Testing hypotheses of a coevolutionary key innovation reveals a complex suite of traits involved in defusing the mustard oil bomb.

Coevolutionary interactions are responsible for much of the Earth's biodiversity, with key innovations driving speciation bursts on both sides of the interaction. One persistent question is whether macroevolutionary traits identified as key innovations accurately predict functional performance and selection dynamics within species, as this necessitates characterizing their function, investigating their fitness consequences, and exploring the selection dynamics acting upon them. Here, we used CRISPR-Cas9 mediating nonhomologous end joining (NHEJ) in the butterfly species Pieris brassicae to knock out and directly assess the function and fitness impacts of nitrile specifier protein (NSP) and major allergen (MA). These are two closely related genes that facilitate glucosinolate (GSL) detoxification capacity, which is a key innovation in mustard feeding Pierinae butterflies. We find NSP and MA are both required for survival on plants containing GSLs, with expression differences arising in response to variable GSL profiles, concordant with detoxification performance. Importantly, this concordance was only observed when using natural host plants, likely reflecting the complexity of how these enzymes interact with natural plant variation in GSLs and myrosinases. Finally, signatures of positive selection for NSP and MA were detected across Pieris species, consistent with these genes' importance in recent coevolutionary interactions. Thus, the war between these butterflies and their host plants involves more than the mere presence of chemical defenses and detoxification mechanisms, as their regulation and activation represent key components of complex interactions. We find that inclusion of these dynamics, in ecologically relevant assays, is necessary for coevolutionary insights in this system and likely others.

Accessing the transcriptome: how to normalize mRNA pools.

As advances in next generation sequencing continue to provide increasing access to the genomics -revolution for research systems having few or no genomic resources, transcriptome sequencing will only increase in importance as a fast and direct means of accessing the genes themselves. However, constructing a comprehensive cDNA library for deep sequencing is very difficult, as highly abundant transcripts hamper de novo identification of low-expressed genes, and genes expressed only under very specific conditions will remain elusive. The reduction of variance in gene expression levels to within a tenfold range of differences by cDNA normalization provides an important means of allocating sequencing across a greater fraction of genes, directly translating into a more even coverage across genes. Here, we outline two different normalization methods, addressing many of the important issues we think need consideration when going from RNA isolation to the cDNA material required for sequencing. This will provide coding gene information across thousands of genes from any organism, providing rapid insights into topics such as gene family member identification and genetic variation that may be associated with a studied phenotype.