Rapid Evolution of Gained Essential Developmental Functions of a Young Gene via Interactions with Other Essential Genes

Abstract

New genes are of recent origin and only present in a subset of species in a phylogeny. Accumulated evidence suggests that new genes, like old genes that are conserved across species, can also take on important functions and be essential for the survival and reproductive success of organisms. Although there are detailed analyses of the mechanisms underlying new genes' gaining fertility functions, how new genes rapidly become essential for viability remains unclear. We focused on a young retro-duplicated gene (CG7804, which we named Cocoon) in Drosophila that originated between 4 and 10 Ma. We found that, unlike its evolutionarily conserved parental gene, Cocoon has evolved under positive selection and accumulated many amino acid differences at functional sites from the parental gene. Despite its young age, Cocoon is essential for the survival of Drosophila melanogaster at multiple developmental stages, including the critical embryonic stage, and its expression is essential in different tissues from those of its parental gene. Functional genomic analyses found that Cocoon acquired unique DNA-binding sites and has a contrasting effect on gene expression to that of its parental gene. Importantly, Cocoon binding predominantly locates at genes that have other essential functions and/or have multiple gene-gene interactions, suggesting that Cocoon acquired novel essential function to survival through forming interactions that have large impacts on the gene interaction network. Our study is an important step toward deciphering the evolutionary trajectory by which new genes functionally diverge from parental genes and become essential.

Keywords: Drosophila; development; gene network; lethality; retrogene evolution.

Fig. 1.

Structural and evolutionary history of CG7804 and TBPH. (A) Exon–intron structure of TBPH (blue) and CG7804 (orange). Filled boxes represent exons (darker color, coding sequence; lighter color, UTRs), whereas lines represent introns. Because CG7804 originated through a retrotransposition event, it lacks most of the introns of TBPH and some of its noncoding sequences do not share homology with TBPH. (B) The duplication event of CG7804 from TBPH is denoted as a dashed line in the phylogeny. The clades of CG7804 and TBPH are in orange and blue, respectively. The number of amino acid substitutions to number of synonymous substitutions inferred by PAML (see text) is denoted at right to branches. The dating of the species phylogeny is from Obbard et al. [2012]. (C) The structures of the amino acid sequences of TBPH and CG7804 are shown at the upper panel. In the second panel, the divergent amino acids of CG7804 from TBPH are denoted as vertical lines. Different colors indicate different changes in amino acid chemical properties (N, noncharged; A, acidic; B, basic). The third and fourth panels show the results of sliding window MK test of CG7804 (window size 99 bp and step 9 bp), including −log 10 P value of the MK test and the estimated proportion of amino acid fixations between Drosophila melanogaster and D. simulans that were driven by positive selection (α). Note that the coordinates of four panels are aligned. (D) Predicted structures of the RRMs of TBPH and CG7804 are shown. The 3D structures of the first (RRM1) and the second (RRM2) RNA-recognition motifs were predicted using Phyre (Kelley et al. 2015). The rainbow color is from N (red) to C (blue) termini.

Fig. 2.

Stage-specific lethality associated with CG7804 knockdown and knockout. (A) Expression knockdown of CG7804 using Tub-GAL4 driver results in different lethality rates at different developmental stages. (B) Expression knockdown of CG7804 leads to eclosion lethal. (C) The outcome of pupae with CG7804 expression knockdown. (D) CG7804 expression is essential in different tissues from those of its parental gene, TBPH. Lethality rate is significantly different between CG7804 and TBPH knockdown when using elav, Dll, and en GAL4 drivers. Because the lethality rate is estimated relative to wildtype genotype (see Materials and Methods), negative lethality rate (e.g., en driver knocking down CG7804) means higher survival rate of that particular genotype than the wildtype. (E) Expression knockdown of CG7804 using Dll-GAL4 driver results in completely fused leg joint (red arrow) or semifused leg joint (black arrow). Legs of wildtype (Dll-GAL4 driver strain) individuals are shown side by side with those of knockdown individuals. T1–T3 are first, second, and third legs, respectively. (F) CG7804 knockout homozygotes have significantly higher lethality rate from embryo to larva and from larva to pupa than wildtype individuals. CG7804 knockout heterozygotes have similar lethality rates to those of wildtype individuals. E, embryo; L, third instar larvae (L3); P, pupae. KD, individuals with CG7804 knockdown genotype; nonKD, wildtype individuals; OP, pupae with pupa cased removed (open pupae). Mann–Whitney U test: *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 3.

Differential expression upon CG7804 knockout. Volcano plots for the log 2 fold change in expression level (x axis) and −log 10 FDR (y axis) for the embryonic stage (A), the larval stage (B), and the pupal stage (C). Red dots represent genes that are differentially expressed with FDR < 0.01 and blue dots for FDR < 0.05. Gray dots represent genes that are not differentially expressed. Numbers under each panel are the number of genes analyzed/number of upregulated genes/number of downregulated genes (D) heatmap for the log 2 fold change in expression level at three developmental stages (E, embryo; L, larva; P, pupa). Each horizontal row represents one gene. Oranges are for positive log 2 fold change (i.e., upregulated with CG7804 knockout), whereas greens are the opposite.

Fig. 4.

Genes with CG7804-binding enrichment are different from other genes in the genome. (A) CG7804 has nuclear localization in the salivary gland of third instar larva. Green, CG7804; blue, DNA; red, cytoskeleton. (B) Examples of a binding region that is unique to CG7804 (1) or is shared between CG7804 and TBPH (2). This example shows that even for some genes bound by both paralogs, CG7804 may still have gained unique binding sites (here, upstream of an essential gene, MyC). (C) Bar plots for the proportion of genes that are differentially expressed (either downregulated or upregulated) for genes with or without CG7804-binding enrichment. Different shade of green colors is whether a gene is differentially expressed upon CG7804 knockout. (D–G) Comparing (D) known mutant phenotype, (E) degree (number of protein–protein interaction/genetic interaction a gene is involved in), (F) α (the proportion of adaptive amino acid substitution), and (G) dN/dS ratio (Drosophila melanogaster linage-specific substitution rates) between genes with/without CG7804-binding enrichment and differential expression upon CG7804 knockout. Three comparisons were performed: 1) between genes with and without CG7804-binding enrichment, 2) among genes with CG7804-binding enrichment, genes with and without differential expression upon CG7804 knockout, and 3) among genes with CG7804-binding enrichment and differential expression upon CG7804 knockout, genes with increased and decreased expression. Mann–Whitney U test: *P < 0.05, **P < 0.01, and ***P < 0.001.