Stability of soil microbial communities to applications of the fungal biological control agent Metarhizium brunneum

Naturally occurring entomopathogens, including fungi, bacteria, protozoa, nematodes and viruses, are important for natural regulation of insect populations and therefore used in biological control of insect pests. Biological control using entomopathogens is an alternative pest management strategy to chemical control, is considered to be more environmental friendly than chemical control and is often part of integrated pest management (IPM). Beneficial effects of entomopathogens as compared to chemicals may include conservation of other natural enemies, reduction of pesticide residues in soil and plants as well as safety for non-target organisms. The development of existing biological control for important soil dwelling pests in Europe by exploiting novel application strategies and synergistic effects of entomopathogenic fungi, entomopathogenic nematodes and natural substances, has been the major aim of the EU-project supporting this thesis (INBIOSOIL). Biological control using entomopathogenic fungi may include soil applications of large quantities of fungal propagules, often resulting in densities of up to 1014 propagules per ha. Such mass applications and introductions of microorganisms to soil may affect indigenous soil microbial communities and the ecosystem functions they fulfil. Assessment of potential effects of applied microorganisms on native soil microbial communities is therefore important and also required in the registration process of novel products by European and Swiss regulations. Species of the entomopathogenic fungal genus Metarhizium are widely used in biological control and potential effects of Metarhizium on native soil microbial communities have barely been investigated. As a basic principal of risk assessment of mass applications it is mandatory to verify that soil microbial communities are exposed to the applied fungal biological control agent (BCA). It is important to be able to determine both exposure (presence and abundance) as well as effects, e.g., changes in microbial community structures, of a BCA in soil as basis of a risk assessment. Therefore, the aims of this thesis were I) to improve molecular tools to assess presence and abundance of the BCA in soil in order to track exposure of soil microbial communities and II) to assess potential effects of applications of two strains of the BCA on soil microbial communities in pot and field experiments and III) to investigate potential analytical constraints of the presence of one highly abundant microbial strain on the assessment of changes in soil microbial community structures. Exposure of soil microbial communities to the BCA was assessed by isolation of Metarhizium spp. from soil on selective medium and subsequently identifying the genotype of the applied strains using simple sequence repeat (SSR) marker analyses. To optimize the typing tool the transferability of 41 existing SSR markers to different Metarhizium spp. were investigated (Chapter 2). This was particularly important since structure and taxonomy within the genus Metarhizium have been drastically revised in the recent past. Among other changes, the species M. anisopliae has been recognized as a species complex and has been divided into 10 species. Suitable markers were selected for genotyping of the applied strains and discriminating them from native Metarhizium strains present in soils. Biocontrol experiments were performed to control Agriotes obscurus larvae in a pot and a field experiment and Diabrotica virgifera virgifera larvae in a pot experiment using the BCA M. brunneum ART2825 and M. brunneum EAMa01/58-Su over a period of four months (Chapter 3 and 4). Treatments included three different formulations of the BCA, i.e. fungus colonized barley kernels (FCBK), fungal capsules (Fcap) and fungal granules (Fgran), unformulated fungal spores, combinations of the BCA and garlic extract, garlic extract alone and controls. Potential effects of M. brunneum on fungal and prokaryotic soil communities were assessed using next generation sequencing (NGS) of ribosomal marker regions, i.e., the internal transcribed spacer 2 (ITS2) for fungal soil communities and part of the small subunit of the ribosomal RNA gene (16S V3-V4) for prokaryotic communities. One of the biases occurring in NGS is preferential amplification of target sequences in the PCR due to the occurrence of one highly abundant sequence. In order to assess if the sequence of the applied strain, which was highly abundant, will affect the assessment of changes of soil microbial communities, an experiment was performed in which different concentration of a plasmid containing the target sequence of the BCA was added to a soil DNA extract (Chapter 5). Results revealed that M. brunneum ART2825 formulated as FCBK was the most efficient treatment against A. obscurus larvae in the pot experiment, which resulted in 77 % reduction of damaged potato tubers compared to the untreated control (Chapter 3). No biocontrol effect was detected in the Agriotes-field experiment as well as in the Diabrotica-pot experiment (Chapter 3 and 4). Selective plating and SSR marker analyses confirmed exposure of soil microbial communities to the applied strains in the pot and field experiments and built the basis for effect analyses (Chapter 3 and 4). In the Agriotes-pot experiment, M. brunneum ART2825 affected the fungal community structures only slightly and only if formulated. These effects were in the same range as effects caused by the carrier material only and no effects were detected if pure fungal spores were applied to soil. In contrast, prokaryotic community structures were not affected by any fungal applications. Also, fungal and prokaryotic communities did not change upon the application of the BCA in the Diabrotica-pot and the Agriotes-field experiment. Finally, assessment of effects, i.e., changes in soil microbial community structures, were not impeded by the presence of one highly abundant sequence or sequencing errors that potentially corroborate the reliability of the results (Chapter 5). In conclusion, genotyping using SSR marker analyses was a valuable tool for confirming inoculation success and exposure of soil microbial communities to the applied strains. Although, constraints of amplicon sequencing, such as PCR bias, erroneous sequences and limitations in species identification, may occur, it is currently the most suitable technique, for the sequencing depth and the large number of samples and replicates required in experiments as performed in this study. Natural fluctuations including spatial and temporal differences of soil microbial community structures were similar or greater than any treatment-effects detected by applying different formulations of the BCA or unformulated fungal spores. Therefore, it is unlikely that applications of M. brunneum ART2825 and M. brunneum EAMa01/58-Su adversely affect soil microbial communities beyond the extent induced by fluctuations occurring in natural habitats. However, the findings of this study are restricted to specific BCA strains, soils, experimental designs, and analytical procedures and may require confirmation in other systems or if new analytical tools become available