The key role of arbuscular mycorrhizal (AM) fungi in maintaining soil fertility and ecosystem functioning and their general sensitivity to pesticides make them good candidate bioindicators in pesticide soil microbial toxicity assessment. We investigated the impact of the herbicide nicosulfuron on mycorrhizal colonization and community structure of AM fungi via a pot-to-field experimental approach. This allowed the assessment of nicosulfuron toxicity (i) at extreme exposure schemes (pot experiment, Tier I) invoked by the repeated application of a range of dose rates (x0, x10, x100, x1000 the recommended dose) and (ii) under realistic exposure scenarios (x0, x1, x2, x5 the recommended dose) in the field (Tier II). In the pot experiment, the x100 and x1000 dose rates significantly reduced plant biomass, mycorrhizal colonization and AM fungal richness as determined by DGGE. This coincided with the progressive accumulation of herbicide concentrations in soil. In contrast, no effects on AM fungi were observed at the nicosulfuron dose rates tested in the field. Clone libraries showed that the majority of AM fungi belonged to the Glomus group and were sensitive to the high levels of nicosulfuron accumulated in soil at the latter culture cycles. In contrast, a Paraglomeraceae and a Glomus etunicatum ribotype were present in maize roots in all cycles and dose rates implying a tolerance to nicosulfuron-induced stress. Overall, the deleterious effects of nicosulfuron on AM fungi induced by the highest dose rates in the pot experiment could be attributed either to fungal-driven toxicity or to plant-driven effects which have subsequent implications for mycorrhizal symbiosis. We suggest that the tiered pot-to-field experimental approach followed in our study combined with classic and standardized molecular tools could provide a realistic assessment of the toxicity of pesticides onto AM fungi as potential bioindicators.