Symbiotic bacterial communities in ants are modified by invasion pathway bottlenecks and alter host behavior.

Biological invasions are a threat to global biodiversity and provide unique opportunities to study ecological processes. Population bottlenecks are a common feature of biological invasions and the severity of these bottlenecks is likely to be compounded as an invasive species spreads from initial invasion sites to additional locations. Despite extensive work on the genetic consequences of bottlenecks, we know little about how they influence microbial communities of the invaders themselves. Due to serial bottlenecks, invasive species may lose microbial symbionts including pathogenic taxa (the enemy release hypothesis) and/or may accumulate natural enemies with increasing time after invasion (the pathogen accumulation and invasive decline hypothesis). We tested these alternate hypotheses by surveying bacterial communities of Argentine ants (Linepithema humile). We found evidence for serial symbiont bottlenecks: the bacterial community richness declined over the invasion pathway from Argentina to New Zealand. The abundance of some genera, such as Lactobacillus, also significantly declined over the invasion pathway. Argentine ants from populations in the United States shared the most genera with ants from their native range in Argentina, while New Zealand shared the least (120 vs. 57, respectively). Nine genera were present in all sites around the globe possibly indicating a core group of obligate microbes. In accordance with the pathogen accumulation and invasive decline hypothesis, Argentine ants acquired genera unique to each specific invaded country. The United States had the most unique genera, though even within New Zealand these ants acquired symbionts. In addition to our biogeographic sampling, we administered antibiotics to Argentine ants to determine if changes in the micro-symbiont community could influence behavior and survival in interspecific interactions. Treatment with the antibiotics spectinomycin and kanamycin only slightly increased Argentine ant interspecific aggression, but this increase significantly decreased survival in interspecific interactions. The survival of the native ant species also decreased when the symbiotic microbial community within Argentine ants was modified by antibiotics. Our work offers support for both the enemy release hypothesis and that invasive species accumulate novel microbial taxa within their invaded range. These changes appear likely to influence invader behavior and survival.