Fish-hunting cone snail disrupts prey's glucose homeostasis with weaponized mimetics of somatostatin and insulin

Abstract

Venomous animals have evolved diverse molecular mechanisms to incapacitate prey and defend against predators. Most venom components disrupt nervous, locomotor, and cardiovascular systems or cause tissue damage. The discovery that certain fish-hunting cone snails use weaponized insulins to induce hypoglycemic shock in prey highlights a unique example of toxins targeting glucose homeostasis. Here, we show that, in addition to insulins, the deadly fish hunter, Conus geographus, uses a selective somatostatin receptor 2 (SSTR₂) agonist that blocks the release of the insulin-counteracting hormone glucagon, thereby exacerbating insulin-induced hypoglycemia in prey. The native toxin, Consomatin nG1, exists in several proteoforms with a minimized vertebrate somatostatin-like core motif connected to a heavily glycosylated N-terminal region. We demonstrate that the toxin's N-terminal tail closely mimics a glycosylated somatostatin from fish pancreas and is crucial for activating the fish SSTR₂. Collectively, these findings provide a stunning example of chemical mimicry, highlight the combinatorial nature of venom components, and establish glucose homeostasis as an effective target for prey capture.

Fig. 1. The net-hunting cone snail, Conus geographus, uses mimetics of the peptide hormones insulin and somatostatin (SS) to disrupt glucose homeostasis in fish prey.

A Images of C. geographus releasing venom into the water to catch fish. Photographs taken by Dylan Taylor. B Schematic overview of the central hypothesis: venom insulins (con-insulins) induce dangerously low blood glucose in fish. Consomatins block the secretion of the insulin-counteracting hormone glucagon via potent and selective activation of the somatostatin receptor 2 (SSTR₂). Together, these toxins induce sustained hypoglycemia facilitating prey capture. C Sequence alignment of the predicted C. geographus consomatins, pG1 and pG2, with the SS drug analog octreotide and human and fish somatostatin-14 (SS-14). Residues important for SSTR activation are highlighted in red. Disulfide bonds are depicted as connecting lines. The chemical structure of Consomatin pG1 is shown below the sequences. Modifications: f, D-phenylalanine; w, D-tryptophan; -ol, alcohol. D Con-insulins and consomatins show similar patterns of expression in the C. geographus venom gland. Both classes of toxins are some of the most highly expressed transcripts and predominantly expressed in the distal region of the venom gland (segment B4), closest to the pharynx (inset).

Fig. 2. Consomatin pG1 is a potent full agonist of the human SSTR₂ (hSSTR₂) that displays G_o protein bias.

A Consomatin pG1 selectively activated hSSTR₂ when measured in the PRESTO-Tango β-arrestin recruitment assay. X indicates non-determinable potency values due to a lack of or weak SSTR activation. B Representative dose-response curves of Consomatin pG1 at the hSSTR₂ in comparison to octreotide and SS-14. Error bars represent means $\pm$ S.E.M. Experiments were conducted in technical triplicate, n = 4 biological replicates. C–G Ability of Consomatin pG1 to induce dissociation of G_i1 (n = 4)_, G_i2 (n = 5)_, G_i3 (n = 4), G_oA (n = 4) and G_oB (n = 3) subunits in comparison to octreotide and SS-14. Representative dose-response curves are shown. Error bars represent means $\pm$ S.E.M. All experiments were conducted in technical duplicates. H Bias plot in terms of ∆Log(τ/K_A) using SS-14 as reference showing that Consomatin pG1 displays G protein bias towards G_o subunits.

Fig. 3. Consomatin pG1 suppresses glucagon secretion in mouse islets of Langerhans and perfused rat pancreas.

A Schematic overview of cell types and SSTR subtypes involved in hormone secretion in pancreatic islets. B Consomatin pG1 suppresses glucagon secretion in isolated mouse islets under low glucose conditions (1 mM glucose) (n = 9 biological replicates). Data represents means $\pm$ S.E.M. ** shows p < 0.01 (two-way repeated measure ANOVA, Šidák test). p = 0.0049 for comparing control vs pG1 (1 nM) and p = 0.0024 for comparing control vs SS-14 (1 nM) at 1 mM glucose condition; n.s.: not statistically significant. C 0.1–10 nM of Consomatin pG1 potently suppresses glucagon secretion in perfused rat pancreas (n = 10 biological replicates). Error bars represent means $\pm$ S.E.M. D Bar plot illustrates the comparison of average glucagon output (fmol / min) of different doses of pG1 with baseline. Data represents means $\pm$ S.E.M. measured for n = 10 rats. ** shows p < 0.01. p = 0.0032 for comparing baseline vs pG1 (1 nM) and p = 0.0061 for comparing baseline vs pG1 (10 nM) (one-way repeated measure ANOVA, Dunnett’s test); n.s.: not statistically significant.

Fig. 4. Activity of Consomatin pG1 and crude C. geographus venom at the fish and human receptors.

A Gene tree of the human (Hs), mouse (Mm), and fish (Danio rerio, Dr; Salmo salar, Ss) SSTR subtypes with SSTR₂ highlighted in yellow. B, C Consomatin pG1 activates the zebrafish Dr-sstr2b (n = 3) but not Dr-sstr2a (n = 6) isoforms when measured in the PRESTO-Tango β-arrestin recruitment assay in technical triplicates while D crude C. geographus venom is capable of activating both Dr-sstr2 isoforms while retaining selectivity for the human SSTR₂. Representative dose-response curves are shown. Error bars represent means $\pm$ S.E.M. of n = 4 biological replicates. Experiments were conducted in technical duplicates.

Fig. 5. Activity-guided bioassays lead to identification of C. geographus subfraction 32-20 as the most potent subfraction in activating hSSTR₂ and fish Dr-sstr2 isoforms.

A Reversed-phase high performance liquid chromatography (RP-HPLC) of C. geographus venom used for activity-based identification of native consomatins. B Screening of venom fractions identified fraction # 32 as the most potent fraction in activating the G_oA subunit at the hSSTR₂. Error bars represent means $\pm$ S.E.M. (n = 2 biological replicates) in technical duplicates. C Further subfractionation using RP-HPLC of fraction # 32 identified D subfraction # 32–20 as the most potent in activating hSSTR₂, and both Dr-sstr2a and Dr-sstr2b. Error bars represent means $\pm$ S.E.M. (n = 3 biological replicates) in technical duplicates.

Fig. 6. Consomatin nG1 and nG2 exist in several proteoforms that carry various different core-1 glycans.

A Sequences of nG1 and nG2 and intact masses identified in the active venom fraction, # 32-20 by MS, MS/MS and Edman sequencing (see Supp. Fig. S11-S14 for MS/MS spectra, intact mass calculations and Edman sequencing results). Tryptic fragments are indicated above the sequences. Glycan structures correspond to those highlighted in panel B. Residues that differ between nG1 and nG2 are shown in green. Disulfide bonds are depicted as connecting lines. Modifications: Z, pyroglutamic acid; w, predicted D-tryptophan; O, hydroxyproline, T = O-glycosylated threonine. B Deconvoluted mass spectrum of precursor ions in subfraction # 32-20 that correspond to differentially glycosylated proteoforms of Consomatin nG1 and nG2 (indicated by red arrows). Minor peaks correspond to proteoforms containing additional hydroxylations on proline and potentially valine and lysine residues ( + 15.9949, peaks not labeled). Sequences including calculations for modifications are shown in Supp. Fig. S12. C LC-MS/MS analysis of released O-glycans after labeling with 2-aminobenzamide (2-AB) revealed that the venom fraction mainly contains core-1 mucin type O-glycans with various degrees of fucosylation, as well xylose (Xyl) and hexose (Hex; possibly mannose) initiated glycan species.

Fig. 7. Consomatin G1 evolved from an endogenous somatostatin and related peptides (SSRP)-like gene to mimic both stable SSTR₂-selective drug analogs and the glycosylated tail of the fish ss4 peptide family.

A Alignment of the full-length precursors of the endogenous SSRP-like signaling peptide from Conus geographus, a canonical consomatin toxin from Conus betulinus, and Consomatin G1. Signal sequences are highlighted in green and the mature peptide regions are shown in yellow. B Gene structures of the sequences shown in panel B show an additional exon for Consomatin G1 that encodes the extended, glycosylated N-terminal tail. C Sequence alignment of the therapeutic SSTR₂-selective drug analog octreotide, native Consomatin nG1 and nG2, the endogenous SSRP-like peptide from C. geographus, and ss4 sequences retrieved from various fish species, including the glycosylated peptide from the channel catfish. The glycosylated pig SS-28 which has the same sequence as human SS-28 is shown for comparison. Glycosylated residues are underlined and highlighted in blue. Residues that may be glycosylated are depicted in light blue. Amino acids known to be important for the activation of the vertebrate SSTRs are highlighted in red. Modifications, including disulfide bonds, are indicated as shown in Fig. 6.