Dentre as espécies do gênero Yersinia, Yersinia enterocolitica é a espécie mais prevalente como causa de doença em humanos e animais. Y. enterocolitica é dividida em seis biotipos. Os biotipos 1B, 2, 3, 4 e 5 compreendem linhagens associadas à doença em humanos e animais, enquanto o biotipo 1A consiste de linhagens consideradas não patogênicas. Apesar de Y. enterocolitica biotipo 2 ser de importância clínica, há uma escassez de estudos no país, o que dificulta avaliar o envolvimento dessa bactéria como causa de doença em humanos e em animais, bem como, determinar o impacto de sua presença no meio-ambiente. O objetivo deste trabalho foi investigar o potencial patogênico, determinar o perfil de suscetibilidade a antimicrobianos e verificar a diversidade genotípica de linhagens de Y. enterocolitica biotipo 2 isoladas no Brasil. Foram estudadas 40 linhagens de Y. enterocolitica biotipo 2, isoladas de humanos (5), ambiente (34) e animal (1), entre os anos de 1979 e 1998. Ademais, nas análises filogenéticas, foram acrescidas 26 linhagens de Y. enterocolitica pertencentes aos outros biotipos, com o intuito de comparar as linhagens de Y. enterocolitica biotipo 2 aos biotipos 1A, 1B, 3, 4 e 5. As linhagens de humanos e animal foram sensíveis a todos os 14 antimicrobianos testados. Dentre as 34 linhagens de ambiente, sete (20,6%) foram resistentes a um ou dois antimicrobianos, sendo esses, amicacina, cefoxitina, gentamicina, e sulfametoxazol - trimetoprima. Todas as linhagens apresentaram os genes inv, ail, ystA, hreP, tccC e myfA. Os genes fepD e fes foram detectados em 39 (97,5%) linhagens, o gene virF foi encontrado em três (7,5%) linhagens, os genes ystB e fepA não foram detectados em nenhuma linhagem. Todas as linhagens apresentaram comportamento relacionado à virulência frente aos testes fenotípicos de atividade da pirazinamidase, hidrólise da esculina e fermentação da salicina. O dendrograma de similaridade genética de Enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) agrupou as linhagens de Y. enterocolitica biotipo 2 em cinco grupos denominados A, B, C, D e E. Todas as linhagens, com exceção de duas, apresentaram similaridade genética superior a 88,3%. O dendrograma de similaridade genética de Pulsed field gel electrophoresis (PFGE) agrupou as linhagens de Y. enterocolitica biotipo 2 em três grupos denominados I, J e K. A maioria das linhagens (72,5%) apresentou similaridade ii genética superior a 78,3%. O dendrograma de similaridade genética de Multilocus variable number tandem repeat analysis (MLVA) agrupou as linhagens de Y. enterocolitica biotipo 2 em dois grupos denominados O e P com similaridade genética superior a 37,7%. Pode-se concluir que o potencial patogênico das linhagens de Y. enterocolitica biotipo 2 foi evidenciado pela prevalência da maioria dos marcadores de virulência, bem como, pelo comportamento relacionado à virulência frente aos testes fenotípicos pesquisados. Algumas linhagens apresentaram-se resistentes a antimicrobianos de primeira escolha no tratamento de yersiniose, o que pode acarretar em falha terapêutica. Os resultados de ERIC-PCR e PFGE mostraram a alta similaridade entre as linhagens de Y. enterocolitica biotipo 2, sugerindo que as mesmas pouco se diferenciaram ao longo dos 19 anos e que possivelmente o meio ambiente tem sido uma fonte de contaminação para humanos e animais no Brasil. A técnica de MLVA agrupou as linhagens de Y. enterocolitica biotipo 2 quanto à sua origem e a técnica de ERIC-PCR agrupou as linhagens de Y. enterocolitica biotipos 1A, 1B, 2, 3, 4, e 5 quanto às diferentes patogenicidades características de cada biotipo. / Among the species of the genus Yersinia, Yersinia enterocolitica is the most prevalent species that cause illness in humans and animals. Y. enterocolitica is divided into six biotypes. Biotypes 1B, 2, 3, 4 e 5 comprise strains associated to illness in humans and animals, while biotype 1A comprise strains considered nonpathogenic. Despite of the fact that Y. enterocolitica biotype 2 is of clinical importance, there is a paucity of studies in this country, which makes difficult to assess the involvement of this bacteria as a cause of illness in humans and animals, as well as to determine the impact of its presence in the environment. The aim of this work was to investigate the pathogenic potential, to determine the antimicrobial resistance profile and to verify the genetic diversity of Y. enterocolitica biotype 2 strains isolated in Brazil. Forty strains of Y. enterocolitica biotype 2 isolated from humans (5), environment (34) and animal (1), between 1979 and 1998 were studied. Besides, in the phylogenetic analyzes it was added 26 Y. enterocolitica strains belonging to the other biotypes, in order to compare the Y. enterocolitica biotype 2 strains to biotypes 1A, 1B, 3, 4 e 5. Humans and animals strains showed susceptibility to all 14 antibiotics tested. Among the 34 environment strains, seven (20.6%) were resistant to one or two antibiotics used such as amikacin, cefoxitin, gentamicin and sulfamethoxazole-trimethoprim. All the strains presented the genes inv, ail, ystA, hreP, tccC and myfA. Genes fepD and fes were detected in 39 (97.5%) strains, virF was found in three (7.5%) strains, and ystB and fepA were not detected in any strains. All the strains exhibited behavior related to virulence against the phenotypic tests of pyrazinamidase activity, esculin hydrolysis and salicin fermentation. The dendrogram of genetic similarity of Enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) grouped the Y. enterocolitica biotype 2 strains in five groups, designated A, B, C, D and E. All the strains, except two, showed a genetic similarity of more than 88.3%. The dendrogram of genetic similarity of Pulsed field gel electrophoresis (PFGE) grouped the Y. enterocolitica biotype 2 strains in three groups, designated I, J and K. The majority of the strains (72.5%) showed a genetic similarity of more than 78.3%. The dendrogram of genetic similarity of Multilocus variable number tandem repeat analysis (MLVA) grouped the Y. enterocolitica iv biotype 2 strains in two groups, designated O and P with a genetic similarity of more than 37.7%. It is possible to conclude that the pathogenic potential of the Y. enterocolitica biotype 2 strains was highlighted by the prevalence of the majority of the virulence markers searched, as well as by the behavior related to virulence against the phenotypic tests. Some strains were resistant to antimicrobials that are the first choice for yersiniosis treatment, which can result in therapeutic failure. The results of ERIC-PCR and PFGE showed a high genetic similarity between the Y. enterocolitica biotype 2 strains, suggesting that the strains differed little over 19 years, and that the environment has been possibly a source of humans and animals infections in Brazil. The MLVA technique grouped the Y. enterocolitica biotype 2 strains according their origins, and the ERIC-PCR technique grouped the Y. enterocolitica biotypes 1A, 1B, 2, 3, 4 and 5 strains according to the different pathogenicity characteristics of each biotype.

Shigella spp. está entre as quatro bactérias mais isoladas de fezes diarreicas no Brasil. No mundo cerca de 164,7 milhões de casos de shigelose ocorrem anualmente, sendo a maioria em países em desenvolvimento. O gênero Shigella spp. possui quatro espécies, sendo Shigella sonnei e Shigella flexneri as espécies mais frequentemente isoladas no Brasil e no mundo. O monitoramento de linhagens resistentes de Shigella spp. é essencial, pois este garante uma terapia eficiente quando necessária. Especificamente, a maioria dos estudos realizados com linhagens de S. sonnei no país verificaram apenas a ocorrência dessa e há poucos estudos que investigaram o potencial patogênico e a diversidade genotípica dessa espécie. Os objetivos desse projeto foram analisar o potencial patogênico, o perfil de resistência a antimicrobianos e a diversidade genotípica de linhagens de S. sonnei isoladas durante três décadas no Estado de São Paulo. No total foram caracterizadas 72 linhagens de S. sonnei isoladas de humanos, entre os anos de 1983 a 2014, quanto à presença de 12 genes de virulência por PCR, perfil de suscetibilidade frente a 16 antimicrobianos por disco difusão e tipagem molecular por Pulsed-field gel electrophoresis (PFGE), Enterobacterial repetitve intergenic consensus PCR (ERIC-PCR), Multiple-locus variable-number tandem-repeat analysis (MLVA) e 20 linhagens tipadas por Multi-locus sequence typing (MLST). Todas as linhagens apresentaram os genes de virulência ipaH, iuc e sigA. O gene ipaBCD foi encontrado em 14 (19%) linhagens, os genes ial e virF em 13 (18%) linhagens e o gene sen em sete (10%) linhagens. Os genes set1A, set1B, pic, sat e sepA não foram detectados. As mais altas taxas de resistência foram frente à sulfametoxazol-trimetoprim encontrada em 42 (58,3%) linhagens e frente à tetraciclina encontrada em 30 (41,6%) linhagens. Onze (15,5%) linhagens foram resistentes à ampicilina e piperacilina. Três (4,2%) linhagens foram resistentes à cefotaxima. Três (4,2%) linhagens foram resistentes ao cloranfenicol. Duas (2,8%) linhagens foram resistentes à ampicilina-sulbactam. Duas (2,8%) linhagens foram resistentes ao ácido nalidíxico. Uma (1,4%) linhagem foi resistente à amoxicilina-ácido clavulânico. Cinco (7%) linhagens foram multidroga resistentes (MDR). O dendrograma gerado pelo PFGE agrupou as 72 linhagens de S. sonnei em dois clusters designados PFGE-A e PFGE-B. O cluster PFGE-A agrupou 39 linhagens isoladas entre 1983-2014 com uma similaridade >=73,6% e mais especificamente 35 dessas linhagens apresentaram uma similaridade >=80,3%. O cluster PFGE-B agrupou 33 linhagens de S. sonnei isoladas entre 1984-2014 com uma similaridade >=74,7% e 27 dessas linhagens exibiram uma similaridade >=83,0%. Similarmente, o dendrograma gerado pelo ERIC-PCR agrupou as 72 linhagens de S. sonnei em dois clusters designados ERIC-A e ERIC-B. O cluster ERIC-A agrupou 37 linhagens isoladas entre 1983-2014 que exibiram uma similaridade >=78,8% e mais especificamente 36 dessas linhagens apresentaram uma similaridade >=82,3%. O cluster ERIC-B agrupou 34 linhagens de S. sonnei isoladas entre 1987-2014 que exibiram uma similaridade >=84,0%. Também por MLVA as linhagens foram agrupadas em dois clusters designados MLVA-A e MLVA-B. O cluster MLVA-A agrupou 31 linhagens isoladas entre 1983-2014 com uma similaridade >=40%. O cluster MLVA-B agrupou 41 linhagens isoladas entre 1983-2014 com uma similaridade >=21,6%. Todas as 20 S. sonnei foram tipadas por MLST como ST152. Conclui-se que o potencial patogênico das linhagens estudadas foi destacado pela presença de importantes genes de virulência. A alta porcentagem de resistência para alguns antimicrobianos testados, tais como, sulfametoxazol-trimetoprim e tetraciclina é preocupante e pode levar à falha terapêutica. Os resultados da tipagem molecular sugerem que existam dois subtipos prevalentes nas linhagens de S. sonnei estudadas que se diferenciaram pouco geneticamente e contaminaram humanos durante 31 anos na região metropolitana de Ribeirão Preto no Estado de São Paulo. O resultado do MLST indica que as linhagens estudadas de Shigella sonnei isoladas no Brasil descendem de um precursor comum / Shigella spp. is among the four most isolated bacteria from diarrheal faeces in Brazil. In the world about 164.7 million cases of shigellosis occur annually, mostly in developing countries. The genus Shigella spp. comprises four species, being Shigella sonnei and Shigella flexneri the most frequently isolated species in Brazil and worldwide. The monitoring of resistant strains of Shigella spp. is essential and ensures an effective therapy when necessary. Specifically, the majority of the studies with S. sonnei performed in the country verified only the occurrence of this bacterium and there are few studies that investigated the pathogenic potential and genotypic diversity of this species. The aims of this project were to analyze the pathogenic potential, antimicrobial resistance profile and genotypic diversity of S. sonnei strains isolated during three decades in the State of São Paulo. In total, 72 of S. sonnei strains isolated from humans, between the years 1983-2014, were characterized for the presence of 12 virulence genes by PCR, resistance profile against 16 antimicrobials by disk diffusion and molecular typing by Pulsed-field gel electrophoresis (PFGE), Enterobacterial repetitve intergenic consensus PCR (ERIC-PCR), Multiple-locus variable-number tandem-repeat analysis (MLVA) and 20 strains typed by Multi-locus sequence typing (MLST). All the strains contained the ipaH, iuc and sigA genes. The ipaBCD gene was detected in 14 (19%) strains, the ial and virF genes in 13 (18%) strains and the sen gene in seven (10%) strains. The set1A, set1B, pic, sepA and sat genes were not detected. The highest resistance rates were against trimethoprim-sulfamethoxazole found in 42 (58.3%) strains and against tetracycline found in 30 (41.6%) strains. Eleven (15.5%) strains were resistant to ampicillin and piperacillin. Three (4.2%) strains were resistant to cefotaxime. Three (4.2%) strains were resistant to chloramphenicol. Two (2.8%) strains were resistant to ampicillin-sulbactam. Two (2.8%) strains were resistant to nalidixic acid. One (1.4%) strain was resistant to amoxicillin-clavulanic acid. Five (7%) strains were multidrug resistant (MDR). The dendrogram generated by PFGE grouped the 72 S. sonnei strains into two clusters designated PFGE-A and PFGE-B. The PFGE-A cluster comprised, 39 S. sonnei strains isolated between 1983 and 2014 with a similarity above 73.6% and more specifically 35 of those strains exhibited a similarity >= 80.3%. The PFGE-B cluster grouped, 33 S. sonnei strains isolated between 1984 and 2014 with a similarity above 74.7%, and 27 of those strains exhibited a similarity above 83.0.Similarly, the dendrogram generated by ERIC-PCR grouped the 72 S. sonnei strains into two clusters designated ERIC-A and ERIC-B. The ERIC-A cluster comprised, 37 S. sonnei strains isolated between 1983 and 2014 that exhibited a similarity above 78.8% and specifically 36 strains of those exhibited a similarity >= 82.3%. The ERIC-B cluster grouped, 34 S. sonnei strains isolated between 1987 and 2014 that exhibited a similarity above 84.0%. Also, by MLVA strains were grouped into two clusters designated MLVA-A and MLVA-B. The MLVA-A cluster comprised 31 strains isolated between 1983 and 2014 with a similarity >=40%. The MLVA-B cluster comprised 41 strains isolated between 1983 and 2014 with a similarity >=21.6%. All the 20 S. sonnei were typed by MLST as ST152. In conclusion, the possible pathogenic potential of the strains studied was highlighted by the presence of important virulence genes. The high percentage of resistance to some of the antimicrobials tested such as trimethoprim-sulfamethoxazole and tetracycline is worrying and may lead to therapeutic failure. Molecular typing results may suggest that there are two prevalent subtypes of S. sonnei strains studied that differed little genetically and have been contaminating humans over 31 years in the metropolitan region of Ribeirão Preto in the São Paulo State in Brazil. The result of MLST indicates that the Shigella sonnei strains studied isolated in Brazil descended from a common precursor

L'industrie alimentaire cherche constamment à améliorer les processus de conservation des aliments. La biopréservation s’impose comme une alternative à l’utilisation de conservateurs dits «chimiques» et fait appel à des micro-organismes sélectionnés en tant que cultures protectrices pour leurs propriétés antimicrobiennes naturels. L'objectif de cette étude était de tirer profit de la diversité de Carnobacterium maltaromaticum afin d’identifier des souches capables d'inhiber la croissance du pathogène alimentaire Listeria monocytogenes. Pour cela, une méthode de typage moléculaire permettant la reconnaissance fine de souches a été établie, puis une méthode de criblage d’activités antibactériennes à robustesse élevée a été mise au point. La méthode de typage moléculaire est une méthode MLVA (Multiple-locus variable-number tandem repeat (VNTR) analysis) avec 3 loci de VNTR (VNTR-A, VNTR-B, VNTR-C) permet de discriminer 15 génotypes différents au sein d’une collection de 24 souches. Cela confirme la grande diversité génotypique et phénotypique présente au sein de la population de C. maltaromaticum et la performance de la méthode. La méthode de criblage est basée sur la réalisation d’essais de compétition à haut débit où chaque membre d’une collection de C. maltaromaticum a été co-cultivé avec une souche bioluminescente de L. monocytogenes. Il a été montré que la production de luminescence est le reflet de la croissance de L. monocytogenes en micro-culture et qu’il est ainsi possible de tester la stabilité du pouvoir antagoniste de chaque membre de la collection dans de multiples conditions de cultures différentes. Cette méthode a ainsi permis d’identifier des souches de C. maltaromaticum dont la propriété anti-L. monocytogenes est peu influencée par les conditions de cultures. Outre l’intérêt fondamental que présente cette méthode pour l’étude des interactions compétitives dans l’aliment, elle a permis d’identifier des souches présentant un fort potentiel de valorisation dans le domaine de la biopréservation alimentaire / The food industry is constantly seeking to improve food preservation processes. Biopreservation imposes itself as an alternative to the use of chemical preservatives and involves micro-organisms selected as protective cultures for their natural antimicrobial properties. The aim of this study was to benefit from the diversity of Carnobacterium maltaromaticum in order to identify strains capable of inhibiting the growth of Listeria monocytogenes, a food-borne pathogen. For this purpose, a molecular typing method that enables fine strain identification has been established, then a method of screening of antibacterial activities with high robustness has been developed. The molecular typing method is a MLVA (Multiple-locus variable-number tandem repeat (VNTR) analysis) with 3 VNTR loci (VNTR-A, VNTR-B, VNTR-C). It allowed to discriminate 15 different genotypes among a collection of 24 C. maltaromaticum strains. These results confirmed the high strain diversity within this species and showed the high performance of this method. The screening method is based on high throughput competition assays where each member of a collection of C. maltaromaticum strains was co-cultivated with a bioluminescent strain of L. monocytogenes. It was established that the bioluminescence of this strains is highly correlated to bacterial growth in micro-cultures, indicating that it is possible to investigate the stability of antagonistic properties of each collection member under multiple different culture conditions. The use of this method allowed to identify C. maltaromaticum strains with bioprotection properties that are poorly influenced by culture conditions. Besides the fundamental interest of this method for investigations of competitive interactions in food, it allowed to identify strains with high potential for bioprotection applications

Dans les îles du sud-ouest de l'océan Indien (SOOI) (Comores, Maurice, Mayotte, Réunion, Rodrigues et Seychelles), le flétrissement bactérien causé par le complexe d'espèces Ralstonia solanacearum (ceRs) est une phytobactériose considérée comme l'une des plus nuisibles pour les productions vivrières ou d'exportation. Les travaux de thèse présentés dans ce manuscrit avaient pour principal objectif l'exploration du niveau et de la distribution de la diversité génétique du ce Rs et de la structure génétique de ses populations dans le SOOI. Nous avons mené de vastes campagnes d'échantillonnage qui ont permis de constituer une large collection de 1704 isolats, principalement à partir de Solanacées (tomate, pomme de terre, piment, aubergine, poivron) et de géranium rosat. L'assignation phylogénétique des isolats a montré une très forte prévalence du phylotype I (88 %), qui est distribué dans chaque île du SOOI, tandis que les phylotypes II (9 %) et III (3 %) ne sont trouvés qu'à La Réunion. Deux souches de phylotype IV ont par ailleurs été signalées à l'île Maurice, représentant le premier rapport de ce groupe phylogénétique dans le SOOI. Une approche phylogénétique et de génotypage (MLSA/MLST) basée sur l'analyse de séquences de 6 gènes de ménage et 1 gène associé à la virulence (egl) a permis de révéler les relations génétiques entre 145 souches représentatives (diversité géographique + hôte d'isolement) du SOOI et 90 souches mondiales de référence. Le développement et l'application d'un schéma MLVA basé sur 17 séquences répétées en tandem (VNTR) sur près de 1300 souches a permis de révéler que les populations de phylotype I sont organisées en complexes clonaux dans le SOOI et que le niveau de diversité génétique est très contrasté selon les îles, Maurice présentant la plus forte diversité génétique. Un résultat majeur de cette thèse est la mise en évidence du déploiement d’une lignée génétique (sequevar I-31 ; STI-13 ; MT-035), surreprésentée dans les îles du SOOI, qui pourrait avoir été introduite via du matériel végétal contaminé depuis l'Afrique de Sud ou l’Afrique de l'Ouest. Nos études préliminaires montrent que l'haplotype majoritaire MT-035 (i) est le probable haplotype fondateur du complexe clonal le plus prévalent dans le SOOI, (ii) présente un pouvoir pathogène élevé (large gamme d'hôtes comprenant des plantes cultivées et des adventices, et forte agressivité sur Solanacées) et (iii) possède une forte aptitude à la compétition dans l'environnement via la production de bactériocines. Ces travaux permettront in fine de renforcer l'épidémiosurveillance et orienter les stratégies de lutte vis-à-vis de cet agent phytopathogène, notamment via le déploiement de cultivars résistants. / In the southwest Indian Ocean (SWIO) islands (Comoros, Mauritius, Mayotte, Réunion, Rodrigues and Seychelles), bacterial wilt caused by the Ralstonia solanacearum species complex (Rssc) is considered one of the most harmful plant disease for food crops or export. The main objective of this work presented in this manuscript was to explore the level and the distribution of the genetic diversity of Rssc and the genetic structure of its populations in SWIO. We conducted extensive sampling campaigns that resulted in a large collection of 1704 isolates, mainly from Solanaceae (tomato, potato, chilli, eggplant, pepper) and geranium rosat. The phylogenetic assignment of the isolates showed a very high prevalence of phylotype I (88 %), which is distributed in each island of the SWIO, while phylotypes II (9 %) and III (3 %) are found only in Réunion. Two phylotype IV strains have also been reported in Mauritius, representing the first report of this phylogenetic group in SWIO. A phylogenetic and genotyping approach (MLSA/MLST) based on sequence analysis of 6 housekeeping genes and 1 gene associated with virulence (egl) revealed the genetic relationships between 145 representative SWIO strains (geographic diversity + host) and 90 global reference strains. The development and application of MLVA scheme based on 17 variable number of tandem repeat sequences (VNTR) on nearly 1300 strains revealed that phylotype I populations are organized into clonal complexes in SWIO and that the level of genetic diversity is highly contrasted according to the islands, with Mauritius having the highest genetic diversity. This work highlights the deployment of a genetic lineage (Sequevar I-31, STI-13, MT-035), overrepresented in SWIO islands, which could have been introduced via contaminated plant material from South Africa or West Africa. Our preliminary studies show that the main haplotype MT-035 (i) is the probable founding haplotype of the most prevalent clonal complex in SWIO, (ii) has high pathogenicity (wide range of hosts including cultivated plants and weeds, and high aggressiveness on Solanaceae) and (iii) has a strong ability to compete in the environment via the production of bacteriocins. This work will ultimately strengthen epidemiosurveillance and guide control strategies of this plant pathogen, including the deployment of resistant cultivars.

Ralstonia solanacearum, une béta-proteobactérie d'origine tellurique, est l'une des phytobactérioses les plus nuisibles au niveau mondial. Cette bactérie est capable d'infecter plus de 250 espèces différentes dont certaines présentent un intérêt économique majeur (tomate, pomme de terre, tabac). R. solanacearum est divisée en 4 phylotypes distincts présentant des origines géographiques différentes : I (asiatique), IIA et IIB (américain), III (africain), IV (indonésien). Parmi ces phylotypes, le phylotype I est en expansion démographique, hautement recombinogène, réparti mondialement et possède une large gamme d'hôtes. Il possède donc un fort potentiel évolutif (sensu McDonald et Linde, 2002). Afin de contrôler cette bactérie, la lutte génétique reste la méthode la plus prometteuse : elle consiste à déployer des cultivars possédant différents sources de résistance (i.e., des gènes de résistance). La variété d'aubergine AG91-25 (E6) possède un gène majeur de résistance (ERs1) lui permettant de contrôler certaines souches de R. solanacearum de phylotype I. Cependant, la gestion de cette résistance requiert d'étudier au préalable sa durabilité afin d'en éviter le contournement. Cette durabilité peut être estimée en étudiant le potentiel évolutif d'un agent pathogène face à cette source de résistance, ainsi qu'en décryptant les mécanismes moléculaires de l'interaction entre l'hôte (gène R) et le pathogène (effecteur de types trois). Afin d'étudier la dynamique évolutive de R. solanacearum sous une pression de sélection exercée par la variété résistante E6, nous avons mis en place un essai d'évolution expérimentale au champ. Cet essai est composé de trois couples de microparcelles d'aubergines résistantes E6 et d'aubergines sensibles E8, implantées deux fois par an, pendant trois ans (soit 5 cycles). Un schéma MLVA (« Multi-Locus VNTR Analysis ») composé de 8 loci minisatellites a été développé afin de caractériser les souches extraites de ces cycles de cultures. Ces VNTR sont spécifiques aux souches de R. solanacearum de phylotype I, hautement polymorphes et discriminants à toutes les échelles : mondiale, régionale et locale. Nos résultats démontrent une absence de contournement de la résistance d'E6 par les populations parcellaires de R. solanacearum, confirmant le caractère durable de cette résistance. Cette variété aurait fortement réduit les populations bactériennes du sol, ne leur permettant plus d'infecter l'hôte résistant. Parallèlement, 100% des plants d'E8 sont morts à partir du cycle 2. La maladie au sein des microparcelles semble progresser selon une dynamique de « plante-à-plante ». Une baisse de la diversité génétique a aussi été observée au cours des cycles de culture répétés d'E8, associée à l'augmentation en fréquence de deux haplotypes. Cependant, aucune structuration génétique claire n'a été observée à l'échelle de la parcelle entière ou de la microparcelle. En revanche, les données d'isolement par la distance semblent indiquer qu'une structure spatiale semble être en cours d'établissement. L'ensemble de nos résultats suggère une structure épidémique clonale de nos populations parcellaires. Nous nous sommes aussi intéressés à l'implication de 10 ET3 dans l'interaction R. solanacearum vs aubergine résistante (E6). La distribution des 10 ET3 candidats est variable au sein d'une collection de souches phylogénétiquement diverses (91 souches) : ripAJ et ripE1 sont les ET3 les plus partagés alors que ripP1 et ripP2 sont les moins fréquemment. Certains ET3 présentent peu (ripAJ) voire pas (ripE1 et ripP2) de polymorphisme de taille, alors que d'autres (ripAU) sont extrêmement polymorphes. Cependant la composition en effecteurs d'une souche ne semble pas être corrélée à un phénotype sur aubergine E6. Nous avons identifié le gène d'effecteur ripAX2 comme ayant une fonction d'avirulence sur aubergine résistante E6. Sa reconnaissance par E6 semble s'opérer au niveau de la zone hypocotylaire. / Ralstonia Solanacearum is a soilborn beta-proteobacterium responsible of bacterial wilt on Solanaceaous crops. This bacterium is considered as one of the most harmful plant disease worldwide. This bacterium possesses the ability to infect more than 250 different species, including crops with major economic importance (tomato, potato, tobacco, eucalyptus…). R. solanacearum is divided into four phylotypes originated from different areas: I (Asian), IIA and IIB (American), III (African), IV (Indonesian). Among these phylotype, phylotype I is currently in demographic expansion, is highly recombinogenic and has a wide hosts range. Thus, altogether, these characteristics demonstrated that this phylotype has a high evolutionary potential (sensu McDonald and Linde, 2002). In order to control this bacterium, genetic plant resistance seems to be the most promising method. This method consists in using cultivars with different source of resistance such as resistance genes and/or resistant QTLs. The AG91-25 (E6), an eggplant cultivar possessing a major resistance gene (ERs1), is capable to control some of phylotype I strains of R. solanacearum. However, in order to optimize the management of this resistance and to avoid its fast breakdown, we need to deeply investigate the durability of this resistant gene. Durability can be estimated by studying the evolutionary potential of our pathogen faced to E6 source of resistance and by understanding the molecular mechanisms underlying the interaction between the host (R gene) and its pathogene (Type III Effector – T3E). In order to study R. solanacearum evolutionary dynamics under selective pressure from E6 resistant cultivar, we set up an experimental evolution trial in the field. This trial consisted of three couples of resistant (E6) and susceptible eggplants (E8) microplots, implanted twice a year during three years, hence consisting of 5 cycles. A Multi-Locus VNTR Analysis (MLVA) scheme, consisting of 8 minisatellite loci, was developed in order to characterize the strains extracted from these crop cycles. These VNTRs were specific to R. solanacearum phylotype I strains, they were highly polymorphic and discriminatory at different scale: globally, regionally and locally.Our results showed no breakdown of E6 resistance by R. solanacearum populations, which confirms that this resistance is durable. It seemed that this cultivar reduced the soil bacterial population, preventing bacterial population to infest the resistant host. At the same time, 100% of the E8 plants have died, starting at cycle 2. Bacterial wilt seemed to spread with a “plant-to-plant” dynamics within each microplot. Genetic diversity reduction was also observed during the successive cycle of susceptible eggplant, associated with the increase of frequency of two main haplotypes. However, we failed to identify a clear genetic structuration, neither at the plot scale nor at the microplot scale. Nevertheless, isolation-by-distance data seemed to show that a spatial structure is currently establishing. Altogether, our results suggested that our plot populations appeared to have a clonal epidemic structure.We also looked into 10 T3Es' involvement in the interaction between R. solanacearum and the resistant eggplant (E6). Their distribution was completely different within a collection of phylogenetically diverse strains (91 strains): ripAJ and ripE1 are the most shared T3Es whereas ripP1 and ripP2 were the less common T3E whithin our collection of strains. Some T3Es showed few (ripAJ) or no length polymorphism at all (ripE1 and ripP2) whereas some other (ripAU) are extremely polymorphic. Nevertheless, the T3E effector repertoire did not seemed to be correlated to a specific phenotype on E6 eggplant. Its recognition by E6 seemed to occur in the hypocotyle region rather than in the mesophyll, highlighting a possible organ-specificity of the interaction between ERs1 and ripAX2.

La compréhension de l'émergence des maladies infectieuses végétales causées par les bactéries passe par l'identification des populations sources, des routes d'invasion et des voies de dissémination, ainsi que par l'estimation des paramètres biotiques et abiotiques associés. Xanthomonas citri pv. citri (Xcc) est l'agent pathogène responsable du chancre Asiatique des agrumes. La bactérie est distribuée dans plusieurs pays agrumicoles du monde et listée comme organisme de quarantaine par ceux où elle est absente. Nous avons abordé l'épidémiologie moléculaire de Xcc à deux échelles spatio-temporelles grâce à un schéma de 14 microsatellites (MLVA-14) et un schéma de 31 marqueurs minisatellites (MLVA-31). Le typage MLVA-14 s'est montré adapté au génotypage d'une bactérie monomorphe comme Xcc. Le typage MLVA-31 a permis de diviser le pathovar Xcc en quatre groupes génétiques distincts correspondant aux différences de gammes d'hôtes mise en évidence chez cette bactérie. Le pathotype A (souches à large gamme d'hôtes parmi les rutacées) est séparé en deux groupes génétiques, tandis que les pathotypes A* et Aw (souches à gamme d'hôtes restreinte au limettier Mexicain et quelques espèces proches) constituent chacun un groupe génétique. Alors que l'expansion géographique de Xcc depuis son aire d'origine dans la première moitié du XXème siècle a quasi exclusivement concerné un seul groupe génétique, trois des quatre groupes décrits ont contribué à l'émergence de Xcc en Afrique au cours de la dernière décennie. La bactérie est pré-adaptée et a été introduite avec son hôte depuis sa population d'origine, faisant de la barrière migratoire la seule étape à franchir pour rencontrer un succès d'invasion. L'objectif de cette thèse a été de décrire les différentes populations émergentes grâce à des approches d'épidémiologie moléculaire et inférentielles, et identifier les différentes origines, routes et acteurs de la dissémination. Nous avons dans un premier temps montré la coexistence de deux groupes génétiques distincts au Mali : DAPC1 qui est dispersé dans quatre provinces du pays et DAPC2 qui est resté cantonné à l'espace péri-urbain de Bamako. L'analyse de l'émergence de Xcc au Sénégal a révélé le succès invasif de DAPC2 dans un autre environnement. La structure des populations DAPC1 du Mali et DAPC2 du Sénégal suggèrent que les plants de pépinières constituent une voie de dissémination majeure dans ces pays. À l'opposé, DAPC2 de Bamako n'est pas détecté en pépinières au Mali et n'a pas montré de caractère invasif. L'existence d'une population « tête de pont » invasive de souches DAPC1 au Mali donnant lieu à une épidémie au Burkina Faso a été mise en évidence par une approche ABC (calcul Bayésien approché). Les populations DAPC2 du Mali et du Sénégal ne présentent pas de lien épidémiologique direct mais partagent des liens de parenté avec des souches présentes dans le sous-continent Indien. Dans un deuxième temps, l'analyse d'une population de souches appartenant au pathotype A* en Ethiopie nous a permis de procéder à des estimations de paramètres démographiques, tels que les tailles efficaces. Nous avons montré que l'approche inférentielle nous permettait d'éclairer l'histoire démographique de Xcc dans un cas d'émergence, et de mettre en avant une dynamique saisonnière accentuant probablement le déséquilibre mutation-dérive lié à la situation d'émergence. L'émergence de Xcc en Afrique est principalement associée aux activités humaines. Sa dissémination locale et globale peut alors être considérablement limitée par des mesures de gestion plus stricte au niveau des pépinières et des flux de commerces. / Several plant emerging infectious diseases are caused by bacteria. To improve our understanding of their emergence, a description of the emerging populations, the reconstruction of invasion routes and pathways, as well as the identification of involved biotic and abiotic parameters are fundamental. The bacterium Xanthomonas citri pv. citri (Xcc) is responsible for Asiatic citrus canker. It is present in many citrus producing countries and listed as a quarantine organism in canker-free countries.Two MLVA schemes were used for molecular epidemiological analyses of Xcc. The first one, MLVA-14, targeted 14 microsatellite markers, and is useful to describe the genetic diversity of this monomorphic bacterium. The second, MLVA-31, targeted 31 minisatellite markers, is suited to global epidemiology analyses. Based on MLVA-31 data, Xcc is divided in four genetic groups (referred to as DAPC clusters) corresponding to Xcc pathotype classification based on host range. Three pathotypes were described: pathotype A strains are able to infect most citrus species, while pathotypes A* and AW strains are naturally restricted to fewer host species.DAPC 1 is responsible for almost all cases of geographical expansion of Xcc over the first half of the 20th century, we show that three Xcc genetic clusters have emerged in Africa over the last decade. Xcc is pre-adapted to its host species. Invasive success of Xcc is then mostly conditioned by migration events. Our objectives were to describe these invasive populations using a molecular epidemiology approach and to assess the origin, routes and actors of this dissemination in Africa. Two genetic clusters were found in Mali: DAPC1 is present in four provinces while DAPC2 is restricted to the Bamako urban environment. In contrast, DAPC2 emerging populations in Senegal showed an invasive succes in an other environment. Populations structures of DAPC1 in Mali and DAPC2 in Senegal highlighted the role of nurseries in Xcc dissemination. On the contrary, DAPC2 strains in Bamako were not detected in Malian nurseries and showed a limited invasive success. Approximate Bayesian Computation highlighted an invasive bridgehead scenario between DAPC1 in Mali and Burkina Faso. DAPC2 populations in Mali and Senegal were not found epidemiologically related but were genetically related to strains previously reported from the Indian subcontinent.Demographic parameters inference, such as effective population sizes, were inferred from Ethiopian pathotype A* populations. The inference approach was useful to decipher the demographic history of this emerging population, and suggested seasonal fluctuations in population sizes over time.Emergence of Xcc in Africa was found strongly related to human activities. Therefore, the local and global dispersion could be limited by a better management of nurseries and trade.

Dentre as espécies do gênero Yersinia, Yersinia enterocolitica é a espécie mais prevalente como causa de doença em humanos e animais. Y. enterocolitica é dividida em seis biotipos. Os biotipos 1B, 2, 3, 4 e 5 compreendem linhagens associadas à doença em humanos e animais, enquanto o biotipo 1A consiste de linhagens consideradas não patogênicas. Apesar de Y. enterocolitica biotipo 2 ser de importância clínica, há uma escassez de estudos no país, o que dificulta avaliar o envolvimento dessa bactéria como causa de doença em humanos e em animais, bem como, determinar o impacto de sua presença no meio-ambiente. O objetivo deste trabalho foi investigar o potencial patogênico, determinar o perfil de suscetibilidade a antimicrobianos e verificar a diversidade genotípica de linhagens de Y. enterocolitica biotipo 2 isoladas no Brasil. Foram estudadas 40 linhagens de Y. enterocolitica biotipo 2, isoladas de humanos (5), ambiente (34) e animal (1), entre os anos de 1979 e 1998. Ademais, nas análises filogenéticas, foram acrescidas 26 linhagens de Y. enterocolitica pertencentes aos outros biotipos, com o intuito de comparar as linhagens de Y. enterocolitica biotipo 2 aos biotipos 1A, 1B, 3, 4 e 5. As linhagens de humanos e animal foram sensíveis a todos os 14 antimicrobianos testados. Dentre as 34 linhagens de ambiente, sete (20,6%) foram resistentes a um ou dois antimicrobianos, sendo esses, amicacina, cefoxitina, gentamicina, e sulfametoxazol - trimetoprima. Todas as linhagens apresentaram os genes inv, ail, ystA, hreP, tccC e myfA. Os genes fepD e fes foram detectados em 39 (97,5%) linhagens, o gene virF foi encontrado em três (7,5%) linhagens, os genes ystB e fepA não foram detectados em nenhuma linhagem. Todas as linhagens apresentaram comportamento relacionado à virulência frente aos testes fenotípicos de atividade da pirazinamidase, hidrólise da esculina e fermentação da salicina. O dendrograma de similaridade genética de Enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) agrupou as linhagens de Y. enterocolitica biotipo 2 em cinco grupos denominados A, B, C, D e E. Todas as linhagens, com exceção de duas, apresentaram similaridade genética superior a 88,3%. O dendrograma de similaridade genética de Pulsed field gel electrophoresis (PFGE) agrupou as linhagens de Y. enterocolitica biotipo 2 em três grupos denominados I, J e K. A maioria das linhagens (72,5%) apresentou similaridade ii genética superior a 78,3%. O dendrograma de similaridade genética de Multilocus variable number tandem repeat analysis (MLVA) agrupou as linhagens de Y. enterocolitica biotipo 2 em dois grupos denominados O e P com similaridade genética superior a 37,7%. Pode-se concluir que o potencial patogênico das linhagens de Y. enterocolitica biotipo 2 foi evidenciado pela prevalência da maioria dos marcadores de virulência, bem como, pelo comportamento relacionado à virulência frente aos testes fenotípicos pesquisados. Algumas linhagens apresentaram-se resistentes a antimicrobianos de primeira escolha no tratamento de yersiniose, o que pode acarretar em falha terapêutica. Os resultados de ERIC-PCR e PFGE mostraram a alta similaridade entre as linhagens de Y. enterocolitica biotipo 2, sugerindo que as mesmas pouco se diferenciaram ao longo dos 19 anos e que possivelmente o meio ambiente tem sido uma fonte de contaminação para humanos e animais no Brasil. A técnica de MLVA agrupou as linhagens de Y. enterocolitica biotipo 2 quanto à sua origem e a técnica de ERIC-PCR agrupou as linhagens de Y. enterocolitica biotipos 1A, 1B, 2, 3, 4, e 5 quanto às diferentes patogenicidades características de cada biotipo. / Among the species of the genus Yersinia, Yersinia enterocolitica is the most prevalent species that cause illness in humans and animals. Y. enterocolitica is divided into six biotypes. Biotypes 1B, 2, 3, 4 e 5 comprise strains associated to illness in humans and animals, while biotype 1A comprise strains considered nonpathogenic. Despite of the fact that Y. enterocolitica biotype 2 is of clinical importance, there is a paucity of studies in this country, which makes difficult to assess the involvement of this bacteria as a cause of illness in humans and animals, as well as to determine the impact of its presence in the environment. The aim of this work was to investigate the pathogenic potential, to determine the antimicrobial resistance profile and to verify the genetic diversity of Y. enterocolitica biotype 2 strains isolated in Brazil. Forty strains of Y. enterocolitica biotype 2 isolated from humans (5), environment (34) and animal (1), between 1979 and 1998 were studied. Besides, in the phylogenetic analyzes it was added 26 Y. enterocolitica strains belonging to the other biotypes, in order to compare the Y. enterocolitica biotype 2 strains to biotypes 1A, 1B, 3, 4 e 5. Humans and animals strains showed susceptibility to all 14 antibiotics tested. Among the 34 environment strains, seven (20.6%) were resistant to one or two antibiotics used such as amikacin, cefoxitin, gentamicin and sulfamethoxazole-trimethoprim. All the strains presented the genes inv, ail, ystA, hreP, tccC and myfA. Genes fepD and fes were detected in 39 (97.5%) strains, virF was found in three (7.5%) strains, and ystB and fepA were not detected in any strains. All the strains exhibited behavior related to virulence against the phenotypic tests of pyrazinamidase activity, esculin hydrolysis and salicin fermentation. The dendrogram of genetic similarity of Enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) grouped the Y. enterocolitica biotype 2 strains in five groups, designated A, B, C, D and E. All the strains, except two, showed a genetic similarity of more than 88.3%. The dendrogram of genetic similarity of Pulsed field gel electrophoresis (PFGE) grouped the Y. enterocolitica biotype 2 strains in three groups, designated I, J and K. The majority of the strains (72.5%) showed a genetic similarity of more than 78.3%. The dendrogram of genetic similarity of Multilocus variable number tandem repeat analysis (MLVA) grouped the Y. enterocolitica iv biotype 2 strains in two groups, designated O and P with a genetic similarity of more than 37.7%. It is possible to conclude that the pathogenic potential of the Y. enterocolitica biotype 2 strains was highlighted by the prevalence of the majority of the virulence markers searched, as well as by the behavior related to virulence against the phenotypic tests. Some strains were resistant to antimicrobials that are the first choice for yersiniosis treatment, which can result in therapeutic failure. The results of ERIC-PCR and PFGE showed a high genetic similarity between the Y. enterocolitica biotype 2 strains, suggesting that the strains differed little over 19 years, and that the environment has been possibly a source of humans and animals infections in Brazil. The MLVA technique grouped the Y. enterocolitica biotype 2 strains according their origins, and the ERIC-PCR technique grouped the Y. enterocolitica biotypes 1A, 1B, 2, 3, 4 and 5 strains according to the different pathogenicity characteristics of each biotype.

Les répétitions en tandem sont constituées de successions de motifs d'ADN. Ces structures présentes dans tous les organismes, procaryotes comme eucaryotes, ont des applications dans de nombreux domaines. Depuis quelques années seulement, les répétitions en tandem sont étudiées chez les bactéries. Le polymorphisme associé à ces séquences peut être utilisé pour le génotypage de bactéries pathogènes, permettant une identification précise au niveau de la souche. Le polymorphisme des séquences répétées est de deux types : polymorphisme de longueur et mutations internes aux motifs. Les génomes des deux bactéries pathogènes responsables d'infections nosocomiales, Staphylococcus aureus et Pseudomonas aeruginosa, ont été étudiés dans le but d'identifier des séquences répétées polymorphes. Un ensemble de marqueurs polymorphes a été validé expérimentalement pour ces deux espèces permettant un typage dit MLVA (pour « Multiple Locus VNTR Analysis »). Le travail plus classique de typage par la taille de la répétition a été complété par un travail de séquençage de certains allèles. Les résultats obtenus montrent comment le typage « MLVA » complété si nécessaire par le séquençage d'allèles, pourraient constituer de nouvelles méthodes peu coûteuses participant au contrôle des infections bactériennes.

Dans le cadre de cette thèse, nous avons mis au point un outil de différenciation moléculaire performant, simple et transférable pour Bartonella henselae, basé sur la technique MLVA. 5 VNTR dits " principaux " ont été sélctionnés pour leur polymorphisme (BHVA- E). Ceci nous a permis d'évaluer la diversité des souches et/ou isolats de B. henselae. Avec ces 5 VNTR, et pour 178 isolats et /ou souches testés, un index de diversité de 0.98 et 99 profils ont été obtenus. Ces profils se répartissent en groupes A et B. Le groupe A n'inclut que des soldats félins, alors que le groupe B est constitué d'isolats félins, d'un isolat canin et de la totalité des isolats humains testés. Une étude réalisée sur des isolats de chats et de leurs propriétaires a montré que la technique MLVA est un outil efficace pour la traçabilité. Les VNTR les moins polymorphes semblent pouvoir jouer le rôle de marqueur géographique, alors que pour les BHV les plus polymorphes, certains allèles sont particulièrement associés aux isolats humains. Aucun profil commun aux génotypes I et II n'a été rencontré. Nos observations suggèrent par ailleurs que tous les isolats du groupe B, c'est-à-dire les isolats de génotype I (d'origine humaine et féline) ainsi que tous les isolats humains appartenant à l'un ou l'autre des 2 génotypes, pourraient être dérivés d'isolats félins de génotype II (groupe A). En outre, la technique MLVA s'est avérée capable de typer des " variants " de B. henselae issus de félidés sauvages. La comparaison des performances de la technique MLVA versus les autres techniques déjà développées telles que ECP, MLST et MST, utilisant des souches communes, a montré que la technique MLVA est plus discriminante que l'ensemble de ces techniques et par ailleurs plus stable que l'ECP. Enfin, nos résultats suggèrent que seul le groupe B serait zoonotique, et que parmi les 5 VNTR polymorphes intragéniques que nous avons utilisés, certains au moins joueraient un rôle dans le potentiel zoonotique, la persistance chez le chat et/ou la virulence pour l'Homme.

L'utilisation de virus de bactéries ou bactériophages pourrait être un complément efficace à l'antibiothérapie. Mon travail a porté sur la caractérisation de bactériophages dirigés contre l'espèce Pseudomonas aeruginosa, pathogène opportuniste responsable d'infections des voies respiratoires des patients atteints de mucoviscidose.J'ai tout d'abord déterminé la sensibilité des souches mucoviscidosiques au Pyophage (un cocktail de phages thérapeutiques Géorgien) et identifié six phages lytiques de quatre genres différents. Environ 15% des souches sont résistantes au Pyophage. Ensuite, en utilisant les souches cliniques multi-résistantes aux phages comme bactérie d'enrichissement, 32 phages ont été obtenus à partir des eaux usées de France et Côte d'Ivoire. Tous les phages analysés sont caudés et distribués au sein de dix genres parmi lesquels six exclusivement lytiques. J'ai identifié des souches bactériennes qui demeurent insensibles à tous les phages. J'ai montré que le système CRISPRs-Cas n'est pas associé à la résistance des souches aux phages lytiques.