Project description:<p><strong>INTRODUCTION:</strong> Larvae of the invasive emerald ash borer [EAB, <em>Agrilus planipennis</em> Fairmaire (Coleoptera: Buprestidae)], kill over 99% of green ash (<em>Fraxinus pennsylvanica</em> Marshall) trees they infest, yet a small percentage of green ash (“lingering ash”) survive years of heavy EAB attack. In the face of an ongoing invasion that threatens multiple North American <em>Fraxinus</em> species with extinction, any evidence for reproducible defensive responses in the native species merits investigation.</p><p><strong>METHODS: </strong>We evaluated the capacity of 3 families of green ash F1 progeny to kill EAB larvae when challenged in greenhouse studies by infestation with a uniform density of EAB eggs followed by dissection 8 weeks post-infestation and comparison of the host metabolomic profiles.</p><p><strong>RESULTS:</strong> The mean proportions of host-killed larvae in the 2 families of F1 progeny from lingering ash parents were significantly higher than that of host-killed larvae in the family of F1 progeny from susceptible ash parents (<em>p</em> < 0.001). Untargeted metabolomics comparing F1 progeny in the quartile with the highest percent host-killed larvae (HHK) to F1 progeny in the quartile with the lowest percent host-killed larvae (LHK) and to the uninfested F1 progeny within each family revealed evidence for induced biochemical responses to EAB. Infested trees produced significantly higher levels of select secoiridoids than uninfested trees, and LHK progeny produced significantly higher levels of select secoiridoids than the HHK progeny. HHK progeny produced significantly higher abundances of 3 metabolites annotated as aromatic alkaloids than the LHK and uninfested individuals.</p><p><strong>DISCUSSION: </strong>Based on these results, we hypothesize that green ash responds to EAB infestation. However, only certain trees have the genetic capacity to tailor a response that kills enough EAB larvae to prevent lethal damage to the vascular system. Rigorous tests of this hypothesis will require 15–20 years of additional crossing, phenotyping, and omics analyses. The results of this investigation will encourage the establishment and continuation of breeding programs that, in concert with biocontrol and management, could provide trees that slow, if not halt, the decimation of the <em>Fraxinus</em> gene pool. At the same time, ongoing work on host-insect interaction will contribute to our understanding of how forest trees recognize and defend themselves against phloem-feeding insects.</p>