Plants produce an array of defensive compounds with toxic or deterrent effects on insect herbivores. Pollen can contain relatively high concentrations of such defense compounds, but the causes and consequences of this enigmatic phenomenon remain mostly unknown. These compounds could potentially protect pollen against antagonists but could also reduce flower attractiveness to pollinators. We combined field observations of the pollen-rewarding Lupinus argenteus with chemical analysis and laboratory assays to test three hypotheses for the presence of pollen defense compounds: (1) these compounds are the result of spillover from adjacent tissues, (2) they protect against pollen thieves, and (3) they act as antimicrobial compounds. We also tested whether pollen defense compounds affect pollinator behavior. We found a positive relationship between alkaloid concentrations in pollen and petals, supporting the idea that pollen defense compounds partly originate from spillover. However, pollen and petals exhibited quantitatively (but not qualitatively) distinct alkaloid profiles, suggesting that plants can adjust pollen alkaloid composition independently from that of adjacent tissues. We found no relationship between pollen alkaloid concentration and the abundance of pollen thieves in Lupinus flowers. However, pollen alkaloids were negatively associated with bacterial abundance. Finally, plants with more alkaloids in their pollen received more pollinator visits, but these visits were shorter, resulting in no change in the overall number of flowers visited. We propose that pollen defense compounds are partly the result of spillover from other tissues, while they also play an antimicrobial role. The absence of negative effects of these compounds on pollinator visitation likely allows their maintenance in pollen at relatively high concentrations. Taken together, our results suggest that pollen alkaloids affect and are mediated by the interplay of multiple interactions.