Direct evidence that specific alleles are adaptive requires the functional connection of genotype, phenotype and fitness. Using CRISPR–Cas9 genome editing, a recent Nature study retraces the convergent evolution of resistance to cardiac glycosides in insects, pinpointing the adaptive consequences of individual alleles.

Credit: Martin Shields/Alamy

Cardiac glycosides are highly toxic plant secondary metabolites that act by inhibiting the sodium–potassium ATPase. Nonetheless, more than 100 insect species across six orders have evolved parallel amino acid substitutions in the α-subunit (ATPα) that enable them to feed from these plants and in some cases sequester the toxins.

Using maximum likelihood ancestral state reconstruction across a species phylogeny, the authors identified three amino acid residues (at sites 111, 119 and 122) within ATPα that frequently underwent substitutions associated with cardiac glycoside specialization. Comparison of a random permutation null model with the mutational order across 21 specialized lineages revealed that the order of substitutions was unlikely to be random.

Next, the team focused on the monarch lineage (Danainae), which includes butterfly species that do not feed on cardiac glycoside-producing plants and those that sequester the toxins, such as the monarch butterfly (Danaus plexippus). They generated Drosophila lines with precise substitutions at ATPα residues 111, 119 and 122 to reconstruct four genotypes that have evolved sequentially in the monarch lineage (Q111L, A119S, L111V and, lastly, N122H).

The Q111L substitution only mildly increased survival of knock-in fly lines fed the cardiac glycoside ouabain at lower concentrations, with no effects seen at higher concentrations. By contrast, Q111L flies that also harboured the A119S substitution showed an increase in survival also at higher ouabain concentrations, as did those with the subsequent L111V mutation. In flies with the ‘monarch’ genotype (that is, all four substitutions), even the highest ouabain concentrations had no impact on survival. Each consecutive substitution had a neutral-to-positive effect on target-site insensitivity, as determined by sodium pump enzymatic assays using head extracts, with ‘monarch flies’ exhibiting the same insensitivity to ouabain as monarch butterflies.

Finally, the authors generated single-substitution knock-in lines for N122H, often the last mutation to evolve, and for A119S, a substitution also present in non-specialized insects. These lines helped decipher the roles of antagonistic pleiotropy and epistasis of individual substitutions, which shape the resistance-conferring mutational path. For example, A119S was found to counter the pleiotropic ‘costs’ of substitutions at sites 111 and 122 throughout the adaptive walk.

antagonistic pleiotropy and epistasis … shape the resistance-conferring mutational path

Taken together, this study functionally validates the adaptive alleles underlying cardiac glycoside specialization in the monarch butterfly, providing “in vivo validation of a multi-step adaptive walk in a multicellular organism”.