Understanding the molecular genetic basis of virulence in the parasitic weed, Striga asiatica

Parasitic weeds of the genus Striga are a major constraint to an efficient and profitable agricultural system in Sub-Saharan Africa, causing up to 100% losses in crop yields which disproportionately affect the poorest subsistence farmers. The mode of action for root parasitic plants such as Striga is in penetration of the host root cortex to form a xylem-xylem connection through which water and nutrients can be derived, while evading or suppressing the host immune system. A promising strategy for control of Striga spp. is the use of resistant crop varieties, but to develop a durable and broad resistance to Striga it is also necessary to understand the genetic basis of the parasite’s virulence and how virulence can vary between and within populations.

The species Striga asiatica is particularly devastating in Madagascar, one of the major rice growing countries in Africa and a major focus for agricultural research. Striga asiatica is preferentially autogamous, and it is known that naturally selfing species tend to show a greater level of host adaptation to their sympatric hosts. Therefore, gene-for-gene interactions between virulence factors of S. asiatica and resistance genes in rice hosts in this evolutionary arms race
were considered to be likely. In this study, S. asiatica accessions, based on field site location, were sampled across Africa with a focus on Madagascar. Rhizotron-based virulence screens showed that variation in Striga virulence was influenced by spatial separation of S. asiatica accessions and, to a greater extent, by host variety. However, this variation in virulence
did not appear to reflect a race-like structure across the sampled S. asiatica accessions, as is seen with the autogamous species, S. gesnerioides.

Whole genome resequencing was performed on 47 individuals. Genotyping revealed well defined phylogenetic lineages between countries. Accessions within Madagascar contained multiple selfing lineages suggesting gene flow between sites. Selfing rates across the species within Madagascar were estimated to be at least 95%. A novel genotype-environment association study using redundancy analysis was employed to test associations of 400,000 genic SNPs with virulence phenotypes across rice hosts. A large number of candidate adaptive loci were discovered. Utilising homologous Arabidopsis annotations, highly enriched gene ontology terms for S. asiatica virulence candidates included several commonly found cell wall degrading enzymes such as pectin acetylesterases and glycosyl hydrolases.

Overall, this thesis has combined phenotypic and genotypic techniques to more clearly characterise the variation in S. asiatica. The major findings concluded that virulence is influenced by both spatial and host variety factors, suggesting that in the host-parasite relationship the virulence of S. asiatica accessions is not determined in a race-like manner. Whole genome resequencing identified distinct phylogenetic lineages among S. asiatica accessions with multiple selfing lineages, and candidate adaptive loci related to virulence, including cell wall degrading enzymes. This study effectively utilised the recent S. asiatica reference genome and redundancy analysis as tools for prediction of virulence-associated genes all of which can be built upon for future resistance breeding efforts