Previous work in our laboratory validated a set of 12 microsatellite markers to distinguish between global populations of Ae. aegypti mosquitoes (6). We screened these markers in 8 mosquito specimens from the 2010 invasion in the Netherlands and compared their genotypes with those from 736 Ae. aegypti mosquito specimens from 15 reference populations around the world, including 4 Florida locations.

We analyzed 8 mosquitoes from 2 tire yards in the Netherlands, 2 mosquitoes from site 1 and 6 from site 2 (2). The samples consisted of individual legs preserved in 70% ethanol. These samples were compared with previously screened Ae. aegypti mosquito populations from 14 locations worldwide: Palm Beach County, Vaca Key, and Conch Key, Florida, USA; Houston, Texas, USA; Pijijiapan and Coatzacoalcos, Mexico; Dominica; Bolivar and Zulia, Venezuela; Rayong and Prachuabkhirikan, Thailand; Tahiti, French Polynesia; and Cairns and Townsville, Queensland, Australia. The number of mosquitoes analyzed per reference population is indicated in Brown et al. (6). We also included in the analyses 47 newly acquired Ae. aegypti mosquito samples from Miami. Collection methods are described elsewhere (2,6).

Genomic DNA was extracted from each mosquito by using DNeasy kits (QIAGEN, Valencia, CA, USA). The samples from the Netherlands and Miami were screened for variation at 12 microsatellite loci following published methods (6,7). Chord distances between each pair of populations were calculated in GENETIX (8) and used in 2 distance-based cluster analyses: a principal components analysis using PAST (9) and a neighbor-joining network using MEGA4 (10). The Bayesian clustering algorithm in the program STRUCTURE (11) was used to identify genetic clusters and assign individual mosquitoes to these clusters with no a priori information about sampling locations. To determine the best genetic match for samples from the Netherlands, we conducted 5 independent runs for each assumed number of populations, K, 1–17. For all runs, we assumed an admixture model and correlated allele frequencies and used a burn-in value of 100,000 iterations followed by 500,000 replications. Results from STRUCTURE were visualized using DISTRUCT (12). A group assignment test was implemented in GENECLASS2 (13) to assign the mosquitoes in the Netherlands of unknown origin back to the reference populations with relative probabilities.

Figure

Figure. A) Principal components analysis based on pairwise population chord distances. The Adedes aegypti mosquito population in the Netherlands is represented by a blue square, the Florida, USA, populations by red crosses,...

Population-level (Figure, panels A, B) and individual-level (Figure, panel C) analyses suggest that the Ae. aegypti mosquito samples from the Netherlands are in the same genetic group as populations from southern Florida. Among these Florida populations, the group assignment test (13) identified Miami as the likely source of the samples from the Netherlands, with a relative probability of 100% compared with the other 14 reference populations. The recorded import of tires from the Miami area to the sites in the Netherlands where Ae. aegypti mosquitoes were discovered strongly corroborates the results from our genetic data, clearly indicating introduction of Ae. aegypti mosquitoes from Miami.

Our findings suggest that 1 of the world’s most dangerous vector arthropods entered Europe through a tire shipment from Miami. Although the importation of mosquitoes into the United States through the used tire trade has received considerable focus, our results indicate that equal caution should be exercised when tires are exported out of the southern United States, particularly into regions where Ae. aegypti mosquitoes are absent. Because vector exportation from the United States has now occurred multiple times (5,14), tires should be included as part of a comprehensive surveillance system to prevent future incidents.

Given the recent reemergence of dengue fever in Florida (15), we know that populations of Ae. aegypti mosquitoes from that region are fully capable of causing outbreaks of arboviral diseases. In the temperate climate of northern Europe, the epidemiologic risk is higher during the warm summer months, when viruses could be introduced to these new vector populations by travelers from tropical locations. This scenario would likely require close human–mosquito interactions at the site of the introductions. Overall, the risk is much greater in southern Europe, where the climate allows for year-round establishment of Ae. aegypti mosquito populations (3). Vector surveillance will prove crucial to prevent reinvasion of the region by this species of mosquitoes. In addition, cooperation between government scientists, policy makers, and companies involved in international trade is necessary domestically and internationally to determine the origins of exotic mosquito vector invasions, rather than fighting diseases as they occur.