UC Berkeley

In this dissertation I expand upon our knowledge in regards to the utility of population genetic approaches to be used for the study of the evolution of introduced biological control agents and their target pests. If biological control methods are to provide sustainable pest management services then more long-term studies will be necessary, and these studies should also include the use of population genetic approaches. For existing biological control programs, post-release population genetic studies could be initiated using museum voucher specimens for baseline data. In Chapter 2, I explored what factors influence our ability to extract usable genomic material from dried museum specimens, and whether we could use non-destructive techniques for parasitic hymenoptera. I found that the age of the specimen was the most important determinant for the amplification of PCR products, with nuclear loci having a higher probability of amplification from older specimens than mitochondrial loci. With these sequence results I was able to differentiate voucher specimens of different strains of the biological control agent Trioxys pallidus and I was able to confirm the identification of an unknown parasitoid reared from the invasive light brown apple moth.

For population genetic surveys to be conducted more frequently in biological control programs, some of the barriers to developing molecular markers that are variable enough for these types of surveys need to be overcome. One barrier is the time required to develop polymorphic microsatellite markers. Therefore, in Chapter 3, I developed a novel bioinformatics pipeline that searches through next-generation sequence (NGS) data and uses the raw sequencing reads to identify polymorphic loci. Using this approach I was able to rapidly develop microsatellite markers for two of my study species (T. pallidus and Chromaphis juglandicola). For both species more than 60% of the target markers amplified and were found to be polymorphic, compared to previous approaches where the success rates were much lower (published studies often show rates between 1 and 20%).

I then examined evolutionary factors that may affect the sustainability of two classical biological control programs; 1) the biological control of walnut aphids, and 2) the biological control of invasive knotweeds. The walnut aphid biological control program is a textbook successful biological control program, but has shown recent evidence of localized breakdowns, whereas the biological control program for invasive knotweeds is currently under review in the United States and Canada. In Chapter 4, I explored whether hybridization between introduced “strains” of T. pallidus is responsible for recent breakdowns in this control program. In that study I found low levels of hybridization – thus it is unlikely hybridization is playing an important role in these breakdowns – as well as evidence that two of the strains may actually be cryptic species; one being a specialist and another being a generalist. In Chapter 5, I explored whether the geographic mosaic theory of coevolution might help explain these localized breakdowns. In that chapter I found evidence for a geographic mosaic in the walnut aphid biological control program, and commented on how components of the geographic mosaic theory of coevolution can help us predict what systems we might expect localized breakdowns to occur in. In Chapter 6, I explored whether endosymbionts might play a role in shaping the host-associations of two strains of the candidate biological control agent for invasive knotweeds. I found that while strains of the psyllid Aphalara itadori showed no barriers to hybridization of their nuclear genomes, there were curious patterns of horizontal transmission of their primary endosymbiont. I also found that one haplotype of the secondary endosymbiont Sodalis sp. dramatically changed in frequency during the hybrid crosses reared on giant knotweed. When compared with previous studies of this species, the results I observe suggest that endosymbionts may play an important role in the differing fitness levels of these two strains.

In conclusion, population genetic approaches provide valuable tools for the study of post-release dynamics in biological control settings. While biological control programs promise to be useful study systems for evolutionary interactions, post-release studies will allow for that promise to come to fruition. In my future research endeavors I would like to continue to monitor the effects of hybridization and the frequency of geographic mosaics of coevolution in biological control settings. In addition, I would like to conduct post-release population genetic studies of both previous successful introductions and programs that resulted in failures. I believe these post-release studies will allow us to better determine what evolutionary factors affect the sustainability of biological control services and will allow for better management practices.