Decoding deception: the binding affinity of cuttlefish ink on shark smell receptors

Abstract

Chemical signaling can play a crucial role in predator–prey dynamics. Here, we present evidence that ink from the common cuttlefish (Sepia officinalis) targets olfactory receptor proteins in sharks, potentially acting as a predator deterrent. We apply in silico 3D docking analysis to investigate the binding affinity of various odorant molecules to shark olfactory receptors of 2 shark species: cloudy catshark (Scyliorhinus torazame) and white shark (Carcharodon carcharias). Pavoninin-4 (a known shark repellent compound) displayed selectivity in binding to receptors in the white shark. In contrast, the primary component of cuttlefish ink, melanin, displayed the highest binding affinities to all shark olfactory receptor proteins in both species. Taurine, another important ink component, exhibited standard to strong bindings for both species. Trans-4,5-epoxy-(E)-2-decenal (“blood decenal”), an odorant associated with the smell of blood, displayed strong binding affinities to all shark olfactory receptors, similar to that of melanin. These findings provide new insights into the molecular interplay between cephalopod inking behavior and their shark predators, with cuttlefish ink likely exploiting the narrow band of the shark olfactory repertoire.

Graphical Abstract

ligand–receptor interaction, protein structure prediction, olfactory genes

Introduction

Chemical signals carry information about the environment, allowing animals to locate food, find mates, and avoid danger. As such, olfaction, the chemical sensory mode called “smell,” is essential for survival in both predator and prey species. For example, barnacles secrete glycoprotein compounds as a mechanical defense against predation, yet many of their predators (e.g. sea stars, whelks, and snails) have evolved olfactory receptors to detect these glycoproteins, allowing them to locate the barnacles more efficiently (Zimmer et al. 2021). However, reliance on olfaction can also present a vulnerability for predators. For example, the California sea hare, Aplysia californica, expels a dark liquid (“ink”) as a defensive strategy against spiny lobsters, which is thought to deter predation through a combination of phagomimicry and chemical sensory disruption (Derby 2007).

Olfactory receptors are chemosensory proteins belonging to the family of G-protein-coupled receptors and, in vertebrates, are often expressed in the cell membrane of the nasal epithelium (Firestein 2001). They play a vital role in processing chemical signals from the external environment. Activation, by binding with an agonist or antagonist molecule, initiates a signaling cascade that ultimately results in perception of a chemical stimulus (“smell”) (Peterlin et al. 2008). The 4 primary olfactory receptor gene families responsible for vertebrate olfaction are (1) odorant receptors (ORs), (2) trace amine-associated receptors (TAARs), (3) olfactory receptors related to class A (ORA, also: V1Rs), and (4) vomeronasal type 2 receptors (V2Rs, also: OlfCs). These receptor families were initially characterized in mammals (Buck and Axel 1991; Dulac and Axel 1995; Matsunami and Buck 1997; Liberles and Buck 2006) and have subsequently been recognized in other vertebrate groups, including birds (Hillier et al. 2004; Niimura and Nei 2005) and other tetrapod clades (Dehara et al. 2012), ray-finned fishes (Niimura 2009) and more recently sharks (Hara et al. 2018; Marra et al. 2019; Sharma et al. 2019; Syed et al. 2023).

Coleoid cephalopods (squid, octopus, and cuttlefish) are thought to have diverged from other mollusks ∼500 MYA (Derby 2014). While the loss of external shells has increased their mobility relative to other mollusk groups, it has also increased predation risk (Wells 1995). In response, coleoid cephalopods have evolved alternative predator deterrence mechanisms, notably ink (Derby et al. 2013). Ink released by the common cuttlefish (Sepia officinalis) is principally composed of melanin, the pigment responsible for its dark color, and multiple free amino acids, of which the nonproteinogenic amino acid taurine is found in the highest concentration (Derby 2007). Sharks (Chondrichthyes and Selachii) are significant cephalopod predators, and there is evidence showing that cuttlefish ink has an adverse effect on shark swimming behavior (Simonitis 2021).

Sharks have informally been referred to as “swimming noses.” The relatively large olfactory bulbs (OBs) observed in sharks have led to the popular notion of their having an acute sense of smell (Hodgson and Mathewson 1978). Recent sequencing of multiple shark genomes (Hara et al. 2018; Marra et al. 2019; Tan et al. 2021; Stanhope et al. 2023) has revealed a surprisingly low number of sensory-specific olfactory receptors compared with other vertebrate clades. Sharks exhibit an average of 43 total olfactory receptor genes, whereas ray-finned fishes typically average >200, and mammals ∼850 (Policarpo et al. 2023). As such, potential olfactory acuity and molecular breadth of total olfactory ability appear decoupled in this clade.

Here, we investigate the exploitation of predator olfactory receptors via ink in the shark–cuttlefish system. Using computational approaches, we modeled the 3D structures of olfactory receptor proteins from 2 ecologically distinct shark species, the cloudy catshark (Scyliorhinus torazame), and the white shark (Carcharodon carcharias). By examining monomolecular binding affinities for cuttlefish ink components, the known shark repellent compound pavoninin-4 (Tachibana et al. 1984), 2 odorants associated with putrefaction (cadaverine and putrescine) and the odorant responsible for the smell of blood (trans-4,5-epoxy-(E)-2-decenal, hereafter: “blood decenal”), we elucidate how a prey species may disrupt shark sensory perception through simultaneous binding across the entire suite of multiple divergent olfactory receptors. Our results highlight how 3D protein modeling can complement our understanding of the life history and ecology of predator–prey interactions.

Materials and methods

Gene selection, protein structure prediction, and odor docking

Olfactory receptor gene sequences (OR, TAAR, V1R/ORA, and V2R/OlfC) were selected from Sharma et al. (2019) and Syed et al. (2023), with receptor names assigned based on the small-spotted catshark (Scyliorhinus canicula) gene nomenclature. Genes were aligned using Clustal Omega (Sievers et al. 2014). Each alignment was used to generate phylogenetic trees using W-IQ-TREE, a server for maximum likelihood analysis using the IQ-TREE software, with 1,000 ultrafast nonparametric bootstrap replicates (Nguyen et al. 2014; Trifinopoulos et al. 2016) and the inferred best-fit model of sequence evolution. One representative receptor per subclade was then chosen across all olfactory gene families. This provided targets to examine odorant–receptor binding combinations using a variety of chemical compounds (volatile and nonvolatile odorants). Given that olfactory receptors are membrane-bound, it is difficult to isolate and resolve their atomic structure via X-ray crystallography. Therefore, binding combinations were investigated using in silico methods. 3D odorant structures used in this experiment were downloaded from PubChem (Kim et al. 2022; Supplementary Table 1). The following odorants were selected: (1) melanin secreted from common cuttlefish (S. officinalis) which gives ink the dark color, (2) taurine, a free nonproteinogenic amino acid found at millimolar concentrations in cuttlefish ink (Derby et al. 2013), (3) pavoninin-4, a toxin secreted from the Pacific sole (Pardachirus pavoninus; Williams et al. 2005), (4) cadaverine, (5) putrescine (Hussain et al. 2013), and (6) blood decenal (Sarrafchi and Laska 2016). Cadaverine was also used as a positive control to test the reliability of our odor docking assay, as it has previously been shown to have a high binding affinity for TAAR13c in zebrafish (Danio rerio; NP_001076509.1), the binding of which causes an avoidance response (Hussain et al. 2013). Based on this, the score obtained from initially docking cadaverine to the zebrafish TAAR13c 3D structure was used as the relative threshold for a high binding affinity.

Receptor 3D structures were predicted using their amino acid sequence (Supplementary material) as input to the Iterative Threading ASSEmbly Refinement (I-TASSER) protein modeling server (Yang and Zhang 2015; Zheng et al. 2021; Zhou et al. 2022). I-TASSER identifies structural templates from the PDB using multiple threading alignment approaches and attempts to infer the best possible structural prediction using homology modeling. Up to 5 structural models are predicted, each with an estimated global confidence score (C-score) that quantifies the reliability and quality of predicted models. Typically ranging between −5 and 2, a higher C-score indicates greater confidence in the model. We selected the model with the highest C-score for subsequent protein docking. Predicted ligand-binding sites were identified with COFACTOR (Roy et al. 2012; Zhang et al. 2017) and COACH (Yang et al. 2013) and implemented in I-TASSER. Custom cavity docking was performed by CB-dock2 (Liu et al. 2022), which predicts odorant docking contact sites with the protein. UCSF Chimera (Pettersen et al. 2004) was used for visualization and subsequent analyses. To investigate the olfactory potential of shark species, we performed custom molecular docking simulations for 6 odorant molecules: 3 environmental odorants potentially under selection (putrescine, cadaverine, and blood decenal); a known shark repellent odorant (pavoninin-4); and 2 molecules found in cuttlefish ink (taurine and melanin).

Results

Gene trees and selecting targets to model

A total of 146 olfactory receptor gene sequences were sourced from previously published data (Supplementary Table 2; Sharma et al. 2019; Syed et al. 2023). This dataset included multiple species, specifically the cloudy catshark, the white shark for subsequent protein modeling, and the small-spotted catshark, which served as a naming reference point for the receptors used in this study (Supplementary Fig. 1). The dataset included 29 ORs, 14 ORAs, 9 TAARs, and 94 V2Rs across all species. The mean amino acid residue length varied between the olfactory families: ORs have an average length of 303 amino acids, while the (multiexon) V2Rs are significantly longer at 643 amino acids. For subsequent protein modeling efforts, the final selection of receptors consisted of 1 OR, 1 TAAR, 1 ORA, and 4 V2Rs sourced from both shark species analyzed (Supplementary Fig. 1).

In silico binding of the death-associated odor cadaverine to zebrafish TAAR13c

The death-associated odor compound cadaverine has been shown to bind strongly in vivo to the zebrafish TAAR13c receptor (Hussain et al. 2013). We modeled the 3D structure of zebrafish TAAR13c, and docked cadaverine, in silico, as a positive control for the reliability of our computational docking assay. COFACTOR (Roy et al. 2012; Zhang et al. 2017) and COACH (Yang et al. 2013) identified 38 contact residues within the upper third portion of the transmembrane domain of TAAR13c (Supplementary Table 3). The docking results highlighted Asp112 and Trp296 (located on transmembrane 7 and transmembrane 8, respectively) as the key residues for cadaverine binding, in agreement with previous studies that predicted these residues as binding partners (Huang 2003; Sharma et al. 2016; Supplementary Fig. 2). Docking cadaverine with zebrafish TAAR13c resulted in a docking score of −4.0 (where more negative scores represent greater binding affinities). We used this score as a threshold indicator of high ligand–receptor binding affinity. It is important to note 1 limitation of in silico docking experiments here. Geometric fit of ligands at the active site is measured under a static model (Terefe and Ghosh 2022); therefore, dockings generally do not account for protein flexibility, which in turn can affect the physiochemical environment in the binding region.

In silico binding assay to shark ORs

Protein structure predictions for each shark olfactory receptor were generated with I-TASSER (Fig. 1; Yang and Zhang 2015; Zheng et al. 2021; Zhou et al. 2022). The reliability of each model was assessed using the C-score, with higher scores indicating greater confidence (Supplementary Table 4). Predicted ligand-binding sites for shark olfactory receptors are given in Supplementary Table 5. There were minute differences in binding patterns between the 2 species (Supplementary Table 6 and Fig. 2). Odorant docking scores correlate with binding free energy; therefore, more negative scores indicate stronger binding of the odorant to the olfactory receptor protein (García-Ortegón et al. 2022). Melanin and blood decenal showed the highest binding affinities, with docking scores ≤−4.0, across all receptors for both shark species (Fig. 2). Taurine and cadaverine exhibited moderate-to-high binding affinities for different receptor–odorant combinations. Putrescine displayed moderate binding affinities across all receptors in both shark species. One discrepancy in the pattern of binding between sharks was observed for pavoninin-4. The white shark displayed very low binding affinity (docking score: 3.9) for the OR1 receptor, whereas this OR1 showed high affinity (−7.8) in the cloudy catshark (Supplementary Table 6).

Fig. 1.

Representative predictive 3D receptor structures of 4 major olfactory receptor types found in the white shark OR, TAAR, and V2R genes are generally encoded by a single exon and consist of 7 transmembrane helices, whereas V2R genes are generally encoded by 6–7 exons and possess long N-terminal tails. OR, odorant receptor; TAAR, trace amine-associated receptor; ORA, olfactory receptors related to class A; V2R, vomeronasal type 2 receptor.

Fig. 2.

Odorant–receptor heat map. Each shark receptor was docked with 6 different odorants (putrescine, cadaverine, taurine, melanin, blood decenal, and pavoninin-4) to examine the binding affinity. Each row represents an odorant, and each column represents an olfactory receptor. The zebrafish/cadaverine complex was used to indicate successful binding (−4 shown in white). High binding affinity is represented by warmer tones, lower binding affinity by cooler tones and non-functional genes are indicated by a tone that a distinct from the gradient extremes.

Discussion

Sharks exhibit a range of relative OB sizes, depending on ecology and habitat (Yopak et al. 2010, 2014), with the white shark, a large-bodied, high-speed predator in coastal/oceanic habitats, possessing one of the largest relative OB sizes among sharks (Yopak et al. 2014). Among vertebrates, a larger relative size for sensory processing regions of the brain (e.g. OB and optic tectum) is thought to reflect an increased reliance on that particular sensory mode (Cobb 1960) and increased relative OB size is often assumed to be evidence for increased olfactory acuity (Hodgson and Mathewson 1978). Similarly, there is an assumption on the genomic side that olfactory acuity is related to overall olfactory receptor gene count (Trimmer et al. 2019); that is, acuity and the potential breadth of molecules that can be detected are coupled. There is some evidence for a correlation between OB size and total OR count for some vertebrae groups, notably birds and mammals (Hughes et al. 2018; Hughes and Finarelli 2019). Yet sensory gene mining in recently sequenced shark genomes (Hara et al. 2018; Marra et al. 2019; Sharma et al. 2019; Syed et al. 2023) has demonstrated that sharks do not adhere to this pattern, with large olfactory architecture, acute sense of smell, and breadth of the deployed protein repertoire not being disassociated in this clade. One possible hypothesis is that sharks employ a very restricted set of finely tuned receptors, in lieu of the highly expanded repertoires observed in other vertebrate groups (e.g. mammals; Hughes et al. 2018). It should also be noted that neuron density in the olfactory structures of sharks exceeds that in mammals, which suggests a highly specialized olfactory system (Aicardi et al. 2024).

Understanding how different odor molecules interact with a broad range of olfactory receptors to trigger specific responses can inform our understanding of predator–prey interactions. Pavoninin-4 is an ichthyotoxin secreted by the peacock sole (P. pavoninus) to repel predators and was researched in the wake of World War II as a potential shark repellant (Tachibana et al. 1984). Pavoninin-4 has been shown to cause a state of tonic immobility when introduced into the nares of the lemon shark (Negaprion brevirostris; Tachibana et al. 1984, 1985; Tachibana and Gruber 1988; Hart and Collin 2015). Therefore, the potential exists for prey species to employ molecules to adversely affect the sensory systems of predator species as a deterrent. 3D modeling of shark olfactory receptor proteins allows us to estimate binding affinities with potential odorant components in cuttlefish ink. Unsurprisingly, pavoninin-4 has strong binding affinities across most of the shark olfactory receptor proteins, with only OR1 and TAAR1a in the white shark showing weak binding (Fig. 2). Both melanin and taurine bind across the entire repertoire of olfactory receptor proteins, with melanin displaying very high binding scores uniformly across both species (Fig. 2). This is despite the fact that the cloudy catshark and the white shark are ecologically distinct (Park et al. 2019; Guttridge et al. 2024), and their lineages diverged ∼150 MYA (Marjanović 2021). This agrees with previous research showing generally low genetic mutation rates in chondrichthyans (Martin et al. 1992; Hara et al. 2018; Sendell-Price et al. 2023). This is particularly intriguing, as the limited receptor repertoire in sharks should present fewer potential binding pocket configurations. Therefore, there is a limited, but also highly conserved, space of physiochemical characteristics presented by shark ORs across species.

The uniformly high binding affinities observed for blood decenal demonstrate its potential importance as an agonist in the shark olfactory system and also reveal the potential vulnerability to hack the entire (narrow) range of shark olfactory proteins with a single molecule. Cuttlefish have likely exploited this with melanin in their ink. Cuttlefish ink has been shown to induce avoidant swimming responses in sharks (Simonitis 2021). Melanin exhibits the same pattern of docking with all available receptor proteins as blood decenal, but it does so with stronger computed affinities. Receptor binding to OR proteins is not a binary “on/off” switch, but rather a complex, nonlinear process, with stronger binding affinities hypothesized to cause prolonged and more intense activations (Buchwald 2019; Du et al. 2023; Alsedfy et al. 2024). This suggests that melanin may induce a comparably stronger olfactory sensation than blood in sharks, that this sensation is aversive, and that it effectively obstructs the entire olfactory apparatus. Alternatively, melanin rich in amino acids may act as a phagomimic (Derby 2007; Wood et al. 2010). By activating similar olfactory receptors and signaling pathways as prey molecules, melanin could trigger feeding-associated responses in the shark, potentially misdirecting its attention toward the ink cloud and away from the inking animal. However, given the observed aversion response of sharks to cephalopod ink (Simonitis 2021), it is more likely that melanin primarily functions as a sensory disruptor.

The substantially lower counts of olfactory receptor genes in shark genomes, compared with other vertebrate groups, combined with slow rates of genomic evolution (Martin et al. 1992; Hara et al. 2018; Sendell-Price et al. 2023), suggest that olfactory capabilities are effectively fixed in this group and are not influenced by ecological factors. Cuttlefish ink has evolved to exploit a restricted and slowly evolving set of genes that presents an effectively stationary target in an evolutionary context. Our docking experiment results imply that melanin is the likely molecular mechanism of action. The demonstrated potential for ink to overcrowd the shark's olfactory perception, leading to repulsion and avoidance, represents a key element in the evolution of shark–cephalopod predatory interaction, highlighting the importance of computational biology and protein modeling in understanding prey defense at a molecular level.

Data availability

All supplementary files are available as part of the Supplementary material in this document. PDB files for 3D protein structures are available in Figshare, along with the alignment files at: https://doi.org/10.6084/m9.figshare.27934341.v1.

Supplemental material available at G3 online.

Acknowledgments

The authors thank University College Dublin for funding this research through Ad Astra grant and Ad Astra PhD studentship grand.

Funding

This work was supported by University College Dublin. G.M.H. was supported by a University College Dublin Ad Astra grant. C.L. was supported by an Ad Astra Studentship grant.

Author contributions

C.L.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, and writing—original draft; L.E.S.: writing—review and editing; J.A.F.: conceptualization, supervision, and writing—review and editing; G.M.H.: conceptualization, data curation, supervision, and writing—review and editing.

Author notes

Conflicts of interest: The authors declare no conflicts of interest.