The Genome of the Softshell Clam Mya arenaria and the Evolution of Apoptosis

Abstract

Apoptosis is a fundamental feature of multicellular animals and is best understood in mammals, flies, and nematodes, with the invertebrate models being thought to represent a condition of ancestral simplicity. However, the existence of a leukemia-like cancer in the softshell clam Mya arenaria provides an opportunity to re-evaluate the evolution of the genetic machinery of apoptosis. Here, we report the whole-genome sequence for M. arenaria which we leverage with existing data to test evolutionary hypotheses on the origins of apoptosis in animals. We show that the ancestral bilaterian p53 locus, a master regulator of apoptosis, possessed a complex domain structure, in contrast to that of extant ecdysozoan p53s. Further, ecdysozoan taxa, but not chordates or lophotrochozoans like M. arenaria, show a widespread reduction in apoptosis gene copy number. Finally, phylogenetic exploration of apoptosis gene copy number reveals a striking linkage with p53 domain complexity across species. Our results challenge the current understanding of the evolution of apoptosis and highlight the ancestral complexity of the bilaterian apoptotic tool kit and its subsequent dismantlement during the ecdysozoan radiation.

Keywords: Mya; Ecdysozoa; apoptosis; gene family evolution; softshell clam.

Fig. 1.

The Mya arenaria protein-coding genome complement is similar to other lophotrochozoans. Multidimensional scaling plot of showing ordination of genomic similarity between high-quality metazoan genomes including M. arenaria. All lophotrochozoan genomes are shown in red.

Fig. 2.

The evolutionary history and functional diversification of choanozoan p53s. (A) Phylogeny of choanozoan p53 sequences obtained from whole-genome sequences. Maximum likelihood tree was estimated under the best fit VT+F+R6 model (Quang le et al. 2008). Support is given by aLRT (Anisimova and Gascuel 2006) and ultrafast bootstrapping (Hoang et al. 2018). Vertebrates p53, p63, and p73 clades are indicated. Lophotrochozoan sequences are consistently more similar to mammalian sequences and reside on relatively short branches in contrast to ecdyzozoan sequences which reside on characteristically long branches. (B) Multiple sequence alignment showing the domain structure of p53 gene family sequences on the phylogeny. Approximate locations of the TAD, p53, TET, and SAM domains are shown. (C) Phylogeny of the whole-genome taxa utilized in this study. Clades of interest include the deuterostomes including mammals, lophotrochozoans including Mya arenaria, and ecdysozoans including Drosophila melanogaster and Caenorhabditis elegans. Dollo parsimony was used to map the evolutionary histories of the individual p53 gene family domains onto the phylogeny. This analysis portrays an increase in domain complexity reaching its peak at the bilaterian ancestor. Much of this domain complexity is retained in deuterostomes and lophotrochozoans but has been lost in nonpriapulid ecdysozoans. Colors are as shown in (B).

Fig. 3.

The evolutionary history of apoptotic genes as revealed by ortholog enrichment analysis. (A) Time calibrated phylogeny of the whole-genome taxa included in the study. Pie charts indicate additive gene totals from apoptosis orthogroups that are significantly depleted in ecdysozoans as revealed by Dollo parsimony analyses. (B) Heatmap showing gene counts across all 57 orthogroups derived from the analysis of KEGG apoptosis genes. (C) The subset of genes that is significantly depleted in ecdysozoans (red branches) as compared with deuterostomes and lophotrochozoans (blue branches). The cells in the heatmaps are ordered based on hierarchical clustering, shown at top.

Fig. 4.

The relationship between genome size and number of apoptosis genes across choanozoan phylogeny. We observed marginal phylogenetic signal for genome size when analyzed on the phylogeny (K = 0.65; λ=0.71). (A) This relationship is visible when the numbers of genes recovered in our orthology analysis are represented on the tree. (B) We therefore wished to examine potential for our observation of significant apoptosis gene depletion in ecdysozoan genomes to be an artifact of global gene depletion. The distribution of total, nonapoptotic genes (gray) and apoptotic genes (red) that were lost in either Lophotrochozoa (left) or Ecdysozoa (right) are shown. In addition to demonstrating the global trend in gene loss in ecdysozoans as compared with lophotrochozoans, the distributions of apoptotic gene loss in ecdysozoans significantly exceeds that of nonapoptosis genes. We conclude that our finding of apoptosis gene reduction is not an artifact of global genome reduction.

Fig. 5.

The functional implications of the evolutionary dynamics of the apoptosis gene set. Top. The KEGG apoptosis human pathway is shown redrawn from (Kanehisa et al. 2019). Genes that are significantly depleted in ecdysozoans are highlighted in yellow. This analysis shows that the genes depleted in ecdysozoans function in every aspect of apoptosis including external signaling, mitochondrial disruption, and DNA fragmentation.

Fig. 6.

Canonical correspondence and phylogenetic least squares analysis indicate correlations between apoptosis gene loss and p53 domain complexity on the tree. (A) Canonical correspondence analysis (CCA) reveals a significant global relationship between apoptosis gene richness and p53 domain richness across taxa (F = 2.09, P = 0.002). This can be seen by a subset of apoptosis genes (red) pulled in the direction of the p53 domain (TET, p53, TAD, and SAM) ordination vectors. Note the difference between the SAM ordination vector and those of the other domains. The distribution of taxa in Euclidian space as a function of both their apoptotic gene richness and p53 domain complexity is also shown (black). (B) We addressed correlations between specific genes and p53 domain richness using phylogenetic least squares analyses. Several domains show overlapping significance with specific genes with the TAD, p53, and TET domains showing the greatest similarity. Colors in (A) and (B) follow figure 2 (see also supplementary figure 2, Supplementary Material online).