Highly Dynamic Gene Family Evolution Suggests Changing Roles for PON Genes Within Metazoa

Abstract

Change in gene family size has been shown to facilitate adaptation to different selective pressures. This includes gene duplication to increase dosage or diversification of enzymatic substrates and gene deletion due to relaxed selection. We recently found that the PON1 gene, an enzyme with arylesterase and lactonase activity, was lost repeatedly in different aquatic mammalian lineages, suggesting that the PON gene family is responsive to environmental change. We further investigated if these fluctuations in gene family size were restricted to mammals and approximately when this gene family was expanded within mammals. Using 112 metazoan protein models, we explored the evolutionary history of the PON family to characterize the dynamic evolution of this gene family. We found that there have been multiple, independent expansion events in tardigrades, cephalochordates, and echinoderms. In addition, there have been partial gene loss events in monotremes and sea cucumbers and what appears to be complete loss in arthropods, urochordates, platyhelminths, ctenophores, and placozoans. In addition, we show the mammalian expansion to three PON paralogs occurred in the ancestor of all mammals after the divergence of sauropsida but before the divergence of monotremes from therians. We also provide evidence of a novel PON expansion within the brushtail possum. In the face of repeated expansions and deletions in the context of changing environments, we suggest a range of selective pressures, including pathogen infection and mitigation of oxidative damage, are likely influencing the diversification of this dynamic gene family across metazoa.

Keywords: brushtail possum; gene family evolution; gene family expansion; gene loss; paraoxonase (PON); phylogenetics.

Fig. 1.

Evolution of PON in metazoans. Phylogenetic tree of PON family proteins in metazoans determined by RAxML based on multiple sequence alignment. Bootstrap support values are shown as percentages out of 1,000 bootstraps. If species have multiple PON genes and they are located on different chromosome/scaffold or are sufficiently far from one another, then they are indicated by different alphabet characters. If they PON genes are located on the same chromosome/scaffold and are within one 100 kb of another PON gene, this is indicated by a number. Taxa mentioned in the results section are labeled. Specific nodes highlighting instances ancient, tandem, or retroduplication are indicated by a circle, triangle, or cross, respectively, either underneath or on the branch.

Fig. 2.

Evolution of PON in tetrapods. (A) Phylogenetic tree of PON family proteins in tetrapods determined by PhyML based on multiple sequence alignment. Bootstrap support values are shown as percentages out of 200. Brushtail PON4 was split into two separate genes based on RNA-seq evidence. Human sequences were underlined to help orient the reader. (B) Phylogenetic tree models represent the scenario in which each of the PONs is ancestral compared with the other two. *P <0.05. The underlined log likelihood indicates the significantly better model. (C, D) Two sets of models compare the placement of the monotremes either basally along with birds and lizards (i.e., sauropsida) or within PON3 with either PON1 being ancestral (C) or PON3 being ancestral (D). ** P < 1e-8. The underlined log likelihood indicates the significantly better model. Rooted and unrooted trees produced the same log likelihood score.

Fig. 3.

Brushtail PON3 and recent PON4A/B under positive selection. (A) Screenshot from IGV showing a mapped reads to a concatenated transcriptome for all brushtail possum PON genes. Tissue samples include four brain, five liver, and one heart sample. The order of the concatenated PON genes is listed above the screenshot. Order was chosen to mimic the chromosomal order found in the brushtail genome. (B) Phylogenetic tree of marsupial PON3 genes used for PAML and BUSTED analysis shown. The bolded branches indicate the foreground sequences tested for positive selection. (C) Log likelihood values were determined by PAML. The marsupial tree in figure 3B was used in the marsupial analysis (D) 3D protein image of rabbit PON1 (PDB:1V04). Corresponding sites under positive selection in brushtail are highlighted and show their atomic structure. Residues which are part of the active site also show their atomic structure.

Fig. 4.

Overview of observed changes in the number of PON genes throughout metazoa. Phylogenetic tree represent the approximate time when changes in the number of PON genes occurred in metazoan broadly (Laumer et al. 2019; Philippe et al. 2019; Kapli and Telford 2020). More detailed phylogenetic tree is available as supplementary fig. S3, Supplementary Material online. The relative timing of the changes is indicated by the number and sign above the branch. Number in between parenthesis after the phylum name indicates the number of ancestral genes which could be detected for the terminal branch. The lack of changes noted in the deeper portions of this tree are not meant to indicate that there was no change in gene copy number in those branches. Rather, it is merely a reflection of the limitation of this study to probe those branches. Ctenophora and Porifera are shown as a polytomy as their exact relationship has not yet been elucidated (Laumer et al. 2019; Li et al. 2021; Redmond and McLysaght 2021). Placement of dicyemida is not well resolved at the time of this publication (Zverkov et al. 2019).