Genomic analysis of variation in reproductive and stress traits in eucalypts

Eucalypts are a group of trees and shrubs distributed in Australia and some islands to its north, which encompass the most planted hardwood forestry species in the world. They occur naturally across many environmental conditions, including those where drought is frequent. Wildfire is also frequent in the Australian landscape, and eucalypts have evolved alongside fire, developed both adaptations to resist fire damage, along with reproductive and vegetative regenerative strategies dealing with wildfire. As climate change is increasing the frequency and intensity of drought and wildfire, improving our understanding of their interactions with the eucalypts on a genetic level is important. However, most traditional genetic research in eucalypts has focused on traits with direct economic importance, and to date we have limited knowledge on any aspect of the genetics underlying responses to these stressors. To improve our understanding on these issues, this thesis reports the application of genomic tools to the analysis of reproductive and stress traits in eucalypts, with a focus on the widely planted temperate species Eucalyptus globulus. Among the several notable features of the eucalypts are their petaline operculum and woody capsular fruit, which protect reproductive organs and seeds, respectively, from stressors such as fire. To study the genetic control of eucalypt operculum and capsule morphology, samples of these reproductive structures were collected from 206 trees that belong to a large outcrossed F2 Eucalyptus globulus mapping population growing in two field trials in Tasmania, Australia. The morphological variation in these structures was characterized by measuring six operculum and five capsule traits. Quantitative trait loci (QTL) analysis was performed using these data and a linkage map consisting of 480 markers. A total of 27 QTL were detected for operculum traits and 28 for capsule traits, with the logarithm of odds ranging from 2.8 to 11.8. There were many co-located QTL associated with operculum or capsule traits, generally reflecting allometric relationships. A key finding was five genomic regions with co-located QTL affecting both operculum and capsule morphology, and the overall trend for these QTL to affect elongation of both organs. Several candidate genes were found associated with the QTL. The results suggest that the reproductive structures of eucalypts share a common genetic pathway during flower bud development. Eucalypts are well-known for their ability to regenerate post-fire. To improve our understanding of the underlying genetic control of such responses, the genetic architecture of fire-related damage and recovery traits was studied in E. globulus. Fourteen traits were evaluated in an outcrossed F2 population, which was burnt by a wildfire 14 years after planting. The trial also included open-pollinated families of the grandparental dwarf and tall ecotypes used to produce the F2 population. Ecotype comparisons revealed that almost all traits were under genetic control, with trees of the dwarf ecotype significantly more damaged and mainly recovering from lignotubers, whereas tall ecotype trees mainly recovered from epicormic resprouts extending for a variable height up the stem. Genetic control of fire-related traits was confirmed by the detection of 38 QTL in the F2 population. Several QTL co-located and likely reflect pleiotropic effects. However, many independent QTL were detected, including QTL for fire damage (crown consumption and trunk scorch), epicormic resprouting, resprout herbivory, and seedling establishment. These results argue that many genetically controlled mechanisms are responsible for variation in fire damage and recovery. The diverse environments occupied by the eucalypts mean that their exposure to drought is quite variable, even within species. To identify the underlying genetic basis of drought resistance in E. globulus, a greenhouse experiment was conducted with 262 open-pollinated families planted in ten complete replicates. Plant morphology, glaucousness and the presence of a physiological disorder known as intumescence were scored on each seedling. Sixty-eight days after planting the seed, watering was stopped and repeated visual measurements of drought damage were taken on each seedling until two thirds of the trial reached mortality. Genome-wide association studies (GWAS) were performed using 13,611 single nucleotide polymorphisms (SNP) and different statistical approaches. Results of the marker-trait associations revealed 1 significant SNP associated with glaucousness and 2 with intumescence. A single SNP was also strongly associated with drought damage, but was no longer significant when total leaf area was used as a covariable. A deep learning neural network confirmed that genetic variation among open-pollinated families had a minimal contribution to drought resistance in the experiment while plant morphological variables most explained the drought stress response of E. globulus seedlings. To improve our understanding of the genetic basis of drought stress in eucalypts across different species, the evolution of specific genes linked to drought tolerance was studied. The genes ECERIFERUM1 (CER1) and ECERIFERUM3 (CER3) encode enzymes involved in the biosynthesis of alkane waxes, a key component of the plant cuticle that contributes to drought tolerance. A genome-wide survey of these genes was performed using recently released genome assemblies of 28 Myrtaceae species, with 22 species from the main eucalypt lineage and 6 non-eucalypt Myrtaceae tree species. 250 genes and pseudogenes were manually annotated, with a near-ubiquitous single copy of CER3 and 2 to 10 CER1 copies per Myrtaceae species identified. Phylogenetic analysis suggested copy number variation in eucalypts is due to multiple tandem duplication events; some ancient (shared by all Myrtaceae species studied), and others more recent (present only in eucalypts). Inter-chromosomal translocations were discovered for both CER1 and CER3, along with recurrent loss of CER1 introns in the WAX2 domain. Despite the varied native environments occupied by the eucalypt species in this study, there were no statistically significant associations between intra-genic structural changes in CER1 or CER1 copy number and aspects of the environment they occupy (including aridity). These studies demonstrate the complex genetic control of reproductive and stress traits in eucalypts. Although several key loci were discovered for different traits, additional research will be required to thoroughly understand the genetics governing the eucalypt drought stress response and the complex mechanisms encoded by the specific genes involved. Overall, this thesis provides a foundation of knowledge to further our understanding of the genetic architecture of key adaptive traits in eucalypts, which will be of key importance given the uncertain environments of the future.