Le génome du bananier (Musa sp.) est envahi par un nombre important de séquences de Banana streak virus (BSV), virus à ADN double brin de la famille des Caulimoviridae qui n'a aucune étape d'intégration au génome hôte au cours de son cycle de multiplication. La majorité de ces intégrations eBSV (endogenous BSV) est défective mais certaines sont restées fonctionnelles et peuvent être à l'origine de particules virales suite à des stress. L'objectif de ce travail de thèse est de préciser si les eBSV sont maintenus ou non dans le génome de Musa balbisiana des bananiers et d'étudier les conséquences évolutives que cela engendre. Nous avons tout d'abord caractérisé les eBSV pour trois espèces de BSV (Banana streak goldfinger virus (BSGFV), Banana streak obino l'ewai virus (BSOLV), Banana streak imove virus (BSImV) présents dans le génome du bananier modèle M. balbisiana cv Pisang Klutuk Wulung (PKW). Nous avons montré que les intégrations eBSGFV et eBSOLV étaient di-alléliques avec un seul allèle fonctionnel à chaque fois, contrairement à eBSImV qui est mono-allélique et pour lequel nous n'avons pas pu identifier l'allèle à l'origine de l'infection. Leur contexte génomique d'intégration diffère avec une co-localisation d'eBSGFV et d'eBSOLV sur le chromosome 1 et d'eBSImV sur le chromosome 2. Ces résultats nous ont permis de développer les outils moléculaires nécessaires à la caractérisation de ces trois eBSV dans la diversité de M. balbisiana. Cette caractérisation a révélé la diversité de structures des eBSV et éclairé une partie encore inconnue de la phylogénie de l'espèce M. balbisiana. Dans un second temps nous avons étudié les mécanismes de régulation des eBSV. Ce travail a porté sur les mécanismes d'ARN interférent pouvant expliquer le maintien des eBSV dans le génome des bananiers. Cette analyse révèle que les eBSV sont effectivement sous contrôle d'un mécanisme de type ARNi et la forte production de petits ARNs de 24nt ciblant les eBSV suggère qu'il s'agit d'un silencing au niveau transcriptionnel (TGS). En parallèle, nous avons aussi recherché les mécanismes mis en place par les bananiers non-porteurs d'eBSV en cas d'infection afin de connaître les défenses constitutives des bananiers face à une attaque virale BSV. Nous avons, sur la base de ces résultats, proposé un modèle de régulation des eBSV et des BSV et discuté de l'impact que ces mécanismes auraient pu avoir sur l'évolution des eBSV. L'ensemble des données de ce travail ont permis de préciser les étapes évolutives qu'ont connues les eBSV dans le génome du bananier, expliquant le maintien que l'on observe aujourd'hui. / The nuclear genome of banana plants is invaded by numerous viral sequences of banana streak virus (BSV), a DNA virus belonging to the family Caulimoviridae which does not require integration for its replication. These endogenous BSV (eBSV) are mostly defective; however, some can release a functional viral genome following activating stresses. The objectives of this work were to identify if the eBSV are maintained or not in the M. balbisiana genome and to study the impacts of this on the evolution of banana plants. First, we characterized three functional eBSV sequences present within the Musa balbisiana cv PKW genome: (Banana streak goldfinger virus (BSGFV) ; Banana streak obino l'ewai virus (BSOLV) ; and, Banana streak imove virus (BSImV). We show that eBSOLV and eBSGFV are di-allelic with just one functional allele contrary to eBSImV which are mono-allelic and for which we cannot reveal the functional allele. Their genomic areas of integration are different and we also observe that eBSOLV and eBSVGFV are both on chromosome 2 whereas eBSImV is on chromosome 1. These results allowed us to develop the molecular tools required for the characterization of these 3 functional eBSVs within the diversity of M. balbisiana. This characterization has revealed the structural diversity of eBSV and has thus clarified previously unresolved details of M. balbisiana phylogeny. Secondly, we studied the regulatory mechanism of eBSV expression. This work investigated if RNA interference (RNAi) mechanisms could explain the maintenance of eBSV in the Musa genome. Our analyses have shown that, as expected, eBSV was under the control of RNAi mechanisms and the strong production of 24nt small RNAs that target eBSV suggests that Transcriptional Gene Silencing (TGS) was involved in this control. In parallel, we investigated the mechanisms implicated in the anti-viral defense during a BSV infection on a banana plant without eBSV in order to understand the constitutive defense of banana plants. On the basis of these results we have proposed a regulation model of eBSV and BSV and we discuss the impact of silencing regulation on eBSV evolution. Data accumulated during this work have clarified several steps in the co-evolutionary history of Musa sp. and eBSV and explain the maintenance of eBSVs in Musa genomes that we observe today.

Background: Cacao swollen shoot virus (CSSV), Cacao swollen shoot CD virus (CSSCDV), and Cacao swollen shoot Togo A virus (CSSTAV) cause cacao swollen shoot disease (CSSD) in West Africa. During 2000-2003, leaf and shoot-swelling symptoms and rapid tree death were observed in cacao in Cote d'Ivoire and Ghana. Molecular tests showed positive infection in only similar to 50-60% of symptomatic trees, suggesting the possible emergence of an unknown badnavirus. Methods: The DNA virome was determined from symptomatic cacao samples using Illumina-Hi Seq, and sequence accuracy was verified by Sanger sequencing. The resultant 14, and seven previously known, full-length badnaviral genomic and RT-RNase H sequences were analyzed by pairwise distance analysis to resolve species relationships, and by Maximum likelihood (ML) to reconstruct phylogenetic relationships. The viral coding and non-coding sequences, genome organization, and predicted conserved protein domains (CPDs) were identified and characterized at the species level. Results: The 21 CSSD-badnaviral genomes and RT-RNase H sequences shared 70-100% and 72-100% identity, respectively. The RT-RNase H analysis predicted four species, based on an >= 80% species cutoff. The ML genome sequence tree resolved three well-supported clades, with >= 70% bootstrap, whereas, the RT-RNase H phylogeny was poorly resolved, however, both trees grouped CSSD isolates within one large clade, including the newly discovered Cacao red vein virus (CRVV) proposed species. The genome arrangement of the four species consists of four, five, or six predicted open reading frames (ORFs), and the CPDs have similar architectures. By comparison, two New World cacao-infecting badnaviruses encode four ORFs, and harbor CPDs like the West African species. Conclusions: Three previously recognized West African cacao-infecting badnaviral species were identified, and a fourth, previously unidentified species, CRVV, is described for the first time. The CRVV is a suspect causal agent of the rapid decline phenotype, however Koch's Postulates have not been proven. To reconcile viral evolutionary with epidemiology considerations, more detailed information about CSSD-genomic variability is essential. Also, the functional basis for the multiple genome arrangements and subtly distinct CPD architectures among cacao-infecting badnaviruses is poorly understood. New knowledge about functional relationships may help explain the diverse symptomatologies observed in affected cacao trees.

Sugarcane bacilliform virus (SCBV) is an economically important pathogen of sugarcane in Australia which limits access to foreign sugarcane germplasm. Although SCBV is present in the major cane growing regions worldwide, very little is known about its variability, virulence and the yield losses resulting from infection. The limited information on SCBV has resulted in quarantine measures being introduced to protect the Australian sugarcane industry, with a major consequence being restricted access to imported sugarcane germplasm for breeding programs. Foreign sugarcane germplasm plays an important role in breeding of new commercial varieties for the Australian sugar industry and is essential for the long term productivity, profitability and sustainability of the sugar industry. This study was aimed at characterising Australian isolates of SCBV to enable the development of reliable and robust molecular and/or antibody-based diagnostic tests which could be used to not only assess the impact of SCBV on the Australian sugarcane industry, but could also be used to screen imported sugarcane germplasm for the virus. SCBV virions (SCBV-IM) were purified from the sugarcane accession "Ireng Maleng" and the dsDNA genome was cloned and sequenced. The genome of SCBV-IM comprised 7687 bp with an organisation typical of other badnaviruses. When the entire nucleotide sequence of SCBV-IM was compared to that of the Moroccan SCBV isolate (SCBV-Mo), less than 75% similarity was present. Within the coding regions, ORF I, ORF II and ORF III had 83%, 71% and 73% nucleotide similarity to SCBV-Mo, respectively. At the amino acid level, ORFs I, II and III from SCBV-IM showed 91%, 84% and 85% similarity to the equivalent regions in SCBV-Mo, respectively. To further investigate the level of sequence variability within Australian SCBV isolates, virions were purified from three further sugarcane accessions and a 220 bp fragment of the reverse transcriptase-coding region was amplified. Five clones from each sub-population were selected and sequenced. Analysis of these sequences revealed considerable variability in the virus population with variability within one plant as great as it was between isolates. However, since the use of specific primers could also be selecting for a sub-population of SCBV sequences, it was possible that the variability may actually be greater than that reported. These results indicated that SCBV isolates are complex and variable and may represent a continuum of genetic variability. High molecular weight DNA species larger than the SCBV 7.6 kbp unit-length genome were found in DNA extracted from purified SCBV-IM virions. We confirmed that these high molecular weight nucleic acids were virus-specific and open circular in conformation. Using field inversion gel electrophoresis (FIGE), the SCBV-IM DNA was separated into four discrete bands with sizes ranging from between 1 to 4 genome copies. The DNA was shown to comprise overlapped individual genome-length molecules and not covalently-bonded continuous DNA strands. We presume that these DNA molecules are concatamers formed during replication as a result of a terminal overlap on the sense strand. The presence of these concatamers within virions may explain the observation of particles with lengths corresponding to one, two or three times the modal length of 130 nm. Four SCBV-infected Saccharum officinarum plants were examined for the presence of integrated viral DNA. Southern blot analysis of viral DNA and total DNA extracted from the same plant source were compared with, or without, restriction digestion. The resulting restriction patterns from viral and total DNA were almost identical suggesting that there were no integrated SCBV sequences in the sugarcane cultivars tested. Although larger-than-single-genome copy bands were detected in both the viral and the total DNA samples, this was probably due to the presence of genomic concatamers. SCBV integration studies using Southern analyses were further complicated by high sequence variability which precluded the restriction digestion of all viral DNA species. As such, some of the SCBV DNA species remain as concatamers which appear as larger-than-unit-length SCBV products. An antiserum derived from a mixture of purified SCBV isolates has been used routinely in the past to screen for SCBV infection, but the heterogeneity reported for badnaviruses has cast doubt on the ability of this antiserum to detect all SCBV isolates. We attempted to determine whether antiserum generated against proteins other than the viral capsid could be used to detect SCBV infections, thus improving the reliability and robustness of SCBV diagnosis. The complete coding regions of SCBV ORF I and ORF II were bacterially expressed and used as antigens for antiserum production. Both ORF I and II proteins were found to be highly immunogenic and generated high-titre antisera, designated AS-I and AS-II, respectively. The diagnostic utility of both antisera to detect SCBV in six different infected sugarcane plants was tested using both immunosorbent electron microscopy (ISEM) and western blots. The currently used SCBV antiserum (AS-V), generated against a mixture of purified SCBV isolates, was included for comparison. In western analyses, neither AS-I nor AS-V was able to conclusively detect SCBV in any of the six infected plants due to reactivity with numerous non-specific proteins. In contrast, AS-II reacted specifically with a protein of the expected size (~13.5 kDa) in 2/6 infected plants. When compared using ISEM, AS-V, AS-I and AS-II trapped virions from 6/6, 6/6 and 2/6 SCBV-infected plants, respectively. However, the number of virions trapped using AS-V was approximately 30-fold more than that trapped using either AS-I or AS-II. These results highlight the variability between SCBV isolates and suggest that ISEM with antisera raised against mixtures of viral proteins may be a useful tool for the detection of viral isolates.

No description available.