The durability of a control method for plant protection is defined as the persistence of its efficacy in space and time. It depends on (i) the selection pressure exerted by it on populations of plant pathogens and (ii) on the capacity of these pathogens to adapt to the control method. Erosion of effectiveness of conventional plant protection methods has been widely studied in the past. For example, apparition of resistance to chemical pesticides in plant pathogens or pests has been extensively documented. The durability of biological control has often been assumed to be higher than that of chemical control. Results concerning pest management in agricultural systems have shown that this assumption may not always be justified. Resistance of various pests to one or several toxins of Bacillus thuringensis and apparition of resistance of the codling moth Cydia pomonella to the Cydia pomonella granulovirus have, for example, been described. In contrast with the situation for pests, the durability of biological control of plant diseases has hardly been studied and no scientific reports proving the loss of efficiency of biological control agents against plant pathogens in practice has been published so far. Knowledge concerning the possible erosion of effectiveness of biological control is essential to ensure a durable efficacy of biological control agents on target plant pathogens. This knowledge will result in identifying risk factors that can foster the selection of strains of plant pathogens resistant to biological control agents. It will also result in identifying types of biological control agents with lower risk of efficacy loss i.e. modes of action of biological control agents that does not favor the selection of resistant isolates in natural populations of plant pathogens. An analysis of the scientific literature was then conducted to assess the potential for plant pathogens to become resistant to biological control agents.

After more than 70 years of chemical pesticide use, modern agriculture is increasingly using biological control products. Resistances to conventional insecticides are wide spread, while those to bio-insecticides have raised less attention, and resistance management is frequently neglected. However, a good knowledge of the limitations of a new technique often provides greater sustainability. In this review, we compile cases of resistance to widely used bio-insecticides and describe the associated resistance mechanisms. This overview shows that all widely used bio-insecticides ultimately select resistant individuals. For example, at least 27 species of insects have been described as resistant to Bacillus thuringiensis toxins. The resistance mechanisms are at least as diverse as those that are involved in resistance to chemical insecticides, some of them being common to bio-insecticides and chemical insecticides. This analysis highlights the specific properties of bio-insecticides that the scientific community should use to provide a better sustainability of these products.

During plaque purification of Spodoptera littoralis nucleopolyhedrovirus in S. littoralis Sl52 cell culture, a deletion mutant virus was isolated. Analysis of the biological properties of this mutant virus revealed an absence of per os infectivity of the occluded virus. Infectivity by injection of the non-occluded (budded) virus is not different between the wild-type and the deleted virus. Restriction analysis of the mutant virus genome revealed a 4·5 kb deletion within the NotI D fragment. The observed phenotype was mapped to the deleted region by rescue experiments. The deletion was characterized and the equivalent DNA fragment on the wild-type virus was sequenced. By co-transfecting the DNA of the deleted virus with plasmids derived from the wild-type virus, it was possible to determine that ORF 7 in this fragment is responsible for the observed phenotype. ORF 7, called pif (per os infectivity factor), is homologous to ORF 119 of Autographa californica nucleopolyhedrovirus. Similar ORFs are present in all sequenced baculoviruses. The product of this gene is an occlusion body-derived virion structural protein required only for the first steps of larva infection, as viruses being produced in cells expressing the gene but not containing it in their genomes are able to produce successful infections.