The production of allelochemicals by the toxigenic dinoflagellate Alexandrium catenella is one of the suggested mechanisms to facilitate its bloom formation and persistence by outcompeting other phototrophic protists and reducing grazing pressure. In Southern California, toxic events caused by A. catenella and paralytic shellfish toxins (PSTs) regularly impact coastal ecosystems; however, the trophic interactions and mechanisms promoting this species in a food web context are still not fully understood. In the present study, we combined a dynamical mathematical model with laboratory experiments to investigate potential toxic and allelochemical effects of an A. catenella strain isolated off the coast of Los Angeles, Southern California, on competitors and a common zooplankton consumer. Experiments were conducted using three toxigenic strains of A. catenella, comparing the new Californian isolate (Alex Cal) to two strains previously described from the North Sea, a lytic (Alex2) and non-lytic (Alex5) strain, testing for donor density-dependent effects on two phytoplankton species (Rhodomonas salina, Tetraselmis sp.) and on the rotifer Brachionus plicatilis. Bioassays revealed a steep decline in competitor and consumer populations with increasing Alex Cal concentrations, indicating an intermediate lytic activity compared to the North Sea strains (lytic Alex2 and non-lytic Alex5). The rotifer fed and grew well on the PST- toxic, but non-lytic Alex5 strain, while its survival significantly decreased with increasing concentrations of the two lytic strains Alex Cal and Alex 2, indicating that negative effects on the rotifer were mediated by allelochemicals rather than PST-toxins. Mixed culture experiments including both competitors and consumers demonstrated that the intensity of allelochemical effects not only depended on the A. catenella density but also on the target density. Negative effects on grazers were alleviated by co-occurring competitors with a lower sensitivity to allelochemicals, thus reducing harmful compounds and allowing grazing control on the dinoflagellate to come into effect again. Results from mixed culture experiments were supported by the mathematical approach used in this study which was calibrated with data from simple monoculture growth, pairwise competition and predator-prey experiments, demonstrating the applicability of this model approach to predict the outcome of more complex food web dynamics at the community level.