Résumé
CaMV is the type member of the genus Caulimovirus, infecting plants belonging to the Brassicaceae family, and is transmitted by aphid vectors through the noncirculative manner. The interaction between virus and vector during the acquisition phase requires the viral non-structural P2 protein, named Helper Component (HC) that links the virions to the aphid stylets. This specific molecular mechanism of transmission (presence of a HC) allows that one genome X may be transmitted by the HC coded by another genome Y. We baptized this phenomenon HC-trans-complementation. We propose here the hypothesis that HC-transcomplementation represent a strategy that allow the virus a better gestion of the constrains imposed by each demographic bottlenecks surveining during each transmission event.Following the bottleneck, the decrease in fitness may be due to random fixation of deleterous mutations. We made four series of experiments in which each bottleneck was constituted by one, two, five, or ten viral genomes. Despite four bottlenecks, our experiment could not be carried out, due to technical difficulties.Recombination could be a way to re-create free-mutation viral genomes. We have estimated the frequency of recombination of CaMV during an infection cycle within one host plant. Viral genomes coming from five plants co-inoculated with two quasi-isogenic clones were analysed. The proportion of recombinant genomes present in the viral population of CaMV seems to be very high and the recombination rate seems evenly distributed along the genome present in one population.Our results on pre-transcriptional control of P2 allow us to review the splicing process during the CaMV cycle. we show that, in absence of the protein P2 in planta, splicing of 70% viral 35S RNA is no more essential for virus infectivity, which is the case for the wild-type virus. Thus, HC may provoke, in an indirect manner, a selective cost through (i) the effort involved in regulation of the complex controlling splicing of all viral RNAs, and (ii) the synthesis of 70 % of the 35S RNA which will not serve as matrix during reverse transcription. On the other hand, our experimental results obtained from competition experiments between a self-transmissible and a non-self-transmissible isolate (with a deletion within the gene II), did not allow us to show directly a selective cost for producing a HC. Possible explanations for this unexpected result are discussed.