Gene expression is more strongly influenced by age than caste in the ant Lasius niger

Abstract

Phenotypic plasticity, where a single genome can give rise to different phenotypes, underlies many remarkable features of the natural world and occurs in a wide range of organisms. Understanding the transcriptional differences that underlie plastic phenotypes remains a major unsolved problem in biology. In many ants, females can develop into either queens or workers, two phenotypes with different morphology, behaviour and longevity. In comparison with workers, queens are larger, more fecund and longer lived. Here, we study gene expression differences between queens and workers in the ant Lasius niger. The analysis of age- and tissue-specific RNA sequencing showed that patterns of caste-biased gene expression vary considerably between ages and tissues. Expression was more tightly linked to age than caste despite the important morphological and behavioural differences between queens and workers. Our data allowed us to identify genes that are consistently biased across biological contexts. Caste-biased genes showed faster rates of molecular evolution, lower levels of DNA methylation and greater variability in expression than unbiased genes. Our results indicate that a substantial proportion of caste-biased expression is ephemeral and that taking account of age and tissue is critical to understanding the transcriptomic basis of plastic phenotypes. By contrast, the biological context of expression bias did not broadly affect methylation or the rate of evolution. The faster rate of evolution and greater variability of expression of caste-biased genes indicate that caste-biased genes evolve from loosely regulated genes that can be co-opted for caste-specific tasks because of the lax control over their expression.

Gene expression is stronger associated with behaviour than with age and fertility in ant workers

“…Our findings strongly contrast with studies investigating caste (queen vs . worker) gene expression across different developmental stages and ages, in which more DEGs were found between different ages or developmental stages than between castes [68,91,106–108]. The larger gene expression differences between brood carers and foragers in combination with no upregulation of pathways associated with longevity indicate strong physiological differences between the two behavioural worker castes, and suggest weak changes with age in the investment in body repair mechanisms [109].…”

“…However, additional analyses, such as GO enrichment, WGCNA and metabolic pathway comparisons are commonly used and provide valuable insights into broader patterns of gene expression [27,30,31,67,68,75,87,90,96]. Hence, results and conclusions depend on the number and identity of large sets of genes, and partly their FDR values and raw read counts provided as an input.…”

“…Gene expression patterns associated with worker behaviour are typically identified by comparing gene expression of brood carers and foragers. As age and fertility additionally influence gene expression [22,28,67,68], studies that did not experimentally control for age and physiology when comparing gene expression between brood carers and foragers [e.g. 28,31,46,56,57,69–73] might have produced results driven not only by behaviour, but also by age and fertility.…”

“…This allowed us to properly assess how behaviour, age, and fertility are associated with gene expression in ant workers. A recent study revealed stronger differences in gene expression between ant queens and workers [68] than between different age-cohorts of the same caste. As we sampled from two clearly distinct age-cohorts differing in age by at least one year, which is estimated to be the mean T. longispinosus worker lifespan in the field [74], we expected age to have a stronger impact on gene expression than behaviour.…”

Convergent eusocial evolution is based on a shared reproductive groundplan plus lineage-specific plastic genes

“…Previous work has mainly focused on identifying whether there is significant overlap of genes or gene pathways associated with caste-based division of labor between independent lineages 23,26,27 , but there has been little effort to quantify the relative importance of shared versus unshared genes to the convergent evolution of caste-based division of labor. Most of these studies have either focused on brain or whole body samples 15,22,24,26,33,[35][36][37] , although expression bias between queens and workers has been shown to be dependent upon developmental stage and tissue type 34,[38][39][40] . Finally, the transcriptomic signatures of reproductive physiology are strongest in the abdomen 34,41 , the location of reproductive organs, but no past study has explicitly compared caste bias in abdominal tissues in species from lineages representing independent origins of eusociality.…”