type III effectors of pathogenic chlamydiae, and that their common domain targets Hrs, and likely that the bacteria manipulate ESCRT-driven processes. Considering the ancient history of intracellular parasitism of chlamydiae it is maybe not surprising that they acquired sophisticated tools to interact with a very ancestral machinery, implicated in several essential functions of their eukaryotic host.

127

BIBLIOGRAPHY

128 Abdelrahman, Y. M., & Belland, R. J. (2005). The chlamydial developmental cycle. FEMS Microbiology Reviews, 29(5), 949–959. doi:10.1016/j.femsre.2005.03.002

Abdelrahman, Y. M., Rose, L. A., & Belland, R. J. (2011). Developmental expression of non- coding RNAs in Chlamydia trachomatis during normal and persistent growth. Nucleic Acids Research, 39(5), 1843–1854. doi:10.1093/nar/gkq1065

Abromaitis, S., & Stephens, R. S. (2009). Attachment and Entry of Chlamydia Have Distinct Requirements for Host Protein Disulfide Isomerase. PLoS Pathogens, 5(4), e1000357. doi:10.1371/journal.ppat.1000357.s003

Agaisse, H., & Derré, I. (2013). A C. trachomatis Cloning Vector and the Generation of C. trachomatis Strains Expressing Fluorescent Proteins under the Control of a C. trachomatis Promoter. PLoS ONE, 8(2), e57090. doi:10.1371/journal.pone.0057090.s011

Agaisse, H., & Derré, I. (2014). The expression of the effector protein IncD in C. trachomatis mediates the recruitment of the lipid transfer protein CERT and the ER-resident protein VAPB to the inclusion membrane. Infection and Immunity. doi:10.1128/IAI.01530-14

Agromayor, M., & Martin-Serrano, J. (2013). Knowing when to cut and run: mechanisms that control cytokinetic abscission. Trends in Cell Biology, 23(9), 433–441. doi:10.1016/j.tcb.2013.04.006

Al-Younes, H. M., Al-Zeer, M. A., Khalil, H., Gussmann, J., Karlas, A., Machuy, N., et al. (2011). Autophagy-independent function of MAP-LC3 during intracellular propagation of Chlamydia trachomatis. Autophagy, 7(8), 814–828. doi:10.4161/auto.7.8.15597

Asao, H., Sasaki, Y., Arita, T., Tanaka, N., Endo, K., Kasai, H., et al. (1997). Hrs is associated with STAM, a signal-transducing adaptor molecule. Its suppressive effect on cytokine-induced cell growth. The Journal of Biological Chemistry, 272(52), 32785– 32791.

Babst, M. (2005). A protein's final ESCRT. Traffic, 6(1), 2–9. doi:10.1111/j.1600- 0854.2005.00246.x

Babst, M., Katzmann, D. J., Estepa-Sabal, E. J., Meerloo, T., & Emr, S. D. (2002a). Escrt-III: an endosome-associated heterooligomeric protein complex required for mvb sorting. Developmental Cell, 3(2), 271–282.

Babst, M., Katzmann, D. J., Snyder, W. B., Wendland, B., & Emr, S. D. (2002b). Endosome- associated complex, ESCRT-II, recruits transport machinery for protein sorting at the multivesicular body. Developmental Cell, 3(2), 283–289.

Bache, K. G. (2003). Hrs regulates multivesicular body formation via ESCRT recruitment to endosomes. The Journal of Cell Biology, 162(3), 435–442. doi:10.1083/jcb.200302131

Balañá, M.-E., Niedergang, F., Subtil, A., Alcover, A., Chavrier, P., & Dautry-Varsat, A. (2005). ARF6 GTPase controls bacterial invasion by actin remodelling. Journal of Cell Science, 118(Pt 10), 2201–2210. doi:10.1242/jcs.02351

129 Balsara, Z. R., Misaghi, S., Lafave, J. N., & Starnbach, M. N. (2006). Chlamydia trachomatis infection induces cleavage of the mitotic cyclin B1. Infection and Immunity, 74(10), 5602–5608. doi:10.1128/IAI.00266-06

Bannantine, J. P., Griffiths, R. S., Viratyosin, W., Brown, W. J., & Rockey, D. D. (2000). A secondary structure motif predictive of protein localization to the chlamydial inclusion membrane. Cellular Microbiology, 2(1), 35–47.

Barker, J. R., Koestler, B. J., Carpenter, V. K., Burdette, D. L., Waters, C. M., Vance, R. E., & Valdivia, R. H. (2013). STING-Dependent Recognition of Cyclic di-AMP Mediates Type I Interferon Responses during Chlamydia trachomatis Infection. mBio, 4(3), e00018–13–e00018–13. doi:10.1128/mBio.00018-13

Barry, C. E., Brickman, T. J., & Hackstadt, T. (1993). Hc1-mediated effects on DNA structure: a potential regulator of chlamydial development. Molecular Microbiology, 9(2), 273–283. doi:10.1111/j.1365-2958.1993.tb01689.x

Bastidas, R. J., Elwell, C. A., Engel, J. N., & Valdivia, R. H. (2013). Chlamydial Intracellular Survival Strategies. Cold Spring Harbor Perspectives in Medicine, 3(5), a010256– a010256. doi:10.1101/cshperspect.a010256

Batteiger, B. E., Tu, W., Ofner, S., Van Der Pol, B., Stothard, D. R., Orr, D. P., et al. (2010). Repeated Chlamydia trachomatis Genital Infections in Adolescent Women. The Journal of Infectious Diseases, 201(1), 42–51. doi:10.1086/648734

Bauler, L. D., & Hackstadt, T. (2014). Expression and Targeting of Secreted Proteins from Chlamydia trachomatis. Journal of Bacteriology, 196(7), 1325–1334. doi:10.1128/JB.01290-13

Beatty, W. L. (2006). Trafficking from CD63-positive late endocytic multivesicular bodies is essential for intracellular development of Chlamydia trachomatis. Journal of Cell Science, 119(Pt 2), 350–359. doi:10.1242/jcs.02733

Beatty, W. L. (2008). Late Endocytic Multivesicular Bodies Intersect the Chlamydial Inclusion in the Absence of CD63. Infection and Immunity, 76(7), 2872–2881. doi:10.1128/IAI.00129-08

Bedson, S. P. (1936). Observation bearing on the antigenic composition of the psittacosis virus. Journal of Experimental Pathology.

Bedson, S. P., & Bland, J. (1932). A morphological study of psittacosis virus, with the description of a developmental cycle. British Journal of Experimental Pathology, 13(5), 461.

Belland, R. J., Zhong, G., Crane, D. D., Hogan, D., Sturdevant, D., Sharma, J., et al. (2003). Genomic transcriptional profiling of the developmental cycle of Chlamydia trachomatis. Proceedings of the National Academy of Sciences of the United States of America, 100(14), 8478–8483. doi:10.1073/pnas.1331135100

Betts-Hampikian, H. J., & Fields, K. A. (2010). The Chlamydial Type III Secretion

130 Mechanism: Revealing Cracks in a Tough Nut. Frontiers in Microbiology, 1, 114. doi:10.3389/fmicb.2010.00114

Betts-Hampikian, H. J., & Fields, K. A. (2011). Disulfide bonding within components of the Chlamydia type III secretion apparatus correlates with development. Journal of Bacteriology, 193(24), 6950–6959. doi:10.1128/JB.05163-11

Bissig, C., & Gruenberg, J. (2014). ALIX and the multivesicular endosome: ALIX in Wonderland. Trends in Cell Biology, 24(1), 19–25. doi:10.1016/j.tcb.2013.10.009

Boleti, H., Benmerah, A., Ojcius, D. M., Cerf-Bensussan, N., & Dautry-Varsat, A. (1999). Chlamydia infection of epithelial cells expressing dynamin and Eps15 mutants: clathrin- independent entry into cells and dynamin-dependent productive growth. Journal of Cell Science, 112 ( Pt 10), 1487–1496.

Boncompain, G., Müller, C., Meas-Yedid, V., Schmitt-Kopplin, P., Lazarow, P. B., & Subtil, A. (2014). The Intracellular Bacteria Chlamydia Hijack Peroxisomes and Utilize Their Enzymatic Capacity to Produce Bacteria-Specific Phospholipids. PLoS ONE, 9(1), e86196. doi:10.1371/journal.pone.0086196.s002

Borriello, S. P., Murray, P. R., & Funke, G. (Eds.). (2005). Topley and Wilson's principles of bacteriology, virology and immunity (10 ed., Vol. 2, pp. 2006–2025).

Bouamr, F., Houck-Loomis, B. R., De Los Santos, M., Casaday, R. J., Johnson, M. C., & Goff, S. P. (2007). The C-terminal portion of the Hrs protein interacts with Tsg101 and interferes with human immunodeficiency virus type 1 Gag particle production. Journal of Virology, 81(6), 2909–2922. doi:10.1128/JVI.01413-06

Brinkworth, A. J., Malcolm, D. S., Pedrosa, A. T., Roguska, K., Shahbazian, S., Graham, J. E., et al. (2011). Chlamydia trachomatis Slc1 is a type III secretion chaperone that enhances the translocation of its invasion effector substrate TARP. Molecular Microbiology, 82(1), 131–144. doi:10.1111/j.1365-2958.2011.07802.x

Brown, H. M., Knowlton, A. E., & Grieshaber, S. S. (2012). Chlamydial infection induces host cytokinesis failure at abscission. Cellular Microbiology, 14(10), 1554–1567. doi:10.1111/j.1462-5822.2012.01820.x

Bulut, Y., Shimada, K., Wong, M. H., Chen, S., Gray, P., Alsabeh, R., et al. (2009). Chlamydial Heat Shock Protein 60 Induces Acute Pulmonary Inflammation in Mice via the Toll-Like Receptor 4- and MyD88-Dependent Pathway. Infection and Immunity, 77(7), 2683–2690. doi:10.1128/IAI.00248-09

Capmany, A., & Damiani, M. T. (2010). Chlamydia trachomatis Intercepts Golgi-Derived Sphingolipids through a Rab14-Mediated Transport Required for Bacterial Development and Replication. PLoS ONE, 5(11), e14084. doi:10.1371/journal.pone.0014084.g008

Carabeo, R. (2011). Bacterial subversion of host actin dynamics at the plasma membrane. Cellular Microbiology, 13(10), 1460–1469. doi:10.1111/j.1462-5822.2011.01651.x

Carabeo, R. A., & Hackstadt, T. (2001). Isolation and Characterization of a Mutant Chinese

131 Hamster Ovary Cell Line That Is Resistant to Chlamydia trachomatis Infection at a Novel Step in the Attachment Process. Infection and Immunity, 69(9), 5899–5904. doi:10.1128/IAI.69.9.5899-5904.2001

Carabeo, R. A., Grieshaber, S. S., Hasenkrug, A., Dooley, C., & Hackstadt, T. (2004). Requirement for the Rac GTPase in Chlamydia trachomatis invasion of non-phagocytic cells. Traffic, 5(6), 418–425. doi:10.1111/j.1398-9219.2004.00184.x

Carabeo, R. A., Mead, D. J., & Hackstadt, T. (2003). Golgi-dependent transport of cholesterol to the Chlamydia trachomatis inclusion. Proceedings of the National Academy of Sciences of the United States of America, 100(11), 6771–6776. doi:10.1073/pnas.1131289100

Chen, A. L., Johnson, K. A., Lee, J. K., Sütterlin, C., & Tang, F. F. (2012). CPAF: A Chlamydial Protease in Search of an Authentic Substrate. PLoS Pathogens, 8(8), e1002842. doi:10.1371/journal.ppat.1002842.t001

Chen, C.-Y., Chi, K.-H., Alexander, S., Martin, I. M. C., Liu, H., Ison, C. A., & Ballard, R. C. (2006). The Molecular Diagnosis of Lymphogranuloma Venereum: Evaluation of a Real- Time Multiplex Polymerase Chain Reaction Test Using Rectal and Urethral Specimens. Sexually Transmitted Diseases, PAP. doi:10.1097/01.olq.0000245957.02939.ea

Chen, D., Lei, L., Lu, C., Flores, R., DeLisa, M. P., Roberts, T. C., et al. (2010). Secretion of the chlamydial virulence factor CPAF requires the Sec-dependent pathway. Microbiology, 156(10), 3031–3040. doi:10.1099/mic.0.040527-0

Chen, Y.-S., Bastidas, R. J., Saka, H. A., Carpenter, V. K., Richards, K. L., Plano, G. V., & Valdivia, R. H. (2014). The Chlamydia trachomatis Type III Secretion Chaperone Slc1 Engages Multiple Early Effectors, Including TepP, a Tyrosine-phosphorylated Protein Required for the Recruitment of CrkI-II to Nascent Inclusions and Innate Immune Signaling. PLoS Pathogens, 10(2), e1003954. doi:10.1371/journal.ppat.1003954.s012

Chin, L. S., Raynor, M. C., Wei, X., Chen, H. Q., & Li, L. (2001). Hrs interacts with sorting nexin 1 and regulates degradation of epidermal growth factor receptor. The Journal of Biological Chemistry, 276(10), 7069–7078. doi:10.1074/jbc.M004129200

Clausen, J. D., Christiansen, G., Holst, H. U., & Birkelund, S. (1997). Chlamydia trachomatis utilizes the host cell microtubule network during early events of infection. Molecular Microbiology, 25(3), 441–449.

Clifton, D. R., Dooley, C. A., Grieshaber, S. S., Carabeo, R. A., Fields, K. A., & Hackstadt, T. (2005). Tyrosine phosphorylation of the chlamydial effector protein Tarp is species specific and not required for recruitment of actin. Infection and Immunity, 73(7), 3860– 3868.

Clifton, D. R., Fields, K. A., Grieshaber, S. S., Dooley, C. A., Fischer, E. R., Mead, D. J., et al. (2004). A chlamydial type III translocated protein is tyrosine-phosphorylated at the site of entry and associated with recruitment of actin. Proceedings of the National Academy of Sciences of the United States of America, 101(27), 10166–10171. doi:10.1073/pnas.0402829101

132

Cocchiaro, J. L., Kumar, Y., Fischer, E. R., Hackstadt, T., & Valdivia, R. H. (2008). Cytoplasmic lipid droplets are translocated into the lumen of the Chlamydia trachomatis parasitophorous vacuole. Proceedings of the National Academy of Sciences of the United States of America, 105(27), 9379–9384.

Collier, L., Balows, A., & Sussman, M. (1998). Topley and Wilson's principles of bacteriology, virology and immunity. (W. J. Hausler & M. Sussman, Eds.) (9 ed., Vol. 3, pp. 977–1011).

Collingro, A., Tischler, P., Weinmaier, T., Penz, T., Heinz, E., Brunham, R. C., et al. (2011). Unity in Variety--The Pan-Genome of the Chlamydiae. Molecular Biology and Evolution, 28(12), 3253–3270. doi:10.1093/molbev/msr161

Colombo, M., Moita, C., van Niel, G., Kowal, J., Vigneron, J., Benaroch, P., et al. (2013). Analysis of ESCRT functions in exosome biogenesis, composition and secretion highlights the heterogeneity of extracellular vesicles. Journal of Cell Science, 126(Pt 24), 5553–5565. doi:10.1242/jcs.128868

Conant, C. G., & Stephens, R. S. (2007). Chlamydia attachment to mammalian cells requires protein disulfide isomerase. Cellular Microbiology, 9(1), 222–232. doi:10.1111/j.1462- 5822.2006.00783.x

Coombes, B. K., & Mahony, J. B. (2002). Identification of MEK- and phosphoinositide 3- kinase-dependent signalling as essential events during Chlamydia pneumoniae invasion of HEp2 cells. Cellular Microbiology, 4(7), 447–460.

Cox, J. V., Naher, N., Abdelrahman, Y. M., & Belland, R. J. (2012). Host HDL biogenesis machinery is recruited to the inclusion of Chlamydia trachomatis-infected cells and regulates chlamydial growth. Cellular Microbiology, 14(10), 1497–1512. doi:10.1111/j.1462-5822.2012.01823.x de Chaumont, F., Dallongeville, S., Chenouard, N., Hervé, N., Pop, S., Provoost, T., et al. (2012). Icy: an open bioimage informatics platform for extended reproducible research. Nature Methods, 9(7), 690–696. doi:10.1038/nmeth.2075

Dean, P. (2011). Functional domains and motifs of bacterial type III effector proteins and their roles in infection. FEMS Microbiology Reviews, 35(6), 1100–1125. doi:10.1111/j.1574-6976.2011.00271.x

Dehoux, P., Flores, R., Dauga, C., Zhong, G., & Subtil, A. (2011). Multi-genome identification and characterization of chlamydiae-specific type III secretion substrates: the Inc proteins. BMC Genomics, 12, 109. doi:10.1186/1471-2164-12-109

Delevoye, C., Nilges, M., Dehoux, P., Paumet, F., Perrinet, S., Dautry-Varsat, A., & Subtil, A. (2008). SNARE protein mimicry by an intracellular bacterium. PLoS Pathogens, 4(3), e1000022. doi:10.1371/journal.ppat.1000022

Derré, I., Swiss, R., & Agaisse, H. (2011). The lipid transfer protein CERT interacts with the Chlamydia inclusion protein IncD and participates to ER-Chlamydia inclusion membrane

133 contact sites. PLoS Pathogens, 7(6), e1002092. doi:10.1371/journal.ppat.1002092.s016

Deschamps, C., Echard, A., & Niedergang, F. (2013). Phagocytosis and cytokinesis: do cells use common tools to cut and to eat? Highlights on common themes and differences. Traffic, 14(4), 355–364. doi:10.1111/tra.12045

Dobro, M. J., Samson, R. Y., Yu, Z., McCullough, J., Ding, H. J., Chong, P. L.-G., et al. (2013). Electron cryotomography of ESCRT assemblies and dividing Sulfolobus cells suggests that spiraling filaments are involved in membrane scission. Molecular Biology of the Cell, 24(15), 2319–2327. doi:10.1091/mbc.E12-11-0785

Du, X., Kazim, A. S., Brown, A. J., & Yang, H. (2012). An Essential Role of Hrs/Vps27 in Endosomal Cholesterol Trafficking. Cell Reports, 1(1), 29–35. doi:10.1016/j.celrep.2011.10.004

Dumoux, M., Clare, D. K., Saibil, H. R., & Hayward, R. D. (2012). Chlamydiae assemble a pathogen synapse to hijack the host endoplasmic reticulum. Traffic, 13(12), 1612–1627. doi:10.1111/tra.12002

Ellis, T. N., & Kuehn, M. J. (2010). Virulence and Immunomodulatory Roles of Bacterial Outer Membrane Vesicles. Microbiology and Molecular Biology Reviews, 74(1), 81–94. doi:10.1128/MMBR.00031-09

Elwell, C. A., & Engel, J. N. (2012). Lipid acquisition by intracellular Chlamydiae. Cellular Microbiology, 14(7), 1010–1018. doi:10.1111/j.1462-5822.2012.01794.x

Elwell, C. A., Ceesay, A., Kim, J. H., Kalman, D., & Engel, J. N. (2008). RNA Interference Screen Identifies Abl Kinase and PDGFR Signaling in Chlamydia trachomatis Entry. PLoS Pathogens, 4(3), e1000021. doi:10.1371/journal.ppat.1000021.g010

Elwell, C. A., Jiang, S., Kim, J. H., Lee, A., Wittmann, T., Hanada, K., et al. (2011). Chlamydia trachomatis Co-opts GBF1 and CERT to Acquire Host Sphingomyelin for Distinct Roles during Intracellular Development. PLoS Pathogens, 7(9), e1002198. doi:10.1371/journal.ppat.1002198.s008

Emerson, P. M., Lindsay, S. W., Alexander, N., Bah, M., Dibba, S.-M., Faal, H. B., et al. (2004). Role of flies and provision of latrines in trachoma control: cluster-randomised controlled trial. The Lancet, 363(9415), 1093–1098. doi:10.1016/S0140-6736(04)15891-1

Er, E. E., Mendoza, M. C., Mackey, A. M., Rameh, L. E., & Blenis, J. (2013). AKT facilitates EGFR trafficking and degradation by phosphorylating and activating PIKfyve. Science Signaling, 6(279), ra45. doi:10.1126/scisignal.2004015

Ettema, T. J., & Bernander, R. (2009). Cell division and the ESCRT complex: A surprise from the archaea. Communicative & Integrative Biology, 2(2), 86–88.

Field, M. C., Sali, A., & Rout, M. P. (2011). On a bender—BARs, ESCRTs, COPs, and finally getting your coat. The Journal of Cell Biology, 193(6), 963–972. doi:10.1016/j.ceb.2010.10.005

134 Fields, K. A., & Hackstadt, T. (2000). Evidence for the secretion of Chlamydia trachomatis CopN by a type III secretion mechanism. Molecular Microbiology, 38(5), 1048–1060.

Fields, K. A., Mead, D. J., Dooley, C. A., & Hackstadt, T. (2003). Chlamydia trachomatis type III secretion: evidence for a functional apparatus during early-cycle development. Molecular Microbiology, 48(3), 671–683.

Filimonenko, M., Stuffers, S., Raiborg, C., Yamamoto, A., Malerod, L., Fisher, E. M. C., et al. (2007). Functional multivesicular bodies are required for autophagic clearance of protein aggregates associated with neurodegenerative disease. The Journal of Cell Biology, 179(3), 485–500. doi:10.1083/jcb.200702115

Freney, J., Renaud, F., Hansen, W., & Bollet, C. (Eds.). (1994). Manuel de Bactériologie Cllinique (2nd ed., Vol. 3).

Gasparrini, F., Molfetta, R., Quatrini, L., Frati, L., Santoni, A., & Paolini, R. (2012). Syk- dependent regulation of Hrs phosphorylation and ubiquitination upon Fc%RI engagement: Impact on Hrs membrane/cytosol localization. European Journal of Immunology, 42(10), 2744–2753. doi:10.1002/eji.201142278

Giles, D. K., Whittimore, J. D., LaRue, R. W., Raulston, J. E., & Wyrick, P. B. (2006). Ultrastructural analysis of chlamydial antigen-containing vesicles everting from the Chlamydia trachomatis inclusion. Microbes and Infection, 8(6), 1579–1591. doi:10.1016/j.micinf.2006.01.018

Gordon, F. B., Quan, A. L., & Trimmer, R. W. (1960). Morphologic Observations on Trachoma Virus in Cell Cultures. Science (New York, NY).

Greene, W., & Zhong, G. (2003). Inhibition of host cell cytokinesis by Chlamydia trachomatis infection. Journal of Infection, 47(1), 45–51. doi:10.1016/S0163- 4453(03)00039-2

Grieshaber, N. A., Fischer, E. R., Mead, D. J., Dooley, C. A., & Hackstadt, T. (2004). Chlamydial histone-DNA interactions are disrupted by a metabolite in the methylerythritol phosphate pathway of isoprenoid biosynthesis. Proceedings of the National Academy of Sciences of the United States of America, 101(19), 7451–7456. doi:10.1073/pnas.0400754101

Grieshaber, N. A., Grieshaber, S. S., Fischer, E. R., & Hackstadt, T. (2005). A small RNA inhibits translation of the histone-like protein Hc1 in Chlamydia trachomatis. Molecular Microbiology, 59(2), 541–550. doi:10.1111/j.1365-2958.2005.04949.x

Grieshaber, N. A., Sager, J. B., Dooley, C. A., Hayes, S. F., & Hackstadt, T. (2006). Regulation of the Chlamydia trachomatis histone H1-like protein Hc2 is IspE dependent and IhtA independent. Journal of Bacteriology, 188(14), 5289–5292. doi:10.1128/JB.00526-06

Grieshaber, S. S., Grieshaber, N. A., & Hackstadt, T. (2003). Chlamydia trachomatis uses host cell dynein to traffic to the microtubule-organizing center in a p50 dynamitin- independent process. Journal of Cell Science, 116(Pt 18), 3793–3802.

135 doi:10.1242/jcs.00695

Hackstadt, T., Rockey, D. D., Heinzen, R. A., & Scidmore, M. A. (1996). Chlamydia trachomatis interrupts an exocytic pathway to acquire endogenously synthesized sphingomyelin in transit from the Golgi apparatus to the plasma membrane. The EMBO Journal, 15(5), 964.

Hackstadt, T., Scidmore, M. A., & Rockey, D. D. (1995). Lipid metabolism in Chlamydia trachomatis-infected cells: directed trafficking of Golgi-derived sphingolipids to the chlamydial inclusion. Proceedings of the National Academy of Sciences of the United States of America, 92(11), 4877–4881.

Hackstadt, T., Scidmore-Carlson, M. A., Shaw, E. I., & Fischer, E. R. (1999). The Chlamydia trachomatis IncA protein is required for homotypic vesicle fusion. Cellular Microbiology, 1(2), 119–130.

Hafner, L. M., Wilson, D. P., & Timms, P. (2014). Development status and future prospects for a vaccine against Chlamydia trachomatis infection. Vaccine, 32(14), 1563–1571. doi:10.1016/j.vaccine.2013.08.020

Hanyaloglu, A. C., McCullagh, E., & Zastrow, von, M. (2005). Essential role of Hrs in a recycling mechanism mediating functional resensitization of cell signaling. The EMBO Journal, 24(13), 2265–2283. doi:10.1038/sj.emboj.7600688

Hatch, G. M., & McClarty, G. (1998). Phospholipid composition of purified Chlamydia trachomatis mimics that of the eucaryotic host cell. Infection and Immunity, 66(8), 3727– 3735.

Hatch, T. P. (1975). Utilization of L-cell nucleoside triphosphates by Chlamydia psittaci for ribonucleic acid synthesis. Journal of Bacteriology, 122(2), 393–400.

Hatch, T. P., Al-Hossainy, E., & Silverman, J. A. (1982). Adenine nucleotide and lysine transport in Chlamydia psittaci. Journal of Bacteriology, 150(2), 662–670.

Hayakawa, A. (2003). Structural Basis for Endosomal Targeting by FYVE Domains. Journal of Biological Chemistry, 279(7), 5958–5966. doi:10.1074/jbc.M310503200

Hefty, P. S., & Stephens, R. S. (2007). Chlamydial type III secretion system is encoded on ten operons preceded by sigma 70-like promoter elements. Journal of Bacteriology, 189(1), 198–206. doi:10.1128/JB.01034-06

Heine, H., Muller-Loennies, S., Brade, L., Lindner, B., & Brade, H. (2003). Endotoxic activity and chemical structure of lipopolysaccharides from Chlamydia trachomatis serotypes E and L2 and Chlamydophila psittaci 6BC. European Journal of Biochemistry / FEBS, 270(3), 440–450. doi:10.1046/j.1432-1033.2003.03392.x

Henne, W. M., Buchkovich, N. J., & Emr, S. D. (2011). The ESCRT Pathway. Developmental Cell, 21(1), 77–91. doi:10.1016/j.devcel.2011.05.015

Heuer, D., Rejman Lipinski, A., Machuy, N., Karlas, A., Wehrens, A., Siedler, F., et al.

136 (2009). Chlamydia causes fragmentation of the Golgi compartment to ensure reproduction. Nature, 457(7230), 731–735. doi:10.1038/nature07578

Hirano, S., Kawasaki, M., Ura, H., Kato, R., Raiborg, C., Stenmark, H., & Wakatsuki, S. (2006). Double-sided ubiquitin binding of Hrs-UIM in endosomal protein sorting. Nature Structural; Molecular Biology, 13(3), 272–277. doi:10.1038/nsmb1051

Ho, T. D., & Starnbach, M. N. (2005). The Salmonella enterica serovar typhimurium-encoded type III secretion systems can translocate Chlamydia trachomatis proteins into the cytosol of host cells. Infection and Immunity, 73(2), 905–911. doi:10.1128/IAI.73.2.905- 911.2005

Hogan, R. J., Mathews, S. A., Mukhopadhyay, S., Summersgill, J. T., & Timms, P. (2004). Chlamydial Persistence: beyond the Biphasic Paradigm. Infection and Immunity, 72(4), 1843–1855. doi:10.1128/IAI.72.4.1843-1855.2004

Horn, M., Collingro, A., Schmitz-Esser, S., Beier, C. L., Purkhold, U., Fartmann, B., et al. (2004). Illuminating the evolutionary history of chlamydiae. Science (New York, NY), 304(5671), 728–730. doi:10.1126/science.1096330

Hu, V. H., Harding-Esch, E. M., Burton, M. J., Bailey, R. L., Kadimpeul, J., & Mabey, D. C. W. (2010). Epidemiology and control of trachoma: systematic review. Tropical Medicine & International Health, 15(6), 673–691. doi:10.1111/j.1365-3156.2010.02521.x

Hurley, J. H. (2010). The ESCRT complexes. Critical Reviews in Biochemistry and Molecular Biology, 45(6), 463–487. doi:10.3109/10409238.2010.502516

Hurley, J. H., & Hanson, P. I. (2010). Membrane budding and scission by the ESCRT machinery: it's all in the neck. Nature Reviews Molecular Cell Biology, 11(8), 556–566. doi:10.1038/nrm2937

Hybiske, K., & Stephens, R. S. (2007a). Mechanisms of Chlamydia trachomatis Entry into Nonphagocytic Cells. Infection and Immunity, 75(8), 3925–3934. doi:10.1128/IAI.00106- 07

Hybiske, K., & Stephens, R. S. (2007b). Mechanisms of host cell exit by the intracellular bacterium Chlamydia. Proceedings of the National Academy of Sciences of the United States of America, 104(27), 11430–11435. doi:10.1073/pnas.0703218104

Iliffe-Lee, E. R., & McClarty, G. (1999). Glucose metabolism in Chlamydia trachomatis: the “energy parasite” hypothesis revisited. Molecular Microbiology, 33(1), 177–187.

Jamison, W. P., & Hackstadt, T. (2008). Induction of type III secretion by cell-free Chlamydia trachomatis elementary bodies. Microbial Pathogenesis, 45(5-6), 435–440. doi:10.1016/j.micpath.2008.10.002

>.$(1! D01! ! E0! E0! F'+.%1! E0! E0! "234250! G::%H'($&':(! I.&=..(! JEK! L6M$;.! $(H! N7:;N7:'(:;'&'H.!%.+,/$&':(!$(H!9,(-&':(;1!!"#$%&'$()*$+&**$,-)*!!"1!OPQROC30

Jewett, T. J., Dooley, C. A., Mead, D. J., & Hackstadt, T. (2008). Chlamydia trachomatis tarp

137 is phosphorylated by src family tyrosine kinases. Biochemical and Biophysical Research Communications, 371(2), 339–344. doi:10.1016/j.bbrc.2008.04.089

Jewett, T. J., Fischer, E. R., Mead, D. J., & Hackstadt, T. (2006). Chlamydial TARP is a bacterial nucleator of actin. Proceedings of the National Academy of Sciences of the United States of America, 103(42), 15599–15604. doi:10.1073/pnas.0603044103

Jiwani, S., Alvarado, S., Ohr, R. J., Romero, A., Nguyen, B., & Jewett, T. J. (2013). Chlamydia trachomatis Tarp Harbors Distinct G and F Actin Binding Domains That Bundle Actin Filaments. Journal of Bacteriology, 195(4), 708–716. doi:10.1128/JB.01768-12

Johnson, C. M., & Fisher, D. J. (2013). Site-specific, insertional inactivation of incA in Chlamydia trachomatis using a group II intron. PLoS ONE, 8(12), e83989. doi:10.1371/journal.pone.0083989

Johnson, K. A., Tang, F. F., & Sütterlin, C. (2009). Centrosome abnormalities during a Chlamydia trachomatis infection are caused by dysregulation of the normal duplication pathway. Cellular Microbiology, 11(7), 1064–1073. doi:10.1111/j.1462- 5822.2009.01307.x

Jouvenet, N. (2012). Dynamics of ESCRT proteins. Cellular and Molecular Life Sciences : CMLS, 69(24), 4121–4133. doi:10.1007/s00018-012-1035-0

Jutras, I., Abrami, L., & Dautry-Varsat, A. (2003). Entry of the Lymphogranuloma Venereum Strain of Chlamydia trachomatis into Host Cells Involves Cholesterol-Rich Membrane Domains. Infection and Immunity, 71(1), 260–266. doi:10.1128/IAI.71.1.260-266.2003

Kabeiseman, E. J., Cichos, K., Hackstadt, T., Lucas, A., & Moore, E. R. (2013). Vesicle- Associated Membrane Protein 4 and Syntaxin 6 Interactions at the Chlamydial Inclusion. Infection and Immunity, 81(9), 3326–3337. doi:10.1128/IAI.00584-13

Kanazawa, C., Morita, E., Yamada, M., Ishii, N., Miura, S., Asao, H., et al. (2003). Effects of deficiencies of STAMs and Hrs, mammalian class E Vps proteins, on receptor downregulation. Biochemical and Biophysical Research Communications, 309(4), 848– 856. doi:10.1016/j.bbrc.2003.08.078

Kari, L., Goheen, M. M., Randall, L. B., Taylor, L. D., Carlson, J. H., Whitmire, W. M., et al. (2011). Generation of targeted Chlamydia trachomatis null mutants. Proceedings of the National Academy of Sciences of the United States of America, 108(17), 7189–7193. doi:10.1073/pnas.1102229108

Katzmann, D. J., Babst, M., & Emr, S. D. (2001). Ubiquitin-dependent sorting into the multivesicular body pathway requires the function of a conserved endosomal protein sorting complex, ESCRT-I. Cell, 106(2), 145–155.

Kim, J. H., Jiang, S., Elwell, C. A., & Engel, J. N. (2011). Chlamydia trachomatis Co-opts the FGF2 Signaling Pathway to Enhance Infection. PLoS Pathogens, 7(10), e1002285. doi:10.1371/journal.ppat.1002285.s006

138 Kiselev, A. O., Skinner, M. C., & Lampe, M. F. (2009). Analysis of pmpD expression and PmpD post-translational processing during the life cycle of Chlamydia trachomatis serovars A, D, and L2. PLoS ONE, 4(4), e5191. doi:10.1371/journal.pone.0005191

Kiselev, A. O., Stamm, W. E., Yates, J. R., & Lampe, M. F. (2007). Expression, processing, and localization of PmpD of Chlamydia trachomatis Serovar L2 during the chlamydial developmental cycle. PLoS ONE, 2(6), e568. doi:10.1371/journal.pone.0000568

Komada, M., & Kitamura, N. (1995). Growth factor-induced tyrosine phosphorylation of Hrs, a novel 115-kilodalton protein with a structurally conserved putative zinc finger domain. Molecular and Cellular Biology, 15(11), 6213–6221.

Komada, M., Masaki, R., Yamamoto, A., & Kitamura, N. (1997). Hrs, a tyrosine kinase substrate with a conserved double zinc finger domain, is localized to the cytoplasmic surface of early endosomes. The Journal of Biological Chemistry, 272(33), 20538–20544.

Kumar, Y., Cocchiaro, J., & Valdivia, R. H. (2006). The Obligate Intracellular Pathogen Chlamydia trachomatis Targets Host Lipid Droplets. Current Biology, 16(16), 1646– 1651. doi:10.1016/j.cub.2006.06.060

Kun, D., Xiang-lin, C., Ming, Z., & Qi, L. (2013). Chlamydia inhibit host cell apoptosis by inducing Bag-1 via the MAPK/ERK survival pathway. Apoptosis, 18(9), 1083–1092. doi:10.1007/s10495-013-0865-z la Maza, de, L. M., & Peterson, E. M. (1982). Scanning electron microscopy of McCoy cells infected with Chlamydia trachomatis. Experimental and Molecular Pathology, 36(2), 217–226.

Lane, B. J., Mutchler, C., Khodor, Al, S., Grieshaber, S. S., & Carabeo, R. A. (2008). Chlamydial Entry Involves TARP Binding of Guanine Nucleotide Exchange Factors. PLoS Pathogens, 4(3), e1000014. doi:10.1371/journal.ppat.1000014.s004

Le Negrate, G., Krieg, A., Faustin, B., Loeffler, M., Godzik, A., Krajewski, S., & Reed, J. C. (2008). ChlaDub1 of Chlamydia trachomatis suppresses NF-kappaB activation and inhibits IkappaBalpha ubiquitination and degradation. Cellular Microbiology, 10(9), 1879–1892. doi:10.1111/j.1462-5822.2008.01178.x

Lee, J.-A., Beigneux, A., Ahmad, S. T., Young, S. G., & Gao, F.-B. (2007). ESCRT-III Dysfunction Causes Autophagosome Accumulation and Neurodegeneration. Current Biology, 17(18), 1561–1567. doi:10.1016/j.cub.2007.07.029

Leiva, N., Capmany, A., & Damiani, M. T. (2013). Rab11-family of interacting protein 2 associates with chlamydial inclusions through its Rab-binding domain and promotes bacterial multiplication. Cellular Microbiology, 15(1), 114–129. doi:10.1111/cmi.12035

Liechti, G. W., Kuru, E., Hall, E., Kalinda, A., Brun, Y. V., VanNieuwenhze, M., & Maurelli, A. T. (2013). nature12892. Nature, 1–17. doi:10.1038/nature12892

Litwin, J. (1962). GROWTH OF THE AGENT OF TRACHOMA IN THE EMBRYONATED EGG*. Annals of the New York Academy of Sciences, 98(1), 145–162.

139

Lu, Q., Hope, L. W., Brasch, M., Reinhard, C., & Cohen, S. N. (2003). TSG101 interaction with HRS mediates endosomal trafficking and receptor down-regulation. Proceedings of the National Academy of Sciences of the United States of America, 100(13), 7626–7631. doi:10.1073/pnas.0932599100

Martin-Iguacel, R., Llibre, J. M., Nielsen, H., Heras, E., Matas, L., Lugo, R., et al. (2010). Lymphogranuloma venereum proctocolitis: a silent endemic disease in men who have sex with men in industrialised countries. European Journal of Clinical Microbiology & Infectious Diseases, 29(8), 917–925. doi:10.1007/s10096-010-0959-2

Matsumoto, A. (1981). Isolation and electron microscopic observations of intracytoplasmic inclusions containing Chlamydia psittaci. Journal of Bacteriology, 145(1), 605–612.

Matsumoto, A., Bessho, H., Uehira, K., & Suda, T. (1991). Morphological studies of the association of mitochondria with chlamydial inclusions and the fusion of chlamydial inclusions. Journal of Electron Microscopy, 40(5), 356–363.

McCullough, J., Colf, L. A., & Sundquist, W. I. (2013). Membrane Fission Reactions of the Mammalian ESCRT Pathway. Annual Review of Biochemistry, 82(1), 663–692. doi:10.1146/annurev-biochem-072909-101058

McCullough, J., Row, P. E., Lorenzo, Ó., Doherty, M., Beynon, R., Clague, M. J., & Urbé, S. (2006). Activation of the endosome-associated ubiquitin isopeptidase AMSH by STAM, a component of the multivesicular body-sorting machinery. Current Biology : CB, 16(2), 160–165. doi:10.1016/j.cub.2005.11.073

McDonald, B., & Martin-Serrano, J. (2007). Regulation of Tsg101 Expression by the Steadiness Box: A Role of Tsg101-associated Ligase. Molecular Biology of the Cell, 19(2), 754–763. doi:10.1091/mbc.E07-09-0957

Mehra, A., Zahra, A., Thompson, V., Sirisaengtaksin, N., Wells, A., Porto, M., et al. (2013). Mycobacterium tuberculosis Type VII Secreted Effector EsxH Targets Host ESCRT to Impair Trafficking. PLoS Pathogens, 9(10), e1003734. doi:10.1371/journal.ppat.1003734.s013

Misaghi, S., Balsara, Z. R., Catic, A., Spooner, E., Ploegh, H. L., & Starnbach, M. N. (2006). Chlamydia trachomatis-derived deubiquitinating enzymes in mammalian cells during infection. Molecular Microbiology, 61(1), 142–150. doi:10.1111/j.1365- 2958.2006.05199.x

Mizuno, E., Kawahata, K., Okamoto, A., Kitamura, N., & Komada, M. (2004). Association with Hrs is required for the early endosomal localization, stability, and function of STAM. Journal of Biochemistry, 135(3), 385–396. doi:10.1093/jb/mvh046

Moore, E. R., Fischer, E. R., Mead, D. J., & Hackstadt, T. (2008). The Chlamydial Inclusion Preferentially Intercepts Basolaterally Directed Sphingomyelin-Containing Exocytic Vacuoles. Traffic, 9(12), 2130–2140. doi:10.1111/j.1600-0854.2008.00828.x

Moorhead, A. M., Jung, J.-Y., Smirnov, A., Kaufer, S., & Scidmore, M. A. (2010). Multiple

140 Host Proteins That Function in Phosphatidylinositol-4-Phosphate Metabolism Are Recruited to the Chlamydial Inclusion. Infection and Immunity, 78(5), 1990–2007. doi:10.1128/IAI.01340-09

Moulder, J. W. (1966). The relation of the psittacosis group (Chlamydiae) to bacteria and viruses. Annual Review of Microbiology, 20, 107–130. doi:10.1146/annurev.mi.20.100166.000543

Mölleken, K., Becker, E., & Hegemann, J. H. (2013). The Chlamydia pneumoniae invasin protein Pmp21 recruits the EGF receptor for host cell entry. PLoS Pathogens, 9(4), e1003325. doi:10.1371/journal.ppat.1003325.s007

Nguyen, B. D., & Valdivia, R. H. (2012). Virulence determinants in the obligate intracellular pathogen Chlamydia trachomatis revealed by forward genetic approaches. Proceedings of the National Academy of Sciences of the United States of America, 109(4), 1263–1268. doi:10.1073/pnas.1117884109

Nikko, E., & André, B. (2007). Split-ubiquitin two-hybrid assay to analyze protein-protein interactions at the endosome: application to Saccharomyces cerevisiae Bro1 interacting with ESCRT complexes, the Doa4 ubiquitin hydrolase, and the Rsp5 ubiquitin ligase. Eukaryotic Cell, 6(8), 1266–1277. doi:10.1128/EC.00024-07

Nunes, A., & Gomes, J. P. (2014). Infection, Genetics and Evolution. Infection, Genetics and Evolution, 23(C), 49–64. doi:10.1016/j.meegid.2014.01.029

Olive, A. J., Haff, M. G., Emanuele, M. J., Sack, L. M., Barker, J. R., Elledge, S. J., & Starnbach, M. N. (2014). Chlamydia trachomatis-induced alterations in the host cell proteome are required for intracellular growth. Cell Host & Microbe, 15(1), 113–124. doi:10.1016/j.chom.2013.12.009

Omsland, A., Sixt, B. S., Horn, M., & Hackstadt, T. (2014). Chlamydial metabolism revisited: interspecies metabolic variability and developmental stage-specific physiologic activities. FEMS Microbiology Reviews, n/a–n/a. doi:10.1111/1574-6976.12059

Ouellette, S. P., & Carabeo, R. A. (2010). A Functional Slow Recycling Pathway of Transferrin is Required for Growth of Chlamydia. Frontiers in Microbiology, 1, 112. doi:10.3389/fmicb.2010.00112

Ouellette, S. P., Karimova, G., Subtil, A., & Ladant, D. (2012). Chlamydia co-opts the rod shape-determining proteins MreB and Pbp2 for cell division. Molecular Microbiology, 85(1), 164–178. doi:10.1111/j.1365-2958.2012.08100.x

Paavonen, J. (2012). Chlamydia trachomatis infections of the female genital tract: State of the art. Annals of Medicine, 44(1), 18–28. doi:10.3109/07853890.2010.546365

Paschen, S. A., Christian, J. G., Vier, J., Schmidt, F., Walch, A., Ojcius, D. M., & Hacker, G. (2008). Cytopathicity of Chlamydia is largely reproduced by expression of a single chlamydial protease. The Journal of Cell Biology, 182(1), 117–127. doi:10.1083/jcb.200804023

141 Pascolini, D., & Mariotti, S. P. (2012). Global estimates of visual impairment: 2010. The British Journal of Ophthalmology, 96(5), 614–618. doi:10.1136/bjophthalmol-2011- 300539

Peabody, C. R. (2003). Type II protein secretion and its relationship to bacterial type IV pili and archaeal flagella. Microbiology, 149(11), 3051–3072. doi:10.1099/mic.0.26364-0

Pedersen, L. N. R., Herrmann, B. R., & MÃ ller, J. K. L. (2009). Typing Chlamydia trachomatis: from egg yolk to nanotechnology. FEMS Immunology & Medical Microbiology, 55(2), 120–130. doi:10.1111/j.1574-695X.2008.00526.x

Pfeffer, S. R. (2013). Rab GTPase regulation of membrane identity. Current Opinion in Cell Biology, 25(4), 414–419. doi:10.1016/j.ceb.2013.04.002

Philips, J. A., Porto, M. C., Wang, H., Rubin, E. J., & Perrimon, N. (2008). ESCRT factors restrict mycobacterial growth. Proceedings of the National Academy of Sciences of the United States of America, 105(8), 3070–3075. doi:10.1073/pnas.0707206105

Pilhofer, M., Aistleitner, K., Biboy, J., Gray, J., Kuru, E., Hall, E., et al. (2013). Discovery of chlamydial peptidoglycan reveals bacteria with murein sacculi but without FtsZ. Nature Communications, 4, 2856. doi:10.1038/ncomms3856

Pirbhai, M., Dong, F., Zhong, Y., Pan, K. Z., & Zhong, G. (2006). The Secreted Protease Factor CPAF Is Responsible for Degrading Pro-apoptotic BH3-only Proteins in Chlamydia trachomatis-infected Cells. Journal of Biological Chemistry, 281(42), 31495– 31501. doi:10.1074/jbc.M602796200

Pizarro-Cerdá, J., & Cossart, P. (2006). Bacterial Adhesion and Entry into Host Cells. Cell, 124(4), 715–727. doi:10.1016/j.cell.2006.02.012

Popoff, V., Mardones, G. A., Bai, S.-K., Chambon, V. R., Tenza, D. L., Burgos, P. V., et al. (2009). Analysis of Articulation Between Clathrin and Retromer in Retrograde Sorting on Early Endosomes. Traffic, 10(12), 1868–1880. doi:10.1111/j.1600-0854.2009.00993.x

Pornillos, O., Higginson, D. S., Stray, K. M., Fisher, R. D., Garrus, J. E., Payne, M., et al. (2003). HIV Gag mimics the Tsg101-recruiting activity of the human Hrs protein. The Journal of Cell Biology, 162(3), 425–434. doi:10.1083/jcb.200302138

Prag, G., Watson, H., Kim, Y. C., Beach, B. M., Ghirlando, R., Hummer, G., et al. (2007). The Vps27/Hse1 complex is a GAT domain-based scaffold for ubiquitin-dependent sorting. Developmental Cell, 12(6), 973–986. doi:10.1016/j.devcel.2007.04.013

Rajalingam, K., Sharma, M., Paland, N., Hurwitz, R., Thieck, O., Oswald, M., et al. (2006). IAP-IAP Complexes Required for Apoptosis Resistance of C. trachomatis–Infected Cells. PLoS Pathogens, 2(10), e114. doi:10.1371/journal.ppat.0020114.sg007

Raposo, G., & Stoorvogel, W. (2013). Extracellular vesicles: exosomes, microvesicles, and friends. The Journal of Cell Biology, 200(4), 373–383. doi:10.1083/jcb.201211138

Raymond, C. K., Howald-Stevenson, I., Vater, C. A., & Stevens, T. H. (1992). Morphological

142 classification of the yeast vacuolar protein sorting mutants: evidence for a prevacuolar compartment in class E vps mutants. Molecular Biology of the Cell, 3(12), 1389–1402.

Rejman Lipinski, A., Heymann, J., Meissner, C., Karlas, A., Brinkmann, V., Meyer, T. F., & Heuer, D. (2009). Rab6 and Rab11 Regulate Chlamydia trachomatis Development and Golgin-84-Dependent Golgi Fragmentation. PLoS Pathogens, 5(10), e1000615. doi:10.1371/journal.ppat.1000615.s012

Rosmarin, D. M., Carette, J. E., Olive, A. J., Starnbach, M. N., Brummelkamp, T. R., & Ploegh, H. L. (2012). Attachment of Chlamydia trachomatis L2 to host cells requires sulfation. Proceedings of the National Academy of Sciences of the United States of America, 109(25), 10059–10064. doi:10.1073/pnas.1120244109/-/DCSupplemental

Rusten, T. E., & Stenmark, H. (2009). How do ESCRT proteins control autophagy? Journal of Cell Science, 122(Pt 13), 2179–2183. doi:10.1242/jcs.050021

Rusten, T. E., Vaccari, T., & Stenmark, H. (2012). Shaping development with ESCRTs. Nature Cell Biology, 14(1), 38–45. doi:10.1038/ncb2381

Rusten, T. E., Vaccari, T., Lindmo, K., Rodahl, L. M. W., Nezis, I. P., Sem-Jacobsen, C., et al. (2007). ESCRTs and Fab1 regulate distinct steps of autophagy. Current Biology : CB, 17(20), 1817–1825. doi:10.1016/j.cub.2007.09.032

Rzomp, K. A., Moorhead, A. R., & Scidmore, M. A. (2006). The GTPase Rab4 Interacts with Chlamydia trachomatis Inclusion Membrane Protein CT229. Infection and Immunity, 74(9), 5362–5373. doi:10.1128/IAI.00539-06

Rzomp, K. A., Scholtes, L. D., Briggs, B. J., Whittaker, G. R., & Scidmore, M. A. (2003). Rab GTPases Are Recruited to Chlamydial Inclusions in Both a Species-Dependent and Species-Independent Manner. Infection and Immunity, 71(10), 5855–5870. doi:10.1128/IAI.71.10.5855-5870.2003

Saier, M. H., Jr. (2006). Protein secretion and membrane insertion systems in gram-negative bacteria. The Journal of Membrane Biology, 214(1-2), 75–90. doi:10.1007/s00232-006- 0049-7

Saka, H. A., & Valdivia, R. H. (2010). Acquisition of nutrients by Chlamydiae: unique challenges of living in an intracellular compartment. Current Opinion in Microbiology, 13(1), 4–10. doi:10.1016/j.mib.2009.11.002

Saka, H. A., Thompson, J. W., Chen, Y.-S., Kumar, Y., Dubois, L. G., Moseley, M. A., & Valdivia, R. H. (2011). Quantitative proteomics reveals metabolic and pathogenic properties of Chlamydia trachomatis developmental forms. Molecular Microbiology, 82(5), 1185–1203. doi:10.1111/j.1365-2958.2011.07877.x

Schlosser, K. (2013). Trachoma through history.

Schwoppe, C., Winkler, H. H., & Neuhaus, H. E. (2002). Properties of the Glucose-6- Phosphate Transporter from Chlamydia pneumoniae (HPTcp) and the Glucose-6- Phosphate Sensor from Escherichia coli (UhpC). Journal of Bacteriology, 184(8), 2108–

143 2115. doi:10.1128/JB.184.8.2108-2115.2002

Scidmore, M. A. (2005). Cultivation and Laboratory Maintenance of Chlamydia trachomatis. Current Protocols in Microbiology, Chapter 11, Unit 11A.1. doi:10.1002/9780471729259.mc11a01s00

Scidmore, M. A., Fischer, E. R., & Hackstadt, T. (1996). Sphingolipids and glycoproteins are differentially trafficked to the Chlamydia trachomatis inclusion. The Journal of Cell Biology, 134(2), 363–374.

Scidmore, M. A., Fischer, E. R., & Hackstadt, T. (2003). Restricted Fusion of Chlamydia trachomatis Vesicles with Endocytic Compartments during the Initial Stages of Infection. Infection and Immunity, 71(2), 973–984. doi:10.1128/IAI.71.2.973-984.2003

Scidmore-Carlson, M. A., Shaw, E. I., Dooley, C. A., Fischer, E. R., & Hackstadt, T. (1999). Identification and characterization of a Chlamydia trachomatis early operon encoding four novel inclusion membrane proteins. Molecular Microbiology, 33(4), 753–765.

Scott, A., Chung, H.-Y., Gonciarz-Swiatek, M., Hill, G. C., Whitby, F. G., Gaspar, J., et al. (2005). Structural and mechanistic studies of VPS4 proteins. The EMBO Journal, 24(20), 3658–3669. doi:10.1038/sj.emboj.7600818

Sharma, M., Machuy, N., Böhme, L., Karunakaran, K., Mäurer, A. P., Meyer, T. F., & Rudel, T. (2011). HIF-1& is involved in mediating apoptosis resistance to Chlamydia trachomatis-infected cells. Cellular Microbiology, 13(10), 1573–1585. doi:10.1111/j.1462-5822.2011.01642.x

Shaw, E. I., Dooley, C. A., Fischer, E. R., Scidmore, M. A., Fields, K. A., & Hackstadt, T. (2000). Three temporal classes of gene expression during the Chlamydia trachomatis developmental cycle. Molecular Microbiology, 37(4), 913–925.

Shemer, Y., & Sarov, I. (1985). Inhibition of growth of Chlamydia trachomatis by human gamma interferon. Infection and Immunity, 48(2), 592–596.

Sixt, B. S., Siegl, A., Müller, C., Watzka, M., Wultsch, A., Tziotis, D., et al. (2013). Metabolic features of Protochlamydia amoebophila elementary bodies--a link between activity and infectivity in Chlamydiae. PLoS Pathogens, 9(8), e1003553. doi:10.1371/journal.ppat.1003553

Spaeth, K. E., Chen, Y.-S., & Valdivia, R. H. (2009). The Chlamydia Type III Secretion System C-ring Engages a Chaperone-Effector Protein Complex. PLoS Pathogens, 5(9), e1000579. doi:10.1371/journal.ppat.1000579.g006

Steiper, M. E., & Young, N. M. (2006). Primate molecular divergence dates. Molecular Phylogenetics and Evolution, 41(2), 384–394. doi:10.1016/j.ympev.2006.05.021

Stenmark, H. (2009). Rab GTPases as coordinatorsof vesicle traffic. Nature Reviews Molecular Cell Biology, 10(8), 513–525. doi:10.1038/nrm2728

Stephens, R. S., Kalman, S., Lammel, C., Fan, J., Marathe, R., Aravind, L., et al. (1998).

144 Genome Sequence of an Obligate Intracellular Pathogen of Humans: Chlamydia trachomatis. Science (New York, NY), 282(5389), 754–759. doi:10.1126/science.282.5389.754

Stephens, R. S., Myers, G., Eppinger, M., & Bavoil, P. M. (2009). Divergence without difference: phylogenetics and taxonomy of Chlamydia resolved. FEMS Immunology & Medical Microbiology, 55(2), 115–119. doi:10.1111/j.1574-695X.2008.00516.x

Stern, K. A., Visser Smit, G. D., Place, T. L., Winistorfer, S., Piper, R. C., & Lill, N. L. (2007). Epidermal Growth Factor Receptor Fate Is Controlled by Hrs Tyrosine Phosphorylation Sites That Regulate Hrs Degradation. Molecular and Cellular Biology, 27(3), 888–898. doi:10.1128/MCB.02356-05

Subtil, A. (2004). Analysis of Chlamydia caviae entry sites and involvement of Cdc42 and Rac activity. Journal of Cell Science, 117(17), 3923–3933. doi:10.1242/jcs.01247

Subtil, A., Blocker, A., & Dautry-Varsat, A. (2000). Type III secretion system in Chlamydia species: identified members and candidates. Microbes and Infection, 2(4), 367–369.

Subtil, A., Collingro, A., & Horn, M. (2014). Tracing the primordial Chlamydiae: extinct parasites of plants? Trends in Plant Science, 19(1), 36–43. doi:10.1016/j.tplants.2013.10.005

Subtil, A., Delevoye, C., Balañá, M.-E., Tastevin, L., Perrinet, S., & Dautry-Varsat, A. (2005). A directed screen for chlamydial proteins secreted by a type III mechanism identifies a translocated protein and numerous other new candidates. Molecular Microbiology, 56(6), 1636–1647. doi:10.1111/j.1365-2958.2005.04647.x

Subtil, A., Parsot, C., & Dautry-Varsat, A. (2001). Secretion of predicted Inc proteins of Chlamydia pneumoniae by a heterologous type III machinery. Molecular Microbiology, 39(3), 792–800.

Suchland, R. J., Rockey, D. D., Bannantine, J. P., & Stamm, W. E. (2000). Isolates of Chlamydia trachomatis that occupy nonfusogenic inclusions lack IncA, a protein localized to the inclusion membrane. Infection and Immunity, 68(1), 360–367.

Sun, W., Yan, Q., Vida, T. A., & Bean, A. J. (2003). Hrs regulates early endosome fusion by inhibiting formation of an endosomal SNARE complex. The Journal of Cell Biology, 162(1), 125–137. doi:10.1083/jcb.200302083

Taboriski, J. (1952). Historic and Ethnologic Factors in the Distribution of Trachoma. American Journal of Opthalmology.

Tamai, K., Tanaka, N., Nara, A., Yamamoto, A., Nakagawa, I., Yoshimori, T., et al. (2007). Role of Hrs in maturation of autophagosomes in mammalian cells. Biochemical and Biophysical Research Communications, 360(4), 721–727. doi:10.1016/j.bbrc.2007.06.105

Tang, F. F., & Bavoil, P. M. (Eds.). (2012). Intracellular Pathogens I: Chlamydiales (1st ed.).

Tang, F. F., Chang, H. L., & Huang, Y. T. (1957). Studies on the etiology of trachoma with

145 special reference to isolation of the virus in chick embryo. Chinese Medical Journal.

Thylefors, B., Dawson, C. R., Jones, B. R., West, S. K., & Taylor, H. R. (1987). A simple system for the assessment of trachoma and its complications. Bulletin of the World Health Organization, 65(4), 477–483.

Urbé, S., Sachse, M., Row, P. E., Preisinger, C., Barr, F. A., Strous, G., et al. (2003). The UIM domain of Hrs couples receptor sorting to vesicle formation. Journal of Cell Science, 116(Pt 20), 4169–4179. doi:10.1242/jcs.00723 van Ooij, C., Apodaca, G., & Engel, J. (1997). Characterization of the Chlamydia trachomatis vacuole and its interaction with the host endocytic pathway in HeLa cells. Infection and Immunity, 65(2), 758–766. Retrieved from http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=9009339&ret mode=ref&cmd=prlinks

Visser Smit, G. D., Place, T. L., Cole, S. L., Clausen, K. A., Vemuganti, S., Zhang, G., et al. (2009). Cbl controls EGFR fate by regulating early endosome fusion. Science Signaling, 2(102), ra86. doi:10.1126/scisignal.2000217

Votteler, J., & Sundquist, W. I. (2013). Virus Budding and the ESCRT Pathway. Cell Host & Microbe, 14(3), 232–241. doi:10.1016/j.chom.2013.08.012

Wang, Y., Kahane, S., Cutcliffe, L. T., Skilton, R. J., Lambden, P. R., & Clarke, I. N. (2011). Development of a Transformation System for Chlamydia trachomatis: Restoration of Glycogen Biosynthesis by Acquisition of a Plasmid Shuttle Vector. PLoS Pathogens, 7(9), e1002258. doi:10.1371/journal.ppat.1002258.g008

Ward, M. E., & Murray, A. (1984). Control Mechanisms Governing the Infectivity of Chlamydia trachomatis for HeLa Cells: Mechanisms of Endocytosis. Microbiology (Reading, England), 130(7), 1765–1780.

Webb, S. G. (1990). Prehistoric eye disease (trachoma?) in Australian Aborigines. Wehrl, W., Brinkmann, V., Jungblut, P. R., Meyer, T. F., & Szczepek, A. J. (2004). From the inside out--processing of the Chlamydial autotransporter PmpD and its role in bacterial adhesion and activation of human host cells. Molecular Microbiology, 51(2), 319–334. doi:10.1046/j.1365-2958.2003.03838.x

Weiss, E. R., Popova, E., Yamanaka, H., Kim, H. C., Huibregtse, J. M., & Göttlinger, H. (2010). Rescue of HIV-1 release by targeting widely divergent NEDD4-type ubiquitin ligases and isolated catalytic HECT domains to Gag. PLoS Pathogens, 6(9), e1001107. doi:10.1371/journal.ppat.1001107

WHO. (2013). who's weekly epidemiological recod, 1–16. Wickstrum, J., Sammons, L. R., Restivo, K. N., & Hefty, P. S. (2013). Conditional Gene Expression in Chlamydia trachomatis Using the Tet System. PLoS ONE, 8(10), e76743. doi:10.1371/journal.pone.0076743.s004

Wilson, D. P., Whittum-Hudson, J. A., Timms, P., & Bavoil, P. M. (2009). Kinematics of Intracellular Chlamydiae Provide Evidence for Contact-Dependent Development. Journal

146 of Bacteriology, 191(18), 5734–5742. doi:10.1128/JB.00293-09

Wollert, T., & Hurley, J. H. (2010). Molecular mechanism of multivesicular body biogenesis by ESCRT complexes. Nature, 464(7290), 864–869. doi:10.1038/nature08849 World Health Organization. (2012). Global incidence and prevalence of selected curable sexually transmitted infections-2008.

Wylie, J. L., Hatch, G. M., & McClarty, G. (1997). Host cell phospholipids are trafficked to and then modified by Chlamydia trachomatis. Journal of Bacteriology, 179(23), 7233– 7242.

Wyrick, P. B. (2000). Intracellular survival by Chlamydia. Cellular Microbiology, 2(4), 275– 282.

Wyrick, P. B. (2010). Chlamydia trachomatis Persistence In Vitro: An Overview. The Journal of Infectious Diseases, 201(S2), 88–95. doi:10.1086/652394

Yan, Q., Sun, W., Kujala, P., Lotfi, Y., Vida, T. A., & Bean, A. J. (2005). CART: an Hrs/actinin-4/BERP/myosin V protein complex required for efficient receptor recycling. Molecular Biology of the Cell, 16(5), 2470–2482. doi:10.1091/mbc.E04

Zhang, J. P., & Stephens, R. S. (1992). Mechanism of C. trachomatis attachment to eukaryotic host cells. Cell, 69(5), 861–869.

Zhang, J., Du, J., Lei, C., Liu, M., & Zhu, A. J. (2014). Ubpy controls the stability of the ESCRT-0 subunit Hrs in development. Development (Cambridge, England), 141(7), 1473–1479. doi:10.1242/dev.099564

147

ANNEXE

148

Available online at www.sciencedirect.com

ScienceDirect

Exploitation of host lipids by bacteria

1,2,3 1,2

Franc¸ois Vromman and Agathe Subtil

Bacteria that interact with eukaryotic cells have developed a phosphatidylcholine, or sphingolipids, have been

variety of strategies to divert host lipids, or cellular processes described in several bacteria, even before the advent

driven by lipids, to their beneﬁt. Host lipids serve as building of lipidomics. In some instance, a host-synthesized

blocks for bacterial membrane formation and as energy source. straight-chain fatty acid is replaced with a branched-

They promote the formation of speciﬁc microdomains, chain fatty acid [2]. The technological advances in mass

facilitating interactions with the host. Host lipids are also critical spectrometry of the last decade have been exploited in

players in the entry of bacteria or toxins into cells, and, for only a few studies yet (Table 1), enough though to

bacteria growing inside parasitophorous vacuoles, in building a broaden our perception of the diversity in lipid proﬁles

secure shelter. Bacterial dissemination is often dependent on of specialized pathogens [3,4]. The advantages procured

enzymatic activities targeting host lipids. Finally, on a larger by host-derived lipids are often not known. As discussed

scale, long lasting parasitic association can disturb host lipid below, they might enable the formation of microdo-

metabolism so deeply as to ‘reprogram’ it, as proposed in the mains, or change the ﬂuidity of the bacterial membrane.

case of Mycobacterium infection. In many cases, incorporation of host lipids is probably the

Addresses result of a ‘scavenger’ behavior, with bacteria ‘feeding’

Institut Pasteur, Unite´ de Biologie des Interactions Cellulaires, Paris, on the host even for structural blocks. It is worth noting

France

2 that many bacteria containing a signiﬁcant proportion of

CNRS URA 2582, Paris, France

3 lipids of host origin grow normally in vitro even when the

Universite´ Pierre et Marie Curie, Paris, France

supply of this particular category is experimentally shut

Corresponding author: Subtil, Agathe ([email protected]) down (Table 1). These results should be taken cau-

tiously, because in vivo conditions could reveal more

stringent requirements, in particular in key steps of

Current Opinion in Microbiology 2014, 17:38–45

bacterial infection such as the establishment of the in-

This review comes from a themed issue on Host–microbe

fection, or the entry into a persistent state. Nevertheless,

interactions: bacteria

they indicate that in terms of growth and division, most

Edited by Olivia Steele-Mortimer and Agathe Subtil

bacteria show a high level of plasticity regarding their

For a complete overview see the Issue and the Editorial lipid composition.

Available online 10th December 2013

Utilizing host lipids to strengthen interactions

with the plasma membrane

http://dx.doi.org/10.1016/j.mib.2013.11.003

Helicobacter pylori is one of these few prokaryotes that

have cholesterol in its outer membrane (Table 2). This

extracellular bacterium incorporates cholesterol from its

Introduction host, and converts it to cholesteryl glucosides, an import-

Lipids have long been associated to two general func- ant step to escape phagocytosis and T cell activation [5].

tions: a structural role, based on their ability to assemble Recent work suggests that accumulated cholesterol and

into membranes, and a role in energy storage. However, cholesteryl glucosides facilitate the selective lateral-

the several thousands of different structures behind the phase segregation and induce the membrane assemblage

word ‘lipid’ hide multiple other essential biological func- and raft coalescence on the host–bacterium contact sites,

tions, in particular in signaling and in the regulation of which may serve as a signal to trigger secretion of bacterial

membrane trafﬁc. Bacteria that interact with eukaryotic proteins into the host by a type IV secretion system [6].

cells have developed a variety of strategies to divert host Similarly, recent data show direct transfer of cholesterol

lipids, or cellular processes driven by lipids, to their from the plasma membranes of epithelial cells to the

beneﬁt. Here we will review some of the recent data that extracellular bacterium Borrelia burgdorferi, where it can

illustrate the exploitation of host lipids by bacteria. For be glycosylated by bacterial enzymes, and transferred

earlier studies we highly recommend a very comprehen- back to the host cell. It is speculated that the particular

sive review [1]. lipid microdomains created on the surface of both bacter-

ium and host could create an opportunity for lipid–lipid

Host lipids as structural elements of interactions and facilitate B. burgdorferi’s contact with

prokaryotic membranes host cells [7]. Indeed, although they lack sphingolipids,

Several bacteria integrate lipids originating from the cholesterol-rich microdomains with properties compar-

host into their own membrane (Table 1). Thus, lipids able to the lipid rafts described in eukaryotic membranes

usually not present in prokaryotes, such as cholesterol, were recently described [8].

Current Opinion in Microbiology 2014, 17:38–45 www.sciencedirect.com

Exploitation of host lipids by bacteria Vromman and Subtil 39

Table 1

Host-derived lipids in bacteria and, when indicated, in parasitophorous vacuoles

Bacteria Host lipids Tools used for experimental Origin and acquisition Requirement?

imported evidence pathways

Anaplasma phagocytophilum Cholesterol Microscopy [45], multidrug- Subversion of the Cholesterol required for growth

(bacteria and vacuole), resistant (MDR) HL-60 cell line Niemann–Pick type C1 [45]

Ehrlichia chaffeensis defective in Niemann–Pick pathway of intracellular

type C1 regulation [27] cholesterol transport

and homeostasis [27]

Brucella abortus Cholesterol Microscopy [46] Cyclic b-1,2-glucans (CbG)

target lipid rafts on the

phagosome; CbG-deﬁcient

mutants fail to prevent

phagosome-lysosome fusion

and do not replicate [46]

Borrelia burgdorferi Cholesterol Radioactive tracers, thin layer Plasma membrane of Unknown

chromatography (TLC) epithelial cells [7]

[7,47]; mass spectrometry

and NMR [48]; microscopy

[7]

Chlamydia trachomatis Sphingolipids Microscopy, TLC [49] Transport from the No impact of a signiﬁcant

(bacteria and vacuole) Golgi [50], Rab decrease in host sphingomyelin

mediated transport supply on bacterial growth [54]

(reviewed in [23]),

translocation of lipid

bodies [24], subversion

of the CERT mediated

transport of ceramide at

ER-vacuole contact

sites [51,52], subversion

of the host cell lipid

transport system

involved in the

formation of HDL [53]

Cholesterol Microscopy, radioactive Entry and growth in the absence

tracers and HPLC [55], cell line of cholesterol [9]

deﬁcient for cholesterol

synthesis [9]

Phosphatidylcholine Radioactive tracers, TLC [2], No impact of a signiﬁcant

and other cell line deﬁcient for CDP- decrease in phosphatidylcholine

phospholipids choline synthetase [54] supply on bacterial growth [54]

Coxiella burnetii Cholesterol Microscopy, biochemistry CD63 positive Absence of cholesterol per se

parasitophorous [40], cell line deﬁcient for multivesicular bodies impacts entry, intracellular

vacuole cholesterol synthesis [9] might be implicated [9] replication is not affected [9];

precursors of cholesterol are

required for optimal growth [40]

Helicobacter pylori Cholesterol TLC [5,56], microscopy [56] Plasma membrane of Cholesterol enhances growth but

epithelial cells [5] is not absolutely required [57];

cholesterol glycosides protect

from phagocytosis [5]

Mycobacterium avium Cholesterol Electron microscopy [58] Host stores, pathway Cholesterol depletion results in

unknown phagosome maturation and

fusion with lysosomes [58]

Mycobacterium Cholesterol Radioactive tracers, mutant in Host stores, pathway Required for persistence in mice

tuberculosis cholesterol import [37]. unknown [37]

Mycobacterium is one of the

few bacteria whose lipidome

has been investigated [59]

Triacylglycerol Microscopy, radioactive Lipid droplets [38] Link to dormancy? [38]

tracers and TLC [38]

Mycoplasma penetrans Cholesterol, TLC [60] Body ﬂuids? Needed for attachment to host

sphingolipids and cells [61]

fatty acids

Salmonella enterica Cholesterol Microscopy, HPLC [62], cell Host stores, pathway Entry and growth in the absence

serovar typhimurium line deﬁcient for cholesterol unknown of cholesterol [9], need for non-

vacuole synthesis [9] sterol precursors [63]

www.sciencedirect.com Current Opinion in Microbiology 2014, 17:38–45

40 Host–microbe interactions: bacteria

Table 2

Lipid categories discussed in this paper

Lipids Principal localization(s) in the host Principal function(s) in the host Described usage by bacteria

Cholesterol Plasma membrane, intracellular Membrane structure, formation Membrane component,

compartments, lipoproteins and of micro-domains (signaling formation of micro-domains at

lipid bodies platforms), precursor of steroid the vacuole or bacterial surface

hormones and bile salts

Diacylglycerols Low abundance on speciﬁc Signaling molecules Signaling at the vacuole

membranes membrane

Phospholipids All membranes Structural (main component of all Membrane component (possibly

membranes) modiﬁed by bacterial enzymes)

Phosphatidyl inositol Low abundance on speciﬁc Signaling molecules Signaling at the vacuole

phosphates (PIPs, membranes (depending on the type membrane

belong to the category of PIPs)

‘phospholipids’)

Sphingolipids Plasma membrane, intracellular Membrane structure, formation Membrane component (possibly

compartments and lipoproteins of micro-domains (signaling modiﬁed by bacterial enzymes),

platforms) formation of micro-domains at

the vacuole surface

Triacylglycerides Lipid bodies, lipoproteins Energy storage Energy source, precursor for lipid

synthesis

Gates for entry Host lipids are not only used for bacterial entry, but also

Cholesterol-rich microdomains at the host plasma mem- for the attachment and uptake of a number of bacterial

brane have also been implicated in the internalization toxins, which need speciﬁc host lipids to penetrate in cells

process of a number of intracellular bacteria (see [1] and and reach their targets [1]. Finally, a recent report

references therein). Drugs removing cholesterol from the suggests that lipid rafts might be targeted for bacterial

plasma membrane reduced bacterial entry, and speciﬁc exit from host cell [13] (Figure 1).

lipids and proteins known to be enriched in these micro-

domains were observed at the bacteria–host contact Establishing a suitable intracellular niche

sites. Mouse embryonic ﬁbroblasts lacking the enzyme Internalization of intracellular bacteria proceeds through

required for the ﬁnal step in cholesterol biosynthesis invagination of the plasma membrane, leading to the

were recently used to investigate the requirement for formation of a vacuole made of host lipids and proteins.

cholesterol for three intracellular pathogens. While Cox- By default, this compartment is meant to mature into a

iella burnetii entry was signiﬁcantly decreased in the phagosome, fuse with late endocytic compartments, acid-

absence of cholesterol, internalization of Salmonella ify and fuse with lysosomes, eventually resulting in

typhimurium and Chlamydia trachomatis was unaffected, bacterial death. This is particularly efﬁcient in pro-

demonstrating that in these cases, cholesterol per se is not fessional phagocytes, but even epithelial cells have innate

required for entry [9]. However, in the absence of cho- immune defense against intravacuolar intruders. To

lesterol, the precursor desmosterol accumulated in the escape those, many bacteria use pore-forming enzymes

cell, and although it did not fully compensate for cho- or phospholipases to puncture the vacuole membrane and

lesterol loss, it could still have contributed to the for- reach the cytosol [1]. Others remain in the membrane-

mation of lipid microdomains at the cell surface and bounded compartment, and actively transform it into a

thereby have participated in bacterial internalization. suitable intracellular niche. Many proteins possess

In support for a role of cholesterol or related lipids in domains that have afﬁnity for speciﬁc lipids. Thus, mod-

entry, mitotic cells show a transient enrichment of cho- ifying the lipid composition of the vacuole impacts the

lesterol in the outer leaﬂet of the plasma membrane, association of proteins to its surface, and thereby its

which correlates with an increase in Salmonella invasion maturation. Phosphatidylinositol phosphates (PIPs) in

[10]. Mobilization of cholesterol at the entry sites could particular serve as docking sites for protein domains.

signiﬁcantly deplete other intracellular compartments of Moreover, their hydrolysis yields second messengers that

this lipid. It was recently proposed that such a scenario transmit downstream signals. They are therefore privi-

takes place during Shigella invasion, resulting in the leged targets of bacterial secreted proteins, called effec-

fragmentation of the Golgi apparatus and reorganization tors (see [1,14] and references therein). Bacteria can

of the recycling compartment [11]. Cleavage of the directly modify host lipids by secreting enzymes, such

myristyl anchor of the ADP-ribosylation factor ARF1 as phospholipases or phosphatidylinositol phosphatases.

by the effector IpaJ has also been implicated in Golgi Identiﬁcation of such enzymes among candidate effector

fragmentation [12]. proteins and demonstration of their activity is, however,

Current Opinion in Microbiology 2014, 17:38–45 www.sciencedirect.com

Exploitation of host lipids by bacteria Vromman and Subtil 41

Figure 1

Extracellular sites Host lipid transfer to bacterial membranes Bidirectional lipid exchange Host lipids as energy and carbon source Host lipids as targets of bacterial toxins Interaction with specific lipid microdomains (induction of bacterial secretion machineries, promotion of bacterial entry) Extracellular Extracellular lipids

plasma membrane

Bacteria escapingpg in the cytosol y Bacteria residingg in a vacuole

Host compartments

Membrane lysis throughthrough the action of Modification of the lipidlipid comcompositionposition of the phospholipases & pore formingforming vacuole toxinestoxines HijackingHijacking host lipids to the vacuolvacuolee Intracellular HostH tl lipidiidt transferf tto btilbacterial membranes b Host lipids as energy and carbon source Signalling through host lipids Reprogramming of the host metabolism Intracellular sites

Current Opinion in Microbiology

Utilization of host lipids by bacteria. Bacteria exploits host lipids both from extracellular and intracellular locations. They target lipids integrated in

membranes (in the plasma membrane, in vacuole or organelle membranes) or associated to proteins (see e.g., [27,53]) in body fluids or in the cytosol.

Many of the interactions listed here have only be described for a few bacteria, but could apply to others.

not trivial. While the Legionella pneumophila vacuole was PIPs are not the only lipid targets of bacterial effectors.

known to be enriched in PI(4)P [15], it is only recently For instance L. pneumophila secrete two enzymes, LpdA

that the type IV system secreted effector SidF was and LecE, whose combined activities are expected to

recognized as a phosphatidylinositol phosphatase respon- result in the conversion of phosphatidylcholine to diacyl-

sible for this effect [16]. In contrast, the phosphoinosi- glycerol in the vacuole membrane [21].

tide phosphatase of S. typhimurium, SopB, whose activity

was identiﬁed more than 10 years earlier [17], has already The presence of cholesterol and sphingolipids in the

attracted a lot of attention. One of the outcomes of SopB vacuolar membrane might permit to create speciﬁc micro-

activity is the generation of PI(3)P at the vacuole surface, environments, concentrating in a small area molecules

inducing the recruitment of several PI(3)P-binding involved in one signaling or enzymatic process. By help-

proteins [18]. In addition, several activation pathways ing organizing the vacuolar membrane into such plat-

relying on phosphorylation cascades are turned on during forms, host lipids might thus contribute to the creation of

Salmonella infection, and it was estimated that about half a niche favorable to bacterial growth. In agreement with

of them required SopB [19]. In particular, the Akt this hypothesis, domains enriched in speciﬁc bacterial

mediated stimulation of Wnt signaling induces an epi- proteins and host kinases were observed in C. trachomatis

thelial–mesenchymal transition of enterocytes into M containing vacuoles [22].

cells [20]. This activity depends on the phosphoinositide

phosphatase activity of SopB, implying that modiﬁcation Hijacking host lipids for vacuolar extension

of host lipids by bacterial effectors can have con- and bacterial growth

sequences far beyond the mere composition of the For bacteria that develop in a vacuole, proliferation

vacuole membrane. is coupled with an increase in the vacuolar membrane

www.sciencedirect.com Current Opinion in Microbiology 2014, 17:38–45

42 Host–microbe interactions: bacteria

surface. It is not known whether lipids made in the was recently proposed that phospholipases secreted by a

bacteria are transferred into the vacuolar membrane; type VI mechanism target phosphatidylethanolamine, the

analysis of the membrane composition via lipidomics major component of bacterial membranes and serve as

might answer this question because some bacterial lipids speciﬁc antibacterial effectors [32].

do not exist in eukaryotic cells. What is clear is that

several host-derived lipids accumulate in phagosomal Feeding on host lipids

membranes (Table 1). A multiplicity of pathways can One motivation to ‘steal’ lipids from the host is that they

be intercepted by a single pathogen, as strikingly illus- can be converted into a carbon source and energy, pro-

trated by the obligate intracellular bacterium C. tracho- vided that the bacterium has the required enzymes, or can

matis ([23], see Table 1). In particular, lipid bodies are make use of the host enzymatic capacity. This is best

translocated into the lumen of the C. trachomatis contain- illustrated by M. tuberculosis, that has acquired several

ing vacuole [24]. The mechanism of capture and translo- genes involved in lipolysis and lipogenesis, and for which

cation of these large structures into the vacuole is not host lipids appear to be the primary source of carbon

known. Several bacteria and viruses interact with lipid [33]. Feeding on host lipids is, however, not restricted to

bodies, and how these interactions might beneﬁt patho- intracellular pathogens [34,35]. A recent study on Vibrio

gens is discussed in two recent reviews [25,26]. Anaplasma cholerae uncovered a surface exposed lipoprotein that

phagocytophilum gives an additional example of an obligate functions as a lipase liberating fatty acids from the host

intracellular bacterium, that subverts a speciﬁc pathway lysophosphatidylcholines. The freed fatty acids are trans-

of intracellular cholesterol transport and homeostasis for ported into the bacteria, serving as a carbon source as well

vacuole biogenesis and cholesterol capture by the bacteria as a building block for bacterial membrane assembly

[27]. [36]. How commonly do bacteria use host lipids to build

their own membrane or as a source of carbon is not known.

Damaging host lipids for dissemination Host lipids might sometimes only be required for bac-

For some pathogens that need to escape vacuoles, desta- terial survival under certain conditions that are not

bilizing host membranes through phospholipase activity necessarily reproduced by in vitro culture models. For

or pore-forming proteins is essential for survival and example, M. tuberculosis has a speciﬁc cholesterol import

proliferation. For instance, Clostridium perfringens uses system that enables to derive both carbon and energy

two enzymes, a phospholipase C/sphingomyelinase and from host cholesterol. Cholesterol import is not required

a cholesterol-dependent cytolysin to escape from the for establishing infection in mice or for growth in resting

phagosomes of macrophages and to persist in host tissues macrophages, but is essential for persistence [37].

[28]. Recent work shows that L. pneumophila secretes a

protein, SdhA, to counteract the deleterious effect of Taking control of host metabolism

another secreted protein, the phospholipase PlaA, on Macrophages infected by M. tuberculosis acquire a ‘foamy’

vacuole membrane stability. In the absence of SdhA, phenotype that is characterized by the intracellular

PlaA activity leads to host cell death and bacterial degra- accumulation of lipid bodies. This phenomenon has crucial

dation [29]. The beneﬁt of this deadly enzyme is still not implications since foamy macrophages form a secure reser-

clear. It may be that under certain circumstances, rupture voir for the bacilli and facilitate their persistence in the

of the vacuole is not followed by bacterial death and human host. Accumulation of lipid bodies is likely the

contributes to bacterial survival. Such is the case for the result of several consequences of infection on host metab-

major human pathogen Mycobacterium tuberculosis: it olism. M. tuberculosis inside the lipid-loaded macrophages

breaks the phagosomal vacuole at late stages of infection imports fatty acids derived from host triacylglycerol and

by employing effector proteins secreted by the type VII incorporates them intact into bacterial triacylglycerol [38].

secretion system ESX-1. Phagosomal rupture results in In addition, host glycolytic activity appears to be diverted

host cell death and gives the bacteria a chance to escape towards ketone body synthesis. This deregulation enables

innate host defense [30]. feedback activation of the anti-lipolytic G protein-coupled

receptor GPR109A, leading to perturbations in lipid

Many extracellular pathogens also have acquired phos- homeostasis and consequent accumulation of lipid bodies

pholipases activities. These enzymes, such as ExoU from in the macrophage [39]. Altogether, the accumulation

Pseudomonas aeruginosa, often have pleiotropic effects or sequestration of host lipids and their utilization by

(due to the destabilization of the membrane as well as M. tuberculosis are part of a metabolic reprogramming of

to the signaling properties of the moieties released) and the host cell, as illustrated by an elegant recent study

can be cytotoxic. There are clear evidences that these examining the fate of propionyl-CoA generated by the

enzymes increase colonization and virulence, indicating degradation of cholesterol [33].

that their activities favor dissemination [31]. However, in

some cases, acquisition of these activities might not have Even without a spectacular phenotype such as the ‘foamy’

been driven by host–pathogen interactions, but by intras- cells, metabolic reprogramming of the host is likely not

pecies and interspecies bacterial interactions. Indeed, it limited to M. tuberculosis infection. For instance, Coxiella

Current Opinion in Microbiology 2014, 17:38–45 www.sciencedirect.com

Exploitation of host lipids by bacteria Vromman and Subtil 43

infected cells have 70% more cholesterol, and genes References and recommended reading

Papers of particular interest, published within the period of review,

involved in both host cell cholesterol biosynthesis and

have been highlighted as:

exogenous cholesterol uptake are upregulated [40]. It is

of special interest

not clear if this is simply a response by the host cell to

of outstanding interest

maintain cholesterol homeostasis. It is likely that in many

instances, infection inﬂuences host lipid metabolism,

1. van der Meer-Janssen YPM, van Galen J, Batenburg JJ, Helms JB:

either directly, with the secretion of bacterial enzymes Lipids in host–pathogen interactions: pathogens exploit

the complexity of the host cell lipidome. Prog Lipid Res 2010,

with enzymatic activity that will interfere with the host’s,

49:1-26.

or indirectly, by feeding the host metabolism with bac-

2. Wylie JL, Hatch GM, McClarty G: Host cell phospholipids are

terial metabolites, or by interfering with regulation loops.

trafﬁcked to and then modiﬁed by Chlamydia trachomatis. J

For instance, it was recently shown that the action of Bact 1997, 179:7233-7242.

tumor necrosis factor, which constitutes a critical host

3. Parsons JB, Rock CO: Bacterial lipids: metabolism and

defense against M. tuberculosis, also results in the pro- membrane homeostasis. Prog Lipid Res 2013, 52:249-276.

duction of ceramide. Increase in ceramide has important 4. Layre E, Moody DB: Lipidomic proﬁling of model organisms

consequences since it acts as an inducer of both apoptosis and the world’s major pathogens. Biochimie 2013, 95:109-115.

and RIP1-dependent programmed necrosis [41]. 5. Wunder C, Churin Y, Winau F, Warnecke D, Vieth M, Lindner B,

Zahringer U, Mollenkopf HJ, Heinz E, Meyer TF: Cholesterol

glucosylation promotes immune evasion by Helicobacter

At tissue level, the inﬂuence of bacterial infection or pylori. Nat Med 2006, 12:1030-1038.

colonization of a host on its lipid metabolism is beginning

6. Wang H-J, Cheng W-C, Cheng H-H, Lai C-H, Wang W-C:

to emerge. A very early work had reported the changes in Helicobacter pylori cholesteryl glucosides interfere with host

membrane phase and affect type IV secretion system function

phospholipid fatty acid composition and triacylglycerol

during infection in AGS cells. Mol Microbiol 2012, 83:67-84.

content in mouse tissues after infection of mice with the

7. Crowley JT, Toledo AM, LaRocca TJ, Coleman JL, London E,

vaccinal strain Bacillus Calmette–Gue´rin [42], indicating

Benach JL: Lipid exchange between Borrelia burgdorferi and

that host colonization can modify the lipids not only in the host cells. PLoS Pathog 2013, 9:e1003109.

This study describes a bidirectional exchange of lipids between the

infected cell, but at the level of the whole tissue. Long

extracellular bacterium Borrelia burgdorferi and its host. Spirochetes

range consequences of the interactions between microbes acquire cholesterol from the plasma membrane of epithelial cells. Free

cholesterol and cholesterol-glycolipids are transferred back to epithelial

and host are well illustrated by the inﬂuence of microbiota

cells through direct contact and through outer membrane vesicles.

on fatty acid absorption in the gut [43]. Another striking

8. Larocca TJ, Pathak P, Chiantia S, Toledo A, Silvius JR, Benach JL,

illustration is given by the parasite Leishmania donovani,

London E: Proving lipid rafts exist: membrane domains in the

which targets pre-miRNA processor Dicer1 to prevent prokaryote Borrelia burgdorferi have the same properties as

eukaryotic lipid rafts. PLoS Pathog 2013, 9:e1003353.

miRNP formation in hepatic cells interacting with the

Following on their previous description of lipid rafts in the membrane of

parasite. As a consequence, Leishmania infection reduces Borrelia burgdorferi, the authors use Fo¨ rster resonance energy transfer

(FRET) and transmission electron microscopy to show that, despite a lack

liver miR-122 and lowers serum cholesterol [44].

of sphingolipids, these domains have characteristic properties of lipid

rafts described in eukaryotes.

9. Gilk SD, Cockrell DC, Luterbach C, Hansen B, Knodler LA,

Conclusion Ibarra JA, Steele-Mortimer O, Heinzen RA: Bacterial colonization

of host cells in the absence of cholesterol. PLoS Pathog 2013,

Exploitation of host lipids by bacteria has consequences

9:e1003107.

at different scales. At the scale of the infectious process

10. Santos AJ, Meinecke M, Fessler MB, Holden DW, Boucrot E:

per se, critical steps of infection such as colonization,

Preferential invasion of mitotic cells by Salmonella reveals

replication, dissemination can depend on the optimal that cell surface cholesterol is maximal during metaphase. J

Cell Sci 2013, 126:2990-2996.

availability of host lipids, sometimes through only small

changes of their distribution or composition. At the scale 11. Mounier J, Boncompain G, Senerovic L, Lagache T, ChrEtien F,

Perez F, Kolbe M, Olivo-Marin J-C, Sansonetti PJ, Sauvonnet N:

of the tissue itself, long lasting parasitic associations can

Shigella effector IpaB-induced cholesterol relocation disrupts

disturb host lipid metabolism so deeply as to ‘reprogram’ the golgi complex and recycling network to inhibit host cell

secretion. Cell Host Microbe 2012, 12:381-389.

it. This aspect of exploitation of host lipids is much less

studied. The possibility to compare lipidomes of puriﬁed 12. Burnaevskiy N, Fox TG, Plymire DA, Ertelt JM, Weigele BA,

Selyunin AS, Way SS, Patrie SM, Alto NM: Proteolytic elimination

bacteria, and of infected versus non-infected tissue, will

of N-myristoyl modiﬁcations by the Shigella virulence factor

certainly help us appreciate these large scale modiﬁ- IpaJ. Nature 2013, 496:106-109.

cations, and their consequence on host physiology.

13. Kim MJ, Kim MK, Kang JS: Involvement of lipid rafts in the

budding-like exit of Orientia tsutsugamushi. Microb Pathog

Acknowledgements 2013, 63C:37-43.

14. Alix E, Mukherjee S, Roy CR: Subversion of membrane transport

We are very grateful to Roland Brosch (Institut Pasteur) for critical reading pathways by vacuolar pathogens. J Cell Biol 2011, 195:943-952.

of the manuscript. FV is funded by the Ministe`re de l’Education Nationale,

15. Weber SS, Ragaz C, Reus K, Nyfeler Y, Hilbi H: Legionella

de la Recherche et de la Technologie (Universite´ Pierre et Marie Curie) and

pneumophila exploits PI(4)P to anchor secreted effector

by the Fondation pour la Recherche Me´dicale. Our research is funded by an

proteins to the replicative vacuole. PLoS Pathog 2006, 2:e46.

ERC Starting Grant to AS (NUChLEAR No. 282046), the ANR (Me´nage a`

trois, ANR-12-BSV2-0009-02), the Institut Pasteur and the Centre National 16. Hsu FS, Zhu WH, Brennan L, Tao LL, Luo ZQ, Mao YX: Structural

de la Recherche Scientiﬁque. basis for substrate recognition by a unique Legionella

www.sciencedirect.com Current Opinion in Microbiology 2014, 17:38–45

44 Host–microbe interactions: bacteria

phosphoinositide phosphatase. Proc Natl Acad Sci U S A 2012, This paper challenges the prevailing view that bacterial lipases evolution

109:13567-13572. was driven by host–bacteria interactions and proposes that the funda-

This paper brings the ﬁrst description of a phosphatidylinositol phospha- mental role of the vgrG-associated lipase families is to serve roles in

tase in Legionella, including structure resolution. The enzyme participates interbacterial competition as effectors of the type VI secretion system

to the conversion of the vacuole in a PI(4)P rich compartment, facilitating translocation apparatus. One of the enzymes examined achieves its

the anchoring of PI(4)P binding effectors to the phagosome membrane. antibacterial activity by degrading phosphatidylethanolamine, the major

component of bacterial membranes.

17. Norris FA, Wilson MP, Wallis TS, Galyov EE, Majerus PW: SopB a

protein required for virulence of Salmonella dublin, is an 33. Lee W, VanderVen BC, Fahey RJ, Russell DG: Intracellular

inositol phosphate phosphatase. Proc Natl Acad Sci U S A 1998, Mycobacterium tuberculosis exploits host-derived fatty

95:14057-14059. acids to limit metabolic stress. J Biol Chem 2013, 288:

6788-6800.

18. Mallo GV, Espina M, Smith AC, Terebiznik MR, Aleman A,

Degradation of cholesterol by Mycobacterium tuberculosis generates

Finlay BB, Rameh LE, Grinstein S, Brumell JH: SopB promotes

propionyl-CoA with potentially toxic consequences. The intracellular

phosphatidylinositol 3-phosphate formation on Salmonella

bacteria can exploit host lipid stores to expand the acetyl-CoA pool

vacuoles by recruiting Rab5 and Vps34. J Cell Biol 2008,

and alleviate propionate-mediated stress. The biochemical data beauti-

182:741-752.

fully illustrate the tight interplay between the lipid metabolism of the host

and the bacteria, a feature which is key to mycobacterium success as

19. Rogers LD, Brown NF, Fang Y, Pelech S, Foster LJ: pathogen.

Phosphoproteomic analysis of Salmonella-infected cells

identiﬁes key kinase regulators and SopB-dependent host

34. Son MS, Matthews WJ Jr, Kang Y, Nguyen DT, Hoang TT: In vivo

phosphorylation events. Sci Signal 2011, 4:rs9.

evidence of Pseudomonas aeruginosa nutrient acquisition

and pathogenesis in the lungs of cystic ﬁbrosis patients. Infect

20. Tahoun A, Mahajan S, Paxton E, Malterer G, Donaldson DS,

Immun 2007, 75:5313-5324.

Wang D, Tan A, Gillespie TL, O’Shea M, Roe AJ et al.: Salmonella

transforms follicle-associated epithelial cells into M cells to

35. Krivan HC, Franklin DP, Wang W, Laux DC, Cohen PS:

promote intestinal invasion. Cell Host Microbe 2012, 12:645-

Phosphatidylserine found in intestinal mucus serves as a sole

656.

source of carbon and nitrogen for salmonellae and

Escherichia coli. Infect Immun 1992, 60:3943-3946.

21. Viner R, Chetrit D, Ehrlich M, Segal G: Identiﬁcation of two

Legionella pneumophila effectors that manipulate host

36. Pride AC, Herrera CM, Guan Z, Giles DK, Trent MS: The outer

phospholipids biosynthesis. PLoS Pathog 2012, 8:e1002988.

surface lipoprotein VolA mediates utilization of exogenous

lipids by Vibrio cholerae. mBio 2013, 4:e00305-e00313.

22. Mital J, Miller NJ, Fischer ER, Hackstadt T: Speciﬁc chlamydial

In contrast to most Gram-negative organisms Vibrio cholerae is capable

inclusion membrane proteins associate with active Src family

of utilizing very-long-chain fatty acids from the surrounding environment.

kinases in microdomains that interact with the host

A surface-exposed lipoprotein VolA is required for processing exogenous

microtubule network. Cell Microbiol 2010, 12:1235.

lysophosphatidylcholine. VolA functions as a lipase liberating a fatty acid,

which is then transported into the bacteria, serving as a carbon source, or

23. Elwell CA, Engel JN: Lipid acquisition by intracellular

shunted into phospholipid synthesis for membrane assembly.

Chlamydiae. Cellular Microbiol 2012, 14:1010-1018.

37. Pandey AK, Sassetti CM: Mycobacterial persistence requires

24. Cocchiaro JL, Kumar Y, Fischer ER, Hackstadt T, Valdivia RH:

the utilization of host cholesterol. Proc Natl Acad Sci U S A

Cytoplasmic lipid droplets are translocated into the lumen of

2008, 105:4376-4380.

the Chlamydia trachomatis parasitophorous vacuole. Proc

Natl Acad Sci U S A 2008, 105:9379-9384.

38. Daniel J, Maamar H, Deb C, Sirakova TD, Kolattukudy PE:

25. Melo RCN, Dvorak AM: Lipid body–phagosome interaction in Mycobacterium tuberculosis uses host triacylglycerol to

macrophages during infectious diseases: host defense or accumulate lipid droplets and acquires a dormancy-like

pathogen survival strategy? PLoS Pathog 2012, 8:e1002729. phenotype in lipid-loaded macrophages. PLoS Pathog 2011,

7:e1002093.

26. Saka HA, Valdivia R: Emerging roles for lipid droplets in

immunity and host–pathogen interactions. Annu Rev Cell Dev 39. Singh V, Jamwal S, Jain R, Verma P, Gokhale R, Rao KVS:

Biol 2012, 28:411-437. Mycobacterium tuberculosis-driven targeted recalibration of

macrophage lipid homeostasis promotes the foamy

27. Xiong Q, Rikihisa Y: Subversion of NPC1 pathway of cholesterol phenotype. Cell Host Microbe 2012, 12:669-681.

transport by Anaplasma phagocytophilum. Cell Microbiol 2012,

14:560-576. 40. Howe D, Heinzen RA: Coxiella burnetii inhabits a cholesterol-

Anaplasma phagocytophilum gets cholesterol from Niemann–Pick type C rich vacuole and inﬂuences cellular cholesterol metabolism.

(NPC1) vesicles of the host, which are translocated into the vacuole in Cell Microbiol 2006, 8:496-507.

which the bacteria grow. Not only do the bacteria capture preexisting

41. Roca FJ, Ramakrishnan L: TNF dually mediates resistance and

NPC1 vesicles, they induce the expansion of a subpopulation of vesicles,

susceptibility to mycobacteria via mitochondrial reactive

thereby ensuring cholesterol supply

oxygen species. Cell 2013, 153:521-534.

28. O’Brien DK, Melville SB: Effects of Clostridium perfringens

42. Jackson SK, Stark JM, Taylor S, Harwood JL: Changes in

alpha-toxin (PLC) and perfringolysin O (PFO) on cytotoxicity to

phospholipid fatty acid composition and triacylglycerol

macrophages, on escape from the phagosomes of

content in mouse tissues after infection with bacille Calmette–

macrophages, and on persistence of C. perfringens in host

Guerin. Br J Exp Pathol 1989, 70:435-441.

tissues. Infect Immun 2004, 72:5204-5215.

43. Semova I, Carten JD, Stombaugh J, Mackey LC, Knight R,

29. Creasey EA, Isberg RR: The protein SdhA maintains the

Farber SA, Rawls JF: Microbiota regulate intestinal absorption

integrity of the Legionella-containing vacuole. Proc Natl Acad

and metabolism of fatty acids in the zebraﬁsh. Cell Host

Sci U S A 2012, 109:3481-3486.

Microbe 2012, 12:277-288.

30. Simeone R, Bobard A, Lippmann J, Bitter W, Majlessi L, Brosch R, This study in zebraﬁsh revealed that the microbiota stimulated fatty acid

Enninga J: Phagosomal rupture by Mycobacterium uptake and formation of lipid bodies in the intestinal epithelium and liver.

tuberculosis results in toxicity and host cell death. PLoS Microbiota increased the number of lipid bodies in a diet-dependent

Pathog 2012, 8:e1002507. manner and different members of the intestinal microbiota promoted fatty

acid absorption via distinct mechanisms.

31. Sitkiewicz I, Stockbauer KE, Musser JM: Secreted bacterial

phospholipase A2 enzymes: better living through 44. Ghosh J, Bose M, Roy S, Bhattacharyya SN: Leishmania

phospholipolysis. Trends Microbiol 2007, 15:63-69. donovani targets Dicer1 to downregulate miR-122, lower

serum cholesterol, and facilitate murine liver infection. Cell

32. Russell AB, LeRoux M, Hathazi K, Agnello DM, Ishikawa T, Host Microbe 2013, 13:277-288.

Wiggins PA, Wai SN, Mougous JD: Diverse type VI secretion

phospholipases are functionally plastic antibacterial 45. Lin M, Rikihisa Y: Ehrlichia chaffeensis and Anaplasma

effectors. Nature 2013, 496:508-512. phagocytophilum lack genes for lipid A biosynthesis and

Current Opinion in Microbiology 2014, 17:38–45 www.sciencedirect.com

Exploitation of host lipids by bacteria Vromman and Subtil 45

incorporate cholesterol for their survival. Infect Immun 2003, 54. Hatch GM, McClarty G: Phospholipid composition of puriﬁed

71:5324-5331. Chlamydia trachomatis mimics that of the eucaryotic host cell.

Infect Immun 1998, 66:3727-3735.

46. Arellano-Reynoso B, Lapaque N, Salcedo S, Briones G,

Ciocchini AE, Ugalde R, Moreno E, Moriyon I, Gorvel JP: Cyclic 55. Carabeo RA, Mead DJ, Hackstadt T: Golgi-dependent transport

beta-1,2-glucan is a Brucella virulence factor required for of cholesterol to the Chlamydia trachomatis inclusion. Proc

intracellular survival. Nat Immunol 2005, 6:618-625. Natl Acad Sci U S A 2003, 100:6771-6776.

47. Livermore BP, Bey RF, Johnson RC: Lipid metabolism of 56. Haque M, Hirai Y, Yokota K, Mori N, Jahan I, Ito H, Hotta H, Yano I,

Borrelia hermsi. Infect Immun 1978, 20:215-220. Kanemasa Y, Oguma K: Lipid proﬁle of Helicobacter spp.:

presence of cholesteryl glucoside as a characteristic feature.

48. Schroder NW, Schombel U, Heine H, Gobel UB, Zahringer U,

J Bacteriol 1996, 178:2065-2070.

Schumann RR: Acylated cholesteryl galactoside as a novel

immunogenic motif in Borrelia burgdorferi sensu stricto. J Biol 57. Testerman TL, McGee DJ, Mobley HL: Helicobacter pylori growth

Chem 2003, 278:33645-33653. and urease detection in the chemically deﬁned medium Ham’s

F-12 nutrient mixture. J Clin Microbiol 2001, 39:3842-3850.

49. Hackstadt T, Scidmore MA, Rockey DD: Lipid metabolism in

Chlamydia trachomatis-infected cells: directed trafﬁcking of 58. de Chastellier C, Thilo L: Cholesterol depletion in

Golgi-derived sphingolipids to the chlamydial inclusion. Proc Mycobacterium avium-infected macrophages overcomes the

Natl Acad Sci U S A 1995, 92:4877-4881. block in phagosome maturation and leads to the reversible

sequestration of viable mycobacteria in phagolysosome-

50. Hackstadt T, Rockey DD, Heinzen RA, Scidmore MA: Chlamydia

derived autophagic vacuoles. Cell Microbiol 2006, 8:242-256.

trachomatis interrupts an exocytic pathway to acquire

endogenously synthesized sphingomyelin in transit from the 59. Layre E, Sweet L, Hong S, Madigan CA, Desjardins D, Young DC,

Golgi apparatus to the plasma membrane. EMBO J 1996, Cheng T-Y, Annand JW, Kim K, Shamputa IC et al.: A comparative

15:964-977. lipidomics platform for chemotaxonomic analysis of

Mycobacterium tuberculosis. Chem Biol 2011, 18:1537-1549.

51. Derre I, Swiss R, Agaisse H: The lipid transfer protein CERT

interacts with the Chlamydia inclusion protein IncD and 60. Salman M, Rottem S: The cell membrane of Mycoplasma

participates to ER-Chlamydia inclusion membrane contact penetrans: lipid composition and phospholipase A1 activity.

sites. PLoS Pathog 2011, 7:e1002092. Biochim Biophys Acta 1995, 1235:369-377.

52. Elwell CA, Jiang S, Kim JH, Lee A, Wittmann T, Hanada K, 61. Zeiman E, Tarshis M, Rottem S: Mycoplasma penetrans under

Melancon P, Engel JN: Chlamydia trachomatis co-opts GBF1 nutritional stress: inﬂuence on lipid and lipoprotein proﬁles

and CERT to acquire host sphingomyelin for distinct roles and on the binding to and invasion of HeLa cells. FEMS

during intracellular development. PLoS Pathog 2011, Microbiol Lett 2008, 287:243-249.

7:e1002198.

62. Catron DM, Sylvester MD, Lange Y, Kadekoppala M, Jones BD,

53. Cox JV, Naher N, Abdelrahman YM, Belland RJ: Host

Monack DM, Falkow S, Haldar K: The Salmonella-containing

HDL biogenesis machinery is recruited to the

vacuole is a major site of intracellular cholesterol

inclusion of Chlamydia trachomatis-infected cells and

accumulation and recruits the GPI-anchored protein CD55.

regulates chlamydial growth. Cell Microbiol 2012,

Cell Microbiol 2002, 4:315-328.

14:1497-1512.

Several pathways for host lipid acquisition by chlamydiae have been 63. Catron DM, Lange Y, Borensztajn J, Sylvester MD, Jones BD,

reported, yet this study unveils another trick. The bacteria co-opt the host Haldar K: Salmonella enterica serovar Typhimurium requires

cell lipid transport system involved in the formation of HDL to acquire nonsterol precursors of the cholesterol biosynthetic pathway

lipids such as phosphatidylcholine. for intracellular proliferation. Infect Immun 2004, 72:1036-1042.

www.sciencedirect.com Current Opinion in Microbiology 2014, 17:38–45