Testing the role of arbuscular mycorrhizal fungi in plant adaptation to serpentine soil

AbstractTesting the role of arbuscular mycorrhizal fungi in plant adaptation to serpentine soilbyShannon Peters SchechterDoctor of Philosophy in MicrobiologyUniversity of California, BerkeleyProfessor Thomas D. Bruns, ChairThis dissertation explores a new theoretical and experimental framework in which plant edaphic adaptation is mediated through arbuscular mycorrhizal fungi (AMF). The first chapter describes the primary ecological relationship between adapted plants and AMF by examining the AMF assemblages associated with field populations of serpentine and non-serpentine adapted ecotypes of California native plant Collinsia sparsiflora. The second chapter tests for plant-fungal specificity between C. sparsiflora ecotypes and serpentine and non-serpentine AMF using a common garden greenhouse experiment. Chapter three tests if soil edaphic factors alone could shape distinct serpentine and non-serpentine AMF assemblages by sampling non-C. sparsiflora root AMF assemblages from adjacent serpentine and non-serpentine sites. The final chapter addresses the functional role of AMF in serpentine adaptation with a greenhouse experiment using serpentine and non-serpentine AMF and C. sparsiflora ecotypes grown in sterilized serpentine soil. I found that serpentine and non-serpentine adapted ecotypes of C. sparsiflora associate with distinct AMF assemblages- an Acaulospora 1OTU- dominated serpentine, and a Glomus 1 OTU-dominated non-serpentine plant ecotype AMF assemblage (Chapter 1). However, I also found a relationship between plant ecotype AMF assemblage and soil nutrients. Thus, this distinction between plant ecotype AMF assemblages might be explained two ways: 1) the plant ecotypes have a high specificity for particular AM fungi within a ubiquitous soil assemblage or 2) the plant ecotypes were tapping non-specifically into AMF assemblages shaped by edaphic factors. I tested the first scenario in Chapter 2 by growing C. sparsiflora serpentine and non-serpentine ecotypes in a common pool of serpentine and non-serpentine AMF and then identified the root AMF of each plant ecotype. I found that the mixing of serpentine and non-serpentine AMF soil inoculum as the source of the common garden resulted in a non-serpentine soil type. Consequently, while the C. sparsiflora ecotypes associated with distinct AMF assemblages within the common garden, overall the ecotype AMF assemblages resembled that of a non-serpentine soil (i.e. Glomus 1 dominated). Therefore, I found no evidence of host-specificity between C. sparsiflora serpentine and non-serpentine ecotypes and serpentine and non-serpentine AMF. However, these results do indicate that the soil may select the AM fungi and potential for host choice of AMF based on soil type is present. \tI tested the second scenario in Chapter 3. I found that serpentine and non-serpentine AMF assemblages are distinct from each other. Variance partitioning analysis showed that both soil edaphic factors (33.5%), and plant assemblages (25.6%) drove the distinction between serpentine and non-serpentine AMF assemblages. This study confirms that there is a strong ecological relationship between AMF and plant tolerance to serpentine soil - plants growing in serpentine soil associate with serpentine-tolerant AMF taxa.\tFinally, I tested for a functional difference between serpentine and non-serpentine AMF assemblages that directly impact C. sparsiflora growth and fitness on serpentine (Chapter 4). Only shoot dry weight showed a significant response to AMF source. I found that serpentine AMF significantly increased growth of hosts over non-serpentine AMF and AMF-free controls. This indicates that serpentine AMF have a specialized adaptation to serpentine conferring growth enhancement to hosts, but it is still unclear what this adaptation is or which function is contributing to growth enhancement. I also found trends that imply that C. sparsiflora serpentine adapted ecotypes have a greater response to AMF than non-serpentine ecotypes, but these trends were not significant.