Evolution of the Cytolytic Pore-Forming Proteins (Actinoporins) in Sea Anemones

Abstract

Sea anemones (Cnidaria, Anthozoa, and Actiniaria) use toxic peptides to incapacitate and immobilize prey and to deter potential predators. Their toxin arsenal is complex, targeting a variety of functionally important protein complexes and macromolecules involved in cellular homeostasis. Among these, actinoporins are one of the better characterized toxins; these venom proteins form a pore in cellular membranes containing sphingomyelin. We used a combined bioinformatic and phylogenetic approach to investigate how actinoporins have evolved across three superfamilies of sea anemones (Actinioidea, Metridioidea, and Actinostoloidea). Our analysis identified 90 candidate actinoporins across 20 species. We also found clusters of six actinoporin-like genes in five species of sea anemone (Nematostella vectensis, Stomphia coccinea, Epiactis japonica, Heteractis crispa, and Diadumene leucolena); these actinoporin-like sequences resembled actinoporins but have a higher sequence similarity with toxins from fungi, cone snails, and Hydra. Comparative analysis of the candidate actinoporins highlighted variable and conserved regions within actinoporins that may pertain to functional variation. Although multiple residues are involved in initiating sphingomyelin recognition and membrane binding, there is a high rate of replacement for a specific tryptophan with leucine (W112L) and other hydrophobic residues. Residues thought to be involved with oligomerization were variable, while those forming the phosphocholine (POC) binding site and the N-terminal region involved with cell membrane penetration were highly conserved.

Keywords: Cnidaria; cytolysins; target recognition; toxin.

Figure 1

(A) Maximum Likelihood tree of actinoporins and actinoporin-like proteins produced in FastTree2. Numbers on branches represent bootstrap values of 1000 replicates. Bootstrapping values greater than 50 are shown at the nodes. Branch labels of clustered gene groups represent lineage common names followed by percent identity (identical amino acid residues) within gene cluster and percent identity when compared to actinoporins in bold. Labeled individual branches show GenBank accession followed by species, and protein name (if applicable). Labels denoted with PREDICTED or GENOME indicate that they were derived bioinformatically in GenBank and are not validated proteins. Branch labels with Genbank accession and species for sea anemones are indicated in bold. The colored box indicates which actinoporin sequences were used in subsequent analyses; (B) Phylogenetic tree with branch colors depicting superfamily associations (Blue: Edwardsioidea, Yellow: Actinostoloidea, Red: Actinioidea, Green: Metridioidea) [42].

Figure 2

(A) Functionally important residues identified on EqII [43]. Functional sites are as follows: (B) site of bend when N-terminus comes into contact with the cell membrane, (POC) residues involved with the POC binding site, (O) residues involved with oligomerization, (S) key sphingomyelin binding site. Colors are used to aid in determining the orientation between the two views shown; (B) Characterization of amino acid variation for the different gene clusters (Clade 1, Clade 2M, Clade 2A) identified in our analysis. SeqLogo graphs for residues that have been identified previously as functionally important above the alignment with the size of each amino acid residue representing the frequency in which these residues occurred in the alignment. Numbers along the bottom correspond to positions of specific amino acid residues in EqII; (C) Maximum Likelihood actinoporin gene tree produced in FastTree2. Colored branches depict superfamily associations (see Figure 1, Yellow: Actinostoloidea, Red: Actinioidea, Green: Metridioidea). Bootstrapping values greater than 50 are shown at the nodes. Branch labels include GenBank ID (when applicable) and the species from which the toxin gene was derived. Bold labels indicate that the mature protein sequence was recovered. Sequences derived from genomic data are indicated with “G” following species in sequence IDs. The superfamily association for Actineria villosa may be incorrect and is noted with an asterisk.

Figure 3

Edmundson wheel projections were determined in HeliQuest. Associated Sequence IDs are shown below the species name.

Figure 3

Edmundson wheel projections were determined in HeliQuest. Associated Sequence IDs are shown below the species name.