Shared Features Underlying Compact Genomes and Extreme Habitat Use in Chironomid Midges

Abstract

Nonbiting midges (family Chironomidae) are found throughout the world in a diverse array of aquatic and terrestrial habitats, can often tolerate harsh conditions such as hypoxia or desiccation, and have consistently compact genomes. Yet we know little about the shared molecular basis for these attributes and how they have evolved across the family. Here, we address these questions by first creating high-quality, annotated reference assemblies for Tanytarsus gracilentus (subfamily Chironominae, tribe Tanytarsini) and Parochlus steinenii (subfamily Podonominae). Using these and other publicly available assemblies, we created a time-calibrated phylogenomic tree for family Chironomidae with outgroups from order Diptera. We used this phylogeny to test for features associated with compact genomes, as well as examining patterns of gene family evolution and positive selection that may underlie chironomid habitat tolerances. Our results suggest that compact genomes evolved in the common ancestor of Chironomidae and Ceratopogonidae and that this occurred mainly through reductions in noncoding regions (introns, intergenic sequences, and repeat elements). Significantly expanded gene families in Chironomidae included biological processes that may relate to tolerance of stressful environments, such as temperature homeostasis, carbohydrate transport, melanization defense response, and trehalose transport. We identified several positively selected genes in Chironomidae, notably sulfonylurea receptor, CREB-binding protein, and protein kinase D. Our results improve our understanding of the evolution of small genomes and extreme habitat use in this widely distributed group.

Keywords: extremophile; gene family evolution; genome compaction; positive selection; repeat elements.

Fig. 1.

A) Time-calibrated phylogenomic tree showing relationships among chironomids and outgroups within order Diptera. Wide, gray bars indicate 95% credible intervals for node ages, and lowercase letters indicate fossil calibrations. Geological periods are shown above the x-axis. Species label colors and positions are the same as in the lower panels; chironomids are the top-most 9 species in shades of red/orange. B) Time-calibrated phylogeny next to genome size, number of protein-coding genes, total intergenic content, and mean intron length for all species in our phylogeny. C) Phylogeny alongside repeat content by class. B, C) Gray vertical lines indicate the mean estimates from the phylogenetic linear regressions for Chironomidae and Ceratopogonidae and for all other species. Gray envelopes indicate the 95% confidence interval bounds for these estimates computed via parametric bootstrapping. All measures are the log₁₀-transformed totals across each species’ entire genome except for intron length, which is the mean of log₁₀-transformed intron lengths.

Fig. 2.

Genome size versus protein-coding genes, intergenic content, mean intron length, and repeat element content by class for dipterans in our phylogeny. Point colors and data transformations for all panels are as described in Fig. 1. Numbers in each panel indicate the estimate of the Pearson correlation coefficient between each variable and genome size using cor_phylo, and “*” indicates the correlation's 95% confidence interval did not overlap zero (supplementary table S3, Supplementary Material online).

Fig. 3.

A) Treemap showing hierarchical structure of GO terms for HOGs that expanded significantly in Chironomidae. B) For each targeted GO term related to extreme physiology, we show the percent of single-copy HOGs that are significant for gene-wide positive selection (BUSTED), intensification of selection (RELAX_intensify), relaxation of selection (RELAX_relax), or both BUSTED and RELAX (either RELAX_intensify or RELAX_relax). Numbers indicate total HOGs per GO term, and some GO terms share HOGs. C) List of HOGs with evidence for both positive selection and change in selection intensity in chironomids, in order of decreasing evidence for both tests (i.e. P_BUSTED × P_RELAX). The GO column indicates the GO term(s) listed in B) associated with each HOG. Parameter k indicates the relative selection intensity for Chironomidae compared to outgroups (k > 1 means intensified selection, k < 1 means relaxed). HOG descriptions were extracted from the representative gene in Culex quinquefasciatus.