Genome-wide association study in Eucalyptus grandis for resistance to Leptocybe invasa

Abstract

Eucalyptus species are commercially grown for their desirable growth, wood properties as well as a potential source for biofuel production making it a highly valuable commodity in the forest industry. However, this species is being threatened by pests and pathogens, and one of the most important pests in the plantations is the galling insect, Leptocybe invasa Fisher and La Salle (Hymenoptera: Eulophidae), also known as the blue gum chalcid. This pest has caused severe damages on Eucalyptus species and hybrids planted commercially and it is a major problem in South Africa and other parts of the world. The speed at which the pest has spread and the extent of the damage it can cause has created a need to improve preventative measures against it. Several studies have shown that significant variation exists within and between Eucalyptus species and could serve as an opportunity to breed for L. invasa resistance. Variations in resistance and susceptibility have been reported within and between Eucalyptus species, suggesting that selection of resistant genotypes maybe a great option in reducing the incidence of L. invasa. Eucalyptus leaves have been reported to contain high proportions of terpenes and some functions as defence mechanisms against pests. Some studies have reported that susceptibility to L. invasa in Eucalyptus trees was associated with increased concentrations of terpenes such as γ-terpinene and resistance was associated with high concentrations of iso–pinocarveol. These terpenes can act as direct defence or indirect by attracting parasitoids of the pest, L. invasa. Furthermore, several studies have reported on qualitative and quantitative variations in terpene traits suggesting that these traits may confer resistance to L. invasa. A breeding programme is viable if there is existing variation in resistance and knowing how much of the variation exists for a particular trait can be addressed by performing a Genome–wide association studies (GWAS). GWAS have been successful in identifying SNP markers, genomic regions and candidate genes of important traits, including those for pest resistance in forest trees with some currently in place to drive an effective Marker Assisted Selection breeding programme. Therefore, the goal of this study was to investigate the genetic architecture of L. invasa resistance and terpene traits in a E. grandis half–sib population. This is an important step in Eucalyptus genetic research and will assist in breeding programmes of this globally important forestry genus. In the initial study, visual scoring was done on 563 insect–challenged Eucalyptus grandis trees, from 61 half–sib families using the following scale, 0– not infested, 1– infested showing evidence of oviposition but no gall development, 2– infested with galls on leaves, mid–ribs or petioles and 3– stunting and lethal gall formation. The trees were genotyped using the EUChip60K SNP chip and 15,445 informative SNP markers were identified in the test population. A GWAS was performed using a Multi–Locus Mixed Model (MLMM) analysis and identified 35 SNP markers on three genomic regions putatively associated with resistance to L. invasa. SNP analysis of a validation population of 494 E. grandis trees confirmed seven SNP markers that were also detected in the initial association analysis. Based on transcriptome profiles of resistant and susceptible genotypes from an independent experiment we identified several putative candidate genes including Resistant (R)–genes, NB–ARC and TIR–NBS–LRR genes. Our results suggest that Leptocybe resistance in E. grandis may be influenced by a few large–effect loci in combination with a few minor loci segregating in our test and validation populations. From the two putative R–genes identified, NB–ARC and TIR–NBS–LRR, we genotyped protein–coding regions of 15 trees which scored 0 thereafter referred to as resistant and 15 trees which scored 3 thereafter referred to as susceptible and identified non–synonymous and synonymous polymorphisms. Our results suggest that the variation observed can either contribute to or disrupt recognition of the pest by the host by changing the regulation of R–protein activity. In the second study, we used GWAS to identify SNP markers and putative candidate genes for terpene traits. The association was performed on three terpene traits, 1,8–cineole, γ–terpinene, p–cymene and 416 E. grandis trees from 75 half–sib families were genotyped using the EUChip60K SNP chip, 15,387 informative SNP markers were identified. Using a MLMM we identified SNP markers associated with the three terpene traits. Based on the detailed annotation for the E. grandis reference genome, putative candidate genes of the significant SNP markers were identified. Genes suggested to be involved in terpene biosynthesis for example GGPPS and IPPS were identified. The study identified a cluster of UDP–Glycosyltransferase genes which may be involved in terpene transport and storage. Our results showed that variation underlying the three–terpene profiles is influenced by a few minor loci in combination with a few major effect loci suggesting an oligogenic nature of the traits. The last chapter is a perspective article which explored the value of GWAS studies in forest trees against pest resistance, in particular L. invasa also the adoption and reported cases of Marker Assisted Selection (MAS). The study further discusses the limitations and prospects of GWAS in Eucalyptus. The perspective presented herein provides an important next step in Eucalyptus genetic research and will assist in breeding programmes of this globally important forestry genus.