Multiomics analysis of the giant triton snail salivary gland, a crown-of-thorns starfish predator

The giant triton snail (Charonia tritonis) is one of the few natural predators of the adult Crown-of-Thorns starfish (COTS), a corallivore that has been damaging to many reefs in the Indo-Pacific. Charonia species have large salivary glands (SGs) that are suspected to produce either a venom and/or sulphuric acid which can immobilize their prey and neutralize the intrinsic toxic properties of COTS. To date, there is little information on the types of toxins produced by tritons. In this paper, the predatory behaviour of the C. tritonis is described. Then, the C. tritonis SG, which itself is made up of an anterior lobe (AL) and posterior lobe (PL), was analyzed using an integrated transcriptomics and proteomics approach, to identify putative toxin- and feeding-related proteins. A de novo transcriptome database and in silico protein analysis predicts that ~3800 proteins have features consistent with being secreted. A gland-specific proteomics analysis confirmed the presence of numerous SG-AL and SG-PL proteins, including those with similarity to cysteine-rich venom proteins. Sulfuric acid biosynthesis enzymes were identified, specific to the SG-PL. Our analysis of the C. tritonis SG (AL and PL) has provided a deeper insight into the biomolecular toolkit used for predation and feeding by C. tritonis.

through which they bore a wide hole using their radula apparatus and long proboscis 22 . In Cassidaria echinophora and Tonna galea sulfuric acid is secreted through their proboscis 24,25 . The saliva of Cassis tuberosa can immobilize the spines of sea urchin 18 . Within the Tonnidae group of marine snails, which also predate crustaceans and bivalve molluscs, large SGs have been identified that produce complex salivary secretions 26 .
The giant triton snail Charonia tritonis (superfamily Tonnoidea) is found on reefs throughout the Indo-Pacific where it predates upon echinoderms including starfish, sea cucumbers and sea urchins [27][28][29] . Giant tritons rely on their highly developed olfactory sense to track and locate prey 30,31 . Upon contact with prey, they initiate a 'tapping' behaviour using their cephalic tentacles. Although the prey attempts to escape, the giant triton immobilizes it initially by mechanical means, positioning its large muscular foot over the aboral surface. This is rapidly followed by insertion of the proboscis and most likely injection of venom(s) that paralyses the prey 27,30 . At least several species within the Ranellidae (Tonnidea) are known to produce sulphuric acid to access their prey 32 . The marine gastropoda, Gyrineum natator uses sulfuric acid to capture their bivalve prey and also use this acid for their defense 32 . The turban shell, Lunella coronate possesses sulphuric acid-producing glands which are used for their defense, to externally digest its accessed prey or to attack less accessible prey 33 . It has been proposed that the Atlantic triton snail (Charonia variegate) possesses toxins derived from its foot or mouth which may assist to immobilize prey 34 . In the knobbed triton snail (Charonia lampas rubicunda), the proboscis delivers a SG secretion that induces instant paralysis of the sea star Patriella brevispina 35 . Recently a peptide venom was identified within the SG of the hairy triton snail, Monoplex parthenopeus (Subfamily Cymatiinae) 36,37 . Conversely, another study on C. lampas feeding behavior found no evidence of the injection of either venom or acid 30 , although it should be noted that to date, SGs of C. lampas have not been analysed for the presence of such biomolecules 30 .
Based on feeding trials with C. tritonis, their preferred prey includes the crown-of-thorns starfish (COTS), Acanthaster planci, a corallivore asteroid that has contributed to mass coral loss throughout Indo-Pacific coral reefs 35 . However, C. tritonis are either naturally rare or endangered due to unregulated harvesting, with many countries now prohibiting their collection. Whatever the case, it has been proposed that the giant triton, as a major predator of COTS, has a role sin regulating the populations of COTS. For this reason, it is desirable to assess basic processes of their biology to assist in the potential development of captive breeding programs for release of giant tritons onto reefs infested with COTS. Consumption of COTS would require a special digestive strategy especially as COTS not only present a formidable physical defense in the form of sharp aboral spines but also contain and secrete various sulfated steroidal saponins and toxins 38 . Saponins are toxic and can deter their predators. Therefore, it is likely that giant triton snails have evolved specific traits to combat their effects 35 . Although an early study by Endean (1972) indicated that Charonia sp. uses venom to paralyze their prey, questions remain as to the chemical nature of these, and what role they play in the capture, immobilization and digestion of COTS 35 .
In this study, we describe the behaviour associated with C. tritonis COTS predation. Following an anatomical analysis of the C. tritonis SG, we have performed next-generation transcriptome sequencing and annotation of ensuing transcripts in association with proteomic analyses. We report for the first time the existence of numerous secreted proteins, including a diverse array of putative toxin-and feeding-like protein families in C. tritonis.

Materials and Methods
Triton behaviour in response to COTS. C. tritonis (N = 8) were collected from the Great Barrier Reef under special permit (G13/36390.1) and held in a 4,000 L indoor 4 m 2 diameter holding tank at ambient temperature (26-28 °C) and salinity (32-35 ppt) with simulated natural photoperiod at the Australian Institute of Marine Science (AIMS). Water current in the tank (clockwise) was induced through airlifts via three 5 cm diameter PVC pipes with water intake at the base of the tank and expulsion through a 90 degree elbow at the surface. The giant tritons were periodically presented with live COTS, between 1 to 2 COTS per giant triton per week. General observations were made on the reaction of the giant tritons as well as the COTS and video recorded on GoPro over 8 h.
Anatomy and tissue collection. For tissue collection, wild C. tritonis were collected from Kavieng, Papua New Guinea and temporarily held and fed on echinoderms at the Marine Research Station, Kavieng. Animals were anaesthetized with isotonic MgCl 2 and the anterior portion removed from the shell. During the dissection the proboscis and SG were photographed using an iPhone 6 (8 MP, phase detection autofocus, dual-LED, Apple Inc. USA). For analysis of SG cell composition, the gland was spread onto a slide, then viewed and photographed using a Leica microscope equipped with a CCD camera. SGs were dissected and separated into the anterior lobe (AL) and posterior lobe (PL). Tissues collected for RNA isolation were stored in RNAlater (Ambion, California). For protein isolation, tissues were processed immediately as described below.
RNA isolation, sequencing and transcriptome assembly. RNA was extracted from tissue using TRIzol Reagent (Invitrogen Corp., Carlsbad, CA, USA), as per the manufacturers protocol. Following extraction, RNA was assessed for quality by visualisation on a 1.2% agarose gel, and quantified using a Nanodrop spectrophotometer (Thermo scientific). Total RNA samples were sent to Australian Genome Research Facility (Brisbane, Australia) for library construction and sequenced (paired-end) using an Illumina HiSeq 2500 sequencing platform. Raw sequence reads (100 bp) were assembled into contigs (>200 bp) using the CLC genomics software (Qiagen). Protein coding regions were determined using the open reading frame (ORF) predictor 39 . Relative expression of genes in each tissue transcriptome was determined based on RPKM (Reads Per Kilobase of exon per Million mapped reads) values, utilizing the commercially available CLC Genomic Workbench 7 software 40 .
Gene annotation, protein models and prediction of secreted proteins (exoproteome). A BLASTp search was used to annotate proteins from each C. tritonis transcriptome. Schematic diagrams of protein domain structures were prepared using IBS illustrator (IBS, version 1.0) software 41 . Multiple sequence alignments were performed using the MEGA 6.0 platform with the ClustalW protocol and the Gonnet protein weight matrix 42 . SWISS-MODEL 43 was used to predict the 3D protein structure of an echotoxin-like protein identified from giant triton SG 44 . First, BLASTp analysis was used to identify a template that shared significant sequence similarity to a C. tritonis echotoxin sequence. The best match was selected based on the presence of similar domains and plausibility quality control. Finally, based on the alignment, the coordinates of the model were constructed for the structurally conserved regions of the model. N-terminal signal sequences were predicted using the SignalP 4.1 45 , Predisi 46 and TMHMM 47 . A protein was designated as secreted only when it met the criteria of both SignalP and Predisi, and did not have a transmembrane domain predicted by TMHMM. Simple Modular Architecture Research Tool (SMART) was used to identify conserved domains in SG proteins 48 . Glycosylation sites were identified by using NetNGlyc 1.0 server 49 . Protein isolation from salivary gland lobes and nanoHPLC-ESI-Triple TOF. Frozen samples of SG lobes were homogenized in protein extraction buffer (8 M urea, 4 M thiourea, 0.8 M NH 4 HCO 3 , pH 8.0) in a 1:5 w:v ratio. Crude extracts were then centrifuged for 20 min (12,000 xg, 4 °C), then supernatant was collected, fractionated by 1D SDS-PAGE and stained using Coomassie Blue (GE Healthcare, city). Gel bands were excised and digested with trypsin following the protocol described previously 50 . Before LC-MS analysis, Zip-tip C18 (Merck Millipore, USA) was used to desalt and concentrate peptides and small proteins.
Tryptic peptides were further analysed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) on a Shimadzu Prominance Nano HPLC (Japan) coupled to a Triple-TOF 5600 mass spectrometer (ABSCIEX, Canada) equipped with a nano electrospray ion source. Aliquots (6 µL) of each extract were injected onto a 50 mm × 300 µm C18 trap column (Agilent Technologies, Australia) at 30 µL/min. The samples were desalted on the trap column for 5 min using solvent A [0.1% formic acid (aq)] at 30 µL/min. The trap column was then placed in-line with a 150 mm × 75 µm 300SBC18 3.5 µm analytical nano HPLC column (Agilent Technologies) for mass spectrometry analysis. Peptides were eluted with a linear gradient of 1-40% solvent B [90:10 acetonitrile: 0.1% formic acid (aq)] over 35 min at 300 nL/min flow rate, followed by a steeper gradient from 40% to 80% solvent B over 5 min. Solvent B was then held at 80% for 5 min to wash the column and then returned to 1% solvent B for equilibration before the next sample injection. The ionspray voltage was set to 2400 V, declustering potential (DP) 100 V, curtain gas flow 25, nebuliser gas 1 (GS1) 12 and interface heater at 150 °C. The mass spectrometer acquired 500 ms full scan TOF-MS data followed full scan (20 × 50 ms) product ion data in an Information Dependent Acquisition (IDA) mode. Full scan TOF-MS data was acquired over the mass range 350-1800 m/z and for product ion MS/MS 100-1800 m/z. Ions observed in the TOF-MS scan exceeding a threshold of 100 counts and a charge state of +2 to +5 were set to trigger the acquisition of product ion, MS/MS spectra of the resultant 20 most intense ions. The data was acquired and processed using Analyst TF 1.5.1 software (ABSCIEX, Canada).
Proteins were identified by database searching using PEAKS v7.0 (BSI, Canada) against the protein database built from the C. tritonis SG lobe transcriptomes. Search parameters were as follows: precursor ion mass tolerance, 0.1 Da; fragment ion mass tolerance, 0.1 Da; fully tryptic enzyme specificity with two possible missed cleavage sites; monoisotopic precursor mass; a fixed modification of cysteine carbamidomethylation; and variable modifications which included methionine oxidation, conversion of glutamine and glutamic acid to pyroglutamic acid, acetylation of lysine and deamidation of asparagine; false discovery rate (FDR) was set to ≤ 1%, and (−10*lgP) was calculated accordingly where P is the probability that an observed match is a random event.

Results and Discussion
C. tritonis hunting behavior. Aquatic invertebrates primarily rely on their olfactory sense to detect and locate potential prey 51 . The otherwise mainly sedentary COTS exhibited extreme agitation and movement when placed into the tank holding giant tritons. Similarly, sedentary giant tritons become active within a few minutes of exposure to the scent of COTS exhibiting a sweeping motion of both tentacles and forward movement. In a 4 m diameter tank with a clockwise water flow tritons would track the odour of COTS by moving counter-clockwise into the current even if the prey item was within 0.5 m of it down current. The COTS themselves are known to detect the odour of giant tritons and, rather than being normally sedentary, would exhibit pronounced movement 31 . A previous study had observed that C. tritonis hunting behavior (upon various echinoderm species, including COTS) was most prominent in the evening (57.1%), compared to morning (35.7%) and afternoon (7.1%) 52 . We observed that the entire hunt, attack and consumption of COTS could be completed within 4 h.
Tritons were documented hunting, attacking and feeding on COTS (Supplementary Video S1). Upon initial contact with COTS, the snails' large muscular foot covers the COTS arms to suppress movement. Simultaneously, C. tritonis retract then elongated their proboscis and moves it gradually around the base of the spines until insertion into the central disc area (Fig. 1A-C). It is possible that the proboscis seeks to cut the circumoral neural ring, as the COTS, while highly alarmed at this stage, become uncoordinated with no directional movement. At this point, the C. tritonis begin to saw into the COTS flesh using their radula. Closer inspection of the internal anatomy of the C. tritonis proboscis reveals the muscle, ducts, and buccal mass housing the radula at the proboscis tip (Fig. 1D). The radulae have been described in significant detail for other triton species, through scanning electron microscopy, showing the presence of cuspless marginal teeth and variations within the shape of central teeth 36 . Giant triton snails produce a prolific amount of mucus during the attack and consumption of the COTS, which may be associated with sequestration and possibly detoxification of saponins released by COTS or to absorb saponins before they reach the interior of the shell and make contact with the gills. Saponins readily cross the gills of fish and lyse red blood cells causing respiratory distress and in high enough concentration can cause death 53 . The terrestrial slug, Arion lusitanicus, has been shown to sequester and detoxify alkaloids from a variety of plants 54 . Given that giant tritons are not only exposed to secreted saponins during the attack but also ingest them in high concentrations, it is likely that they can metabolise saponins. Metabolic pathways for saponin detoxification mechanisms have been described for plant-fungi interactions, where the fungi contains genes that encode enzymes that break down plant saponins, leading to disease resistance 55, 56 . C. tritonis salivary gland anatomy and transcriptome assembly. In those Tonnoidea investigated to date, SGs are paired, with each gland divided into two parts: the smaller tubular or acinous anterior lobes (AL) and the voluminous posterior lobes (PL) that putatively secrete sulphuric acid 19,24,25,36 . Their size, shape and structure may vary significantly between genera and those species within 36 . The C. tritonis paired SGs were identified within the region of the foregut, similarly split morphologically into anterior and posterior lobes; the AL is larger and has an orange appearance, while the PL is smaller and white ( Fig. 2A-D). This is in contrast to what is observed in other Tonnoidea, where the PL is the larger lobe. The histology of the SG has been described for two Tonnoidea, the Argobuccinum pusulosum and Monoplex intermedius, showing a posterior salivary duct entering the inside of the anterior lobe 21,57 . In this study, no histology was performed, however, a cell smear of the AL revealed a mixture of cells, with large and clear cells supported by mucin-like molecules being the most prominent (Fig. 2E). In C. intermedius, the fine structure analysis of the salivary glands has revealed posterior acid secreting and acinous anterior lobes 21 . Transverse sections through the posterior SG of the C. lampas show many cells that appear to contain a basophilic mucus 30 .
Raw sequence RNA-seq reads were obtained for both AL and PL of the C. tritonis SG using Illumina technologies (NCBI PRJNA383875), then assembled into transcriptome library contigs. In the SG-AL, there were 84,807,390 total filtered reads assembled into 105,955 contigs, 105,297 ORFs (Supplementary Data S1) and 18 In silico exoproteome prediction and proteomic analysis. An in silico analysis was performed to identify putative SG secreted proteins, which predicted that the SG-AL and SG-PL secrete 3,805 and 3,860 proteins, respectively; all contain signal peptides and no transmembrane domains (Supplementary Data S2b and S4b). Due to technical limitations, proteins were extracted from crude intact SG-AL and SG-PL, rather than performing gland 'milking' which is now recognized as a more efficient method for obtaining cone snail venom 58 . Crude protein extracts were initially separated by SDS-PAGE and viewed by Coomassie staining, demonstrating the presence of a large number of proteins with a high-low molecular weight distribution (Fig. 3A). Bands that were extracted from frac-    family database): 122 and 502 Pfam families were identified from SG-AL and SG-PL, respectively, with 77 families common to both (Fig. 3B).
Supplementary Data S7 provides a summary of the MS analysis of the SG-AL and SG-PL extracts. Proteins that were highly represented in MS analysis for either or both the SG-AL and SG-PL, are shown in Table 1. In the C. tritonis SG-PL, the most highly represented protein was the lectin L6-like protein. Glycan-binding proteins, commonly known as lectins, play a crucial role in innate (and adaptive) immunity. Binding of potential pathogens by lectins leads to phagocytosis, complement activation, and antigen processing but also to regulation of adaptive immune functions 59,60 . In addition to their role in pathogen recognition, some lectins act as direct defense effectors by intoxicating the antagonist upon binding 61 . In the SG-AL, the most highly represented protein was aminopeptidase N, suggesting high abundance in this lobe. Aminopeptidase N is widely distributed in both plant and animal species where it is known to selectively break down amino acids from the amino-terminus of proteins or oligopeptides 62 . In animals, this enzyme is most abundant along the brush border membrane of the intestines, facilitating the digestion of ingested protein 63 . However, in the midgut of insects, the enzyme can act as a receptor binding to toxin proteins produced by viruses and bacteria [64][65][66] .
The hemorrhagic metalloprotease disintegrin-like proteins and cysteine-rich secretory Mr30 were also highly represented in our MS analysis (Table 1). Hemorrhagic metalloprotease disintegrin-like proteins have been isolated from snake venom where have been shown to inhibit the process of collagen-and ADP-induced platelet aggregation 67 , break-down of coagulation factors and the initiation of apoptosis 68 . The cysteine-rich secretory Mr30 has been identified from Conus snails and reported to have similar properties to Tex31, a protease responsible for processing of pro-conotoxins 69,70 . The C. tritonis SG does not contain conotoxins, yet the cysteine-rich secretory Mr30 may help to pre-process other types of pro-toxin proteins, including the venom serine carboxypeptidase, a member of the peptidase S10 family. This enzyme is known to have various physiological functions, including enhancing massive release of histamine, degradation of neurotransmitters, mediation of immunity-related processes, phosphorylation of venom proteins, and influencing the nervous system as a neurotoxic polypeptide 71 . The biological role of this peptidase in C. tritonis venom is unclear, but given that members of the peptidase S10 family are active at acidic pH 72 , they may perform their function in the venom before it is injected into the prey.
Arylsulfatase hydrolyses the sulfate group of ingested biomolecules, such as glycosaminoglycans (GAGs), which are large sugar molecules 73,74 . In molluscs, arylsulfatase activity has been reported in various tissues of ivory barnacles (Balanus eburneus), including the mantle, suggesting a probable relationship between arylsulfatase activity and the cyclic formation and hardening of the exoskeleton 75 . In snails of the family Muricidae, arylsulfatase is found in the hypobranchial gland where it plays a role in Tyrian purple biosynthesis 76 ; it helps in the conversion of tyrindoxyl sulphate into the biologically active precursors of Tyrian purple 77,78 .
A single arylsulfatase was identified in the C. tritonis SG-AL through proteomic analysis, although another 29 transcripts were present in the transcriptomes that encode for arylsulfatases. No arylsulfatase proteins were identified in the SG-PL; however, 36 transcripts were identified that encode for arylsulfatases. This finding provides further evidence that arylsulfatase genes are encoded within the animal's genome, rather than being obtained from dietary sources or through symbiotic microorganisms, as has been suggested from other molluscan studies 79,80 . It is speculated that SG arylsulfatase is involved in the breakdown of prey biomolecules during feeding. Echinoderms are well known to contain a diverse array of saponin molecules (to deter predators, parasites and microbes) including sulfated saponins; this subclass may be neutralized or broken down by arylsulfatase-mediated hydrolysis depending on the position of sulfation on either the aglycone or sugar moiety 81 . C. tritonis arylsulfatases contain all conserved arylsulfatase domains, and besides a variable signal sequence, the mature enzyme is highly conserved with arylsulfatases from five other species (Aplysia californica, Octopus bimaculosides, Helix pomatia, Biomphalaria glabrata and Strongylocentrotus purpuratus (Fig. 4).
We report the identification of putative toxin-related proteins from the transcriptome, with some being supported by the proteome MS identification, as shown in Table 2; their gene expression level (RPKM) is provided in Supplementary Data S8.
The cysteine-rich venom proteins (CRVPs) belong to the larger family of proteins known as the cysteine-rich secretory proteins (CRISPs), and are found in the venoms of diverse species, including snakes, cone snails, coleoids, stinging insects, scorpions and spiders [82][83][84][85][86][87][88] . Proteins from this family are also commonly found in the mammalian male reproductive tract 89 and are associated with a broad range of functions, such as fertilization and sperm-egg interaction 86,88 . In the current study, 8 transcripts that encode venom-like CRISPs were identified within the SG-AL, and 2 were confirmed through proteomic analysis ( Table 2). In addition, 3 transcripts that encode venom-like CRISPs were identified from the SG-PL, and an additional 2 confirmed through proteomics analysis ( Table 2). Domain analysis predicts regions characteristic of CRISPs, including the first class of pathogenesis-related proteins (Pr-1) and a cysteine-rich domain (CRD) (Fig. 4) 90 . Multiple sequence alignment with CRISPs from 2 other species, Conus textile and Conus marmoreus, showed that there is a high level of diversity in C. tritonis venom-like CRISPs, besides spatial arrangement of cysteine (C) residues (Fig. 4). CRVPs are well known to exert multiple activities through the blocking of L-type Ca 2+ channels and K + channel inhibitors 87 , which can reduce smooth muscle contraction and cause myonecrosis. Recent studies have revealed that CRISPs in snake venoms inhibit smooth muscle contraction and cyclic nucleotide-gated ion channels 83 . Considering that the CRISPs found in the venom of snakes could function as ion channel blockers 91,92 , the triton CRISPs might have a similar function by targeting the ion channels of prey.
Metalloproteinases are a family of proteolytic enzymes that are involved in a large number of biological processes. A variety of metalloproteinases are found in the venoms of spiders, scorpions, centipedes, cone snails and the platypus [93][94][95][96][97][98] . In snakes, these enzymes cause hemorrhaging upon envenomation 99 . The biochemical basis for metalloproteinase activity is through proteolytic destruction of tissue basement membranes and of the extracellular matrix surrounding capillaries and small vessels. They may also interfere with coagulation, thus   complementing loss of blood from the vasculature. The variety of hemorrhagic toxins found in snake venoms is due to the presence of structurally related proteins composed of various domains 99 . The type of domains found in each toxin plays a major role in the hemorrhagic potency of the protein.
Three transcripts were identified that encode venom-related metalloproteinases in the C. tritonis SG-AL with a single zinc metalloproteinase confirmed by proteome analysis. In the SG-PL, seven transcripts were found which  Table 2. Identification of putative toxin-related proteins from the Charonia tritonis salivary gland (SG) anterior lobe (AL) and posterior lobe (PL).
encode for metalloproteinease proteins and two proteins in the in silico exoproteome analysis. The zinc-dependent metalloproteinase contains a conserved metalloproteinease domain (Fig. 4). Multiple sequence alignment shows the conservation of cysteines and glycosaminoglycan attachment sites (serine motifs) in C. tritonis-derived metalloproteineases with four other species i.e. Lottia gigantea, Crassostrea gigas, Biomphalaria glabrata and Octopus bimaculoides. In the parasitic wasp Chelonus inanitus, conservation of high number of serine motifs is thought to be involved in substrate or site specific binding of venom protein 100 . This group of metalloproteineases belong to the MEROPS peptidase family M12, subfamily M12B [adamalysin family, clan (MA(M)]. The adamalysins are zinc-dependent endopeptidases also found in snake venom. The ' A disintegrin and metalloprotease' (ADAM) family of metalloproteases (also referred to as adamalysin-like metalloproteases) contains proteolytic domains from snake venoms, proteases from the mammalian reproductive tract, and the tumor necrosis factor alpha convertase, TACE. ADAMs are glycoproteins, which are involved in cell signaling, cell fusion, and cell-cell interactions. This supports a role for C. tritonis salivary gland metalloproteinase in defense, although further study is required to define its function. Echotoxins (1, 2 and 3) are lethal and hemolytic proteinaceous toxins of approximately 25 kDa, which were identified from the SG of the Tonnoidea, Monoplex echo 44,101,102 . Other echotoxins were subsequently identified in non-Tonnoidea snails through transcriptomics analysis of Conus consors and Conus geographus (superfamily Conoidea) 103,104 , as well as the C. reticulate (superfamily Buccinoidea) 6 . In this study, two deduced echotoxin-like proteins were identified from C. tritonis, present in both the SG-AL and SG-PL, at 275 residues (full-length) and 142 residues (partial-length). Multiple sequence alignment of known species echotoxins shows high conservation throughout the entire precursor besides the N-and C-terminal regions (Fig. 5A). No N-terminal signal peptide was predicted for either C. tritonis echotoxin-like proteins (based on SignalP analysis), yet alignment with M. echo echotoxins suggests that a signal peptide would be cleaved following V 30 EP. Of the 6 aromatic residues that are thought to form an aromatic patch on the surface of these types of protein 105 , 4 are conserved in the C. tritonis echotoxin-like protein. Also, a lysine residue, which is predicted to be involved in the assembly of actinoporin molecules for pore-formation to lipid membranes 105 , is present. A phylogenetic analysis supports, with high confidence, the evolutionary origin of M. echo echotoxins being most closely related to C. tritonis (Fig. 5B).
Echotoxins lyse erythrocytes following binding to gangliosides, which is a similar mechanism to that of some bacterial hemolysins 106 , yet dissimilar to the marine hemolysins, for example, sea anemone hemolysins bind to sphingomonnyelin 107 . Sea anemones are a rich source of lethal pore-forming toxins (PFTs) that may include a combination of peptides and proteins, known as cytolysins or actinoporins 108 . PFTs target cell membranes forming water-filled pores across the lipid bilayer, followed by oligomerization and penetration of the protein subunits through the lipid bilayer. The discovery of actinoporin-like hemolysins (echotoxins) within higher eumetazoans is of particular interest in comparative biochemistry. However, it should be noted that echotoxins and actinoporins have distinct modes of action 101,107 . Interestingly, a sequence homology search using C. tritonis echotoxins revealed some similarity (a three-turn alpha helix and beta sheet) with fragaceatoxin, an actinoporin-type of pore forming hemolytic protein from sea anemone (SWISS-MODEL ID: 4tsn.3.A) (Fig. 5C). The crystal structures of two other actinoporin proteins from sea anemone, equinotoxin II and sticholysin II, both revealed a compact beta-sandwich consisting of ten strands in two sheets flanked on each side by two short alpha-helices, which is a similar topology to osmotin, a plant defense protein belonging to the fifth class of the pathogenesis-related proteins (Pr-5) 109,110 . Studies have reported that the beta sandwich structure attaches to the membrane, while a three-turn alpha helix lying on the surface of the beta sheet may be involved in membrane pore formation, possibly via the penetration of the membrane by the helix [109][110][111][112] . Additionally, computer-aided protein structure prediction identified a ligand-binding site for phospohocholine on C. tritonis echotoxin (Fig. 5C). In sea anemone, small and basic α-pore forming actionporin proteins have a phosphocholine binding site which facilitates binding to the cell membrane and formation of pores, a feature that they share with toxins such as diphtheria and anthrax 111 . The identification of echotoxins from C. tritonis suggests a broad role for these in marine gastropod snails, and are most likely important for prey interaction.
Analysis of sulfuric acid biosynthesis enzyme genes. Many marine gastropod snails, for the purpose of feeding and defense, release strong acids 20,21,[113][114][115][116] . Although acids can facilitate penetration through the prey calcareous body wall, the acids can also serve as an allomone through their action to deter epibiont fouling, kill/ traumatize their prey or avoid predation 112 . Several research studies within the molluscan Pleurobranchoidea family (commonly known as sea slugs) have focused on sulfuric acid production through histochemistry, showing large acid vacuoles in the median buccal gland and the subepithelial glands 113,116,117 .
In the Tonnoidea superfamily, acid production in several families, including the Ranellidae and Cassidae, has been investigated 32 . For example, in the family Ranellidae, extracts of the SG-PL of Cymatium lampas induced immediate paralysis in the sea star Patriella brevispina 35 . The PL-SGs of Cymatuim intermedius and Gyrineum natator both contain specialized acid-producing and protein-secreting epithelial cells and secrete strong acids 20,21,114 . G. natator uses sulfuric acid produced in the salivary glands to make holes in the oyster shells and also uses these secretions to attack prey only when more easily obtained food is not available 32 . These marine gastropods can also use these acid secretions for their defense 30,32,33 . In Charonia species, saliva secreted from C. tritonis putatively immobilizes COTS 35 . However, previous studies were unable to identify the chemical components of secreted saliva from Charonia sp. salivary gland; C. lampas did not inject venom or acid into its prey but rather used its foot to capture and manipulate the prey and its radula to consume the flesh 30 .
Cysteine biosynthesis genes were found in the C. tritonis SG-AL and the in silico exoproteome analysis (Supplementary Data S2a). Proteomics analysis revealed one arylsulfatase B-like protein Arylsulfatase, a sulfur scavenging enzyme, which may play a role in the breakdown of sulphated saponins 81 . Presence of a higher level of arylsulfatase in the digestive organs of the predatory mollusks has been reported 118 . It has also been confirmed that this enzyme catalyzed cleavage of sulfate in the C-4 position of xylose incorporated into carbohydrate chains of saponins from sea cucumbers 81 . In triton snail, after breakdown of saponin by arylsulfatase, released sulfates may be processed in the SG-AL to synthesize sulphuric acid. Three major enzymes required for the biosynthesis of sulfuric acids, serine acetyltransferase, cysteine synthase and cysteine dioxygenase, have been found in C. tritonis (Fig. 6A) and domain conservation for those enzymes in SG-AL tissue subsequently identified (Fig. 6B). Although the present study reveals the genetic metabolic tools required for biosynthesis of sulfuric acids in the SG-AL, further studies are necessary to confirm its production.

Conclusions
We have described through analysis of tank assays the process involved in C. tritonis attack on COTS, including proboscis extension, penetration, and the likely secretion of SG-derived feeding and putative toxin-related proteins. Transcriptome and proteome analysis of the SG-AL and SG-PL have identified putative venom-and feeding-related proteins. This work provides insight into the source of bioactive components used by C. tritonis to predate on COTS.
Data Accessibility. Raw sequence data for transcriptome assemblies can be found at NCBI PRJNA383875.
Protein sequences for all species proteins used in this investigation are provided in File S8.