DÉTERMINANTS DE LA MOBILISATION DE L’ILOT GÉNOMIQUE DE RÉSISTANCE SGI1

par

Romain Durand

Thèse présentée au Département de biologie en vue de l’obtention du grade de docteur ès sciences (Ph.D.)

FACULTÉ DES SCIENCES UNIVERSITÉ DE SHERBROOKE

Sherbrooke, Québec, Canada, novembre 2020

Le 26 novembre 2020

Le jury a accepté la thèse de Monsieur Romain Durand dans sa version finale.

Membres du jury

Professeur Vincent Burrus Directeur de recherche Département de biologie Université de Sherbrooke

Professeur Sébastien Rodrigue Codirecteur de recherche Département de biologie Université de Sherbrooke

Docteur Yoshiharu Yamaichi Évaluateur externe Département de biologie des génomes Institut de Biologie Intégrative de la Cellule

Docteur Benoît Leblanc Évaluateur interne Département de biologie Université de Sherbrooke

Professeure Pascale B. Beauregard Président-rapporteur Département de biologie Université de Sherbrooke

REMERCIEMENTS

J’aimerais tout d’abord remercier mon directeur de thèse, Vincent, qui a cru en moi très tôt, m’a transmis sa passion pour la science et m’a toujours poussé à dépasser mes limites. Merci aux Fonds de Recherche du Québec – Nature et Technologies, pour avoir financé une partie de mon doctorat. Merci à Sébastien, Pascale et Benoît pour leurs conseils avisés.

Merci aux membres passés et présents du laboratoire, particulièrement Nicolas Carraro pour avoir significativement contribué à lancer ma carrière de jeune chercheur. Merci également à Didier de m’avoir mis sur le chemin de la recherche scientifique. Merci à ceux qui m’ont côtoyé et soutenu au cours de ces 5 dernières années. Merci à ceux qui, malgré la distance, m’ont continuellement supporté. Victor, plus de vingt ans d’amitié ont mis du vent dans mes voiles.

Je tiens également à remercier ma (grande) famille qui m’a perpétuellement soutenu au travers des épreuves : mon père Christophe, ma mère Valérie, mes sœurs Lucile et Juliette et mon frère Mathis, ainsi que ma belle-famille de Tahiti pour leur accueil chaleureux.

Une petite pensée pour mon chat CRISPR-Cat9, qui m’a (presque) toujours accueilli avec enthousiasme après de longues journées de travail.

Enfin, un immense merci à mon épouse Teura. Je n’aurais jamais été capable de gravir cette montagne sans ton soutien infatigable. Je ne compte plus les obstacles que tu m’as aidé à surmonter. Maintenant qu’une page se tourne, je n’ai qu’une seule hâte : voir ce que la vie nous réserve à tous les deux.

SOMMAIRE

Les plasmides conjugatifs appartenant au groupe d’incompatibilité C (IncC) sont des acteurs majeurs de la dissémination des gènes de résistance aux antibiotiques chez les Gammaprotéobactéries. Malheureusement, bien qu’ils figurent dans des centaines d’études épidémiologiques, ils ne sont à l’affiche que d’une poignée d’études mécanistiques. La caractérisation d’AcaCD, complexe majeur d’activation du transfert des plasmides IncC, a constitué une étape clé dans la compréhension du cycle de vie de ces éléments. La liaison de cet activateur transcriptionnel en amont de nombreux opérons impliqués dans le transfert conjugatif permet de déclencher conjointement l’assemblage d’un système de sécrétion de type IV et l’initiation du transfert conjugatif. Elle permet aussi de déclencher l’excision et la mobilisation subséquente de plusieurs familles distinctes d’ilots génomiques mobilisables.

L’ilot génomique de résistance Salmonella Genomic Island 1 (SGI1), initialement isolé à partir de Salmonella enterica, est aujourd’hui considéré comme une large famille d’ilots génomiques mobilisables. Un grand nombre de variants, conférant régulièrement des résistances aux antibiotiques, ont été isolés à partir de nombreuses espèces de Gammaprotéobactéries. Bien qu’il n’encode pas de système de sécrétion de type IV complet, SGI1 peut être spécifiquement mobilisé par un plasmide IncC. L’excision de SGI1 est médiée par un gène dont l’expression est activée par AcaCD. Après circularisation, le transfert conjugatif est initié à l’origine de transfert de l’ilot, qui transloque alors au travers du pore de conjugaison encodé par le plasmide.

Ce projet de doctorat a permis de montrer que SGI1 encode trois sous-unités fonctionnelles du système de sécrétion de type IV. Celles-ci sont incorporées à la place des sous-unités homologues encodées par le plasmide IncC. Notre étude a montré que ce remodelage du pore de conjugaison par SGI1 lui permettait d’optimiser sa propagation. Nous avons également montré le rôle clé du complexe SgaCD encodé par l’ilot, homologue au complexe majeur d’activation du transfert des plasmides IncC AcaCD. SgaCD s’est révélé essentiel à l’excision et à la réplication de SGI1, entrainant ultimement une déstabilisation du plasmide IncC résidant dans la même cellule. Enfin, nous présentons des données préliminaires montrant l’existence d’un nouveau déterminant de la mobilisation de SGI1 et d’un répresseur du transfert des plasmides IncC encodé par SGI1.

Ensemble, nos résultats participent à caractériser la relation complexe entre deux familles d’éléments génétiques mobiles. Nous avons pu identifier de nombreux acteurs impliqués dans la dissémination de ces éléments, lesquels constituent des cibles potentielles pour le développement de composés visant à freiner la propagation des résistances aux antibiotiques.

Mots clés : résistance aux antibiotiques, transfert conjugatif, système de sécrétion de type IV, plasmides conjugatifs, ilots génomiques, régulation transcriptionnelle, Gammaprotéobactéries, IncC, SGI1

x

TABLE DES MATIÈRES

CHAPITRE 1 Introduction ...... 1 1.1 Transfert horizontal de gènes ...... 1 1.2 Conjugaison ...... 5 1.2.1. Mécanisme global ...... 5 1.2.2. Classification des T4SS ...... 7 1.2.3. T4SS de type P ...... 10 1.2.3.1 Structure ...... 10 1.2.3.2 Mécanisme de la conjugaison (type P) ...... 12 1.2.3.3 Pilus ...... 14 1.2.4. T4SS de type F ...... 15 1.2.4.1 Structure ...... 16 1.2.4.2 Pilus ...... 18 1.2.4.3 Mpf et Mps ...... 19 1.2.4.4 Exclusion ...... 20 1.2.4.5 Relaxase TraI du plasmide F ...... 21 1.3 Les éléments conjugatifs ...... 22 1.3.1. Régulation ...... 23 1.3.1.1 Régulation du transfert des ICE ...... 23 1.3.1.2 Cycle de vie des ICE ...... 26 1.3.1.3 Régulation du transfert des plasmides conjugatifs ...... 28 1.3.2. Éléments modèles au laboratoire ...... 29 1.3.2.1 Les ICE SXT/R391 ...... 29 1.3.2.2 Les plasmides A/C ...... 31 1.3.2.3 MGIVchHai6 ...... 35 1.3.2.4 SGI1 ...... 36 1.3.3. Bilan des interactions entre différentes familles d’éléments ...... 39 1.4 Hypothèses du projet de doctorat ...... 42 1.5 Objectifs du projet de doctorat ...... 43 CHAPITRE 2 Salmonella Genomic Island 1 (SGI1) Reshapes the Mating Apparatus of IncC Conjugative Plasmids to Promote Self-Propagation ...... 45 2.1 Introduction ...... 45 2.1.1. Pertinence ...... 45 2.1.2. Apport des auteurs ...... 46 2.1.3. Référence ...... 47 2.2 Article ...... 48 2.2.1. Page titre ...... 48 2.2.2. Abstract ...... 49 2.2.3. Author summary ...... 49 2.2.4. Introduction ...... 50 2.2.5. Results ...... 54

2.2.5.1 SGI1 traS genes are induced by the master activator of transfer of IncC plasmids ...... 54

2.2.5.2 TraNC, TraHC and TraGC are essential for IncC plasmid transfer ...... 56

2.2.5.3 The traS gene cluster is important for SGI1 mobilization ...... 58 2.2.5.4 SGI1 complements a defective IncC-encoded T4SS with functional subunits58

2.2.5.5 TraGS is required for optimal SGI1 transfer ...... 59

2.2.5.6 TraHS is specifically required with TraGS for SGI1 optimal transfer ...... 61

2.2.5.7 TraNC and TraNS proteins are exchangeable ...... 63

2.2.5.8 TraNS enhances SGI1 transfer through the hybrid T4SS ...... 63 2.2.5.9 IncC plasmids exert entry exclusion ...... 65

2.2.5.10 TraGS disables entry exclusion between cells bearing IncC plasmids ...... 65 2.2.6. Discussion ...... 69 2.2.7. Materials and methods ...... 73 2.2.7.1 Bacterial strains and media ...... 73 2.2.7.2 Mating assays ...... 73 2.2.7.3 Molecular biology methods ...... 74

xii

2.2.7.4 Plasmid and strain construction ...... 74 2.2.7.5 β-galactosidase assays ...... 76 2.2.8. Acknowledgments ...... 77 2.2.9. References ...... 77 2.2.10. Supporting information ...... 82 CHAPITRE 3 Crucial role of Salmonella Genomic Island 1 (SGI1) Master Activator in Parasitism of IncC plasmids ...... 85 3.1 Introduction ...... 85 3.1.1. Pertinence ...... 85 3.1.2. Apport des auteurs ...... 86 3.1.3. Référence ...... 87 3.2 Article ...... 88 3.2.1. Page titre ...... 88 3.2.2. Abstract ...... 89 3.2.3. Introduction ...... 89 3.2.4. Material and methods ...... 92 3.2.4.1 Bacterial strains and media ...... 92 3.2.4.2 Molecular biology ...... 94 3.2.4.3 Plasmids and strains constructions ...... 94 3.2.4.4 ChIP-exo assays ...... 95 3.2.4.5 Cappable-seq and RNA-seq assays ...... 96 3.2.4.6 Illumina sequencing library preparation ...... 96 3.2.4.7 Bioinformatic analyses ...... 97 3.2.4.8 β-galactosidase assays ...... 98 3.2.4.9 qPCR assays ...... 98 3.2.4.10 Cohabitation assays ...... 99 3.2.4.11 Statistical analyses and figures ...... 99 3.2.5. Results ...... 100 3.2.5.1 SGI1 rescues the transfer of an IncC plasmid lacking its master activator of transfer ...... 100

xiii

3.2.5.2 SgaCD activates all AcaCD-activatable promoters and generates the same binding footprint as AcaCD ...... 102 3.2.5.3 Overexpression of sgaDC significantly affects the transcriptome ...... 107 3.2.5.4 Activation of gene expression by SgaCD is comparable to AcaCD ...... 109 3.2.5.5 SgaCD is essential for SGI1 replication ...... 111 3.2.5.6 SgaCD-activated replication of SGI1 destabilises the helper IncC plasmid.. 113 3.2.5.7 A single nucleotide polymorphism (SNP) abolishes SgaCD activity of SGI1-C ...... 115 3.2.6. Discussion ...... 116 3.2.7. Availability ...... 122 3.2.8. Accession numbers ...... 122 3.2.9. Acknowledgements ...... 123 3.2.10. References ...... 123 3.2.11. Supplementary data ...... 129 3.2.11.1 References ...... 148 CHAPITRE 4 Matériel et méthodes ...... 151 4.1 Souches bactériennes et milieux de culture ...... 151 4.2 Essais de conjugaison ...... 153 4.3 Méthodes de biologie moléculaire ...... 154 4.4 Constructions génétiques ...... 154 4.5 qPCR ...... 157 4.6 Essai de double hybride bactérien ...... 158 4.7 Phylogénie ...... 158 4.8 Génomique comparative ...... 159 CHAPITRE 5 Résultats ...... 161 5.1 Rôle du gène acaB dans la mobilisation de SGI1 ...... 161 5.1.1. acaB est essentiel au transfert des plasmides IncC ou à la mobilisation de SGI1 ...... 161 5.1.2. acaB est impliqué dans la stabilité des plasmides IncC ...... 165 5.1.3. AcaB est conservée chez plusieurs familles d’éléments génétiques mobiles ..... 169

xiv

5.1.4. S063 n’est pas essentiel à la mobilisation de MGIVflInd1 par SXT ...... 171 5.2 Caractérisation du gène S010 encodé par l’ilot génomique de résistance SGI1 ...... 172 5.2.1. S010 réprime le transfert des plasmides IncC et le co-transfert ...... 172 5.2.2. S010 inhibe légèrement le transfert de MGIVchHai6 ...... 175

5.2.3. La protéine encodée par S010 interagit directement avec MobIC ...... 177 5.3 La famille SGI1 étendue ...... 178 5.3.1. Des éléments de type SGI1 intégrés dans des sites alternatifs ...... 178 5.3.2. Caractérisation de MGIVchUSA3 ...... 181 CHAPITRE 6 Discussion et conclusion ...... 185 6.1 Un ilot génomique portant des gènes de transfert ...... 186 6.1.1. L’importance des gènes tra portés par SGI1 ...... 186 6.1.2. GIsul2 ...... 187 6.1.3. Échappement à l’exclusion d’entrée ...... 188

6.1.4. Rôle de TraNS ...... 191 6.2 Étude du gène acaB ...... 191 6.3 Un ilot génomique encodant un inhibiteur de la fertilité des plasmides IncC ...... 195 6.3.1. Initiation du transfert des plasmides IncC et de SGI1 ...... 195 6.3.2. Inhibition de la fertilité ...... 198 6.4 Un ilot génomique encodant un activateur transcriptionnel ...... 199 6.4.1. Déstabilisation des plasmides IncC ...... 199 6.4.2. Comparaisons avec l’activateur flagellaire d’E. coli ...... 202 6.5 Perspectives ...... 207 6.5.1. Caractérisation d’Acr2 ...... 207 6.5.2. Études structurales ...... 209 6.5.3. Éléments génétiques mobiles et fluorescence ...... 210 6.5.4. Conclusion ...... 211 ANNEXE 1 ...... 213

xv

ANNEXE 2 ...... 214 BIBLIOGRAPHIE ...... 216

xvi

LISTE DES TABLEAUX

Chapitre 1 Tableau 1.1. Déterminants pour la mobilisation de quelques MGI ...... 41

Chapitre 2 Table 2.1. Protein homologs encoded by IncC plasmids and SGI1...... 53 Table 2.2. Strains and plasmids used in this study...... 64 Table S2.1. Primers used in this study ...... 83

Chapitre 3 Table 3.1. Strains and elements used in this study...... 93 Table S3.1. Oligonucleotides used in this study...... 129 Table S3.2. Bioinformatic analyses supplemental data...... 131 Table S3.3. Top 10 up- and down-regulated MG1655 genes upon sgaDC overexpression. 134 Table S3.4. Predicted AcaCD/SgaCD-responsive promoters in other families of conjugative plasmids...... 135

Chapitre 4 Tableau 4.1. Souches bactériennes et éléments génétiques utilisés dans cette étude ...... 151

Annexe 2 Tableau A2.1.Liste des amorces utilisées ...... 214

LISTE DES FIGURES

Chapitre 1 Figure 1.1. Représentation schématique d'un « arbre de vie » tenant compte du transfert horizontal de gènes ...... 2 Figure 1.2. Catégories majeures du transfert horizontal de gènes ...... 3 Figure 1.3. Mécanisme global de la conjugaison ...... 6 Figure 1.4. Gènes conservés chez les T4SS de type P, F et I ...... 9 Figure 1.5. Structure d’une partie du T4SS de R388 (type P) observée en microscopie électronique à transmission ...... 11 Figure 1.6. Représentation schématique d’un T4SS de type P ...... 12 Figure 1.7. Synthèse du pilus (T4SS de type P) ...... 15 Figure 1.8. Structure du T4SS de type F ...... 17 Figure 1.9. Élongation et rétraction dynamique du pilus de type F observées en microscopie confocale ...... 18 Figure 1.10. Schémas de régulation du transfert de plusieurs familles d’ICE ...... 25 Figure 1.11. Schéma explicatif du cycle de vie d’un ICE ...... 27 Figure 1.12. Représentation schématique du noyau de gènes conservés chez les ICE SXT/R391 ...... 30 Figure 1.13. Représentation schématique du noyau de gènes conservés chez les plasmides A/C ...... 32 Figure 1.14. Régulation du transfert des plasmides A/C ...... 34 Figure 1.15. Représentation schématique de l’ilot génomique MGIVchHai6 ...... 36 Figure 1.16. Représentation schématique de l’ilot génomique SGI1 ...... 37 Figure 1.17. Modèles proposés pour 3 mécanismes de mobilisation d’ilots génomiques ...... 40

Chapitre 2 Figure 2.1. Linear schematic representation of the core sequence of IncC plasmids and of Salmonella genomic island 1 (SGI1)...... 51

Figure 2.2. Regulation of expression of int, xis, traNS and traHGS of SGI1...... 55

Figure 2.3. Role of TraN/G/HC on conjugative transfer of IncC plasmids...... 57

Figure 2.4. Role of traC and traS genes on conjugative transfer of IncC plasmids and SGI1...... 60

Figure 2.5. Suppression by traGS of IncC entry exclusion...... 67 Figure 2.6. Schematic representation of mating pore configurations and impact on IncC plasmids and SGI1 transfer efficiency...... 71

Figure S2.1. Effect of traC/S genes on cotransfer of IncC plasmids and SGI1...... 82

Chapitre 3 Figure 3.1. Role of SgaCD in conjugative transfer of IncC plasmids and SGI1...... 101 Figure 3.2. In-depth analysis of the AcaCD/SgaCD regulon...... 103 Figure 3.3. VAP profiles of AcaCD/SgaCD footprint...... 105 Figure 3.4. Differential expression analysis...... 108 Figure 3.5. Systematic analysis of AcaCD/SgaCD-dependent promoters...... 110 Figure 3.6. Effect of sgaDC and acaDC deletions on SGI1Kn dynamics and pVCR94Sp stability...... 112 Figure 3.7. Effect of sgaDC deletion on incompatibility between SGI1 and IncC plasmids in the absence of selective pressure...... 114 Figure 3.8. Model of regulation of IncC plasmids and SGI1 gene expression...... 118 Figure S3.1. Content of pOPlacZ-derivative vectors used for β-galactosidase assays...... 138 Figure S3.2. Correlation between ChIP-exo and RNA-seq replicates...... 139 Figure S3.3. VAP aggregate profiles of AcaCD/SgaCD footprint...... 142 Figure S3.4. Differential expression analysis...... 144 Figure S3.5. Predicted structures of AcaC, SgaC, SetC, AsaC and AqaC...... 145 Figure S3.6. Interactions between SGI1 and IncC plasmids...... 147

xx

Chapitre 5 Figure 5.1. Rôle d’acaB dans le transfert conjugatif des plasmides IncC ...... 161 Figure 5.2. Mobilisation de SGI1Kn par pVCR94Sp ΔacaB ...... 163 Figure 5.3. Impact de la délétion d’acaB sur les dynamiques de pVCR94Sp et SGI1Kn ...... 166 Figure 5.4. Instabilité de pVCR94Sp ΔacaB en présence de SGI1Kn ...... 168 Figure 5.5. Analyse phylogénétique de la protéine AcaB ...... 170 Figure 5.6. Mobilisation de MGIVflInd1 par SXTΔS063 ...... 171 Figure 5.7. Transfert d’un plasmide IncC en présence de SGI1Kn ΔS010 ...... 173 Figure 5.8. Impact de S010 sur le transfert conjugatif des plasmides IncC ...... 174 Figure 5.9. Effet de S010 sur la mobilisation de MGIVchHai6 ...... 176

Figure 5.10. Interaction directe entre les protéines encodées par S010 et mobIC évaluée par double hybride bactérien ...... 177 Figure 5.11. Représentation schématique des ilots atypiques apparentés à SGI1 ...... 180 Figure 5.12. Mobilisation de MGIVchUSA3 ...... 182

Chapitre 6 Figure 6.1. Rôle adaptatif de TraG/VirB6 ...... 189 Figure 6.2. Comparaison des structures d’AcaB et d’autres protéines RHH ...... 194 Figure 6.3. Mécanisme du système de partition ParMRC ...... 200 Figure 6.4. Structure du complexe FlhDC ...... 203 Figure 6.5. Initiation de la transcription bactérienne ...... 205

xxi

LISTE DES ABRÉVIATIONS ET DES SIGLES

ADN Acide désoxyribonucléique ADN-T ADN de transfert, ne s’applique que pour le plasmide Ti d’Agrobacterium tumefaciens ARN Acide ribonucléique. ARNm pour ARN messager. ATP Adénosine triphosphate att Sites d’attachement (attL et attR : gauche et droite, attB : bactérien, attP : du phage, attI : de l’ICE) c-di- Cyclic dimeric guanosine monophosphate. Molécule signal impliquée dans GMP d’importantes voies de régulation cellulaire. Cas CRISPR associated protein (protéine associée au système CRISPR) ChIP-exo Chromatin immunoprecipitation coupled with exonuclease treatment (immunoprécipitation de la chromatine liée à un traitement aux exonucléases) CRISPR Clustered Regularly Interspaced Short Palindromic Repeats. Système immunitaire adaptatif bactérien. Dtr DNA transfer and replication factors (facteurs de transfert et de réplication de l’ADN) GGI Gonococcal Genetic Island (ilot génétique de Neisseria gonorrhoeae) GTA Gene Transfer Agents. Désigne une famille d’éléments génétiques mobiles. H-NS Histone-like nucleoid structuring protein (protéine impliquée dans la structure du nucléoïde) ICE Integrative and Conjugative Element (Élément intégratif et conjugatif) Inc Incompatibilité (les plasmides appartenant au groupe d’incompatibilité C sont abrégés « IncC ») LB Lysogeny broth, milieu de culture utilisé en routine MGI Mobilizable Genomic Island (ilot génomique mobilisable) Mpf Mating pair formation (formation de la paire conjugative) Mps Mating pair stabilization (stabilisation de la paire conjugative) oriT Origine de transfert pb Paires de bases (nucléotidiques). Unité désignant la longueur d’une séquence d’ADN. qPCR Quantitative polymerase chain reaction (réaction de polymérisation en chaine quantitative) RBS Ribosome binding site (site de liaison au ribosome) RDF Recombination directionality factor (facteur d’orientation de la recombinaison) RHH Ribbon-helix-helix. Désigne une famille de facteurs de transcription bactériens. RT Rétrotranscription. Lorsque celle-ci est suivie d’une qPCR on parle de RT-qPCR. SGI1 Salmonella Genomic Island 1 (ilot génomique 1 de Salmonella) SPRI Solid phase reversible immobilisation (immobilisation réversible sur phase solide) SXT Sulfaméthoxazole triméthoprime. Désigne l’ICE éponyme et une partie de son profil de résistance T4SS Type 4 Secretion System (système de sécrétion de type IV) Ti Tumor-inducing plasmid (Plasmide induisant des tumeurs). Le plasmide Ti est le principal déterminant de la virulence d’Agrobacterium tumefaciens.

TnSeq Mutagenèse par transposition aléatoire couplée à du séquençage à haut débit TS Translocation signals (Signaux de translocation) UFC Unité formant colonie X-gal 5-bromo-4-chloro-3-indolyl- β-D-galactopyranoside

xxiv

CHAPITRE 1 INTRODUCTION

1.1 Transfert horizontal de gènes

Les bactéries sont en perpétuelle évolution, non seulement capables de transmettre leur matériel génétique à leur descendance (transfert vertical de gènes) mais également entre elles (transfert horizontal de gènes). Ces échanges constants permettent notamment l’adaptation des bactéries à de nouveaux environnements. Il s’agit aussi du seul moyen permettant de compenser l’accumulation de mutations qui pourraient engendrer un impact négatif sur la viabilité. Contrairement aux eucaryotes, pour lesquels la recombinaison lors de la méiose permet un réassortiment du patrimoine génétique, les procaryotes ne peuvent compter que sur le transfert horizontal. Celui-ci va permettre d’assurer leur diversité et joue donc un rôle fondamental dans leur survie. On observe dès lors une grande différence entre, par exemple, les bactéries ayant développé une relation endosymbiotique ou endoparasitique, lesquelles arborent un génome très stable et pratiquement fermé aux échanges, et certains pathogènes dont la virulence repose uniquement sur un plasmide acquis par transfert horizontal. Le modèle darwinien de bifurcation à partir d’un ancêtre commun n’est donc pas un modèle suffisamment robuste pour expliquer l’évolution des bactéries. Cette théorie est d’autant plus difficile à argumenter que l’utilisation de deux marqueurs phylogénétiques différents suffit à révéler deux histoires évolutives distinctes au sein d’un même génome. L’évolution des bactéries résulte plutôt d’une combinaison de la bifurcation et de la réticulation liée aux nombreux échanges de gènes (Figure 1.1).

Figure 1.1. Représentation schématique d'un « arbre de vie » tenant compte du transfert horizontal de gènes Tiré de (Martin, 1999).

Le transfert horizontal de gènes peut impliquer différents mécanismes, généralement groupés en trois catégories : la transformation, la transduction et la conjugaison (Figure 1.2).

2

Figure 1.2. Catégories majeures du transfert horizontal de gènes Le transfert horizontal de gènes regroupe trois catégories principales. La conjugaison est le transfert d’ADN nécessitant un contact entre deux cellules à l’aide d’une machinerie dédiée. La transduction implique le transfert d’ADN encapsidé. La transformation est l’acquisition d’ADN exogène libre dans le milieu extracellulaire. Tiré de (Sommer et al., 2017).

La transformation bactérienne nécessite deux « partenaires » : de l’ADN exogène et une cellule réceptrice dans un état physiologique particulier appelé « compétence ». Elle a été découverte pour la première fois chez Streptococcus pneumoniae mais a depuis été montrée possible chez de nombreuses autres espèces. Les protéines nécessaires à l’acquisition de l’ADN libre sont rarement exprimées en tout temps. Au contraire, pour la plupart des bactéries, un traitement est nécessaire pour entrer dans un état de compétence, lequel est un prérequis à la transformation (Johnston et al., 2014).

3 La transduction est un terme très global, regroupant plusieurs mécanismes différents. Le point de départ est l’infection d’une bactérie par un phage. Le phage peut être « virulent », un terme majoritairement utilisé pour désigner un phage lytique obligatoire (Hobbs and Abedon, 2016). Dans ce cas, l’infection cause la lyse de la cellule hôte, libérant immédiatement les virions de phage. Le phage peut également être « tempéré », auquel cas l’infection peut être suivie d’un cycle lysogénique, dans lequel le phage est intégré au chromosome. Dès lors, on parle de « prophage », lequel ne mène pas immédiatement à la production/libération de virions. La conversion lysogénique se matérialise par des changements phénotypiques liés à des fonctions encodées par le prophage ou le phage. Le prophage est également capable de s’exciser et d’empaqueter de l’ADN dans une capside. La transduction correspond au transfert de gènes de l’hôte par ce mécanisme vers d’autres bactéries (Touchon et al., 2017). On distingue la transduction généralisée, qui implique l’encapsidation accidentelle de fragments du génome de l’hôte au lieu de l’ADN du phage, de la transduction spécialisée, résultant d’une mauvaise excision et qui entraine l’encapsidation des deux types d’ADN. À noter également, la domestication de prophage pourrait potentiellement être à l’origine d’autres éléments apparentés, comme les « GTA » (Gene Transfer Agents) (Lang et al., 2012a).

Enfin, la conjugaison requiert un contact cellule-cellule et dans la majeure partie des cas un système de sécrétion de type IV (T4SS), lequel va permettre de transloquer un brin d’ADN depuis la cellule donneuse vers la cellule réceptrice. Chez les Actinobactéries cependant, la conjugaison est assurée par un multimère d’une unique translocase d’ADN double brin de type FtsK/SpoIIIE (Vogelmann et al., 2011).

Il est important de noter que d’autres modes de transfert horizontal existent, résultant parfois de combinaisons de modes. C’est le cas du phénomène de « transjugaison » découvert chez la bactérie Thermus thermophilus qui combine la transformation et la conjugaison (Blesa et al., 2017). Plus spécifiquement, un évènement de « transjugaison » implique un contact cellule- cellule, activant la sécrétion d’ADN par la cellule donneuse au travers de la translocase TdtA, puis son acquisition active par transformation par la cellule réceptrice compétente. Chez la

4 même espèce, l’appareil de transformation peut aussi acquérir de l’ADN transféré via des vésicules extracellulaires (Blesa and Berenguer, 2015). Il a également été rapporté que ce mode de transfert pouvait entrainer la dissémination de déterminants métaboliques chez des bactéries à Gram positif (Klieve et al., 2005), ou même des résistances aux antibiotiques chez des bactéries à Gram négatif (Rumbo et al., 2011).

1.2 Conjugaison

La conjugaison est le sujet de nombreuses études depuis plus de 70 ans (Arutyunov and Frost, 2013; Tatum and Lederberg, 1947). Elle représente un intérêt fondamental d’un point de vue épidémiologique, puisque les éléments conjugatifs sont les acteurs principaux de la dissémination de gènes de résistances aux antibiotiques et aux métaux lourds, et de facteurs de virulence. Ces éléments sont regroupés en familles distinctes, comme les plasmides conjugatifs, des plasmides encodant un T4SS nécessaire à leur transfert, ou encore les éléments intégratifs et conjugatifs (ICE) qui diffèrent des plasmides conjugatifs en un point majeur : ils résident intégrés dans le chromosome de leur hôte. Les éléments conjugatifs présentent également un intérêt d’un point de vue évolutif, puisqu’ils peuvent conférer des fonctions importantes comme la fixation d’azote ou encore la dégradation de certaines molécules. Malgré toutes ces années d’études, les mécanismes au cœur de la conjugaison restent cependant encore mal compris.

1.2.1. Mécanisme global

Le transfert conjugatif est initié au niveau d’une région spécifique qu’on appelle « origine de transfert » ou « oriT » (Figure 1.3). Dans le cas d’un transfert simple brin, des protéines dédiées vont permettre le clivage d’un brin d’ADN et sa liaison covalente avec une relaxase. Le complexe ainsi formé constitue le relaxosome. Cette première étape semble plutôt conservée chez de nombreuses bactéries. Une protéine de couplage va permettre la mise en contact du relaxosome avec le T4SS, aussi appelé pore de conjugaison. Le T4SS est composé d’un nombre variable d’unités protéiques, dont plusieurs sont présentes en multiples copies. Il est possible de

5 grouper les protéines présentant une fonction commune (ATPases, formation du pilus) et/ou celles présentant une localisation commune : les protéines situées au niveau de la membrane interne forment le « complexe de la membrane interne » (IMC), celles au niveau de la membrane externe forment le « complexe de la membrane externe » (OMC). La machinerie assemblée s’étend du côté cytoplasmique de la membrane interne jusqu’à la membrane externe et constitue une sorte de « tunnel » au travers duquel ADN et protéines vont pouvoir être transloqués (Grohmann et al., 2018).

Figure 1.3. Mécanisme global de la conjugaison Le mécanisme global de la conjugaison est représenté de façon schématique en prenant un plasmide conjugatif comme famille d’élément modèle.

6 Figure 1.3. Mécanisme global de la conjugaison (suite) L’évènement de transfert est généralement réprimé mais, à la suite d’un stimulus en particulier, la répression est levée et plusieurs opérons clés sont activés (1) : l’opéron mob encode les protéines nécessaires à la formation du relaxosome, tandis que l’opéron mpf encode le T4SS. Le transfert est initié à l’oriT (étoile orange) par la relaxase, qui clive un brin d’ADN et s’y lie de façon covalente (2). La relaxase liée à l’ADN forme le relaxosome, qui est mis en contact avec le T4SS par l’intermédiaire de la protéine de couplage (3). L’ADN est ensuite transloqué au travers du T4SS, entrant alors dans la cellule réceptrice. La conjugaison est un processus réplicatif : le plasmide dans la cellule donneuse est répliqué et continue d’être maintenu de façon habituelle. L’ADN transféré dans la cellule réceptrice est circularisé et le second brin est synthétisé (4) : le plasmide peut dès lors être maintenu dans la cellule réceptrice. Créée avec BioRender.com.

1.2.2. Classification des T4SS

Les T4SS comprennent deux familles majeures : les systèmes de conjugaison et les translocateurs d’effecteurs. Une troisième famille regroupe les échanges (ADN ou protéines) indépendants d’un contact (Cascales and Christie, 2003).

Parmi les translocateurs les mieux décrits, on peut citer Agrobacterium tumefaciens dont le plasmide Ti encode un T4SS permettant la translocation d’ADN-T (ADN transféré) et de protéines effectrices. Ces dernières entrainent la synthèse d’opines et induisent la formation de tumeurs chez la plante infectée (Cascales and Christie, 2003). La protéine CagA, transloquée au travers d’un T4SS par Helicobacter pylori, affecte de nombreux processus impliqués dans la différenciation, la prolifération et la motilité de cellules eucaryotes (Amieva et al., 2003). La protéine RalF, transloquée au travers d’un T4SS par Legionella pneumophila, permet in fine de promouvoir la survie intracellulaire du pathogène (Nagai et al., 2002). Enfin, la sécrétion de la

7 toxine Pertussis de Bordetella pertussis, directement dans le milieu extracellulaire, s’effectue également via un T4SS (Weiss et al., 1993). À ce jour, les translocateurs d’effecteurs n’ont été décrits que chez des pathogènes à Gram négatif ou des symbiotes à Gram négatif (Bhatty et al., 2013). Les échanges indépendants d’un contact sont un peu plus rares. On peut citer le T4SS encodé par l’ilot GGI de Neisseria gonorrhoeae, lequel permet la sécrétion d’ADN dans le milieu en vue d’une acquisition par transformation (Dillard and Seifert, 2001). L’acquisition d’ADN dans le milieu peut également se faire via un T4SS, comme l’attestent les systèmes Cjp/VirB de Campylobacter jejuni et ComB d’H. pylori (Bacon et al., 2000; Hofreuter et al., 2001).

Les systèmes conjugatifs au contraire, sont présents chez pratiquement toutes les espèces bactériennes, ainsi que certaines archées (Guglielmini et al., 2013). De nombreux systèmes ont été décrits jusqu’à présent. Il est important de distinguer le T4SS en lui-même du pilus extracellulaire (le pilus, non retrouvé chez les Gram +, faisant partie du système conjugatif mais pas du T4SS). Plusieurs classifications des T4SS existent, une des plus courantes ayant été établie par Peter Christie de l’Université du Texas. Dans cette classification, on distingue 3 types principaux de T4SS : les types P, F et I (Figure 1.4). Le prototype pour les T4SS de type P est le système présent chez A. tumefaciens, composé des unités « VirB » et de la protéine de couplage VirD4. Puisque ce système est largement caractérisé, la même nomenclature est souvent utilisée pour décrire les unités appartenant à d’autres T4SS. Pour le type F en revanche, on utilise plus souvent la nomenclature « Tra ». Ainsi, dans la Figure 1.4, on constate une homologie fonctionnelle entre VirD4 (type P) et la protéine de couplage TraD (type F). Le type I tient son nom de la machinerie encodée par les plasmides IncI. Il qualifie également le système Dot/Icm retrouvé chez les pathogènes Legionella et Coxiella.

8

Figure 1.4. Gènes conservés chez les T4SS de type P, F et I Les gènes encodant des homologues fonctionnels des unités VirB/VirD4 sont en couleur. Type F : les gènes en gris foncé ont une fonction relative au pilus, tandis que les gènes en gris clair sont impliqués dans la Mps (stabilisation de la paire) ou l’exclusion de surface. Les minuscules correspondent aux gènes trb tandis que les majuscules correspondent aux gènes tra. Type I : les gènes en beige encodent des protéines de la membrane interne. Les gènes en violet clair encodent des protéines interagissant avec la protéine de couplage DotL. Tiré de (Christie, 2016).

Les T4SS de type P sont relativement bien conservés entre eux (A. tumefaciens, R388, pKM101) : on retrouve un homologue pour chaque protéine. Ils sont également plus simples et donc mieux décrits que les types F et I, lesquels regroupent des T4SS plus disparates. Cette diversité implique des machineries de tailles différentes, composées de protéines présentant des

9 fonctions différentes. Il est donc souvent délicat d’extrapoler un mécanisme moléculaire d’un T4SS à un autre. Cependant, le mécanisme global reste similaire et plusieurs comparaisons sont régulièrement réalisées à différentes échelles. Dans les sections qui vont suivre, les types P et F sont décrits un peu plus en détail.

1.2.3. T4SS de type P

1.2.3.1 Structure

Il s’agit du système le mieux caractérisé, il est basé sur le prototype encodé par le plasmide Ti d’A. tumefaciens. Par convention, on distingue quatre composantes principales :

1) La protéine de couplage VirD4 2) Le complexe de la membrane interne avec : a) VirB4, VirB11 (ATPases) b) VirB3, VirB6 et VirB5, cette dernière faisant également partie du pilus c) VirB8 3) Le complexe de la membrane externe avec : a) VirB7 et VirB9, qui constituent ensemble le complexe cœur b) VirB10, qui s’étend jusqu’au complexe de la membrane interne 4) Le pilus conjugatif avec : a) VirB2 (piline) b) VirB1 (transglycosylase) c) VirB5 (extrémité du pilus), qu’on retrouve également associée à la membrane interne

Des résultats de microscopie électronique à transmission ont permis de visualiser pour la première fois la structure du T4SS encodé par R388, à l’exception du pilus et des ATPases VirB4 et VirB11 (Figure 1.5) (Low et al., 2014).

10

Figure 1.5. Structure d’une partie du T4SS de R388 (type P) observée en microscopie électronique à transmission Les structures du complexe cœur (« core complex ») ainsi que du complexe de la membrane interne (« IMC ») ont été combinées pour obtenir les images représentées de face (a) et de profil (c). Le panneau (b) indique également la densité électronique (bleu ; faible densité, rouge ; forte densité). Tiré de (Low et al., 2014).

Cette avancée a permis d’établir les bases pour l’architecture typique d’un T4SS de type P, représentée en Figure 1.6.

11

Figure 1.6. Représentation schématique d’un T4SS de type P Cette représentation schématique est basée sur la structure présentée en Figure 1.5. Le nombre de copies des unités est indiqué entre parenthèses. B1* désigne un fragment de la transglycosylase VirB1. Le code couleur correspond à celui décrit en Figure 1.4. Sur la droite, les protéines impliquées dans la translocation de l’ADN sont présentées dans l’ordre du chemin emprunté proposé (Cascales and Christie, 2004a). Tiré de (Christie, 2016).

1.2.3.2 Mécanisme de la conjugaison (type P)

L’origine de transfert oriT, spécifique de l’élément, est reconnue par des protéines parfois appelées Dtr (DNA transfer and replication). Une des protéines Dtr, la relaxase VirD2, clive le brin destiné au transfert au niveau du site nic dans la séquence oriT. La relaxase est dès lors liée de façon covalente à l’extrémité 5’ du brin clivé (Guzmán-Herrador and Llosa, 2019). D’autres

12 protéines Dtr ont pour fonction de guider la relaxase, mais aussi de faciliter l’accès à l’oriT, puisque l’ADN est surenroulé. L’ensemble du complexe (brin d’ADN-T lié à la relaxase, protéines auxiliaires) constitue le relaxosome. La protéine de couplage VirD4 va permettre la mise en contact du relaxosome avec la machinerie de sécrétion.

Étant donné le peu de diffusion des protéines dans le cytoplasme, il est supposé que l’initiation du transfert s’effectue proche du système de sécrétion assemblé. Il n’y a cependant aucune preuve à ce jour de signaux permettant de coordonner la localisation du système de sécrétion et celle du relaxosome. Des signaux de translocation (TS) correspondant à des motifs retrouvés sur la relaxase peuvent conférer une certaine spécificité quant à la mise en contact du substrat avec la protéine de couplage (Alperi et al., 2013; Meyer, 2015; Parker and Meyer, 2007; Redzej et al., 2013).

L’ADN est d’abord pris en charge au niveau de la membrane interne, en particulier au niveau des trois ATPases VirD4, VirB4 et VirB11 (Atmakuri et al., 2004). Par la suite, l’ADN passe au travers du complexe de la membrane interne, constitué de multiples monomères des protéines VirB3, VirB6, VirB8, mais aussi VirB5 qu’on retrouve également à l’extrémité du pilus (Low et al., 2014). La protéine VirB10 fait le lien entre les complexes de la membrane interne et externe, non seulement d’un point de vue structural mais aussi potentiellement en tant que médiateur de signaux d’un bout à l’autre du T4SS (Banta et al., 2011; Cascales and Christie, 2004b).

Le complexe de la membrane externe est constitué de nombreux monomères de VirB7, VirB9 et VirB10. Enfin, le pilus est principalement composé de VirB2, mais aussi d’un fragment protéolytique de VirB1 (VirB1*) ainsi que de VirB5 à l’extrémité. Bien que VirB1 soit une transglycosylase, son activité semble être également impliquée dans la formation du pilus, en plus de la dégradation du peptidoglycane (Zupan et al., 2007). De nombreuses études illustrant les avancées en microscopie électronique dans la dernière décennie ont permis de résoudre avec

13 précision la structure du complexe de la membrane externe (Chandran et al., 2009; Fronzes et al., 2009; Gordon et al., 2017; Low et al., 2014; Sgro et al., 2018).

Il est important de noter que, malgré la composition du T4SS en différents complexes, de nombreuses interactions sont importantes pour assembler et stabiliser la machinerie, y compris des interactions impliquant des protéines apparemment éloignées. Le dynamisme du processus est aussi illustré par les changements de conformation nécessaires de différents acteurs. Par exemple, une récente étude réalisée à l’aide du plasmide modèle R388, a montré que la relaxase (liée au brin d’ADN à transférer) devait être dépliée avant de pouvoir être transloquée au travers du T4SS (Trokter and Waksman, 2018). Un autre exemple de processus dynamique nécessitant une série de modifications est celui de la genèse du pilus.

1.2.3.3 Pilus

Dans les T4SS de type P, les pili sont petits, rigides et aucune dynamique d’élongation ou de rétraction n’a encore été montrée (Christie et al., 2005). Pour cette raison, les échanges se font rarement en milieu liquide. En revanche, l’accumulation de pili brisés forme une sorte de tamis contribuant à l’agrégation des cellules donneuses et réceptrices, ce qui facilite les échanges sur milieu solide (Christie, 2016).

Qu’il s’agisse d’un pilus de type F ou P, la piline (TraA ou VirB2) est d’abord produite sous forme de propiline dont le peptide leader est clivé après insertion dans la membrane interne (Majdalani et al., 1996; Paiva et al., 1992) (Étape 1 de la Figure 1.7). Les monomères sont cyclisés et accumulés dans la membrane interne jusqu’à réception d’un signal inconnu qui va déclencher la synthèse du pilus. VirB4 et VirB11 aident à l’acheminement des monomères dans le périplasme tandis que le peptidoglycane est digéré par la transglycosylase VirB1 (Étape 2 de la Figure 1.7). L’élongation du pilus a été montrée être effectuée par la base (Clarke et al., 2008). Les monomères seraient acheminés et accumulés au complexe de la membrane externe afin que le pilus puisse y être assemblé (Étapes 3 et 4 de la Figure 1.7).

14

Figure 1.7. Synthèse du pilus (T4SS de type P) Tiré de (Cabezón et al., 2015).

1.2.4. T4SS de type F

Les T4SS de type F diffèrent d’une part des T4SS de type P, mais également entre eux. Certaines protéines trouvent leur homologue dans le système P même si la plupart du temps il s’agit d’une

15 homologie fonctionnelle. Néanmoins, le mécanisme global de la conjugaison reste, à de nombreux égards, assez similaire à la conjugaison via un T4SS de type P. Dans la section suivante, plusieurs éléments et mécanismes spécifiques au type F sont décrits, tout en comparant les éléments communs aux deux systèmes.

1.2.4.1 Structure

Le T4SS de type F est nettement plus complexe que celui de type P. Bien que les deux systèmes partagent plusieurs homologues, certaines protéines comme l’ATPase VirB11 semblent uniques au type P, tandis que de nombreuses autres sont caractéristiques du type F (Lawley et al., 2003).

Une étude mettant en œuvre de la tomographie cryoélectronique in situ a cependant permis de révéler pour la première fois les composantes principales de ce système de sécrétion (Hu et al., 2019). De façon surprenante, les auteurs ont rapporté l’existence de plusieurs structures, révélant la dynamique d’assemblage (Figure 1.8).

16

Figure 1.8. Structure du T4SS de type F Le T4SS a pu être observé par tomographie cryoélectronique in situ sous plusieurs formes : F1 désigne le « tunnel » constitué des complexes de la membrane externe (« OMC ») et de la membrane interne (« IMC ») liés par un cylindre. F2 correspond à F1 avec le pilus ancré à la membrane externe. F3 correspond au pilus attaché à une tige qui s’étend de la membrane interne jusqu’à la membrane externe. Modifié à partir de (Hu et al., 2019).

Le T4SS une fois assemblé (F1) pourrait constituer la plateforme d’extension du pilus (F2). Après élongation, celui-ci serait capable d’ancrer une bactérie réceptrice, entrainant une transduction du signal qui engendrerait la rétraction du pilus ainsi que le recrutement du relaxosome. Une fois le transfert effectué, les cellules se sépareraient et le T4SS reviendrait à son état initial quiescent. Dépendant du mode de croissance (planctonique, agrégation, biofilm), le pilus pourrait être transloqué sur d’autres plateformes (F3, avec tige périplasmique ou membrane externe sans module périplasmique). Dans ce cas, le pilus n’aurait plus comme

17 fonction de transduire un signal d’activation du transfert mais simplement de promouvoir l’agrégation non spécifique des cellules aux alentours.

1.2.4.2 Pilus

Contrairement aux pili associés aux T4SS de type P, les pili associés aux T4SS de type F sont épais, longs, flexibles, et peuvent s’allonger et se rétracter de façon dynamique (Arutyunov and Frost, 2013) (Figure 1.9). Ils sont assemblés à partir de monomères de piline, lesquels sont acétylés (et non cyclisés comme pour le type P) et s’associent à une molécule de phospholipide de la membrane interne avant d’être extraits de celle-ci (Costa et al., 2016). Ils permettent un échange de matériel aussi efficace en milieu liquide que sur milieu solide.

Figure 1.9. Élongation et rétraction dynamique du pilus de type F observées en microscopie confocale Les cellules (Escherichia coli) et les pili sont marqués au DsRed et Alexa-488, respectivement. La durée en secondes suivant l’initiation de la synthèse est indiquée. Le pilus mesure environ 2 μm sur la deuxième image. La rétraction est visible sur la dernière image. Tiré de (Clarke et al., 2008).

Le pilus peut mesurer jusqu’à 20 μm sous certaines conditions (peu d’oxygène, faible agitation) (Arutyunov and Frost, 2013). Pendant longtemps sa fonction dans le mécanisme de transfert de l’ADN a été débattue. Aujourd’hui on sait que le lumen du pilus mesure environ 3 nm de diamètre, ce qui est assez large pour accueillir un brin d’ADN ainsi que la relaxase liée (Arutyunov and Frost, 2013). Cependant, la rétraction du pilus et le contact étroit entre les deux

18 cellules permettent d’augmenter très nettement l’efficacité du transfert. Ainsi il est peu probable que l’ADN soit transloqué au travers d’un pilus allongé. D’autre part, il est difficile d’expliquer le mécanisme d’acquisition de l’ADN : est-ce que le pilus pénètre au travers de la membrane externe ? Cela suggèrerait que l’ADN est libéré dans le périplasme, ce qui n’est a priori pas le cas. La transglycosylase a-t-elle un rôle dans la dégradation du peptidoglycane de la cellule réceptrice ? Si oui, est-elle elle-même transloquée au travers du pilus ?

De nombreux systèmes ont été étudiés à ce jour et de nombreuses hypothèses ont été avancées dans chaque cas, mais il est particulièrement délicat d’extrapoler un mécanisme d’un système à un autre. Ce qui est sûr, c’est que les études ayant été menées sur les pili et la translocation de l’ADN ont souvent débouché sur une impasse (Arutyunov and Frost, 2013). Dans les dernières années, les chercheurs ont reporté leur attention sur l’assemblage de la machinerie de conjugaison ou encore la reconnaissance de l’origine de transfert, deux domaines dans lesquels de nombreuses découvertes ont pu être effectuées.

1.2.4.3 Mpf et Mps

Le T4SS de type F se distingue du type P principalement par la complexité du mécanisme « Mpf » (pour « Mating-pair formation »). Pour permettre des échanges en milieu liquide par exemple, de nombreuses protéines viennent s’ajouter au T4SS « de base » pour permettre l’élongation et la rétraction du pilus : TraF, TraH, TraU, TraW et TrbI (Maneewannakul et al., 1992).

Ces échanges sont également permis grâce à la stabilisation de la paire (« Mps » pour « Mating- pair stabilization »). La stabilisation pour un T4SS de type F est principalement assurée par TraN et TraG. TraN est une protéine de la membrane externe, de grande taille (> 600 acides aminés), composée d’environ 20 segments transmembranaires (Klimke et al., 2005). Une portion à l’extrémité N-terminale interagit avec OmpA et le lipopolysaccharide à la surface des cellules réceptrices, tandis que l’extrémité C-terminale est importante pour la conjugaison

19 (Klimke et al., 2005). TraG est également grande (> 900 acides aminés) et principalement constituée d’un domaine périplasmique. Elle possède également un domaine au niveau de la membrane interne, lequel est impliqué dans l’exclusion d’entrée (Anthony et al., 1999). La partie N-terminale de la protéine est assez similaire à VirB6 du T4SS de type P, tandis que le domaine périplasmique C-terminal a pour homologue fonctionnel VirB8 (malgré une divergence au niveau des séquences) (Lawley et al., 2003). Enfin, il est à noter que des mutations dans le dernier tiers de la protéine affectent la Mpf mais pas la formation du pilus (Firth and Skurray, 1992).

Grâce à cette stabilisation de la paire, le transfert est généralement très rapide. Lorsqu’un plasmide conjugatif est transféré (dans le cas de F, à une vitesse de 45 kb/min), la bactérie ayant acquis l’élément devient donneuse à son tour après seulement 30 minutes (Arutyunov and Frost, 2013).

1.2.4.4 Exclusion

L’exclusion est un phénomène permettant de limiter le transfert conjugatif redondant, c’est-à- dire vers une cellule contenant déjà l’élément conjugatif. Il est nécessaire d’en distinguer deux formes; l’exclusion d’entrée et l’exclusion de surface. L’exclusion d’entrée permet d’empêcher la translocation d’un élément au travers du pore de conjugaison, tandis que l’exclusion de surface empêche la mise en contact entre les deux cellules.

Chez le plasmide F, ces mécanismes reposent sur TraS et TraT, respectivement. L’exclusion de surface est un mécanisme relativement peu efficace, malgré un haut niveau d’expression de TraT, laquelle vient « tapisser la membrane externe ». La Mpf est affectée mais le mécanisme est mal compris. Il se peut que des interactions impliquent TraN et OmpA (Riede and Eschbach, 1986).

20 L’exclusion d’entrée est nettement plus efficace (jusqu’à 3 logs de diminution de la fréquence de transfert (Audette et al., 2007)) et ne nécessite qu’un faible niveau d’expression de la protéine. Des études ont pu montrer que des régions cytoplasmiques précises de TraS et TraG étaient impliquées dans le mécanisme (Audette et al., 2007; Marrero and Waldor, 2007). Il est toutefois difficile de comprendre comment s’effectue une interaction entre deux domaines cytoplasmiques, l’un chez la donneuse et l’autre chez la réceptrice. Des hypothèses de translocation d’un fragment de TraG (ou TraS) ont notamment été avancées (Firth and Skurray, 1992).

1.2.4.5 Relaxase TraI du plasmide F

Tout comme pour le type P, la conjugaison via un T4SS de type F est initiée à l’origine de transfert par plusieurs protéines dont la relaxase, qui tient le rôle principal. Les relaxases peuvent être phylogénétiquement très éloignées et pour cette raison, sont regroupées en familles distinctes (Garcillán-Barcia et al., 2009). Les combinaisons d’un type de T4SS et d’une famille de relaxase sont nombreuses. Par exemple, le plasmide R388 encode un T4SS de type P et la relaxase TrwC qui appartient à la famille MOBF, tandis que le plasmide F, encode un T4SS de type F et la relaxase TraI, appartenant également à la famille MOBF.

La relaxase TraI du plasmide F comprend quatre domaines : un domaine trans-estérase qui va permettre le clivage au niveau de l’oriT ainsi que la liaison covalente au brin d’ADN-T (Datta et al., 2003), un domaine hélicase vestigial qui permet la liaison à l’ADN simple brin (Dostál and Schildbach, 2010), un domaine hélicase (5’-3’) actif, ainsi qu’un domaine C-terminal impliqué dans le recrutement de facteurs auxiliaires du relaxosome (Guogas et al., 2009; Matson and Ragonese, 2005). Malgré la complexité d’une telle protéine, une étude a récemment permis de résoudre la structure complète de TraI encodée par le plasmide R1, par microscopie cryoélectronique (Ilangovan et al., 2017). Dans cette même étude, les auteurs dévoilent deux conformations possibles de la relaxase : ouverte et fermée. En outre, les auteurs montrent que deux monomères de TraI (chacun dans une conformation différente) peuvent se lier à l’oriT en

21 même temps. Le clivage au site nic est possible lorsque TraI est dans sa forme ouverte et est liée à l’ADN via son domaine trans-estérase. La liaison à l’oriT de TraI dans sa forme fermée inhibe le clivage, protégeant l’ADN. Sous cette conformation, TraI peut également se lier via son domaine hélicase à l’extrémité 3’ du site de clivage, et permettre ainsi le déroulement de l’ADN. La séparation des brins est assurée par le domaine hélicase vestigial. Le fait que TraI seule puisse assurer deux activités pourrait être expliqué par un rôle à l’arrivée dans la cellule réceptrice. TraI, toujours liée à l’ADN, agirait comme une translocase, tirant le brin pour assurer un transfert complet au travers du T4SS. Chez les plasmides de type F, la liaison du relaxosome à la protéine de couplage TraD implique des interactions très spécifiques entre le facteur auxiliaire TraM et TraD (Sastre et al., 1998).

1.3 Les éléments conjugatifs

La conjugaison est médiée par des éléments génétiques mobiles variés, qu’il n’est pas toujours évident de classifier de façon claire. Il y a moins de trente ans, on pensait encore que le transfert horizontal de gènes n’était médié que par des plasmides ou des phages (Delavat et al., 2017). La découverte des ilots génomiques et de leur diversité a grandement changé notre compréhension générale de la mobilité génétique (Dobrindt et al., 2004). On définit un ilot génomique comme étant une région d’ADN d’origine étrangère, c’est-à-dire récemment acquise par transfert horizontal. Elle peut différer du reste du génome par son contenu en GC ou par sa présence dans seulement quelques souches d’une même espèce ou de deux espèces proches (Dobrindt et al., 2004). Les ilots génomiques sont généralement classés selon leur fonction : pathogénicité, résistance, catabolisme, symbiose (Juhas et al., 2009). Mais d’autres caractéristiques inhérentes à leur structure peuvent permettre de les distinguer. Les éléments intégratifs et conjugatifs, plus communément appelés ICE, sont des ilots génomiques qui comme leur nom l’indique, sont intégrés au génome mais sont également capables de transférer de façon autonome par conjugaison (Burrus et al., 2002). Pour cela, ils nécessitent d’être excisés et circularisés, après quoi leur comportement devient pratiquement identique à celui des plasmides conjugatifs.

22 Dans la suite du document, une attention particulière sera portée à trois familles d’éléments : les ICE, les plasmides conjugatifs et les ilots génomiques mobilisables (MGI). Ces derniers sont non autonomes et nécessitent la présence d’un ICE ou d’un plasmide conjugatif pour transférer.

1.3.1. Régulation

Le transfert conjugatif est un évènement particulièrement coûteux pour la cellule, en termes énergétiques. Il nécessite la production de nombreuses protéines, l’assemblage d’une nanomachine de grande complexité et une importante coordination des processus mis en jeu. Pour cette raison, l’activation du transfert est un phénomène régulé de façon stricte. Chez certaines familles d’ICE, l’activation futile peut même entrainer des dommages importants chez l’hôte, pouvant aller jusqu’à causer sa mort (Delavat et al., 2016; Menard and Grossman, 2013; Pradervand et al., 2014; Reinhard et al., 2013).

De nombreux systèmes de régulation ont été décrits à ce jour. Dans de nombreux cas, le transfert est activé par un ou plusieurs des stimuli suivants : dommages à l’ADN (réponse SOS), sécrétion de molécules par des réceptrices potentielles ou encore phase de croissance de l’hôte (Johnson and Grossman, 2015).

1.3.1.1 Régulation du transfert des ICE

L’activation du transfert des ICE appartenant à la famille SXT/R391, mais aussi de ICEBs1 et ICESt3, a été montrée être induite par la réponse SOS (Auchtung et al., 2005; Beaber et al., 2004; Bellanger et al., 2007; Burrus and Waldor, 2003). Cette activation correspond en fait à une dérépression, résultant du clivage du répresseur (Figure 1.10A et C). Dans le cas des ICE SXT/R391 par exemple, le répresseur SetR est inactivé par autoprotéolyse (Beaber et al., 2004). Par la suite, le régulateur CroS vient se lier au promoteur contrôlant l’expression de setR pour empêcher la reprise prématurée de l’expression du répresseur (Poulin-Laprade and Burrus, 2015). Chez ICEBs1, c’est la protéase ImmA encodée par l’élément qui assure le clivage du

23 répresseur ImmR (Bose et al., 2008). La réponse SOS ayant été rapportée n’être activée que dans une petite proportion de la population cellulaire, il est probable que seul un sous-groupe de cellules portant l’élément constitue la population de donneuses actives (Kamenšek et al., 2010; Pennington and Rosenberg, 2007).

Chez d’autres familles d’éléments, la régulation dépend d’autres signaux. Les ICE Tn916 et CTnDOT répondent à la présence de tétracycline dans le milieu (Figure 1.10B et D) (Celli and Trieu‐Cuot, 1998; Waters and Salyers, 2013). Le transfert de l’ICE de symbiose ICEMlSymR7A de Mesorhizobium loti est activé par le quorum-sensing (Figure 1.10F) (Ramsay et al., 2006, 2009, 2013). Enfin, la régulation de l’élément ICEclc a récemment été montrée reposer sur un système complexe permettant ultimement d’activer la compétence chez une sous-population dédiée de cellules (Carraro et al., 2020). Dans ces deux cas, l’élément répond à l’atteinte de la phase stationnaire de croissance de l’hôte. Ultimement, la dérépression d’un ICE mène à son excision, première étape dans son processus de transfert.

Ces différents modèles semblent indiquer que le transfert d’un ICE est majoritairement activé lorsque la cellule ne constitue plus un hôte viable. Dès lors, le transfert horizontal semble être une option de sauvegarde, en comparaison avec une transmission héréditaire dans une population apparemment « condamnée » (Johnson and Grossman, 2015). La nécessité de stabilité et la capacité de sacrifice de ces éléments sont illustrées par le cycle de vie extraordinaire de l’ICE tripartite ICEMcSym1271 (Haskett et al., 2016). Constitué de trois fragments intégrés dans trois loci distincts de M. ciceri, son excision et sa circularisation nécessitent une coordination de multiples processus de recombinaison, incluant l’inversion de régions chromosomiques complètes (Haskett et al., 2016). Les ICE tripartites sont aujourd’hui reconnus comme une famille à part entière d’éléments ayant persisté dans la nature et dont la structure (et non le cargo) confèrerait un avantage sélectif à son hôte sur le long terme (Haskett et al., 2017).

24

Figure 1.10. Schémas de régulation du transfert de plusieurs familles d’ICE

25 Figure 1.10. Schémas de régulation du transfert de plusieurs familles d’ICE (suite) Pour chaque modèle présenté, les panneaux supérieurs au fond gris clair illustrent l’état quiescent de l’ICE, tandis que les panneaux inférieurs au fond gris foncé décrivent les conditions nécessaires à leur activation. Les promoteurs sont représentés par des flèches noires à angle droit. Une activation ou une répression est indiquée par une flèche noire ou une ligne bloquée, respectivement. Les flèches grises ondulées représentent des interactions protéine-protéine. Les gènes sont représentés par des flèches vides, tandis que les ARNm sont représentés par des vagues. Les protéines clés de chaque système sont représentées par des cercles ou des hexagones de couleur. Leur éventuelle dégradation est schématisée par un cercle brisé. La phosphorylation est indiquée par un « P » encerclé. Les phases de croissance exponentielle et stationnaire sont abrégées « EXPO » et « STAT », respectivement. « AI » correspond à l’homosérine lactone auto-inductrice.

1.3.1.2 Cycle de vie des ICE

L’excision d’un ICE est médiée par l’intégrase Int de concert avec un facteur d’orientation de la recombinaison (RDF) encodé par le gène xis (par abus de langage on parle d’excisionase). Elle se matérialise par une recombinaison des bornes attL (attachement gauche) et attR (attachement droit) délimitant l’ICE, pour former le site attP (attachement du phage, ou attI; attachement de l’ICE) et ainsi circulariser l’élément. Une fois sous forme circulaire, l’élément doit assurer son maintien de façon épisomique, c’est-à-dire indépendamment du chromosome. La réplication des ICE s’effectue en cercle roulant, la synthèse du brin retardé étant initiée à partir d’une origine de réplication simple brin appelée sso. Cette région, déjà décrite pour de nombreux plasmides et phages se répliquant en cercle roulant et caractérisée chez ICEBs1 de B. subtilis, est impliquée dans la stabilité de l’ICE autant chez la cellule donneuse que dans la cellule réceptrice (Wright et al., 2015). Chez les ICE de la famille SXT/R391, il a été montré que l’élément circulaire répliqué est réparti dans les cellules filles lors de la division cellulaire

26 grâce au système de partition active qu’il encode (Carraro et al., 2015a). Toutefois, ce mécanisme est loin d’être conservé chez tous les ICE.

Comme c’est le cas pour un plasmide conjugatif, l’ICE encode toutes les fonctions nécessaires à son transfert. Une fois le relaxosome pris en charge par la protéine de couplage, l’élément est transloqué en simple brin (ou double brin chez les Actinobactéries) au travers d’un T4SS (ou multimère d’une unique translocase chez les Actinobactéries). Une fois dans la bactérie réceptrice, l’intégrase exprimée de façon constitutive va permettre la réintégration de l’ICE au niveau du site attB (attachement bactérien) dans le chromosome, assurant ainsi son maintien et sa transmission verticale. Les différentes étapes constituant le cycle de vie d’un ICE sont décrites en Figure 1.11.

Figure 1.11. Schéma explicatif du cycle de vie d’un ICE L’ICE est excisé (1) ou intégré (2) au moyen de l’intégrase Int qu’il encode et d’un facteur d’orientation de la recombinaison (RDF). Le transfert et la réplication en cercle roulant sont initiés par la relaxase à l’oriT, qui sert également d’origine double brin (dso) (3). Après clivage d’un brin de l’ADN, le second brin est synthétisé (4).

27 Figure 1.11. Schéma explicatif du cycle de vie d’un ICE (suite) Le brin clivé et lié à la relaxase peut être amené au système de sécrétion pour être transféré ou bien être circularisé (5) en vue d’être répliqué dans la cellule donneuse (6). L’ICE répliqué est soit maintenu grâce à différents systèmes de stabilisation (7), soit réintégré au chromosome (8). Lorsque l’ICE est transféré (9), l’ADN transloque au travers du système de sécrétion. Une fois dans la cellule réceptrice, l’ADN est circularisé (10) et la synthèse du second brin est initiée à partir du sso, comme décrit ci-dessus (6). À ce stade, il est possible que l’ICE soit répliqué à nouveau. Lorsque le transfert est à nouveau réprimé, l’ICE est intégré au chromosome. Adapté de (Burrus, 2017).

1.3.1.3 Régulation du transfert des plasmides conjugatifs

Le transfert des plasmides conjugatifs est également contrôlé de façon précise. Les conditions menant à son activation diffèrent toutefois de celles des ICE en un point majeur : les plasmides conjugatifs doivent en tout temps assurer leur propre maintien indépendamment du chromosome de leur hôte.

Plusieurs similarités ont tout de même pu être observées. Le transfert intracellulaire du plasmide Ti ou la mobilisation du plasmide mobilisable RSF1010 ont été montrés être activés par un système de quorum-sensing (White and Winans, 2007). De façon similaire, le transfert du plasmide pCF10 d’Enterococcus faecalis est activé suite à la détection d’une phéromone produite par la cellule réceptrice (Dunny, 2007). Le transfert du plasmide linéaire SLP2 de Streptomyces lividans dépend du mode de croissance de son hôte, puisqu’il est disséminé dans l’hyphe préalablement à la sporulation (Hsu and Chen, 2010). De manière générale, le transfert de nombreux plasmides conjugatifs est activé lorsque des changements environnementaux et/ou physiologiques sont détectés (Frost and Koraimann, 2010). De nombreuses études ont permis de montrer que le transfert conjugatif était affecté par des paramètres comme la température, le pH ou la concentration d’oxygène (Frost, 2009). Une étude a même testé la capacité de transfert

28 d’un plasmide conjugatif entre deux souches de B. thuringiensis, ainsi que sa capacité de mobilisation d’un plasmide mobilisable, dans des conditions simulant la microgravité ressentie dans l’espace (Beuls et al., 2009).

1.3.2. Éléments modèles au laboratoire

1.3.2.1 Les ICE SXT/R391

Les ICE de la famille SXT/R391 ont fait l’objet de nombreuses études épidémiologiques ces dernières années. Elle regroupe des ICE variés, lesquels sont régulièrement isolés à partir de souches cliniques et environnementales de Gammaprotéobactéries. Ils ont également largement contribué à la dissémination de multiples résistances aux antibiotiques depuis la fin des années 1980 dans le cadre de la 7e pandémie de choléra (Spagnoletti et al., 2014; Weill et al., 2017). L’élément prototype SXT a été isolé pour la première fois en 1992 à partir d’une souche clinique de cholerae O139 (Waldor et al., 1996), à laquelle il conférait des résistances au sulfaméthoxazole et au triméthoprime (d’où son nom), mais aussi à la streptomycine et au chloramphénicol (Hochhut et al., 2001). Encodant un T4SS de type F, les ICE SXT/R391 peuvent transférer de façon autonome mais sont également capables de mobiliser en trans toute une famille d’ilots génomiques non autonomes incluant MGIVflInd1 de V. fluvialis (Daccord et al., 2013). L’initiation du transfert est assurée par la relaxase TraI appartenant à la famille

MOBH1, mais aussi la protéine auxiliaire MobI, qui aiderait à la reconnaissance de l’oriT et/ou au recrutement de la relaxase, bien que sa fonction exacte reste à caractériser.

Leur noyau de gènes conservés (Figure 1.12) comprend entre autres :

- Le module d’intégration-excision - Les deux gènes encodant le complexe activateur transcriptionnel SetCD - Le gène encodant le répresseur SetR - Le gène encodant le facteur auxiliaire MobI, spécifique de l’origine de transfert

29 - L’opéron srpRM, encodant un système de partition active - Les gènes ssb, bet et exo, qui sont impliqués dans la réparation des cassures double brin de l’ADN suite au transfert conjugatif (Roy et al., 2020) - Les gènes encodant le T4SS

Figure 1.12. Représentation schématique du noyau de gènes conservés chez les ICE SXT/R391 Les gènes sont représentés par des flèches, à chaque couleur correspond une fonction, comme indiqué dans l’encadré. Certains variants présentent une inversion au niveau des deux sites « INV » encadrés et fléchés. Tiré de (Bioteau et al., 2018).

De façon intéressante, certains ICE de la même famille encodent une intégrase alternative, catalysant une intégration dans des gènes encodant un ARN de transfert pour la sérine (ARNtSer) et non dans le gène prfC comme c’est le cas pour le prototype SXT (Bioteau et al., 2018; Luo et al., 2016; Taviani et al., 2012). Dans une étude jointe en Annexe 1, nous avons montré qu’il convenait d’intégrer ces éléments atypiques à la famille des ICE SXT/R391.

Les gènes encodant le T4SS, quant à eux, restent particulièrement conservés. De façon intéressante, l’organisation des opérons peut être retrouvée chez une autre famille d’éléments : les plasmides conjugatifs des groupes d’incompatibilité A et C (A/C), lesquels encodent également un T4SS de type F. Bien que les séquences des gènes tra des deux familles d’éléments ne présentent que très peu d’identité, leur ordre synténique semble indiquer une divergence à partir d’un ancêtre commun (Poulin-Laprade et al., 2015a).

30 Une telle organisation nécessite un système d’activation permettant l’expression simultanée de tous les gènes impliqués dans le transfert. Chez les ICE SXT/R391, cette activation est permise grâce au complexe SetCD. En temps normal, l’expression de setDC est inhibée par le répresseur transcriptionnel SetR, apparenté au régulateur cI du phage λ. Lorsque certains agents (rayons UV, mitomycine C, ciprofloxacine) entrainent des dommages à l’ADN et une activation de la réponse SOS, SetR est inactivé par autoprotéolyse, ce qui permet la levée de l’inhibition (Poulin- Laprade et al., 2015b). SetCD reconnait par la suite un site consensus, situé en amont du site d’initiation de la transcription de 11 opérons. Ces mêmes sites de liaison SetCD sont également retrouvés chez les ilots génomiques mobilisés par les ICE SXT/R391. Ainsi les gènes impliqués dans le transfert de l’ICE vont être activés par SetCD, tandis que l’excision des ilots en question va également être activée par SetCD. En parallèle, un second répresseur CroS vient réprimer l’expression de SetR jusqu’à complétion du transfert. La reprise de l’inhibition par SetR replace l’ICE dans son état quiescent (Poulin-Laprade and Burrus, 2015).

1.3.2.2 Les plasmides A/C

Appelés les plasmides « IncA/C » jusqu’en 2015 (désormais « A/C »), ils ont été parmi les premiers associés à des résistances aux antibiotiques chez des Gammaprotéobactéries (Harmer and Hall, 2015). Les plasmides IncA et IncC ont initialement été regroupés sous le terme de « complexe A/C » en raison de la compatibilité des réplicons (Hedges, 1974), un phénomène longtemps admis mais confirmé de façon expérimentale seulement récemment (Ambrose et al., 2018). La nomenclature « A/C » a été déformée pour devenir « IncA/C » pour des raisons inconnues et s’est imposée dans la littérature (Harmer and Hall, 2015). Aujourd’hui on distingue les plasmides IncA des plasmides IncC sur la base du mécanisme d’exclusion d’entrée, récemment caractérisé par notre équipe (Humbert et al., 2019). Les plasmides IncA ne regroupent que très peu de représentants (RA1, pIncAC-KP4898), semblent être confinés à des niches écologiques restreintes et sont rarement isolés à partir de souches cliniques (Arcari et al., 2020). Il a cependant déjà été rapporté qu’un plasmide IncA était responsable de la dissémination du gène blaVIM-1, qui encode une carbapénémase, chez plusieurs entérobactéries

31 dans un hôpital en Italie (Arcari et al., 2020). Les plasmides IncC représentent une famille nettement plus grande. Ils sont largement disséminés chez des entérobactéries pathogènes et ont été montrés avec les ICE SXT/R391 avoir été les acteurs principaux de la dissémination de multiples résistances chez les souches africaines de V. cholerae O1 de la 7e pandémie de choléra (Weill et al., 2017). En plus de conférer couramment des résistances aux aminoglycosides, au chloramphénicol ou aux sulfonamides, de plus en plus d’études épidémiologiques montrent que les plasmides IncC peuvent porter le gène blaNDM-1, qui encode une carbapénémase (Hancock et al., 2017; Martino et al., 2019; Wailan et al., 2015; Wu et al., 2019). Ce type de gène conférant des résistances à la majorité des β-lactamines, la dissémination des plasmides IncC limite de plus en plus les possibilités de traitement.

La famille des plasmides IncC est elle-même divisée en deux sous-groupes : les plasmides IncC de type 1 et 2. Ces deux types ne diffèrent que d’un pourcent au niveau de la séquence nucléotidique de leur noyau de gènes conservés, mais cette différence révèle une évolution distincte datant d’au moins 5000 ans (bien avant l’accumulation de gènes de résistance) (Harmer and Hall, 2015). Le noyau de gènes conservés comprend tous les gènes nécessaires à la réplication, à la stabilité, au transfert conjugatif et à la régulation (Figure 1.13) (Harmer and Hall, 2015; Poulin-Laprade et al., 2015a).

Figure 1.13. Représentation schématique du noyau de gènes conservés chez les plasmides A/C Les gènes sont représentés par des flèches, à chaque couleur correspond une fonction. Rouge : répresseurs transcriptionnels, vert : activateurs transcriptionnels, bleu : transfert conjugatif, violet : recombinaison homologue, orange : stabilité. Adapté de (Carraro et al., 2017a).

32 La stabilité des plasmides A/C est assurée en premier lieu par RepA, qui assure l’initiation de la réplication à l’origine de réplication (oriV), mais aussi par le système de partition active ParAB (Hancock et al., 2017). D’autres protéines ont été montrées être impliquées dans la stabilité des plasmides A/C, notamment un système toxine-antitoxine (Carraro et al., 2014a), dont l’opéron semble être fortement exprimé (Lang et al., 2012b). Un second système de partition putatif a également été rapporté, mais aucune étude n’a encore permis de confirmer sa fonction de façon expérimentale (Harmer and Hall, 2015). Il semblerait cependant que dans des conditions « normales » de maintenance du plasmide, le système ne soit pas nécessaire (Hancock et al., 2017). Dans les dernières années, les progrès des techniques de séquençage à haut débit ont permis de caractériser un peu plus en détail les mécanismes impliqués dans la dissémination de ces éléments.

Récemment, une étude publiée par le laboratoire a permis de caractériser la fonction du système de recombinaison homologue encodé par les gènes conservés ssb, bet et exo (Roy et al., 2020). Homologue à celui des ICE SXT/R391, ce complexe assure la réparation des cassures double brin de l’ADN occasionnées par les systèmes de défense endogènes CRISPR-Cas en réponse à l’entrée d’ADN exogène. Un tel mécanisme permet non seulement de déjouer le système d’immunité adaptative de la bactérie, mais promeut également la plasticité des plasmides IncC (et des ICE SXT/R391), comme l’illustrent les multiples patrons de cicatrisation engendrés.

Tout comme pour les ICE SXT/R391, le facteur auxiliaire MobIC a été montré indispensable au transfert conjugatif des plasmides IncC et serait impliqué, par analogie avec MobIS, dans la reconnaissance de l’oriT et/ou le recrutement de la relaxase TraI (Carraro et al., 2014b; Ceccarelli et al., 2008). En se basant également sur les systèmes mieux décrits jusqu’à présent, TraI viendrait cliver le brin destiné au transfert au niveau du site nic, puis le relaxosome

(composé entre autres du brin d’ADN à transférer, de TraI, MobIC, et éventuellement TraJ) serait mis en contact avec la machinerie de transfert grâce à la protéine de couplage TraD.

33 En 2014, une étude pionnière en matière de séquençage à haut débit d’un génome bactérien, a permis d’identifier le complexe majeur d’activation du transfert conjugatif des plasmides A/C : AcaCD (Carraro et al., 2014a). Homologue au complexe SetCD encodé par les ICE SXT/R391, cet hétéromultimère reconnait un site consensus situé environ 50 nucléotides en amont du site d’initiation de la transcription de 18 opérons. Environ un tiers des gènes activés par AcaCD sont impliqués dans le transfert conjugatif, toutefois beaucoup sont également de fonction encore inconnue.

L’expression d’acaDC est dirigée à partir du promoteur Pacr1, lequel est régulé par plusieurs acteurs (Figure 1.14) :

- Acr1, un répresseur encodé à partir du même transcrit qu’AcaCD (Carraro et al., 2014a) - Acr2, un répresseur encodé à partir d’un opéron convergent localisé directement en aval (Carraro et al., 2014a; Lang and Johnson, 2016)

- AcaB, un activateur nouvellement identifié dont le gène est localisé en aval de traNC (Hancock et al., 2020)

De façon intéressante, l’expression d’acaB est dépendante d’AcaCD, ce qui entraine la formation d’une boucle de rétroaction positive (Carraro et al., 2014a; Hancock et al., 2020).

Figure 1.14. Régulation du transfert des plasmides A/C

34 Figure 1.14. Régulation du transfert des plasmides A/C (suite) Le schéma de régulation est résumé au moyen de flèches vertes et rouges, symbolisant l’activation et la répression transcriptionnelle, respectivement. Adapté de (Carraro et al., 2015b).

Contrairement aux ICE SXT/R391, les signaux nécessaires à la levée de l’inhibition chez les plasmides A/C sont encore inconnus à ce jour. En revanche, tout comme les ICE SXT/R391, les plasmides A/C sont également capables de mobiliser des ilots génomiques non autonomes appartenant à des familles distinctes tels que MGIVmi1, MGIVchHai6 ou encore SGI1 (Carraro et al., 2014a, 2015b, 2016; Douard et al., 2010). Ces derniers n’encodent pas de T4SS mais sont délimités par des sites attL et attR, encodent leur propre module d’intégration-excision et possèdent une oriT, le plus souvent associée à un gène mobI. Dans la majorité des cas, la reconnaissance de l’oriT est assurée par la protéine MobI encodée soit par l’ilot soit par l’ICE ou le plasmide. Pour être mobilisé, l’ilot doit être excisé et circularisé. Pour cette raison, les ilots mobilisés par un plasmide IncC présentent le plus souvent un site AcaCD en amont du gène xis, ce qui va permettre une activation par AcaCD de l’excision de l’ilot. D’autres sites peuvent également être présents en amont d’opérons stratégiques pour sa mobilité. Une fois excisé et circularisé, l’ilot emprunte le T4SS de l’ICE/du plasmide pour transférer.

1.3.2.3 MGIVchHai6

MGIVchHai6 est un ilot génomique spécifiquement mobilisable par les plasmides A/C, retrouvé intégré à l’extrémité 3’ du gène trmE dans le chromosome I de la souche V. cholerae HC-36A1 (Carraro et al., 2016). Cette souche clinique multirésistante (en raison de l’acquisition de l’ilot), isolée en 2010 à Haïti, présente la particularité d’être non-O1/non-O139, c’est-à-dire reconnue comme non épidémique. Toutefois la simple mobilisation de l’ilot par un plasmide A/C peut permettre son acquisition par des souches d’entérobactéries ou des souches de V. cholerae O1 et/ou O139 donc épidémiques. D’une taille d’environ 47 kb, MGIVchHai6 contient un intégron

35 In36A1 conférant de multiples résistances aux antibiotiques, ainsi qu’un transposon Tn6310 conférant une résistance au mercure (Figure 1.15).

Figure 1.15. Représentation schématique de l’ilot génomique MGIVchHai6 Les gènes sont représentés par des flèches, à chaque couleur correspond une fonction. Noir : recombinaison spécifique de site, brun : DUF927, jaune :

régulateurs transcriptionnels, orange : MobIM, violet : virulence RhuM, rose : restriction-modification de type I, gris : déterminants de résistances, blanc : inconnue. L’intégron In36A1 confère des résistances à l’ampicilline, au chloramphénicol, à la streptomycine, à la spectinomycine, au sulfaméthoxazole, au triméthoprime et à la tétracycline (ACSSuTmT). Tiré de (Carraro et al., 2016).

La séquence de l’ilot contient deux sites AcaCD, l’un en amont de mobIM, l’autre en amont d’un opéron de trois gènes : vchc36a1_0085 encodant un régulateur transcriptionnel putatif, vchc36a1_0086 encodant une protéine possédant un domaine de fonction inconnue (DUF927) et xis. Une étude récemment publiée par le laboratoire a confirmé que ces promoteurs sont bien activés par AcaCD et que MobIM est indispensable à la mobilisation de MGIVchHai6 (Rivard et al., 2020). Il semblerait toutefois que l’ilot utilise également MobIC pour un transfert optimal, même si le facteur encodé par le plasmide IncC n’est pas nécessaire pour mobiliser l’ilot.

1.3.2.4 SGI1

SGI1 (Salmonella Genomic Island 1) est un ilot génomique spécifiquement mobilisable par les plasmides A/C (Douard et al., 2010). Il a été décrit pour la première fois en 2000 chez la souche épidémique Salmonella enterica serovar Typhimurium DT104 (Boyd et al., 2001, 2000). Il a

36 depuis été rapporté chez de très nombreuses autres espèces de Gammaprotéobactéries : Proteus (PGI1), Acinetobacter (AGI1), Morganella, Providencia, Enterobacter, Escherichia, Vibrio (GI-15) et Klebsiella (Cummins et al., 2020; Hamidian et al., 2015; Schultz et al., 2017; Siebor and Neuwirth, 2013; Siebor et al., 2016). Jusqu’à présent, SGI1 et ses variants ont toujours été retrouvés intégrés à l’extrémité 3’ du gène trmE.

SGI1 a une taille d’environ 43 kb et contient un intégron complexe de classe 1 conférant des résistances à l’ampicilline, au chloramphénicol, à la streptomycine, à la spectinomycine, au sulfaméthoxazole ainsi qu’à la tétracycline (Figure 1.16). Il partage avec ses variants un noyau de gènes conservés de 26 kb composé entre autres des gènes :

- int et xis, respectivement impliqués dans l’intégration et l’excision de l’ilot - rep, impliqué dans la réplication de l’ilot (Huguet et al., 2020) - sgaDC, codant pour le complexe activateur transcriptionnel SgaCD homologue à AcaCD (Poulin-Laprade et al., 2015a) - sgiAT, codant pour un système toxine-antitoxine (Huguet, 2016)

- traNS, traGS et traHS, codant pour des protéines de transfert (Carraro et al., 2017a) - mpsAB, impliqués dans l’initiation du transfert à l’oriT (Kiss et al., 2019)

Figure 1.16. Représentation schématique de l’ilot génomique SGI1 Les gènes sont représentés par des flèches, à chaque couleur correspond une fonction. Noir : recombinaison spécifique de site, orange : réplication, jaune : régulateurs transcriptionnels, bleu : transfert conjugatif, rose : toxine-antitoxine, gris : déterminants de résistances, brun : transposition, blanc : inconnue.

37 Figure 1.16. Représentation schématique de l’ilot génomique SGI1 (suite) L’intégron In104, apparenté à l’intégron In36A1 de MGIVchHai6, confère des résistances à l’ampicilline, au chloramphénicol, à la streptomycine, à la spectinomycine, au sulfaméthoxazole et à la tétracycline (ACSSuT). Adapté de (Carraro et al., 2017a).

SGI1 est un élément caractérisé par une grande stabilité, comme l’illustre sa rétention dans une population de cellules sans pression de sélection après 350 générations (Kiss et al., 2012). Bien que l’ilot encode un système toxine-antitoxine, celui-ci n’a été démontré essentiel à la stabilité de SGI1 que dans un contexte où un plasmide A/C serait présent dans la même cellule (Huguet, 2016). D’autre part, la mobilisation de SGI1 par les plasmides A/C a été montrée dépendre de l’excision de l’ilot, médiée par le gène xis, dont l’expression est elle-même dépendante de la présence d’un plasmide A/C ou d’une activation par AcaCD (Carraro et al., 2014a; Kiss et al., 2015). En comparaison, le gène int encodant l’intégrase de SGI1, est exprimé à partir d’un promoteur constitutif fort, ce qui permet à la fois de compenser pour une éventuelle excision spontanée, mais permet aussi à l’ilot de se réintégrer immédiatement après avoir transféré dans une nouvelle cellule.

Des motifs de liaison AcaCD ont été prédits en amont de quatre autres opérons de SGI1 : S004- rep, traNS, traHGS et le gène de fonction inconnue S018 (Carraro et al., 2014a). Des essais rapporteurs basés sur une fusion transcriptionnelle des promoteurs au gène lacZ, ont permis de confirmer qu’ils sont effectivement activables par AcaCD, mais aussi à moindre niveau par le complexe homologue SgaCD encodé par SGI1 (Kiss et al., 2015; Murányi et al., 2016).

L’oriT de SGI1 est localisée en aval du gène S023 et mesure 124 pb (Kiss et al., 2019). Elle s’accompagne de deux gènes importants pour la mobilisation de l’ilot : mpsB et mpsA, ce dernier encodant une recombinase site-spécifique à tyrosine qui jouerait le rôle d’une relaxase atypique.

SGI1 ne porte aucun gène homologue à mobI. Il semblerait également que MobIC encodé par

38 les plasmides A/C ne soit pas nécessaire pour reconnaitre l’oriT (Hegyi et al., 2017). L’initiation de la mobilisation de SGI1 reste cependant un phénomène encore très mal connu.

À ce jour, SGI1 est avec GIsul2 le seul ilot génomique mobilisable encodant des protéines du pore de conjugaison (Harmer et al., 2017). Les protéines TraNS, TraHS et TraGS encodées par SGI1 sont homologues aux sous-unités du T4SS encodé par les plasmides IncC. Leur caractérisation a fait l’objet d’un article qui constitue le CHAPITRE 2 de cette thèse.

La relation entre SGI1 et les plasmides IncC est également marquée par une déstabilisation de ces derniers par l’ilot (Harmer et al., 2016), un phénomène récemment attribué à l’état réplicatif de SGI1 (Huguet et al., 2020). La description de la famille de variants SGI1-HKL, incapables de déstabiliser les plasmides IncC, a soulevé la possibilité que des gènes encodés par SGI1 soient impliqués dans le mécanisme. En effet, ces variants présentant une délétion de 2.8 kb allant de traNS à S009, ils n’encodent donc pas le complexe SgaCD. Son rôle dans le cycle de vie de SGI1, incluant le phénomène de déstabilisation des plasmides IncC, a fait l’objet d’un second article qui constitue le CHAPITRE 3 de cette thèse.

Enfin, la relation entre SGI1 et les plasmides IncC est également marquée par un co-transfert assez rare, un phénomène jusqu’à présent jamais expliqué (Carraro et al., 2014a; Harmer et al., 2016). Des expériences préliminaires de mutagenèse par transposition aléatoire couplée à du séquençage à haut débit (TnSeq) a cependant permis d’identifier le gène S010 porté par SGI1, comme un déterminant affectant négativement le co-transfert (communication personnelle de Kévin Huguet).

1.3.3. Bilan des interactions entre différentes familles d’éléments

La caractérisation de la mobilisation de divers ilots génomiques par les plasmides A/C ou les ICE SXT/R391 a révélé une diversité de mécanismes, qu’il est important de bien distinguer. Dans la Figure 1.17, des modèles ont été proposés pour :

39 - La mobilisation de SGI1 par les plasmides A/C (panneau A) - La mobilisation de MGIVchHai6 par les plasmides A/C (panneau B) - La mobilisation de MGIVflInd1 par les ICE SXT/R391 (panneau C)

Figure 1.17. Modèles proposés pour 3 mécanismes de mobilisation d’ilots génomiques Tiré de (Carraro et al., 2017b).

40 Dans le cas de SGI1, AcaCD active l’expression de xis et de S004-rep, déclenchant respectivement son excision et sa réplication subséquente. L’initiation de la mobilisation est un phénomène nécessitant de plus amples recherches.

Dans le cas de MGIVchHai6, AcaCD active l’expression de xis et des deux gènes mobI. L’oriT de l’ilot est spécifiquement reconnue par MobIM, mais la présence de MobIC permet une mobilisation optimale (Rivard et al., 2020).

Dans le cas de MGIVflInd1, c’est SetCD, le complexe activateur des ICE SXT/R391, qui active les gènes int et rdfM pour permettre l’excision. L’oriT de ce MGI étant pratiquement identique

à celle des ICE SXT/R391, les protéines MobIS et TraI encodées par l’ICE sont nécessaires à la reconnaissance et au clivage de l’oriT.

Le Tableau 1.1 ci-dessous résume les déterminants établis jusqu’à présent pour la mobilisation de différents MGI.

Tableau 1.1. Déterminants pour la mobilisation de quelques MGI

SGI1 MGIVchHai6 MGIVmi1 MGIVflInd1 Site trmE trmE yicC yicC d’intégration Mobilisé par Plasmides A/C Plasmides A/C Plasmides A/C ICE SXT/R391 Activateur AcaCD AcaCD AcaCD SetCD xis, S004-rep, xis, mobI, Gènes activés traN , traH et xis, mobI, 85, 86 int, rdfM S S cds400, cds420 S018. Spécifique, en Spécifique, en Spécifique, en Similaire à oriT amont de mpsA amont de mobIM amont de mobI l’oriT de SXT

41 1.4 Hypothèses du projet de doctorat

De très nombreuses études épidémiologiques sont publiées chaque année, alertant sur l’impact à grande échelle des plasmides conjugatifs IncC et de l’ilot génomique SGI1. Cependant, à l’image de tant d’autres familles d’éléments génétiques mobiles, relativement peu d’études s’intéressent aux mécanismes moléculaires au cœur de leur transfert. C’est la raison pour laquelle le laboratoire s’attèle à identifier le plus précisément possible les déterminants nécessaires à la mobilisation de ces vecteurs de résistances.

La caractérisation d’AcaCD et la description de son rôle clé dans le transfert des plasmides IncC et dans la mobilisation de SGI1 ont constitué une avancée significative dans la compréhension de ces familles d’éléments. Mais elles ont également soulevé de nombreuses questions, en particulier quant aux étapes précises du cycle de vie de l’ilot génomique.

Nos hypothèses pour ce projet de doctorat sont les suivantes :

1) Les gènes traNS, traHS et traGS portés par SGI1 encodent des protéines fonctionnelles, lesquelles sont incorporées au pore de conjugaison du plasmide IncC et permettent d’augmenter spécifiquement la fréquence de transfert de l’ilot.

2) Le gène traGS est impliqué dans le mécanisme d’exclusion d’entrée exercé par les plasmides IncC. 3) Le complexe activateur transcriptionnel SgaCD encodé par SGI1 est capable de reconnaitre le même motif qu’AcaCD et d’activer tous les promoteurs dépendants d’AcaCD. 4) SgaCD est important pour l’excision et la réplication de SGI1 et est à l’origine du mécanisme de déstabilisation des plasmides IncC. 5) AcaB est essentiel à la mobilisation de SGI1. 6) Le gène de fonction inconnue S010 porté par SGI1 encode une protéine fonctionnelle dont le rôle est de réprimer le transfert conjugatif des plasmides IncC.

42 7) SGI1 peut également être retrouvé dans un ou plusieurs sites d’intégration différents du site canonique trmE.

1.5 Objectifs du projet de doctorat

L’objectif global du projet de doctorat est d’identifier de nouveaux déterminants de la mobilisation de SGI1. Afin de valider les hypothèses émises précédemment, nous avons fixé les objectifs suivants :

1) Caractériser les gènes traNS, traHS et traGS portés par SGI1 a. Évaluer les activités de deux promoteurs correspondants avec ou sans induction par AcaCD; b. Évaluer l’impact de leur délétion sur la fréquence de transfert de SGI1 et du plasmide IncC;

c. Évaluer l’impact de la délétion des gènes orthologues traNC, traHC et traGC encodés par les plasmides IncC sur la fréquence de transfert des deux éléments; d. Évaluer l’impact de délétions combinées sur le transfert des deux éléments;

e. Déterminer le rôle de TraGS dans le mécanisme d’exclusion d’entrée exercé par les plasmides IncC.

2) Caractériser la fonction du complexe activateur transcriptionnel SgaCD encodé par SGI1 a. Comparer l’impact des délétions uniques et combinées de sgaDC et acaDC sur la fréquence de transfert et sur le nombre de copies de SGI1 et du plasmide IncC, ainsi que sur le taux d’excision de SGI1; b. Établir le régulon de SgaCD i. Vérifier si SgaCD reconnait le même motif qu’AcaCD et/ou active les mêmes promoteurs qu’AcaCD; ii. Comparer l’activité de tous les promoteurs avec ou sans induction par AcaCD ou SgaCD;

43 iii. Évaluer l’impact d’une induction par SgaCD sur le profil transcriptomique d’une souche contenant SGI1 et un plasmide IncC; c. Établir le rôle de SgaCD dans le maintien de SGI1 et d’un plasmide IncC.

3) Caractériser le rôle d’AcaB dans la mobilisation de SGI1 a. Évaluer l’impact de la délétion d’acaB sur la fréquence de transfert, le taux d’excision et le nombre de copies de SGI1; b. Vérifier si AcaB est important pour la stabilité d’un plasmide IncC.

4) Caractériser la fonction du gène S010 encodé par SGI1 a. Évaluer l’impact de la délétion de S010 sur la fréquence de transfert de SGI1 et d’un plasmide IncC; b. Identifier les interactants du produit de traduction de S010 et établir son mode d’action.

5) Montrer que SGI1 peut être retrouvé dans plusieurs sites d’intégration a. Identifier dans les bases de données en ligne les ilots qui encodent les mêmes protéines clés que SGI1 et identifier leur site d’intégration; b. Vérifier que les ilots identifiés sont mobilisables par un plasmide IncC.

44 CHAPITRE 2 SALMONELLA GENOMIC ISLAND 1 (SGI1) RESHAPES THE MATING APPARATUS OF INCC CONJUGATIVE PLASMIDS TO PROMOTE SELF- PROPAGATION

2.1 Introduction

2.1.1. Pertinence

L’ilot génomique de résistance SGI1 est spécifiquement mobilisé par les plasmides conjugatifs IncC. Bien que l’ilot n’encode pas son propre T4SS, il porte néanmoins trois gènes encodant des sous-unités du pore de conjugaison. À l’aide de mutants de délétion pour les gènes traN, traG et traH chez un plasmide IncC (traC) et SGI1 (traS), nous avons montré par des essais de conjugaison que ces gènes codent pour des protéines de transfert fonctionnelles. Des essais rapporteurs ont permis de confirmer que l’expression de ces gènes sur SGI1 est activée par AcaCD. Par la suite, nous avons montré que les protéines codées par SGI1 sont incorporées au pore de conjugaison, remplaçant les unités correspondantes encodées par les plasmides IncC. Enfin, nous avons montré que cette incorporation, impliquant notamment un remplacement de l’unité TraG, permet à SGI1 d’échapper à l’exclusion d’entrée exercée par les plasmides IncC.

Les gènes traNS, traGS et traHS sont essentiels à la mobilisation de SGI1. Non seulement ils sont activés par AcaCD à l’arrivée d’un plasmide IncC dans la souche, mais traHGS est également exprimé de façon constitutive à un faible niveau. Il est envisageable que l’accumulation d’unités

TraHS aide à son incorporation avant l’unité homologue encodée par le plasmide IncC. D’autres résultats montrent que les unités TraHS et TraGC sont incompatibles, rendant le pore de conjugaison inutilisable pour les deux éléments. Une incorporation précoce de TraHS pourrait constituer un prérequis au recrutement de TraGS. Une fois les deux unités incorporées, TraNS jouerait le rôle d’adaptateur en aidant à la stabilisation du pore nouvellement remodelé. La

45 substitution de TraG empêchant l’interaction habituelle avec la protéine d’exclusion encodée par les plasmides IncC, SGI1 peut dès lors transférer vers une réceptrice même si celle-ci porte également un plasmide IncC.

Cette étude a d’importantes implications épidémiologiques, non seulement puisqu’elle décrit un mécanisme permettant une optimisation de la propagation d’un ilot génomique de résistance, mais en plus parce qu’elle remet en question notre vision de ces éléments supposés être passifs. En vérité, SGI1 n’attend que l’arrivée d’un plasmide IncC dans la cellule pour enclencher des mécanismes promouvant activement sa dissémination. Ces données, ainsi que la conservation des gènes traS chez la majorité des variants décrits de SGI1, laissent entendre qu’ils ont été sélectionnés pour l’avantage évolutif conféré à cette famille d’éléments.

2.1.2. Apport des auteurs

L’approche expérimentale a été définie par Nicolas Carraro et Vincent Burrus. Les premières expériences ont été réalisées par Charley Anquetil sous la supervision de Nicolas Carraro et Catherine Barrette sous la supervision de Nicolas Rivard. Pour ma part, j’ai réalisé les expériences et analyses des Figures 2.2 et 2.5.

L’article a subi plusieurs rondes de révision, nécessitant de réitérer et d’ajouter de nombreuses expériences, dont Nicolas Rivard, Malika Humbert et moi-même avons été en charge. Par la suite, j’ai réalisé toutes les analyses statistiques et ai produit les nouvelles versions de toutes les figures.

Nicolas Carraro a rédigé la première version du manuscrit, puis j’ai participé à son édition avec Nicolas Carraro, Nicolas Rivard et Vincent Burrus. Finalement, le manuscrit a été relu par Alain Lavigueur.

46 2.1.3. Référence

Carraro N*, Durand R*, Rivard N, Anquetil C, Barrette C, Humbert M, Burrus V. (2017) Salmonella Genomic Island 1 (SGI1) reshapes the mating apparatus of IncC conjugative plasmids to promote self-propagation. PLoS Genet 13(3): e1006705. https://doi.org/10.1371/journal.pgen.1006705

*co-1er auteur

Mots clés : IncC, SGI1, T4SS, conjugaison, exclusion

47 2.2 Article

2.2.1. Page titre

Salmonella genomic island 1 (SGI1) reshapes the mating apparatus of IncC conjugative plasmids to promote self-propagation

Nicolas Carraro*, Romain Durand*, Nicolas Rivard, Charley Anquetil, Catherine Barrette, Malika Humbert, Vincent Burrus.

* These authors contributed equally to this work.

48 2.2.2. Abstract

IncC conjugative plasmids and Salmonella genomic island 1 (SGI1) and relatives are frequently associated with multidrug resistance of clinical isolates of pathogenic Enterobacteriaceae. SGI1 is specifically mobilized in trans by IncA and IncC plasmids (commonly referred to as A/C plasmids) following its excision from the chromosome, an event triggered by the transcriptional activator complex AcaCD encoded by these helper plasmids. Although SGI1 is not self- transmissible, it carries three genes, traNS, traHS and traGS, coding for distant homologs of the predicted mating pore subunits TraNC, TraHC and TraGC, respectively, encoded by A/C plasmids. Here we investigated the regulation of traNS and traHGS and the role of these three genes in the transmissibility of SGI1. Transcriptional fusion of the promoter sequences of traNS and traHGS to the reporter gene lacZ confirmed that expression of these genes is inducible by AcaCD. Mating experiments using combinations of deletion mutants of SGI1 and the helper IncC plasmid pVCR94 revealed complex interactions between these two mobile genetic elements. Whereas traNC and traHGC are essential for IncC plasmid transfer, SGI1 could rescue null mutants of each individual gene revealing that TraNS, TraHS and TraGS are functional proteins. Complementation assays of individual traC and traS mutants showed that not only do

TraNS/HS/GS replace TraNC/HC/GC in the mating pore encoded by IncC plasmids but also that traGS and traHS are both required for SGI1 optimal transfer. In fact, remodeling of the IncC- encoded mating pore by SGI1 was found to be essential to enhance transfer rate of SGI1 over the helper plasmid. Furthermore, traGS was found to be crucial to allow DNA transfer between cells bearing IncC helper plasmids, thereby suggesting that by remodeling the mating pore SGI1 disables an IncC-encoded entry exclusion mechanism. Hence traS genes facilitate the invasion by SGI1 of cell populations bearing IncC plasmids.

2.2.3. Author summary

Acquisition and dissemination of multidrug resistance genes among Enterobacteriaceae is in part driven by IncA and IncC (A/C) conjugative plasmids and Salmonella genomic island 1

49 (SGI1). Although unrelated, SGI1 relies on the self-transmissible A/C plasmids to disseminate within bacterial populations. The mechanisms allowing SGI1 to hijack the mating apparatus synthesized by A/C plasmids have not been previously established. Here, we show that IncC plasmids trigger the expression of three SGI1-borne genes that code for functional mating pore subunits distantly related to those encoded by the IncC helper plasmids. Our results indicate that these subunits alter the mating pore encoded by IncC plasmids to ensure optimal transfer of SGI1 and promote SGI1 dissemination in cell populations harboring IncC plasmids. Apart from SGI1 and relatives, documented mobilizable genomic islands are not known to code for mating pore components, possibly because of redundancy with those encoded by helper conjugative elements. Instead they usually code for mobilization proteins such as a relaxase and auxiliary factors involved in DNA recognition, processing and docking to the mating pore encoded by their helper conjugative element. From an ecological and epidemiological perspective, the strategy used by SGI1 likely confers a strong competitive advantage to SGI1 over IncC plasmids in clinical settings and could account for the high prevalence of SGI1 and relatives in multidrug- resistant Salmonella enterica and Proteus mirabilis.

2.2.4. Introduction

Conjugation is a nearly ubiquitous mechanism of horizontal gene transfer in , allowing the exchange of the largest number of genes per transfer event, often across taxonomical barriers [1-3]. In Gram-negative and most Gram-positive bacteria, conjugation is mediated by a complex nano-machine called type IV secretion system (T4SS). Conjugative T4SSs are multiprotein complexes that span the cell envelope and translocate DNA substrates from a donor to a recipient cell [4]. Conjugative plasmids and integrative and conjugative elements (ICEs) code for T4SS to promote their own dissemination by conjugation. These self-transmissible mobile genetic elements often bear a gene cargo of metal and antibiotic resistance genes, virulence determinants and other traits with potential selective advantages for the bacterial host [2,5-7].

50 Conjugative plasmids of the incompatibility group C (IncC) have been found in a broad range of Enterobacteriaceae and in Vibrio cholerae in which they can replicate and efficiently transfer [8-11]. IncC plasmids are closely related to IncA plasmids, and together are collectively referred to as A/C plasmids [8]. IncC plasmids are often recovered from clinical isolates of major pathogenic bacteria to which they confer resistance against multiple drugs, including last-resort antibiotics such as carbapenems [8,12,13]. IncC plasmids share a common scaffold of genes necessary for their replication, stability, conjugative transfer and regulation (Fig 2.1) [8,14]. Expression of the conjugative transfer genes of IncC plasmids is controlled by repressors Acr1 and Acr2 [15]. These two transcriptional repressors control the transcription of an operon containing acaC and acaD that code for the two subunits of the master activator AcaCD [15,16]. In IncC plasmids, AcaCD specifically binds to and activates 18 promoters that drive the expression of multiple genes and operons, most of which are of unknown function [15]. A third of these promoters drive the expression of tra genes coding for F-type T4SS assembly

(traLEKB, traVA, dsbC/traC/trhF/traWU, traFHG), mating pair stabilization (traN), and the relaxase and type IV coupling protein (traIDJ) necessary for conjugative transfer (Fig 2.1) [17,18].

Figure 2.1. Linear schematic representation of the core sequence of IncC plasmids and of Salmonella genomic island 1 (SGI1). The position and orientation of open reading frames (ORFs) are indicated by arrowed boxes. Colors depict the function deduced from functional analyses

51 and BLAST comparisons. AcaCD binding sites are represented by green angled arrows. SGI1 is flanked by the attL (vertical blue line on the left) and attR (vertical blue line on the right) attachment sites when integrated into the 3' end of the trmE gene in the chromosome of Salmonella enterica serovar Typhimurium DT104. ACSSuT, resistance to ampicillin, chloramphenicol, streptomycin/spectinomycin, sulfamethoxazole and tetracycline.

AcaCD also activates the expression of genes carried by unrelated mobilizable genomic islands (MGIs), thereby triggering their excision from the chromosome and their IncC-dependent dissemination into new bacterial hosts. Such MGIs include MGIVmi1 from Vibrio mimicus, MGIVchHai6 from Vibrio cholerae and Salmonella genomic island 1 (SGI1) from Salmonella enterica serovar Typhimurium DT104 [15,19-22]. SGI1 and its multiple variants are frequently found in S. enterica and Proteus mirabilis clinical isolates [23-25]. SGI1 and relatives all share a common 26-kb core region disrupted by a complex class 1 integron conferring multidrug resistance (Fig 2.1) [23,26,27].

SGI1 is thought to hijack the conjugative apparatus encoded by IncA and IncC plasmids to transfer to a new host cell by a mechanism that remains unknown as no origin of transfer (oriT) or mobilization protein such as a relaxase has been identified in SGI1 to date [28]. Remarkably, SGI1 was reported to be mobilized at a much higher rate than MGIVmi1 and MGIVchHai6 (~3 logs higher) when mobilized by the same IncC plasmid, even outperforming the helper plasmid by 10 fold [15,20]. Unlike MGIVmi1 and MGIVchHai6, the core region conserved in SGI1- related elements codes for two putative T4SS subunits, TraGS and TraHS, as well as a putative mating pair stabilization protein, TraNS, that are distantly related to the counterparts TraGC

(Vcrx144), TraHC (Vcrx143) and TraNC (Vcrx084) encoded by IncC plasmids (Fig 2.1, Table 2.1) [18,29-33].

52 Table 2.1. Protein homologs encoded by IncC plasmids and SGI1.

Name Length Identity Similarity Coverage Signal Pfam Predicted function (aa) peptide domain

TraNC/TraNS 933/920 78 % 88 % 97 % yes PF06986 Mating pair stabilization, adhesin

TraHC/TraHS 478/475 64 % 78 % 92 % yes PF06122 Mating apparatus formation/stabilization

TraGC/TraGS 1205/1135 37 % 57 % 99 % no PF07916 Mating apparatus stabilization, entry exclusion

While these observations suggest that the putative T4SS subunits encoded by SGI1 could be involved in SGI1 spread, they have been shown to be dispensable for its mobilization [28]. Like xis, a gene coding for the recombination directionality factor Xis that facilitates the excision of SGI1 from the chromosome, expression of the three tra genes of SGI1 has recently been reported to be under the control of AcaCD (Fig 2.1) [15,34,35]. This raises the question of the functional role of the putative tra genes of SGI1.

In this study, we investigated whether SGI1 could alter the mating pore encoded by IncC plasmids to enhance its own transfer. First, we confirmed that the genes traNS, and traHGS of SGI1 are under AcaCD control and that this cluster of tra genes is important for SGI1 mobilization. Using combinations of deletion mutants and complementation assays, we explored the role of each TraC/TraS subunit on the formation of the mating pore and its proficiency to mediate transfer of SGI1 and/or the helper IncC plasmid. Finally, we demonstrated that substitution of the TraG subunit enables SGI1 to escape an entry exclusion mechanism encoded by IncC plasmids.

53 2.2.5. Results

2.2.5.1 SGI1 traS genes are induced by the master activator of transfer of IncC plasmids

Expression of traNC and traHGC of IncC plasmids is activated by AcaCD, which recognizes a specific AcaCD box upstream of the -35 sequence in the promoters PtraNc and PtraF, respectively

(Fig 2.1) [15]. Two AcaCD boxes were also predicted in intergenic sequences upstream of traNS and traHGS. To test whether expression of SGI1 tra genes is AcaCD-dependent, promoters of traNS (PtraNs) and traHGS (PtraHs) were cloned upstream of a promoterless lacZ gene and their activity was measured by β-galactosidase assays. Controls included the constitutive promoter

Pint and the AcaCD-inducible promoter Pxis that drive the expression of the integrase- and Xis- coding genes of SGI1, respectively.

Activity of PtraNs was below the level of detection in absence of AcaCD, whereas in the same condition, PtraHs exhibited a weak yet detectable constitutive activity that remained weaker than the constitutive expression of Pint (Fig 2.2A). AcaCD increased the activity of PtraNs and PtraHs by 6 and 10 fold, respectively, as observed for Pxis (Fig 2.2A and 2.2B). Therefore, like their

IncC counterparts, expression of SGI1-borne traS genes is directly stimulated by AcaCD, suggesting that TraNS, TraHS and TraGS of SGI1 are produced alongside with TraNC, TraHC and TraGC of IncC plasmids. Thus, the SGI1-coded TraS subunits might complement or even compete with those encoded by IncC plasmids and lead to the synthesis of a hybrid T4SS with altered properties.

54

Figure 2.2. Regulation of expression of int, xis, traNS and traHGS of SGI1.

(A) Activity of Pint, Pxis, PtraNs and PtraHs was monitored from single-copy, chromosomally integrated lacZ transcriptional fusions in E. coli BW25113 Nx. Colorimetric assays of β-galactosidase activity were carried out on LB medium

supplemented with (+) or without (-) arabinose to express acaCD from PBAD on

pacaCD. (B) Induction of Pint, Pxis, PtraNs and PtraHs in response to AcaCD. β- galactosidase assays were carried out using the same strains as in panel A.

55 Figure 2.2. Regulation of expression of int, xis, traNS and traHGS of SGI1. (continued) Ratios between normalized OD420 values in the arabinose-induced over non- induced pacaCD, and noninduced pacaCD over arabinose-induced empty pBAD30 vector are shown. The bars represent the mean and standard deviation values obtained from at least three independent experiments. Statistical analyses were performed using the one-way ANOVA followed by Sidak's post- test to compare each induction ratio to its corresponding control. Statistical significance is indicated as followed: ****, P < 0.0001; ***, P < 0.001; ns, not significant.

2.2.5.2 TraNC, TraHC and TraGC are essential for IncC plasmid transfer

While predicted to be part of the mating apparatus of IncC plasmids, whether TraNC, TraHC or

TraGC are necessary for conjugative transfer has not yet been established [8]. To investigate this, we carried out mating experiments using a set of deletion mutants of pVCR94ΔX2 (Su Sp) as well as complementation assays aimed at evaluating their importance for conjugative transfer.

Individual deletion of traNC, traHC or traGC completely abolished conjugative transfer of pVCR94ΔX2 (Fig 2.3). Trans-complementation of each deletion mutant by expressing the missing gene from the medium-copy plasmid pBAD30 under the control of the arabinose- inducible PBAD promoter restored pVCR94ΔX2 transfer to wild-type level, thereby confirming that the mutations were non-polar (Fig 2.3). Therefore, the predicted T4SS subunits TraNC,

TraGC and TraHC are essential for conjugative transfer of IncC plasmids.

56

Figure 2.3. Role of TraN/G/HC on conjugative transfer of IncC plasmids.

Effect of traNC, traGC and traHC on conjugative transfer of pVCR94ΔX2. Conjugation assays were carried out using E. coli BW25113 Nx containing the indicated elements as donor strains and E. coli CAG18439 (Tc) as the recipient strain. Wild-type (WT) or derivative mutants of pVCR94ΔX2 as well as the gene expressed from pBAD30 are indicated below each graph; (-) indicates that

the plasmid is not present in the donor cell. For clarity, gene names traXY were

shortened XY. Transfer frequencies are expressed as the number of transconjugants per NxR KnR SpR donor CFUs and the ratio of transfer frequencies relative to WT is indicated at the base of each bar. The bars represent the mean and standard deviation values obtained from at least three independent experiments. “x” indicates that the frequency of transfer was below the detection limit (<10−7). Statistical analyses were carried out on the logarithm of the values using the one-way ANOVA with Tukey's multiple comparison test. Statistical significance is indicated as followed: ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant.

57 2.2.5.3 The traS gene cluster is important for SGI1 mobilization

To facilitate further investigations on SGI1, we substituted its multidrug resistance locus In104 with a kanamycin-resistance (Kn) cassette, while preserving the core genes conserved in all SGI1-like elements (Fig 2.1). Resulting SGI1ΔIn104 (Kn) was then used with pVCR94ΔX2 to investigate the role of traNS and traHGS, and of their respective IncC-borne homologs in the dissemination of SGI1 and IncC plasmids.

As shown in Fig 2.2, expression of traNS, traHS and traGS in SGI1 is triggered by AcaCD of IncC plasmids. Nevertheless, whether these genes code for proteins able to contribute to the formation of a functional T4SS was unclear. Kiss et al. [28] reported that mobilization of a natural SGI1 variant lacking ~10 kb (from 3,610 to 13,537 bp) including open reading frames from traNS to traHS was not significantly impacted by the deletion. This suggests the region encompassing traNS to traHS does not contain indispensable functions for SGI1 mobilization.

To verify this, we constructed a ΔtraS mutant of SGI1ΔIn104 lacking the whole traNS to traHS region and compared the mobilization by pVCR94ΔX2 of SGI1ΔIn104 and its ΔtraS mutant.

Unlike previously reported [28], we observed that ΔtraS led to ~4,000-fold decrease of transfer −4 −4 (2.27 ± 1.42 for SGI1ΔIn104 vs 5.46×10 ± 1.20×10 for its ΔtraS mutant), thereby suggesting that genes included in the traS gene cluster are important for SGI1 mobilization.

2.2.5.4 SGI1 complements a defective IncC-encoded T4SS with functional subunits

We took advantage of the non-transmissible traNC, traHC and traGC null mutants of pVCR94ΔX2 that likely produce a non-functional mating apparatus to test whether SGI1ΔIn104 could rescue such mutants. If the putative T4SS subunits encoded by SGI1ΔIn104 can replace the ones missing in the T4SS encoded by pVCR94ΔX2, transmissibility of the plasmid should be restored. Remarkably, SGI1ΔIn104 could restore conjugative transfer of each individual mutant to levels that were comparable to wild-type pVCR94ΔX2, while also allowing its own transfer (second pair of bars in Fig 2.4A, 2.4B and 2.4C). These results indicate that SGI1

58 produces functional mating pore components that can replace the missing corresponding parts in the IncC T4SS.

To confirm this hypothesis, trans-complementation of the traNC, traGC and traHC null mutants of pVCR94ΔX2 was carried out by providing donor cells with traNS, traGS and traHS, respectively, expressed from pBAD30. Expression of traNS restored conjugative transfer of pVCR94ΔX2 ΔtraNC to wild-type level (Fig 2.3). Surprisingly, despite its strong divergence from TraGC (Table 2.1), TraGS restored conjugative transfer of pVCR94ΔX2 ΔtraGC with only a mere 5-fold reduction compared to wild-type (Fig 2.3). In contrast, although TraHC and TraHS share 64% identity, expression of traHS failed to restore conjugative transfer of pVCR94ΔX2

ΔtraHC (Fig 2.3). This latter observation was puzzling because SGI1ΔIn104 could complement the traHC null mutant of pVCR94ΔX2.

2.2.5.5 TraGS is required for optimal SGI1 transfer

As shown above, TraGS is a functional substitute for TraGC as it restored transfer of pVCR94ΔX2 ΔtraGC (Fig 2.3). In contrast, when the reciprocal experiment was carried out, we found that TraGC is a poor substitute for TraGS for mediating transfer of SGI1ΔIn104. While

SGI1ΔIn104 transfer was unaffected by the ΔtraGC mutation in pVCR94ΔX2 (Fig 2.4A),

SGI1ΔIn104 ΔtraGS transfer was strongly reduced despite the presence of wild-type pVCR94ΔX2. Taken together, these results suggest that TraGS is required for optimal transfer of SGI1 whereas efficient transmissibility of IncC plasmids can be mediated by TraGC and/or

TraGS (Figs 2.3 and 2.4A).

59

Figure 2.4. Role of traC and traS genes on conjugative transfer of IncC plasmids and SGI1.

60 Figure 2.4. Role of traC and traS genes on conjugative transfer of IncC plasmids and SGI1. (continued)

(A) Effect of traGC and traGS on conjugative transfer of pVCR94ΔX2 (light

gray bars) and SGI1ΔIn104 (dark gray bars). (B) Effect of traHC and traHS. (C)

Effect of traNC and traNS. (D) Effect of combinatory deletions. For details, refer to legend of Fig 2.3. Two one-way ANOVA with Tukey's multiple comparison test were carried out separately for each element. The unpaired t- test (two-tailed) was performed to compare the bars of different elements. WT frequencies of transfer for both elements come from a single set of experimental replicates but are displayed throughout panels A to D as a reference in each statistical analysis.

Combination of both traGC and traGS mutations completely abolished transfer of both elements, confirming the key role of this T4SS component (Fig 2.4A and Fig S2.1A). In this context, overexpression of either traGC or traGS nearly restored full transfer of pVCR94ΔX2 ΔtraGC.

However, traGC was inefficient at complementing SGI1ΔIn104 ΔtraGS and resulted in a 250- fold decrease of transfer compared to SGI1ΔIn104 (Fig 2.4A). In contrast, providing traGS in trans fully restored transfer of SGI1ΔIn104 ΔtraGS. Therefore, traGS seems to be a key factor for enhancing the transmissibility of SGI1 relatively to IncC plasmids.

2.2.5.6 TraHS is specifically required with TraGS for SGI1 optimal transfer

Overexpression of traHS was unable to complement a traHC null mutant of pVCR94ΔX2, whereas wild-type SGI1ΔIn104 rescued pVCR94ΔX2 ΔtraHC (Figs 2.3 and 2.4B). As traHS and traGS seem to be part of the same operon in SGI1, like their homologs traHC and traGC in IncC plasmids (Fig 2.1), we suspected that the products of these genes might be interacting partners in the T4SS [14,15]. Substitution of a cognate partner within a pair of interacting proteins by a homologous protein encoded by another element is likely to impair these interactions and affect the functionality of the resulting hybrid mating pore. To test this

61 hypothesis, complementation assays were performed using ptraHGS to coexpress both traHS and traGS in cells bearing either pVCR94ΔX2 ΔtraHC or pVCR94ΔX2 ΔtraHGC (Fig 2.3). In both cases, conjugative transfer of pVCR94ΔX2 was partially restored, thereby confirming that

TraGS and TraHS work together within the mating pore. Together with the lack of complementation of pVCR94ΔX2 ΔtraHC by ptraHS, this latter observation likely reflects the inability of the SGI1-encoded TraHS subunit to interact with the IncC plasmid-encoded TraGC subunit to form a functional T4SS.

Although SGI1ΔIn104 ΔtraHS combined with pVCR94ΔX2 did not prevent the formation of a functional mating pore since pVCR94ΔX2 transferred at wild-type level, we observed that transfer of SGI1ΔIn104 ΔtraHS was reduced by 100 fold (Fig 2.4B). Because TraGS is required for optimal transfer of SGI1ΔIn104 (Fig 2.4A), this observation suggests that although TraHS does not work in association with TraGC, TraHC associated with TraGS could form a functional mating pore.

To investigate further these interactions, we combined pVCR94ΔX2 and SGI1ΔIn104 mutants, as well as expression of TraC/TraS subunits from pBAD30 vectors in conjugative transfer experiments. Neither element transferred when the traHC/traHS combination of mutants was used (Fig 2.4B and Fig S2.1B). However, complementation by providing either traHC or traHS in trans partially restored transfer of both elements. When pVCR94ΔX2 ΔtraHC was combined with SGI1ΔIn104 ΔtraGS, no transfer was detected for either element (Fig 2.4D), thereby confirming that TraGC and TraHS are incompatible and unable to form a functional T4SS. In contrast, the reciprocal association of pVCR94ΔX2 ΔtraGC with SGI1ΔIn104 ΔtraHS allowed transfer of both elements to near wild-type levels, thereby confirming that association of TraHC with TraGS can sustain formation of a functional and efficient hybrid T4SS (Fig 2.4B).

Altogether, these results revealed that while SGI1 can use a mating pore entirely encoded by

IncC plasmids, the expression and association of TraHS with TraGS are necessary for its optimal transfer, which largely surpasses the transfer rate of the helper plasmid.

62 2.2.5.7 TraNC and TraNS proteins are exchangeable

Additional experiments were performed to assess the impact of traNC and traNS on the transfer ability of pVCR94ΔX2 and SGI1. Transfer assays using a donor bearing pVCR94ΔX2 and

SGI1ΔIn104 ΔtraNS revealed that traNS was dispensable for SGI1 transfer if the IncC helper plasmid provided TraNC (Fig 2.4C). Combination of ΔtraNC and ΔtraNS mutations abolished pVCR94ΔX2 transfer, although it allowed residual transfer of SGI1ΔIn104 ΔtraNS (>4 logs below SGI1ΔIn104 level). Complementation of both mutations using either ptraNC or ptraNS restored the transmissibility of both elements (Fig 2.4C and Fig S2.1C).

2.2.5.8 TraNS enhances SGI1 transfer through the hybrid T4SS

Attempts to rescue pVCR94ΔX2 ΔtraHGC with ptraHGS only partially restored transfer of the mutant plasmid (250-fold reduction of transfer compared to wild-type) (Fig 2.3), whereas wild- type SGI1ΔIn104 restored transfer of pVCR94ΔX2 ΔtraHGC to near wild-type level (Fig 2.4D).

This observation suggests that TraNS of SGI1 interacts specifically with the TraHGS-containing

T4SS. This prompted us to test whether traNS could act cooperatively with traHGS to enhance the transmissibility of SGI1ΔIn104. We found that although SGI1ΔIn104 ΔtraNS transfer is not affected by the mutation in the context of wild-type pVCR94ΔX2, the concomitant absence of traHGC in pVCR94ΔX2 resulted in a 330-fold decrease of SGI1ΔIn104 ΔtraNS, despite the presence of SGI1ΔIn104-borne traHGS and pVCR94ΔX2-borne traNC (Fig 2.4D). In contrast, transfer of pVCR94ΔX2 was not affected in this context. These results showed that all three SGI1 Tra subunits seem to work together to promote its optimal transfer.

63 Table 2.2. Strains and plasmids used in this study.

Strains or element Relevant genotype or phenotype Source or reference E. coli BW25113 F- Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), λ-, rph-1, Δ(rhaD-rhaB)568, [45] hsdR514 BW25113 Nx Nx-derivative of BW25113 [11,46] CAG18439 MG1655 lacZU118 lacI42::Tn10 (Tc) [47,48] SM10λpir F- recA::RP4-2-Tc::Mu Km λpir [48]

Plasmids pVCR94ΔX2 Sp-derivative of the IncC plasmid pVCR94 (Su Sp) [15] pVCR94ΔX3 Kn-derivative of the IncC plasmid pVCR94 (Su Kn) [15] pVCR94ΔX4 Cm-derivative of the IncC plasmid pVCR94 (Su Cm) This study pVCR94ΔX2 ΔtraNC traNC deletion mutant of pVCR94ΔX2 (Su Sp) This study pVCR94ΔX2 ΔtraGC traGC deletion mutant of pVCR94ΔX2 (Su Sp) This study pVCR94ΔX2 ΔtraHC traHC deletion mutant of pVCR94ΔX2 (Su Sp) This study pVCR94ΔX2 ΔtraHGC traHGC deletion mutant of pVCR94ΔX2 (Su Sp) This study pSIM6 Thermo-inducible expression of λRed recombination (Ts Ap) [49] pSIM18 Thermo-inducible expression of λRed recombination (Ts Hy) [49] pKD3 Cm template for one-step chromosomal gene inactivation [45] pKD13 Kn template for one-step chromosomal gene inactivation [45] pCP20 Thermo-inducible expression of Flp recombinase (Ts Ap Cm) [50] pBAD30 orip15A araC PBAD (Ap) [51] pacaCD pBAD30::acaCD (Ap) [15] ptraNC pBAD30::traNC (Ap) This study ptraGC pBAD30::traGC (Ap) This study ptraHC pBAD30::traHC (Ap) This study ptraNS pBAD30::traNS (Ap) This study ptraGS pBAD30::traGS (Ap) This study ptraHS pBAD30::traHS (Ap) This study ptraHGS pBAD30::traHGS (Ap) This study pOPlacZ oriVR6Kγ; attPλ; promoterless lacZ (Kn) [15] pPromint pOPlacZ Pint-lacZ (Kn) This study pPromxis pOPlacZ Pxis-lacZ (Kn) This study pPromtraNS pOPlacZ PtraNs-lacZ (Kn) This study pPromtraHS pOPlacZ PtraHs-lacZ (Kn) This study pINT-Ts oriR101; cI857; λpR-intλ (Ap Ts) [52] pSU4628 CloDF13::TnAΔEcoRV (Ap) [38]

Genomic islands SGI1 Wild-type SGI1 integrated into the 3’ end of trmE (Ap Cm Sp Sm Su [15] Tc) SGI1ΔIn104 ΔIn104::aph, Kn-derivative of SGI1 This study SGI1ΔIn104 ΔtraNS traNS deletion mutant of SGI1ΔIn104 (Kn) This study SGI1ΔIn104 ΔtraGS traGS deletion mutant of SGI1ΔIn104 (Kn) This study SGI1ΔIn104 ΔtraHS traHS deletion mutant of SGI1ΔIn104 (Kn) This study SGI1ΔIn104 ΔtraHGS traHGS deletion mutant of SGI1ΔIn104 (Kn) This study SGI1ΔIn104 ΔtraS traS::cat deletion mutant of SGI1ΔIn104 (Kn Cm) This study

64 Table 2.2. Strains and plasmids used in this study. (continued) Ap, ampicillin; Cm, chloramphenicol; Hy, hygromycin B; Kn, kanamycin; Sm, streptomycin; Sp, spectinomycin; Su, sulfamethoxazole; Tc, tetracycline; Tm, trimethoprim; Ts, thermosensitive.

2.2.5.9 IncC plasmids exert entry exclusion

In F-type T4SSs, TraG is known to be a determinant of entry exclusion in the donor cell [36]. Entry exclusion is a process by which DNA transport from the donor cell is blocked by a recipient cell that contains a plasmid belonging to the same exclusion group. IncA and IncC plasmids seem to have been combined into A/C based on entry exclusion rather than incompatibility [37]. However, whether A/C plasmids exert entry exclusion had yet to be demonstrated. To test this, we monitored the mobilization of pSU4628, a derivative of the broad- host-range mobilizable plasmid CloDF13 (Table 2.2).

Although pSU4628 lacks T4SS genes, it codes for a mobilization protein that recognizes its cognate oriT and enables its translocation through P- and F-type T4SSs [38]. As a control, we first tested whether pSU4628 could be efficiently mobilized from E. coli SM10λpir which bears conjugative plasmid RP4 (P-type) to E. coli bearing or lacking pVCR94ΔX4. pSU4628 transferred at high frequency regardless of the presence of the IncC plasmid in the recipient (Fig 2.5A), thereby indicating that IncC plasmids do not exclude entry of DNA mediated by IncP RP4. In contrast, when pVCR94ΔX2 was used to mobilize pSU4628 to the same recipient strains, a 160-fold reduction of transfer (Exclusion index (EI)) was observed if pVCR94ΔX4 was present in the recipient, thereby confirming that IncC plasmids exert entry exclusion.

2.2.5.10 TraGS disables entry exclusion between cells bearing IncC plasmids

Genes mediating entry exclusion of A/C plasmids have not yet been characterized; however, by analogy with other F-type T4SSs [36], TraGC is likely the determinant of entry exclusion in

65 donor cells. Since SGI1 codes for its own TraGS subunit, we hypothesized that SGI1 would escape IncC entry exclusion, thereby facilitating DNA exchange between cells bearing IncC plasmids. To test this, we monitored pSU4628 mobilization from a donor bearing both pVCR94ΔX2 and SGI1ΔIn104 to recipient cells bearing or lacking pVCR94ΔX4. pSU4628 mobilization by pVCR94ΔX2 was enhanced by SGI1ΔIn104 at a rate comparable to mobilization by RP4 regardless of the presence of pVCR94ΔX4 in the recipient (EI = 1.6) (Fig 2.5A), thereby confirming that SGI1 disables IncC entry exclusion. Deletion of traHS had no significant impact (EI = 1.3) suggesting it plays no role in disabling IncC entry exclusion. In contrast, when SGI1ΔIn104 ΔtraGS was used, mobilization of pSU4628 was much reduced and the presence of the IncC plasmid in the recipient resulted in exclusion (EI = 27) (Fig 2.5A).

Mobilization of pSU4628 in the presence of SGI1ΔIn104 ΔtraHGS was comparable with donors lacking SGI1ΔIn104 (EI = 19).

66

Figure 2.5. Suppression by traGS of IncC entry exclusion. (A) IncC entry exclusion inhibits mobilization of pSU4628. E. coli SM10λpir (KnR) bearing pSU4628 (ApR) (hatched bars) or E. coli BW25113 Nx bearing pSU4628 and pVCR94ΔX2 (SpR) in the absence or presence of SGI1ΔIn104 (KnR) or its mutants were used as donors. E. coli CAG18439 (TcR) bearing or lacking pVCR94ΔX4 (CmR) was used as the recipient. Transfer frequencies are expressed as the number of TcR ApR transconjugants per KnR ApR donor CFUs (hatched bars) or per NxR (KnR) SpR ApR donor CFUs. Exclusion index (EI) is

indicated at the bottom of each gray bar. (B) Effect of traGS on mobilization of SGI1ΔIn104 when the helper IncC plasmid is in the recipient.

67 Figure 2.5. Suppression by traGS of IncC entry exclusion. (continued) R E. coli BW25113 Nx bearing SGI1ΔIn104 (WT or ΔtraGS, Kn ) and E. coli R R CAG18439 (Tc ) bearing pVCR94ΔX4 (Cm ) were crossed. traGS and traGC

complementation assays were carried out with ptraGS or ptraGC, respectively. Transfer frequencies are expressed as the number of NxR CmR transconjugants per TcR CmR donor CFUs for pVCR94ΔX4 and TcR KnR transconjugants per NxR CmR donor CFUs for SGI1ΔIn104. The bars represent the mean and standard deviation values obtained from at least three independent experiments. “x” indicates that the frequency of transfer was below the detection limit (<10−7). Statistical analyses were carried out on the logarithm of the values using the one-way ANOVA with Sidak's post-test (A) to compare each bar to its corresponding control using an empty recipient, and with Tukey's multiple comparison test (B) for each element. Statistical significance is indicated as followed: ****, P < 0.0001; *, P < 0.05; ns, not significant.

Recently, Sibor et al. [24] showed SGI1 mobilization when the IncC helper plasmid resides in the recipient strain. Since SGI1 is not self-transmissible, this observation suggests that prior to SGI1 mobilization, the donor strain acquires the IncC plasmid, which can then mobilize SGI1 toward the recipient. Such a two-step transfer of SGI1 can only occur if SGI1 disables IncC entry exclusion. We verified this using as a donor E. coli BW25113 Nx containing either

SGI1ΔIn104 (Kn) or its ΔtraGS mutant, and as a recipient E. coli CAG18439 bearing pVCR94ΔX4. Selection of intermediate BW25113 Nx transconjugants bearing pVCR94ΔX4 showed that the helper plasmid transferred efficiently regardless of the absence or presence of traGS (Fig 2.5B). Furthermore, while SGI1ΔIn104 transferred at high frequency to CAG18439 with pVCR94ΔX4, we failed to detect transfer of SGI1ΔIn104 ΔtraGS (Fig 2.5B). Mobilization was restored to wild-type level when the ΔtraGS mutant was complemented with traGS, whereas overexpression of traGC partially rescued transfer to levels 4 logs below the wild-type. This confirms that SGI1 fails to transfer to cells harboring an IncC helper if it must rely on a TraGC-

68 based T4SS. TraGS-based T4SS is critical for SGI1 propagation across bacterial population bearing IncC plasmids.

2.2.6. Discussion

Most known MGIs are opportunistic passengers riding the T4SS encoded by their conjugative helper element [39]. IncC conjugative plasmids have hitherto been shown to mobilize in trans three different MGIs: MGIVmi1 from V. mimicus, MGIVchHai6 from V. cholerae and SGI1 from S. enterica. Mobilization of both MGIVmi1 and MGIVchHai6 would rely on the auxiliary mobilization protein MobI, which would play the role of adaptor between the oriT of the MGIs and the relaxase of IncC plasmids [15,19,20]. However, this mechanism of mobilization is suboptimal with transfer rates 150 to 200 times lower than the IncC helper plasmid. In strong contrast, SGI1 was reported to transfer more than 10 times better than the same IncC helper plasmid [15,40]. Hence, SGI1 is not merely a free rider of the T4SS encoded by IncC plasmids, but rather tweaks the engine to its own benefit.

This study confirmed the key role of TraNC, TraGC and TraHC in the formation of the T4SS of

IncC plasmids. Expression of each protein restored the transfer of the corresponding mutant, although it did not enhance its transfer rate above wild-type level. Therefore, unlike the master regulator AcaCD [15], individual production of TraNC, TraGC or TraHC is not a limiting step for plasmid transfer. Moreover, our study of the three putative tra genes of SGI1 confirmed that not only is expression of these genes AcaCD-dependent as recently shown by Murányi et al. [35], but each one also codes for a fully functional T4SS subunit; together the Tra subunits of

SGI1 could complement individual traGC, traHC and traNC deletion mutants of pVCR94ΔX2. Heterologous complementation of T4SS functions by subunits encoded by different plasmids has already been reported for P-type T4SSs. For instance, the peptide hydrolase TraL from the IncN plasmid pKM101 can replace VirB1 of the VirB T4SS of Agrobacterium tumefaciens despite low sequence identity (31%) [41]. Likewise, their TraC and VirB5 components can also be exchanged [42]. Furthermore, the VirB10 homolog TrwE of IncW plasmid R388 can be

69 partially substituted for conjugation by TrwE of Bartonella tribocorum, a component of a T4SS involved in pathogenicity [43]. However, such exchanges of T4SS subunits between mobile genetic elements is uncommon and usually prevented in natural systems. For instance, interference between multiple functionally divergent T4SSs that co-occur in Bartonella is avoided by tight spatiotemporal regulation of expression or rapid diversification of the T4SS components [44].

Optimal SGI1 transfer depends on which of the subunits are composing the mating pore. Fig 2.6 illustrates the possible combinations of T4SS subunits and their outcome on SGI1 transfer efficiency inferred from our results (Figs 2.3 and 2.4). Our findings challenge a previous report by Kiss et al. [28] suggesting that SGI1 traS genes are not involved in SGI1 mobilization. In fact, our results indicate that traGS (collaboratively with traHS and traNS) enhances the transfer rate of SGI1ΔIn104 over the helper IncC plasmid (Fig 2.4). Moreover, traGC substitution by traGS enables SGI1 to invade cell populations bearing IncC plasmids likely by evading IncC entry exclusion (Fig 2.5B). Since SGI1 has also been shown to destabilize IncC plasmids [40], this mechanism sheds a new light on the ecological and epidemiological significance of SGI1 and relatives in the propagation of multidrug resistance. In fact, we predict that combination of entry exclusion escaping and IncC plasmid destabilization would result in displacement of IncC plasmids by SGI1 in enterobacterial cell populations bearing IncC plasmids upon contact with a small subpopulation of cell bearing only SGI1.

70

Figure 2.6. Schematic representation of mating pore configurations and impact on IncC plasmids and SGI1 transfer efficiency.

Each mating pore configuration shows only a single subunit of TraNC/S,

TraGC/S, TraHC/S, which are represented as described in the figure, and color- coded in purple for IncC plasmid-encoded subunits and blue for SGI1-encoded subunits. The rest of the mating apparatus is not shown and is assumed to be provided by the IncC plasmid. Efficiency of transfer of both elements is indicated by a green circle (optimal or efficient transfer for both elements), an orange triangle (impaired transfer for SGI1 but not IncC plasmid) or a red cross (abolished transfer for both elements) under the corresponding mating pore configuration as inferred from the results (Figs 2.3 and 2.4).

71 Deletion of both traNC and traNS revealed that, despite the lack of putative adhesin that seems to be required for transfer of IncC plasmids, SGI1 can still transfer at low frequency. The

TraNC/TraNS adhesins, which are thought to stabilize the mating cell pair [18,30,31], are likely required for the transfer of a large DNA molecule such as pVCR94ΔX2 (~120 kb), while the smaller size of SGI1 (~26 kb for SGI1ΔIn104) would render it less vulnerable to premature separation of the mating partners due to the shorter transfer time required to transfer the whole element. In addition, traNC and traNS could be easily exchanged without drastic impairment of

SGI1ΔIn104 or pVCR94ΔX2 transfer. This result is not surprising considering that TraNC and

TraNS are the least divergent proteins of the three orthologous pairs (78% identity) (Fig 2.1,

Table 2.1). An H (TraHC or TraHS) and a G (TraGC or TraGS) subunit are both required for assembly of a functional mating apparatus. However, we showed that all combinations are not functionally equivalent, as TraHS and TraGC appeared to be incompatible (Fig 2.4B). Thus,

SGI1-encoded subunit TraHS specifically interacts with TraGS, strongly enhancing the efficiency of SGI1 transfer. Altogether, our observations suggest that the TraHGS association allows a specific interaction with protein(s) and/or DNA of SGI1 to optimize its transfer. One candidate could be the relaxosome, i.e. the machinery that processes DNA at the SGI1-borne origin of transfer (oriT) to allow its transfer. oriT of SGI1 and components of the relaxosome that process it remain to be identified and could be partly encoded by SGI1 to confer specificity to the altered mating pore.

Chromosomally integrated SGI1 is not entirely quiescent. While expression of IncC plasmid- borne traC genes strictly depends on AcaCD activation [15], traHGS are constitutively transcribed at low level in an AcaCD-independent fashion (Fig 2.2). Basal expression of traHGS strongly suggests that SGI1 primes the bacterial cell to accumulate TraGS and TraHS subunits. While production of these T4SS subunits is likely vain when SGI1 is alone, upon arrival of a helper IncC plasmid, cells primed with TraGS and TraHS might be more prone to rapidly incorporate SGI1-encoded subunits in lieu of IncC-plasmid-encoded ones, thereby favoring the transfer of SGI1 over the helper plasmid. In contrast, expression of traNS is strictly dependent upon AcaCD activation (Fig 2.2) [35]. Tight control over traNS expression may have been

72 selected to prevent futile expression of the cell surface-exposed adhesin, which could potentially serve as a receptor for infection by bacteriophages.

In conclusion, unlike any other known mobilizable genomic island described to date, SGI1 not only hijacks the mating pore encoded by IncC plasmids, but also customizes it by inserting its T4SS subunits. This strategy enhances the propagation of SGI1 in bacterial populations as a result of enhanced transfer rates and expansion of its host range to recipient cells bearing IncC plasmids. This study takes us one step further into the comprehension of the intimate relation that links the mobility of unrelated classes of multidrug resistance-conferring mobile genetic elements.

2.2.7. Materials and methods

2.2.7.1 Bacterial strains and media

Bacterial strains and plasmids used in this study are described in Table 2.2. Strains were routinely grown in lysogeny broth (LB-Miller, EMD) at 37°C in an orbital shaker/incubator and were preserved at -80°C in LB broth containing 15% (vol/vol) glycerol. Antibiotics were used at the following concentrations: ampicillin (Ap), 100 μg/ml; chloramphenicol (Cm), 20 μg/ml; hygromycin B (Hy), 50 μg/ml; kanamycin (Kn), 50 μg/ml or 10 μg/ml for single copy integrants of pOPlacZ; nalidixic acid (Nx), 40 μg/ml; spectinomycin (Sp), 50 μg/ml; streptomycin (Sm), 200 μg/ml; sulfamethoxazole (Su), 160 μg/ml; tetracycline (Tc), 12 μg/ml; trimethoprim (Tm), 32 μg/ml. When required, bacterial cultures were supplemented with either 0.02 or 0.2% L- arabinose.

2.2.7.2 Mating assays

Conjugation assays were performed by mixing 100 μl of donor cells and 100 μl of recipient cells (typically ~2×109 cells/ml each) that were grown overnight in LB broth at 37°C with suitable

73 antibiotics to ensure retention of the plasmid and SGI1 derivatives. Cells were pelleted by centrifugation for 3 min at 1,200 g, washed once in 200 μl of LB broth and resuspended in 10 μl of LB broth. Mating mixtures were then deposited as drops on LB agar plates and incubated at 37°C for 6 hours. The cells were recovered from the plates in 800 μl of LB broth, vortexed and diluted via serial 10-fold dilutions before plating on LB agar plates containing suitable antibiotics. Donors were selected using a chromosomal marker, and as necessary a marker for pVCR94, SGI1ΔIn104 and/or pSU4628. To induce expression of tra genes in complementation assays, mating experiments were carried out onto LB agar plates with 0.02% arabinose. Frequency of transfer was calculated as transconjugants/donor from data obtained from at least 3 parallel mating experiments.

2.2.7.3 Molecular biology methods

Plasmid DNA was prepared using the EZ-10 Spin Column Plasmid DNA Minipreps Kit (Bio Basic) according to manufacturer's instructions. All enzymes used in this study were purchased from New England Biolabs. PCR assays were performed with the primers described in Table S2.1. PCR conditions were as follows: (i) 3 min at 94°C; (ii) 30 cycles of 30 sec at 94°C, 30 sec at the appropriate annealing temperature, and 1 minute/kb at 68°C; and (iii) 5 min at 68°C. When necessary, PCR products were purified using an EZ-10 Spin Column PCR Products Purification Kit (Bio Basic) according to manufacturer's instructions. E. coli was transformed by electroporation as described by Dower et al. [53] in a Bio-Rad GenePulser Xcell apparatus set at 25 μF, 200 Ω and 1.8 kV using 1-mm gap electroporation cuvettes. Sequencing reactions were performed by the Plateforme de Séquençage et de Génotypage du Centre de Recherche du CHUL (Québec, QC, Canada).

2.2.7.4 Plasmid and strain construction

Plasmids and oligonucleotides used in this study are listed in Table 2.2 and Table S2.1. Plasmids used for complementation assays were derived from pBAD30. traNC, traGC, and traHC were

74 amplified using primer pairs 94traN84EcoRI.for/94traN84EcoRI.rev, 94traG144EcoRI.for/94traG144EcoRI.rev, 94traH143EcoRI.for/94traH143EcoRI.rev, and genomic DNA of E. coli BW25113 Nx containing pVCR94ΔX2 as the template. Amplicons were digested by EcoRI and cloned into EcoRI-digested pBAD30 using T4 DNA ligase, generating ptraNC, ptraGC and ptraHC. Likewise, traNS, traGS, traHS and traHGS were amplified using primer pairs SGI105traNSalI.for/SGI105traNSalI.rev, SGI111traGSalI.for/SGI111traGSalI.rev, SGI1s012EcoRI.for/SGI1s012EcoRI.rev, and SGI1s012SalI.for/SGI111traGSalI.rev, and genomic DNA of E. coli BW25113 Nx containing SGI1 as template. Amplicons were digested by SalI or EcoRI and cloned into SalI or EcoRI- digested pBAD30 using the T4 DNA ligase, generating ptraNS, ptraGS, ptraHS and ptraHGS.

PCR fragments containing the promoter region upstream of int, xis, traNS, traHGS were amplified using primer pairs SGI1promintPstI.for/SGI1promintPstI.rev, SGI1promxisPstI.for/SGI1promxisPstI.rev, SGI1promtraNPstI.for/SGI1promtraNPstI.rev, SGI1promtraHPstI.for/SGI1promtraHPstI.rev and cloned into the PstI restriction site of pOPlacZ to produce pPromint, pPromxis, pPromtraNS, pPromtraHS, respectively [15]. The resulting plasmids were verified by restriction profiling and DNA sequencing. These vectors were integrated in single copy into the chromosomal site attBλ of E. coli BW25113 Nx using pINT-Ts [52].

Deletion mutants of pVCR94ΔX2 and SGI1 were constructed using the one-step chromosomal gene inactivation technique with pSIM6 or pSIM18 (Table 2.2) [45]. For pVCR94ΔX2, deletions of traNC, traGC, traHC and traHGC were obtained using primer pairs, 94del84traN.for/94del84traN.rev, 94del144traG.for/94del144traG.rev, 94del143traH.for/94del143traH.rev, 94del143traH.for/94del144traG.rev, respectively, and pKD3 as the template (Table 2.2 and Table S2.1). SGI1 derivative SGI1ΔIn104 was obtained using primer pair SGI1delVar.for/SGI1delVar.rev and pKD13 as the template. Subsequent deletions of traNS, traGS, traHS, traHGS and traNS-traHS region in SGI1ΔIn104 were obtained using primer pairs, SGI1delS005.for/SGI1del05traN.rev, SGI1delS011.for/SGI1delS011.rev,

75 SGI1delS012.for/SGI1delS012.rev, SGI1delS012.for/SGI1delS011.rev and SGI1delS012.for/SGI1del05traN.rev, respectively, and pKD3 as the template. Substitution of the aph (Kn) resistance gene with the cat (Cm) resistance gene in pVCR94ΔX3 was carried out using the same approach with primers 94DelXnoFRTcm.for and 94DelXnoFRTcm.rev, and pKD3 as the template, yielding pVCR94ΔX4.

When possible, the antibiotic resistance cassette was removed from the resulting construction by Flp-catalyzed excision using the pCP20 vector [50]. All deletions were verified by PCR and antibiotic resistance profiling.

2.2.7.5 β-galactosidase assays

Qualitative assays on solid LB agar plate were done using 40 μg/ml 5-bromo-4-chloro-3- indolyl-β-D-galactopyranoside (X-gal) as the substrate with or without 0.02% arabinose. Plates were observed after overnight incubation at 37°C.

Quantitative liquid assays using o-2-nitrophenyl-β-D-galactopyranoside (ONPG) as the substrate were done according to a protocol adapted from Miller [54]. After overnight incubation at 37°C in 4 ml LB broth supplemented with appropriate antibiotics, cultures were refreshed 1:100 in 4 ml LB broth supplemented with 10 μg/ml kanamycin, 25 μg/ml ampicillin and 0.2% arabinose except for the non-induced controls. Cultures were incubated for 5 hours at 37°C with shaking prior to sampling for enzymatic assays. OD measurements for enzymatic assays were performed using a Multiskan Go Microplate Spectrophotometer (Thermo Scientific). Each experiment was performed in at least three independent biological replicates. Induction ratios were calculated by dividing the “induced” values by the “non-induced” values whereas the control ratios were calculated by dividing the “non-induced” values by the control values.

76 2.2.8. Acknowledgments

We are grateful to Reynold Gerardo Farrera Calderón for his technical assistance and Alain Lavigueur for his insightful comments on the manuscript.

2.2.9. References

1. Smillie C, Garcillán-Barcia MP, Francia MV, Rocha EPC, de la Cruz F. Mobility of plasmids. Microbiol Mol Biol Rev MMBR. 2010;74: 434–452. doi:10.1128/MMBR.00020-10

2. Norman A, Hansen LH, Sorensen SJ. Conjugative plasmids: vessels of the communal gene pool. Philos Trans R Soc B Biol Sci. 2009;364: 2275–2289. doi:10.1098/rstb.2009.0037

3. Guglielmini J, Quintais L, Garcillán-Barcia MP, de la Cruz F, Rocha EPC. The Repertoire of ICE in Prokaryotes Underscores the Unity, Diversity, and Ubiquity of Conjugation. PLoS Genet. 2011;7: e1002222. doi:10.1371/journal.pgen.1002222

4. Christie PJ. The Mosaic Type IV Secretion Systems. EcoSal Plus. 2016;7. doi:10.1128/ecosalplus.ESP-0020-2015

5. Carraro N, Burrus V. Biology of Three ICE Families: SXT/R391, ICEBs1, and ICESt1/ICESt3. Microbiol Spectr. 2014;2. doi:10.1128/microbiolspec.MDNA3-0008-2014

6. Johnson CM, Grossman AD. Integrative and Conjugative Elements (ICEs): What They Do and How They Work. Annu Rev Genet. 2015;49: 577–601. doi:10.1146/annurev-genet- 112414-055018

7. Wozniak RAF, Waldor MK. Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Nat Rev Microbiol. 2010;8: 552–563. doi:10.1038/nrmicro2382

8. Harmer CJ, Hall RM. The A to Z of A/C plasmids. Plasmid. 2015;80: 63–82. doi:10.1016/j.plasmid.2015.04.003

9. Fricke WF, Welch TJ, McDermott PF, Mammel MK, LeClerc JE, White DG, et al. Comparative genomics of the IncA/C multidrug resistance plasmid family. J Bacteriol. 2009;191: 4750–4757. doi:10.1128/JB.00189-09

10. Welch TJ, Fricke WF, McDermott PF, White DG, Rosso M-L, Rasko DA, et al. Multiple antimicrobial resistance in plague: an emerging public health risk. PLoS One. 2007;2: e309. doi:10.1371/journal.pone.0000309

77 11. Carraro N, Sauvé M, Matteau D, Lauzon G, Rodrigue S, Burrus V. Development of pVCR94ΔX from Vibrio cholerae, a prototype for studying multidrug resistant IncA/C conjugative plasmids. Front Microbiol. 2014;5: 44. doi:10.3389/fmicb.2014.00044

12. Wailan AM, Sartor AL, Zowawi HM, Perry JD, Paterson DL, Sidjabat HE. Genetic contexts of blaNDM-1 in patients carrying multiple NDM-producing strains. Antimicrob Agents Chemother. 2015;59: 7405–7410. doi:10.1128/AAC.01319-15

13. Wang R, Yu D, Zhu L, Li J, Yue J, Kan B. IncA/C plasmids harboured in serious multidrug-resistant Vibrio cholerae serogroup O139 strains in China. Int J Antimicrob Agents. 2015;45: 249–254. doi:10.1016/j.ijantimicag.2014.10.021

14. Poulin-Laprade D, Carraro N, Burrus V. The extended regulatory networks of SXT/R391 integrative and conjugative elements and IncA/C conjugative plasmids. Front Microbiol. 2015;6: 837. doi:10.3389/fmicb.2015.00837

15. Carraro N, Matteau D, Luo P, Rodrigue S, Burrus V. The master activator of IncA/C conjugative plasmids stimulates genomic islands and multidrug resistance dissemination. PLoS Genet. 2014;10: e1004714. doi:10.1371/journal.pgen.1004714

16. Lang KS, Johnson TJ. Characterization of Acr2, an H-NS-like protein encoded on A/C2- type plasmids. Plasmid. 2016; doi:10.1016/j.plasmid.2016.07.004

17. Guglielmini J, Néron B, Abby SS, Garcillán-Barcia MP, de la Cruz F, Rocha EPC. Key components of the eight classes of type IV secretion systems involved in bacterial conjugation or protein secretion. Nucleic Acids Res. 2014;42: 5715–5727. doi:10.1093/nar/gku194

18. Lawley TD, Klimke WA, Gubbins MJ, Frost LS. F factor conjugation is a true type IV secretion system. FEMS Microbiol Lett. 2003;224: 1–15. doi:10.1016/S0378-1097(03)00430-0

19. Carraro N, Matteau D, Burrus V, Rodrigue S. Unraveling the regulatory network of IncA/C plasmid mobilization: when genomic islands hijack conjugative elements. Mob Genet Elem. 2015;5: 1–5. doi:10.1080/2159256X.2015.1045116

20. Carraro N, Rivard N, Ceccarelli D, Colwell RR, Burrus V. IncA/C Conjugative Plasmids Mobilize a New Family of Multidrug Resistance Islands in Clinical Vibrio cholerae Non- O1/Non-O139 Isolates from Haiti. mBio. 2016;7: e00509-16. doi:10.1128/mBio.00509-16

21. Douard G, Praud K, Cloeckaert A, Doublet B. The Salmonella genomic island 1 is specifically mobilized in trans by the IncA/C multidrug resistance plasmid family. PLoS One. 2010;5: e15302. doi:10.1371/journal.pone.0015302

22. Doublet B, Boyd D, Mulvey MR, Cloeckaert A. The Salmonella genomic island 1 is an integrative mobilizable element. Mol Microbiol. 2005;55: 1911–1924. doi:10.1111/j.1365- 2958.2005.04520.x

78 23. Hall RM. Salmonella genomic islands and antibiotic resistance in Salmonella enterica. Future Microbiol. 2010;5: 1525–1538. doi:10.2217/fmb.10.122

24. Siebor E, de Curraize C, Amoureux L, Neuwirth C. Mobilization of the Salmonella genomic island SGI1 and the Proteus genomic island PGI1 by the A/C2 plasmid carrying blaTEM- 24 harboured by various clinical species of Enterobacteriaceae. J Antimicrob Chemother. 2016;71: 2167–2170. doi:10.1093/jac/dkw151

25. Boyd DA, Shi X, Hu Q, Ng LK, Doublet B, Cloeckaert A, et al. Salmonella genomic island 1 (SGI1), variant SGI1-I, and new variant SGI1-O in Proteus mirabilis clinical and food isolates from China. Antimicrob Agents Chemother. 2008;52: 340–344. doi:10.1128/AAC.00902-07

26. Boyd D, Cloeckaert A, Chaslus-Dancla E, Mulvey MR. Characterization of variant Salmonella genomic island 1 multidrug resistance regions from serovars Typhimurium DT104 and Agona. Antimicrob Agents Chemother. 2002;46: 1714–1722.

27. Levings RS, Lightfoot D, Partridge SR, Hall RM, Djordjevic SP. The genomic island SGI1, containing the multiple antibiotic resistance region of Salmonella enterica serovar Typhimurium DT104 or variants of it, is widely distributed in other S. enterica serovars. J Bacteriol. 2005;187: 4401–4409. doi:10.1128/JB.187.13.4401-4409.2005

28. Kiss J, Nagy B, Olasz F. Stability, entrapment and variant formation of Salmonella genomic island 1. PLoS One. 2012;7: e32497. doi:10.1371/journal.pone.0032497

29. Mulvey MR, Boyd DA, Olson AB, Doublet B, Cloeckaert A. The genetics of Salmonella genomic island 1. Microbes Infect Inst Pasteur. 2006;8: 1915–1922. doi:10.1016/j.micinf.2005.12.028

30. Klimke WA, Frost LS. Genetic analysis of the role of the transfer gene, traN, of the F and R100-1 plasmids in mating pair stabilization during conjugation. J Bacteriol. 1998;180: 4036–4043.

31. Klimke WA, Rypien CD, Klinger B, Kennedy RA, Rodriguez-Maillard JM, Frost LS. The mating pair stabilization protein, TraN, of the F plasmid is an outer-membrane protein with two regions that are important for its function in conjugation. Microbiol Read Engl. 2005;151: 3527–3540. doi:10.1099/mic.0.28025-0

32. Arutyunov D, Arenson B, Manchak J, Frost LS. F plasmid TraF and TraH are components of an outer membrane complex involved in conjugation. J Bacteriol. 2010;192: 1730–1734. doi:10.1128/JB.00726-09

33. Harris RL, Silverman PM. Tra proteins characteristic of F-like type IV secretion systems constitute an interaction group by yeast two-hybrid analysis. J Bacteriol. 2004;186: 5480–5485. doi:10.1128/JB.186.16.5480-5485.2004

79 34. Kiss J, Papp PP, Szabó M, Farkas T, Murányi G, Szakállas E, et al. The master regulator of IncA/C plasmids is recognized by the Salmonella genomic island SGI1 as a signal for excision and conjugal transfer. Nucleic Acids Res. 2015;43: 8735–8745. doi:10.1093/nar/gkv758

35. Murányi G, Szabó M, Olasz F, Kiss J. Determination and Analysis of the Putative AcaCD-Responsive Promoters of Salmonella Genomic Island 1. PLOS ONE. 2016;11: e0164561. doi:10.1371/journal.pone.0164561

36. Garcillán-Barcia MP, de la Cruz F. Why is entry exclusion an essential feature of conjugative plasmids? Plasmid. 2008;60: 1–18. doi:10.1016/j.plasmid.2008.03.002

37. Hedges RW. R factors from Providence. J Gen Microbiol. 1974;81: 171–181. doi:10.1099/00221287-81-1-171

38. Cabezón E, Sastre JI, de la Cruz F. Genetic evidence of a coupling role for the TraG protein family in bacterial conjugation. Mol Gen Genet MGG. 1997;254: 400–406.

39. Bellanger X, Payot S, Leblond-Bourget N, Guédon G. Conjugative and mobilizable genomic islands in bacteria: evolution and diversity. FEMS Microbiol Rev. 2014;38: 720–760. doi:10.1111/1574-6976.12058

40. Harmer CJ, Hamidian M, Ambrose SJ, Hall RM. Destabilization of IncA and IncC plasmids by SGI1 and SGI2 type Salmonella genomic islands. Plasmid. 2016; doi:10.1016/j.plasmid.2016.09.003

41. Höppner C, Liu Z, Domke N, Binns AN, Baron C. VirB1 orthologs from Brucella suis and pKM101 complement defects of the lytic transglycosylase required for efficient type IV secretion from Agrobacterium tumefaciens. J Bacteriol. 2004;186: 1415–1422.

42. Schmidt-Eisenlohr H, Domke N, Baron C. TraC of IncN plasmid pKM101 associates with membranes and extracellular high-molecular-weight structures in Escherichia coli. J Bacteriol. 1999;181: 5563–5571.

43. de Paz HD, Sangari FJ, Bolland S, García-Lobo JM, Dehio C, de la Cruz F, et al. Functional interactions between type IV secretion systems involved in DNA transfer and virulence. Microbiol Read Engl. 2005;151: 3505–3516. doi:10.1099/mic.0.28410-0

44. Gillespie JJ, Phan IQH, Scheib H, Subramanian S, Edwards TE, Lehman SS, et al. Structural Insight into How Bacteria Prevent Interference between Multiple Divergent Type IV Secretion Systems. mBio. 2015;6: e01867-01815. doi:10.1128/mBio.01867-15

45. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97: 6640–6645. doi:10.1073/pnas.120163297

80 46. Grenier F, Matteau D, Baby V, Rodrigue S. Complete Genome Sequence of Escherichia coli BW25113. Genome Announc. 2014;2. doi:10.1128/genomeA.01038-14

47. Singer M, Baker TA, Schnitzler G, Deischel SM, Goel M, Dove W, et al. A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli. Microbiol Rev. 1989;53: 1–24.

48. Miller VL, Mekalanos JJ. A novel suicide vector and its use in construction of insertion mutations: osmoregulation of outer membrane proteins and virulence determinants in Vibrio cholerae requires toxR. J Bacteriol. 1988;170: 2575–2583.

49. Datta S, Costantino N, Court DL. A set of recombineering plasmids for Gram-negative bacteria. Gene. 2006;379: 109–115. doi:10.1016/j.gene.2006.04.018

50. Cherepanov PP, Wackernagel W. Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene. 1995;158: 9–14. doi:10.1016/0378-1119(95)00193-A

51. Guzman LM, Belin D, Carson MJ, Beckwith J. Tight regulation, modulation, and high- level expression by vectors containing the arabinose PBAD promoter. J Bacteriol. 1995;177: 4121–4130.

52. Haldimann A, Wanner BL. Conditional-replication, integration, excision, and retrieval plasmid-host systems for gene structure-function studies of bacteria. J Bacteriol. 2001;183: 6384–6393. doi:10.1128/JB.183.21.6384-6393.2001

53. Dower WJ, Miller JF, Ragsdale CW. High efficiency transformation of E. coli by high voltage electroporation. Nucleic Acids Res. 1988;16: 6127–6145. doi:10.1093/nar/16.13.6127

54. Miller JH. A short course in bacterial genetics: a laboratory manual and handbook for Escherichia coli and related bacteria. CSHL Press, Plainview NY. 1992;

81 2.2.10. Supporting information

Figure S2.1. Effect of traC/S genes on cotransfer of IncC plasmids and SGI1.

82 Figure S2.1. Effect of traC/S genes on cotransfer of IncC plasmids and SGI1. (continued)

Effect of traGC and traGS (A), and traHC and traHS (B) and traNC and traNS (C) and combinatory mutants (D), on cotransfer of pVCR94ΔX2 and SGI1ΔIn104. For details, refer to legend of Fig 2.3.

Table S2.1. Primers used in this study

Name Nucleotide sequence (5’ to 3’)a 94traN84EcoRI.for NNNNNNGAATTCAAGGAGGAATAATAAATGCGAAGTCATAATTACTTTATGA 94traN84EcoRI.rev NNNNNNGAATTCTTATCCACCACCACCAGTTGGCGCG 94traG144EcoRI.for NNNNNNGAATTCAAGGAGGAATAATAAATGGGATCTTTTTCAATCCACTCTA 94traG144EcoRI.rev NNNNNNGAATTCTTAGAACCTTTGATTTGCTATGTT 94traH143EcoRI.for NNNNNNGAATTCAAGGAGGAATAATAAATGGTCACGCACAAGACATTAAAA 94traH143EcoRI.rev NNNNNNGAATTCTTAGTTCGTGCTAGGAGGATTGGA SGI105traNSalI.for NNNNNNGTCGACAAGGAGGAATAATAAATGTCCACTATGCCGCCCAT SGI105traNSalI.rev NNNNNNGTCGACTCAACCACCATTGCCATCCAA SGI111traGSalI.for NNNNNNGTCGACAAGGAGGAATAATAAATGGATTTTAGTATTTATTCCGT SGI111traGSalI.rev NNNNNNGTCGACTTAACGGCCTCCTTTGGCTA SGI1s012EcoRI.for NNNNNNGAATTCAAGGAGGAATAATAAATGAGAGTCTCTCCCCCTTA SGI1s012EcoRI.rev NNNNNNGAATTCTTACTCCTTTTTCCCTGATGCTT SGI1s012SalI.for NNNNNNGTCGACAAGGAGGAATAATAAATGAGAGTCTCTCCCCCTTA SGI1promintPstI.for NNNNNNCTGCAGACACCTTGAGCAGGGCAAA SGI1promintPstI.rev NNNNNNCTGCAGCGTTACTCCAAAATTTTAACT SGI1promxisPstI.for NNNNNNCTGCAGGTTATTGATAGACCTAGTTTAT SGI1promxisPstI.rev NNNNNNCTGCAGGGCCAATGTGCCGGTTT SGI1promtraNPstI.for NNNNNNCTGCAGCCAGCTTTTTAGTTTGGATA SGI1promtraNPstI.rev NNNNNNCTGCAGGATAGAGCATTGCGAGCAAT SGI1promtraHPstI.for NNNNNNCTGCAGTAGGTTTCGTGTCACCCGAA SGI1promtraHPstI.rev NNNNNNCTGCAGTTAAAGCTCCTCTTTTAGAA TCTTCTCAGATAGCATGGGAAGGTTGAAATGGAGAACACAGTGTAGGCTGGAGCTG 94del84traN.for CTTC AATCAATCTACATGAGAAAGGGGCCGAAAGGCCCCTTTTTTTATTACATATGAATA 94del84traN.rev TCCTCCTTA TTGTCCAATCCTCCTAGCACGAACTAAGGAGCTATTGGAGTGTAGGCTGGAGCTGC 94del144traG.for TTC TAAGGGGGCCTGCTGGCCCCCTTATTATTTGGTTGGTCTTTTATTACATATGAATA 94del144traG.rev TCCTCCTTA CTCATGAAGTACATCCGGGACAAATTACAGGAGAACTAAGGTGTAGGCTGGAGCTG 94del143traH.for CTTC GAAAAGCAGAGTCACCGATAGAGTGGATTGAAAAAGATCCCATATGAATATCCTCC 94del143traH.rev TTA

83 Table S2.1. Primers used in this study (continued)

TATTTTAACATGTATATAGATTAAATTCCAATCAACTTGTTCGGAATAGGAACTTC SGI1delVar.for AAGA CAAACCACCATTTTCTTAACAGTTCATAACTTAGCAAGTTAGAGCGCTTTTGAAGC SGI1delVar.rev TCA CGACATAGCATTATCCAAACTAAAAAGCTGGAGAAATGCTGTGTAGGCTGGAGCTG SGI1delS005.for CTTC ACAAAAACTTTCACATGTGAAAGTTTCTCATTGATAAACATCATTACATATGAATA SGI1del05traN.rev TCCTCCTTA TTTACTTGCGGCCCAAGCATCAGGGAAAAAGGAGTAACTAGTGTAGGCTGGAGCTG SGI1delS011.for CTTC TAAGCTCTTTAATTGCGCGTCTATATTCAGCTCGTTTGCTCATATGAATATCCTCC SGI1delS011.rev TTA GTAGGCTTTCGGGTGACACGAAACCTATTGGAGCAACAGTGTGTAGGCTGGAGCTG SGI1delS012.for CTTC CTAGAAATGCGGCATCACCGACGGAATAAATACTAAAATCCATATGAATATCCTCC SGI1delS012.rev TTA GGGTAGAATAAGCCTCGATATAGTCATGTGACTAAAAGGCGGAATAGGAACTTCAT 94DelXnoFRTcm.for TTA TCGTTAACTGCACATTCGGGATATTTCTCTATATTCGCGGGCCTACCTGTGACGGA 94DelXnoFRTcm.rev AGAT

a restriction sites are underline

84 CHAPITRE 3 CRUCIAL ROLE OF SALMONELLA GENOMIC ISLAND 1 (SGI1) MASTER ACTIVATOR IN PARASITISM OF INCC PLASMIDS

3.1 Introduction

3.1.1. Pertinence

Le transfert des plasmides conjugatifs IncC et de l’ilot génomique de résistance SGI1 est activé par le complexe transcriptionnel AcaCD encodé par les plasmides IncC. SGI1 porte deux gènes : sgaD et sgaC, qui encodent ensemble le complexe homologue SgaCD. Celui-ci a été montré capable d’activer les promoteurs AcaCD-dépendants présents sur SGI1, mais son rôle restait jusqu’à récemment nébuleux. À l’aide d’une approche intégrative combinant ChIP-exo, Cappable-seq et RNA-seq, nous avons montré que SgaCD reconnaissait le même motif qu’AcaCD. Une analyse exhaustive de l’activité de tous les promoteurs dépendants d’AcaCD a permis de confirmer qu’ils étaient également activables par SgaCD. Des essais de qPCR et de cytométrie en flux ont permis de montrer que SgaCD jouait un rôle essentiel dans l’activation de l’excision et de la réplication de SGI1, menant ultimement à une déstabilisation du plasmide IncC porté par la même cellule.

Cette étude confirme la robustesse d’une approche intégrative visant à déterminer de façon exhaustive l’empreinte d’un complexe activateur transcriptionnel, la liste des sites d’initiation de la transcription et l’impact sur les niveaux des transcrits. C’est seulement la troisième fois qu’une telle approche est utilisée avec succès chez un modèle procaryote (les deux fois précédentes avaient permis de caractériser les régulons d’AcaCD et SetCD). On montre cependant pour la première fois que deux complexes phylogénétiquement éloignés, encodés par des familles distinctes d’éléments génétiques mobiles, reconnaissent le même motif. En outre, nos données montrent qu’AcaCD et SgaCD ne sont pas redondants, puisque SgaCD permet de compenser parfaitement la perte d’AcaCD, tandis que l’inverse entraine une importante diminution de la fréquence de mobilisation de SGI1. Ce résultat a de graves ramifications puisque plusieurs plasmides IncC portant un opéron acaDC défectueux ont été rapportés. Puisque SGI1 peut transférer vers une souche possédant déjà un plasmide IncC, l’ilot peut redonner un second souffle à ces plasmides habituellement non transférables et causer leur mobilisation temporaire. En raison des ressemblances entre les régulons d’AcaCD et SgaCD, nous proposons également que les ilots de type SGI1 soient également mobilisables par des plasmides non apparentés encodant des complexes transcriptionnels phylogénétiquement proches. Enfin, nous proposons un mécanisme par lequel SGI1 serait capable de déstabiliser les plasmides IncC, lequel reposerait sur une activation par SgaCD d’un système de partition encodé par ces plasmides.

3.1.2. Apport des auteurs

L’approche expérimentale a été définie conjointement avec Vincent Burrus. J’ai adapté l’approche intégrative sous les conseils de Sébastien Rodrigue et Dominick Matteau. J’ai réalisé la quasi-totalité des expériences et analyses qui ont mené aux conclusions de cet article. Kévin Huguet a établi le protocole pour l’expérience de cytométrie en flux et réalisé une grande partie des constructions génétiques destinées à cette expérience. Nicolas Rivard et Nicolas Carraro ont construit plusieurs mutants et réalisé les essais préliminaires de conjugaison.

Le séquençage haut débit et les réactions de qPCR ont été réalisés par la Plateforme RNomique de l’Université de Sherbrooke, en particulier Elvy Lapointe et Mathieu Durand. J’ai réalisé toutes les analyses de bio-informatique, faisant parfois appel à Jean-François Lucier, en adaptant des pipelines établis par Dominick Matteau.

J’ai rédigé la première version du manuscrit, ai produit toutes les figures puis ai participé à l’édition du manuscrit avec Vincent Burrus. Le manuscrit a été critiqué par Kévin Huguet, Nicolas Rivard, Nicolas Carraro et Sébastien Rodrigue et relu par Alain Lavigueur.

86 3.1.3. Référence

Durand R, Huguet KT, Rivard N, Carraro N, Rodrigue S, Burrus V. (2020) Crucial role of Salmonella Genomic Island 1 (SGI1) master activator in parasitism of IncC plasmids. Nucleic Acids Research, en révision (NAR-02797-V-2020). Disponible sur bioRxiv : https://doi.org/10.1101/2020.08.27.269225

Mots clés : IncC, SGI1, transcription, conjugaison, réplication, incompatibilité

87 3.2 Article

3.2.1. Page titre

Crucial Role of Salmonella Genomic Island 1 Master Activator in Parasitism of IncC plasmids

Romain Durand, Kévin T. Huguet, Nicolas Rivard, Nicolas Carraro, Sébastien Rodrigue, Vincent Burrus.

88 3.2.2. Abstract

IncC conjugative plasmids and the multiple variants of Salmonella Genomic Island 1 (SGI1) are two functionally interacting families of mobile genetic elements commonly associated with multidrug resistance in . SGI1 and its siblings are specifically mobilised in trans by IncC conjugative plasmids. Conjugative transfer of IncC plasmids is activated by the plasmid-encoded master activator AcaCD. SGI1 carries five AcaCD-responsive promoters that drive the expression of genes involved in its excision, replication, and mobilisation. SGI1 encodes an AcaCD homologue, the transcriptional activator complex SgaCD (also known as

FlhDCSGI1) that seems to recognise and activate the same SGI1 promoters. Here, we investigated the relevance of SgaCD in SGI1’s lifecycle. Mating assays revealed the requirement for SgaCD and its IncC-encoded counterpart AcaCD in the mobilisation of SGI1. An integrative approach combining ChIP-exo, Cappable-seq, and RNA-seq confirmed that SgaCD activates each of the 18 AcaCD-responsive promoters driving the expression of the plasmid transfer functions. A comprehensive analysis of the activity of the complete set of AcaCD-responsive promoters in both SGI1 and IncC plasmid was performed through reporter assays. qPCR and flow cytometry assays revealed that SgaCD is essential for the excision and replication of SGI1, and the destabilisation of the helper IncC plasmid.

3.2.3. Introduction

Multidrug resistant bacteria are increasingly recognised as an economic burden and a global threat to public health (1). Their emergence is being fuelled by diverse mobile genetic elements including genomic islands and conjugative plasmids (2) that often carry many antibiotic resistance genes. A better understanding of the mechanisms promoting the dissemination of mobile genetic elements is thus urgently needed. Mobilisable genomic islands (MGIs), also referred to as Integrated Mobilisable Elements (IMEs), have recently been getting renewed attention as they are increasingly recognised as important contributors to the propagation of multidrug resistance (3–7). MGIs usually carry diverse cargos conferring antibiotic or heavy

89 metal resistance, bacteriocin synthesis or resistance to phage infection (7–9). Salmonella Genomic Island 1 (SGI1) is a 42.4-kb MGI that confers resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides and tetracycline (ACSSuT) (10). SGI1 variants, which often bear a class 1 integron with diverse sets of antibiotic resistance gene cassettes, are found integrated at the 3’ end of trmE (also known as mnmE or thdF) in the chromosome of several species of Gammaproteobacteria, including human pathogens such as Salmonella enterica serovars, Proteus mirabilis, Morganella morganii, Providencia stuartii or Klebsiella pneumoniae (11–15). SGI1 and its variants are specifically mobilised in trans by the IncC and the closely related IncA conjugative plasmids (16). IncC plasmids are large, broad-host-range, and globally distributed plasmids that contribute to the propagation of multidrug resistance genes. For instance, IncC plasmids are frequently associated with New Delhi metallo-β- lactamase genes (blaNDM) that confer resistance against most β-lactams including carbapenems (17). IncC plasmids have also been a key driver of antibiotic resistance dissemination in the seventh-pandemic Vibrio cholerae lineage in Africa (18). Despite a lower epidemiological success compared to IncC plasmids, IncA plasmids have been found responsible for the spread of the carbapenemase gene blaVIM-1 in several species of Enterobacterales in hospitalised patients in Italy (19).

Conjugative transfer of IncC plasmids is controlled by two loci, the acr1-acaDC-acr2 region, and the acaB gene located near traNC. acaDC encodes the heteromeric complex AcaCD distantly related to SetCD, the transcriptional activator of transfer genes of the integrative and conjugative elements (ICEs) of the SXT/R391 family (20, 21). AcaCD and SetCD are distantly related to FlhCD, the transcriptional activator of flagellar operons in Escherichia coli and

Salmonella (20, 22, 23). Expression of acaDC is driven from the promoter Pacr1 that is repressed by Acr1 and Acr2 (20, 24). Recently, Hancock et al. demonstrated that Pacr1 is activated by AcaB, a novel transcriptional activator that exhibits structural similarity to bacterial transcription factors from the ribbon–helix–helix (RHH) superfamily (25). Interestingly, acaB has been shown to be part of the AcaCD regulon, wherein AcaCD and AcaB promote mutual expression, generating a positive feedback loop that activates conjugation. AcaCD turns on a

90 total of 18 AcaCD-activatable promoters driving the expression of major operons involved in the formation of the mating pore and initiation of conjugative transfer, as well as genes of unknown function (20, 26). Furthermore, AcaCD also activates the expression of genes carried by distinct families of MGIs, including five operons in SGI1 (7, 26–28). The functions of the five AcaCD-activatable genes xis, rep, traNS, traHS and traGS of SGI1 have been characterised. xis encodes the recombination directionality factor that facilitates the excision of SGI1 from the chromosome catalysed by Int (29). rep encodes the replication initiator protein that initiates

SGI1 replication at the origin of replication (oriV) (30). traNS, traHS and traGS encode three type IV secretion system (T4SS) subunits that replace their respective counterparts TraNC,

TraHC and TraGC in the mating pore encoded by the helper IncC plasmid (31).

MGIs are usually thought of as relatively passive mobile genetic elements that await the arrival of a helper self-transmissible element (a conjugative plasmid or an ICE) to escape their quiescent state and ride along with their helper element. To transfer to a new host, MGIs first need to excise from the host’s chromosome, and next to be translocated through the mating pore encoded by their helper element. For both processes, IncC-mobilised MGIs take advantage of the AcaCD regulon (7, 20, 27). Surprisingly, SGI1 has regularly challenged the assumption of passivity of MGIs (3). SGI1 has notably been shown to actively reshape the mating pore encoded by IncC conjugative plasmids to enhance its own propagation (31). SGI1 also actively replicates and carries a functional toxin-antitoxin system (sgiAT) that enhances its stability when an IncC plasmid is concomitantly present in the host (30, 32). In addition, SGI1 has been reported to destabilise helper IncA and IncC plasmids (33), a mechanism only recently attributed to the activation of the replicative state of excised SGI1 (30). Remarkably, SGI1-K, a member of the

SGI1 family, is unable to destabilise IncC plasmids (33). SGI1-K lacks the 3’ half of traNS (S005) and two upstream genes, sgaD (S007) and sgaC (S006), that code for an AcaCD ortholog complex named SgaCD (also known as FlhDCSGI1) (28, 34, 35). SgaC and SgaD share 79 and 46% identity with AcaC and AcaD, respectively. Although SgaCD was reported to activate all five AcaCD-activatable promoters of SGI1 and to complement the acaDC deletion of the IncC plasmid R16a, its biological role in SGI1’s lifecycle remains unclear (28). Deletion of sgaDC

91 was reported to have no impact on the mobilisation of SGI1-C by the helper IncC plasmid R55 (27). SGI1-C is a variant that differs from SGI1 by its smaller integron conferring resistance to spectinomycin, streptomycin and sulfonamides (36). Furthermore, deletion of acaDC of the helper IncC plasmid R16a was reported to completely abolish both self-transfer and SGI1-C mobilisation (27, 28).

Nevertheless, the conservation of sgaDC in most members of the SGI1 family suggests an important role in their lifecycle rather than simple redundancy for activation of AcaCD- activatable promoters. In this study, we investigated the relevance of SgaCD. First, we showed using mating assays that SgaCD is important for SGI1 mobilisation by IncC plasmids. Using a combination of ChIP-exo, Cappable-seq and RNA-seq approaches, we compared transcriptional activation by SgaCD and AcaCD of AcaCD-responsive promoters in SGI1 and in a model IncC plasmid. A systematic assessment of promoter activities through β-galactosidase reporter assays allowed us to measure the differential response of these promoters to SgaCD and AcaCD complexes. Finally, we unveiled the crucial role of SgaCD in the excision and replication of SGI1, ultimately leading to the destabilisation of IncC conjugative plasmids.

3.2.4. Material and methods

3.2.4.1 Bacterial strains and media

Bacterial strains, plasmids and genomic islands used in this study are described in Table 3.1. Strains were routinely grown in lysogeny broth (LB) at 37°C in an orbital shaker/incubator and were preserved at −75°C in LB broth containing 20% (vol/vol) glycerol. Antibiotics were used at the following concentrations: ampicillin (Ap), 100 μg/ml; chloramphenicol (Cm), 20 μg/ml; kanamycin (Kn), 50 μg/ml or 10 μg/ml for single copy integrants of pOPlacZ; nalidixic acid (Nx), 40 μg/ml; spectinomycin (Sp), 50 μg/ml; tetracycline (Tc), 12 μg/ml; rifampicin (Rf), 50 μg/ml. To induce expression from pBAD30 and pAH56, LB medium was supplemented with

92 0.2% L-arabinose or 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), respectively. Conjugation assays were performed as described previously (31).

Table 3.1. Strains and elements used in this study.

Strains, plasmids or Relevant genotype or phenotypea Source or elements reference E. coli BW25113 F- Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), λ-, rph-1, Δ(rhaD- (39) rhaB)568, hsdR514 VB113 Nx-derivative of BW25113 (58,66) KH95 BW25113 rpoB526 (Rf) (30) CAG18439 MG1655 lacZU118 lacI42::Tn10 (Tc) (67) MG1655 Rf Rf-derivative of MG1655 (68) Plasmids pVCR94Sp SpR derivative of pVCR94 (Sp Su) (20) pVCR94Sp ΔacaDC ΔacaDC mutant of pVCR94Sp (20) pVCR94Sp2 Derivative of pVCR94Sp carrying 8 more genes recently found to be This study conserved pVCR94Sp2 ΔacaDC mutant of pVCR94Sp2 This study ΔacaDC GreenSp Sp pVCR94 pVCR94 traGcΩ(PBAD-mNeonGreen-FRT) (Su Sp) (30) pMS1 pSIM5 Δcat::gen; Thermo-inducible expression of λRed recombination (58) (Ts, Gn) pSIM6 Thermo-inducible expression of λRed recombination (Ts Ap) (69) pVI36 SpR PCR template for one-step chromosomal gene inactivation (68) pKD3 CmR PCR template for one-step chromosomal gene inactivation (39) pKD4 KnR PCR template for one-step chromosomal gene inactivation (39) pCP20 Thermo-inducible expression of Flp recombinase (Ts Ap Cm) (70) pBAD30 orip15A araC PBAD (Ap) (71) pBAD-acaDC pBAD30::acaDC (20) pBAD-sgaDC pBAD30::sgaDC This study pacaDC3XFLAG pAH56::acaDC3XFLAG (Kn) (20) psgaDC3XFLAG pAH56::sgaDC3XFLAG (Kn) This study pOPlacZ pAH56 lacZ (Kn) (20) pPromacr1 pOPlacZ Pacr1-lacZ (Kn) (20) pINT-TS oriR101; cI857; λpR-intλ (Ap Ts) (40) Genomic Islands SGI1Kn ΔIn104::aph mutant of SGI1 devoid of the integron In104 (Kn) (31) SGI1Kn Δxis Δxis mutant of SGI1Kn This study SGI1Kn ΔsgaD ΔsgaD mutant of SGI1Kn This study SGI1Kn ΔsgaC ΔsgaC mutant of SGI1Kn This study SGI1Kn ΔsgaDC ΔsgaDC mutant of SGI1Kn This study Red Cm SGI1 SGI1 S009Ω(PBAD-mCherry-FRT) (Cm) (30) SGI1Red ΔsgaDC ΔsgaDC::aph mutant of SGI1Red (Cm Kn) This study

93 Table 3.1. Strains and elements used in this study. (continued) a Ap, ampicillin; Cm, chloramphenicol; Gn, gentamycin; Kn, kanamycin; Rf, rifampicin; Sp, spectinomycin; Su, sulfamethoxazole; Tc, tetracycline; Ts, thermosensitive.

3.2.4.2 Molecular biology

Plasmid DNA was prepared using either the EZ-10 Spin Column Plasmid DNA Miniprep Kit (Bio Basic) or the QIAprep Spin Miniprep Kit (Qiagen), according to manufacturer's instructions. Genomic DNA was prepared using the QIamp DNA Mini Kit (Qiagen), according to manufacturer’s instructions. Restriction enzymes used in this study were purchased from New England Biolabs. Several DNA polymerases were used: Q5 (New England Biolabs), Taq (New England Biolabs) and Easy Taq (Civic Bioscience). PCR products were purified using either the EZ-10 Spin Column PCR Products Purification Kit (Bio Basic) or the QIAquick PCR Purification Kit (Qiagen), according to manufacturer’s instructions. E. coli was transformed by electroporation as described by Dower et al. (37) in a Bio-Rad GenePulser Xcell apparatus set at 25 μF, 200 Ω and 1.8 kV using 1-mm gap electroporation cuvettes. Sanger sequencing reactions were performed by the Plateforme de Séquençage et de Génotypage du Centre de Recherche du CHUL (Québec, QC, Canada).

3.2.4.3 Plasmids and strains constructions

Oligonucleotides used in this study are listed in Table S3.1. New sequencing data obtained from pVCR94 (NZ_CP033514.1) revealed that pVCR94Sp (38) was missing eight conserved IncC genes. Therefore a new derivative, pVCR94Sp2, was constructed using the one-step chromosomal gene inactivation technique with pMS1, primer pair pVCR94delY2.f/94DelXnoFRT.rev and pVI36 as the template (39). Deletion mutants of pVCR94Sp2, SGI1Kn and SGI1Red were constructed in a similar fashion. Deletion of acaDC in pVCR94Sp2 was obtained using pSIM6, primer pair 94DelacaD.for/94DelacaC.rev and pKD4 as

94 the template. Deletions of xis, sgaD, sgaC and sgaDC in SGI1Kn were obtained using pSIM6, primer pairs SGI1delxis.for/SGI1delxis.rev, SGI1delsgaD07.for/SGI1delsgaD07.rev, SGI1delsgaC06.for/SGI1delsgaC06.rev and SGI1delsgaD07.for/SGI1delsgaC06.rev, respectively, and pKD3 as the template. Deletion of sgaDC in SGI1Red was obtained using pMS1, primer pair SGI1delsgaD07.for/SGI1delsgaC06.rev and pKD4 as the template. When possible, the antibiotic resistance cassette was removed from the resulting construction by Flp- catalysed excision using pCP20. All deletions were verified by PCR and antibiotic resistance profiling. sgaDC was amplified using primer pair SGI1sgaDEcoRI.for/SGI1sgaCEcoRI.rev and genomic DNA of E. coli VB113 containing SGIKn as the template. The amplicon was then digested with EcoRI and cloned into EcoRI-digested pBAD30 using T4 DNA ligase (NEB), generating pBAD-sgaDC. psgaDC3XFLAG was derived from pacaDC3XFLAG. sgaDC was amplified using primer pair pAH56sgaCDinsF/pAH56sgaCDinsR and pacaDC3XFLAG was linearized using primer pair pAH56sgaCDvecF/pAH56sgaCDvecR. These four primers were designed using NEBuilder® Assembly Tool (NEB). psgaDC3XFLAG was obtained by ligating both amplicons using the Gibson Assembly® Cloning Kit (NEB), to replace acaDC with sgaDC. PCR fragments containing the promoter region upstream of vcrx012, vcrx035-36, vcrx059, vcrx068, traA, vcrx076, dsbC, traNC, acaB-vcrx087, vcrx098, vcrx114, vcrx128, vcrx140, S004 and S018 were amplified using the corresponding primer pairs listed in Table S3.1, and cloned into pOPlacZ using either PstI or PstI and XhoI to produce the corresponding pOPlacZ derivatives (Figure S3.1). These vectors were ultimately integrated in single copy into the chromosomal site attBλ of E. coli BW25113 using pINT-Ts (40).

3.2.4.4 ChIP-exo assays

LB medium supplemented with 50 μg/ml rifampicin, 10 μg/ml kanamycin, 50 μg/ml spectinomycin and, when necessary, 20 μg/ml chloramphenicol was inoculated with E. coli MG1655 Rf bearing psgaDC3XFLAG, pVCR94Sp2 ΔacaDC and SGI1Red ΔsgaDC. Induction of 3XFLAG sgaDC expression was done by adding 0.1 mM IPTG to cultures grown to an OD600 of 0.2, followed by a 1-h incubation at 37°C with shaking. 10 mL of culture was used for the ChIP-

95 exo experiment, which was carried out as described previously (20), except for a shorter blocking step of magnetic beads used for immunoprecipitation: Dynabeads™ Protein A (Invitrogen™) were blocked by washing three times for 5 minutes each, in PBS/BSA (5 mg/mL). Replicates are detailed in Table S3.2.

3.2.4.5 Cappable-seq and RNA-seq assays

RNA was extracted from 5 mL cultures prepared as detailed above. Cultures were centrifuged for 10 min at 3,700 g, and the pellet was thoroughly resuspended in 1 mL TRI Reagent® (Sigma- Aldrich). Total RNA was then extracted using the Direct-zol RNA MiniPrep Kit (Zymo Research) with the recommended DNase I treatment, according to manufacturer's instructions. Cappable-seq was carried out as described elsewhere (41), with the following modifications. Approximately 10 μg of clean RNA was used for each assay. The very first RNA clean-up step was performed using Agencourt® RNAClean® XP Beads (Beckman) to ensure maximum elimination of unincorporated DTB-GTP. The removal of 3’ phosphates from fragmented RNA was performed using the Thermo Scientific™ T4 Polynucleotide Kinase (Thermo Fisher Scientific) and its supplied ATP-free buffer. For RNA-seq samples, 800 ng of total RNA was fragmented in 5X RNA Fragmentation Buffer (200 mM Tris-Acetate, pH 8.1, 500 mM KOAc, 150 mM MgOAc) by incubating at 95°C for 7 minutes and quenching immediately on ice. RNA was then purified using the RNA Clean & Concentrator-5 Kit (Zymo Research), according to manufacturer’s instructions. The quality and concentration of RNA before and after fragmentation were evaluated using a 2100 Bioanalyzer instrument (Agilent Technologies). Replicates are detailed in Table S3.2.

3.2.4.6 Illumina sequencing library preparation

ChIP-exo libraries were prepared as described previously (20), except for the second strand synthesis step that was performed using the Bst X DNA polymerase (Enzymatics) with ThermoPol® Buffer (NEB). Cappable-seq and RNA-seq libraries were prepared using the

96 NEBNext® Small RNA Library Prep Set for Illumina® (NEB), according to manufacturer’s instructions, except NEBNext 5’ SR Adaptor for Illumina was replaced by previously described 5'-hybrid-A0 oligo (20). DNA molecules corresponding to the rRNA transcripts were depleted using the duplex-specific nuclease (Evrogen), as described elsewhere (42). All libraries were amplified and checked as previously described (20), then ultimately pooled. Illumina sequencing was performed on two different sequencing lines on a NextSeq® 500/550 High Output system at the Plateforme Rnomique de l'Université de Sherbrooke (Sherbrooke, QC, Canada).

3.2.4.7 Bioinformatic analyses

Reads were trimmed with Trimmomatic (43) to discard nucleotides with a quality score below 30 and reads with a length below 36 bp. Quality was assessed before and after using FastQC (44). Trimmed reads were aligned on the E. coli MG1655 genome (NC_000913), pVCR94Sp2 ΔacaDC and SGI1Red ΔsgaDC using Bowtie 2 (45). Alignment quality was assessed using SAMStat (46), and reads with a quality score below 10 were discarded using SAMtools view (47). Information on the number of reads before and after trimming, the quality of mapping and coverage for each sample, is detailed in Table S3.2. Reads were then compressed, sorted and indexed using SAMtools (47). ChIP-exo and Cappable-seq reads were chopped to their first nucleotide and density was calculated separately for each DNA strand using BEDTools genomecov (48). Density files were ultimately compressed to BigWig format and visualised on the UCSC Genome Browser.

The footprint profile for each transcription start site of interest was analysed using the Versatile Aggregate Profiler (VAP) (49). Three pairs of divergent AcaCD-dependent promoters (vcrx035- 036, vcrx059-traI and acaB-vcrx087) were excluded from the analysis to prevent potentially overlapping ChIP-exo signals on the positive and negative strands. Replicates of a given condition were pooled using SAMtools and reads were treated as described above. Densities for any condition and strand were normalised by the total signal obtained for a given reference

97 (pVCR94Sp2 ΔacaDC or SGI1Red ΔsgaDC). The start and end coordinates of each transcription start site of interest were used as reference points. The signal was reported for each base pair and represented as median and either 1st-9th deciles or full range without smoothing. RPKM values were calculated for each DNA strand separately, using a script adapted from EDGE-pro (50). Differential expression analysis was performed using DESeq2 (51). Signal was also calculated for each condition on 1-kb intervals using BEDTools multicov (48). Log-transformed values were used to represent the variability between replicates (Figure S3.2).

Search for SgaC homologues was carried out using Blastp (52) against the Genbank non- redundant protein sequence (nr) database restricted to the Enterobacteriaceae (taxid:543).

3.2.4.8 β-galactosidase assays

The assays were carried out as described previously, using o-nitrophenyl-β-D- galactopyranoside (ONPG) as substrate (53). Cultures were prepared with LB medium supplemented with 10 μg/ml kanamycin to select the strain and 50 μg/ml ampicillin to maintain pBAD-acaDC or pBAD-sgaDC. Induction of acaDC or sgaDC was done by adding 0.2% arabinose to a refreshed culture grown to an OD600 of 0.2, followed by a 2-h incubation at 37°C with shaking prior to cell sampling.

3.2.4.9 qPCR assays

Genomic DNA was obtained from 1 mL of cell cultures of E. coli VB113 bearing pVCR94Sp, SGI1Kn or their mutants, grown for 16 h in LB medium supplemented with 40 μg/ml nalidixic acid, 50 μg/ml spectinomycin and 50 μg/ml kanamycin. Genomic DNA purity and concentration were measured with an ND-1000 NanoDrop spectrophotometer (Thermo Fisher Scientific). attB (236 bp), S026 (234 bp) from SGI1Kn, repA (237 bp) from pVCR94Sp, as well as reference genes dnaB (235 bp), hicB (235 bp) and trmE (238 bp) from E. coli, were quantified using primer pairs qAttBFw/qAttBRv, qS026Fw/qS026Rv, qFwpVCR/qRvpVCR, qdnaBFw/qdnaBRv,

98 qhicBFw/qhicBRv and qthdFFw/qthdFRv, respectively (Table S3.1). Excision of SGI1Kn was calculated as the ratio of attB per chromosome, SGI1Kn and pVCR94Sp copy numbers were calculated as the ratios of S026 and repA per chromosome, respectively. Normalisation was done using all three reference genes (54). For clarity, the strain VB113 bearing both pVCR94Sp and SGI1Kn was considered to exhibit a 100% excision rate. VB113 bearing either pVCR94Sp or SGI1Kn alone, were presumed to contain only one copy of a given element. qPCR experiments were performed in biological triplicate at the Plateforme Rnomique de l'Université de Sherbrooke (Sherbrooke, QC, Canada).

3.2.4.10 Cohabitation assays

The assays were carried out as described previously (30). Culture samples were diluted 1:1,000 in 1 mL of PBS. Fluorescence intensity of mNeonGreen and mCherry in cells was monitored by flow cytometry analysis on a BD FACSJazz (BD Biosciences), and data were acquired with the BD FACS Sortware. mNeonGreen and mCherry were excited with 488 and 561 nm solid- state lasers, and their emission was detected using 513/17 and 610/20 nm emission filters, respectively. For each sample, fluorescence of 20,000 cells was captured, and the data was analysed using FCS Express 7 (De Novo Software).

3.2.4.11 Statistical analyses and figures

Prism 8 (GraphPad Software) was used to plot graphics and to carry out statistical analyses. All figures were prepared using Inkscape 1.0 (https://inkscape.org/) or BioRender (https://biorender.com).

99 3.2.5. Results

3.2.5.1 SGI1 rescues the transfer of an IncC plasmid lacking its master activator of transfer

Given the similarity between AcaCD and SgaCD, and the essentiality of AcaCD for IncC plasmid transfer activation, we first assessed the importance of sgaDC. Transfer rates of SGI1Red and the coresident IncC helper plasmid pVCR94Sp2 were assessed in mating assays using combinations of wildtype and ΔDC mutants of both elements. Deletion of acaDC had no impact on transfer of the helper plasmid, SGI1Red or cotransfer of both elements, confirming that SGI1 can complement the loss of acaDC (Figure 3.1A). In contrast, deletion of sgaDC reduced transfer of SGI1Red nearly 2,000-fold compared to the wildtype, whereas it had no impact on transfer of the helper plasmid or cotransfer of both elements, showing that sgaDC is important for SGI1 transfer and/or stability (Figure 3.1A). Deletion of both acaDC and sgaDC abolished transfer of both elements. Finally, overexpression of a single chromosomal copy of 3XFLAG sgaDC under the control of the IPTG-inducible Ptac promoter activated transfer of both elements beyond wildtype levels (Figure 3.1A). To confirm that this phenotype was exclusively 3XFLAG attributable to SgaCD, trans-complementation with chromosomal Ptac-sgaDC was also performed using donors lacking SGI1. While deletion of acaDC abolished pVCR94Sp2 transfer in this context, it was fully restored by sgaDC3XFLAG overexpression (Figure 3.1B).

100

Figure 3.1. Role of SgaCD in conjugative transfer of IncC plasmids and SGI1. Effect of sgaDC on conjugative transfer of pVCR94Sp2 (light grey bars) in the presence (A) or absence (B) of SGI1Red (dark grey bars). Conjugation assays were carried out using MG1655 Rf containing the indicated elements as donor strains and E. coli CAG18439 (Tc) as the recipient strain. Wild-type (WT) or derivative mutants of both elements are indicated under each graph. For clarity, acaDC is shortened DC in (B). Transfer frequencies are expressed as the number of transconjugants per RfR SpR CmR donor CFUs in (A), or RfR SpR donor CFUs in (B). The bars represent the mean and standard error of the mean obtained from a biological triplicate. ¤ indicates that the frequency of transfer was below the detection limit (<10−7). For each panel and each element, a one- way ANOVA with Dunnett's multiple comparison test were performed on the logarithm of the values to compare each bar to the WT control (first bar). Statistical significance is indicated as follows: ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant.

101 3.2.5.2 SgaCD activates all AcaCD-activatable promoters and generates the same binding footprint as AcaCD

The SgaCD regulon was characterised using a combination of ChIP-exo, Cappable-seq and RNA-seq approaches to identify the binding sites, transcriptional start sites (TSS) and quantify mRNA transcript levels, respectively. In these experiments, either sgaDC3XFLAG or 3XFLAG acaDC was expressed from Ptac instead of the native promoter to ensure homogenous expression in cell populations. Otherwise isogenic cells lacking sgaDC3XFLAG and acaDC3XFLAG were used as negative controls. We also used ΔDC mutants of pVCR94Sp2 and SGI1Red to prevent interference from the native unlabelled complexes. The ChIP-exo experiment revealed that SgaCD binds all known AcaCD-activatable promoters on pVCR94Sp2 and SGI1 (Figure 3.2). SgaCD notably activates the promoter of acaB whose translation product activates acaDC expression (25). No additional promoter was found to be bound by SgaCD, showing that AcaCD and SgaCD complexes have similar activities and recognise identical sequence motifs. ChIP- exo peaks usually paired with those of Cappable-seq on the positive DNA strand, except for the control condition, confirming that SgaCD binding correlates with transcription of each promoter (Figure 3.2). Transcriptomic data obtained from RNA-seq assays with pVCR94Sp2 confirmed that sgaDC expression leads to an overall increase of mRNA levels, especially for operons associated with conjugative transfer functions (Figure 3.2A). Likewise, expression of most SGI1 genes increased upon expression of sgaDC or acaDC (Figure 3.2B).

102

Figure 3.2. In-depth analysis of the AcaCD/SgaCD regulon. Results of ChIP-exo, Cappable-seq and RNA-seq experiments on E. coli MG1655 Rf carrying both pVCR94Sp2 ΔacaDC (A) and SGI1Red ΔsgaDC (B) with or without a single chromosomal copy of either psgaDC3XFLAG expressing the native SgaD subunit along with a C-terminal 3XFLAG-tagged SgaC subunit induced by IPTG, or pacaDC3XFLAG. The circular map of the plasmid was linearised after the repA gene. The location and orientation of ORFs are depicted by arrowed boxes, which are color-coded by function as indicated.

103 Figure 3.2. In-depth analysis of the AcaCD/SgaCD regulon. (continued) Green arrows indicate previously identified (A) or predicted (B) AcaCD- dependent promoters. Each track (one representative replicate per condition as detailed in Table S3.2) plots the number of mapped reads as a function of the position in each element. Read densities are displayed as black bars, or blue and orange bars for Cappable-seq densities on the positive and negative DNA strands (linear scale for ChIP-exo and Cappable-seq densities, log scale for RNA-seq densities). Pink dots at the summit of peaks indicate a signal beyond the represented y-axis maximal value.

To compare the binding footprint of both complexes, two aggregated profiles were built by compiling ChIP-exo and Cappable-seq signals of pVCR94Sp2 and SGI1 promoters. No major difference could be observed between the SgaCD and AcaCD profiles (Figure 3.3 upper panels). Both complexes bind to the same site located -64 to -37 bp upstream of the transcription start site (TSS) (20). The large protected sequence starts -42 bp upstream of the TSS, that is 4 bp downstream of the GCCCNDWWWGGGC palindromic motif and ends 22 bp after the TSS. This protection is compatible with the promoter-bound RNA polymerase holoenzyme complex, as previously reported for class II activation where the bound activator complex abuts the promoter -35 element (Figure 3.3 upper panels and Figure S3.3A-D) (20, 21, 55, 56). The sequence overlapping the binding motif, where two peaks are observed immediately after GCCC and GGGC, corresponds to the AcaCD/SgaCD footprint. The AcaCD profile displayed sharper boundaries than that of SgaCD with a stronger signal at these positions (Figure 3.3 upper panels and Figure S3.3A-D). Assuming comparable crosslink efficiencies, this observation suggests better recruitment and tighter binding of AcaCD compared to SgaCD. Remarkably, the Cappable-seq signal of SgaCD on SGI1 promoters was weaker than that of AcaCD. Nevertheless, these results show that SgaCD and AcaCD, though encoded by two unrelated mobile genetic elements, recognise and bind identical sequence motifs.

104

Figure 3.3. VAP profiles of AcaCD/SgaCD footprint. Upper panels: ChIP-exo and Cappable-seq profiles obtained with SgaCD (A) and AcaCD (B) were aggregated from 13 AcaCD-activatable IncC promoters and all 5 AcaCD-activatable SGI1 promoters. The normalised density of ChIP- exo reads mapping on the positive DNA strand (left border, in blue) and the negative strand (right border, in red), as well as the normalised density of Cappable-seq reads mapping on the positive strand (purple), are plotted as functions of the distance in nucleotides to the aggregated transcription start site (TSS). For all reads, only the first nucleotide was conserved, in order to better visualize the protection borders. Data are represented as median, first and ninth deciles. The binding site is indicated as a black line, with its GCCCNDWWWGGGC motif delimitated by black triangles.

105 Figure 3.3. VAP profiles of AcaCD/SgaCD footprint. (continued) The region protected by the transcriptional complex, delimitated by the last and first peaks on the ChIP-exo left and right borders, respectively, is depicted in light grey. The region protected by either SgaCD or AcaCD, starting at the binding site, is depicted in dark grey. Lower panels show the ChIP-exo signal (left border) for each promoter as a heatmap, where pink indicates a signal beyond 5,000.

Both complexes appeared to bind better to Pxis relatively to other AcaCD-activatable promoters

(Figure 3.3 lower panels). Comparison of individual profiles obtained for Pxis and PS004 promoters with AcaCD and SgaCD, revealed clear differences. AcaCD and SgaCD profiles at

Pxis were similar to the aggregated profiles (Figure S3.3E-F), supporting the hypothesis that SgaCD binds less efficiently than AcaCD, leading to a weaker transcription initiation, as suggested by the weaker Cappable-seq signal. In strong contrast, the individual profiles obtained with SGI1-borne PS004 promoter were unique (Figure S3.3G-H). Comparison of ChIP-exo signals revealed efficient binding of SgaCD but not AcaCD. Detection of a strong AcaCD ChIP- exo signal on the right border shortly after the binding site (at position -26) suggests premature detachment or poor recruitment of the transcriptional activator (Figure S3.3H). The extremely weak Cappable-seq signal supports the hypothesis of a faulty transcriptional initiation by AcaCD. Except for a stretch of five Gs located between the binding site and -10 element, no specific feature could be found that could explain the weak activation of PS004 by AcaCD (Figure S3.1). In contrast, the Cappable-seq signal obtained with SgaCD was higher and displayed 3 peaks, at the predicted TSS, and 5 and 8 bp after the TSS. In summary, SgaCD seems to activate

PS004 with much higher efficiency than its IncC-encoded homolog.

106 3.2.5.3 Overexpression of sgaDC significantly affects the transcriptome

A differential expression analysis was performed on transcriptomic data to identify up- and down-regulated genes following sgaDC overexpression. This analysis considers the different sequencing depths of each condition and provides statistical metrics. A vast majority of differentially expressed genes were found to be activated by SgaCD (Figure 3.4A). For instance, expression of the IncC plasmid genes traL or traK that encode predicted inner and outer membrane subunits of the conjugal T4SS, respectively, increased 500 to 1,000-fold (Figure 3.4A). In contrast, few plasmid genes were found to be downregulated upon sgaDC overexpression. The most significant reduction of expression was observed for vcrx119 and vcrx120, two genes of unknown function. Since no binding signal was detected upstream of these two genes, repression by SgaCD is likely indirect (Figure 3.2). Furthermore, expression of the mobilisation gene mobI, which encodes a key factor for initiation of conjugative transfer at oriT, remained unchanged, confirming its independence from AcaCD and SgaCD (Figure 3.4A) (20, 57, 58). Finally, none of the 8 genes of pVCR94Sp2 that are absent in pVCR94Sp were expressed under the control of an AcaCD-activatable promoter (Figures 3.2 and 3.4A).

916 chromosomal genes were differentially expressed upon sgaDC overexpression in the presence of pVCR94Sp2 ΔacaDC (Figure S3.4 and Table S3.3). Several of the most up-regulated genes have been shown to be involved in resistance to stress and membrane transport. Down- regulated genes are involved in cysteine and enterobactin synthesis, outer membrane transport and anaerobic respiration. As no SgaCD binding could be detected by ChIP-exo in the promoter region of the affected genes, SgaCD influence on chromosomal gene expression was likely indirect. Four additional differential expression analyses were conducted to evaluate the impact of SGI1 and to compare activation by SgaCD and AcaCD. Expression of mobI was repressed upon overexpression of sgaDC in the presence of SGI1Red ΔsgaDC (Figures 3.4B and 3.4D). Since expression of mobI is independent of SgaCD, such repression likely results from expression of an SGI1-encoded factor.

107

Figure 3.4. Differential expression analysis.

Volcano plot of IncC (A-D) and SGI1 (E-F) genes using the Log2 of fold- change and adjusted p-value. Conditions are described at the top of each panel. The vertical lines indicate a 2-fold change in expression, and the horizontal line demarcates statistical significance at p = 0.05.

108 Figure 3.4. Differential expression analysis. (continued) Differentially expressed genes are marked as blue dots. For clarity, IncC genes are labeled "001" instead of "vcrx001".

Few genes were found to be differentially expressed when overexpressing sgaDC compared to acaDC (Figures 3.4C and 3.4F). Therefore, a weaker transcriptional initiation as suggested by Cappable-seq does not necessarily lead to lower levels of transcripts in the artificial context of overexpressing these transcriptional activators. Finally, we confirmed that most SGI1 genes are up-regulated by SgaCD, including xis and rep, with S004 and traHS presenting the highest fold-change values (Figure 3.4E).

3.2.5.4 Activation of gene expression by SgaCD is comparable to AcaCD

Since transcriptomic data only show the outcome of potentially multiple regulatory processes, ChIP-exo and Cappable-seq results were confirmed using an expression reporter assay based on transcriptional lacZ fusions to each of the 23 AcaCD-activatable promoters inserted in single copy into the chromosome. β-galactosidase assays were carried out with both AcaCD and SgaCD to quantify the relative expression level with each activator complex (Figure 3.5 and Figure S3.1). While vcrx035, vcrx036 and vcrx087 were constitutively expressed, virtually all other promoters were directly activated by AcaCD and SgaCD (Figure 3.5A-C). Induction ratios of SgaCD were almost invariably lower than those of AcaCD, confirming previously reported results obtained with the 5 AcaCD-activatable promoters of SGI1 (Figure 3.5D) (28). However, the difference was not statistically significant for 15 out of 23 promoters. The gene of unknown function vcrx087 was expressed both constitutively, which can be explained by the presence of nearly canonical σ70 −35 and −10 elements in its promoter sequence, and under the control of AcaCD/SgaCD (Figure 3.5B and Figure S3.1). In contrast, although the promoters driving expression of vcrx035 and vcrx036 seemed constitutively ON, both lacked canonical −35 and −10 elements.

109

Figure 3.5. Systematic analysis of AcaCD/SgaCD-dependent promoters.

110 Figure 3.5. Systematic analysis of AcaCD/SgaCD-dependent promoters. (continued) Activity of all AcaCD-dependent promoters on IncC plasmids and SGI1 was monitored from single-copy, chromosomally integrated transcriptional lacZ fusions in E. coli BW25113. (A) Qualitative assay on LB medium supplemented with X-Gal. Promoters are identified by the first gene of the corresponding operon. AcaCD-independent promoters are labelled in black. (B, C) Miller units. (D) Induction ratios. β-galactosidase assays were carried out in LB medium supplemented with (grey bars in A, B) or without (white bars in A,

B) arabinose to express acaDC or sgaDC from PBAD on pBAD-acaDC (A) or pBAD-sgaDC (B). The bars represent the mean and standard error of the mean obtained from a biological triplicate. In panel D, the x axis crosses the y axis at y = 1, indicating no induction. For panels B-D, three independent two-way ANOVA with Sidak's multiple comparison test were performed on the logarithm of the values to compare the pair of bars for each promoter. Statistical significance is indicated as follows: ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant.

3.2.5.5 SgaCD is essential for SGI1 replication

To investigate the role of SgaCD in SGI1’s lifecycle, we conducted qPCR assays using different mutants of pVCR94Sp and SGI1Kn. As shown previously (20, 29), SGI1 excision was undetectable in the absence of the helper plasmid or when a Δxis mutant was used (Figure 3.6A). Surprisingly, deletion of acaDC had a rather limited, statistically non-significant impact on the excision rate of SGI1Kn. On the contrary, deletion of sgaDC reduced the excision rate nearly 2,900-fold compared to the wild-type level, supporting a major role of SgaCD compared to AcaCD in SGI1’s lifecycle. Deletion of both resulted in the abolition of excision, indicating that residual excision of SGI1Kn ΔsgaDC is triggered by the helper plasmid-encoded AcaCD. A comparable phenotype was observed when measuring the copy number of SGI1Kn (Figure 3.6B). Deletion of sgaDC completely abolished SGI1 replication, whereas deletion of acaDC reduced

111 it only 2.7-fold (Figure 3.6B). None of the deletions had any statistically significant impact on pVCR94Sp copy number (Figure 3.6C).

Figure 3.6. Effect of sgaDC and acaDC deletions on SGI1Kn dynamics and pVCR94Sp stability. All targets were amplified by qPCR alongside three chromosomal reference genes: trmE, hicB and dnaB. (A) SGI1Kn excision rate corresponds to the attB/chromosome ratio. (B) SGI1Kn copy number corresponds to the s026/chromosome ratio. (C) pVCR94Sp copy number corresponds to the repA/chromosome ratio. For each panel, all ratios were normalised using the control set to 1 and displayed in white. The bars represent the mean and standard error of the mean obtained from a biological triplicate. ¤ indicates that the excision rate was below the detection limit (< 0.001%). Statistical analyses were performed (on the logarithm of the values for panel A) using three independent one-way ANOVA with Tukey's multiple comparison test. "a" indicates the bars showed no significant difference between one another. For panels A and C, statistical significance indicates comparisons to the normalisation control. Statistical significance is indicated as follows: ****, P < 0.0001; ***, P < 0.001; *, P < 0.05; ns, not significant.

112 Together, these results account for the reduced mobilisation of SGI1Kn ΔsgaDC by pVCR94Sp (Figure 3.1A). Residual excision and replication of SGI1 in the presence of pVCR94Sp ΔacaDC but not in its absence suggest that an unidentified plasmid-encoded factor is able to trigger or derepress the expression of sgaDC.

3.2.5.6 SgaCD-activated replication of SGI1 destabilises the helper IncC plasmid

Incompatibility between SGI1 and a coresident helper IncC plasmid has been shown to be linked to the replicative state of SGI1 (30). To test whether sgaDC had any role to play in SGI1 replication and incompatibility, we used a previously designed flow cytometry assay aimed at assessing the evolution of a population of cells carrying red fluorescence-producing SGI1Red and green fluorescence-producing pVCR94GreenSp over time (30). Cells producing red fluorescence segregate into two populations. Low-intensity red-fluorescent cells carry a single copy of SGI1 usually integrated in the chromosome, whereas cells producing high-intensity red fluorescence contain excised replicating SGI1. pVCR94GreenSp alone remained relatively stable and was retained in 71% of cells at T48 (Figure 3.7A). Cells carrying SGI1Red or its ΔsgaDC mutant produced mostly low-intensity red fluorescence and remained steady (Figure 3.7B-C). Overexpression of acaDC was sufficient to promote SGI1 excision and replication in the absence of the helper IncC plasmid as 71% of the cells produced high-intensity red fluorescence at T0 (Figure 3.7D). However, SGI1Red ΔsgaDC was rapidly lost as most cells failed to produce any fluorescence at T24. Overexpression of sgaDC resulted in a more widespread activation of SG1 replication as more than 90% of cells were highly red fluorescent at T0 (Figure 3.7E). SGI1 loss was delayed with sgaDC compared to acaDC, with more than 6% of cells retaining replicating SGI1 at T48. Together, these results show that SGI1 excision and replication can occur in the absence of the helper IncC plasmid when provided with AcaCD or SgaCD, although the latter seemed to promote stronger activation of rep expression. However, after excision, replication of SGI1 was insufficient to ensure its inheritance in the cell population in the absence of the IncC plasmid, despite the presence of the sgiAT toxin-antitoxin system.

113

Figure 3.7. Effect of sgaDC deletion on incompatibility between SGI1 and IncC plasmids in the absence of selective pressure. Evolution of the percentage of E. coli KH95 cells bearing (A) pVCR94GreenSp (IncC) or (B) SGI1Red or (C-E) SGI1Red ΔsgaDC (SGI1) or the indicated combination of both elements (F-I), in the absence or presence of pBAD- sgaDC (D, H) or pBAD-acaDC (E, I), over 48 hours in the absence of antibiotics (except for ampicillin for maintenance of pBAD derivatives) as monitored by flow cytometry. Plots show the mean and standard error of the mean values obtained from a biological triplicate.

114 When SGI1Red was in the presence of pVCR94GreenSp, most cells produced high red fluorescence at T0, confirming SGI1 replication (Figure 3.7F). At T48, only 16% of the cells produced high red fluorescence, a striking decrease that correlated with the loss of pVCR94GreenSp. Conversely, low red fluorescence signal increased, consistent with chromosomally integrated SGI1Red. In contrast, when SGI1Red ΔsgaDC was used, less than 4% of cells exhibited a high red fluorescence signal at T0, showing reduced SGI1 replication (Figure 3.7G). In addition, green and low red fluorescence remained steady throughout the experiment confirming that SGI1Red remained chromosomally integrated and that pVCR94GreenSp persisted in the cells.

Remarkably and consistent with the qPCR assays (Figure 3.6B), the presence of acaDC on the helper plasmid failed to trigger SGI1 replication in this context, confirming the importance of sgaDC in SGI1’s lifecycle. Complementation of ΔsgaDC using pBAD-sgaDC restored SGI1Red replication and a strong incompatibility phenotype (Figure 3.7H). Remarkably, IncC plasmid loss correlated with decreased SGI1 replication and increased SGI1 integration, confirming that plasmid instability is caused by the rescue of sgaDC expression (Figure 3.7H). Finally, complementation using pBAD-acaDC triggered SGI1 excision but failed to restore SGI1 replication, resulting in progressive loss of both SGI1Red and pVCR94GreenSp (Figure 3.7I). Altogether, these data indicate that sgaDC, not acaDC, controls SGI1 excision and replication, and the subsequent destabilisation of the helper IncC plasmid.

3.2.5.7 A single nucleotide polymorphism (SNP) abolishes SgaCD activity of SGI1-C

Contrasting with our observations with SGI1, deletion of sgaDC was shown to have no impact on the mobilisation of SGI1-C, whereas deletion of acaDC completely abolished both plasmid self-transfer and SGI1-C mobilisation (27, 28). Hence, sgaDC of SGI1-C is unable to complement an acaDC null mutant of its helper plasmid, unlike sgaDC of SGI1. To identify the cause of this discrepancy, we aligned S008-sgaDC genes and upstream sequences of SGI1 and SGI1-C and found only two SNPs in SGI1-C located at positions 6,600 (C to A) and 6,655 bp (G to A) relative to SGI1. While the second SNP is a silent mutation in sgaC, the first SNP

115 changes the CTG codon to an ATG codon, resulting in a L139M substitution in SgaC C-terminal moiety immediately upstream of the predicted Zinc finger (Figure S3.5F). A Blastp search of SgaC homologues revealed that either L or M amino acid residues are found at position 139. However, the M residue at this position is extremely rare in SGI1 variants (3 out the first 100 hits), with the notable exception of SGI2, formerly known as SGI1-J, of S. enterica serovar Emek from the UK (1999) (59). Furthermore, distantly related C subunits of other conjugative elements also have an L residue at the corresponding position (Figure S3.5F), suggesting that the L139M substitution could be detrimental to their activity. FlhC, the most distant homologue, has a V residue at the corresponding position.

3.2.6. Discussion

SGI1 encodes the transcriptional activator complex SgaCD, whose biological relevance remained unclear until now (28). The similarity of SgaCD to other activator complexes, such as AcaCD encoded by IncC conjugative plasmids and SetCD encoded by SXT/R391 ICEs, suggested an important, yet perhaps redundant, role in SGI1 mobilisation (27, 34). Here we characterised the SgaCD regulon using ChIP-exo experiments, transcriptomic analyses, and β- galactosidase reporter assays. ChIP-exo experiments showed that SgaCD recognises and binds to the same sites as AcaCD both on the IncC plasmid and on SGI1, leading to activation of the transfer genes and operons (Figures 3.3 and 3.5). We also showed that despite the evolutionary distance, both complexes bind the same DNA motifs. However, sgaDC and acaDC are not exchangeable in their natural context, as deletion of either gene set leads to drastically different phenotypes. While acaDC is essential to activate IncC plasmid transfer in the absence of SGI1, it becomes dispensable in SGI1+ cells. In contrast, suppression of sgaDC of SGI1 in the presence of the IncC plasmid abolishes SGI1 replication and allows peaceful coexistence of both elements despite the presence of a fully functional copy of acaDC. Consistent with our results, SGI1-K that lacks sgaDC due to a 2,779-bp deletion extending from S008 to the 5’ half of traNS (S005) is compatible with IncC plasmids (33, 35). Furthermore, SGI1-C was reported to fail to complement an acaDC null mutant of its helper plasmid and the sgaDC deletion to have no

116 impact on SGI1-C mobilisation (27, 28). We showed here that a single rare mutation in sgaC of SGI1-C, also found in SGI2, could be responsible for this phenotype, likely rendering SgaCD unable to act as a transcriptional activator at its physiological expression level.

The ability of SGI1 to complement an acaDC null IncC plasmid could have important epidemiological consequences. Occurrence of naturally acaDC-defective IncC plasmids has previously been reported (20). Although probably unable to activate self-transfer, entry of SGI1 in the host could resuscitate such “zombie” plasmids that would transiently regain their capacity to transfer, and mediate mobilisation of SGI1 (Figure S3.6). This process would be facilitated by the ability of SGI1 to escape entry exclusion exerted by IncA and IncC plasmids that normally prevents or strongly reduces redundant transfer between cells that contain plasmids of the same entry exclusion group (31, 38).

The pathway allowing SGI1 to complement an acaDC null IncC plasmid is unclear, and could involve previously reported low-level, constitutive expression of sgaDC (27, 60). Rare spontaneous excision of SGI1 suggests that the promoter of sgaDC is either mostly repressed under normal conditions or drastically activated by the presence of an IncA or IncC plasmid (61). Low SgaCD level produced by integrated SGI1 could be unable to switch on xis expression, preventing SGI1 excision and replication in the absence of the helper plasmid. Likewise, low SgaCD level is probably insufficient to trigger expression of the entire IncC plasmid tra gene set and initiate transfer. However, since the rates of mobilisation of SGI1 and self-transfer of the ΔacaDC helper plasmid are comparable to wild-type (Figure 3.1A), we propose that an IncC plasmid-encoded factor, triggers a positive feed-back loop in response to low levels of SgaCD likely via activation of sgaDC expression. Alternatively, plasmid entry itself could eventually act as a trigger for activation of sgaDC expression (Figure 3.8), perhaps via activation of the SOS response by the invading single DNA strand during conjugation (62).

117

Figure 3.8. Model of regulation of IncC plasmids and SGI1 gene expression. Major genes and operons involved in conjugative transfer and regulation are depicted as color-coded arrowed boxes (or a star for IncC origin of transfer oriT) based on the function: green, purple or grey, transcriptional activation; red, transcriptional repression; blue, type IV secretion system; orange, relaxosome; black, site-specific recombination; yellow, replication. Promoters are depicted by angled arrows and color-coded based on the corresponding activator: grey, unknown; green, AcaCD/SgaCD-activatable; purple, AcaB- activatable. Activation is represented by faded arrows. Repression is represented by red blocked dashed arrows. A bidirectional grey arrow indicates potential interactions. The chromosome is depicted as a large knotted DNA structure. SGI1 is represented in its initial integrated form, with its attachment sites depicted as two black lines. We propose a model in 6 main steps: 1) activation of sgaDC expression, 2) activation of SGI1 excision, 3) activation of SGI1 replication, 4) activation of acaB, 5) activation of acaDC and 6) activation of all the AcaCD/SgaCD-dependent promoters.

118 Figure 3.8. Model of regulation of IncC plasmids and SGI1 gene expression. (continued) For clarity, the activation of only two IncC transfer operons is depicted. Timing

of the alleviation of repression by Acr1 and Acr2 is unknown. Created with BioRender.com.

Recent discovery of AcaB added an important piece to the regulatory switch that controls the “On/Off state” of conjugative transfer of IncA and IncC plasmids (25). Mutual activation of acaB and acaDC has been proposed to be the trigger or amplifier of the conjugative transfer “On state”. We showed here that, like AcaCD, SgaCD activates expression of AcaB (Figures

3.2 and 3.5), hence low level of SgaCD could initiate derepression of Pacr1 via AcaB activation. This initial gentle push would then be amplified by AcaCD, promoting excision and replication of SGI1, which in return would raise SgaCD levels. However, although AcaB activates acaDC expression, it probably does not activate sgaDC expression, since no AcaCD- or AcaB-binding site was found upstream of sgaDC (20, 25). Yet another IncC-encoded factor is likely at play. Consistent with this hypothesis, SGI1’s rep gene was shown to be slightly expressed in the presence of pVCR94Sp ΔacaDC, but not in its absence (30). Given the crucial role of sgaDC for SGI1 excision, replication, mobilisation and incompatibility with IncC plasmids (Figures 3.1A, 3.6A, 3.6B and 3.7G), this observation supports the existence of a plasmid-encoded activator of sgaDC expression. Hence, the promoter of sgaDC instead of Pxis as previously suggested (27) would act as the sensor for IncC plasmid entry. Identification of the putative factor promoting such a feedback loop is on-going.

By analogy with the flagellar activator FlhCD to which they are distantly related, the C subunit of AcaCD and SgaCD likely binds to DNA immediately upstream of the -35 element, whereas the D subunit would stabilise the transcriptional complex (63). In support of this hypothesis, the primary sequence of D subunits is more divergent than the one of C subunits (27). Furthermore, AcaC and SgaC also share predicted tertiary structures that are remarkably similar to that of FlhC, including the Zinc-finger domain and its four conserved cysteine residues (Figure S3.5). However, the region located in between the two inner cysteine residues differs greatly in FlhC

119 compared to AcaC / SgaC. The similarity of SgaC and AcaC in this region likely accounts for the cross recognition of the same set of binding sites and the difference with FlhCD binding site. Considering that SgaC and AcaC derived from a common ancestor, the apparent weaker promoter activation by SgaCD could have evolved during its domestication by SGI1-like elements to reduce the risk of futile excision in the absence of a helper plasmid, enhancing SGI1 stability. Weak promoter activation would be compensated by the ability of SGI1 to replicate. IncC plasmid maintains as a single-copy replicon per cell, whereas the replicative cycle of SGI1 in IncC+ cells generates over 7 copies per cell, likely boosting sgaDC expression (Figure 3.6B) (30). Several properties of the expression of the transcriptional activator genes remain to be characterised to fully understand the crosstalk between SGI1 and its helper plasmid, including the conditions of activation and strength of sgaDC and acaDC promoters, the half-life of the corresponding mRNA transcripts, their translation rates and the relative stability of each activator complex. Assuming that SgaCD is produced in a much larger quantity than AcaCD due to SGI1’s high copy number, with comparable turnover rates of both complexes, AcaCD’s role could become negligible and even dispensable. This hypothesis is supported not only by the full complementation of an acaDC null mutant by SGI1, but also by the significant increase of plasmid transfer upon sgaDC expression (Figure 3.1). We cannot rule out that chimerical activator complexes SgaC-AcaD or AcaC-SgaD also form and play a role in the regulation of gene activation on the helper plasmid and on SGI1. If such chimera exist, and if SgaCD and

AcaCD tend to form heterohexamers like E. coli FlhD4C2 (63), then a heterogeneous bestiary of activator complexes regulates gene expression when SGI1 and its helper plasmid occupy the same cell.

Incompatibility between SGI1 and IncC plasmids could have emerged as a defence mechanism deployed by SGI1 to prevent excessive activation of excision and replication, which ultimately results in SGI1 loss in the absence of selective pressure (Figure 3.7). Destabilisation of the plasmid would allow the island to revert to its quiescent state after transferring to a new host, ensuring its stability by residing integrated in the host’s chromosome. Although the exact mechanism lying underneath the destabilisation is still unknown, we recently suggested titration

120 of endogenous replication proteins (30). Alternatively, SGI1 could be interfering with the partitioning of the IncC plasmid in daughter cells. Being single-copy large plasmids (Figure 3.6C), IncC replicons are likely extremely susceptible to perturbation of their partition process. In fact, IncC plasmids encode two partitioning systems, a type I parABS that has been shown to be essential for plasmid maintenance and a putative type II parMRC-like partitioning system distantly related to srpRMC of SXT/R391 ICEs and encoded by vcrx151-152 (64, 65). We show here that expression of vcrx151-152 is activated by AcaCD and SgaCD. Remarkably, forced excision and replication of SGI1 via overexpression of acaDC or sgaDC in the absence of an IncC plasmid resulted in rapid elimination of SGI1 from the cell population (Figure 3.7D-E). In stark contrast, when the plasmid was present, SGI1 persisted and reintegrated in the chromosome while promoting IncC plasmid loss (Figure 3.7H). This observation hints at SGI1 taking control of the IncC partitioning system, perhaps to enhance its equal segregation into daughter cells during cell division. These lines of inquiry will need to be examined in detail to better understand the processes at stake.

Despite the divergence of the primary sequences of AcaC and SgaC, we showed that both transcriptional activator complexes bind to the same DNA motif. Other conjugative plasmids that confer multidrug resistance to marine-dwelling bacteria and have not yet been ascribed to an incompatibility group also encode AcaCD orthologues. The C-terminus of the C subunits encoded by these plasmids (e.g. AqaC or AsaC) diverge significantly from AcaC and SgaC. Nevertheless, their Zinc-finger domain is relatively well conserved (Figure 3.5F-G). Unsurprisingly, these plasmids contain AcaCD-like binding sites upstream genes involved in conjugative transfer (Table S3.4). Based on these observations, the breadth of IncC/SGI1-like interactions is likely broader than anticipated, with SGI1 and its variants being possibly activated and mobilized by such plasmids that do not belong to the IncA nor the IncC group (16). Furthermore, since AcaCD-responsive promoters have recently been found in MGIs integrated at trmE, yicC and dusA in the chromosome of several species of Gammaproteobacteria (9), our observations support a complex network of mobilisation events involving diverse families of MGIs and conjugative plasmids of multiple incompatibility groups

121 (IncA, IncC and untyped) encoding AcaCD-like transcriptional activators with identical DNA- binding motifs.

Interactions between members of the SGI1 family and their helper plasmids are undoubtedly complex. The use of naturally occurring variants such as SGI1-C, SGI1-F, SGI1-I, SGI1-K or SGI2, with diverse naturally occurring IncC and even IncA plasmids, while providing interesting hints and clues, also renders comparisons between studies challenging and could lead to erroneous conclusions. These shortcomings result from SNPs that may affect the expression of key effectors that are important for the regulation, replication, stability, and conjugative transfer of the two interacting partners. Considering the discrepancies between others’ results and ours, we urge the establishment of a reliable, robust system based on a selected subset of model SGI1 variants and helper plasmids to properly decipher the complex biology and interactions between SGI1 and IncC plasmids.

3.2.7. Availability

Complete data from aligned reads for ChIP-Exo, Cappable-Seq and RNA-Seq experiments can also be visualized using the UCSC genome browser at http://bioinfo.ccs.usherbrooke.ca/sgaCD.html

3.2.8. Accession numbers

Raw sequencing data were submitted to Genbank under Bioproject accession number PRJNA648047 with the following Biosample accession numbers: for ChIP-exo assays, from SAMN15617565 to SAMN15617572, respectively; for Cappable-Seq assays, from SAMN15617573 to SAMN15617580; for RNA-Seq assays, from SAMN15617581 to SAMN15617604.

122 3.2.9. Acknowledgements

We thank Calcul Quebec (www.calculquebec.ca) and Compute Canada (www.computecanada.ca) for access to bioinformatics resources and support, as well as Jean- François Lucier for his help with bioinformatic analyses. We also thank the Centre de calcul scientifique de l’Université de Sherbrooke for technical assistance and the Université de Sherbrooke RNomics Plateform for assistance with Illumina sequencing. We are grateful to Alain Lavigueur for his insightful comments on the manuscript and to Dominick Matteau for thoughtful discussions and advice.

3.2.10. References

1. CDC (2019) Antibiotic Resistance Threats in the United States, 2019. Atlanta, GA: U.S. Department of Health and Human Services, CDC.

2. MacLean,R.C. and Millan,A.S. (2019) The evolution of antibiotic resistance. Science, 365, 1082–1083.

3. Guédon,G., Libante,V., Coluzzi,C., Payot,S. and Leblond-Bourget,N. (2017) The Obscure World of Integrative and Mobilizable Elements, Highly Widespread Elements that Pirate Bacterial Conjugative Systems. Genes, 8, 337.

4. Bellanger,X., Payot,S., Leblond-Bourget,N. and Guédon,G. (2014) Conjugative and mobilizable genomic islands in bacteria: evolution and diversity. FEMS Microbiology Reviews, 38, 720–760.

5. Carraro,N., Rivard,N., Burrus,V. and Ceccarelli,D. (2017) Mobilizable genomic islands, different strategies for the dissemination of multidrug resistance and other adaptive traits. Mob Genet Elements, 7, 1–6.

6. Piña-Iturbe,A., Ulloa-Allendes,D., Pardo-Roa,C., Coronado-Arrázola,I., Salazar- Echegarai,F.J., Sclavi,B., González,P.A. and Bueno,S.M. (2018) Comparative and phylogenetic analysis of a novel family of Enterobacteriaceae-associated genomic islands that share a conserved excision/integration module. Sci Rep, 8, 10292.

7. Carraro,N., Rivard,N., Ceccarelli,D., Colwell,R.R. and Burrus,V. (2016) IncA/C conjugative plasmids mobilize a new family of multidrug resistance islands in clinical Vibrio cholerae non- O1/non-O139 isolates from Haiti. mBio, 7, pii: e00509-16.

123 8. Brouwer,M.S.M., Warburton,P.J., Roberts,A.P., Mullany,P. and Allan,E. (2011) Genetic organisation, mobility and predicted functions of genes on integrated, mobile genetic elements in sequenced strains of Clostridium difficile. PLoS ONE, 6, e23014.

9. Rivard,N., Colwell,R.R. and Burrus,V. (2020) Antibiotic Resistance in Vibrio cholerae: Mechanistic Insights from IncC Plasmid-Mediated Dissemination of a Novel Family of Genomic Islands Inserted at trmE. mSphere, 5, e00748-20.

10. Boyd,D., Peters,G.A., Cloeckaert,A., Boumedine,K.S., Chaslus-Dancla,E., Imberechts,H. and Mulvey,M.R. (2001) Complete nucleotide sequence of a 43-kilobase genomic island associated with the multidrug resistance region of Salmonella enterica serovar Typhimurium DT104 and its identification in phage type DT120 and serovar Agona. J. Bacteriol., 183, 5725– 5732.

11. Cummins,M.L., Hamidian,M. and Djordjevic,S.P. (2020) Salmonella Genomic Island 1 is Broadly Disseminated within Gammaproteobacteriaceae. Microorganisms, 8, 161.

12. Schultz,E., Barraud,O., Madec,J.-Y., Haenni,M., Cloeckaert,A., Ploy,M.-C. and Doublet,B. (2017) Multidrug Resistance Salmonella Genomic Island 1 in a Morganella morganii subsp. morganii Human Clinical Isolate from France. mSphere, 2, e00118-17.

13. Soliman,A.M., Shimamoto,T., Nariya,H. and Shimamoto,T. (2019) Emergence of Salmonella Genomic Island 1 Variant SGI1-W in a Clinical Isolate of Providencia stuartii from Egypt. Antimicrob. Agents Chemother., 63, e01793-18.

14. Siebor,E. and Neuwirth,C. (2013) Emergence of Salmonella genomic island 1 (SGI1) among Proteus mirabilis clinical isolates in Dijon, France. J. Antimicrob. Chemother., 68, 1750–1756.

15. Soliman,A.M., Ramadan,H., Ghazy,E., Yu,L., Hisatsune,J., Kayama,S., Sugai,M., Nariya,H., Shimamoto,T., Jackson,C.R., et al. (2020) Emergence of Salmonella genomic island 1 variant SGI1-C in a multidrug-resistant clinical isolate of Klebsiella pneumoniae ST485 from Egypt. Antimicrob. Agents Chemother., 64, e01055-20.

16. Douard,G., Praud,K., Cloeckaert,A. and Doublet,B. (2010) The Salmonella genomic island 1 is specifically mobilized in trans by the IncA/C multidrug resistance plasmid family. PLoS ONE, 5, e15302.

17. Wu,W., Feng,Y., Tang,G., Qiao,F., McNally,A. and Zong,Z. (2019) NDM Metallo-β- Lactamases and Their Bacterial Producers in Health Care Settings. Clin. Microbiol. Rev., 32, e00115-18.

18. Weill,F.-X., Domman,D., Njamkepo,E., Tarr,C., Rauzier,J., Fawal,N., Keddy,K.H., Salje,H., Moore,S., Mukhopadhyay,A.K., et al. (2017) Genomic history of the seventh pandemic of cholera in Africa. Science, 358, 785–789.

124 19. Arcari,G., Di Lella,F.M., Bibbolino,G., Mengoni,F., Beccaccioli,M., Antonelli,G., Faino,L. and Carattoli,A. (2020) A Multispecies Cluster of VIM-1 Carbapenemase-Producing Enterobacterales Linked by a Novel, Highly Conjugative, and Broad-Host-Range IncA Plasmid Forebodes the Reemergence of VIM-1. Antimicrob Agents Chemother, 64, e02435-19, /aac/64/4/AAC.02435-19.atom.

20. Carraro,N., Matteau,D., Luo,P., Rodrigue,S. and Burrus,V. (2014) The master activator of IncA/C conjugative plasmids stimulates genomic islands and multidrug resistance dissemination. PLoS Genet, 10, e1004714.

21. Poulin-Laprade,D., Matteau,D., Jacques,P.-É., Rodrigue,S. and Burrus,V. (2015) Transfer activation of SXT/R391 integrative and conjugative elements: unraveling the SetCD regulon. Nucleic Acids Res., 43, 2045–2056.

22. Liu,X. and Matsumura,P. (1994) The FlhD/FlhC complex, a transcriptional activator of the Escherichia coli flagellar class II operons. J. Bacteriol., 176, 7345–7351.

23. Beaber,J.W., Hochhut,B. and Waldor,M.K. (2002) Genomic and functional analyses of SXT, an integrating antibiotic resistance gene transfer element derived from Vibrio cholerae. J. Bacteriol., 184, 4259–4269.

24. Lang,K.S. and Johnson,T.J. (2016) Characterization of Acr2, an H-NS-like protein encoded on A/C2-type plasmids. Plasmid, 87–88, 17–27.

25. Hancock,S.J., Phan,M.-D., Luo,Z., Lo,A.W., Peters,K.M., Nhu,N.T.K., Forde,B.M., Whitfield,J., Yang,J., Strugnell,R.A., et al. (2020) Comprehensive analysis of IncC plasmid conjugation identifies a crucial role for the transcriptional regulator AcaB. Nat Microbiol, 10.1038/s41564-020-0775-0.

26. Carraro,N., Matteau,D., Burrus,V. and Rodrigue,S. (2015) Unraveling the regulatory network of IncA/C plasmid mobilization: When genomic islands hijack conjugative elements. Mob Genet Elements, 5, 1–5.

27. Kiss,J., Papp,P.P., Szabó,M., Farkas,T., Murányi,G., Szakállas,E. and Olasz,F. (2015) The master regulator of IncA/C plasmids is recognized by the Salmonella Genomic island SGI1 as a signal for excision and conjugal transfer. Nucleic Acids Res., 43, 8735–8745.

28. Murányi,G., Szabó,M., Olasz,F. and Kiss,J. (2016) Determination and Analysis of the Putative AcaCD-Responsive Promoters of Salmonella Genomic Island 1. PLoS ONE, 11, e0164561.

29. Doublet,B., Boyd,D., Mulvey,M.R. and Cloeckaert,A. (2005) The Salmonella genomic island 1 is an integrative mobilizable element. Mol. Microbiol., 55, 1911–1924.

125 30. Huguet,K.T., Rivard,N., Garneau,D., Palanee,J. and Burrus,V. (2020) Replication of the Salmonella Genomic Island 1 (SGI1) triggered by helper IncC conjugative plasmids promotes incompatibility and plasmid loss. PLoS Genet., 16, e1008965.

31. Carraro,N., Durand,R., Rivard,N., Anquetil,C., Barrette,C., Humbert,M. and Burrus,V. (2017) Salmonella genomic island 1 (SGI1) reshapes the mating apparatus of IncC conjugative plasmids to promote self-propagation. PLOS Genetics, 13, e1006705.

32. Huguet,K.T., Gonnet,M., Doublet,B. and Cloeckaert,A. (2016) A toxin antitoxin system promotes the maintenance of the IncA/C-mobilizable Salmonella Genomic Island 1. Sci Rep, 6, 32285.

33. Harmer,C.J., Hamidian,M., Ambrose,S.J. and Hall,R.M. (2016) Destabilization of IncA and IncC plasmids by SGI1 and SGI2 type Salmonella genomic islands. Plasmid, 87–88, 51–57.

34. Poulin-Laprade,D., Carraro,N. and Burrus,V. (2015) The extended regulatory networks of SXT/R391 integrative and conjugative elements and IncA/C conjugative plasmids. Front Microbiol, 6, 837.

35. Hamidian,M., Holt,K.E. and Hall,R.M. (2015) The complete sequence of Salmonella genomic island SGI1-K. J. Antimicrob. Chemother., 70, 305–306.

36. Boyd,D., Cloeckaert,A., Chaslus-Dancla,E. and Mulvey,M.R. (2002) Characterization of variant Salmonella genomic island 1 multidrug resistance regions from serovars Typhimurium DT104 and Agona. Antimicrob. Agents Chemother., 46, 1714–1722.

37. Dower,W.J., Miller,J.F. and Ragsdale,C.W. (1988) High efficiency transformation of E. coli by high voltage electroporation. Nucleic Acids Res., 16, 6127–6145.

38. Humbert,M., Huguet,K.T., Coulombe,F. and Burrus,V. (2019) Entry Exclusion of Conjugative Plasmids of the IncA, IncC, and Related Untyped Incompatibility Groups. J. Bacteriol., 201, e00731-18.

39. Datsenko,K.A. and Wanner,B.L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A, 97, 6640–6645.

40. Haldimann,A. and Wanner,B.L. (2001) Conditional-replication, integration, excision, and retrieval plasmid-host systems for gene structure-function studies of bacteria. J. Bacteriol., 183, 6384–6393.

41. Ettwiller,L., Buswell,J., Yigit,E. and Schildkraut,I. (2016) A novel enrichment strategy reveals unprecedented number of novel transcription start sites at single base resolution in a model prokaryote and the gut microbiome. BMC Genomics, 17, 199.

126 42. Christodoulou,D.C., Gorham,J.M., Herman,D.S. and Seidman,J.G. (2011) Construction of normalized RNA-seq libraries for next-generation sequencing using the crab duplex-specific nuclease. Curr Protoc Mol Biol, Chapter 4, Unit4.12.

43. Bolger,A.M., Lohse,M. and Usadel,B. (2014) Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics, 30, 2114–2120.

44. Andrews,S. (2010) FastQC: A Quality Control tool for High Throughput Sequence Data.

45. Langmead,B. and Salzberg,S.L. (2012) Fast gapped-read alignment with Bowtie 2. Nat Methods, 9, 357–359.

46. Lassmann,T., Hayashizaki,Y. and Daub,C.O. (2011) SAMStat: monitoring biases in next generation sequencing data. Bioinformatics, 27, 130–131.

47. Li,H., Handsaker,B., Wysoker,A., Fennell,T., Ruan,J., Homer,N., Marth,G., Abecasis,G. and Durbin,R. (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25, 2078–2079.

48. Quinlan,A.R. and Hall,I.M. (2010) BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics, 26, 841–842.

49. Coulombe,C., Poitras,C., Nordell-Markovits,A., Brunelle,M., Lavoie,M.-A., Robert,F. and Jacques,P.-É. (2014) VAP: a versatile aggregate profiler for efficient genome-wide data representation and discovery. Nucleic Acids Res., 42, W485-493.

50. Magoc,T., Wood,D. and Salzberg,S.L. (2013) EDGE-pro: Estimated Degree of Gene Expression in Prokaryotic Genomes. Evol Bioinform Online, 9, 127–136.

51. Love,M.I., Huber,W. and Anders,S. (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15, 550.

52. Altschul,S.F., Gish,W., Miller,W., Myers,E.W. and Lipman,D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403–410.

53. Miller,J.H. (1992) A Short Course in Bacterial Genetics: A Laboratory Manual and Handbook for Escherichia coli and Related Bacteria Cold Spring Harbor Laboratory Press, Cold Spring Harbor.

54. Hellemans,J., Mortier,G., De Paepe,A., Speleman,F. and Vandesompele,J. (2007) qBase relative quantification framework and software for management and automated analysis of real- time quantitative PCR data. Genome Biology, 8, R19.

55. Browning,D.F. and Busby,S.J. (2004) The regulation of bacterial transcription initiation. Nat. Rev. Microbiol., 2, 57–65.

127 56. Latif,H., Federowicz,S., Ebrahim,A., Tarasova,J., Szubin,R., Utrilla,J., Zengler,K. and Palsson,B.O. (2018) ChIP-exo interrogation of Crp, DNA, and RNAP holoenzyme interactions. PLoS ONE, 13, e0197272.

57. Hegyi,A., Szabó,M., Olasz,F. and Kiss,J. (2017) Identification of oriT and a recombination hot spot in the IncA/C plasmid backbone. Sci Rep, 7, 10595.

58. Carraro,N., Sauvé,M., Matteau,D., Lauzon,G., Rodrigue,S. and Burrus,V. (2014) Development of pVCR94ΔX from Vibrio cholerae, a prototype for studying multidrug resistant IncA/C conjugative plasmids. Front Microbiol, 5, 44.

59. Hamidian,M., Holt,K.E. and Hall,R.M. (2015) The complete sequence of Salmonella genomic island SGI2. Journal of Antimicrobial Chemotherapy, 70, 617–619.

60. Golding,G.R., Olson,A.B., Doublet,B., Cloeckaert,A., Christianson,S., Graham,M.R. and Mulvey,M.R. (2007) The effect of the Salmonella genomic island 1 on in vitro global gene expression in Salmonella enterica serovar Typhimurium LT2. Microbes Infect., 9, 21–27.

61. Kiss,J., Nagy,B. and Olasz,F. (2012) Stability, entrapment and variant formation of Salmonella genomic island 1. PLoS ONE, 7, e32497.

62. Baharoglu,Z., Bikard,D. and Mazel,D. (2010) Conjugative DNA transfer induces the bacterial SOS response and promotes antibiotic resistance development through integron activation. PLoS Genet., 6, e1001165.

63. Wang,S., Fleming,R.T., Westbrook,E.M., Matsumura,P. and McKay,D.B. (2006) Structure of the Escherichia coli FlhDC complex, a prokaryotic heteromeric regulator of transcription. J. Mol. Biol., 355, 798–808.

64. Carraro,N., Poulin,D. and Burrus,V. (2015) Replication and Active Partition of Integrative and Conjugative Elements (ICEs) of the SXT/R391 Family: The Line between ICEs and Conjugative Plasmids Is Getting Thinner. PLoS Genet, 11, e1005298.

65. Hancock,S.J., Phan,M.-D., Peters,K.M., Forde,B.M., Chong,T.M., Yin,W.-F., Chan,K.-G., Paterson,D.L., Walsh,T.R., Beatson,S.A., et al. (2017) Identification of IncA/C Plasmid Replication and Maintenance Genes and Development of a Plasmid Multilocus Sequence Typing Scheme. Antimicrob. Agents Chemother., 61, e01740-16.

66. Grenier,F., Matteau,D., Baby,V. and Rodrigue,S. (2014) Complete genome sequence of Escherichia coli BW25113. Genome Announc, 2, pii: e01038-14.

67. Singer,M., Baker,T.A., Schnitzler,G., Deischel,S.M., Goel,M., Dove,W., Jaacks,K.J., Grossman,A.D., Erickson,J.W. and Gross,C.A. (1989) A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli. Microbiol Rev, 53, 1–24.

128 68. Ceccarelli,D., Daccord,A., René,M. and Burrus,V. (2008) Identification of the Origin of Transfer (oriT) and a New Gene Required for Mobilization of the SXT/R391 Family of Integrating Conjugative Elements. J. Bacteriol., 190, 5328–5338.

69. Datta,S., Costantino,N. and Court,D.L. (2006) A set of recombineering plasmids for gram- negative bacteria. Gene, 379, 109–115.

70. Cherepanov,P.P. and Wackernagel,W. (1995) Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene, 158, 9–14.

71. Guzman,L.M., Belin,D., Carson,M.J. and Beckwith,J. (1995) Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J. Bacteriol., 177, 4121–4130.

3.2.11. Supplementary data

Table S3.1. Oligonucleotides used in this study.

Primer Sequence Reference AACTAGGTCTATCAATAACTGGTGTGTAACGGGTAGGCT SGI1delxis.for This study Ggtgtaggctggagctgcttc CTCACTGCGGAAAGTCCGTATTGAATAAAACACTTTAAG SGI1delxis.rev This study AttattaCATATGAATATCCTCCTTA TATTAAAAAAGATTCAGCAATCTTTGGTGTGGAGATAATC SGI1delsgaC06.for This study gtgtaggctggagctgcttc GCTATTTGCCCTTTTGCGGCATACGCGGATGTATTCAttaCA SGI1delsgaC06.rev This study TATGAATATCCTCCTTA CAATCTGGGCCGCAGAAAAAAAGGTAAGGAGGACTACTG SGI1delsgaD07.for This study Agtgtaggctggagctgcttc TTACCCAGCGAGATAACGAACCAGATGTAAAACATTCAC SGI1delsgaD07.rev This study Tcatatgaatatcctcctta TTTAATATTTACAGCATCACATAATGATACTATGATGTTGc pVCR94delY2.f This study ggaataggaacttcaagaat TCGTTAACTGCACATTCGGGATATTTCTCTATATTCGCGGa 94DelXnoFRT.rev (1) gagcgcttttgaagctca GCTATCTATCGCAACCTTCGTGATTTGTGAGGGGGGCGGA 94DelacaD.for (2) gtgtaggctggagctgcttc GCGCTCATCTTCTGGTCCGAAATGTCATAGTCTACTCAttaC 94DelacaC.rev (2) ATATGAATATCCTCCTTA NNNNNNGAATTCaaggaggaataataaATGTATGCCTTAGAGCCG SGI1sgaDEcoRI.for This study TT SGI1sgaCEcoRI.rev NNNNNNGAATTCTCAGGCAGCTTTTGAGA This study pAH56sgaCDvecF aagctgccAAGCTTGACTACAAAGAC This study pAH56sgaCDvecR taaggcataCATATGTATATCTCCTTCTTACAAG This study pAH56sgaCDinsF tatacatatgTATGCCTTAGAGCCGTTAGAAC This study

129 Table S3.1. Oligonucleotides used in this study. (continued)

pAH56sgaCDinsR gtcaagcttGGCAGCTTTTGAGACCCG This study promvcrx012PstIF NNCTGCAGGTGGCGGAAAGTTTCACAT This study promvcrx012XhoIR NNNNCTCGAGTATTCACCTCCTATACCAACCAG This study promvcrx36PstIF NNCTGCAGAACAAAGCTCCTTAATTGCTTG This study promvcrx35PstIF NNCTGCAGACATATCACCTTTACAGTATCATT This study promvcrx060traIpstI.for NNNNNNCTGCAGCATCAAAAATTGTCGATGA (2) promvcrx060traIpstI.rev NNNNNNCTGCAGCTATCGTATTTCTCGTCGCTA (2) promvcrx059XhoIR NNNNCTCGAGCATCAAAAATTGTCGATGA This study promvcrx068PstIF NNCTGCAGCCTCCATAAAGTGCCCAAAATG This study promvcrx068XhoIR NNNNCTCGAGAGACCCTCCAAATAACTTGTCGT This study promtraLpstI.for NNNNNNCTGCAGCCGATCCAGTAAACGCAA (2) promtraLpstI.rev NNNNNNCTGCAGTTCTTTTCCTAGTGGCCTGTA (2) promtraVPstIF NNCTGCAGTGTTCTCCGACCAGATGTTG This study promtraAXhoIR NNNNCTCGAGCCGACAAAATAAGTAGCGCTG This study promvcrx076PstIF NNCTGCAGAAGACACCTTGCCGTT This study promvcrx076XhoIR NNNNCTCGAGCGCATTCTCCATTCGTTTTG This study promdsbCPstIF NNCTGCAGGGTTCATTCCTCTCAAAGC This study promdsbCXhoIR NNNNCTCGAGGTCTATCCCCTAAGATTGGATTA This study promtraNcPstIF NNCTGCAGGCAGACCCTAACGAATTTG This study promtraNcXhoIR NNNNCTCGAGTGTGTTCTCCATTTCAACCTT This study promvcrx087PstIF NNCTGCAGTAGAGATACTCCTAAGTTGAGGTA This study promvcrx086PstIF NNCTGCAGGGTGTTGCTCCTATAAGAA This study promvcrx098PstIF NNCTGCAGTGCGTCTTTGAGCTTTG This study promvcrx098XhoIR NNNNCTCGAGGTGTGCTCCTTAAAGTAACTTG This study promvcrx114PstIF NNCTGCAGAAGGCCCTCCTTCG This study promvcrx114XhoIR NNNNCTCGAGGCTTCTCTCCTTTGTTTGTG This study promvcrx128PstIFv2 NNCTGCAGGTAACCTGCTTCTCATTTGTTCTG This study promvcrx128XhoIR NNNNCTCGAGGAAGGAATGCGCGTCTATG This study promvcrx140PstIF NNCTGCAGCAGGATCACCTGGAGCAAG This study promvcrx140XhoIR NNNNCTCGAGGAGCTGCTCTTTCAACGATG This study promtraFpstI.for NNNNNNCTGCAGGTTTGGTCAATTCTTTCACAT (2) promtraFpstI.rev NNNNNNCTGCAGCCATGCTCGATATGGTAGA (2) promacaQpstI.for NNNNNNCTGCAGTACTCTTTACCTCCAGTTTACCA (2) promacaQpstI.rev NNNNNNCTGCAGTTAAACTGCGTTGTTAGCCAT (2) promvcrx152/001pstI.for NNNNNNCTGCAGCAAAGATTGCTTTTAGATTGCTT (2) promvcrx152/001pstI.re NNNNNNCTGCAGCAATTCCTATCCAATTCCTA (2) v SGI1promxisPstI.for NNNNNNCTGCAGGTTATTGATAGACCTAGTTTAT (3) SGI1promxisPstI.rev NNNNNNCTGCAGGGCCAATGTGCCGGTTT (3) SGI1promS004PstI.for NNNNNNCTGCAGGGCCTCTGGAAGTGCCCTAT This study NNNNCTCGAGTAAATTTCTCCAGCATCATCATTGATTTAC proms004XhoIR This study TG SGI1promtraNPstI.for NNNNNNCTGCAGCCAGCTTTTTAGTTTGGATA (3) SGI1promtraNPstI.rev NNNNNNCTGCAGGATAGAGCATTGCGAGCAAT (3) SGI1promtraHPstI.for NNNNNNCTGCAGTAGGTTTCGTGTCACCCGAA (3) SGI1promtraHPstI.rev NNNNNNCTGCAGTTAAAGCTCCTCTTTTAGAA (3) SGI1proms018PstI.for NNNNNNCTGCAGAATTAGTTGGAATTGAAGGGGTT This study SGI1proms018PstI.rev NNNNNNCTGCAGGGCTCTGCTGATTAAAAGCCAT This study

130 Table S3.1. Oligonucleotides used in this study. (continued)

qAttBFw AACATCTACAACAGGGCAAAG (4) qAttBRv GAGGAATAACAGGAGTGGTAAC (4) qS026Fw TGTCATCAGAAAGAACAAGCTC (4) qS026Rv GCGTTTTTATTCTGTTGCCC (4) qFwpVCR AAGAGAACCAAAGACAAAGACC (4) qRvpVCR CACCTTCACCGTGAAATGC (4) qdnaBFw ACGATTTTTACACCCGCCCAC (4) qdnaBRv ATCATCTCACGGACAACGGCAC (4) qhicBFw GCTTATCCCTTTACCTTCGCC (4) qhicBRv TAACTCTTTGCCAAGCGCC (4) qthdFFw GATAATGACACTATCGTAGCCC (4) qthdFRv GCAGTTCCAGCACATCTTC (4)

Table S3.2. Bioinformatic analyses supplemental data.

Approx. Desired Obtained Coverage Reads Reads on coverage Reads on Coverage B T # of # of Trimmed on on Sample L MG1655 on pVCR94 on SGI1 # # reads reads reads pVCR94 SGI1 (Q10) MG1655 (Q10) (x) (M) (M) (x) (Q10) (x) ChIPexo_sgaDC_ 1 1 1 1 0,32 287634 72096 2 94018 114 na na pVCRdeltaacaDC_1 ChIPexo_sgaDC_ 2 1 1 1 0,38 322648 78768 3 121940 147 na na pVCRdeltaacaDC_2 ChIPexo_sgaDC_ pVCRdeltaacaDC_ 1 1 1 1 0,30 255144 47288 2 68166 82 57771 276 SGI1deltasgaDC_1 ChIPexo_sgaDC_ pVCRdeltaacaDC_ 1 2 1 1 0,34 308145 76561 2 82964 100 69895 334 SGI1deltasgaDC_2 ChIPexo_sgaDC_ pVCRdeltaacaDC_ 2 1 1 1 0,35 316230 60961 2 92783 112 79977 382 SGI1deltasgaDC_3 ChIPexo_acaDC_ pVCRdeltaacaDC_ 1 1 2 1 0,51 480195 48119 2 105826 128 146783 701 SGI1deltasgaDC_1 ChIPexo_acaDC_ pVCRdeltaacaDC_ 1 2 2 1 0,51 478999 45566 1 110312 133 170019 812 SGI1deltasgaDC_2 ChIPexo_acaDC_ pVCRdeltaacaDC_ 2 1 2 1 0,75 708642 78579 3 117709 142 196265 938 SGI1deltasgaDC_3 CapSeq_sgaDC_ 1 1 1 1 1,04 977102 750797 24 19660 24 na na pVCRdeltaacaDC CapSeq_pVCRdelta 1 1 1 1 2,39 2269504 2220355 72 5232 6 na na acaDC CapSeq_acaDC_ pVCRdeltaacaDC_ 1 1 2 1 19,70 789203 621680 20 4978 6 6483 31 SGI1deltasgaDC_1

131 Table S3.2. Bioinformatic analyses supplemental data. (continued)

CapSeq_acaDC_ pVCRdeltaacaDC_ 1 2 2 1 13,65 513887 375291 12 2420 3 7972 38 SGI1deltasgaDC_2 CapSeq_pVCRdelta acaDC_ 1 1 2 1 17,62 849366 81919 3 4859 6 11052 53 SGI1deltasgaDC_1 CapSeq_pVCRdelta acaDC_ 2 1 2 1 19,83 1052271 795003 26 2659 3 3152 15 SGI1deltasgaDC_2 CapSeq_sgaDC_ pVCRdeltaacaDC_ 1 1 2 1 6,19 922345 648794 21 13476 16 13514 65 SGI1deltasgaDC_2 CapSeq_sgaDC_ pVCRdeltaacaDC_ 2 1 2 1 7,07 2596089 1855165 60 44565 54 51553 246 SGI1deltasgaDC_3 RNASeq_pVCRdelta 1 1 1 5 9,87 9283646 5640354 182 73883 89 na na acaDC_2 RNASeq_pVCRdelta 2 1 1 5 9,06 8523516 6158345 199 73232 89 na na acaDC_3 RNASeq_pVCRdelta 1 2 1 5 8,89 8414979 5941814 192 102720 124 na na acaDC_4 RNASeq_pVCRdelta 2 2 1 5 9,76 9190911 6434217 208 64363 78 na na acaDC_5 RNASeq_induced_ sgaDC_ 1 1 1 5 10,11 9597871 8848617 286 150922 182 na na pVCRdeltaacaDC_1 RNASeq_induced_ sgaDC_ 1 2 1 5 9,42 8922661 8088735 261 181030 219 na na pVCRdeltaacaDC_2 RNASeq_induced_ sgaDC_ 2 1 1 5 8,98 8504670 4800404 155 222304 269 na Na pVCRdeltaacaDC_3 RNASeq_induced_ sgaDC_ 2 2 1 5 7,56 7150962 4350985 140 198722 240 na na pVCRdeltaacaDC_4 RNASeq_induced_ sgaDC_ 2 3 1 5 3,97 3741495 1933340 62 87977 106 na na pVCRdeltaacaDC_5 RNASeq_induced_ sgaDC_ 1 1 1 5 9,33 8802470 4970555 160 277279 335 447965 2141 pVCRdeltaacaDC_ SGI1deltasgaDC_1 RNASeq_induced_ sgaDC_ 2 1 1 5 8,23 7778553 5578376 180 331062 400 469022 2241 pVCRdeltaacaDC_ SGI1deltasgaDC_3 RNASeq_induced_ sgaDC_ 2 2 1 5 8,84 8349834 5202798 168 296246 358 554795 2651 pVCRdeltaacaDC_ SGI1deltasgaDC_4 RNASeq_induced_ sgaDC_ 2 3 1 5 8,38 7946552 5473657 177 413753 500 619336 2960 pVCRdeltaacaDC_ SGI1deltasgaDC_5

132 Table S3.2. Bioinformatic analyses supplemental data. (continued)

RNASeq_induced_ acaDC_ 1 1 2 5 8,47 8198741 5144563 166 365399 442 309008 1477 pVCRdeltaacaDC_ SGI1deltasgaDC_1 RNASeq_induced_ acaDC_ 1 2 2 5 8,96 8619107 5226629 169 399512 483 477398 2281 pVCRdeltaacaDC_ SGI1deltasgaDC_3 RNASeq_induced_ acaDC_ 2 1 2 5 8,87 8594800 5062466 163 338885 410 481110 2299 pVCRdeltaacaDC_ SGI1deltasgaDC_4 RNASeq_induced_ acaDC_ 2 2 2 5 9,35 9038305 5045461 163 321642 389 384855 1839 pVCRdeltaacaDC_ SGI1deltasgaDC_5 RNASeq_induced_ acaDC_ 2 3 2 5 11,37 10958195 5090333 164 316293 382 367010 1754 pVCRdeltaacaDC_ SGI1deltasgaDC_6 RNASeq_pVCRdelta acaDC_ 1 1 2 5 8,91 8631906 7449854 240 94677 114 129364 618 SGI1deltasgaDC_1 RNASeq_pVCRdelta acaDC_ 1 2 2 5 8,23 7972683 7046746 227 118205 143 120607 576 SGI1deltasgaDC_2 RNASeq_pVCRdelta acaDC_ 1 3 2 5 6,88 6682701 5759123 186 110485 134 196802 940 SGI1deltasgaDC_3 RNASeq_pVCRdelta acaDC_ 2 1 2 5 6,92 6723083 6128823 198 65507 79 106946 511 SGI1deltasgaDC_4 RNASeq_pVCRdelta acaDC_ 2 2 2 5 8,96 8687731 7982967 258 94389 114 98811 472 SGI1deltasgaDC_5 RNASeq_pVCRdelta acaDC_ 2 3 2 5 8,00 7762858 7093720 229 78430 95 118898 568 SGI1deltasgaDC_6

B; biological replicate, T; technical replicate, L; sequencing line. Samples framed in bold were chosen to represent the tracks displayed in Figure 3.2.

133 Table S3.3. Top 10 up- and down-regulated MG1655 genes upon sgaDC overexpression.

Log2(fold Adjusted GO terms Other notable Gene GO terms (biological process) -change) p-value (molecular function) mentions GO:0015473 - Deletion mutant shows GO:0055085 - transmembrane transport ydeT 6,9 2,95E-21 fimbrial usher porin increased biofilm GO:0009297 - pilus assembly activity formation (5) GO:0005515 - protein GO:0010350 - cellular response to binding magnesium starvation iraM 5,9 5,12E-09 GO:0043856 - anti- GO:0071468 - cellular response to acidic sigma factor pH antagonist activity GO:0005515 - protein YhjB belongs to the GO:0006355 - regulation of transcription, binding yhjB 5,1 5,91E-09 NarL family of response DNA-templated GO:0003677 - DNA regulators (6) binding GO:0008982 - protein-N(PI)- GO:0034219 - carbohydrate phosphohistidine- agaB 5,0 1,16E-05 transmembrane transport sugar phosphotransferase activity GO:0016787 - yaiS 4,9 2,71E-06 hydrolase activity Located within a remnant of an ETT2 ygeF 4,8 2,33E-06 (T3SS) pathogenicity island (7) GO:0016740 - cmtB 4,8 2,74E-06 GO:0008643 - carbohydrate transport transferase activity GO:0071468 - cellular response to acidic Overexpression pH increases fitness and ycgZ 4,7 3,58E-11 GO:0045892 - negative regulation of resistance to several transcription, DNA-templated antibiotics (8) GO:1990929 - sulfoquinovosidase activity GO:0005975 - carbohydrate metabolic yihQ 4,7 1,97E-05 GO:0016798 - process hydrolase activity, acting on glycosyl bonds GO:0003700 - DNA- YdeO plays an binding transcription important role in GO:0045893 - positive regulation of ydeO 4,7 9,69E-07 factor activity survival under both transcription, DNA-templated acidic and anaerobic conditions (9) GO:0050311 - sulfite reductase (ferredoxin) GO:0000103 - sulfate assimilation activity cysI -5,1 2,60E-10 GO:0019344 - cysteine biosynthetic GO:0004783 - sulfite process reductase (NADPH) activity GO:0045900 - negative regulation of GO:0043022 - translational elongation ribosome binding raiA -4,7 4,70E-15 GO:0045947 - negative regulation of translational initiation GO:0009409 - response to cold

134 Table S3.3. Top 10 up- and down-regulated MG1655 genes upon sgaDC overexpression. (continued) yibI -4,7 1,01E-06 GO:0008897 - holo- GO:0009239 - enterobactin biosynthetic entD -4,6 7,82E-06 [acyl-carrier-protein] process synthase activity GO:0008867 - galactarate GO:0046392 - galactarate catabolic garD -4,5 2,95E-21 dehydratase activity process GO:0016829 - lyase activity Encodes a porin associated with nmpC -4,5 4,57E-07 increased sensitivity to several colicins and phages (10) GO:0009055 - electron transfer narGHJ GO:0009061 - anaerobic respiration -4,3 7,31E-16 activity I GO:0042128 - nitrate assimilation GO:0008940 - nitrate reductase activity GO:0042279 - nitrite GO:0019645 - anaerobic electron reductase nrfA -4,3 1,39E-19 transport chain (cytochrome, GO:0042128 - nitrate assimilation ammonia-forming) activity GO:0008942 - nitrite GO:0009061 - anaerobic respiration nirD -4,3 3,29E-05 reductase [NAD(P)H] GO:0042128 - nitrate assimilation activity GO:0102025 - GO:0015709 - thiosulfate transport ATPase-coupled cysA -4,2 4,19E-05 GO:1902358 - sulfate transmembrane thiosulfate transport transmembrane transporter activity

Table S3.4. Predicted AcaCD/SgaCD-responsive promoters in other families of conjugative plasmids.

Gene/ Plasmid Accession # Locus Start Stop Strand Scoreb p-valueb q-valueb Matched sequenceb taga AAACTGCCCGGATTGGGC pAsa4c NZ_KT033470.1 - 12326 12353 - 27,0102 8,64E-10 4,69E-05 ACCAACACCG AAACTGCCCAAATTGGAC pAsa4c NZ_KT033470.1 traI 36563 36590 + 31,5000 3,28E-11 5,34E-06 AGTTACGCAG TTTGCGCCCAATTCGGGC pAsa4c NZ_KT033470.1 - 43729 43756 - 25,3061 2,38E-09 9,06E-05 AGTTACAGCG TAACTGCCCGAATTGGGC pAsa4c NZ_KT033470.1 traL 43734 43761 + 23,0816 7,83E-09 2,55E-04 GCAAAGCCGT ATTGTGACCAAAATGGGC pAsa4c NZ_KT033470.1 traV 46946 46973 + 30,5612 7,11E-11 7,72E-06 AGTTTCACAG

135 Table S3.4. Predicted AcaCD/SgaCD-responsive promoters in other families of conjugative plasmids. (continued)

GATGTGACCAAAATGGGC pAsa4c NZ_KT033470.1 dsbC 53669 53696 + 32,3980 1,48E-11 4,81E-06 CGTTACACAG HTG22_ AAGTTGCCCAAAATGGAC pAsa4c NZ_KT033470.1 66272 66299 + 21,9286 1,39E-08 4,11E-04 RS00315 GTCTACAGCG HTG22_ ATTTCACCCGAAAAGGGC pAsa4c NZ_KT033470.1 76581 76608 + 25,6122 2,00E-09 9,06E-05 RS00370 AGCTTGAGCG HTG22_ TTTGCGCCCCAAAAAGGC pAsa4c NZ_KT033470.1 89711 89738 + 27,1735 7,80E-10 4,69E-05 RS00460 AGTTACAGCG AATGCGCCCTTAATGGGC pAsa4c NZ_KT033470.1 traF 150655 150682 + 25,2143 2,50E-09 9,06E-05 GCCAACACCT TAAATGCCCATTTTGGGC pAsa4c NZ_KT033470.1 parM 161774 161801 - 27,8878 4,92E-10 4,01E-05 AGTTCCAGAG AAGCTGACCAGAAAGGGC pVPS91 NZ_KX957972.1 traV 10285 10312 + 22,816 8,95E-09 5,83E-04 AGTTTCACAG HTL46_ AAAACGCTCAAATTGGGC pVPS91 NZ_KX957972.1 11772 11799 + 29,622 1,46E-10 2,38E-05 RS00070 AGTTACACAG TAATTGACCAAACAGGGC pVPS91 NZ_KX957972.1 dsbC 17409 17436 + 22 1,34E-08 7,27E-04 AACTACAGAT GTGTTGTCCAAAATGGAC pVPS91 NZ_KX957972.1 traN 24534 24561 + 20,878 2,28E-08 9,29E-04 AGTTCCACAG HTL46_ CAAACGCCCTAAAAGGGC pVPS91 NZ_KX957972.1 58179 58206 + 20,939 2,22E-08 9,29E-04 RS00310 AAATAGACGT HTL46_ AAACTGCCCAAAATGGAC pVPS91 NZ_KX957972.1 79965 79992 + 24,959 2,89E-09 2,35E-04 RS00450 CATTTCACGA HTL46_ TATTTGCCCAAAATGGGC pVPS91 NZ_KX957972.1 80013 80040 + 18,99 5,34E-08 1,74E-03 RS00450 TGATGCAGAG TAATTGCCCTAAATGGGA pVPS91 NZ_KX957972.1 traF 83903 83930 + 20,286 3,00E-08 1,08E-03 GGTTTCAGAT HTL46_ TATGCGCCCGTATAGGGC pVPS91 NZ_KX957972.1 113346 113373 + 25,449 2,19E-09 2,35E-04 RS00640 AGTTAAGCAT HTL46_ ATTGCGCCCAAAAAGGGC pVPS91 NZ_KX957972.1 156001 156028 + 34,959 1,02E-12 3,33E-07 RS00920 AGTTACAGGT AAACTGCCCAAAATGGAC pVPS91 NZ_KX957972.1 traI 162964 162991 + 18,5 6,61E-08 1,96E-03 AGGAGGGGAG AhyD4_ TTTGCACCCCAAAAAGGC pAhD4-1 NZ_CP013966.1 38016 38043 - 24,776 3,20E-09 1,25E-04 RS23175 AGTTACAGCG AhyD4_ ATTGCACCCGAAAAGGGC pAhD4-1 NZ_CP013966.1 50883 50910 - 27,99 4,60E-10 2,87E-05 RS23255 AGTTTGAGCG AhyD4_ AAGTTGCCCAAAATGGAC pAhD4-1 NZ_CP013966.1 61021 61048 - 24,225 4,31E-09 1,49E-04 RS23310 TGTTACAGCG GATGTGACCAAATTGGGC pAhD4-1 NZ_CP013966.1 dsbC 73058 73085 - 30,449 7,77E-11 1,10E-05 CGTTACAGAG ATTGTGACCAAAAAGGGC pAhD4-1 NZ_CP013966.1 traV 79773 79800 - 33,755 3,90E-12 1,21E-06 AGTTACACAG TAACTGCCCTAAATGGGC pAhD4-1 NZ_CP013966.1 traL 82976 83003 - 23,806 5,38E-09 1,68E-04 GCAAAGCCGT TTTGCGCCCATTTAGGGC pAhD4-1 NZ_CP013966.1 - 82981 83008 + 29,541 1,55E-10 1,21E-05 AGTTACAGCG AAACTGCCCAAATTGGAC pAhD4-1 NZ_CP013966.1 traI 93232 93259 - 30,051 1,06E-10 1,10E-05 AGTTAGGCAG GAACTGCCCGGATTGGGC pAhD4-1 NZ_CP013966.1 - 120376 120403 + 25,959 1,63E-09 8,48E-05 ACCAACAGAG

136 Table S3.4. Predicted AcaCD/SgaCD-responsive promoters in other families of conjugative plasmids. (continued)

AATGCGCCCTTAATGGGC pAhD4-1 NZ_CP013966.1 traF 145826 145853 - 25,214 2,50E-09 1,11E-04 GCCAACACCT TTGGCGCCCATTAAGGGC pAhD4-1 NZ_CP013966.1 - 145831 145858 + 19,122 5,04E-08 1,43E-03 GCATTGAGCG PAQU1_ TATGCGCCCGTATAGGGC pAQU1 NC_016983.1 1486 1513 + 25,449 2,19E-09 2,23E-04 002 AGTTAAGCAT AAACTGCCCAAAATGGAC pAQU1 NC_016983.1 traI 48352 48379 + 18,5 6,61E-08 2,69E-03 AGGAGGGGAG AAGCTGACCAGAAAGGGC pAQU1 NC_016983.1 traV 58679 58706 + 22,816 8,95E-09 7,30E-04 AGTTTCACAG PAQU1_ AAAACGCTCAAATTGGGC pAQU1 NC_016983.1 60167 60194 + 29,622 1,46E-10 5,97E-05 074 AGTTACACAG TAATTGACCAAACAGGGC pAQU1 NC_016983.1 dsbC 65753 65780 + 22 1,34E-08 9,10E-04 AACTACAGAT GTGTTGTCCAAAATGGAC pAQU1 NC_016983.1 traN 72878 72905 + 20,878 2,28E-08 1,16E-03 AGTTCCACAG PAQU1_ CAAACGCCCTAAAAGGGC pAQU1 NC_016983.1 113823 113850 + 20,939 2,22E-08 1,16E-03 128 AAATAGACGT AAACTGCCCAAAATGGAC pAQU1 NC_016983.1 - 171835 171862 + 28,51 3,24E-10 6,60E-05 CATTACACGA PAQU1_ TATTTGCCCAAAATGGGC pAQU1 NC_016983.1 171883 171910 + 27,765 5,33E-10 7,25E-05 198 TGATACAGAG TAATTGCCCTAAATGGGA pAQU1 NC_016983.1 traF 175774 175801 + 20,286 3,00E-08 1,36E-03 GGTTTCAGAT

a Gene or locus tag upstream of which the predicted AcaCD box was detected. “-” indicates that no gene or locus tag was found immediately upstream on the same strand.

b Predictions of AcaCD-binding site were carried out by using FIMO (Find Individual Motif Occurrences) V5.1.1 (11) with the AcaCD MEME logo, as described elsewhere (2). Description of the score, q- and p-values is available at http://alternate.meme-suite.org//doc/fimo-output-format.html.

137

Figure S3.1. Content of pOPlacZ-derivative vectors used for β-galactosidase assays. For each vector, the cloned insert corresponds to the sequence between the two restriction sites (either PstI or XhoI) written in orange. AcaCD binding site is written in green. TSS is written in blue and was updated from Carraro et al, 2014 (2), using the Cappable-seq data obtained in this study. Predicted -10 and -35 motifs are indicated in purple and red, respectively. Promoters that displayed a constitutive activity are indicated in brown.

138

Figure S3.2. Correlation between ChIP-exo and RNA-seq replicates.

139

Figure S3.2. Correlation between ChIP-exo and RNA-seq replicates. (continued)

140 Figure S3.2. Correlation between ChIP-exo and RNA-seq replicates. (continued) ChIP-exo and RNA-seq signals for each replicate (both strands) are represented as a log-transformed function of the position in (A) pVCR94Sp2 ΔacaDC or (B) SGI1Red ΔsgaDC, where both elements have been split in 1-kb intervals (I1, I2, etc). For additional details, refer to legend of Fig 3.2. (C) RNA-seq signal obtained with pVCR94Sp2 ΔacaDC was used to compare the replicates for each indicated condition (detailed in Table S3.2). Pearson correlation r values calculated either from signal on 1 kb intervals (panels on the left) or RPKM values (panels on the right) are shown as heatmaps.

141

Figure S3.3. VAP aggregate profiles of AcaCD/SgaCD footprint.

142 Figure S3.3. VAP aggregate profiles of AcaCD/SgaCD footprint. (continued) ChIP-exo and Cappable-seq profiles obtained with SgaCD (A) and ChIP-exo and 5'-RACE profiles obtained with AcaCD (B) were aggregated from 13 AcaCD-activatable IncC promoters. ChIP-exo and Cappable-seq profiles obtained with SgaCD (C) and AcaCD (D) were aggregated from 5 SGI1 AcaCD-activatable promoters. Individual profiles are shown for xis (E-F) and S004 promoters (G-H) The normalized density of ChIP-exo reads mapping on the positive DNA strand (left border, in blue) and the negative strand (right border, in red), as well as the normalized density of Cappable-seq or 5'-RACE reads mapping on the positive strand (purple), are plotted as functions of the distance in nucleotides to the aggregated transcription start site (TSS). Data are represented as median and range. The binding site is indicated as a black line, with its GCCCNDWWWGGGC motif delimitated by black triangles. The region protected by the transcriptional complex, delimitated by the last and first peaks on the ChIP-exo left and right borders, respectively, is depicted in light grey. The region protected by either SgaCD or AcaCD, starting at the binding site, is depicted in dark grey.

143

Figure S3.4. Differential expression analysis.

Volcano plot of E. coli MG1655 genes using the Log2 of fold-change and adjusted p-value. Overexpression of sgaDC is compared to the control in presence of pVCR94Sp2 ΔacaDC. The vertical lines indicate a 2-fold change in expression used as the threshold, and the horizontal line defines statistical significance at p = 0.05. Differentially expressed genes are marked as blue dots.

144

Figure S3.5. Predicted structures of AcaC, SgaC, SetC, AsaC and AqaC.

145 Figure S3.5. Predicted structures of AcaC, SgaC, SetC, AsaC and AqaC. (continued) (A-E) Predicted structures of AcaC, SgaC, SetC, AsaC and AqaC. Query and FlhC structures are shown in cartoon and pink backbone, respectively. Structures were predicted by I-TASSER (12). PDB 2AVU:E (FlhC) was specified as template (without alignment) and was identified as having the closest structural similarity (i.e. the highest TM-score) to the query predicted model. Zn2+ (white sphere) was identified as the most probable ligand by COFACTOR and COACH (13, 14). Corresponding binding cysteines are displayed in blue sticks and labeled. (F) Alignment of the primary sequence of the zinc-finger domain of the AcaC orthologues, displaying an almost conserved leucine residue at position 139, as well as the four conserved cysteines involved in the predicted zinc-finger domain. (G) Phylogenetic analysis of AcaC orthologues inferred by using the Maximum Likelihood method and Whelan And Goldman model (15). Primary sequences were aligned with MUSCLE (16). The tree with the highest log likelihood is shown. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the JTT model, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0,9821)). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. There was a total of 201 positions in the final dataset. Evolutionary analyses were conducted in MEGA X (17). AsaC (ALL42451.1), AqaC (WP_014386839.1), AqaC (VPS91) (WP_014386839.1) and AhaC (WP_047235007.1) are encoded by pAsa4c from Aeromonas salmonicida subsp. Salmonicida (18), pAQU1 from Photobacterium damselae subsp. damselae (19), pVPS91 from Vibrio parahaemolyticus (20), and pAhD4-1 from Aeromonas hydrophila (21), respectively.

146

Figure S3.6. Interactions between SGI1 and IncC plasmids.

147 Figure S3.6. Interactions between SGI1 and IncC plasmids. (continued) Drawings illustrate the incapacity of an IncC plasmid to transfer to a cell already bearing an IncC plasmid, because of entry exclusion (top) and the

escape by SGI1 of such mechanism, allowed by the replacement of TraGC by

TraGS (middle). Following SGI1 transfer to a strain already containing an acaDC-null IncC plasmid, SGI1 is able to complement the transfer of said plasmid to another strain (bottom). Created with BioRender.com.

3.2.11.1 References

1. Carraro,N., Sauvé,M., Matteau,D., Lauzon,G., Rodrigue,S. and Burrus,V. (2014) Development of pVCR94ΔX from Vibrio cholerae, a prototype for studying multidrug resistant IncA/C conjugative plasmids. Front Microbiol, 5.

2. Carraro,N., Matteau,D., Luo,P., Rodrigue,S. and Burrus,V. (2014) The Master Activator of IncA/C Conjugative Plasmids Stimulates Genomic Islands and Multidrug Resistance Dissemination. PLoS Genet, 10, e1004714.

3. Carraro,N., Durand,R., Rivard,N., Anquetil,C., Barrette,C., Humbert,M. and Burrus,V. (2017) Salmonella genomic island 1 (SGI1) reshapes the mating apparatus of IncC conjugative plasmids to promote self-propagation. PLOS Genetics, 13, e1006705.

4. Huguet,K.T., Rivard,N., Garneau,D., Palanee,J. and Burrus,V. (2020) Replication of the Salmonella Genomic Island 1 (SGI1) triggered by helper IncC conjugative plasmids promotes incompatibility and plasmid loss. PLOS Genetics, 16, e1008965.

5. Vlasblom,J., Zuberi,K., Rodriguez,H., Arnold,R., Gagarinova,A., Deineko,V., Kumar,A., Leung,E., Rizzolo,K., Samanfar,B., et al. (2015) Novel function discovery with GeneMANIA: a new integrated resource for gene function prediction in Escherichia coli. Bioinformatics, 31, 306–310.

6. Yamamoto,K., Hirao,K., Oshima,T., Aiba,H., Utsumi,R. and Ishihama,A. (2005) Functional Characterization in Vitro of All Two-component Signal Transduction Systems from Escherichia coli. J. Biol. Chem., 280, 1448–1456.

7. Ren,C.-P., Chaudhuri,R.R., Fivian,A., Bailey,C.M., Antonio,M., Barnes,W.M. and Pallen,M.J. (2004) The ETT2 gene cluster, encoding a second type III secretion system from Escherichia coli, is present in the majority of strains but has undergone widespread mutational attrition. J. Bacteriol., 186, 3547–3560.

148 8. Soo,V.W.C., Hanson-Manful,P. and Patrick,W.M. (2011) Artificial gene amplification reveals an abundance of promiscuous resistance determinants in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A., 108, 1484–1489.

9. Yamanaka,Y., Oshima,T., Ishihama,A. and Yamamoto,K. (2014) Characterization of the YdeO regulon in Escherichia coli. PLoS ONE, 9, e111962.

10. Pugsley,A.P. and Schnaitman,C.A. (1978) Identification of three genes controlling production of new outer membrane pore proteins in Escherichia coli K-12. J. Bacteriol., 135, 1118–1129.

11. Grant,C.E., Bailey,T.L. and Noble,W.S. (2011) FIMO: scanning for occurrences of a given motif. Bioinformatics, 27, 1017–1018.

12. Yang,J., Yan,R., Roy,A., Xu,D., Poisson,J. and Zhang,Y. (2015) The I-TASSER Suite: protein structure and function prediction. Nature Methods, 12, 7–8.

13. Zhang,C., Freddolino,P.L. and Zhang,Y. (2017) COFACTOR: improved protein function prediction by combining structure, sequence and protein–protein interaction information. Nucleic Acids Res, 45, W291–W299.

14. Yang,J., Roy,A. and Zhang,Y. (2013) Protein–ligand binding site recognition using complementary binding-specific substructure comparison and sequence profile alignment. Bioinformatics, 29, 2588–2595.

15. Whelan,S. and Goldman,N. (2001) A General Empirical Model of Protein Evolution Derived from Multiple Protein Families Using a Maximum-Likelihood Approach. Mol Biol Evol, 18, 691–699.

16. Edgar,R.C. (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res, 32, 1792–1797.

17. Kumar,S., Stecher,G., Li,M., Knyaz,C. and Tamura,K. (2018) MEGA X: Molecular Evolutionary Genetics Analysis across Computing Platforms. Mol Biol Evol, 35, 1547–1549.

18. Tanaka,K.H., Vincent,A.T., Trudel,M.V., Paquet,V.E., Frenette,M. and Charette,S.J. (2016) The mosaic architecture of Aeromonas salmonicida subsp. salmonicida pAsa4 plasmid and its consequences on antibiotic resistance. PeerJ, 4, e2595.

19. Nonaka,L., Maruyama,F., Miyamoto,M., Miyakoshi,M., Kurokawa,K. and Masuda,M. (2012) Novel Conjugative Transferable Multiple Drug Resistance Plasmid pAQU1 from Photobacterium damselae subsp. damselae Isolated from Marine Aquaculture Environment. Microbes and Environments, 27, 263–272.

149 20. Humbert,M., Huguet,K.T., Coulombe,F. and Burrus,V. (2019) Entry Exclusion of Conjugative Plasmids of the IncA, IncC, and Related Untyped Incompatibility Groups. J Bacteriol, 201.

21. Zhu,L., Zheng,J.-S., Wang,W.-M. and Luo,Y. (2019) Complete Genome Sequence of Highly Virulent Aeromonas hydrophila Strain D4, Isolated from a Diseased Blunt-Snout Bream in China. Microbiol Resour Announc, 8

150 CHAPITRE 4 MATÉRIEL ET MÉTHODES

4.1 Souches bactériennes et milieux de culture

Les souches bactériennes et éléments génétiques utilisés dans cette étude sont détaillés dans le Tableau 4.1. Les souches ont été cultivées en routine dans du milieu LB à 37°C en incubateur ou dans un agitateur orbital et préservées à -75°C dans du milieu LB liquide contenant 20% (vol/vol) de glycérol. Les antibiotiques ont été utilisés aux concentrations suivantes : ampicilline (Ap), 100 μg/mL; chloramphénicol (Cm), 20 μg/mL; kanamycine (Kn), 50 μg/mL ou 10 μg/mL pour les souches contenant pOPlacZ intégré en simple copie; acide nalidixique (Nx), 40 μg/mL; streptomycine (Sm), 200 μg/mL; spectinomycine (Sp), 50 μg/mL; sulfaméthoxazole (Su), 160 μg/mL; tétracycline (Tc), 12 μg/mL; triméthoprime (Tm), 32 μg/mL. Lorsque nécessaire, les cultures ont été supplémentées avec 0,02 % de L-arabinose.

Tableau 4.1. Souches bactériennes et éléments génétiques utilisés dans cette étude

Souche ou élément Génotype ou phénotype pertinenta Référence E. coli BW25113 F- Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), λ-, (Datsenko and rph-1, Δ(rhaD-rhaB)568, hsdR514 Wanner, 2000) VB113 Dérivé NxR de BW25113 (Carraro et al., 2014b; Grenier et al., 2014) CAG18439 MG1655 lacZU118 lacI42::Tn10 (Tc) (Singer et al., 1989) VB112 Dérivé RfR de MG1655 (Ceccarelli et al., 2008) BTH101 F- cya-99, araD139, galE15, galK16, rpsL1 (Karimova et al., (SmR), hsdR2, mcrA1, mcrB1 2005) β2163 F- RP4-2-Tc::Mu ΔdapA::(erm-pir) (Kn Em) (Demarre et al., 2005) V. cholerae OYP6G08 dusA::MGIVchUSA3 Yann Boucher Plasmides pVCR94Su Dérivé SuR de pVCR94 (Su) (Carraro et al., 2014b) pVCR94Sp Dérivé SpR de pVCR94 (Sp Su) (Carraro et al., 2014a) Tableau 4.1. Souches bactériennes et éléments génétiques utilisés dans cette étude (suite)

pVCR94Sp ΔacaB Mutant ΔacaB de pVCR94Sp (Sp Su) Cette étude pVCR94Sp ΔacaC Mutant ΔacaC de pVCR94Sp (Sp Su) (Carraro et al., 2014a) pVCR94Sp ΔacaDC Mutant ΔacaDC de pVCR94Sp (Sp Su) (Carraro et al., 2014a) pVCR94Sp ΔacaB Double mutant ΔacaB ΔacaDC de pVCR94Sp Cette étude ΔacaDC (Sp Su) pVCR94Kn Δacr2 Mutant Δacr2 de pVCR94Kn (Kn Su) (Roy et al., 2020) pSU4628 CloDF13::TnAΔEcoRV (Ap) (Cabezón et al., 1997) pSIM6 Système de recombinaison λRed thermo- (Datta et al., 2006) inductible (Ap, Ts) pKD3 Matrice CmR pour la méthode de délétion de (Datsenko and gène chromosomique Wanner, 2000) pKD13 Matrice KnR pour la méthode de délétion de (Datsenko and gène chromosomique Wanner, 2000) pVI36 Matrice SpR pour la méthode de délétion de gène (Ceccarelli et al., chromosomique 2008) pCP20 Recombinase Flp thermo-inductible (Ap, Cm, (Cherepanov and Ts) Wackernagel, 1995) pBAD30 orip15A araC PBAD (Ap) (Guzman et al., 1995) pBAD30-acaB pBAD30::acaB (Ap) Cette étude pBAD30-acaDC pBAD30::acaDC (Ap) (Carraro et al., 2014a) pBAD30-sgaDC pBAD30 ::sgaDC (Ap) (Durand et al., 2020) pBAD30-mpsAB pBAD30::mpsAB (Ap) Cette étude pBAD-S063 pBAD-TOPO::S063 (Ap) (Wozniak et al., 2009) pBAD30-S010 pBAD30::S010 (Ap) Cette étude pBAD30-traI907-992 pBAD30::traI907-992 (Ap) Cette étude pBAD30-traI907- pBAD30-traI907-992 présentant une Cette étude C61I 992 substitution C61I (Ap) pACYC177-S010 pACYC177Δbla::PtraHs-S010 (Kn) Cette étude pPromacr1 pOPlacZ Pacr1-lacZ (Kn) (Carraro et al., 2014a) pINT-Ts oriR101; cI857; λpR-intλ (Ap, Ts) (Haldimann and Wanner, 2001) pUT18C Dérivé de pUC19 encodant le sous-domaine T18 (Karimova et al., (résidus 225 à 399 de CyaA de B. pertussis, Ap) 2005) pUT18C-acaB Vecteur encodant une chimère T18-protéine Cette étude encodée par acaB (Ap) pUT18CB8 Vecteur encodant une chimère T18-VirB8 (Sharifahmadian et (T4SS de Brucella suis, Ap) al., 2017) pUT18C-S010 Vecteur encodant une chimère T18-protéine Cette étude encodée par S010 (Ap) pKT25 Dérivé de pSU40 encodant le sous-domaine T25 (Karimova et al., (résidus 1 à 224 de CyaA de B. pertussis, Kn) 2005) pKT25-traGS Vecteur encodant une chimère T25-TraGS (Kn) Cette étude

152 Tableau 4.1. Souches bactériennes et éléments génétiques utilisés dans cette étude (suite)

pKT25B10 Vecteur encodant une chimère T25-VirB10 (Sharifahmadian et (T4SS de B. suis, Kn) al., 2017) pKT25-mobIC Vecteur encodant une chimère T25-MobIC (Kn) Cette étude pSW23T pSW23::oriTRP4; oriVR6Kγ (Cm) (Demarre et al., 2005) ICE SXT (Su Tm Sm Cm) (Waldor et al., 1996) SXTΔS063 Mutant ΔS063 de SXT (Su Tm Sm Cm) (Wozniak et al., 2009) Ilots génomiques SGI1Kn Dérivé KnR de SGI1 (Kn) (Carraro et al., 2017a) SGI1Kn Δxis Mutant Δxis de SGI1Kn (Kn) (Durand et al., 2020) SGI1Kn ΔS004 Mutant ΔS004 de SGI1Kn (Kn) Cette étude Kn Kn SGI1 ΔtraNS Mutant ΔtraNS de SGI1 (Kn) (Carraro et al., 2017a) SGI1Kn ΔsgaC Mutant ΔsgaC de SGI1Kn (Kn) (Durand et al., 2020) SGI1Kn ΔsgaD Mutant ΔsgaD de SGI1Kn (Kn) (Durand et al., 2020) SGI1Kn ΔsgaDC Mutant ΔsgaDC de SGI1Kn (Kn) (Durand et al., 2020) SGI1Kn ΔS010::cat Mutant ΔS010::cat de SGI1Kn (Kn Cm) Cette étude Kn Kn SGI1 ΔtraGS Mutant ΔtraGS de SGI1 (Kn) (Carraro et al., 2017a) Kn Kn SGI1 ΔtraHS Mutant ΔtraHS de SGI1 (Kn) (Carraro et al., 2017a) Kn Kn SGI1 ΔtraHGS Mutant ΔtraHGS de SGI1 (Kn) (Carraro et al., 2017a) SGI1Kn ΔmpsA Mutant ΔmpsA de SGI1Kn (Kn) Cette étude SGI1Kn ΔmpsB Mutant ΔmpsB de SGI1Kn (Kn) Cette étude MGIVflInd1 (Kn) (Daccord et al., 2010) MGIVchHai6Kn Mutant ΔIn36A1 ΔTn6310 de MGIVchHai6 (Rivard et al., 2020) (Kn) MGIVchUSA3 Cette étude MGIVchUSA3Cm MGIVchUSA3::pSW23T (Cm) Cette étude MGIVchUSA3Kn Dérivé KnR de MGIVchUSA3 (Kn) Cette étude MGIVchUSA3Kn Mutant ΔprtN de MGIVchUSA3Kn (Kn) Cette étude ΔprtN

aAp, ampicilline; Cm, chloramphénicol; Em, érythromycine Kn, kanamycine; Nx, acide nalidixique; Sm, streptomycine; Sp, spectinomycine; Su, sulfaméthoxazole; Tc, tétracycline; Tm, triméthoprim; Ts, thermosensible.

4.2 Essais de conjugaison

Les essais de conjugaison ont majoritairement été réalisés sur milieu LB solide en incubant 6h à 37°C, comme décrits précédemment (Carraro et al., 2017a). Sauf mention contraire, les

153 souches d’E. coli VB113 et CAG18439 ont été utilisées comme souches donneuse et réceptrice, respectivement. Les essais de mobilisation de MGIVchUSA3 ont été réalisés par Florence Deschênes sous ma supervision, sur milieu LB solide en incubant 2h à 37°C, avec E. coli CAG18439 et VB112 comme souches donneuse et réceptrice, respectivement.

4.3 Méthodes de biologie moléculaire

L’ADN plasmidique a été préparé à l’aide des kits EZ-10 Spin Column Plasmid DNA Minipreps (Bio Basic) ou QIAprep Spin Miniprep (Qiagen), selon les recommandations des fournisseurs. L’ADN génomique a été préparé à l’aide du kit QIamp DNA Mini (Qiagen) selon les recommandations du fournisseur. Les enzymes de restriction utilisées dans cette étude proviennent de New England Biolabs. Plusieurs polymérases d’ADN ont été utilisées : Q5 (New England Biolabs), Taq (New England Biolabs) et EasyTaq (Civic Bioscience). Les produits de PCR ont été purifiés à l’aide des kits EZ-10 Spin Column PCR Products Purification (Bio Basic) ou QIAquick PCR Purification (Qiagen), selon les recommandations des fournisseurs, ou par immobilisation réversible sur phase solide (SPRI) à l’aide de billes magnétiques HighPrep™ PCR Clean-Up (MJS BioLynx). E. coli a été transformé par électroporation, comme décrit par Dower et al., 1988, dans un appareil GenePulser Xcell (Bio-rad) réglé à 25 μF, 200 Ω et 1,8 kV en utilisant des cuvettes de largeur 1 mm. Les réactions de séquençage de type Sanger ont été réalisées à la Plateforme de Séquençage et de Génotypage du Centre de Recherche du CHUL (Québec, QC, Canada).

4.4 Constructions génétiques

Les amorces utilisées dans cette étude sont listées dans le Tableau A2.1.

La délétion du gène acaB sur pVCR94Sp et sur pVCR94Sp ΔacaDC a été obtenue en appliquant la méthode de délétion de gène chromosomique décrite par Datsenko et Wanner (2000), à l’aide de pSIM6, de la paire d’amorces 94-086_WF/94-086_WR, ainsi que du vecteur matrice pKD3.

154 Les délétions de S004, S010, mpsA et mpsB, ont été obtenues de façon similaire à l’aide de pKD3 et des paires d’amorces SGI1dels004.for/SGI1dels004.rev, SGI1del010.for/SGI1del010.rev, SGI1del020.for/SGI1del020.rev et SGI1del019.for/SGI1del019.rev, respectivement. Lorsque possible, la cassette de résistance a été retirée par excision grâce à pCP20. Toutes les délétions ont été vérifiées par PCR et par profil de résistance. Le gène acaB a été amplifié à l’aide de la paire d’amorces 94-086_CF/ 94-086_CR à partir d’ADN génomique de VB113 contenant pVCR94Su. L’amplicon a été digéré par EcoRI, puis cloné dans pBAD30 à l’aide de la T4 DNA ligase (New England Biolabs) pour générer pBAD30-acaB. De la même façon, l’opéron mpsAB a été amplifié à l’aide de la paire d’amorces mpsABEcoRIf/mpsABSalIr à partir d’ADN génomique de VB113 contenant SGI1Kn. L’amplicon a été digéré par EcoRI et SalI, puis cloné dans pBAD30 pour générer pBAD30-mpsAB. Le vecteur pPromacr1 a été intégré dans le site attBλ de VB113 à l’aide de pINT-Ts. La souche résultante a été utilisée dans un essai rapporteur qualitatif sur milieu LB solide, supplémenté avec 40 μg/ml du substrat 5-bromo-4-chloro-3- indolyl-β-D-galactopyranoside (X-gal) et, lorsque nécessaire, 0,02% d’arabinose.

Le gène S010 a été amplifié à l’aide de la paire d’amorces SGI1S010EcoRI.for/SGI1S010EcoRI.rev à partir d’ADN génomique de VB113 contenant SGI1Kn. L’amplicon a été digéré par EcoRI, puis cloné dans pBAD30 à l’aide de la T4 DNA ligase (New England Biolabs) pour générer pBAD30-S010. De la même façon, la région homologue à S010 sur traI (encodant du résidu 907 au résidu 992) a été amplifiée à l’aide de la paire d’amorces traI907-992_EcoRI.f/traI907-992_EcoRI.r à partir d’ADN génomique de VB113 contenant pVCR94Su. Les amorces ont été conçues pour insérer un site synthétique de liaison au ribosome (RBS) et une adénine pour commencer par un codon d’initiation en phase. L’amplicon a été digéré par EcoRI, puis cloné dans pBAD30 à l’aide de la T4 DNA ligase (New

England Biolabs) pour générer pBAD30-traI907-992. La construction obtenue a par la suite été modifiée par mutagenèse dirigée à l’aide du kit Q5 Mutagenesis (New England Biolabs) pour introduire une substitution d’une cytosine par une isoleucine à la position 61 (C61I). Pour ce faire, on a amplifié le vecteur à l’aide de la paire d’amorces TraISDMf/ TraISDMr et suivi le C61I reste du protocole du kit, générant ultimement pBAD30-traI907-992 .

155 Le gène S010 accompagné de son promoteur natif (PtraHs) a été amplifié à l’aide de la paire d’amorces SGI1promtraHPstI.rev/SGI1s010PstIrev à partir d’ADN génomique de VB113 Kn contenant SGI1 ΔtraHGS. L’amplicon a été digéré par PstI, puis cloné dans pACYC177Δbla à l’aide de la T4 DNA ligase (New England Biolabs) pour générer pACYC177-S010.

Les gènes acaB et S010 ont été amplifiés à l’aide des paires d’amorces Vcrx086XbaIf/94- 086_CR et SGI1S010XbaIf/SGI1S010EcoRI.rev, respectivement. Les amplicons ont été digérés par EcoRI et XbaI, puis clonés dans pUC18C à l’aide de la T4 DNA ligase (New England Biolabs) pour générer pUC18C-acaB et pUC18C-S010, respectivement. De la même façon, les gènes traGS et mobIC ont été amplifiés à l’aide des paires d’amorces 94MobIXbaIf/94mobIEcoRI.rev et SGI1TraGBamHIf/SGI1TraGKpnIr, respectivement. Les amplicons ont été digérés avec EcoRI et XbaI ou BamHI et KpnI, puis clonés dans pKT25 à l’aide de la T4 DNA ligase (New England Biolabs) pour générer pKT25-traGS et pKT25-mobIC, respectivement.

Une région d’homologie a été amplifiée sur MGIVchUSA3 à l’aide de la paire d’amorces dusAigEcoRIF/dusAigEcoRIR. L’amplicon a été digéré par EcoRI, puis cloné dans le vecteur suicide mobilisable pSW23T à l’aide de la T4 DNA ligase (New England Biolabs). Le vecteur a ensuite été introduit dans la souche E. coli β2163 par électroporation. La souche obtenue a été utilisée comme donneuse dans un essai de conjugaison, avec V. cholerae OYP6G08 portant MGIVchUSA3 comme souche réceptrice. L’intégration en copie unique du vecteur suicide pSW23T a permis d’obtenir le variant MGIVchUSA3Cm. On a ensuite apporté pVCR94Kn Δacr2 par conjugaison dans la souche obtenue, puis mobilisé MGIVchUSA3Cm avec succès vers E. coli CAG18439. La réintégration de l’ilot dans le gène dusA d’E. coli a été confirmée en amplifiant les sites attL, attR, attB et attP à l’aide des paires d’amorces EcodusAattLf/dusAattLr, dusAattRf/EcodusAattRr, EcodusAattLf/EcodusAattRr et dusAattRf/ dusAattLr, respectivement. MGIVchUSA3Kn a été construit en substituant la totalité de la région altérée par un unique gène de résistance à la kanamycine, en appliquant la méthode de délétion de gène chromosomique décrite par Datsenko et Wanner (2000), à l’aide de pSIM6, de la paire

156 d’amorces dusAscarNoFRTf/dusAscarNoFRTr et du vecteur matrice pKD13. La délétion de prtN sur MGIVchUSA3Kn a été obtenue de façon similaire à l’aide de pSIM6, de la paire d’amorces DelprtNf/DelprtNr et du vecteur matrice pVI36. La cassette de résistance à la spectinomycine a été retirée comme décrit plus haut à l’aide de pCP20. Les manipulations décrites dans ce dernier paragraphe ont été réalisées par Florence Deschênes sous ma supervision.

4.5 qPCR

Les essais de qPCR ont été réalisés comme décrit précédemment (Durand et al., 2020). L’ADN génomique a été extrait à partir de 1 mL de cultures sur la nuit de VB113 contenant pVCR94Sp ou l’un de ses mutants, et/ou SGI1Kn ou l’un de ses mutants, dans du milieu LB liquide supplémenté avec de l’acide nalidixique, de la spectinomycine et/ou de la kanamycine. Le site attB, le gène S026 de SGI1, le gène repA de pVCR94 et les gènes de référence dnaB, hicB et trmE ont été amplifiés à l’aide des paires d’amorces : qAttBFw/qAttBRv, qS026Fw/qS026Rv, qFwpVCR/qRvpVCR, qdnaBFw/qdnaBRv, qhicBFw/qhicBRv et qthdFFw/qthdFRv, respectivement. Les séquences des amorces sont disponibles dans l’article. Le taux d’excision de SGI1Kn a été calculé en divisant le nombre de sites attB par chromosome. Les nombres de copies de SGI1Kn et de pVCR94Sp ont été calculés en faisant le ratio S026 ou repA par chromosome, respectivement. La normalisation a été réalisée à l’aide des trois gènes de référence (Hellemans et al., 2007). La souche VB113 contenant pVCR94Sp et SGI1Kn a été fixée comme contrôle présentant 100% d’excision, tandis que les souches portant soit pVCR94Sp, soit SGI1Kn, ont été présumées ne contenir qu’une seule copie de chaque élément par chromosome. Les essais de qPCR ont été réalisés en triplicata biologique à la Plateforme Rnomique de l’Université de Sherbrooke (Sherbrooke, QC, Canada).

157 4.6 Essai de double hybride bactérien

Des vecteurs encodant des chimères avec les protéines d’intérêt et les sous-domaines T18 et T25 du domaine catalytique de l’adénylate cyclase (CyaA) de B. pertussis ont été construits comme décrit ci-dessus. L’hétérodimérisation entre deux protéines hybrides entraine la complémentation fonctionnelle entre les fragments T18 et T25, rétablissant la production d’AMP cyclique (AMPc). L’AMPc peut ensuite lier la protéine activatrice des catabolites (CAP), formant un complexe capable d’activer l’expression de plusieurs gènes dont ceux de l’opéron lac. Pour chaque interaction testée, la souche E. coli BTH101, présentant une délétion du gène cya encodant l’adénylate cyclase, a été co-transformée avec un vecteur dérivé de pUC18C et un vecteur dérivé de pKT25. La souche a été incubée à 30°C sur milieu LB solide supplémenté en X-gal, ainsi qu’en ampicilline et kanamycine pour sélectionner les co- transformants. Pour chaque condition testée, on a ainsi pu confirmer que le phénotype de dégradation ou non du X-gal (illustré par des colonies bleues ou blanches, respectivement) était homogène pour toute la population. On a ensuite cultivé les souches obtenues dans du milieu LB liquide jusqu’à atteindre la phase stationnaire, dilué 10 fois et enfin déposé des gouttes de 10 μL sur milieu LB solide supplémenté en X-gal, ampicilline et kanamycine. Les pétris ont été à nouveau incubés sur la nuit à 30°C. La souche BTH101 et les vecteurs dérivés de pUT18C et pKT25 ont été gracieusement fournis par le laboratoire de Christian Baron, avec l’accord de Daniel Ladant.

4.7 Phylogénie

Les séquences en acides aminés des protéines homologues à celle encodée par acaB ont été obtenues à l’aide de l’outil Blastp contre la base de données nr (Altschul et al., 1990). Elles ont ensuite été alignées à l’aide de l’outil Muscle (Edgar, 2004), puis rognées à l’aide de l’outil trimal v1.2 en utilisant l’approche heuristique automatisée (Capella-Gutiérrez et al., 2009). L’analyse phylogénétique a été réalisée à l’aide du logiciel MEGA7 (Kumar et al., 2016). Les données ont été inférées à l’aide de la méthode du maximum de vraisemblance, basée sur le

158 modèle de Le et Gascuel, (2008). Les arbres initiaux pour la recherche heuristique ont été obtenus automatiquement en appliquant les algorithmes Neighbor-Join et BioNJ à une matrice de distances estimée à l’aide d’un modèle JTT, puis en sélectionnant la topologie avec la plus grande valeur de log de vraisemblance. Une distribution Gamma discrète a été utilisée pour modéliser les différences de taux d’évolution entre les sites (5 catégories (+G, paramètre = 0,5980)). Le modèle de variation du taux permettait que certains sites soient évolutivement invariables ([+I], 21,1268% des sites).

4.8 Génomique comparative

Les séquences de nouveaux ilots de type SGI1 ont été obtenues en parcourant la base de données Refseq à l’aide de l’outil Blastp (Altschul et al., 1990) et des séquences en acides aminés de

MpsA, TraGS, SgaC, TraNS (puis IntdusA et IntyicC après avoir identifié les premiers ilots). Les résultats ont été exportés puis classés par numéro d’accession pour être capable d’identifier les ilots qui présentaient l’ordre synténique attendu. Les ilots ont été délimités manuellement en cherchant les séquences répétées caractéristiques des sites d’attachement. Lorsque la séquence d’un ilot était répartie sur deux contigs différents, celle-ci a été assemblée manuellement. Les ilots ont ensuite été regroupés à l’aide de l’outil cd-hit-est, lorsqu’ils présentaient un pourcentage d’identité de séquences nucléotidiques supérieur à 95% (Huang et al., 2010). Certaines séquences ont dû être corrigées manuellement pour annoter des petits cadres de lecture manquants comme mpsB, ou pour homogénéiser les codons d’initiation.

159

160 CHAPITRE 5 RÉSULTATS

5.1 Rôle du gène acaB dans la mobilisation de SGI1

5.1.1. acaB est essentiel au transfert des plasmides IncC ou à la mobilisation de SGI1

Le gène acaB a récemment été montré encoder un activateur de l’expression d’acaDC essentiel au transfert des plasmides IncC (Hancock et al., 2020). Dans un premier temps, nous avons souhaité vérifier que ce résultat pouvait être reproduit en utilisant notre plasmide IncC modèle pVCR94Sp. Nous avons pu confirmer que la délétion d’acaB nuit fortement au transfert de pVCR94Sp (Figure 5.1). La surexpression du gène à partir d’un promoteur inductible à l’arabinose a permis de rétablir la fréquence de transfert du plasmide au-delà de celle du type sauvage. Ces données confirment le rôle d’AcaB dans l’amplification du signal d’activation du transfert des plasmides IncC (Hancock et al., 2020).

Figure 5.1. Rôle d’acaB dans le transfert conjugatif des plasmides IncC

161 Figure 5.1. Rôle d’acaB dans le transfert conjugatif des plasmides IncC (suite) L’essai de conjugaison a été réalisé en utilisant E. coli VB113 portant les éléments indiqués comme souche donneuse et E. coli CAG18439 comme souche réceptrice. Le gène acaB est abrégé « aB ». Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des donneuses, sélectionnées sur Nx, Sp (et Ap pour la condition avec pBAD30-acaB). Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique. Une ANOVA unidirectionnelle suivie d’un test posthoc de Dunnett réalisée sur les valeurs transformées en log, a permis de comparer chaque barre au contrôle de type sauvage. La signification statistique est comme suit : **** ; P < 0,0001.

Dans un second temps, nous avons voulu vérifier si la même délétion affectait également la mobilisation de l’ilot génomique SGI1. De façon intéressante, la délétion d’acaB s’est révélée abolir la mobilisation de SGI1Kn, mais la présence de l’ilot a permis de rétablir le transfert du plasmide (Figure 5.2).

162

Figure 5.2. Mobilisation de SGI1Kn par pVCR94Sp ΔacaB L’essai de conjugaison a été réalisé en utilisant E. coli VB113 portant soit pVCR94Sp (barres blanches) soit pVCR94Sp ΔacaB, avec le mutant indiqué de SGI1Kn. Les gènes ont été abrégés comme suit : aB pour acaB, aDC pour acaDC, C, D et DC pour sgaC, sgaD et sgaDC, et enfin 10 pour S010. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des donneuses, sélectionnées sur Nx, Kn, Sp, (et Ap pour les conditions avec pBAD30). Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique, en gris clair pour les fréquences de transfert de pVCR94Sp ΔacaB (A) et en gris foncé pour les fréquences de transfert de SGI1Kn (B). « ¤ » indique que la fréquence de transfert était sous le seuil de détection (10-7).

163 Figure 5.2. Mobilisation de SGI1Kn par pVCR94Sp ΔacaB (suite) Pour chaque élément, une ANOVA unidirectionnelle suivie d’un test posthoc de Dunnett réalisée sur les valeurs transformées en log, a permis de comparer chaque barre au contrôle de type sauvage. La signification statistique est comme suit : **** ; P < 0,0001, *** ; P < 0,001, ns ; non significatif.

Afin d’identifier le gène à l’origine de la complémentation observée, nous avons tenté de mobiliser plusieurs mutants de délétion de l’ilot. Parmi les conditions testées, seules les délétions de sgaC et sgaDC ont mené à une réabolition du transfert de pVCR94Sp ΔacaB (Figure 5.2A).

Nous avons précédemment montré que ni la délétion de sgaC, ni celle de sgaDC n’affecte le transfert d’un plasmide IncC, tandis que les délétions combinées de sgaDC et acaDC entrainent une abolition totale du transfert (Durand et al., 2020). AcaB étant nécessaire pour activer l’expression d’acaDC (Hancock et al., 2020), en son absence, les niveaux d’AcaCD sont trop bas pour activer le transfert du plasmide. L’activation du transfert repose alors sur le complexe homologue encodé par SGI1 : SgaCD.

Pour confirmer cette hypothèse nous avons surexprimé sgaDC ou acaDC à partir d’un promoteur inductible à l’arabinose : le transfert de pVCR94Sp ΔacaB est alors rétabli et dépasse même le niveau du type sauvage (Figure 5.2A). Ces résultats concordent avec la capacité de SgaCD d’activer le transfert d’un plasmide IncC portant un opéron acaDC non fonctionnel (Durand et al., 2020).

Toutefois, la complémentation par SGI1Kn du transfert de pVCR94Sp ΔacaB vient avec un prix : l’abolition de sa propre mobilisation. La surexpression d’acaB permet de rétablir la fréquence de mobilisation au niveau du type sauvage. De façon intrigante, la délétion du gène de fonction inconnue S010 permet également de rétablir la fréquence de mobilisation à un niveau non significativement différent de celui du type sauvage (Figure 5.2B). Dans le même contexte,

164 pVCR94Sp ΔacaB transfère à une fréquence pratiquement 5000 fois plus élevée que lorsque S010 est présent (Figure 5.2A). Ce résultat laisse supposer que S010 est un répresseur du transfert des plasmides IncC. Des expériences additionnelles visant à caractériser la fonction de ce gène sont détaillées dans le chapitre suivant.

En résumé, il semblerait qu’AcaB soit également essentiel à la régulation de la mobilisation de SGI1. Toutefois, bien que sa perte puisse être compensée par la surexpression d’acaDC ou sgaDC dans le cas du plasmide, ces conditions ne suffisent pas à rétablir la mobilisation de SGI1 (Figure 5.2B).

5.1.2. acaB est impliqué dans la stabilité des plasmides IncC

Afin d’étudier l’impact de la délétion d’acaB sur la stabilité du plasmide, le taux d’excision de SGI1 ainsi que sa réplication, nous avons conduit des expériences de qPCR. Les conditions impliquant acaB s’ajoutent aux conditions précédemment testées dans le cadre de l’article présenté en CHAPITRE 3.

La délétion d’acaB, tout comme celle d’acaDC, n’a pas eu d’effet significatif sur le taux d’excision de SGI1Kn (Figure 5.3A). Ces résultats pourraient potentiellement être expliqués par la répression du promoteur Pacr1 dirigeant l’expression d’acaDC. Si les conditions de l’expérience ne suffisent pas à lever la répression par Acr1 et Acr2, dans ce cas l’expression d’acaDC serait trop faible pour que la délétion de l’opéron entraine un effet significatif. En comparaison, l’impact significatif de la délétion de sgaDC montre que l’opéron est bien plus exprimé qu’acaDC dans ces conditions.

165

Figure 5.3. Impact de la délétion d’acaB sur les dynamiques de pVCR94Sp et SGI1Kn Les expériences ont été réalisées comme décrit précédemment (Durand et al., 2020). Les souches d’E. coli VB113 contenant les éléments indiqués, ont été cultivées sur la nuit dans du milieu LB liquide avec les antibiotiques appropriés. L’ADN génomique a été extrait à partir de 1 mL de culture. Les cibles et les gènes de référence dnaB, hicB et trmE ont été amplifiés par qPCR à l’aide d’amorces spécifiques. Le taux d’excision de SGI1Kn (A) a été calculé en divisant le nombre de sites attB par chromosome. Les nombres de copies de SGI1Kn (B) et de pVCR94Sp (C) ont été calculés en faisant le ratio S026 ou repA par chromosome, respectivement. Les gènes ont été abrégés comme suit : aB pour acaB, DC pour acaDC et sgaDC, ΔΔ correspond à une double délétion d’acaB et acaDC.

166 Figure 5.3. Impact de la délétion d’acaB sur les dynamiques de pVCR94Sp et SGI1Kn (suite) Pour chaque panneau, les valeurs ont été normalisées par celle du contrôle (barre blanche) fixée à 1. Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique. « ¤ » indique que le taux d’excision était sous le seuil de détection (0,001%). (A et C) Deux ANOVA unidirectionnelles suivies d’un test posthoc de Dunnett (réalisés sur les valeurs transformées en log dans le panneau A), ont permis de comparer chaque barre au contrôle. (B) Une ANOVA unidirectionnelle suivie d’un test posthoc de Tukey a permis de comparer toutes les barres deux à deux. « a » indique que les comparaisons annotées étaient non significatives. La signification statistique est comme suit : **** ; P < 0,0001, ** ; P < 0,01, ns ; non significatif.

De façon similaire, les délétions d’acaB ou d’acaDC (individuelles ou combinées) n’ont pas mené à une abolition de la réplication, contrairement à la délétion de sgaDC (Figure 5.3B). En revanche, la délétion d’acaB lorsque SGI1Kn était présent dans la souche, a entrainé une diminution significative du nombre de copies de pVCR94Sp (Figure 5.3C). Ce résultat est d’autant plus surprenant qu’il n’est pas reproduit en l’absence de SGI1 ou lorsqu’on délète acaDC à la place.

Pour vérifier si le phénomène entrainait une perte du plasmide, nous avons cultivé la souche en question en faisant varier la pression de sélection (Figure 5.4). Dans les deux conditions testées, la souche d’E coli VB113 portant pVCR94Sp ΔacaB et SGI1Kn a été cultivée en appliquant une pression de sélection sur les cellules contenant l’ilot. Comme attendu, toutes les colonies testées était résistantes à la kanamycine à la fin de l’expérience, confirmant qu’aucune n’avait perdu SGI1Kn. Dans la condition avec pression de sélection sur les deux éléments, toutes les colonies testées en fin d’expérience étaient capables de croître sur un milieu supplémenté avec les deux antibiotiques. Toutefois le profil des « patches », constitués de multiples microcolonies, laisse supposer que le plasmide subit une perte à haute fréquence. Dans la condition sans pression de

167 sélection sur le plasmide, seulement 12 des 48 colonies testées étaient encore résistantes à la spectinomycine, confirmant l’instabilité du plasmide.

Figure 5.4. Instabilité de pVCR94Sp ΔacaB en présence de SGI1Kn La souche E. coli VB113 contenant pVCR94Sp ΔacaB et SGI1Kn a été cultivée sur la nuit dans du milieu LB liquide supplémenté avec les antibiotiques indiqués (Kn : kanamycine, Sp : spectinomycine, afin de sélectionner les cellules contenant l’ilot et/ou le plasmide, respectivement). Pour chacune des deux conditions, la culture a d’abord été striée par épuisement sur milieu LB solide supplémenté en kanamycine, puis 48 colonies isolées ont été striées par « patches » sur milieu LB solide supplémenté en kanamycine. Enfin, des répliques au velours ont été réalisées sur milieu LB solide supplémenté en kanamycine et spectinomycine. Chaque flèche droite représente une incubation sur la nuit à 37°C.

168 Bien qu’il s’agisse de résultats préliminaires, il semblerait que la délétion d’acaB, en présence de SGI1, entraine une perte à haute fréquence du plasmide.

AcaB est présumé être un déclencheur initial, essentiel à l’activation de l’expression d’acaDC (Hancock et al., 2020). En son absence, l’expression insuffisante d’acaDC pourrait avoir eu pour conséquence un défaut d’excision de l’ilot, ce qui aurait mené au phénotype observé où le plasmide peut transférer mais pas SGI1. Une expérience non présentée ici a d’ailleurs confirmé que la mobilisation de l’ilot MGIVchHai6 par pVCR94Sp est également abolie lorsqu’on délète acaB. Cependant, on a montré par des essais de qPCR que la délétion d’acaB n’entraine pas d’effet significatif sur le taux d’excision de SGI1. De surcroît, l’ilot peut être mobilisé à la même fréquence que le type sauvage lorsqu’on délète en plus le gène de fonction inconnue S010, laissant supposer que dans un tel contexte, ni l’excision, ni l’initiation du transfert ne sont défectueuses. Enfin, la présence de SGI1Kn semble causer la perte à haute fréquence de pVCR94Sp ΔacaB, contrastant avec la capacité du plasmide à transférer dans un tel contexte à des niveaux non significativement différents du type sauvage. Ensemble, ces résultats suggèrent qu’AcaB joue bien plus qu’un rôle d’activateur initial dans le cycle de vie des plasmides IncC et de l’ilot génomique SGI1.

5.1.3. AcaB est conservée chez plusieurs familles d’éléments génétiques mobiles

Afin d’étudier la conservation d’AcaB, nous avons utilisé l’outil Blastp et avons ainsi pu dresser une liste de ses homologues. Après avoir aligné les séquences, un arbre phylogénétique a pu être construit (Figure 5.5).

169

Figure 5.5. Analyse phylogénétique de la protéine AcaB L’arbre présentant le plus grand log de vraisemblance (-1683,6879) est présenté à l’échelle. Les valeurs de bootstraps sont indiquées en pourcentages à la base des branches lorsque supérieures à 80. La longueur des branches représente le nombre de substitutions par site sur 213 acides aminés. Les plasmides IncA sont indiqués en rouge et les plasmides IncC, en orange. Pour la protéine homologue encodée par les ICE SXT/R391, un consensus a été construit à partir de vingt séquences.

La protéine AcaB est conservée chez les plasmides IncC mais aussi chez les plasmides IncA, ainsi que d’autres familles de plasmides conjugatifs comme pAsa4c ou pAhD4-1. Elle possède également un homologue encodé par le gène S063 des ICE SXT/R391, avec laquelle elle partage 54% d’identité de séquence. De façon intéressante, une étude a déjà rapporté que la délétion de S063 entrainait une diminution de la fréquence de transfert de SXT d’environ cent fois (Wozniak et al., 2009).

170 5.1.4. S063 n’est pas essentiel à la mobilisation de MGIVflInd1 par SXT

Nous avons souhaité comparer les rôles des gènes orthologues acaB et S063, encodés par les plasmides IncC et les ICE SXT/R391, respectivement. Le transfert d’un plasmide IncC délété pour acaB étant complémenté par SGI1, nous avons décidé de reproduire l’expérience en mobilisant cette fois-ci l’ilot génomique MGIVflInd1 par SXT ou son mutant de délétion pour le gène S063. La délétion s’est révélée ne pas avoir d’effet significatif sur le transfert de l’ICE ou de l’ilot (Figure 5.6). Comparé aux observations faites en l’absence de MGIVflInd1, ce résultat est curieux, surtout que l’ilot ne semble pas encoder d’homologue du gène.

Figure 5.6. Mobilisation de MGIVflInd1 par SXTΔS063 L’essai de conjugaison a été réalisé en utilisant E. coli CAG18439 portant SXT ou SXTΔS063 et MGIVflInd1, avec ou sans pBAD-S063 ou pBAD30-acaB comme souches donneuses et VB112 comme souche réceptrice. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des donneuses, sélectionnées sur tétracycline. Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique (excepté pour les complémentations), en gris clair pour les fréquences de transfert de SXT, en gris foncé pour celles de MGIVflInd1 et en noir pour les fréquences de co-transfert.

171 Figure 5.6. Mobilisation de MGIVflInd1 par SXTΔS063 (suite) Pour chaque groupe, une ANOVA unidirectionnelle suivie d’un test posthoc de Tukey réalisés sur les valeurs transformées en log, a permis de comparer toutes les barres deux à deux.

Ces essais préliminaires semblent également montrer que la surexpression de S063 ou même de l’orthologue acaB entraine une augmentation des fréquences de transfert de SXT et de MGIVflInd1. Si ces données sont confirmées (y compris en l’absence de l’ilot), S063 pourrait bien s’ajouter à la liste des acteurs de la régulation du transfert des ICE SXT/R391. Il est important de noter cependant que le système de régulation de ces éléments est particulièrement différent du système présent chez les plasmides IncC.

5.2 Caractérisation du gène S010 encodé par l’ilot génomique de résistance SGI1

5.2.1. S010 réprime le transfert des plasmides IncC et le co-transfert

Le gène de fonction inconnue S010, encodé par l’ilot génomique SGI1, est un petit gène situé directement en aval de traGS et exprimé à partir du même promoteur activable par AcaCD (Durand et al., 2020). Il encode une protéine putative de 84 acides aminés qui présente la grande particularité d’être similaire à l’extrémité C-terminale de TraI, la relaxase des plasmides IncC. En effet, le produit de traduction prédit de S010 partage 31% d’identité de séquence avec la région C-terminale 914 à 989 de TraI. À l’aide d’essais de conjugaison, nous avons pu montrer que la délétion de S010 entraine une augmentation drastique de la fréquence de transfert de pVCR94Sp et de celle du co-transfert, mais n’affecte pas celle de SGI1Kn (Figure 5.7).

172

Figure 5.7. Transfert d’un plasmide IncC en présence de SGI1Kn ΔS010 L’essai de conjugaison a été réalisé en utilisant VB113 portant pVCR94Sp et Kn Kn SGI1 ou SGI1 ΔS010, avec ou sans pBAD30-S010 ou pBAD30-traI907-992 comme souches donneuses, ainsi qu’E. coli CAG18439 contenant ou non pBAD30-S010 comme souches réceptrices. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des réceptrices. Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique, en gris clair pour les fréquences de transfert de pVCR94Sp, en gris foncé pour celles de SGI1Kn et en noir pour les fréquences de co-transfert. Pour chaque groupe, une ANOVA unidirectionnelle suivie d’un test posthoc de Dunnett réalisée sur les valeurs transformées en log, a permis de comparer chaque barre à la condition ΔS010 sans complémentation. La signification statistique est comme suit : **** ; P < 0,0001, *** ; P < 0,001, ns ; non significatif.

La complémentation réalisée en surexprimant S010 à partir d’un promoteur inductible à l’arabinose dans la donneuse a permis de rabaisser légèrement la fréquence de transfert de pVCR94Sp et celle du co-transfert. La complémentation dans la réceptrice n’a pas eu d’effet significatif. En revanche, la complémentation dans la donneuse et dans la réceptrice a permis de

173 diminuer encore la fréquence de transfert de pVCR94Sp et celle du co-transfert par un facteur de 27 et 41, respectivement. Ce résultat pourrait s’expliquer par le fait que les réceptrices deviennent rapidement donneuses à leur tour après un premier évènement de transfert.

Pour vérifier si la production d’une protéine correspondant à l’extrémité C-terminale de TraI homologue à la protéine encodée par S010, permettait d’obtenir le même phénotype qu’en surexprimant S010, nous avons cloné la région en question dans le plasmide pBAD30. La complémentation dans la donneuse n’a pas eu d’effet significatif sur le transfert de pVCR94Sp ou le co-transfert (Figure 5.7).

Par la suite, nous avons voulu vérifier si la surexpression de S010 en l’absence de SGI1 était suffisante pour réprimer le transfert de pVCR94Sp. Celle-ci a effectivement permis de réduire significativement la fréquence de transfert de pVCR94Sp, en particulier lorsque S010 était exprimé depuis son promoteur natif (Figure 5.8). On a montré précédemment que S010 était exprimé à partir du promoteur de traHS, c’est-à-dire qu’il est fortement activé par AcaCD. Il semblerait que ce haut niveau d’expression coïncide avec une plus forte répression du transfert de pVCR94Sp.

Figure 5.8. Impact de S010 sur le transfert conjugatif des plasmides IncC

174 Figure 5.8. Impact de S010 sur le transfert conjugatif des plasmides IncC (suite) L’essai de conjugaison a été réalisé en utilisant VB113 portant pVCR94Sp, en l’absence ou non du vecteur d’expression indiqué comme souches donneuses et E. coli CAG18439 contenant ou non pBAD30-S010 comme souches C61I réceptrices. traI907-992 et traI907-992 ont été abrégés I907-992 et I907-992*, respectivement. Le vecteur dérivé de pACYC177 permet d’exprimer S010 à partir de son promoteur natif AcaCD-activable. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des réceptrices. Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique. Une ANOVA unidirectionnelle suivie d’un test posthoc de Dunnett réalisée sur les valeurs transformées en log, a permis de comparer chaque barre au contrôle de type sauvage. La signification statistique est comme suit : **** ; P < 0,0001, *** ; P < 0,001, ns ; non significatif.

L’extrémité C-terminale de TraI présente une structure secondaire prédite différente de la protéine encodée par S010. Elle présente entre autres une troisième hélice dans une région organisée en feuillet dans le cas de la protéine encodée par S010. Nous avons tenté de modifier la séquence de traI907-992 dans le but d’observer un gain de fonction similaire à S010. Pour cela, nous avons substitué une cytosine par une isoleucine dans la région en question, la mutation ayant été prédite pour changer la conformation locale en feuillet. Malheureusement, malgré la C61I mutation introduite, la surexpression de traI907-992 n’a pas permis de réprimer significativement le transfert de pVCR94Sp (Figure 5.8).

5.2.2. S010 inhibe légèrement le transfert de MGIVchHai6

Puisque S010 affecte le transfert d’un plasmide IncC mais pas celui de SGI1, nous avons voulu tester l’effet d’une surexpression sur un autre ilot génomique mobilisable : MGIVchHai6. La surexpression de S010 dans la donneuse et dans la réceptrice a entrainé une diminution

175 significative de la fréquence de mobilisation de l’ilot et de la fréquence de co-transfert (Figure 5.9).

Figure 5.9. Effet de S010 sur la mobilisation de MGIVchHai6 L’essai de conjugaison a été réalisé en utilisant VB113 portant pVCR94Sp, MGIVchHai6Kn et pBAD30-S010 comme souche donneuse, et E. coli CAG18439 portant pBAD30-S010 comme souche réceptrice. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des réceptrices. Les barres représentent la moyenne et l’erreur standard d’un triplicata biologique. Pour chaque groupe, un test de Student non apparié réalisé sur les valeurs transformées en log a permis de comparer chaque barre au contrôle de type sauvage. La signification statistique est comme suit : **** ; P < 0,0001, *** ; P < 0,001, ** ; P < 0,01.

L’impact de S010 sur le transfert d’un plasmide IncC et la mobilisation de MGIVchHai6 mais pas sur la mobilisation de SGI1 laisse supposer un rôle au niveau de l’initiation du transfert. Celle-ci implique une reconnaissance de l’origine de transfert, puis un recrutement de la relaxase TraI. Dans le cas des plasmides IncC, la reconnaissance de l’oriT nécessite la présence du

176 facteur auxiliaire MobIC. Celui-ci a précédemment été montré ne pas être nécessaire pour mobiliser MGIVchHai6, toutefois sa délétion entrainait une diminution de la fréquence de mobilisation de l’ilot (Rivard et al., 2020). SGI1, quant à lui, ne porte aucun gène semblable à mobI. Une étude a même montré que la délétion de mobIC entraine une augmentation de la fréquence de mobilisation de SGI1-C par le plasmide IncC R16a (Hegyi et al., 2017). L’homologie entre S010 et la fin de traI laisse supposer qu’un mécanisme d’inhibition compétitive est mis en place. Pour cette raison, nous avons émis l’hypothèse que la protéine encodée par S010 séquestre le facteur auxiliaire MobIC, entrainant un défaut d’initiation du transfert des plasmides IncC.

5.2.3. La protéine encodée par S010 interagit directement avec MobIC

Pour vérifier l’hypothèse de séquestration de MobIC, nous avons effectué un test préliminaire de double hybride bactérien. Dans cet essai, chaque protéine est fusionnée à un sous-domaine du domaine catalytique de l’adénylate cyclase de B. pertussis. Lorsque les protéines interagissent directement, l’activité de l’adénylate cyclase reconstituée peut être évaluée. Ce premier test qualitatif a confirmé que les protéines encodées par S010 et mobIC interagissent directement (Figure 5.10).

Figure 5.10. Interaction directe entre les protéines encodées par S010 et mobIC évaluée par double hybride bactérien Les paires de vecteurs pUT18C / pKT25, encodant des chimères avec les sous- domaines T18 / T25 du domaine catalytique de l’adénylate cyclase et les protéines indiquées, ont été co-transformées dans la souche E. coli BTH101.

177 Figure 5.10. Interaction directe entre les protéines encodées par S010 et mobIC évaluée par double hybride bactérien (suite) L’interaction forte entre les protéines VirB8 et VirB10 du T4SS de Brucella suis a été utilisée comme contrôle positif (Sharifahmadian et al., 2017), tandis

que l’interaction entre AcaB et TraGS a été utilisée comme contrôle négatif. Les souches obtenues ont été cultivées dans du milieu LB liquide supplémenté avec les antibiotiques appropriés, puis des gouttes de 10 μL de cultures en phase stationnaire ont été déposées sur milieu LB solide supplémenté en X-gal, ampicilline et kanamycine.

Cet essai préliminaire semble confirmer l’hypothèse selon laquelle la protéine encodée par S010 est capable de séquestrer MobIC, ce qui entrainerait un défaut d’initiation du transfert des plasmides IncC.

5.3 La famille SGI1 étendue

5.3.1. Des éléments de type SGI1 intégrés dans des sites alternatifs

Les résultats décrits jusqu’à maintenant ont montré que SGI1 encodait plusieurs gènes impliqués dans des mécanismes essentiels à son cycle de vie, lesquels sont le plus souvent conservés chez de nombreux variants. Jusqu’à encore récemment, l’identification de nouveaux éléments de type SGI1 a toujours impliqué un criblage basé sur l’intégrase de l’ilot et/ou son site d’intégration canonique : le gène chromosomique trmE (Cummins et al., 2020; Siebor et al., 2020; Soliman et al., 2020).

Les ICE SXT/R391, bien qu’ils appartiennent à une famille d’éléments très différente de celle de SGI1, sont tout de même des ilots génomiques, et ont déjà été montrés encoder des intégrases alternatives (Bioteau et al., 2018). Récemment, notre équipe a montré que des ilots génomiques

178 apparentés à MGIVchHai6, canoniquement intégré à l’extrémité 3’ de trmE, pouvaient également être retrouvés intégrés à l’extrémité 3’ du gène yicC (Rivard et al., 2020).

Dès lors, nous avons émis l’hypothèse que SGI1 pouvait encoder une intégrase alternative associée à d’autres sites d’intégration que trmE. À l’aide de l’outil Blastp et des séquences de plusieurs protéines clés de SGI1 (MpsA, TraGS, SgaC et TraNS), nous avons été capables d’identifier de nouveaux ilots apparentés intégrés dans les gènes dusA de plusieurs et yicC (Figure 5.11). Plus spécifiquement, nous avons été en mesure d’identifier 24 ilots génomiques possiblement mobilisables, isolés à partir de 36 souches bactériennes différentes, et dont la taille variait de 22,8 à 37,1 kb. 21 ilots étaient intégrés à l’extrémité 5’ de dusA, tandis que 3 étaient intégrés à l’extrémité 3’ du gène yicC de E. coli, Aeromonas veronii, P. mirabilis, S. enterica serovar Kentucky et Pokkaliibacter plantistimulans (Figure 5.11). Tous portent les gènes mpsAB, traGS, traNS et sgaC dans cet ordre. Comme pour SGI1, le gène encodant l’intégrase est situé directement en aval du site attL. En revanche, l’équivalent du gène xis, rdfM, est situé à l’extrémité opposée des ilots intégrés dans yicC, comme c’est le cas pour MGIVmi1 et MGIVflInd1 (Carraro et al., 2014a; Daccord et al., 2012), et ne semble pas posséder d’homologue chez les ilots intégrés dans dusA.

179

Figure 5.11. Représentation schématique des ilots atypiques apparentés à SGI1 La position et l’orientation des cadres de lecture sont indiquées par des boites fléchées de couleur. La couleur représente la fonction prédite, comme indiqué dans la légende.

180 Figure 5.11. Représentation schématique des ilots atypiques apparentés à SGI1 (suite) Les sites AcaCD sont indiqués par des flèches vertes à angle droit. Chaque ilot est délimité par ses sites attL et attR représentés par des lignes grises verticales. Les régions conservées sont représentées en gris clair.

Aucun de ces ilots ne porte de gènes de résistance et la majorité des gènes prédits pour encoder une transposase semblent être tronqués, suggérant qu’il soit peu probable que ces ilots puissent acquérir facilement des gènes de résistance. En revanche, au moins un ilot identifié, GIVpaBan1, a visiblement subi un évènement d’accrétion ayant mené à l’acquisition d’une séquence dérivée du plasmide mobilisable ColE1, indiquant tout de même une certaine plasticité.

De façon intéressante, plusieurs ilots encodent un système toxine-antitoxine et/ou un système de restriction-modification, participant probablement à leur assurer une plus grande stabilité et/ou à assurer leur protection lorsqu’ils arrivent dans un nouvel hôte. Allant dans ce sens, absolument tous les ilots portent un gène prédit comme étant impliqué dans la réplication de l’élément (Figure 5.11). Une comparaison des séquences protéiques des produits putatifs de ces gènes semble en effet indiquer qu’ils encodent deux types de protéines initiatrices de la réplication. Notre équipe a d’ailleurs précédemment rapporté la possibilité que des ilots de type SGI1 encodent une protéine alternative d’initiation de la réplication (Huguet et al., 2020).

Enfin, plusieurs ilots encodent des gènes prédits comme étant impliqués dans des mécanismes pouvant potentiellement altérer le métabolisme de l’hôte : transport de petites molécules, chimiotaxie, régulation cellulaire et fimbriae (Figure 5.11).

5.3.2. Caractérisation de MGIVchUSA3

Nous avons voulu vérifier que les ilots nouvellement identifiés pouvaient être mobilisés par un plasmide IncC, tout comme SGI1. Pour cela, nous avons entrepris de caractériser un représentant : MGIVchUSA3. En raison de l’absence de marqueur de sélection dans le mutant

181 de type sauvage, nous avons dû réaliser plusieurs modifications génétiques pour rendre l’ilot manipulable. Une première modification a permis d’obtenir un variant de l’ilot conférant une résistance au chloramphénicol. Dès lors, nous avons pu mobiliser MGIVchUSA3Cm à l’aide de pVCR94Sp Δacr2 vers une souche d’E. coli. Les transconjugants ont pu être sélectionnés sur milieu LB supplémenté en chloramphénicol et leur intégration dans le gène dusA d’E. coli a pu être confirmée. Ces résultats préliminaires constituent la preuve de concept confirmant la capacité des ilots identifiés à être mobilisables par un plasmide IncC. Il est important de noter que le transfert de l’ilot a pu être effectué entre deux espèces bactériennes différentes, en particulier vers une espèce n’appartenant pas aux Vibrionaceae, indiquant que la spécificité observée lors de l’identification dans les bases de données n’est pas nécessairement représentative du spectre d’hôtes. Une fois dans E. coli, nous avons pu modifier l’ilot avec plus de facilité pour qu’il ne diffère du type sauvage qu’en un unique marqueur de sélection. Enfin, nous avons pu quantifier précisément la fréquence de mobilisation de l’ilot par pVCR94Sp dans des essais de conjugaison d’E. coli vers E. coli (Figure 5.12). MGIVchUSA3 s’est révélé mobilisable à une fréquence de transfert relativement élevée par notre plasmide IncC modèle (Figure 5.12).

Figure 5.12. Mobilisation de MGIVchUSA3

182 Figure 5.12. Mobilisation de MGIVchUSA3 (suite) L’essai de conjugaison a été réalisé en utilisant E. coli CAG18439 portant les mutants indiqués de pVCR94Sp et MGIVchUSA3Kn comme souches donneuses, et VB112 comme souche réceptrice. Les fréquences de transfert sont exprimées en tant que nombre d’UFC correspondant aux transconjugants, divisé par le nombre d’UFC des donneuses, sélectionnées sur Tc, Kn, Sp. Les barres représentent la moyenne et l’erreur standard obtenues à partir d’au moins trois réplicats biologiques. « ¤ » indique que la fréquence de transfert était sous le seuil de détection (10-7). Pour chaque groupe, une ANOVA unidirectionnelle suivie d’un test posthoc de Dunnett réalisée sur les valeurs transformées en log, a permis de comparer chaque barre au contrôle de type sauvage. La signification statistique est comme suit : ** ; P < 0,01, * ; P < 0,05, ns ; non significatif.

Puisque MGIVchUSA3 présente la particularité de ne pas encoder le gène sgaD, nous avons souhaité tester sa capacité à être mobilisé par un mutant de délétion de pVCR94Sp pour le gène acaC. De façon vraiment intéressante, l’ilot et le plasmide ont pu être transférés à des fréquences comparables à celles du type sauvage (Figure 5.12). Ce résultat constitue la première preuve supportant la formation de complexes chimériques SgaC/AcaD fonctionnels capables d’activer les fonctions de transfert avec la même efficacité qu’AcaCD ou SgaCD dans des essais de mobilisation. L’utilisation de pVCR94Sp ΔacaDC s’est révélée affecter négativement le transfert des deux éléments, allant jusqu’à pratiquement abolir la mobilisation de MGIVchUSA3 et le co- transfert (Figure 5.12). De façon notable, pVCR94Sp ΔacaDC est toujours en mesure de transférer à une fréquence 100 fois supérieure au seuil de détection. Ce résultat contraste les données obtenues en 2014, lesquels montraient une abolition complète du transfert lorsqu’on délétait acaD, acaC ou acaDC (Carraro et al., 2014a). Il semblerait que, non seulement SgaC encodée par MGIVchUSA3 soit suffisante pour compenser l’absence des sous-unités encodées par le plasmide, mais de surcroît qu’elle ne nécessite pas d’unité D pour être fonctionnelle. Ces données semblent donc confirmer le modèle établi récemment, selon lequel les unités D ont un

183 rôle de stabilisation et ne sont pas absolument nécessaires au recrutement du complexe transcriptionnel par les unités C (Durand et al., 2020).

L’omniprésence du gène prtN chez les ilots de type SGI1 intégrés dans dusA (Figure 5.11), mais aussi chez une majorité d’ilots non apparentés intégrés au même site (Farrugia et al., 2015), laissait supposer un rôle important du gène, probablement lié au module d’intégration. Celui-ci étant prédit pour encoder une petite protéine possédant un domaine de liaison à l’ADN, nous avons émis l’hypothèse qu’il encodait le facteur d’orientation de la recombinaison médiant l’excision des ilots intégrés dans dusA. La délétion de prtN s’est révélée abolir complètement la mobilisation de MGIVchUSA3 et le co-transfert mais ne pas affecter le transfert de pVCR94Sp (Figure 5.12). Ces premières données seront prochainement confirmées par des essais de qPCR pour évaluer l’impact de la délétion sur le taux d’excision de MGIVchUSA3, toutefois elles semblent supporter une homologie fonctionnelle avec le gène xis encodé par SGI1.

184 CHAPITRE 6 DISCUSSION ET CONCLUSION

Une relation tumultueuse unit les plasmides conjugatifs IncC et l’ilot génomique de résistance SGI1. Une hypothèse possible serait que les deux familles d’éléments partagent un ancêtre commun ; un élément conjugatif initialement autonome, lequel aurait perdu la majorité des gènes encodant sa machinerie de transfert. Une autre hypothèse serait que SGI1 soit en réalité le descendant d’un phage, lequel aurait hérité de gènes issus d’un plasmide IncC. Dans les deux cas, SGI1 a évolué pour devenir cet ilot au cycle de vie complexe, attendant patiemment l’arrivée d’un plasmide IncC pour être mobilisé, puis mettant tout en œuvre pour optimiser son maintien et sa propagation aux dépens du plasmide.

Ce projet doctoral a permis de caractériser la fonction de plusieurs gènes portés par SGI1. Les gènes traNS, traHS et traGS encodent tous trois des sous-unités du T4SS homologues aux protéines TraNC, TraHC et TraGC encodées par les plasmides conjugatifs IncC. Nous avons montré que les sous-unités encodées par l’ilot sont fonctionnelles et sont incorporées au pore de conjugaison, octroyant à SGI1 une meilleure efficacité de transfert et lui permettant d’échapper au mécanisme d’exclusion d’entrée exercé par les plasmides IncC. L’étude des gènes sgaD et sgaC, encodant le complexe d’activation transcriptionnelle SgaCD, a permis de montrer le rôle unique du complexe dans le cycle de vie de SGI1. Nous avons notamment pu montrer que SgaCD est essentiel à l’excision et à la réplication de l’ilot génomique, l’état réplicatif de SGI1 engendrant ultimement la déstabilisation du plasmide IncC. L’étude du régulateur AcaB nouvellement identifié a permis de montrer qu’il s’agit d’un déterminant essentiel de la mobilisation de SGI1, visiblement impliqué dans d’autres processus comme la stabilité du plasmide IncC en présence de SGI1. Enfin, nous avons fourni les premières preuves que SGI1 encode également un mécanisme de répression du transfert des plasmides IncC, lequel implique vraisemblablement une séquestration d’un facteur indispensable à l’initiation du transfert du

185 plasmide. Ensemble, ces résultats améliorent grandement notre compréhension des mécanismes impliqués dans la mobilisation de l’ilot génomique SGI1.

6.1 Un ilot génomique portant des gènes de transfert

6.1.1. L’importance des gènes tra portés par SGI1

La présence des gènes de transfert traNS, traHS et traGS portés par SGI1 laissait supposer un rôle important, comme l’atteste le faible nombre de variants naturels de SGI1 n’encodant pas ces gènes. Le mutant SGI1-LK1, par exemple, présente une délétion allant de traNS jusqu’à S013 (de Curraize et al., 2020). Une expérience de mobilisation a permis de montrer que la délétion n’empêchait pas le transfert de l’ilot (de Curraize et al., 2020). Des résultats similaires avaient été publiés il y a plusieurs années, lesquels impliquaient un autre mutant naturel de SGI1 avec une délétion de la région allant de traNS à traHS (Kiss et al., 2012). Malheureusement, l’écart-type associé à l’expérience de mobilisation était indiscutablement grand et aucun test statistique n’accompagnait les données. Dans le cadre de notre étude, nous avons pris soin de reproduire cette délétion, laquelle s’est avérée entrainer une diminution drastique de la fréquence de transfert de SGI1 (Carraro et al., 2017a). Nos résultats contrastent ceux de Kiss et al. (2012) dans le sens où la délétion des trois gènes semble avoir un effet plus important qu’initialement estimé. Néanmoins, dans les deux cas, la délétion n’abolit pas la mobilisation de l’ilot. Au cours du processus de revue de notre étude, nous avons ainsi dû défendre nos données en justifiant le plus possible l’intérêt d’étudier des gènes qui, si l’on se fiait à la littérature existante, ne jouaient aucun rôle important dans la mobilisation de SGI1. Heureusement, le défi a été relevé : notre étude a permis de montrer pour la première fois que ces gènes de transfert encodent des protéines fonctionnelles, lesquelles sont incorporées au pore de conjugaison et permettent à SGI1 d’optimiser sa propagation (bien qu’elles ne soient pas indispensables à celle-ci). De plus, l’incorporation de la protéine TraGS au pore de conjugaison permet d’échapper au mécanisme d’exclusion d’entrée exercé par les plasmides IncC.

186 6.1.2. GIsul2

SGI1 est le seul ilot génomique mobilisable connu à ce jour, comme étant capable de remodeler le pore de conjugaison encodé par l’élément conjugatif le mobilisant. Il n’est toutefois pas le seul ilot à porter des gènes de transfert. L’ilot génomique GIsul2 porte également trois gènes de transfert qui encodent des homologues des protéines TrbJ, TrbK et TrbL du plasmide conjugatif RP4 (Chiu and Thomas, 2004; Harmer et al., 2017). Celles-ci constituent toutes les trois des sous-unités du T4SS de type P encodé par RP4, un plasmide appartenant au groupe IncP1-α. TrbJ est une protéine associée à la membrane interne, homologue à VirB5 et qui aurait pour rôle la formation d’un pore membranaire (Lawley et al., 2003; Pansegrau et al., 1994). TrbK est une protéine d’exclusion d’entrée, agissant vraisemblablement au niveau du système de formation de la paire conjugative (Haase et al., 1996). Enfin, TrbL, homologue à VirB6, est une protéine impliquée dans la stabilisation de la paire (Lawley et al., 2003; Pansegrau et al., 1994). Du côté de SGI1, TraNS est une adhésine putative probablement impliquée dans la stabilisation de la paire, TraHS semble également impliquée dans la stabilisation et/ou la formation de la paire et

TraGS est l’acteur principal de l’échappement par SGI1 au mécanisme d’exclusion d’entrée exercé par les plasmides IncC (Carraro et al., 2017a; Humbert et al., 2019). Pour les deux familles d’ilots, ces gènes encodant des protéines aux fonctions similaires semblent avoir été conservés au travers de l’évolution pour permettre à l’élément d’optimiser sa propagation. Ces deux familles diffèrent cependant en un point majeur : leur site d’intégration. SGI1 n’a jusqu’à maintenant été retrouvé intégré que dans des sites chromosomiques. GIsul2 quant à lui, a été rapporté être intégré au sein de plasmides conjugatifs appartenant à au moins 3 groupes d’incompatibilité différents, en plus de sites chromosomiques. De façon notable, l’ilot est intégré dans les plasmides IncC R16a et pIP40a (Harmer et al., 2017; Szabó et al., 2016). Enfin, le mécanisme et la spécificité (si elle existe) de mobilisation de l’ilot sont inconnus. L’hypothèse la plus probable est que GIsul2 serait mobilisable par les plasmides IncP, toutefois cette hypothèse repose uniquement sur l’homologie des gènes de transfert (Harmer et al., 2017).

187 De façon notable, plusieurs études publiées incluent des essais de mobilisation de SGI1 par R16a (Douard et al., 2010; Kiss et al., 2015, 2019; Murányi et al., 2016). Toutefois, malgré la similarité des fonctions encodées par SGI1 et GIsul2, aucune de ces études n’a évalué les potentielles interactions en jeu.

6.1.3. Échappement à l’exclusion d’entrée

L’étude des gènes de transfert de SGI1 a permis de révéler le rôle clé de TraGS. La protéine encodée par SGI1 est incorporée à la place de TraGC dans le T4SS et permet à SGI1 d’échapper au mécanisme d’exclusion d’entrée exercé par les plasmides IncC. Ce mécanisme permet d’empêcher le transfert redondant d’un élément conjugatif entre deux cellules donneuses. Il est à distinguer du mécanisme d’exclusion de surface, ou du phénomène d’incompatibilité, ce dernier intervenant lorsque les deux copies sont déjà dans la même cellule. Chez les plasmides F et R100, le mécanisme d’exclusion d’entrée repose sur l’interaction entre deux protéines : TraG dans la donneuse, et le facteur d’exclusion TraS dans la réceptrice (Audette et al., 2007). Chez les ICE SXT/R391, c’est le facteur Eex qui interagit avec TraG. De façon surprenante, les domaines cytoplasmiques des deux partenaires sont essentiels à l’interaction (Marrero and Waldor, 2005, 2007). Ce résultat a inspiré plusieurs hypothèses quant à la capacité de TraG à contacter le facteur d’exclusion séparé par 4 membranes dans le cytoplasme de la cellule partenaire. Le signal pourrait potentiellement être transduit grâce à une extension ou une translocation du domaine C-terminal de TraG au travers du T4SS (Figure 6.1) (Firth and Skurray, 1992).

Notre équipe a confirmé que l’exclusion d’entrée exercée par les plasmides IncC impliquait

TraGC, mais aussi le facteur d’exclusion EexC (Humbert et al., 2019). Le mécanisme permet d’exclure les plasmides IncA en plus des plasmides C, expliquant l’incompatibilité apparente réfutée par Ambrose et al. (2018). Ainsi, lorsqu’on surexprime eexC dans la souche réceptrice, le transfert du plasmide IncC pVCR94Sp est pratiquement aboli. En revanche, la surexpression

188 conjointe de traGS dans la souche donneuse permet de rétablir le transfert du plasmide à des niveaux comparables au type sauvage (Humbert et al., 2019).

Figure 6.1. Rôle adaptatif de TraG/VirB6 Les T4SS de type F et P sont représentés selon le même code couleur que la Figure 1.4. Le panneau de gauche (T4SS de type F) illustre le mécanisme d’exclusion d’entrée, en particulier l’hypothèse selon laquelle TraG s’étendrait jusqu’au facteur d’exclusion TraS. Alternativement, l’interaction pourrait avoir lieu après qu’un fragment de TraG ait été transloqué au travers du T4SS. Le panneau de droite (T4SS de type P de Rickettsia) illustre l’hypothèse selon laquelle VirB6 aurait un rôle dans le ciblage spécifique des cellules eucaryotes (Gillespie et al., 2009, 2010; Rancès et al., 2008). Tiré de (Christie, 2016).

Dans le cadre de notre étude, nous avons pu montrer que les promoteurs PtraF et PtraHs, dirigeant l’expression de traGC et traGS respectivement, exhibaient des activités très différentes (Durand

189 et al., 2020). En effet, nos essais rapporteurs ont montré qu’en l’absence d’induction, PtraF présente une activité extrêmement faible, parfois très proche du seuil de détection (en pratique, il fallait attendre jusqu’à 5h pour voir apparaître les premiers signes d’hydrolyse du substrat).

De façon contrastée, dans les mêmes conditions, PtraHs présente une activité constitutive faible.

Cette différence observée renforce notre hypothèse initiale selon laquelle les unités TraHS et

TraGS seraient accumulées de façon précoce pour pouvoir être plus rapidement incorporées dans le T4SS à la place des unités encodées par le plasmide (Carraro et al., 2017a). Ces données sont remarquables puisque la faible activité de PtraF est probablement liée à l’imposante place de

TraGC dans le T4SS. Un tel contrôle au niveau transcriptionnel laisse supposer une bonne stabilité de l’ARNm et de la protéine produite, mais aussi que seule une activation par AcaCD constitue la marque d’un assemblage imminent du système de sécrétion. SGI1 a visiblement évolué pour profiter de ces contraintes, tel que l’illustre l’incorporation préférentielle de ses propres sous-unités pour optimiser son transfert. Alternativement, la faible expression basale de traGC pourrait être liée à l’expression constitutive de eexC, encodant le facteur d’exclusion d’entrée des plasmides IncC. La production retardée de TraGC assurerait l’absence d’interaction dans la donneuse avec son facteur d’exclusion conjoint, une interaction n’existant pas lorsque l’unité TraGS est produite de façon précoce.

Cette capacité de SGI1 à échapper au mécanisme d’exclusion d’entrée a d’importantes ramifications. Puisque l’ilot est capable de transférer vers des cellules portant déjà un plasmide IncC, sa propagation dans une population en est grandement accrue. En outre, SGI1 est également capable de déstabiliser les plasmides IncC, ce qui veut dire qu’ultimement, la majorité des cellules ayant acquis les deux éléments vont finir par ne porter que l’ilot. Celles-ci peuvent dès lors acquérir de nouveaux éléments conjugatifs conférant potentiellement de nombreux gènes de résistance. SGI1 participe de cette façon à augmenter le potentiel adaptatif d’une population bactérienne face à des conditions environnementales variables.

190 6.1.4. Rôle de TraNS

SGI1 encode également l’adhésine putative TraNS, probablement impliquée dans la stabilisation de la paire conjugative. Bien que nos résultats aient montré l’importance de TraNS pour stabiliser un pore auquel TraHS et TraGS ont été incorporées, certaines observations n’ont mené qu’à des hypothèses (Carraro et al., 2017a). Par exemple, la double délétion de traNC et traNS n’abolit pas la mobilisation de SGI1. Ce résultat a été attribué à la petite taille de l’ilot, qui permettrait à l’élément de transférer plus rapidement que le plasmide, le rendant peu sensible à un détachement prématuré de la paire conjugative. D’autre part, là où la délétion de traNC n’a Sp Kn qu’un très faible impact sur le transfert de pVCR94 et SGI1 , celle de traNS entraine une augmentation significative de la fréquence de transfert des deux éléments. Il semblerait qu’en présence de sa sous-unité homologue, TraNS ait un effet négatif sur le transfert. Cela pourrait

éventuellement expliquer pourquoi PtraNs est le seul promoteur AcaCD-dépendant qui n’est pas activable par SgaCD (Durand et al., 2020, Figure 3.5). SGI1 pourrait avoir évolué de sorte que la cellule limite l’accumulation d’adhésines à sa surface, lesquelles peuvent constituer des récepteurs facilitant des infections par des bactériophages (Carraro et al., 2017a).

Il est probable que le rôle de protéines comme TraN ou TraH, impliquées dans la stabilisation de la paire, soit d’autant plus important dans un contexte de conjugaison en milieu liquide. Or, toutes les expériences de conjugaison réalisées au laboratoire ont été faites sur milieu LB solide.

6.2 Étude du gène acaB

La caractérisation du gène acaB a permis d’ajouter la protéine qu’il encode à la liste des acteurs de la régulation du transfert des plasmides IncC. Nos données montrent que ce rôle s’étend aux ilots génomiques de type SGI1. En effet, nous avons montré que la délétion d’acaB entrainait une abolition de la mobilisation de SGI1 (Figure 5.2B). Ce résultat est d’autant plus surprenant que, dans les mêmes conditions, le transfert du plasmide est rétabli (Figure 5.2B), signifiant que les opérons de transfert sont bien activés. Toujours en absence d’acaB, si la souche porte un

191 mutant de délétion de SGI1 pour l’opéron sgaDC, le transfert du plasmide est à nouveau aboli (Figure 5.2A), confirmant que SgaCD est responsable de l’activation du transfert du plasmide en l’absence d’AcaB. Ces données montrent également que l’activité du promoteur dirigeant l’expression de sgaDC ne dépend pas d’AcaB. Cette hypothèse est supportée par le faible effet qu’a la surexpression de sgaDC sur le transfert de pVCR94Sp ΔacaB (Figure 5.2A). Il est donc possible qu’AcaB active un ou plusieurs autres gènes spécifiquement nécessaires pour mobiliser SGI1. Puisque la délétion d’acaB n’abolit ni l’excision, ni la réplication de l’ilot (Figure 5.3), il est possible que l’initiation du transfert à l’oriT de SGI1 soit compromise. Cette hypothèse permettrait d’expliquer le phénotype contrasté observé pour le plasmide. Toutefois, elle n’explique pas l’instabilité du plasmide causée par la délétion d’acaB, conjointement avec la présence de SGI1 dans la souche (Figure 5.4). Des expériences plus poussées seront indispensables pour émettre des hypothèses quant aux causes du phénomène observé.

Les conditions nécessaires à l’expression de sgaDC sont encore inconnues. Une étude a précédemment rapporté l’expression constitutive de l’opéron, basée sur des tests impliquant la région en amont de sgaD jusqu’à S009 (incluant le gène de fonction inconnue S008) (Kiss et al., 2015). Ces données doivent être interprétées avec précaution puisque les sites d’initiation de la transcription pour ces opérons n’ont pas été identifiés. Dès lors, on ne peut exclure la possibilité qu’un promoteur dirigeant l’expression de sgaDC chevauche le gène S008. S008 et sgaC semblent d’ailleurs être transcrits à des niveaux très distincts (Golding et al., 2007). D’autre part, l’expression constitutive de sgaDC n’est pas cohérente avec l’absence d’excision spontanée de SGI1 et la capacité de SgaCD d’activer efficacement l’excision de l’ilot (Durand et al., 2020; Kiss et al., 2012). En d’autres termes, la présence d’un plasmide IncC est absolument nécessaire pour activer ou déréprimer le promoteur dirigeant l’expression de sgaDC (Durand et al., 2020).

De façon similaire, l’initiation du transfert de SGI1 pourrait bien être réprimée jusqu’à l’arrivée d’un plasmide IncC dans la souche, puis déréprimée par AcaB. La recherche de séquences conservées dans le promoteur Pacr1 et chez SGI1 a d’ailleurs permis d’identifier un motif présent

192 à la fois dans Pacr1 (en aval du site de liaison Acr2) et dans la région promotrice de l’opéron mpsAB. Ce motif pourrait bien être le site de liaison d’un répresseur putatif de l’initiation du transfert de SGI1. Cette hypothèse nécessite toutefois d’être confirmée de façon expérimentale.

D’autres résultats tendent à montrer un rôle pléiotropique pour AcaB. De façon notable, la délétion du site de liaison d’AcaB dans le promoteur Pacr1 n’entraine pas l’abolition du transfert du plasmide, suggérant que la protéine puisse être impliquée dans d’autres interactions (Hancock et al., 2020).

Les auteurs ont pu résoudre la structure de la protéine par cristallographie à une résolution de 2.9 Å dans sa conformation native. De façon intéressante, AcaB présente des similarités structurales avec les facteurs de transcription bactériens de la famille RHH (Figure 6.2).

Plusieurs d’entre eux ont été caractérisés jusqu’à présent, dont le facteur auxiliaire TraM du plasmide F. Celui-ci a été montré lier plusieurs sites sous forme de tétramères au niveau de l’oriT et interagir avec la protéine de couplage. Ces interactions, caractérisées en profondeur, constituent un barrage permettant d’assurer la spécificité du substrat lors de son contact avec la machinerie de conjugaison (Beranek et al., 2004; Lu and Frost, 2005; Lu et al., 2008; Peng et al., 2014; Wong et al., 2011).

Bien qu’aucun site de liaison AcaB n’ait pu être clairement identifié en amont de mpsAB, il n’est pas impossible que le régulateur soit également impliqué dans l’initiation du transfert à l’oriT de SGI1. Un tel rôle semble toutefois difficilement explicable lorsqu’on considère qu’en absence d’acaB, SGI1ΔS010 peut être mobilisé à la même fréquence que le type sauvage (Figure 5.2B).

193

Figure 6.2. Comparaison des structures d’AcaB et d’autres protéines RHH La première et la seconde hélice en épingle à cheveux d’AcaB sont représentées en jaune et en cyan, respectivement. Les hélices à l’extrémité C-terminale sont représentées en magenta. La région en feuillets β impliquée dans la liaison à l’ADN est représentée en rouge. Le code couleur est conservé pour les autres protéines RHH caractérisées. Tiré de (Hancock et al., 2020).

194 6.3 Un ilot génomique encodant un inhibiteur de la fertilité des plasmides IncC

Le phénotype observé en l’absence du gène S010 encodé par SGI1 est un des phénotypes les plus impressionnants observés dans des essais de mobilisation de mutants de l’ilot. En effet, la délétion de S010 entraine une augmentation drastique de la fréquence de transfert de pVCR94Sp, mais aussi de celle du co-transfert (Figure 5.7). De manière générale, le co-transfert semble être un évènement rare (Carraro et al., 2014a; Harmer et al., 2016), lequel résulte plus vraisemblablement de deux évènements de transfert successifs : l’un impliquant le plasmide, l’autre l’ilot génomique. Il y a plusieurs raisons de croire en cette hypothèse :

- Le transfert de plusieurs éléments au travers d’un même T4SS vers une même cellule réceptrice nécessiterait une extraordinaire stabilité de la paire conjugative. - Des auteurs ont déjà rapporté la possibilité que plusieurs structures dérivées d’un T4SS de type F soient présentes à la surface d’une même cellule (Hu et al., 2019) - La séparation de la paire est probablement enclenchée en réponse au transfert complet d’un élément donné.

En d’autres termes, il est probable que soit le plasmide, soit SGI1, « privatise » l’accès au T4SS. L’observation finale du compte de colonies après un essai de conjugaison est la conséquence de multiples évènements de transferts initiés tantôt à partir de l’oriT du plasmide, tantôt à partir de l’oriT de l’ilot. Dans un contexte naturel, la faible fréquence de co-transfert semble indiquer qu’au moins un facteur nécessaire à l’initiation du transfert est limitant. Le phénotype inverse est observé lorsqu’on délète S010.

6.3.1. Initiation du transfert des plasmides IncC et de SGI1

Plusieurs systèmes ont été étudiés jusqu’à maintenant et la caractérisation d’un système de sécrétion à souvent mené à inférer la fonction de protéines homologues chez d’autres systèmes. Le relaxosome en revanche, est un complexe présentant bien plus de diversité fonctionnelle. De

195 façon notable, la relaxase TraI encodée par les plasmides IncC appartient à la famille MOBH, tandis que les mécanismes les plus souvent décrits s’appliquent en réalité majoritairement aux relaxases MOBF, ces deux familles étant particulièrement éloignées phylogénétiquement (Garcillán-Barcia et al., 2009). Récemment, une étude a rapporté pour la première fois une caractérisation biochimique d’une relaxase de la famille MOBH; la relaxase TraI encodée par l’ilot GGI de N. gonorrhoeae (Heilers et al., 2019). Bien que celle-ci et son homologue encodée par les plasmides IncC appartiennent à deux sous-groupes différents (MOBH2 et MOBH1, respectivement), les deux protéines partagent plusieurs points communs. TraI encodée par l’ilot GGI est constituée de trois domaines : un domaine relaxase en N-terminal, un domaine central désordonné, ainsi qu’un domaine C-terminal de fonction inconnue (Heilers et al., 2019). Le domaine relaxase contient deux motifs : un motif 3H alternatif et un domaine HD- phosphohydrolase. Ces deux motifs sont conservés chez TraI encodée par les plasmides IncC. Ensemble, ils permettent à TraI d’accueillir deux cations métalliques divalents, lesquels sont requis pour cliver l’ADN à l’oriT. La liaison d’un premier cation avec une grande affinité permet de stabiliser la protéine, tandis que la liaison d’un second cation complète le site actif (Heilers et al., 2019). Les cinq résidus ayant été montrés importants pour coordonner la liaison de ces cations sont bien conservés chez TraI encodée par les plasmides IncC. De façon intéressante, TraI encodée par l’ilot GGI ne se lie pas de façon covalente au brin d’ADN à transférer (Heilers et al., 2019), un phénomène ayant été rapporté une seule autre fois pour la relaxase du plasmide

CloDF13 appartenant à la famille MOBC (Núñez and De La Cruz, 2001). Ensemble, ces données suggèrent que des mécanismes similaires puissent être mis en jeu lors de l’initiation du transfert des plasmides IncC.

Tout comme pour les ICE SXT/R391, TraI est essentielle au transfert des plasmides IncC, mais aussi à la mobilisation de MGIVchHai6 (Beaber et al., 2002; Carraro et al., 2015a; Rivard et al., 2020). La surexpression de traI de pVCR94 s’est révélée toxique (Rivard et al., 2020), signe que la protéine n’est probablement produite qu’en petite quantité dans un contexte naturel. La délétion du gène a été rapportée ne pas abolir la mobilisation de SGI1-C, bien qu’elle ait été pratiquement indétectable (10-8) (Kiss et al., 2019), suggérant un rôle essentiel de la relaxase

196 dans la mobilisation de SGI1. La protéine TraJ, également essentielle au transfert des plasmides IncC et des ICE SXT/R391, a récemment été montrée indispensable à la mobilisation de MGIVchHai6 (Rivard et al., 2020; Wozniak et al., 2009). Son implication dans l’initiation du transfert de SGI1 reste toutefois encore en suspens. Enfin, le facteur auxiliaire MobIC a été montré indispensable au transfert des plasmides IncC mais pas à la mobilisation de MGIVchHai6 ou SGI1-C (Carraro et al., 2014b; Hegyi et al., 2017; Rivard et al., 2020). Au contraire, il semblerait que la présence de MobIC ait un effet négatif sur le transfert de SGI1-C Sp (Hegyi et al., 2017), et que celle de MobIM ait un effet négatif sur le transfert de pVCR94 (Rivard et al., 2020), confirmant une compétition entre le plasmide et l’ilot lors de l’initiation du transfert.

La reconnaissance de l’oriT de SGI1 semble donc reposer non pas sur MobIC, mais principalement sur MpsA, une relaxase atypique nécessitant MpsB pour se lier et/ou cliver le brin d’ADN au niveau du site nic (Kiss et al., 2019). La formation du relaxosome a été proposée ne pas être dépendante de TraI, dont le rôle proposé serait la mise en relation du substrat avec la protéine de couplage TraD (Kiss et al., 2019). Cette hypothèse serait supportée par la co- évolution des gènes traI et traD (Garcillán-Barcia et al., 2009), mais aussi par le fait que la majorité des éléments mobilisables encodent les protéines nécessaires à l’initiation de leur transfert (Guédon et al., 2017). Une fois le relaxosome formé, le transfert de ces éléments devient dès lors dépendant du système de sécrétion encodé par leur élément conjugatif mobilisateur.

Il semblerait que SGI1 ait trouvé la solution ultime pour assurer sa mise en contact avec la protéine de couplage encodée par les plasmides IncC : privatiser la relaxase. Plus spécifiquement, nos résultats semblent montrer que le gène S010 porté par SGI1, encoderait une protéine capable de séquestrer le facteur MobIC. En imitant la structure du domaine de TraI vraisemblablement impliqué dans la liaison avec MobIC, la protéine encodée par S010 empêcherait activement la formation du relaxosome au niveau de l’oriT du plasmide. Des essais de retard sur gel seront toutefois nécessaires pour confirmer cette hypothèse.

197 6.3.2. Inhibition de la fertilité

Ce n’est pas la première fois qu’un système de compétition entre deux familles d’éléments conjugatifs non apparentés est rapporté. Comme on a pu le voir au CHAPITRE 1, le transfert est un évènement nécessitant une régulation stricte. Parmi tous les mécanismes visant à contrôler le transfert, il est nécessaire de distinguer les mécanismes suivants :

- Ceux encodés par la cellule réceptrice, comme les systèmes de restriction-modification (Wilkins, 2002) et les systèmes de défense CRISPR-Cas (Marraffini and Sontheimer, 2008), - Ceux encodés par l’élément conjugatif lui-même, comme les mécanismes d’exclusion (Garcillán-Barcia and de la Cruz, 2008) ou les systèmes « classiques » d’inhibition de la fertilité comme le système FinOP des plasmides IncF (Frost and Koraimann, 2010), - Les mécanismes de résistance aux systèmes mentionnés ci-dessus (Carraro et al., 2017a; Roy et al., 2020), - Les systèmes « atypiques » d’inhibition de la fertilité affectant des éléments non apparentés.

Parmi ces derniers, deux systèmes majeurs ont été caractérisés. Le premier, reposant sur le gène osa encodé par les plasmides IncW, inhibe le transfert de l’ADN-T d’A. tumefaciens vers les cellules de plantes (Close and Kado, 1991; Maindola et al., 2014). Le second, reposant sur le gène pifC des plasmides IncFI et IncI1, inhibe le transfert des plasmides IncP et IncW (Getino et al., 2017; Miller et al., 1985). Dans les deux cas, la cible de l’inhibition est la protéine de couplage (Cascales et al., 2005; Getino et al., 2017; Santini and Stanisich, 1998). Dans le cas du système Osa, il a même été montré que l’inhibition s’effectue en limitant la liaison du substrat à la protéine de couplage (Cascales et al., 2005).

Le système reposant sur S010 de SGI1 s’ajoute donc à la liste d’inhibiteurs atypiques de fertilité. Avec le système de déstabilisation des plasmides IncC reposant sur sgaDC, ils sont ensemble responsables de la faible fréquence de co-isolement des deux éléments à partir d’une même souche (Huguet et al., 2016; Weill et al., 2017).

198 6.4 Un ilot génomique encodant un activateur transcriptionnel

6.4.1. Déstabilisation des plasmides IncC

La caractérisation du complexe activateur transcriptionnel SgaCD encodé par SGI1 a entre autres permis d’émettre une seconde hypothèse quant à la faible fréquence de co-isolement de l’ilot avec un plasmide IncC : SGI1 pourrait pirater le système de partition des plasmides IncC.

Les plasmides IncC encodent deux systèmes de partition : un système parABS de type I ayant été montré essentiel à la maintenance du plasmide, ainsi qu’un système parMRC de type II putatif (Carraro et al., 2015a; Hancock et al., 2017). Ce dernier est prédit pour être encodé par l’opéron vcrx151-152 de pVCR94Sp, situé directement en amont d’acr2, juste avant l’origine de transfert du plasmide. Dans le cadre de notre étude, nous avons pu confirmer que le promoteur dirigeant l’expression de cet opéron est activable par AcaCD et SgaCD (Durand et al., 2020).

Les deux systèmes sont composés de trois éléments : une protéine motrice (ParM ou ParA), un adaptateur liant l’ADN (ParR ou ParB) et une région semblable à un centromère reconnue par l’adaptateur (parC ou parS) (Salje et al., 2010). Le système parMRC, originellement isolé à partir du plasmide R1 (Gerdes et al., 1985), est le système actif de ségrégation plasmidique le mieux caractérisé à ce jour. Il repose sur la polymérisation de ParM sous forme de filaments semblables à des filaments d’actine, poussant les plasmides aux pôles opposés de la cellule (Figure 6.3).

199

Figure 6.3. Mécanisme du système de partition ParMRC

200 Figure 6.3. Mécanisme du système de partition ParMRC (suite) a) Le centromère parC est situé dans la région promotrice de l’opéron constitué de parM et parR. ParR, en se liant à parC, réprime par la même occasion la transcription au niveau du promoteur. b) Les différentes étapes de la ségrégation sont résumées brièvement, avec les plasmides représentés par des disques roses, ParR en bleu et ParM en orange. c) L’élongation de ParM sous forme de filaments nécessite la présence d’ATP. Pour ne pas être désassemblés, les filaments nécessitent d’être protégés à chaque extrémité par ParR lié à parC. Cette coiffe permet de stabiliser la structure, qui s’allonge alors de façon bidirectionnelle. Tiré de (Salje et al., 2010).

La dépolymérisation et repolymérisation de ParM se traduit par la formation spontanée de courts filaments qui parcourent aléatoirement la cellule à la recherche de complexes ParR-parC (Garner et al., 2007). Étant donné la présence d’au moins 7 copies de SGI1 dans la cellule, l’existence seule d’une séquence centromère semblable à celle du plasmide IncC favoriserait la ségrégation de SGI1 au détriment de celle du plasmide (Durand et al., 2020; Huguet et al., 2020). Des expériences seront toutefois nécessaires pour confirmer le rôle des gènes vcrx151-152 et identifier le centromère associé.

Dans le cadre de notre étude, nous avons pu observer que la surexpression de sgaDC ou acaDC dans une souche portant uniquement SGI1, entraine l’excision et la réplication de l’ilot mais mène rapidement à sa perte (Durand et al., 2020). Au contraire, la présence additionnelle d’un plasmide IncC semble conférer plus de stabilité à l’ilot. Le système ParMRC putatif encodé par les plasmides IncC ne semble pas être essentiel à la maintenance d’un plasmide IncC dans des conditions où acaDC est réprimé (Hancock et al., 2017). Toutefois, il pourrait devenir essentiel dans des conditions de transfert, comme le suggère l’expression dépendante d’AcaCD de vcrx151-152. Les conditions nécessaires à l’expression de sgaDC restent également encore à être identifiées. Les résultats que nous avons obtenus en qPCR montrent que la délétion de sgaDC a bien plus d’impact sur l’excision et la réplication de SGI1 que la délétion d’acaDC,

201 suggérant qu’acaDC est partiellement réprimé dans ces conditions. Dans ce cas, l’activation par SgaCD d’un système de partition habituellement faiblement activé, pourrait impacter négativement la maintenance du plasmide.

D’autres hypothèses ont été proposées pour expliquer la déstabilisation des plasmides IncC par SGI1, notamment la titration de protéines impliquées dans la réplication des deux éléments (Huguet et al., 2020). Dans tous les cas, cette déstabilisation est abolie en l’absence de SgaCD (Durand et al., 2020; Harmer et al., 2016), confirmant le rôle essentiel du complexe encodé par SGI1.

6.4.2. Comparaisons avec l’activateur flagellaire d’E. coli

L’étude des complexes SetCD, AcaCD et SgaCD a rapidement mené à comparer leurs similarités de séquences avec le complexe FlhDC, activateur flagellaire d’E. coli (Kiss et al., 2015; Poulin-Laprade et al., 2015a). Dans le cadre de notre étude, nous sommes allés un peu plus loin et avons apporté les preuves de similarités structurales entre les unités C, en particulier au niveau du doigt de Zinc, sur lequel repose la fonction de liaison à l’ADN (Durand et al., 2020). Le complexe FlhDC s’assemble sous forme d’hexamère composé de deux unités C contactant l’ADN et de quatre unités D aidant à stabiliser la liaison (Li et al., 2012) (Figure 6.4).

202

Figure 6.4. Structure du complexe FlhDC

Partie supérieure : la structure du complexe FlhD4C2 (PDB : 2AVU) est montrée en mode surface avec le potentiel électrostatique sous vide (rouge : négatif, bleu : positif). La rotation en amenant le dessus du complexe vers le lecteur permet d’obtenir la vue montrée en haut à droite, mettant en évidence les résidus de FlhC impliqués dans la liaison à l’ADN. Si on effectue au contraire une rotation en amenant le dessous du complexe vers le lecteur, on met alors en évidence les surfaces exposées des sous-unités FlhD. Tiré de (Li et al., 2012). Partie inférieure : la liaison du complexe à l’ADN et le recrutement du complexe transcriptionnel sont représentés de façon schématique. Tiré de (Li et al., 2017).

203 Une telle similarité structurale entre SetC, AcaC, SgaC et même AsaC et AqaC respectivement encodées par pAsa4c d’Aeromonas salmonicida subsp. salmonicida et pAQU1 de Photobacterium damselae subs. damselae, explique pourquoi le mécanisme d’activation transcriptionnelle est également similaire. SetC, AcaC et SgaC reconnaissent toutes trois un motif chevauchant la région -35, leur liaison est stabilisée par les unités D correspondantes, aidant ainsi au recrutement du facteur σ70 lié au complexe transcriptionnel (Figure 6.4). Pour cette raison, SetCD, AcaCD et SgaCD sont des activateurs transcriptionnels de classe II au même titre que FlhDC (Browning and Busby, 2004; Poulin-Laprade et al., 2015a).

Dans le cadre de notre étude, nous avons pu établir le profil de l’empreinte des complexes AcaCD / SgaCD lorsqu’ils sont liés à un promoteur comportant un site de liaison. Pour cela, nous avons procédé à des essais d’immunoprécipitation de la chromatine, couplée à un traitement aux exonucléases (ChIP-exo). Cette technique implique entre autres de produire des versions étiquetées des complexes en relativement grande quantité, puis de figer l’interaction de leur liaison à l’ADN à l’aide de formaldéhyde. Cette étape est soumise à plusieurs contraintes techniques, notamment celle d’être réalisée à température pièce. Pour cette raison, l’empreinte qui est capturée n’est pas seulement celle du complexe d’intérêt lié au promoteur, mais aussi celle du complexe transcriptionnel (Carraro et al., 2014a; Durand et al., 2020; Latif et al., 2018; Poulin-Laprade et al., 2015b). En effet, à température physiologique, c’est la conformation ouverte du complexe qui prédomine (Davis et al., 2007; Kovacic, 1987; Schickor et al., 1990; Sclavi et al., 2005) (Figure 6.5). Il est important donc de noter que le signal de ChIP-exo observé résulte d’un processus dynamique, rapide et cumulé sur de nombreux promoteurs dans de très nombreuses cellules. Malgré tout, nous avons été capables de distinguer l’empreinte d’AcaCD et SgaCD, de celle du complexe transcriptionnel débutant 5 pb avant la fin du site de liaison et s’avançant 22 pb après le site d’initiation de la transcription (Durand et al., 2020). Ces données sont supportées par la présence d’un pic important à la position -45, révélant une importante interaction n’ayant toutefois pas mené à une protection de l’ADN.

204

Figure 6.5. Initiation de la transcription bactérienne Le complexe transcriptionnel formé entre autres par la polymérase d’ARN peut prendre deux conformations : une conformation fermée (dans le sens où l’ADN est sous forme double brin), et une conformation ouverte (dans le sens où l’ADN est partiellement simple brin). Ces conformations ne doivent pas être confondues avec l’ouverture et la fermeture des pinces β/β’ qui suivent le schéma inverse. Avant que le complexe soit ouvert, la configuration fluctue entre les 3 premières positions. Tiré de (Robinson and Oijen, 2013).

Ces données sont délicates à interpréter, puisqu’une absence de signal peut signifier soit une absence de liaison par le complexe, soit une liaison si maintenue qu’elle a entrainé une protection parfaite de la séquence d’ADN. Dans notre contexte, il y a si peu de promoteurs, exprimés à un si faible niveau que la surproduction de nos complexes est absolument nécessaire. Malheureusement, comme on a pu le montrer dans le cadre de l’article, on s’écarte du contexte naturel où les complexes sont limitants. Par exemple, un détachement prématuré d’AcaCD n’entrainera pas nécessairement une interruption dans l’initiation de la transcription; le complexe sera simplement recruté à nouveau.

De façon notable, SetC appartient à un clade phylogénétiquement distant de celui qui regroupe AcaC, SgaC, AsaC et AqaC (Durand et al., 2020; Kiss et al., 2015; Poulin-Laprade et al., 2015a). Cela se traduit entre autres par un motif reconnu bien distinct. En effet, le motif reconnu par

205 SetCD est plus petit : 18 pb, contre 28 pb pour le motif reconnu par AcaCD et SgaCD, et ne comporte pas de répétitions palindromiques (Carraro et al., 2014a; Poulin-Laprade et al., 2015b).

FlhDC, quant à lui, appartient à un clade phylogénétique encore plus distant. Son motif de liaison est plus grand que celui d’AcaCD, mais comporte également deux répétitions palindromiques. La région d’interaction directe avec l’ADN est présumée correspondre à ces répétitions, lesquelles sont séparées par une région mesurant entre 10 et 11 pb (Claret and Hughes, 2002). De façon intéressante, cette longueur coïncide avec celle d’un tour complet de la double hélice d’ADN (Wang, 1979). Cette hypothèse supporte un modèle dans lequel deux monomères de FlhC contactent chaque répétition inversée « du même côté » (leurs extrémités N-terminales se faisant face), introduisant par la même occasion une courbure de l’ADN avec un angle de 111° (Campos and Matsumura, 2001; Wang et al., 2006). Notons que seulement 7 pb A/T situées dans les répétitions sont nécessaires à la liaison (Lee et al., 2011).

Si l’on compare avec le motif de liaison d’AcaCD, les répétitions sont plus courtes (4 pb au lieu de 18) et ne sont séparées que par 5 pb, soit un demi-tour de la double hélice. On peut donc émettre l’hypothèse que chaque monomère d’AcaC se lie à un « côté » différent de l’ADN, entrainant probablement un changement de conformation bien distinct de celui induit par FlhC. Cette information, si elle est avérée, est cruciale, puisqu’elle suggère un rôle potentiellement très différent pour les unités D d’un système à l’autre. Les unités D sont d’ailleurs bien plus distantes entre elles que les unités C (Kiss et al., 2015).

FlhD a été montrée importante pour stabiliser la liaison de FlhC à l’ADN (Li et al., 2012). En outre, elle a également été montrée essentielle au recrutement du complexe transcriptionnel (Li et al., 2017). Deux types de mécanismes impliquant les facteurs anti-FlhDC YdiV, STM1344 et STM167 ont été caractérisés (Li et al., 2017). Si l’inhibiteur est en faible concentration, il peut empêcher le recrutement de l’ARN polymérase. Lorsque la concentration de l’inhibiteur est élevée, il peut aller jusqu’à entrainer un détachement du complexe complet, lequel devient alors

206 plus susceptible d’être dégradé par la protéase ClpXP (Takaya et al., 2012). Ultimement, ces mécanismes peuvent causer une abolition de la motilité cellulaire mais peuvent également accroître la virulence d’une souche de Salmonella (Ahmad et al., 2013; Li et al., 2017).

Contrastant avec ces résultats, nous avons pu montrer que SgaC encodée par MGIVchUSA3 est suffisante pour activer les opérons de transfert de pVCR94Sp ΔacaDC (Figure 5.12). Bien que la fréquence résultante fût nettement plus basse que celle du type sauvage, ces données suggèrent qu’AcaD ou SgaD n’est pas aussi essentielle que FlhD. Ce phénotype est potentiellement attribuable à la liaison de SgaC au promoteur ; n’induisant pas la même courbure de l’ADN que FlhC, elle pourrait disposer d’une plus grande stabilité. De surcroît, nos données montrent que le complexe chimérique SgaC/AcaD est capable d’activer les fonctions de transfert de pVCR94Sp et MGIVchUSA3 avec la même efficacité que celle observée dans un contexte de type sauvage Figure 5.12. Ce résultat illustre une nouvelle interaction possible entre une protéine encodée par un plasmide IncC et une protéine encodée par SGI1. À l’avenir, il sera intéressant d’évaluer la prévalence de ces complexes chimériques, mais aussi de caractériser leur impact sur l’activation transcriptionnelle des opérons AcaCD-dépendants. Il sera également important d’étudier les interactions impliquant les autres acteurs de la régulation identifiés jusqu’à présent, en particulier au niveau du promoteur Pacr1 des plasmides IncC. En effet, malgré le rôle important de SgaCD, SGI1 est tout de même soumis à la régulation par AcaCD.

6.5 Perspectives

6.5.1. Caractérisation d’Acr2

Malheureusement, encore relativement peu d’informations sont disponibles quant à la régulation du transfert des plasmides IncC. On sait toutefois qu’il existe au moins un module principal de régulation, constitué de deux opérons convergents. Le premier, dont l’expression est dirigée à partir du promoteur Pacr1, encode le répresseur Acr1 et le complexe majeur d’activation AcaCD

207 (Figure 1.14) (Carraro et al., 2014a). Le second encode le répresseur Acr2, qui, tout comme Acr1, réprime l’expression du premier opéron (Carraro et al., 2014a; Lang and Johnson, 2016).

De façon notable, la délétion d’acr1 sur un plasmide dérivé du plasmide IncC pAR060302 a été rapportée ne pas affecter son transfert (Lang and Johnson, 2016). En outre, des expériences de TnSeq visant à déterminer les gènes essentiels au transfert et à la maintenance de pVCR94Sp ont montré que l’utilisation d’un mutant de type sauvage entrainait un biais d’insertion dans le gène acr2 (communication personnelle de Dominick Matteau). L’utilisation d’un mutant de délétion pour acr2 a permis de rectifier le biais en question (Roy et al., 2020). Pour ces raisons, il est probable qu’Acr2 constitue le répresseur majeur du transfert des plasmides IncC. Son rôle pourrait toutefois être plus étendu qu’initialement estimé. En effet, plusieurs sites de liaison par

Acr2 ont pu être identifiés en plus de celui présent dans Pacr1, notamment dans le promoteur

PtraF dirigeant l’expression de traF, traHC et traGC. Étant donné la pléiotropie de TraGC, on peut imaginer qu’une répression directe par Acr2 permettrait de désamorcer le T4SS, contribuant à un prompt retour à un état quiescent. Cette répression pourrait aussi constituer un moyen additionnel d’assurer un contrôle spatio-temporel de l’expression de traGC.

Une rapide recherche de la séquence consensus du site de liaison par Acr2 dans la séquence de

SGI1 a révélé une liaison potentielle dans le promoteur PS004 (données non présentées), lié à l’activation de la réplication de l’ilot génomique (Huguet et al., 2020). Cette découverte est intrigante puisque la liaison s’effectuerait exactement au même niveau qu’AcaCD ou SgaCD. Si elle est avérée, une répression par Acr2 de la réplication de SGI1 serait un moyen efficace pour empêcher que l’ilot soit répliqué lorsqu’il est piégé dans le chromosome. Alternativement, Acr2 pourrait au contraire jouer un rôle d’activateur dans un tel contexte. En effet, des auteurs ont montré qu’affecter le domaine de liaison à l’ADN de la protéine menait à une augmentation de la fréquence de transfert du plasmide IncC, en comparaison avec celle d’un mutant de délétion pour le gène au complet (Lang and Johnson, 2016). Les auteurs ont donc proposé qu’Acr2 puisse être capable, dans certaines conditions, de promouvoir la transcription plutôt

208 que la réprimer. Cette hypothèse corroborerait l’observation que nous avons faite d’une activité Sp de PS004 en présence de pVCR94 ΔacaDC et non en son absence (Huguet et al., 2020).

Chez les ICE SXT/R391, la caractérisation du répresseur SetR a constitué un point pivot dans la compréhension de leur régulation. Des études plus approfondies prenant pour cible Acr2 permettront certainement d’établir les conditions nécessaires à la dérépression du transfert des plasmides IncC, mais aussi de l’ilot génomique SGI1. Pour l’instant, on sait que la réponse SOS n’est pas impliquée dans la dérépression, puisque la délétion de recA n’impacte pas le transfert des plasmides IncC (Carraro et al., 2014b). Plusieurs autres pistes restent à être explorées pour déterminer les stimuli menant à une activation du transfert. Étant données les similarités entre les plasmides A/C et les ICE SXT/R391 (gènes orthologues, hôtes bactériens), il serait notamment intéressant d’étudier le lien entre transfert, quorum-sensing, formation de biofilm et niveaux de c-di-GMP (Bordeleau et al., 2010; Liu et al., 2014; Waters et al., 2008).

6.5.2. Études structurales

La caractérisation biochimique d’autres déterminants du transfert des plasmides IncC et de SG1 permettrait d’émettre de nouvelles hypothèses quant aux mécanismes mis en jeu. La résolution de la structure d’AcaB a par exemple permis d’établir son appartenance à la famille des facteurs transcriptionnels RHH (Hancock et al., 2020). La caractérisation de la structure du produit de traduction de S010 permettra certainement d’identifier les résidus exposés et d’estimer l’interface d’interaction avec MobIC. Dès lors, la poursuite des essais de mutagenèse sur traI902-

997, mais aussi sur S010, en sera facilitée et permettra d’évaluer les mutations entrainant une perte ou un gain de fonction. Enfin, l’interaction entre les protéines encodées par S010 et mobIC et les différents mutants obtenus devra être quantifiée à l’aide d’essais rapporteurs en milieu liquide. Ces techniques ont déjà fait leurs preuves, notamment pour caractériser l’interaction entre VirB8 et VirB10 du T4SS de B. suis (Sharifahmadian et al., 2017). Il serait également intéressant de réaliser des essais de retard sur gel pour vérifier si l’interaction entre les protéines encodées par S010 et mobIC entraine un défaut de liaison à l’oriT.

209 La cristallographie aux rayons X reste un procédé particulièrement long, nécessitant parfois plusieurs années avant d’obtenir un résultat pour une seule protéine. Toutefois, la patience est parfois récompensée par l’obtention d’une grande quantité d’informations. Le développement croissant de plateformes proposant ce service promeut grandement la collaboration entre les groupes de recherche, permet d’économiser l’achat de machines extrêmement coûteuses et laisse plus de temps pour d’autres expériences en parallèle, le temps que la protéine soit cristallisée. Plusieurs techniques ont fait leurs preuves pour établir la structure de protéines impliquées dans la conjugaison : la microscopie et la tomographie cryoélectronique (in situ, coloration négative…) (Costa et al., 2016; Hu et al., 2019; Ilangovan et al., 2017; Redzej et al., 2017; Sgro et al., 2018) ou encore la diffraction des rayons X à petits angles (Heilers et al., 2019; Kwak et al., 2017).

6.5.3. Éléments génétiques mobiles et fluorescence

La construction de mutants de SG1 et du plasmide IncC pVCR94 encodant des protéines fluorescentes a permis d’effectuer des expériences de cytométrie en flux pour évaluer le maintien des éléments dans des populations bactériennes. Cette technique a permis au laboratoire d’apporter les preuves que l’ilot génomique SGI1 est activement répliqué en présence d’un plasmide IncC et de lier cet état réplicatif à la déstabilisation du plasmide par l’ilot (Huguet et al., 2020). Nous avons également pu montrer qu’AcaCD et SgaCD étaient suffisants pour activer l’excision et la réplication de SGI1, ou encore que le phénomène de déstabilisation était aboli en l’absence de SgaCD (Durand et al., 2020). La poursuite de ces expériences à l’aide de mutants additionnels permettra de détailler le modèle proposé. Il serait notamment intéressant d’évaluer l’impact d’une délétion d’acaDC sur le maintien des deux éléments. Il sera également important de vérifier si la délétion des gènes vcrx151-152, encodant un système de partition active putatif, entraine la perte de SGI1. En outre, il serait intéressant de vérifier si les délétions combinées de vcrx151-152 et sgaDC rétablissent le phénomène d’instabilité observé pour le type sauvage. Si les résultats attendus sont obtenus, il sera nécessaire de caractériser le mécanisme de partition de SGI1 et d’établir son importance dans le

210 cycle de vie de l’ilot. De la même façon, la réplication et éventuellement la partition des ilots apparentés à SGI1 intégrés dans dusA, apportera la preuve que SGI1 constitue une large famille d’éléments partageant des cycles de vie similaires.

Jusqu’à maintenant, nous n’avons pas encore évalué l’effet potentiel des systèmes toxine- antitoxine encodés par le plasmide et par SGI1. Le système SgiAT encodé par SGI1 a pourtant été montré important pour la stabilité de l’ilot en présence d’un plasmide IncC (Huguet et al., 2016). Enfin, des expériences d’évolution accélérée couplées à du séquençage haut débit de génome complet permettraient d’identifier les mutations entrainant une abolition de l’incompatibilité entre un plasmide IncC et SGI1. Si les mutations apparaissent dans des sites de liaison AcaCD / SgaCD, elles pourraient bien permettre d’identifier précisément les bases absolument nécessaires à la liaison par ces complexes transcriptionnels.

Les mêmes mutants (pVCR94Green et SGI1Red) pourraient également être utilisés dans des essais de microscopie à fluorescence en temps réel, afin d’évaluer les dynamiques de propagation des éléments dans une population bactérienne. Cette technique a précédemment permis d’établir le modèle de dissémination de l’élément ICEclc (Delavat et al., 2016).

6.5.4. Conclusion

Nos données ont permis de caractériser un peu plus en détail les mécanismes mis en jeu lors du transfert d’un plasmide IncC et de la mobilisation de l’ilot génomique de résistance SGI1. La grande quantité d’acteurs et de voies de régulation illustre à quel point le cycle de vie de ces éléments est un ensemble de processus complexes, qui nécessite encore d’être investigué. Nous espérons avoir pu convaincre de l’intérêt d’étudier conjointement les plasmides IncC et SGI1, tant les deux éléments interagissent et dépendent l’un de l’autre. Nous espérons également avoir proposé de nouvelles cibles potentielles pour le développement de composés destinés à freiner la dissémination de résistances aux antibiotiques chez les Gammaprotéobactéries.

211 212 ANNEXE 1

Bioteau, A., Durand, R., and Burrus, V. (2018). Redefinition and Unification of the SXT/R391 Family of Integrative and Conjugative Elements. Applied and Environmental Microbiology 84, e00485-18. https://doi.org/10.1128/AEM.00485-18

L’article, formaté tel que publié dans Applied and Environmental Microbiology, est reproduit avec la permission de ASM Journals.

213 EVOLUTIONARY AND GENOMIC MICROBIOLOGY crossm

Redefinition and Unification of the SXT/R391 Family of

Integrative and Conjugative Elements Downloaded from

Audrey Bioteau,a Romain Durand,a Vincent Burrusa aDépartement de biologie, Université de Sherbrooke, Sherbrooke, Quebec, Canada

ABSTRACT Integrative and conjugative elements (ICEs) of the SXT/R391 family are key drivers of the spread of antibiotic resistance in Vibrio cholerae, the infectious http://aem.asm.org/ agent of cholera, and other pathogenic bacteria. The SXT/R391 family of ICEs was defined based on the conservation of a core set of 52 genes and site-specific inte- gration into the 5= end of the chromosomal gene prfC. Hence, the integrase gene int has been intensively used as a marker to detect SXT/R391 ICEs in clinical isolates. ICEs sharing most core genes but differing by their integration site and integrase gene have been recently reported and excluded from the SXT/R391 family. Here we explored the prevalence and diversity of atypical ICEs in GenBank databases and

their relationship with typical SXT/R391 ICEs. We found atypical ICEs in V. cholerae on June 28, 2018 by UNIVERSITE DE SHERBROOKE isolates that predate the emergence and expansion of typical SXT/R391 ICEs in the mid-1980s in seventh-pandemic toxigenic V. cholerae strains O1 and O139. Our anal- yses revealed that while atypical ICEs are not associated with antibiotic resistance genes, they often carry cation efflux pumps, suggesting heavy metal resistance. Atypical ICEs constitute a polyphyletic group likely because of occasional recombina- tion events with typical ICEs. Furthermore, we show that the alternative integration and excision genes of atypical ICEs remain under the control of SetCD, the main ac- tivator of the conjugative functions of SXT/R391 ICEs. Together, these observations indicate that substitution of the integration/excision module and change of specific- ity of integration do not preclude atypical ICEs from inclusion into the SXT/R391 family. IMPORTANCE Vibrio cholerae is the causative agent of cholera, an acute intestinal infection that remains to this day a world public health threat. Integrative and con- jugative elements (ICEs) of the SXT/R391 family have played a major role in spread- ing antimicrobial resistance in seventh-pandemic V. cholerae but also in several spe- cies of Enterobacteriaceae. Most epidemiological surveys use the integrase gene as a marker to screen for SXT/R391 ICEs in clinical or environmental strains. With the re- cent reports of closely related elements that carry an alternative integrase gene, it became urgent to investigate whether ICEs that have been left out of the family are a liability for the accuracy of such screenings. In this study, based on comparative genomics, we broaden the SXT/R391 family of ICEs to include atypical ICEs that are Received 27 February 2018 Accepted 11 April often associated with heavy metal resistance. 2018 Accepted manuscript posted online 13 April 2018 KEYWORDS antibiotic resistance, conjugation, entry exclusion, integrative and Citation Bioteau A, Durand R, Burrus V. 2018. conjugative element, Vibrio cholerae, genetic recombination, site-specific Redefinition and unification of the SXT/R391 recombination family of integrative and conjugative elements. Appl Environ Microbiol 84:e00485-18. https:// doi.org/10.1128/AEM.00485-18. ince the mid-1980s, multidrug resistance has considerably increased in toxicogenic Editor Eric V. Stabb, University of Georgia SVibrio cholerae, the infectious agent of the acute diarrheal disease cholera (1, 2). The Copyright © 2018 American Society for Microbiology. All Rights Reserved. emergence and global spread of resistance in V. cholerae have been exacerbated by Address correspondence to Vincent Burrus, integrative and conjugative elements (ICEs) of the SXT/R391 family (3, 4). ICEs are [email protected]. mobile genetic elements that propagate by conjugative transfer, a process involving a

July 2018 Volume 84 Issue 13 e00485-18 Applied and Environmental Microbiology aem.asm.org 1 Bioteau et al. Applied and Environmental Microbiology direct contact between donor and recipient cells. ICEs are usually integrated into the chromosome of their host and are vertically inherited from one generation to another. Excision of SXT/R391 ICEs from the chromosome and their transfer to a new host are triggered by conditions that induce the host SOS response, including exposure to antibiotics (5). The resistance profile and evolution of V. cholerae seventh-pandemic clones appear to have been shaped by two major independent events of ICE acquisition (4). First, in

the mid-1980s, V. cholerae strains of the O1 El Tor serogroup gained ICEVchInd5, which Downloaded from is now globally distributed in seventh-pandemic clones. In addition to conferring multidrug resistance, ICEVchInd5 seems to have restricted horizontal gene transfer because it codes for the periplasmic DNA endonuclease IdeA, which has been shown to inhibit natural transformation (6, 7). The second event took place in 1992, when SXT was acquired by V. cholerae O139 or its El Tor progenitor, prior to the first O139 cholera outbreak in the Indian subcontinent in early 1993 (4, 8). SXT and ICEVchInd5 likely originate from environmental Vibrio spp. or other Gammaproteobacteria. Now, SXT/ http://aem.asm.org/ R391 ICEs are found in several other species of Vibrio, Proteus, Providencia, Alteromonas, Shewanella, and Marinomonas. Recently, such an ICE was even reported in a drug- resistant isolate of the Pasteurellaceae member Actinobacillus pleuropneumoniae recov- ered from a pneumonic pig in the United Kingdom (9). SXT/R391 ICEs have been grouped together based on the conservation of a core set of 52 genes and their integration into the 5= end of prfC, a gene coding for the peptide chain release factor 3 (10). Over the years, a function has been attributed

to 30 of these genes. Hence, int and xis mediate integration into and excision from on June 28, 2018 by UNIVERSITE DE SHERBROOKE the chromosome at prfC (11, 12), srpR and srpM code for an active partitioning system (13), and mobI and traI initiate at the origin of transfer (oriT) a rolling-circle replication of the excised ICE (13, 14). Together with the putative type IV coupling protein TraD and the putative auxiliary relaxosome component TraJ, they also enable the conjugative translocation of the ICE into the recipient cell, via a type IV secretion system (T4SS) encoded by four operons of tra genes. bet and exo encode a ␭ Red-like homologous recombination system involved in ICE plasticity (15). Eex is the entry exclusion factor in the recipient cell that works jointly with TraG in the donor cell (16). setC and setD code for a class 2 transcriptional regulator that activates the expression of all the aforementioned operons except eex (17). Finally, setR and croS code for two transcriptional repressors that form a bistable switch controlling setCD expression in response to DNA damage (5, 18). In recent years, two reports have described atypical ICEs that are closely related to SXT/R391 ICEs but differ in structure and/or specificity of integration. ICEVchBan8 of V. cholerae O37 carries alternative int and xis genes, is integrated at the 3= end of a serine tRNA gene, and has undergone a large inversion between traG and srpM (19). Therefore, ICEVchBan8 was excluded from the SXT/R391 family (10). Whole- genome sequencing of Vibrio alginolyticus strains from China revealed three other atypical ICEs inserted in the same tRNA-Ser gene as ICEVchBan8 but lacking the large DNA inversion (20). These two reports show that int is not a conserved feature. Thus, the core set of conserved genes that was deduced from sequence comparison of a representative set of 13 ICEs integrated at prfC (10) and used to define the SXT/R391 family is likely smaller than previously described. Moreover, the lack of conservation of int suggests that its systematic use as a marker to detect SXT/R391 ICEs in epidemiological screenings needs to be reconsidered since atypical ICEs closely related to SXT/R391 are not detected when int is used as a marker, thereby jeopardizing the reliability of such screenings. In this study, we conducted a comparative analysis of a large sample of represen- tative elements that helped us redefine the core genome of SXT/R391 ICEs, which now excludes genes involved in integration/excision. We show that despite the change of integration/excision module, expression of the alternative integrase gene retains its dependency on the transcriptional activator SetCD. Based on these analyses and strict

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 2 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology conservation of the regulatory functions, we propose a novel inclusive and compre- hensive typing scheme for SXT/R391 ICEs.

RESULTS AND DISCUSSION Conserved core of genes of the SXT/R391 family of ICEs. To assess the diversity of SXT/R391 ICEs, we searched the GenBank database using R997 from Proteus mirabilis as a reference. We also screened a collection of 275 Canadian Vibrio isolates using primers designed to detect setCD. One positive isolate, V. parahaemolyticus S107-1, was Downloaded from found, and its genome was sequenced and assembled. The sequence of ICEVpaCan1 was then extracted and included in this study. Overall, a set of 68 complete ICEs from different bacterial species and origins were recovered (Table 1). After manual curation of the annotated sequences to correct missing small open reading frames, such as mobI or xis, and inconsistent start codons, the soft core of SXT/R391-like ICEs was built based on genes present in 95% of all sampled ICEs (57 of 61 ICEs). Forty-three genes were found to be strictly conserved, including those involved in DNA recombination and repair (ssb-bet-exo, radC), initiation of conjugative rolling-circle http://aem.asm.org/ replication (mobI and traI) and partition (srpRM), and conjugative DNA translocation (traDJ), as well as T4SS assembly (traLEKB, traVA, dsbC-traC-trhF-traWUN, and traFHG) and entry exclusion (eex)(Fig. 1). The regulation module that contains setCD, setR, and croS was also strictly conserved. Genes coding for a putative cobalamin synthase (cobS) and a putative lytic transglycosylase (s082) as well as 11 genes of unknown function (s091, s093, s063, s089, s088, s068, s069, s092, s072, s083, s084) are also part of the conserved soft core. Although int and xis are annotated in all sampled SXT/R391 ICEs, both genes are on June 28, 2018 by UNIVERSITE DE SHERBROOKE missing in the soft core (Fig. 1). This absence results from the low sequence identity between proteins encoded by the SXT-type int or xis genes and those encoded by ICEVchBan8-type int or xis genes. Int encoded by SXT is only 27% and 25% identical to Int encoded by ICEValA056-2 and ICEVchBan8, respectively, whereas these two alter- native integrases share 90% identity. Atypical ICEs integrated at the same insertion site as VPI-2. Eighteen of the 68 ICEs were found to carry an atypical site-specific recombination module and to be located within the 3= end of a tRNA-Ser gene. Comparison of the attachment sites (attL and attR) that flanked these ICEs revealed that site-specific recombination for integra- tion took place within a 14-bp repeated sequence (5=-CTCACTCACCGCCA-3=) corre- sponding to the 3= end of the tRNA gene (Fig. 2B). In toxicogenic V. cholerae O1 and O139 isolates, this locus is the insertion site of Vibrio pathogenicity island 2 (VPI-2), which encodes sialic acid release, transport, and catabolism and a neuraminidase (21). VPI-2 is flanked by a 22-bp direct repeat (22) containing the same 14-bp conserved sequence (Fig. 2B). Although VPI-2 and atypical ICEs share the same integration site, their integrases share only 60% identity, thereby ruling out the emergence of atypical ICEs through direct substitution of the prfC recombination module of SXT/R391 ICEs by the tRNA-Ser recombination module of VPI-2. This idea is supported by Boyd et al. (23), who demonstrated that genomic island integrases are distinct from those encoded by phages, plasmids, integrons, and ICEs. Four types of SXT/R391 ICEs. Phylogenetic analysis of the 18 atypical integrases suggests the existence of three types with two alternative ICE configurations (Fig. 2A and B). Therefore, SXT/R391 ICEs fall into four distinct types differing in insertion site, structural organization, prevalence, and distribution in bacterial species. Type 1 in- cludes all ICEs inserted at the 5= end of prfC, like SXT from V. cholerae O139 MO10. Type 1 ICEs are found in both Enterobacteriaceae and Vibrionaceae of clinical and environ- mental origin (Fig. 3A; Table 1). Type 1 ICEs of clinical origin are strongly associated with antibiotic resistance, significantly more than with heavy metal resistance (Fig. 3C and D). Interestingly, elements carrying multiple antibiotic resistance genes tend to carry fewer heavy metal resistance genes and vice versa (Fig. 3B). This observation likely reflects the selection of the most beneficial traits for a specific ecological niche. Thus, isolates from clinical settings have higher counts of antibiotic resistance genes (Fig. 3C),

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 3 Bioteau et al. Applied and Environmental Microbiology

TABLE 1 SXT/R391 ICEs used in the study Predicted resistance GenBank Species Strain Origin Yr ICE name Ta Size (bp)b Gc gene(s)d Referencee accession no. Actinobacillus MIDG3553 Pneumonic lung of a 2012 ICEApl2 1R 92,660 ϩ sul2, floR, strAB, 9 MF187965 pleuropneumoniae pig, UK dfrA1 Alteromonas macleodii D7 Adaman Sea, Thailand 2000 ICEAmaD7 1S 117,055 ϩ merTPCA, CP014323 acrAB-tolC, zitB, zntA, copA

Alteromonas mediterranea MED64 Aegean Sea, Lebanon 2000 ICEAmaMED64 1S 80,715 ϩϪ CP004848 Downloaded from Escherichia coli HVH 177 Human blood, 2003 ICEEcoHVH177 1S 89,125 ϩϪ AZJM01000017 Denmark Idiomarinaceae bacterium HL-53 Unknown ND ICEIbaHL53 1S 71,962 ϩϪ LN899469 Methylophaga frappieri JAM7 Seawater treatment ND ICEMfrJAM7 1S 109,780 ϩϪ CP003380 plant, Montreal Biodome, Canada Photobacterium damselae PC554.2 Senegalese sole, 2003 ICEPdaSpa1 1S 102,985 ϩ tetA 10 AJ870986 Galicia, Spain

Proteus mirabilis TJ1809 Stool, Tianjin, China 2013 ICEPmiCHN1809 1R 76,218 ϩϪ 36 KX243413 09MAS2407 Stool, Maansham, 2009 ICEPmiCHN2407 1R 97,078 ϩ tetA, merTPCA, 36 KX243405 http://aem.asm.org/ China czcD 09MAS2410 Stool, Maansham, 2009 ICEPmiCHN2410 1R 93,537 ϩ merTPCA, czcD 36 KX243406 China 09MAS2416 Stool, Maansham, 2009 ICEPmiCHN2416 1R 92,556 ϩ merTPCA, czcD 36 KX243407 China TJ3237 Stool, Tianjin, China 2013 ICEPmiCHN3237 1S 87,215 ϩ 36 KX243414 TJ3300 Stool, Tianjin, China 2013 ICEPmiCHN3300 1R 108,335 ϩ sul2, floR, strAB, 36 KX243415 dfrG TJ3335 Stool, Tianjin, China 2013 ICEPmiCHN3335 1S 89,996 ϩ sul2, floR, strAB, 36 KX243416 tetA MD20140901 Stool, Beijing, China 2014 ICEPmiCHN901 1S 89,493 ϩ sul2, floR, strAB, 36 KX243408 dfrG MD20140902 Stool, Beijing, China 2014 ICEPmiCHN902 1S 89,096 ϩ sul2, floR, strAB 36 KX243409 on June 28, 2018 by UNIVERSITE DE SHERBROOKE dfrG MD20140903 Stool, Beijing, China 2014 ICEPmiCHN903 1S 89,644 ϩ sul2, floR, strAB, 36 KX243410 dfrG MD20140904 Stool, Beijing, China 2014 ICEPmiCHN904 1S 94,942 ϩ sul2, floR, strAB, 36 KX243411 tetA, czcD MD20140905 Stool, Beijing, China 2014 ICEPmiCHN905 1S 94,956 ϩ sul2, floR, strAB, 36 KX243412 tetA, czcD PM13C04 Chicken fecal sample, 2013 ICEPmiChn1 1S 92,752 ϩ sul2, floR, strAB, 32 KT962845 Hubei, China tetA, czcD ϩ PM14C28 Chicken fecal sample, 2014 ICEPmiJpn1 1S 91,091 blaCMY-2 30 KT894734 Heibei, China HI4320 Human urine, MA, USA 1986 ICEPmiUSA1 1S 79,733 ϩϪ 10 AM942759 ϩ Unnamed India 1977 R997 1S 85,368 blaHMS-1, czcD 61 KY433363.1

Proteus vulgaris 08MAS2213 Food, Maansham, 2008 ICEPvuCHN2213 1S 94,340 ϩ sul2, floR, strAB 36 KX243403 China Providencia alcalifaciens P-18 Bangladesh 1999 ICEPalBan1 1R 96,586 ϩ sul2, floR, strAB, 10 GQ463139 dfrA1 Providencia rettgeri 107 Pretoria, South Africa 1967 R391 1R 88,532 ϩ aph, merTPCA 62 AY090559 Providencia stuartii ATCC 33672 ATCC, unknown ND ICEPst33672 1R 75,588 ϩ merTPCA CP008920 Shewanella decolorationis S12 Wastewater treatment 2012 ICESdeCHNS12 1S 70,735 ϩϪ AXZL01000060 plant, Guangdong, China Shewanella sp. W3-18-1 Pacific Ocean 2000 ICESpuPO1 1S 108,623 ϩ acrAB-tolC, zitB, 10 CP000503 zntA, copA, czcA, czcD

Vibrio alginolyticus A056 Whiteleg shrimp, 2003 ICEValA056-1 1S 89,004 ϩ sul2, strAB, 20 KR231688 Guangdong, China acrAB-tolC ICEValA056-2 2R 103,826 ϩϪ 20 KR231689 E0601 Seawater, Guangdong, 2006 ICEValE0601 1R 106,165 ϩϪ 20 KT072768 China HN396 Seawater, Guangxi, 2008 ICEValHN396 2R 86,687 ϩϪ 20 KT072770 China HN437 Seawater, Hainan, 2008 ICEValHN437 2* 94,920 ϩ ugd 20 KT072771 China HN492 Seawater, Hainan, 2008 ICEValHN492 1R 106,164 ϩ 20 KT072769 China ZJ-T Orange-spotted 2005 ICEValZJT1 1S Ͼ88,684 sul2, strAB, CP016224 grouper, acrAB-tolC, Guangdong, China qnrVC5 ICEValZJT2 2R Ͼ100,515 CP016224

Vibrio cholerae O1 El Tor 2055/98 Bangladesh 1998 ICEVchBan5 1R 102,131 ϩ sul2, floR, strAB, 10 GQ463140 dfrA1 (Continued on next page)

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 4 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology

TABLE 1 (Continued) Predicted resistance GenBank Species Strain Origin Yr ICE name Ta Size (bp)b Gc gene(s)d Referencee accession no. MJ-1236 Matlab, Bangladesh 1994 ICEVchBan9 1R 106,124 ϩ sul2, floR, strAB, 10 CP001485 tetA, dfrA1, zitB ICDC-2605 Patient, Guizhou, China 1998 ICEVchCHN2605 1S 98,458 ϩ sul2, floR, strAB, 37 KT151661 dfrG ICDC-4210 Patient, Jiangxi, China 1999 ICEVchCHN4210 1S 110,349 ϩ sul2, floR, strAB, 37 KT151662

tetA, dfrG Downloaded from ICDC-956 Environment, Liaoning, 2001 ICEVchCHN956 1S 109,322 ϩ sul2, floR, strAB, 37 KT151655 China tetA, dfrG wujiang-2 Patient, China ND ICEVchCHNwuji 1R 96,168 ϩ sul2, strAB, KT151664 tetA, dfrA1 CO943 Sevagram, India 1994 ICEVchInd5 1R 97,847 ϩ sul2, floR, strAB, 10 GQ463142 dfrA1 2010EL-1786 Patient, Artibonite, 2010 ICEVchHai1 1R 97,847 ϩ sul2, floR, strAB, CP003069 Haiti dfrA1 2012HC-25 Stool, Haiti 2012 ICEVch2012HC25 3R Ͼ108,600 ϩϪ JSTY01000047 JSTY01000048

V. cholerae O139 AS207 Clinical, Kolkata, India 1997 ICEVchInd4 1S 95,491 ϩ sul2, floR, strAB 10 GQ463141 http://aem.asm.org/ MO10 Clinical, Chennai, India 1992 SXT 1S 99,452 ϩ sul2, floR, strAB, 8 DS990138 dfr18

V. cholerae non-O1/O139 HC-36A1 Stool, Haiti 2010 ICEVchHai2 1S 82,795 ϩϪ 63 AXDR01000007 2012Env-25 Water, Haiti 2012 ICEVch2012Env25 3R Ͼ108,600 ϩϪ JSTE01000047 JSTE01000048 1-010118-075 Sewage, San Luis 2001 ICEVchMex1 1R 82,839 ϩϪ 10 GQ463143 Potosi, Mexico

V. cholerae O37 MZO-3 Clinical, Bangladesh 2001 ICEVchBan8 3R 105,790 ϩ acrAB-tolC 10 JQ345361 V. cholerae O77 8-76 Diarrhea, India 1976 ICEVch8-76-1 1S Ͼ102,800 ND JIDN01000032 JIDN01000021 on June 28, 2018 by UNIVERSITE DE SHERBROOKE ICEVch8-76-2 3R Ͼ122,200 ND JIDN01000012 JIDN01000016

V. cholerae YB8E08 Oyster Pond, MA, USA 2009 ICEVchYB8E08 3R Ͼ104,300 ϩ ND LBGN01000012 LBGN01000009 OYP6E07 Oyster Pond, MA, USA 2009 ICEVchYP6E07 3* 93,439 ϩ acrAB-tolC NMTB01000014 3272-78 Water, MD, USA 1977 ICEVch3272-78 3R 104,112 ϩ acrAB-tolC MIOZ01000052 3223-74 Storm drain, Guam 1974 ICEVch3223-74 3R 104,112 ϩ acrAB-tolC MIZG01000083

Vibrio fluvialis H-08942 Infant diarrhea, 2002 ICEVflInd1 1S 102,862 ϩ sul2, floR, strAB, 10 KM213605 Kolkata, India dfr18 Vibrio mytili CAIM528 Seawater, Spain 1985 ICEVmyCAIM528 2R Ͼ63,900 Ϫ JXOK01000005 JXOK01000034

Vibrio parahaemolyticus CHN25 Shrimps, Shanghai, 2011 ICEVpaCHN25 1R 88,331 ϩ sul2, strAB, CP010883 China tetA, dfrG S163 Seafood, Malaysia 2007 ICEVpaS163 2R Ͼ75,600 ϩϪ AWHQ01000004 AWHQ01000048 S167 Environment, China 2007 ICEVpaS167 2R Ͼ82,000 ϩϪ AWHM01000024 AWHM01000222 ϩ UCM-V493 Sediment, Spain 2002 ICEVpaUCM493 1S 111,578 blaHMS-1 CP007004 S107-1 Oyster, Ladysmith 2005 ICEVpaCan1 1R 81,255 ϩ CP028481 Harbor, BC, Canada

Vibrio vulnificus SC9729 Seawater, South Korea 2011 ICEVvuSC9729 4S Ͼ85,200 ϩ cbiO JZEQ01000086 JZEQ01000106 CladeA-yb158 Tilapia fish, Israel 2005 ICEVvuCladeA158 4R Ͼ87,800 araJ LBNN01000013 LBNN01000014 CG100 Oyster, Taiwan 1993 ICEVvuCG100 4S Ͼ105,300 araJ PDGD01000031 PDGD01000050 a Type of SXT/R391 ICEs corresponds to the integrase type (1 for intprfC and 2, 3, and 4 for inttRNA-Ser) and entry exclusion groups (S or R). An ambiguous entry exclusion group is indicated with an asterisk. Entry exclusion groups were inferred from phylogenetic analyses presented in Fig. 5 and 6. bFor ICEs scattered over two contigs extracted from WGS data, an estimation of the minimum size is provided. cICE included in the Get_Homologs analysis. Other ICEs were excluded due to poor sequencing quality or missing data. dND, not determined; Ϫ, no resistance gene. eReferences are provided for ICEs that have already been described elsewhere.

while isolates originating from wastewater or marine or freshwater sediments are more likely to be exposed to heavy metals such as zinc, cobalt, and cadmium. Type 2, 3, and 4 ICEs are all inserted at the 3= end of the tRNA-Ser gene and are found exclusively in isolates of Vibrio species that are mostly of environmental origin (Fig. 3A; Table 1). These ICEs do not carry antibiotic resistance genes but commonly encode a heavy metal efflux pump, which seems to be a distinctive trait of atypical ICEs

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 5 Bioteau et al. Applied and Environmental Microbiology Downloaded from

FIG 1 Genetic map of the soft core of conserved genes of 61 SXT/R391 ICEs. The soft core map is drawn to scale and based on conservation of genes in at least 57 of the 61 ICE representatives used in the analysis. Positions of the promoters are based on the work of Poulin-Laprade et al. (17, 18). Numbered loci in brackets contain shell genes. Locus 1 includes rumAB and its associated ISCR2 antibiotic resistance element, s024-s026, and the hot spot 5 (HS5) as described by Wozniak et al. (10). Loci 2, 3, and 4 correspond to variable DNA inserted at HS1, HS2, and HS4, respectively (10). Loci 5 and 6 correspond to orfZ and s070, respectively (15, 64). Locus 7 corresponds to s073 and variable DNA inserted at HS3, such as a trimethoprim resistance-conferring integron (10) or diguanylate cyclase genes (65). Locus 8 corresponds to the mercury resistance genes in R391.

(Fig. 3D). Type 2 ICEs, such as ICEValA056-2 from V. alginolyticus A056, retained the http://aem.asm.org/ overall structure of SXT/R391 ICEs despite the alternative integration module. ICEVch- Ban8 from V. cholerae O37 MZO-3 provides an example of type 3 ICEs. In addition to the alternative integration module, type 3 ICEs hold another major rearrangement consist- on June 28, 2018 by UNIVERSITE DE SHERBROOKE

FIG 2 Alternative integration/excision modules. (A) Structural comparison of the region surrounding the regulatory module of SXT from V. cholerae O139 MO10, ICEValA056-2 from V. alginolyticus A056, and ICEVchBan8 from V. cholerae O37 MZO-3. Maps of circularized ICEs are drawn to scale. Color keys are the same as for Fig. 1, except for gray and brown, indicating integration/excision, and black, indicating transposition. (B) Maximum likelihood phylogenetic analysis of 20 integrases targeting tRNA-Ser. The integrases of VPI-2 from V. cholerae N16961 (locus VC1758) and VPI-8-76 from V. cholerae O77 (locus DA89_2501) that target the same tRNA-Ser gene were used as the outgroup. Bootstrap supports, as percentages, are indicated at the branching points only when Ͼ80%. Branch length represents the number of substitutions per site over 406 amino acid positions. The genetic context of int genes in each taxon is represented on the right side of the tree (refer to panel A for color coding). An asterisk indicates the presence of a frameshift in int of ICEVmyCAIM528. Alignment of the attL and attR sites for each corresponding element is also indicated. Shaded sequences correspond to the 3= end sequence of the serine tRNA gene. Nucleotides in bold show identity, while colors represent sequences internal to the elements, color coded by types. The region in which site-specific recombination likely takes place is indicated by a box.

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 6 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology Downloaded from http://aem.asm.org/ on June 28, 2018 by UNIVERSITE DE SHERBROOKE

FIG 3 Distribution of ICEs and resistances. (A) Distribution of ICE types for each isolation niche. (B) Prevalence of heavy metal and antibiotic resistance genes. (C) Resistance profiles by niche of isolation. (D) Resistance profiles according to ICE types. Statistical analyses were performed using the Wilcoxon matched-pairs signed-rank test (two-tailed) to compare the medians of each resistance type for each isolation niche (animal, P ϭ 0.1328; clinical, P Ͻ 0.0001; environmental, P ϭ 0.2227) and according to the type (type 1, P ϭ 0.0002; atypical, P ϭ 0.0078). ing in an inversion of the srpR-to-traG region (Fig. 2A). Currently, type 3 ICEs have been found only in V. cholerae, including environmental isolates recovered in the 1970s. Therefore, the emergence of type 3 ICEs in V. cholerae predates the spread of type 1 ICEs in toxigenic strains O1 and O139 of V. cholerae that began in the mid-1980s (4). Finally, three ICEs all found in Vibrio vulnificus belong to type 4. Although these ICEs retain the same structure as type 2, namely, lacking the large srpR-traG inversion, their integrases share the same common ancestor as type 3 ICEs. This suggests that types 3 and 4 would derive from a common ancestor sharing the same genetic structure as type 2 with a subsequent srpR-traG DNA inversion in the type 3 lineage. However, further analysis based on the soft-core genes suggests that V. vulnificus ICEs strongly diverge from other SXT/R391 ICEs (see below). Therefore, their type 4 integrase might be a more recent acquisition.

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 7 Bioteau et al. Applied and Environmental Microbiology Downloaded from http://aem.asm.org/

FIG 4 The integrase gene of atypical ICEs is under the control of SetCD. (A) Alignment of predicted SetCD-dependent promoters upstream of int in ICEs targeting the tRNA-Ser gene. The SetCD-binding logo and SetCD-dependent promoters of xis and srpM of SXT were characterized previously (17). SetCD boxes are shown in bold green capital letters. The position of known transcription start sites (TSS) is indicated in blue capital letters. Shine-Dalgarno sequences (SD) are underlined, while start codons are in capital letters. The approximate positions of the Ϫ35 and Ϫ10 regions are highlighted in gray. Numbers in brackets indicate the length in base pairs of spacers between the TSS region and the SD sequence. (B) The activity of two promoters representative of type 2 (ICEVch2012HC25) and 3 (ICEVpaS167) inttRNA-Ser genes was monitored from single-copy, chromosomally integrated lacZ transcriptional fusions in E. coli BW25113. Colorimetric

␤ on June 28, 2018 by UNIVERSITE DE SHERBROOKE assays of -galactosidase activity were carried out on LB medium supplemented with or without arabinose to express setCD from PBAD on pGG2B (64).

It is noteworthy that isolation biases exist in databases. For example, ICEs isolated in 2003 and 2005 were carried mostly by strains recovered from animals (see Fig. S1A in the supplemental material). Likewise, ICEs around the year 2000 belong mostly to type 1 (Fig. S1B). To better visualize how these elements cluster together based solely on the isolation perspective, a factorial analysis of mixed data (FAMD) revealed that ICEs found in clinical isolates belong mostly to type 1, whereas atypical ICEs are carried mostly by environmental strains (see Fig. S2 in the supplemental material). The int and xis genes of atypical ICEs remain under the control of the activator SetCD. The site-specific recombination module of type 1 SXT/R391 ICEs consists of two convergent genes, int and xis, each preceded by the SetCD-dependent promoters Pxis and PsrpM, respectively (Fig. 2A). In atypical ICEs, int and xis likely form an operon under the control of a new promoter, Pint, that replaces Pxis. Bioinformatics analysis revealed the presence of a putative SetCD binding motif located 176 to 178 bp upstream of the

GTG start codon of int (Fig. 4A). We introduced the promoter sequences Pint of type 2 ICEVpaS167 and type 3 ICEVch2012HC25 upstream of a promoterless lacZ reporter gene to test whether Pint is able to drive expression in a SetCD-dependent manner. ␤ -Galactosidase assays confirmed that, like Pxis of type 1 ICEs, Pint of type 2 and 3 ICEs are constitutively off and activated by SetCD (Fig. 4B). We observed a 36-fold change in ␤-galactosidase activity (9.6 Ϯ 6.3 arbitrary units [a.u.] versus 0.27 Ϯ 0.25 a.u.) for Ϯ Ϯ type 2 Pint and a 51-fold change (15.4 11.5 a.u. versus 0.30 0.26 a.u.) for type 3 Pint upon induction of setCD expression with 0.2% arabinose. Most atypical ICEs belong to the R entry exclusion group. Entry exclusion is a mechanism by which DNA transport from the donor cell is blocked by the recipient cell to prevent redundant exchange of conjugative elements between cells containing elements belonging to the same exclusion group. For SXT/R391 ICEs, entry exclusion is mediated by two inner membrane proteins: Eex in the recipient and TraG in the donor. SXT/R391 ICEs have been shown to belong to either the S (SXT) or R (R391) entry exclusion group (24). ICEs of the S group exclude the entry of ICEs of the S group but not those of the R group and vice versa. ICEs that have been previously examined

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 8 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology Downloaded from http://aem.asm.org/ on June 28, 2018 by UNIVERSITE DE SHERBROOKE

FIG 5 Maximum likelihood phylogenetic analysis of 37 Eex orthologue proteins. Bootstrap supports, as percentages, are indicated at the branching points only when Ͼ80%. Branch length represents the number of substitutions per site over 143 amino acid positions. Taxa corresponding to type 1, 2, 3, and 4 ICEs are shown in black, green, dark red, and orange, respectively. Taxa highlighted in blue correspond to type 1 ICEs found in strains also bearing a coresident type 2 or 3 ICE. Asterisks indicate ICEs with an ambiguous entry exclusion group. always code for matching pairs of Eex and TraG proteins belonging to the same exclusion group, i.e., EexS/TraGS or EexR/TraGR. Phylogenetic analyses of the Eex and TraG proteins were carried out to deter- mine the exclusion group of the 68 sampled ICEs. Eex proteins reliably clustered into the two main groups, S and R (Fig. 5). In contrast, TraG proteins did not provide a robust tree. While analysis of the conservation of an amino acid triplet at positions 606 to 608 allowed us to assign their exclusion group (PG[E/Q] for S, T[G/D]D for R) (24), TraG proteins of the same exclusion group did not cluster into consistent lineages (see Fig. S3 in the supplemental material). Considering that inter-ICE recombination events probably took place within traG and were blurring phyloge- netic relationships, we built a phylogenetic network by using the DNA sequence of

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 9 Bioteau et al. Applied and Environmental Microbiology Downloaded from http://aem.asm.org/ on June 28, 2018 by UNIVERSITE DE SHERBROOKE

FIG 6 NeigborNet phylogenetic network of 48 traG genes. Labels are color coded as described for Fig. 5. Asterisks indicate ICEs with an ambiguous entry exclusion group.

traG genes. This approach revealed a clear segregation between traG genes of the S and R groups (Fig. 6). Except for type 4 ICEVvuSC9729 and ICEVvuCG100, atypical ICEs tended to cluster within the R exclusion group (Fig. 5 and 6). Interestingly, type 2 ICEVchYP6E07 and type 3 ICEValHN437 are the only ICEs of our sample set that could not be assigned to a specific entry exclusion group. Unlike any other SXT/R391 ICE, both encode an EexR entry exclusion protein and a TraGS subunit, thereby suggesting that they should exclude the entry of ICEs of the S group but not those of the R group or themselves. However, type 2 ICEVchYP6E07 and type 3 ICEValHN437 should be excluded by recipients containing an ICE of the R group. Type 2 ICEs, which are found in multiple Vibrio species, are also uniformly distributed within the R exclusion group. In contrast, EexR proteins of type 3 ICEs from V. cholerae isolates recovered in the 1970s and 2000s exhibit little diversity and form a distinct lineage. The large traG-srpM inversion specific to the type 3 lineage might have isolated genetically these ICEs from other noninverted SXT/ R391 ICEs by decreasing the opportunities of inter-ICE recombination events through loss of local synteny around eex (Fig. 2A). However, this isolation does not seem to occur for traG,asICEVch2012Env25, ICEVch2012HC25 (traGR), and

ICEVchYP6E07 (traGS) strongly diverge from other type 3 ICEs. These divergences also correlate with the presence or absence of a complete copy of an insertion sequence of the IS481 family (Fig. 2B , tnp) that is truncated or absent in the other

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 10 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology type 3 ICEs. This observation suggests that the large traG-srpM inversion could have occurred more than once during the evolution of the type 3 lineage. Overall, we found that exclusion groups are evenly distributed in the different ecological niches among which the ICE-containing isolates were found (Fig. S1C). Therefore, there is no association between a specific niche and a specific exclusion group. Cohabitation of ICEs of distinct types in the same natural isolate. Natural occurrences of isolates containing multiple tandem copies of SXT/R391 ICEs are

strongly counterselected in the environment. Entry exclusion and RecA-dependent and Downloaded from -independent mechanisms prevent or resolve such occurrences (15, 16, 25). Homolo- gous recombination plays a key role in destabilizing tandem arrays of SXT/R391 ICEs, reducing them to a singleton (25, 26). By doing so, it also considerably enhances their diversity by generating new recombinant ICEs (4, 10, 15). Nevertheless, a change of integration site through substitution of the int and xis genes seems to facilitate the cohabitation of a type 1 ICE with another type, likely by lowering the opportunities of homologous recombination. Indeed, two V. alginolyticus isolates and one V. cholerae http://aem.asm.org/ isolate carry a combination of two ICEs: a type 1 ICE inserted at prfC, with either a type 2 or a type 3 ICE inserted at tRNA-Ser. In all instances, the type 1 ICE belongs to the S exclusion group whereas the second ICE belongs to the R exclusion group (Fig. 5 and 6). Atypical SXT/R391 ICEs form a polyphyletic group. To better understand the relationship between typical and atypical SXT/R391 ICEs, we constructed a phyloge- netic tree based on the alignment of the soft-core genes of 61 ICEs of our sample set

(Fig. 7). We found that atypical ICEs do not cluster as a distinct lineage, thereby on June 28, 2018 by UNIVERSITE DE SHERBROOKE indicating that they form a polyphyletic group. Type 4 ICEs, represented by ICEV- vuSC9729, seem to be the most divergent from all other types and constitute the outgroup of the tree. Furthermore, while a lineage of type 3 ICEs clearly diverges from all other ICEs (group 3a in Fig. 7), the other type 3 ICEs (group 3b in Fig. 7) are deeply rooted among type 1 ICEs. This observation is consistent with multiple occurrences of the large srpR-traG inversion. Except for ICEVpaS167, type 2 ICEs form a distinct cluster sharing a common ancestor with two type 1 ICEs found in V. alginolyticus. Altogether, these results indicate that atypical ICEs have undergone recombination events involv- ing type 1 ICEs. Whether type 4 ICEVvuSC9729, the most distant lineage, is still capable of participating in such events remains to be determined. Concluding remarks. Based on the current definition of the SXT/R391 family, int has been routinely used as the initial and often sole marker for PCR screening of SXT/R391 ICEs in epidemiological surveys of multidrug-resistant environmental and clinical iso- lates of several pathogenic species (27–39). Our comparative genomics analyses re- vealed that the core set of genes conserved in SXT/R391 ICEs is smaller than previously described (17). Like entry exclusion, int, xis, and integration sites are variable features of this family. Nevertheless, we showed that the integration and excision functions remain under the control of the transcriptional activator SetCD that governs the expression of the conjugation genes. The lack of conservation of int makes it unfit for detection of SXT/R391 ICEs at large; however, int remains a suitable marker in epidemiological studies, as we showed that multidrug-resistant clinical isolates are associated mostly with type 1 ICEs. Nevertheless, we propose to researchers to renounce the systematic use of int as the first or sole marker for detection of SXT/R391 ICEs. Instead, setCD should be considered a more robust and specific marker given its key role and strict conservation. setCD will allow the comprehensive detection of all types of SXT/R391 ICEs in bacterial samples.

MATERIALS AND METHODS Bacterial strains and media. Bacterial strains were routinely grown in lysogeny broth (LB-Miller; EMD) at 37°C in an orbital shaker/incubator and were preserved at Ϫ80°C in LB broth containing 15% (vol/vol) glycerol. Antibiotics were used at the following concentrations: ampicillin (Ap), 50 ␮g/ml (Vibrio) and 100 ␮g/ml (Escherichia coli); nalidixic acid (Nx), 40 ␮g/ml (E. coli); and kanamycin (Kn), 10 ␮g/ml (E. coli).

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 11 Bioteau et al. Applied and Environmental Microbiology Downloaded from http://aem.asm.org/ on June 28, 2018 by UNIVERSITE DE SHERBROOKE

FIG 7 Maximum likelihood phylogenetic analysis of the soft core of 61 ICEs. Branch length represents the number of substitutions per site over 24,943 nucleotide positions. All gaps were removed prior to analysis. Labels are color coded as described for Fig. 5. Asterisks indicate ICEs with an ambiguous entry exclusion group.

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 12 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology

TABLE 2 DNA sequences of the primers used in this study Primer name Nucleotide sequence (5= to 3=) pOP-VpaS167F TTTGCATCTTATTTTACTCTAAAAGTCGCACTGCAGATTCGACTCGAGCAGGA pOP-VpaS167R GTGTAGGTGGTTAGATAGATGAAAAAATCTAAGCTTGGCACTGGCCACGCA pOP-Vch12HCF TCCGCACTTCGCTTTACGCTTAGAGTCTCACTGCAGATTCGACTCGAGCAGGA pOP-Vch12HCR GTGTATGTGGCTAGATAGATAAAAGAATCTAAGCTTGGCACTGGCCACGCA setDuF ACAACCAATCARCAAGACTTCTATC setCuR ATTTAAGCTGAATGCGCTGTTG Downloaded from

Molecular biology. lacZ reporter fusions were constructed by introducing the promoter sequences of int of ICEVpaS167 and ICEVch2012HC25 with primer pairs pOP-VpaS167F/pOP-VpaS167R and pOP- Vch12HCF/pOP-Vch12HCR, respectively (Table 2), into the PstI restriction site of pOPlacZ using the Q5 site-directed mutagenesis kit (New England BioLabs) according to the manufacturer’s instructions. The resulting plasmids were confirmed by restriction profiling and DNA sequencing and then integrated in single copy into the chromosomal site attB␭ of E. coli BW25113 using pINT-Ts (40, 41). Sequencing reactions (except PacBio sequencing) were performed by the Plateforme de Séquençage et de Géno-

typage du Centre de Recherche du CHUL (Québec, QC, Canada). http://aem.asm.org/ ␤-Galactosidase assays. Qualitative assays on solid LB agar plate were done using 40 ␮g/ml 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside (X-Gal) as the substrate with or without 0.02% arabi- nose. Plates were observed after overnight incubation at 37°C and storage at 4°C. Quantitative liquid assays using o-2-nitrophenyl-␤-D-galactopyranoside (ONPG) as the substrate were done as previously described (42). Optical density at 420 nm (OD420) values were converted to log, and the slope of the linear regression was calculated for the early reaction. The slope was then normalized by the corre- sponding OD600 value and the reaction time. ICE detection with setCD markers. Primer pair setDuF and setCuR (Table 2) for setCD detection was designed using PrimerDesign-M (43) on a multiple-sequence alignment of setCD homologues. setCD loci were recovered from GenBank NT/NR and whole-genome shotgun contigs (WGS) databases restricted to on June 28, 2018 by UNIVERSITE DE SHERBROOKE Gammaproteobacteria (taxid:1236) using blastn with a 97% coverage threshold (44, 45). The resulting set of 521 sequences was aligned using the Muscle multiple-sequence alignment tool (46). A collection of 275 Canadian Vibrio strains (169 V. alginolyticus,2V. vulnificus, and 104 V. parahaemolyticus strains) were screened using the primers setDuF and setCuR to detect potential ICEs regulated by setCD. PCRs were carried out using the EasyTaq polymerase (Civic Bioscience) under the following reaction conditions: (i) 5 min at 94°C; (ii) 30 cycles of 30 s at 94°C, 30 s at the appropriate annealing temperature, and 30 s/kb at 72°C; and (iii) 1 min at 72°C. PacBio sequencing. Whole-genome sequencing of V. parahaemolyticus S107 containing ICEVpaCan1 was carried out from genomic DNA extracted from 2 ml of exponential-phase culture using the Gentra Puregene kit (Qiagen). PacBio RS II single-molecule real-time sequencing (Pacbio SMRTcell) and de novo genome assembly were performed at the McGill University and Génome Québec Innovation Centre. ICE data set and comparative analysis. ICE data sets were obtained by using the NCBI’s blastn algorithm against several databases (NT/NR, WGS, Refseq genomic) on different integrase types and three other key sets of genes (croS-setR, setCD, and sprR-sprM). When complete ICE sequences were not available, the best blast hits found across contigs were extracted from NCBI’s sequence set browser and were aligned against the adequate ICE type reference sequence using MUMmer3 (47). The following putative ICE draft sequences were then manually assembled and curated: ICEVch2012Env25, ICEVch2012HC25, ICEVch3223-74, ICEVch3272-78, ICEVchYB8E08, ICEVpaS163, ICEVpaS167, and ICEV- vuSC9729. Whole ICE sequences were scanned to detect antibiotic and heavy metal resistances using CARD (48). The sequences of 61 assembled ICEs were submitted to the “Get_Homologues” software package (49) (80% minimum coverage in blastn pairwise alignments) to obtain homology information and isolate the soft-core genome (95% conservation). The 43 isolated soft-core genes of each ICE were extracted, reorganized to be syntenic, and concatenated using in-house scripts into 61 streamlined ICE sequences prior to their alignments using Clustal Omega (50). Evolutionary analyses were conducted in MEGA7 (51) and PhyML (52) and inferred by using the maximum likelihood method based on the LG (InttRNA-Ser) or JTT (Eex or TraG proteins) matrix-based models (53, 54). Protein sequences were aligned with Muscle (46). Eex and TraG primary sequences were first clustered with CD-HIT (55) (sequence identity cutoff, 1) prior to alignment. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model (Eex or TraG proteins) and then selecting the topology with the superior log likelihood value. A discrete gamma distribution was used to model evolutionary rate differences among sites (5 categories) for Eex or TraG proteins. A NeighborNet phylogenetic network was built for traG using SplitsTree4 (56) with default parameters (Uncorrected_P method for distances and EqualAngle drawing method) after clustering of the nucleotide sequences of traG genes recovered from the ICE sample set using CD-HIT-est (55) (sequence identity cutoff, 1). The 48 unique DNA sequences were aligned with Muscle prior to phylogenetic analysis. Phylogenetic analysis of the 61 streamlined ICEs was inferred by using the maximum likelihood method based on the general time-reversible (GTR) model (57). A discrete gamma distribution was used to model evolutionary rate differences among sites (5 categories). The rate variation model allowed for some sites to be evolutionarily invariable ([ϩI], 39.0911% sites). Versions Pfam 31.0, Uniprot 2017_11, and HMMER v3.1b2 were used.

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 13 Bioteau et al. Applied and Environmental Microbiology

Statistical analyses. A factorial analysis of mixed data was performed on all 68 ICEs using 8 variables (origin, type, exclusion group, strain species, WHO region of isolation, year of isolation, number of antibiotic resistance genes, and number of heavy metal resistance genes). Data were imported in R version 3.4.3 (58). Missing values were handled with the missMDA package (59). The analysis itself was performed using FactoMineR package (60). Accession number(s). The whole-genome sequence of V. parahaemolyticus S107 was deposited in GenBank under accession numbers CP028481 and CP028482.

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/AEM Downloaded from .00485-18. SUPPLEMENTAL FILE 1, PDF file, 0.7 MB.

ACKNOWLEDGMENTS We thank Kévin Huguet for technical assistance and Swapan Banerjee (Health Canada) for access to his collection of Canadian Vibrio isolates. We are grateful to Fanie Pelletier for her help with statistical analyses and Alain Lavigueur for critical reading of http://aem.asm.org/ the manuscript. Computations were made on the supercomputer Mp2 from Université de Sher- brooke, managed by Calcul Québec and Compute Canada. The operation of this supercomputer is funded by the Canada Foundation for Innovation (CFI), the ministère de l’Économie, de la science et de l’innovation du Québec (MESI) and the Fonds de recherche du Québec—Nature et technologies (FRQ-NT). This work was supported by a Discovery Grant (2016-04365) from the Natural Sciences and Engineering Council of Canada (NSERC) and a Project Grant (PJT-153071) on June 28, 2018 by UNIVERSITE DE SHERBROOKE from the Canadian Institutes of Health Research (CIHR) to V.B.

REFERENCES 1. Ghosh A, Ramamurthy T. 2011. Antimicrobials & cholera: are we 11. Burrus V, Waldor MK. 2003. Control of SXT integration and excision. stranded. Indian J Med Res 133:225–231. J Bacteriol 185:5045–5054. https://doi.org/10.1128/JB.185.17.5045 2. Kitaoka M, Miyata ST, Unterweger D, Pukatzki S. 2011. Antibiotic resis- -5054.2003. tance mechanisms of Vibrio cholerae. J Med Microbiol 60:397–407. 12. Hochhut B, Waldor MK. 1999. Site-specific integration of the conjugal https://doi.org/10.1099/jmm.0.023051-0. Vibrio cholerae SXT element into prfC. Mol Microbiol 32:99–110. https:// 3. Burrus V, Marrero J, Waldor MK. 2006. The current ICE age: biology and doi.org/10.1046/j.1365-2958.1999.01330.x. evolution of SXT-related integrating conjugative elements. Plasmid 55: 13. Carraro N, Poulin D, Burrus V. 2015. Replication and active partition of 173–183. https://doi.org/10.1016/j.plasmid.2006.01.001. integrative and conjugative elements (ICEs) of the SXT/R391 family: the 4. Spagnoletti M, Ceccarelli D, Rieux A, Fondi M, Taviani E, Fani R, line between ICEs and conjugative plasmids is getting thinner. PLoS Colombo MM, Colwell RR, Balloux F. 2014. Acquisition and evolution Genet 11:e1005298. https://doi.org/10.1371/journal.pgen.1005298. of SXT-R391 integrative conjugative elements in the seventh- 14. Ceccarelli D, Daccord A, René M, Burrus V. 2008. Identification of the pandemic Vibrio cholerae lineage. mBio 5:e01356-14. https://doi.org/ origin of transfer (oriT) and a new gene required for mobilization of the 10.1128/mBio.01356-14. SXT/R391 family of integrating conjugative elements. J Bacteriol 190: 5. Beaber JW, Hochhut B, Waldor MK. 2004. SOS response promotes hori- 5328–5338. https://doi.org/10.1128/JB.00150-08. zontal dissemination of antibiotic resistance genes. Nature 427:72. 15. Garriss G, Waldor MK, Burrus V. 2009. Mobile antibiotic resistance en- https://doi.org/10.1038/nature02241. coding elements promote their own diversity. PLoS Genet 5:e1000775. 6. Dalia AB, Seed KD, Calderwood SB, Camilli A. 2015. A globally distributed https://doi.org/10.1371/journal.pgen.1000775. mobile genetic element inhibits natural transformation of Vibrio chol- 16. Marrero J, Waldor MK. 2005. Interactions between inner membrane erae. Proc Natl Acad SciUSA112:10485–10490. https://doi.org/10.1073/ proteins in donor and recipient cells limit conjugal DNA transfer. Dev pnas.1509097112. Cell 8:963–970. https://doi.org/10.1016/j.devcel.2005.05.004. 7. Ceccarelli D, Spagnoletti M, Bacciu D, Danin-Poleg Y, Mendiratta DK, Kashi 17. Poulin-Laprade D, Matteau D, Jacques P-É, Rodrigue S, Burrus V. 2015. Y, Cappuccinelli P, Burrus V, Colombo MM. 2011. ICEVchInd5 is prevalent in Transfer activation of SXT/R391 integrative and conjugative elements: epidemic Vibrio cholerae O1 El Tor strains isolated in India. Int J Med unraveling the SetCD regulon. Nucleic Acids Res 43:2045–2056. https:// Microbiol 301:318–324. https://doi.org/10.1016/j.ijmm.2010.11.005. doi.org/10.1093/nar/gkv071. 8. Waldor MK, Tschäpe H, Mekalanos JJ. 1996. A new type of conjugative 18. Poulin-Laprade D, Burrus V. 2015. A ␭ Cro-like repressor is essential for transposon encodes resistance to sulfamethoxazole, trimethoprim, and the induction of conjugative transfer of SXT/R391 elements in response streptomycin in Vibrio cholerae O139. J Bacteriol 178:4157–4165. https:// to DNA damage. J Bacteriol 197:3822–3833. https://doi.org/10.1128/JB doi.org/10.1128/jb.178.14.4157-4165.1996. .00638-15. 9. Li Y, Li Y, Fernandez Crespo R, Leanse LG, Langford PR, Bossé JT. 2018. 19. Taviani E, Spagnoletti M, Ceccarelli D, Haley BJ, Hasan NA, Chen A, Characterization of the Actinobacillus pleuropneumoniae SXT-related in- Colombo MM, Huq A, Colwell RR. 2012. Genomic analysis of ICEVchBan8: tegrative and conjugative element ICEApl2 and analysis of the encoded an atypical genetic element in Vibrio cholerae. FEBS Lett 586:1617–1621. FloR protein: hydrophobic residues in transmembrane domains contrib- https://doi.org/10.1016/j.febslet.2012.03.064. ute dynamically to florfenicol and chloramphenicol efflux. J Antimicrob 20. Luo P, He X, Wang Y, Liu Q, Hu C. 2016. Comparative genomic analysis Chemother 73:57–65. https://doi.org/10.1093/jac/dkx342. of six new-found integrative conjugative elements (ICEs) in Vibrio 10. Wozniak RAF, Fouts DE, Spagnoletti M, Colombo MM, Ceccarelli D, alginolyticus. BMC Microbiol 16:79. https://doi.org/10.1186/s12866-016 Garriss G, Déry C, Burrus V, Waldor MK. 2009. Comparative ICE genomics: -0692-9. insights into the evolution of the SXT/R391 family of ICEs. PLoS Genet 21. Jermyn WS, Boyd EF. 2002. Characterization of a novel Vibrio pathoge- 5:e1000786. https://doi.org/10.1371/journal.pgen.1000786. nicity island (VPI-2) encoding neuraminidase (nanH) among toxigenic

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 14 Inclusive Redefinition of the SXT/R391 Family of ICEs Applied and Environmental Microbiology

Vibrio cholerae isolates. Microbiology 148:3681–3693. https://doi.org/10 2006. Septaplex PCR assay for rapid identification of Vibrio cholerae .1099/00221287-148-11-3681. including detection of virulence and int SXT genes. FEMS Microbiol Lett 22. Carpenter MR, Rozovsky S, Boyd EF. 2015. Pathogenicity island cross talk 265:208–214. https://doi.org/10.1111/j.1574-6968.2006.00491.x. mediated by recombination directionality factors facilitates excision from the 40. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomal chromosome. J Bacteriol 198:766–776. https://doi.org/10.1128/JB.00704-15. genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 23. Boyd EF, Almagro-Moreno S, Parent MA. 2009. Genomic islands are 97:6640–6645. https://doi.org/10.1073/pnas.120163297. dynamic, ancient integrative elements in bacterial evolution. Trends 41. Haldimann A, Wanner BL. 2001. Conditional-replication, integration, ex- Microbiol 17:47–53. https://doi.org/10.1016/j.tim.2008.11.003. cision, and retrieval plasmid-host systems for gene structure-function 24. Marrero J, Waldor MK. 2007. The SXT/R391 family of integrative conju- studies of bacteria. J Bacteriol 183:6384–6393. https://doi.org/10.1128/ gative elements is composed of two exclusion groups. J Bacteriol 189: JB.183.21.6384-6393.2001. 3302–3305. https://doi.org/10.1128/JB.01902-06. 42. Carraro N, Durand R, Rivard N, Anquetil C, Barrette C, Humbert M, Downloaded from 25. Burrus V, Waldor MK. 2004. Formation of SXT tandem arrays and SXT- Burrus V. 2017. Salmonella genomic island 1 (SGI1) reshapes the R391 hybrids. J Bacteriol 186:2636–2645. https://doi.org/10.1128/JB.186 mating apparatus of IncC conjugative plasmids to promote self- .9.2636-2645.2004. propagation. PLoS Genet 13:e1006705. https://doi.org/10.1371/ 26. Pembroke JT, Murphy DB. 2000. Isolation and analysis of a circular form journal.pgen.1006705. of the IncJ conjugative transposon-like elements, R391 and R997: impli- 43. Yoon H, Leitner T. 2015. PrimerDesign-M: a multiple-alignment based cations for IncJ incompatibility. FEMS Microbiol Lett 187:133–138. multiple-primer design tool for walking across variable genomes. https://doi.org/10.1111/j.1574-6968.2000.tb09149.x. Bioinformatics 31:1472–1474. https://doi.org/10.1093/bioinformatics/ 27. Mohapatra H, Mohapatra SS, Mantri CK, Colwell RR, Singh DV. 2008. btu832. Vibrio cholerae non-O1, non-O139 strains isolated before 1992 from 44. Boratyn GM, Camacho C, Cooper PS, Coulouris G, Fong A, Ma N, Madden http://aem.asm.org/ Varanasi, India are multiple drug resistant, contain intSXT, dfr18 and TL, Matten WT, McGinnis SD, Merezhuk Y, Raytselis Y, Sayers EW, Tao T, aadA5 genes. Environ Microbiol 10:866–873. https://doi.org/10.1111/j Ye J, Zaretskaya I. 2013. BLAST: a more efficient report with usability .1462-2920.2007.01502.x. improvements. Nucleic Acids Res 41:W29–W33. https://doi.org/10.1093/ 28. Taviani E, Ceccarelli D, Lazaro N, Bani S, Cappuccinelli P, Colwell RR, nar/gkt282. Colombo MM. 2008. Environmental Vibrio spp., isolated in Mozambique, 45. NCBI Resource Collaborators. 2013. Database resources of the National contain a polymorphic group of integrative conjugative elements and Center for Biotechnology Information. Nucleic Acids Res 41:D8–D20. class 1 integrons. FEMS Microbiol Ecol 64:45–54. https://doi.org/10.1111/ https://doi.org/10.1093/nar/gks1189. j.1574-6941.2008.00455.x. 46. Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accu- 29. Ceccarelli D, Bani S, Cappuccinelli P, Colombo MM. 2006. Prevalence of racy and high throughput. Nucleic Acids Res 32:1792–1797. https://doi aadA1 and dfrA15 class 1 integron cassettes and SXT circulation in Vibrio .org/10.1093/nar/gkh340.

cholerae O1 isolates from Africa. J Antimicrob Chemother 58:1095–1097. 47. Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, Antonescu C, on June 28, 2018 by UNIVERSITE DE SHERBROOKE https://doi.org/10.1093/jac/dkl352. Salzberg SL. 2004. Versatile and open software for comparing large 30. Harada S, Ishii Y, Saga T, Tateda K, Yamaguchi K. 2010. Chromo- genomes. Genome Biol 5:R12. https://doi.org/10.1186/gb-2004-5-2

somally encoded blaCMY-2 located on a novel SXT/R391-related inte- -r12. grating conjugative element in a Proteus mirabilis clinical isolate. 48. Jia B, Raphenya AR, Alcock B, Waglechner N, Guo P, Tsang KK, Lago BA, Antimicrob Agents Chemother 54:3545–3550. https://doi.org/10 Dave BM, Pereira S, Sharma AN, Doshi S, Courtot M, Lo R, Williams LE, .1128/AAC.00111-10. Frye JG, Elsayegh T, Sardar D, Westman EL, Pawlowski AC, Johnson TA, 31. Rodríguez-Blanco A, Lemos ML, Osorio CR. 2016. Unveiling the pan- Brinkman FSL, Wright GD, McArthur AG. 2017. CARD 2017: expansion genome of the SXT/R391 family of ICEs: molecular characterisation of and model-centric curation of the comprehensive antibiotic resistance new variable regions of SXT/R391-like ICEs detected in Pseudoalteromo- database. Nucleic Acids Res 45:D566–D573. https://doi.org/10.1093/nar/ nas sp. and Vibrio scophthalmi. Antonie Van Leeuwenhoek 109: gkw1004. 1141–1152. https://doi.org/10.1007/s10482-016-0716-3. 49. Contreras-Moreira B, Vinuesa P. 2013. GET_HOMOLOGUES, a versatile 32. Lei C-W, Zhang A-Y, Wang H-N, Liu B-H, Yang L-Q, Yang Y-Q. 2016. Char- software package for scalable and robust microbial pangenome anal- acterization of SXT/R391 integrative and conjugative elements in Proteus ysis. Appl Environ Microbiol 79:7696–7701. https://doi.org/10.1128/ mirabilis isolates from food-producing animals in China. Antimicrob Agents AEM.02411-13. Chemother 60:1935–1938. https://doi.org/10.1128/AAC.02852-15. 50. Sievers F, Higgins DG. 2017. Clustal Omega for making accurate align- 33. Balado M, Lemos ML, Osorio CR. 2013. Integrating conjugative ele- ments of many protein sequences. Protein Sci Publ Protein Soc 27: ments of the SXT/R391 family from fish-isolated encode 135–145. https://doi.org/10.1002/pro.3290. restriction-modification systems that confer resistance to bacterio- 51. Kumar S, Stecher G, Tamura K. 2016. MEGA7: molecular evolutionary phages. FEMS Microbiol Ecol 83:457–467. https://doi.org/10.1111/ genetics analysis version 7.0 for bigger datasets. Mol Biol Evol 33: 1574-6941.12007. 1870–1874. https://doi.org/10.1093/molbev/msw054. 34. Badhai J, Kumari P, Krishnan P, Ramamurthy T, Das SK. 2013. Presence 52. Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, Gascuel O. of SXT integrating conjugative element in marine bacteria isolated 2010. New algorithms and methods to estimate maximum-likelihood from the mucus of the coral Fungia echinata from Andaman Sea. phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: FEMS Microbiol Lett 338:118–123. https://doi.org/10.1111/1574-6968 307–321. https://doi.org/10.1093/sysbio/syq010. .12033. 53. Jones DT, Taylor WR, Thornton JM. 1992. The rapid generation of 35. Rodríguez-Blanco A, Lemos ML, Osorio CR. 2012. Integrating conju- mutation data matrices from protein sequences. Bioinformatics gative elements as vectors of antibiotic, mercury, and quaternary 8:275–282. https://doi.org/10.1093/bioinformatics/8.3.275. ammonium compound resistance in marine aquaculture environ- 54. Le SQ, Gascuel O. 2008. An improved general amino acid replacement ments. Antimicrob Agents Chemother 56:2619–2626. https://doi.org/ matrix. Mol Biol Evol 25:1307–1320. https://doi.org/10.1093/molbev/ 10.1128/AAC.05997-11. msn067. 36. Li X, Du Y, Du P, Dai H, Fang Y, Li Z, Lv N, Zhu B, Kan B, Wang D. 2016. 55. Huang Y, Niu B, Gao Y, Fu L, Li W. 2010. CD-HIT Suite: a web server for SXT/R391 integrative and conjugative elements in Proteus species reveal clustering and comparing biological sequences. Bioinforma Oxf Engl abundant genetic diversity and multidrug resistance. Sci Rep 6:37372. 26:680–682. https://doi.org/10.1093/bioinformatics/btq003. https://doi.org/10.1038/srep37372. 56. Huson DH, Bryant D. 2006. Application of phylogenetic networks in 37. Wang R, Yu D, Yue J, Kan B. 2016. Variations in SXT elements in epidemic evolutionary studies. Mol Biol Evol 23:254–267. https://doi.org/10.1093/ Vibrio cholerae O1 El Tor strains in China. Sci Rep 6:22733. https://doi molbev/msj030. .org/10.1038/srep22733. 57. Nei M, Kumar S. 2000. Molecular evolution and phylogenetics. Oxford 38. Song Y, Yu P, Li B, Pan Y, Zhang X, Cong J, Zhao Y, Wang H, Chen L. 2013. University Press, Oxford, England. The mosaic accessory gene structures of the SXT/R391-like integrative 58. R Core Team. 2016. R: a language and environment for statistical com- and conjugative elements derived from Vibrio spp. isolated from aquatic puting. R Foundation for Statistical Computing, Vienna, Austria. products and environment in the Yangtze River estuary, China. BMC 59. Josse J, Husson F. 2016. missMDA: a package for handling missing values Microbiol 13:214. https://doi.org/10.1186/1471-2180-13-214. in multivariate data analysis. J Stat Softw 70:1–31. https://doi.org/10 39. Mantri CK, Mohapatra SS, Ramamurthy T, Ghosh R, Colwell RR, Singh DV. .18637/jss.v070.i01.

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 15 Bioteau et al. Applied and Environmental Microbiology

60. Lê S, Josse J, Husson F. 2008. FactoMineR: an R package for multi- 2013. A new integrative conjugative element detected in Haitian isolates variate analysis. J Stat Softw 25:1–18. https://doi.org/10.18637/jss of Vibrio cholerae non-O1/non-O139. Res Microbiol 164:891–893. https:// .v025.i01. doi.org/10.1016/j.resmic.2013.08.004. 61. Ryan MP, Armshaw P, O’Halloran JA, Pembroke JT. 2017. Analysis and 64. Garriss G, Poulin-Laprade D, Burrus V. 2013. DNA-damaging agents comparative genomics of R997, the first SXT/R391 integrative and con- induce the RecA-independent homologous recombination functions of jugative element (ICE) of the Indian sub-continent. Sci Rep 7:8562. integrating conjugative elements of the SXT/R391 family. J Bacteriol https://doi.org/10.1038/s41598-017-08735-y. 195:1991–2003. https://doi.org/10.1128/JB.02090-12. 62. Böltner D, MacMahon C, Pembroke JT, Strike P, Osborn AM. 2002. R391: 65. Bordeleau E, Brouillette E, Robichaud N, Burrus V. 2010. Beyond antibi- a conjugative integrating mosaic comprised of phage, plasmid, and otic resistance: integrating conjugative elements of the SXT/R391 family transposon elements. J Bacteriol 184:5158–5169. https://doi.org/10 that encode novel diguanylate cyclases participate to c-di-GMP signal- .1128/JB.184.18.5158-5169.2002. ling in Vibrio cholerae. Environ Microbiol 12:510–523. https://doi.org/10 Downloaded from 63. Ceccarelli D, Spagnoletti M, Hasan NA, Lansing S, Huq A, Colwell RR. .1111/j.1462-2920.2009.02094.x. http://aem.asm.org/ on June 28, 2018 by UNIVERSITE DE SHERBROOKE

July 2018 Volume 84 Issue 13 e00485-18 aem.asm.org 16 ANNEXE 2

Tableau A2.1. Liste des amorces utilisées

Nom Séquencea 94-086_WF TGCGTTTATCTGTACCTCAACTTAGGAGTATCTCTAgtgtaggctggagct gcttcg 94-086_WR GCAAAGGTTTGGGCTGGGTTATGCTGCTTTTGTTTTcatatgaatatcctcct ta SGI1dels004.for TAAATCGCCTAGGTGCCCAAATAGGGCACTTCCAGAGGCCgtgtagg ctggagctgcttc SGI1dels004.rev CCGCCTGATACTGAGCAAGTAGTTGGCTAAATGGCTTcatatgaatatcct cctta SGI1del010.for TGTATCAAGGTATTTGTCCGTAGCCAAAGGAGGCCGTTAAgtgtaggc tggagctgcttc SGI1del010.rev AAAAACAAAGGCTCCTGAGAGCCTTTGTTTTTATTCATTA TTAcatatgaatatcctcctta SGI1del019.for TGGTCGCATCGATGTGACAAATAGTTATTTGGGGTAATTGgtgtaggct ggagctgcttc SGI1del019.rev AAGGTCTTATTAAGGAATATGACATAGGCGTAGATGGCTTcatatgaa tatcctcctta SGI1del020.for AACAAAGTAACAATTTTGATTAACAGAGTTAGGGGGATCAgtgtagg ctggagctgcttc SGI1del020.rev CATAAACATGGCTTCTTTCTCTTACCCATTTTTCCATCAAcatatgaatat cctcctta 94-086_CF GTACGAATTCTTAGGAGTATCTCTAATGGCAG 94-086_CR AGGTGAATTCTGGGTTATGCTGCTTTTGTT mpsABEcoRIf NNNNGAATTCAAGGAGGAATAATAAATGCGTTCAGAGCGGACT mpsABSalIr NNNGTCGACTTAATCAGCAGAGCCGGTG SGI1S010EcoRI.for NNNNNNGAATTCAAGGAGGAATAATAAATGAGCAAACGAGCTG AATA SGI1S010EcoRI.rev NNNNNNGAATTCTTATAGACTTAACGTTAACTGCCTA traI907-992_EcoRI.f NNNNNNGAATTCAAGGAGGAATAATAAATGCCGTCTGGTGTTA traI907-992_EcoRI.r NNNNNNGAATTCTCATGTTTCATTTACCTCTAAATA TraISDMf ACACATCCTCATAGGGATGCTGAGCCG TraISDMr TTGCTGATGCCGATGTTTTC SGI1promtraHPstI.re NNNNNNCTGCAGTTAAAGCTCCTCTTTTAGAA v SGI1s010PstIrev NNNNNNCTGCAGTTATAGACTTAACGTTAACTGCCTA Vcrx086XbaIf NTCTAGAAATGGCAGAAGCACAAG SGI1S010XbaIf NTCTAGAAATGAGCAAACGAGCTGAA 94MobIXbaIf NTCTAGAGGTGAGTCTACCAACAGAGCGATG 94mobIEcoRI.rev NNNNNNGAATTCTCACACCTCGTCGCTATGTGTCTT

214 SGI1TraGBamHIf NNNGGATCCAATGGATTTTAGTATTTATTCCGTCGGT SGI1TraGKpnIr NNGGTACCTTAACGGCCTCCTTTGGCTA dusAigEcoRIF NNNNGAATTCACAAGTTATCGCTCTATCACTG dusAigEcoRIR NNNNGAATTCCCTTTGATGTGGGGCATG EcodusAattLf CAGCCGACATTCAGGTTG dusAattLr AAGCGAATGATGCCTTTACTG dusAattRf GTGTGTGTAGCTTCAGGTG EcodusAattRr GTCGAACCTGGATTGTTTATCATTG dusAscarNoFRTf CAGCAGTCCTATCATGCCCCACATCAAAGGGAATTCagagcgcttttgaag ctca dusAscarNoFRTr GATCATAATCAATATCGATGGAGAAAAGCAATGACAtcggaataggaac ttcaaga

aLes sites P1 et P2 hybridant sur pKD3 sont indiqués en minuscules. Les sites de restriction sont indiqués en gras. Les nucléotides introduisant une mutation ou une insertion sont soulignés.

215 BIBLIOGRAPHIE

Ahmad, I., Wigren, E., Guyon, S.L., Vekkeli, S., Blanka, A., Mouali, Y. el, Anwar, N., Chuah, M.L., Lünsdorf, H., Frank, R., et al. (2013). The EAL-like protein STM1697 regulates virulence phenotypes, motility and biofilm formation in Salmonella typhimurium. Molecular Microbiology 90, 1216–1232.

Alperi, A., Larrea, D., Fernández-González, E., Dehio, C., Zechner, E.L., and Llosa, M. (2013). A Translocation Motif in Relaxase TrwC Specifically Affects Recruitment by Its Conjugative Type IV Secretion System. J. Bacteriol. 195, 4999–5006.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990). Basic local alignment search tool. Journal of Molecular Biology 215, 403–410.

Ambrose, S.J., Harmer, C.J., and Hall, R.M. (2018). Compatibility and entry exclusion of IncA and IncC plasmids revisited: IncA and IncC plasmids are compatible. Plasmid.

Amieva, M.R., Vogelmann, R., Covacci, A., Tompkins, L.S., Nelson, W.J., and Falkow, S. (2003). Disruption of the Epithelial Apical-Junctional Complex by Helicobacter pylori CagA. Science 300, 1430–1434.

Anthony, K.G., Klimke, W.A., Manchak, J., and Frost, L.S. (1999). Comparison of Proteins Involved in Pilus Synthesis and Mating Pair Stabilization from the Related Plasmids F and R100-1: Insights into the Mechanism of Conjugation. J. Bacteriol. 181, 5149–5159.

Arcari, G., Di Lella, F.M., Bibbolino, G., Mengoni, F., Beccaccioli, M., Antonelli, G., Faino, L., and Carattoli, A. (2020). A Multispecies Cluster of VIM-1 Carbapenemase-Producing Enterobacterales Linked by a Novel, Highly Conjugative, and Broad-Host-Range IncA Plasmid Forebodes the Reemergence of VIM-1. Antimicrob Agents Chemother 64, e02435-19, /aac/64/4/AAC.02435-19.atom.

Arutyunov, D., and Frost, L.S. (2013). F conjugation: Back to the beginning. Plasmid 70, 18– 32.

Atmakuri, K., Cascales, E., and Christie, P.J. (2004). Energetic components VirD4, VirB11 and VirB4 mediate early DNA transfer reactions required for bacterial type IV secretion. Mol Microbiol 54.

Auchtung, J.M., Lee, C.A., Monson, R.E., Lehman, A.P., and Grossman, A.D. (2005). Regulation of a Bacillus subtilis mobile genetic element by intercellular signaling and the global DNA damage response. PNAS 102, 12554–12559.

216

Audette, G.F., Manchak, J., Beatty, P., Klimke, W.A., and Frost, L.S. (2007). Entry exclusion in F-like plasmids requires intact TraG in the donor that recognizes its cognate TraS in the recipient. Microbiology 153, 442–451.

Bacon, D.J., Alm, R.A., Burr, D.H., Hu, L., Kopecko, D.J., Ewing, C.P., Trust, T.J., and Guerry, P. (2000). Involvement of a Plasmid in Virulence of Campylobacter jejuni 81-176. Infect Immun 68, 4384–4390.

Banta, L.M., Kerr, J.E., Cascales, E., Giuliano, M.E., Bailey, M.E., McKay, C., Chandran, V., Waksman, G., and Christie, P.J. (2011). An Agrobacterium VirB10 Mutation Conferring a Type IV Secretion System Gating Defect. J. Bacteriol. 193, 2566–2574.

Beaber, J.W., Hochhut, B., and Waldor, M.K. (2002). Genomic and Functional Analyses of SXT, an Integrating Antibiotic Resistance Gene Transfer Element Derived from Vibrio cholerae. Journal of Bacteriology 184, 4259–4269.

Beaber, J.W., Hochhut, B., and Waldor, M.K. (2004). SOS response promotes horizontal dissemination of antibiotic resistance genes. Nature 427, 72.

Bellanger, X., Morel, C., Decaris, B., and Guédon, G. (2007). Derepression of Excision of Integrative and Potentially Conjugative Elements from Streptococcus thermophilus by DNA Damage Response: Implication of a cI-Related Repressor. Journal of Bacteriology 189, 1478– 1481.

Beranek, A., Zettl, M., Lorenzoni, K., Schauer, A., Manhart, M., and Koraimann, G. (2004). Thirty-Eight C-Terminal Amino Acids of the Coupling Protein TraD of the F-Like Conjugative Resistance Plasmid R1 Are Required and Sufficient To Confer Binding to the Substrate Selector Protein TraM. Journal of Bacteriology 186, 6999–7006.

Beuls, E., Houdt, R.V., Leys, N., Dijkstra, C., Larkin, O., and Mahillon, J. (2009). Bacillus thuringiensis Conjugation in Simulated Microgravity. Astrobiology 9, 797–805.

Bhatty, M., Laverde Gomez, J.A., and Christie, P.J. (2013). The expanding bacterial type IV secretion lexicon. Research in Microbiology 164, 620–639.

Bioteau, A., Durand, R., and Burrus, V. (2018). Redefinition and Unification of the SXT/R391 Family of Integrative and Conjugative Elements. Applied and Environmental Microbiology 84, e00485-18.

Blesa, A., and Berenguer, J. (2015). Contribution of vesicle-protected extracellular DNA to horizontal gene transfer in Thermus spp. International Microbiology 177–187.

Blesa, A., Baquedano, I., Quintáns, N.G., Mata, C.P., Castón, J.R., and Berenguer, J. (2017). The transjugation machinery of Thermus thermophilus: Identification of TdtA, an ATPase involved in DNA donation. PLOS Genetics 13, e1006669.

217

Bordeleau, E., Brouillette, E., Robichaud, N., and Burrus, V. (2010). Beyond antibiotic resistance: integrating conjugative elements of the SXT/R391 family that encode novel diguanylate cyclases participate to c-di-GMP signalling in Vibrio cholerae. Environmental Microbiology 12, 510–523.

Bose, B., Auchtung, J.M., Lee, C.A., and Grossman, A.D. (2008). A conserved anti-repressor controls horizontal gene transfer by proteolysis. Mol Microbiol 70, 570–582.

Boyd, D., Peters, G.A., Cloeckaert, A., Boumedine, K.S., Chaslus-Dancla, E., Imberechts, H., and Mulvey, M.R. (2001). Complete Nucleotide Sequence of a 43-Kilobase Genomic Island Associated with the Multidrug Resistance Region of Salmonella enterica Serovar Typhimurium DT104 and Its Identification in Phage Type DT120 and Serovar Agona. J. Bacteriol. 183, 5725– 5732.

Boyd, D.A., Peters, G.A., Ng, L.-K., and Mulvey, M.R. (2000). Partial characterization of a genomic island associated with the multidrug resistance region of Salmonella enterica Typhymurium DT104. FEMS Microbiol Lett 189, 285–291.

Browning, D.F., and Busby, S.J.W. (2004). The regulation of bacterial transcription initiation. Nat Rev Microbiol 2, 57–65.

Burrus, V. (2017). Mechanisms of stabilization of integrative and conjugative elements. Current Opinion in Microbiology 38, 44–50.

Burrus, V., and Waldor, M.K. (2003). Control of SXT Integration and Excision. J. Bacteriol. 185, 5045–5054.

Burrus, V., Pavlovic, G., Decaris, B., and Guédon, G. (2002). Conjugative transposons: the tip of the iceberg. Molecular Microbiology 46, 601–610.

Cabezón, E., Sastre, J.I., and de la Cruz, F. (1997). Genetic evidence of a coupling role for the TraG protein family in bacterial conjugation. Mol. Gen. Genet. 254, 400–406.

Cabezón, E., Ripoll-Rozada, J., Peña, A., Cruz, F. de la, and Arechaga, I. (2015). Towards an integrated model of bacterial conjugation. FEMS Microbiology Reviews 39, 81–95.

Campos, A., and Matsumura, P. (2001). Extensive alanine scanning reveals protein–protein and protein–DNA interaction surfaces in the global regulator FlhD from Escherichia coli. Molecular Microbiology 39, 581–594.

Capella-Gutiérrez, S., Silla-Martínez, J.M., and Gabaldón, T. (2009). trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25, 1972– 1973.

218

Carraro, N., Matteau, D., Luo, P., Rodrigue, S., and Burrus, V. (2014a). The Master Activator of IncA/C Conjugative Plasmids Stimulates Genomic Islands and Multidrug Resistance Dissemination. PLoS Genet 10, e1004714.

Carraro, N., Sauvé, M., Matteau, D., Lauzon, G., Rodrigue, S., and Burrus, V. (2014b). Development of pVCR94ΔX from Vibrio cholerae, a prototype for studying multidrug resistant IncA/C conjugative plasmids. Front Microbiol 5.

Carraro, N., Poulin, D., and Burrus, V. (2015a). Replication and Active Partition of Integrative and Conjugative Elements (ICEs) of the SXT/R391 Family: The Line between ICEs and Conjugative Plasmids Is Getting Thinner. PLoS Genet 11, e1005298.

Carraro, N., Matteau, D., Burrus, V., and Rodrigue, S. (2015b). Unraveling the regulatory network of IncA/C plasmid mobilization: When genomic islands hijack conjugative elements. Mobile Genetic Elements 5, 1–5.

Carraro, N., Rivard, N., Ceccarelli, D., Colwell, R.R., and Burrus, V. (2016). IncA/C Conjugative Plasmids Mobilize a New Family of Multidrug Resistance Islands in Clinical Vibrio cholerae Non-O1/Non-O139 Isolates from Haiti. MBio 7, e00509-16.

Carraro, N., Durand, R., Rivard, N., Anquetil, C., Barrette, C., Humbert, M., and Burrus, V. (2017a). Salmonella genomic island 1 (SGI1) reshapes the mating apparatus of IncC conjugative plasmids to promote self-propagation. PLOS Genetics 13, e1006705.

Carraro, N., Rivard, N., Burrus, V., and Ceccarelli, D. (2017b). Mobilizable genomic islands, different strategies for the dissemination of multidrug resistance and other adaptive traits. Mobile Genetic Elements 7, 1–6.

Carraro, N., Richard, X., Sulser, S., Delavat, F., Mazza, C., and van der Meer, J.R. (2020). An analog to digital converter controls bistable transfer competence development of a widespread bacterial integrative and conjugative element. ELife 9, e57915.

Cascales, E., and Christie, P.J. (2003). THE VERSATILE BACTERIAL TYPE IV SECRETION SYSTEMS. Nat Rev Microbiol 1.

Cascales, E., and Christie, P.J. (2004a). Definition of a Bacterial Type IV Secretion Pathway for a DNA Substrate. Science 304.

Cascales, E., and Christie, P.J. (2004b). Agrobacterium VirB10, an ATP energy sensor required for type IV secretion. PNAS 101, 17228–17233.

Cascales, E., Atmakuri, K., Liu, Z., Binns, A.N., and Christie, P.J. (2005). Agrobacterium tumefaciens oncogenic suppressors inhibit T-DNA and VirE2 protein substrate binding to the VirD4 coupling protein. Molecular Microbiology 58, 565–579.

219

Ceccarelli, D., Daccord, A., René, M., and Burrus, V. (2008). Identification of the Origin of Transfer (oriT) and a New Gene Required for Mobilization of the SXT/R391 Family of Integrating Conjugative Elements. J. Bacteriol. 190, 5328–5338.

Celli, J., and Trieu‐Cuot, P. (1998). Circularization of Tn916 is required for expression of the transposon-encoded transfer functions: characterization of long tetracycline-inducible transcripts reading through the attachment site. Molecular Microbiology 28, 103–117.

Chandran, V., Fronzes, R., Duquerroy, S., Cronin, N., Navaza, J., and Waksman, G. (2009). Structure of the outer membrane complex of a type IV secretion system. Nature 462, 1011– 1015.

Cherepanov, P.P., and Wackernagel, W. (1995). Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene 158, 9–14.

Chiu, C.-M., and Thomas, C.M. (2004). Evidence for past integration of IncP-1 plasmids into bacterial chromosomes. FEMS Microbiology Letters 241, 163–169.

Christie, P.J. (2016). The Mosaic Type IV Secretion Systems. EcoSal Plus 7.

Christie, P.J., Atmakuri, K., Krishnamoorthy, V., Jakubowski, S., and Cascales, E. (2005). Biogenesis, Architecture, and Function of Bacterial Type IV Secretion Systems. Annu Rev Microbiol 59.

Claret, L., and Hughes, C. (2002). Interaction of the Atypical Prokaryotic Transcription Activator FlhD2C2 with Early Promoters of the Flagellar Gene Hierarchy. Journal of Molecular Biology 321, 185–199.

Clarke, M., Maddera, L., Harris, R.L., and Silverman, P.M. (2008). F-pili dynamics by live-cell imaging. PNAS 105, 17978–17981.

Close, S.M., and Kado, C.I. (1991). The osa gene of pSa encodes a 21.1-kilodalton protein that suppresses Agrobacterium tumefaciens oncogenicity. Journal of Bacteriology 173, 5449–5456.

Costa, T.R.D., Ilangovan, A., Ukleja, M., Redzej, A., Santini, J.M., Smith, T.K., Egelman, E.H., and Waksman, G. (2016). Structure of the Bacterial Sex F Pilus Reveals an Assembly of a Stoichiometric Protein-Phospholipid Complex. Cell 166, 1436-1444.e10.

Cummins, M.L., Hamidian, M., and Djordjevic, S.P. (2020). Salmonella Genomic Island 1 is Broadly Disseminated within Gammaproteobacteriaceae. Microorganisms 8. de Curraize, C., Siebor, E., Varin, V., Neuwirth, C., and Hall, R.M. (2020). Two New SGI1-LK Variants Found in Proteus mirabilis and Evolution of the SGI1-HKL Group of Salmonella Genomic Islands. MSphere 5.

220

Daccord, A., Ceccarelli, D., and Burrus, V. (2010). Integrating conjugative elements of the SXT/R391 family trigger the excision and drive the mobilization of a new class of Vibrio genomic islands. Molecular Microbiology 78, 576–588.

Daccord, A., Mursell, M., Poulin-Laprade, D., and Burrus, V. (2012). Dynamics of the SetCD- Regulated Integration and Excision of Genomic Islands Mobilized by Integrating Conjugative Elements of the SXT/R391 Family. J. Bacteriol. 194, 5794–5802.

Daccord, A., Ceccarelli, D., Rodrigue, S., and Burrus, V. (2013). Comparative Analysis of Mobilizable Genomic Islands. J. Bacteriol. 195, 606–614.

Datsenko, K.A., and Wanner, B.L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. PNAS 97, 6640–6645.

Datta, S., Larkin, C., and Schildbach, J.F. (2003). Structural Insights into Single-Stranded DNA Binding and Cleavage by F Factor TraI. Structure 11, 1369–1379.

Datta, S., Costantino, N., and Court, D.L. (2006). A set of recombineering plasmids for gram- negative bacteria. Gene 379, 109–115.

Davis, C.A., Bingman, C.A., Landick, R., Record, M.T., and Saecker, R.M. (2007). Real-time footprinting of DNA in the first kinetically significant intermediate in open complex formation by Escherichia coli RNA polymerase. PNAS 104, 7833–7838.

Delavat, F., Mitri, S., Pelet, S., and van der Meer, J.R. (2016). Highly variable individual donor cell fates characterize robust horizontal gene transfer of an integrative and conjugative element. Proc Natl Acad Sci U S A 113, E3375–E3383.

Delavat, F., Miyazaki, R., Carraro, N., Pradervand, N., Meer, V.D., and Roelof, J. (2017). The hidden life of integrative and conjugative elements. FEMS Microbiol Rev.

Demarre, G., Guérout, A.-M., Matsumoto-Mashimo, C., Rowe-Magnus, D.A., Marlière, P., and Mazel, D. (2005). A new family of mobilizable suicide plasmids based on broad host range R388 plasmid (IncW) and RP4 plasmid (IncPα) conjugative machineries and their cognate Escherichia coli host strains. Res. Microbiol. 156, 245–255.

Dillard, J.P., and Seifert, H.S. (2001). A variable genetic island specific for Neisseria gonorrhoeae is involved in providing DNA for natural transformation and is found more often in disseminated infection isolates. Molecular Microbiology 41, 263–277.

Dobrindt, U., Hochhut, B., Hentschel, U., and Hacker, J. (2004). Genomic islands in pathogenic and environmental microorganisms. Nat Rev Micro 2, 414–424.

Dostál, L., and Schildbach, J.F. (2010). Single-Stranded DNA Binding by F TraI Relaxase and Helicase Domains Is Coordinately Regulated. J. Bacteriol. 192, 3620–3628.

221

Douard, G., Praud, K., Cloeckaert, A., and Doublet, B. (2010). The Salmonella Genomic Island 1 Is Specifically Mobilized In Trans by the IncA/C Multidrug Resistance Plasmid Family. PLoS One 5.

Dower, W.J., Miller, J.F., and Ragsdale, C.W. (1988). High efficiency transformation of E. coli by high voltage electroporation. Nucl. Acids Res. 16, 6127–6145.

Dunny, G.M. (2007). The peptide pheromone-inducible conjugation system of Enterococcus faecalis plasmid pCF10: cell–cell signalling, gene transfer, complexity and evolution. Philosophical Transactions of the Royal Society B: Biological Sciences 362, 1185–1193.

Durand, R., Huguet, K.T., Rivard, N., Carraro, N., Rodrigue, S., and Burrus, V. (2020). Crucial Role of Salmonella Genomic Island 1 Master Activator in Parasitism of IncC plasmids. Nucleic Acids Research under revision, NAR-02797-V-2020.

Edgar, R.C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32, 1792–1797.

Farrugia, D.N., Elbourne, L.D.H., Mabbutt, B.C., and Paulsen, I.T. (2015). A novel family of integrases associated with prophages and genomic islands integrated within the tRNA- dihydrouridine synthase A (dusA) gene. Nucleic Acids Res 43, 4547–4557.

Firth, N., and Skurray, R. (1992). Characterization of the F plasmid bifunctional conjugation gene, traG. Molec. Gen. Genet. 232, 145–153.

Fronzes, R., Christie, P.J., and Waksman, G. (2009). The structural biology of type IV secretion systems. Nat Rev Microbiol 7.

Frost, L.S. (2009). Bacterial conjugation. In Desk Encyclopedia of Microbiology, M. Schaechter, ed. (San Diego, CA, USA: Academic Press), pp. 291–308.

Frost, L.S., and Koraimann, G. (2010). Regulation of bacterial conjugation: balancing opportunity with adversity. Future Microbiology 5, 1057–1071.

Garcillán-Barcia, M.P., and de la Cruz, F. (2008). Why is entry exclusion an essential feature of conjugative plasmids? Plasmid 60, 1–18.

Garcillán-Barcia, M.P., Francia, M.V., and de La Cruz, F. (2009). The diversity of conjugative relaxases and its application in plasmid classification. FEMS Microbiol Rev 33, 657–687.

Garner, E.C., Campbell, C.S., Weibel, D.B., and Mullins, R.D. (2007). Reconstitution of DNA Segregation Driven by Assembly of a Prokaryotic Actin Homolog. Science 315, 1270–1274.

Gerdes, K., Larsen, J.E., and Molin, S. (1985). Stable inheritance of plasmid R1 requires two different loci. Journal of Bacteriology 161, 292–298.

222

Getino, M., Palencia-Gándara, C., Garcillán-Barcia, M.P., and de la Cruz, F. (2017). PifC and Osa, Plasmid Weapons against Rival Conjugative Coupling Proteins. Frontiers in Microbiology 8.

Gillespie, J.J., Ammerman, N.C., Dreher-Lesnick, S.M., Rahman, M.S., Worley, M.J., Setubal, J.C., Sobral, B.S., and Azad, A.F. (2009). An Anomalous Type IV Secretion System in Rickettsia Is Evolutionarily Conserved. PLOS ONE 4, e4833.

Gillespie, J.J., Brayton, K.A., Williams, K.P., Diaz, M.A.Q., Brown, W.C., Azad, A.F., and Sobral, B.W. (2010). Phylogenomics Reveals a Diverse Rickettsiales Type IV Secretion System. Infection and Immunity 78, 1809–1823.

Golding, G.R., Olson, A.B., Doublet, B., Cloeckaert, A., Christianson, S., Graham, M.R., and Mulvey, M.R. (2007). The effect of the Salmonella genomic island 1 on in vitro global gene expression in Salmonella enterica serovar Typhimurium LT2. Microbes and Infection 9, 21–27.

Gordon, J.E., Costa, T.R.D., Patel, R.S., Gonzalez-Rivera, C., Sarkar, M.K., Orlova, E.V., Waksman, G., and Christie, P.J. (2017). Use of Chimeric Type IV Secretion Systems to Define Contributions of Outer Membrane Subassemblies for Contact-Dependent Translocation. Mol Microbiol 105, 273–293.

Grenier, F., Matteau, D., Baby, V., and Rodrigue, S. (2014). Complete Genome Sequence of Escherichia coli BW25113. Genome Announc 2.

Grohmann, E., Christie, P.J., Waksman, G., and Backert, S. (2018). Type IV secretion in Gram- negative and Gram-positive bacteria: Type IV secretion. Molecular Microbiology 107, 455– 471.

Guédon, G., Libante, V., Coluzzi, C., Payot, S., and Leblond-Bourget, N. (2017). The Obscure World of Integrative and Mobilizable Elements, Highly Widespread Elements that Pirate Bacterial Conjugative Systems. Genes 8, 337.

Guglielmini, J., de la Cruz, F., and Rocha, E.P.C. (2013). Evolution of Conjugation and Type IV Secretion Systems. Mol Biol Evol 30, 315–331.

Guogas, L.M., Kennedy, S.A., Lee, J.-H., and Redinbo, M.R. (2009). A Novel Fold in the TraI Relaxase–Helicase C-Terminal Domain Is Essential for Conjugative DNA Transfer. Journal of Molecular Biology 386, 554–568.

Guzman, L.M., Belin, D., Carson, M.J., and Beckwith, J. (1995). Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J. Bacteriol. 177, 4121–4130.

Guzmán-Herrador, D.L., and Llosa, M. (2019). The secret life of conjugative relaxases. Plasmid 104, 102415.

223

Haase, J., Kalkum, M., and Lanka, E. (1996). TrbK, a small cytoplasmic membrane lipoprotein, functions in entry exclusion of the IncPα plasmid RP4. Journal of Bacteriology 178, 6720–6729.

Haldimann, A., and Wanner, B.L. (2001). Conditional-Replication, Integration, Excision, and Retrieval Plasmid-Host Systems for Gene Structure-Function Studies of Bacteria. J. Bacteriol. 183, 6384–6393.

Hamidian, M., Holt, K.E., and Hall, R.M. (2015). Genomic resistance island AGI1 carrying a complex class 1 integron in a multiply antibiotic-resistant ST25 Acinetobacter baumannii isolate. J Antimicrob Chemother 70, 2519–2523.

Hancock, S.J., Phan, M.-D., Peters, K.M., Forde, B.M., Chong, T.M., Yin, W.-F., Chan, K.-G., Paterson, D.L., Walsh, T.R., Beatson, S.A., et al. (2017). Identification of IncA/C Plasmid Replication and Maintenance Genes and Development of a Plasmid Multilocus Sequence Typing Scheme. Antimicrob Agents Chemother 61.

Hancock, S.J., Phan, M.-D., Luo, Z., Lo, A.W., Peters, K.M., Nhu, N.T.K., Forde, B.M., Whitfield, J., Yang, J., Strugnell, R.A., et al. (2020). Comprehensive analysis of IncC plasmid conjugation identifies a crucial role for the transcriptional regulator AcaB. Nature Microbiology 1–9.

Harmer, C.J., and Hall, R.M. (2015). The A to Z of A/C plasmids. Plasmid 80, 63–82.

Harmer, C.J., Hamidian, M., Ambrose, S.J., and Hall, R.M. (2016). Destabilization of IncA and IncC plasmids by SGI1 and SGI2 type Salmonella genomic islands. Plasmid.

Harmer, C.J., Hamidian, M., and Hall, R.M. (2017). pIP40a, a type 1 IncC plasmid from 1969 carries the integrative element GIsul2 and a novel class II mercury resistance transposon. Plasmid 92, 17–25.

Haskett, T.L., Terpolilli, J.J., Bekuma, A., O’Hara, G.W., Sullivan, J.T., Wang, P., Ronson, C.W., and Ramsay, J.P. (2016). Assembly and transfer of tripartite integrative and conjugative genetic elements. PNAS 113, 12268–12273.

Haskett, T.L., Ramsay, J.P., Bekuma, A.A., Sullivan, J.T., O’Hara, G.W., and Terpolilli, J.J. (2017). Evolutionary persistence of tripartite integrative and conjugative elements. Plasmid 92, 30–36.

Hedges, R.W. (1974). R factors from Providence. Microbiology, 81, 171–181.

Hegyi, A., Szabó, M., Olasz, F., and Kiss, J. (2017). Identification of oriT and a recombination hot spot in the IncA/C plasmid backbone. Scientific Reports 7, 10595.

Heilers, J.-H., Reiners, J., Heller, E.-M., Golzer, A., Smits, S.H.J., and van der Does, C. (2019). DNA processing by the MOBH family relaxase TraI encoded within the gonococcal genetic island. Nucleic Acids Res 47, 8136–8153.

224

Hellemans, J., Mortier, G., De Paepe, A., Speleman, F., and Vandesompele, J. (2007). qBase relative quantification framework and software for management and automated analysis of real- time quantitative PCR data. Genome Biology 8, R19.

Hobbs, Z., and Abedon, S.T. (2016). Diversity of phage infection types and associated terminology: the problem with ‘Lytic or lysogenic.’ FEMS Microbiology Letters 363, fnw047.

Hochhut, B., Lotfi, Y., Mazel, D., Faruque, S.M., Woodgate, R., and Waldor, M.K. (2001). Molecular Analysis of Antibiotic Resistance Gene Clusters in Vibrio cholerae O139 and O1 SXT Constins. Antimicrob Agents Chemother 45, 2991–3000.

Hofreuter, D., Odenbreit, S., and Haas, R. (2001). Natural transformation competence in Helicobacter pylori is mediated by the basic components of a type IV secretion system. Molecular Microbiology 41, 379–391.

Hsu, C.-C., and Chen, C.W. (2010). Linear Plasmid SLP2 Is Maintained by Partitioning, Intrahyphal Spread, and Conjugal Transfer in Streptomyces. Journal of Bacteriology 192, 307– 315.

Hu, B., Khara, P., and Christie, P.J. (2019). Structural bases for F plasmid conjugation and F pilus biogenesis in Escherichia coli. PNAS 201904428.

Huang, Y., Niu, B., Gao, Y., Fu, L., and Li, W. (2010). CD-HIT Suite: a web server for clustering and comparing biological sequences. Bioinformatics 26, 680–682.

Huguet, K.T. (2016). Stabilité de Salmonella Genomic Island 1 et son incompatibilité avec les plasmides IncA/C.

Huguet, K.T., Gonnet, M., Doublet, B., and Cloeckaert, A. (2016). A toxin antitoxin system promotes the maintenance of the IncA/C-mobilizable Salmonella Genomic Island 1. Sci Rep 6.

Huguet, K.T., Rivard, N., Garneau, D., Palanee, J., and Burrus, V. (2020). Replication of the Salmonella Genomic Island 1 (SGI1) triggered by helper IncC conjugative plasmids promotes incompatibility and plasmid loss. PLOS Genetics 16, e1008965.

Humbert, M., Huguet, K.T., Coulombe, F., and Burrus, V. (2019). Entry Exclusion of Conjugative Plasmids of the IncA, IncC, and Related Untyped Incompatibility Groups. J Bacteriol 201.

Ilangovan, A., Kay, C.W.M., Roier, S., El Mkami, H., Salvadori, E., Zechner, E.L., Zanetti, G., and Waksman, G. (2017). Cryo-EM Structure of a Relaxase Reveals the Molecular Basis of DNA Unwinding during Bacterial Conjugation. Cell 169, 708-721.e12.

Johnson, C.M., and Grossman, A.D. (2015). Integrative and Conjugative Elements (ICEs): What They Do and How They Work. Annual Review of Genetics 49, 577–601.

225

Johnston, C., Martin, B., Fichant, G., Polard, P., and Claverys, J. (2014). Bacterial transformation: distribution, shared mechanisms and divergent control. Nature Reviews. Microbiology; London 12, 181–196.

Juhas, M., van der Meer, J.R., Gaillard, M., Harding, R.M., Hood, D.W., and Crook, D.W. (2009). Genomic islands: tools of bacterial horizontal gene transfer and evolution. FEMS Microbiol Rev 33, 376–393.

Kamenšek, S., Podlesek, Z., Gillor, O., and Žgur-Bertok, D. (2010). Genes regulated by the Escherichia coli SOS repressor LexA exhibit heterogenous expression. BMC Microbiol 10, 283.

Karimova, G., Dautin, N., and Ladant, D. (2005). Interaction Network among Escherichia coli Membrane Proteins Involved in Cell Division as Revealed by Bacterial Two-Hybrid Analysis. J Bacteriol 187, 2233–2243.

Kiss, J., Nagy, B., and Olasz, F. (2012). Stability, Entrapment and Variant Formation of Salmonella Genomic Island 1. PLoS One 7.

Kiss, J., Papp, P.P., Szabó, M., Farkas, T., Murányi, G., Szakállas, E., and Olasz, F. (2015). The master regulator of IncA/C plasmids is recognized by the Salmonella Genomic island SGI1 as a signal for excision and conjugal transfer. Nucleic Acids Res 43, 8735–8745.

Kiss, J., Szabó, M., Hegyi, A., Douard, G., Praud, K., Nagy, I., Olasz, F., Cloeckaert, A., and Doublet, B. (2019). Identification and Characterization of oriT and Two Mobilization Genes Required for Conjugative Transfer of Salmonella Genomic Island 1. Front Microbiol 10.

Klieve, A.V., Yokoyama, M.T., Forster, R.J., Ouwerkerk, D., Bain, P.A., and Mawhinney, E.L. (2005). Naturally Occurring DNA Transfer System Associated with Membrane Vesicles in Cellulolytic Ruminococcus spp. of Ruminal Origin. Appl Environ Microbiol 71, 4248–4253.

Klimke, W.A., Rypien, C.D., Klinger, B., Kennedy, R.A., Rodriguez-Maillard, J.M., and Frost, L.S. (2005). The mating pair stabilization protein, TraN, of the F plasmid is an outer-membrane protein with two regions that are important for its function in conjugation. Microbiology 151, 3527–3540.

Kovacic, R.T. (1987). The 0°C closed complexes between Escherichia coli RNA polymerase and two promoters, T7-A3 and lacUV5. J. Biol. Chem. 262, 13654–13661.

Kumar, S., Stecher, G., and Tamura, K. (2016). MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol 33, 1870–1874.

Kwak, M.-J., Kim, J.D., Kim, H., Kim, C., Bowman, J.W., Kim, S., Joo, K., Lee, J., Jin, K.S., Kim, Y.-G., et al. (2017). Architecture of the type IV coupling protein complex of Legionella pneumophila. Nature Microbiology 2, nmicrobiol2017114.

226

Lang, K.S., and Johnson, T.J. (2016). Characterization of Acr2, an H-NS-like protein encoded on A/C2-type plasmids. Plasmid 87–88, 17–27.

Lang, A.S., Zhaxybayeva, O., and Beatty, J.T. (2012a). Gene transfer agents: phage-like elements of genetic exchange. Nat Rev Microbiol 10, 472–482.

Lang, K.S., Danzeisen, J.L., Xu, W., and Johnson, T.J. (2012b). Transcriptome Mapping of pAR060302, a blaCMY-2-Positive Broad-Host-Range IncA/C Plasmid. Appl. Environ. Microbiol. 78, 3379–3386.

Latif, H., Federowicz, S., Ebrahim, A., Tarasova, J., Szubin, R., Utrilla, J., Zengler, K., and Palsson, B.O. (2018). ChIP-exo interrogation of Crp, DNA, and RNAP holoenzyme interactions. PLoS One 13.

Lawley, T.D., Klimke, W.A., Gubbins, M.J., and Frost, L.S. (2003). F factor conjugation is a true type IV secretion system. FEMS Microbiology Letters 224, 1–15.

Le, S.Q., and Gascuel, O. (2008). An Improved General Amino Acid Replacement Matrix. Mol Biol Evol 25, 1307–1320.

Lee, Y.-Y., Barker, C.S., Matsumura, P., and Belas, R. (2011). Refining the Binding of the Escherichia coli Flagellar Master Regulator, FlhD4C2, on a Base-Specific Level. Journal of Bacteriology 193, 4057–4068.

Li, B., Li, N., Wang, F., Guo, L., Huang, Y., Liu, X., Wei, T., Zhu, D., Liu, C., Pan, H., et al. (2012). Structural insight of a concentration-dependent mechanism by which YdiV inhibits Escherichia coli flagellum biogenesis and motility. Nucleic Acids Res 40, 11073–11085.

Li, B., Yue, Y., Yuan, Z., Zhang, F., Li, P., Song, N., Lin, W., Liu, Y., Yang, Y., Li, Z., et al. (2017). Salmonella STM1697 coordinates flagella biogenesis and virulence by restricting flagellar master protein FlhD4C2 from recruiting RNA polymerase. Nucleic Acids Res 45, 9976– 9989.

Liu, Z., Que, F., Liao, L., Zhou, M., You, L., Zhao, Q., Li, Y., Niu, H., Wu, S., and Huang, R. (2014). Study on the Promotion of Bacterial Biofilm Formation by a Salmonella Conjugative Plasmid and the Underlying Mechanism. PLoS One 9.

Low, H.H., Gubellini, F., Rivera-Calzada, A., Braun, N., Connery, S., Dujeancourt, A., Lu, F., Redzej, A., Fronzes, R., Orlova, E.V., et al. (2014). Structure of a type IV secretion system. Nature 508, 550–553.

Lu, J., and Frost, L.S. (2005). Mutations in the C-Terminal Region of TraM Provide Evidence for In Vivo TraM-TraD Interactions during F-Plasmid Conjugation. Journal of Bacteriology 187, 4767–4773.

227

Lu, J., Wong, J.J.W., Edwards, R.A., Manchak, J., Frost, L.S., and Glover, J.N.M. (2008). Structural basis of specific TraD–TraM recognition during F plasmid-mediated bacterial conjugation. Molecular Microbiology 70, 89–99.

Luo, P., He, X., Wang, Y., Liu, Q., and Hu, C. (2016). Comparative genomic analysis of six new-found integrative conjugative elements (ICEs) in Vibrio alginolyticus. BMC Microbiol 16, 79.

Maindola, P., Raina, R., Goyal, P., Atmakuri, K., Ojha, A., Gupta, S., Christie, P.J., Iyer, L.M., Aravind, L., and Arockiasamy, A. (2014). Multiple enzymatic activities of ParB/Srx superfamily mediate sexual conflict among conjugative plasmids. Nature Communications 5, 5322.

Majdalani, N., Moore, D., Maneewannakul, S., and Ippen-Ihler, K. (1996). Role of the propilin leader peptide in the maturation of F pilin. J Bacteriol 178, 3748–3754.

Maneewannakul, S., Maneewannakul, K., and Ippen-Ihler, K. (1992). Characterization, localization, and sequence of F transfer region products: the pilus assembly gene product TraW and a new product, TrbI. J. Bacteriol. 174, 5567–5574.

Marraffini, L.A., and Sontheimer, E.J. (2008). CRISPR Interference Limits Horizontal Gene Transfer in Staphylococci by Targeting DNA. Science 322, 1843–1845.

Marrero, J., and Waldor, M.K. (2005). Interactions between Inner Membrane Proteins in Donor and Recipient Cells Limit Conjugal DNA Transfer. Developmental Cell 8, 963–970.

Marrero, J., and Waldor, M.K. (2007). Determinants of Entry Exclusion within Eex and TraG Are Cytoplasmic. J. Bacteriol. 189, 6469–6473.

Martin, W. (1999). Mosaic bacterial chromosomes: a challenge en route to a tree of genomes. Bioessays 21, 99–104.

Martino, F., Tijet, N., Melano, R., Petroni, A., Heinz, E., De Belder, D., Faccone, D., Rapoport, M., Biondi, E., Rodrigo, V., et al. (2019). Isolation of five Enterobacteriaceae species harbouring blaNDM-1 and mcr-1 plasmids from a single paediatric patient. PLoS One 14.

Matson, S.W., and Ragonese, H. (2005). The F-Plasmid TraI Protein Contains Three Functional Domains Required for Conjugative DNA Strand Transfer. J. Bacteriol. 187, 697–706.

Menard, K.L., and Grossman, A.D. (2013). Selective Pressures to Maintain Attachment Site Specificity of Integrative and Conjugative Elements. PLOS Genetics 9, e1003623.

Meyer, R. (2015). Mapping Type IV Secretion Signals on the Primase Encoded by the Broad- Host-Range Plasmid R1162 (RSF1010). J Bacteriol 197, 3245–3254.

228

Miller, J.F., Lanka, E., and Malamy, M.H. (1985). F factor inhibition of conjugal transfer of broad-host-range plasmid RP4: requirement for the protein product of pif operon regulatory gene pifC. Journal of Bacteriology 163, 1067–1073.

Murányi, G., Szabó, M., Olasz, F., and Kiss, J. (2016). Determination and Analysis of the Putative AcaCD-Responsive Promoters of Salmonella Genomic Island 1. PLOS ONE 11, e0164561.

Nagai, H., Kagan, J.C., Zhu, X., Kahn, R.A., and Roy, C.R. (2002). A Bacterial Guanine Nucleotide Exchange Factor Activates ARF on Legionella Phagosomes. Science 295, 679–682.

Núñez, B., and De La Cruz, F. (2001). Two atypical mobilization proteins are involved in plasmid CloDF13 relaxation. Molecular Microbiology 39, 1088–1099.

Paiva, W.D., Grossman, T., and Silverman, P.M. (1992). Characterization of F-pilin as an inner membrane component of Escherichia coli K12. J. Biol. Chem. 267, 26191–26197.

Pansegrau, W., Lanka, E., Barth, P.T., Figurski, D.H., Guiney, D.G., Haas, D., Helinski, D.R., Schwab, H., Stanisich, V.A., and Thomas, C.M. (1994). Complete Nucleotide Sequence of Birmingham IncPα Plasmids: Compilation and Comparative Analysis. Journal of Molecular Biology 239, 623–663.

Parker, C., and Meyer, R.J. (2007). The R1162 relaxase/primase contains two, type IV transport signals that require the small plasmid protein MobB. Molecular Microbiology 66, 252–261.

Peng, Y., Lu, J., Wong, J.J.W., Edwards, R.A., Frost, L.S., and Mark Glover, J.N. (2014). Mechanistic Basis of Plasmid-Specific DNA Binding of the F Plasmid Regulatory Protein, TraM. Journal of Molecular Biology 426, 3783–3795.

Pennington, J.M., and Rosenberg, S.M. (2007). Spontaneous DNA breakage in single living Escherichia coli cells. Nature Genetics 39, 797–802.

Poulin-Laprade, D., and Burrus, V. (2015). A λ Cro-Like Repressor Is Essential for the Induction of Conjugative Transfer of SXT/R391 Elements in Response to DNA Damage. J. Bacteriol. 197, 3822–3833.

Poulin-Laprade, D., Carraro, N., and Burrus, V. (2015a). The extended regulatory networks of SXT/R391 integrative and conjugative elements and IncA/C conjugative plasmids. Front. Microbiol. 837.

Poulin-Laprade, D., Matteau, D., Jacques, P.-É., Rodrigue, S., and Burrus, V. (2015b). Transfer activation of SXT/R391 integrative and conjugative elements: unraveling the SetCD regulon. Nucl. Acids Res. 43, 2045–2056.

Pradervand, N., Sulser, S., Delavat, F., Miyazaki, R., Lamas, I., and Meer, J.R. van der (2014). An Operon of Three Transcriptional Regulators Controls Horizontal Gene Transfer of the

229

Integrative and Conjugative Element ICEclc in Pseudomonas knackmussii B13. PLOS Genetics 10, e1004441.

Ramsay, J.P., Sullivan, J.T., Stuart, G.S., Lamont, I.L., and Ronson, C.W. (2006). Excision and transfer of the Mesorhizobium loti R7A symbiosis island requires an integrase IntS, a novel recombination directionality factor RdfS, and a putative relaxase RlxS. Molecular Microbiology 62, 723–734.

Ramsay, J.P., Sullivan, J.T., Jambari, N., Ortori, C.A., Heeb, S., Williams, P., Barrett, D.A., Lamont, I.L., and Ronson, C.W. (2009). A LuxRI-family regulatory system controls excision and transfer of the Mesorhizobium loti strain R7A symbiosis island by activating expression of two conserved hypothetical genes. Molecular Microbiology 73, 1141–1155.

Ramsay, J.P., Major, A.S., Komarovsky, V.M., Sullivan, J.T., Dy, R.L., Hynes, M.F., Salmond, G.P.C., and Ronson, C.W. (2013). A widely conserved molecular switch controls quorum sensing and symbiosis island transfer in Mesorhizobium loti through expression of a novel antiactivator. Molecular Microbiology 87, 1–13.

Rancès, E., Voronin, D., Tran-Van, V., and Mavingui, P. (2008). Genetic and Functional Characterization of the Type IV Secretion System in Wolbachia. Journal of Bacteriology 190, 5020–5030.

Redzej, A., Ilangovan, A., Lang, S., Gruber, C.J., Topf, M., Zangger, K., Zechner, E.L., and Waksman, G. (2013). Structure of a translocation signal domain mediating conjugative transfer by type IV secretion systems. Molecular Microbiology 89, 324–333.

Redzej, A., Ukleja, M., Connery, S., Trokter, M., Felisberto‐Rodrigues, C., Cryar, A., Thalassinos, K., Hayward, R.D., Orlova, E.V., and Waksman, G. (2017). Structure of a VirD4 coupling protein bound to a VirB type IV secretion machinery. The EMBO Journal 36, 3080– 3095.

Reinhard, F., Miyazaki, R., Pradervand, N., and van der Meer, J.R. (2013). Cell Differentiation to “Mating Bodies” Induced by an Integrating and Conjugative Element in Free-Living Bacteria. Current Biology 23, 255–259.

Riede, I., and Eschbach, M.-L. (1986). Evidence that TraT interacts with OmpA of Escherichia coli. FEBS Letters 205, 241–245.

Rivard, N., Colwell, R.R., and Burrus, V. (2020). Antibiotic Resistance in Vibrio cholerae: Mechanistic Insights from IncC Plasmid-Mediated Dissemination of a Novel Family of Genomic Islands Inserted at trmE. MSphere 5.

Robinson, A., and Oijen, A.M. van (2013). Bacterial replication, transcription and translation: mechanistic insights from single-molecule biochemical studies. Nature Reviews Microbiology 11, nrmicro2994.

230

Roy, D., Huguet, K.T., Grenier, F., and Burrus, V. (2020). IncC conjugative plasmids and SXT/R391 elements repair double-strand breaks caused by CRISPR–Cas during conjugation. Nucleic Acids Res.

Rumbo, C., Fernández-Moreira, E., Merino, M., Poza, M., Mendez, J.A., Soares, N.C., Mosquera, A., Chaves, F., and Bou, G. (2011). Horizontal Transfer of the OXA-24 Carbapenemase Gene via Outer Membrane Vesicles: a New Mechanism of Dissemination of Carbapenem Resistance Genes in Acinetobacter baumannii ▿. Antimicrob Agents Chemother 55, 3084–3090.

Salje, J., Gayathri, P., and Löwe, J. (2010). The ParMRC system: molecular mechanisms of plasmid segregation by actin-like filaments. Nature Reviews Microbiology 8, 683–692.

Santini, J.M., and Stanisich, V.A. (1998). Both the fipA Gene of pKM101 and the pifC Gene of F Inhibit Conjugal Transfer of RP1 by an Effect on traG. Journal of Bacteriology 180, 4093– 4101.

Sastre, J.I., Cabezón, E., and Cruz, F. de la (1998). The Carboxyl Terminus of Protein TraD Adds Specificity and Efficiency to F-Plasmid Conjugative Transfer. J. Bacteriol. 180, 6039– 6042.

Schickor, P., Metzger, W., Werel, W., Lederer, H., and Heumann, H. (1990). Topography of intermediates in transcription initiation of E. coli. EMBO J 9, 2215–2220.

Schultz, E., Barraud, O., Madec, J.-Y., Haenni, M., Cloeckaert, A., Ploy, M.-C., and Doublet, B. (2017). Multidrug Resistance Salmonella Genomic Island 1 in a Morganella morganii subsp. morganii Human Clinical Isolate from France. MSphere 2.

Sclavi, B., Zaychikov, E., Rogozina, A., Walther, F., Buckle, M., and Heumann, H. (2005). Real-time characterization of intermediates in the pathway to open complex formation by Escherichia coli RNA polymerase at the T7A1 promoter. Proc Natl Acad Sci U S A 102, 4706– 4711.

Sgro, G.G., Costa, T.R.D., Cenens, W., Souza, D.P., Cassago, A., Oliveira, L.C., Salinas, R.K., Portugal, R.V., Farah, C.S., and Waksman, G. (2018). Cryo-EM structure of the bacteria killing type IV secretion system core complex from Xanthomonas citri. Nat Microbiol 3, 1429–1440.

Sharifahmadian, M., Nlend, I.U., Lecoq, L., Omichinski, J.G., and Baron, C. (2017). The type IV secretion system core component VirB8 interacts via the β1‐strand with VirB10. FEBS Letters 591, 2491–2500.

Siebor, E., and Neuwirth, C. (2013). Emergence of Salmonella genomic island 1 (SGI1) among Proteus mirabilis clinical isolates in Dijon, France. J Antimicrob Chemother 68, 1750–1756.

231

Siebor, E., Curraize, C. de, Amoureux, L., and Neuwirth, C. (2016). Mobilization of the Salmonella genomic island SGI1 and the Proteus genomic island PGI1 by the A/C2 plasmid carrying blaTEM-24 harboured by various clinical species of Enterobacteriaceae. J. Antimicrob. Chemother. 71, 2167–2170.

Siebor, E., de-CURRAIZE, C., and Neuwirth, C. (2020). New insights concerning Acinetobacter genomic island-related elements. International Journal of Antimicrobial Agents 106117.

Singer, M., Baker, T.A., Schnitzler, G., Deischel, S.M., Goel, M., Dove, W., Jaacks, K.J., Grossman, A.D., Erickson, J.W., and Gross, C.A. (1989). A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli. Microbiol Rev 53, 1–24.

Soliman, A.M., Ramadan, H., Ghazy, E., Yu, L., Hisatsune, J., Kayama, S., Sugai, M., Nariya, H., Shimamoto, T., Jackson, C.R., et al. (2020). Emergence of Salmonella genomic island 1 variant SGI1-C in a multidrug-resistant clinical isolate of Klebsiella pneumoniae ST485 from Egypt. Antimicrobial Agents and Chemotherapy.

Sommer, M.O.A., Munck, C., Toft-Kehler, R.V., and Andersson, D.I. (2017). Prediction of antibiotic resistance: time for a new preclinical paradigm? Nature Reviews Microbiology 15, nrmicro.2017.75.

Spagnoletti, M., Ceccarelli, D., Rieux, A., Fondi, M., Taviani, E., Fani, R., Colombo, M.M., Colwell, R.R., and Balloux, F. (2014). Acquisition and Evolution of SXT-R391 Integrative Conjugative Elements in the Seventh-Pandemic Vibrio cholerae Lineage. MBio 5, e01356-14.

Szabó, M., Nagy, T., Wilk, T., Farkas, T., Hegyi, A., Olasz, F., and Kiss, J. (2016). Characterization of two multidrug-resistant IncA/C plasmids from the 1960s by using Oxford Nanopore MinION sequencer device. Antimicrob. Agents Chemother. AAC.01121-16.

Takaya, A., Erhardt, M., Karata, K., Winterberg, K., Yamamoto, T., and Hughes, K.T. (2012). YdiV: a dual function protein that targets FlhDC for ClpXP-dependent degradation by promoting release of DNA-bound FlhDC complex. Molecular Microbiology 83, 1268–1284.

Tatum, E.L., and Lederberg, J. (1947). Gene Recombination in the Bacterium Escherichia coli. J Bacteriol 53, 673–684.

Taviani, E., Spagnoletti, M., Ceccarelli, D., Haley, B.J., Hasan, N.A., Chen, A., Colombo, M.M., Huq, A., and Colwell, R.R. (2012). Genomic analysis of ICEVchBan8: An atypical genetic element in Vibrio cholerae. FEBS Lett 586, 1617–1621.

Touchon, M., Moura de Sousa, J.A., and Rocha, E.P. (2017). Embracing the enemy: the diversification of microbial gene repertoires by phage-mediated horizontal gene transfer. Current Opinion in Microbiology 38, 66–73.

232

Trokter, M., and Waksman, G. (2018). Translocation through the Conjugative Type IV Secretion System Requires Unfolding of Its Protein Substrate. J Bacteriol 200.

Vogelmann, J., Ammelburg, M., Finger, C., Guezguez, J., Linke, D., Flötenmeyer, M., Stierhof, Y.-D., Wohlleben, W., and Muth, G. (2011). Conjugal plasmid transfer in Streptomyces resembles bacterial chromosome segregation by FtsK/SpoIIIE. EMBO J 30, 2246–2254.

Wailan, A.M., Sartor, A.L., Zowawi, H.M., Perry, J.D., Paterson, D.L., and Sidjabat, H.E. (2015). Genetic Contexts of blaNDM-1 in Patients Carrying Multiple NDM-Producing Strains. Antimicrob. Agents Chemother. 59, 7405–7410.

Waldor, M.K., Tschäpe, H., and Mekalanos, J.J. (1996). A new type of conjugative transposon encodes resistance to sulfamethoxazole, trimethoprim, and streptomycin in Vibrio cholerae O139. J Bacteriol 178, 4157–4165.

Wang, J.C. (1979). Helical repeat of DNA in solution. Proc Natl Acad Sci U S A 76, 200–203.

Wang, S., Fleming, R.T., Westbrook, E.M., Matsumura, P., and McKay, D.B. (2006). Structure of the Escherichia coli FlhDC Complex, a Prokaryotic Heteromeric Regulator of Transcription. Journal of Molecular Biology 355, 798–808.

Waters, J.L., and Salyers, A.A. (2013). Regulation of CTnDOT Conjugative Transfer Is a Complex and Highly Coordinated Series of Events. MBio 4.

Waters, C.M., Lu, W., Rabinowitz, J.D., and Bassler, B.L. (2008). Quorum Sensing Controls Biofilm Formation in Vibrio cholerae through Modulation of Cyclic Di-GMP Levels and Repression of vpsT. Journal of Bacteriology 190, 2527–2536.

Weill, F.-X., Domman, D., Njamkepo, E., Tarr, C., Rauzier, J., Fawal, N., Keddy, K.H., Salje, H., Moore, S., Mukhopadhyay, A.K., et al. (2017). Genomic history of the seventh pandemic of cholera in Africa. Science 358, 785–789.

Weiss, A.A., Johnson, F.D., and Burns, D.L. (1993). Molecular characterization of an operon required for pertussis toxin secretion. Proc Natl Acad Sci U S A 90, 2970–2974.

White, C.E., and Winans, S.C. (2007). Cell–cell communication in the plant pathogen Agrobacterium tumefaciens. Philosophical Transactions of the Royal Society B: Biological Sciences 362, 1135–1148.

Wilkins, B.M. (2002). Plasmid promiscuity: meeting the challenge of DNA immigration control. Environmental Microbiology 4, 495–500.

Wong, J.J.W., Lu, J., Edwards, R.A., Frost, L.S., and Glover, J.N.M. (2011). Structural basis of cooperative DNA recognition by the plasmid conjugation factor, TraM. Nucleic Acids Res 39, 6775–6788.

233

Wozniak, R.A.F., Fouts, D.E., Spagnoletti, M., Colombo, M.M., Ceccarelli, D., Garriss, G., Déry, C., Burrus, V., and Waldor, M.K. (2009). Comparative ICE Genomics: Insights into the Evolution of the SXT/R391 Family of ICEs. PLoS Genet 5, e1000786.

Wright, L.D., Johnson, C.M., and Grossman, A.D. (2015). Identification of a Single Strand Origin of Replication in the Integrative and Conjugative Element ICEBs1 of Bacillus subtilis. PLoS Genet 11.

Wu, W., Feng, Y., Tang, G., Qiao, F., McNally, A., and Zong, Z. (2019). NDM Metallo-β- Lactamases and Their Bacterial Producers in Health Care Settings. Clin. Microbiol. Rev. 32, e00115-18.

Zupan, J., Hackworth, C.A., Aguilar, J., Ward, D., and Zambryski, P. (2007). VirB1* Promotes T-Pilus Formation in the vir-Type IV Secretion System of Agrobacterium tumefaciens. J Bacteriol 189, 6551–6563.

234