Echinoderms produce strong reversible adhesives secreted by unique hydraulic adhesive organs called tube feet. The comprehension of this temporary adhesion has motivated several morphological, biomechanical, and biochemical studies in the last decades (Davey et al., 2021).

Adoral tube feet, in particular, are well adapted for locomotion and attachment. They possess a mobile extensible stem, topped by an adhesive viscoelastic disc (Santos and Flammang, 2005; Santos et al., 2005). In sea stars (Asteroidea), three tube foot morphotypes have been described (simple disc-ending, reinforced disc-ending and knob-ending) based on the histological structure of their disc (Santos et al., 2005). In regular sea urchins (Echinoidea), all tube feet are reinforced-disc ending, but they have been subcategorized based on the size of their disc and thickness of their stem connective tissue and retractor muscle (type 4 > 3 > 2 > 1) (Smith, 1978). These morphological differences were pointed as essential for tube feet attachment strength and consequently, for species distribution (Smith, 1978).

However, more recent biomechanical studies with three Mediterranean sea urchin species, Arbacia lixula (Linnaeus, 1758), Paracentrotus lividus (Lamark, 1816) and Sphaerechinus granularis (Lamark, 1816), and four Indian Ocean species, Colobocentrotus atratus (Linnaeus, 1758), Echinometra mathaei (Blainville, 1825), Heterocentrotus trigonarius (Lamarck, 1816) and Stomopneustes variolaris (Lamarck, 1816), belonging to three orders and five families of the Class Echinoidea, found no correlation between interspecific variations in disc tenacity (force per unit area) and the disc adhesive area (Santos and Flammang, 2006, 2008). In addition, when a tensile force is exerted on a tube foot stem, it is the connective tissue that bears the load, and not the muscle (Santos and Flammang, 2005, 2008).

Furthermore, in the three Mediterranean species significant variations in the ultrastructure of the adhesive secretory granules were reported, suggesting that there might be molecular differences in the composition of their adhesive secretions (Santos and Flammang, 2006). To test this hypothesis an antibody raised against S. granularis adhesive material was tested for cross-reactivity on tube foot histological sections from the above-mentioned species, plus Tripneustes gratilla (Linnaeus, 1776) (Santos and Flammang, 2012). This approach was used to bypass the need for complete characterization of echinoid adhesives and because it successfully evidenced the compositional similarity of sea star adhesive footprints at the class level (Santos et al., 2005). Surprisingly, in sea urchins no cross-reactivity was observed in any species tested, not even in T. gratilla that belongs to the same family as S. granularis (Santos and Flammang, 2012).

After these findings, several biochemical studies attempted to fully characterize the adhesive material but focused exclusively on Paracentrotus lividus. The analysis of adhesive footprints showed that they consisted of a honeycomb-like meshwork of aggregated globular nanostructures (Viana and Santos, 2018) composed of proteins, neutral sugars, lipids and inorganic residues (Santos et al., 2009). The adhesive material proteome revealed a prevalence of five protein groups (actins, tubulins, myosins, ribosomal proteins and histones) and only one cell-adhesion protein, P. lividus Nectin (Lebesgue et al., 2016). A recent re-analysis of this proteome using a tube foot specific transcriptome, combined with a quantitative proteome analysis of the adhesive disc versus the non-adhesive stem, and a validation of gene expression using in situ hybridization (ISH), drastically increased the mapped proteins and highlighted sixteen transcripts potentially involved in bioadhesion (Pjeta et al., 2020). Of these, six transcripts (Nectin, alpha-tectorin, uncharacterized protein, Myeloperoxidase, neurogenic locus notch homolog protein and alpha-macroglobulin) presented a ISH expression pattern consistent with the location of the adhesive secretory cell bodies, and simultaneously possessed an ortholog adhesion-related transcript in the sea star Asterias rubens (Linnaeus, 1758) (Lengerer et al., 2019; Pjeta et al., 2020).

P. lividus-Nectin is a 210-kDa homodimer glycoprotein consisting of two polypeptides with an equal mass of 105 kDa each, joining covalently by S–S bridges (Zito et al., 1998). It contains 6 tandemly repeated discoidin-like (or F5/8 type C) domains predicted to bind molecules bearing galactose and N-acetylglucosamine carbohydrate moieties (Costa et al., 2010), and a LDT motif predicted to be the binding site to an α4/β7 integrin receptor (Zito et al., 2010). It was first discovered in the extracellular matrix (ECM) of P. lividus embryos. It is also present in the unfertilized egg cytoplasm, stored in granules, and is released into the ECM surrounding the embryo after fertilization. In later developmental stages, it polarizes on the apical surface of ectodermal and endodermal cells. Thus, P. lividus-Nectin is involved in cell adhesion processes as an integrin ligand and its contact to ectodermal cells is essential for correct larval skeletogenesis (Zito et al., 2000, 2010; Costa et al., 2010).

Nectin was first related to adult P. lividus adhesion when it was identified both in tube feet and adhesive footprints using antibodies raised against embryonic Nectin (Lebesgue et al., 2016). Although present in both the tube foot disc and stem, it was shown to be highly overexpressed at the mRNA and protein level in the adhesive discs, being regulated by the degree of hydrodynamism to which the sea urchin is exposed (Lebesgue et al., 2016; Toubarro et al., 2016). So far, in adult P. lividus tube feet, three Nectin variants, the embryonic Nectin (variant 1 – Uniprot Q70JA0) plus two others (variant 2 – Uniprot A0A182BBB6; variant 3 – tube foot transcriptome), differing in only a few amino acid substitutions, have been reported (Lebesgue et al., 2016; Toubarro et al., 2016; Pjeta et al., 2020). In addition, several Nectin isoforms have been observed in 2DE gels, presenting different degrees of phosphorylation and glycosylation (Santos et al., 2013).

The remaining five P. lividus transcripts potentially involved in bioadhesion (Pjeta et al., 2020) possess domains that are recurrent in other marine adhesive and cohesive proteins (Davey et al., 2021) but require further investigation. P. lividus alpha-tectorin shares domains present in adhesive proteins from sea stars (Sfp1; Hennebert et al., 2014), flatworms (Mlig-ap1 and −2, Mile-ap1 and Mile-ap2a/b; Pjeta et al., 2019; Wunderer et al., 2019), cnidarians (Rodrigues et al., 2016) and a terrestrial slug (Smith et al., 2017). The identification of Myeloperoxidase in P. lividus (Lebesgue et al., 2016; Pjeta et al., 2020) agrees with reports of peroxidase-like proteins being present in the adhesive secretions of sea star (Hennebert et al., 2015), cnidarians (Rodrigues et al., 2016), caddisfly larvae (Wang et al., 2014), and adult barnacles (So et al., 2017). Peroxidases are believed to act as catalyzers of protein crosslinking within the adhesive, thus, contributing to its high cohesive strength (Pjeta et al., 2020). P. lividus neurogenic locus notch homolog protein contains trypsin inhibitor-like cysteine-rich domains, also present in sea star Sfp1 (Hennebert et al., 2014), that can form disulfide bonds and possibly contribute to echinoderm adhesive insolubility attributed to the presence of proteins with significant amounts of cysteines (Santos et al., 2009; Pjeta et al., 2020). Finally, P. lividus alpha-macroglobulin share common domains with several proteins present in the adhesive secretions of sea stars (Hennebert et al., 2015; Lengerer et al., 2019), limpets (Kang et al., 2020), tunicates (Li et al., 2019) and barnacle larvae (Dreanno et al., 2006; Ferrier et al., 2016).

The glycosidic fraction of the adhesive material has also been studied in P. lividus, demonstrating the involvement of high molecular weight glycoproteins containing N-acetyl glucosamine residues. Five lectins (GSL II, WGA, STL, LEL, and SBA) out of 22 specifically labeled the disc adhesive epidermis and the adhesive footprints and detected several disc specific glycoproteins (Simão et al., 2020). Of these, LEL, that recognizes N-acetyl glucosamine in a chitobiose arrangement [GlcNAc β(1,4)GlcNAc], specifically labeled the adhesive secretory granules within the characteristic sea urchin ‘apical tuft’ secretory cells and produced an intense labeling of the footprint, indicating that a glycoprotein containing chitobiose is most likely a main component of P. lividus adhesive secretion (Simão et al., 2020). This agrees with growing evidence showing the importance of glycoproteins in marine adhesives, from permanent (mussels, barnacles, algal spores) to non-permanent ones, such as transitory (limpets, marsh periwinkle) or temporary (sea stars, flatworms, ascidian larvae) adhesives (Dreanno et al., 2006; Hennebert et al., 2011; Ferrier et al., 2016; Pjeta et al., 2019; Wunderer et al., 2019; Zeng et al., 2019; Kang et al., 2020).

In this context, the present study aims at unraveling the evolutionary history of adhesion among echinoids by comparing, in terms of the adhesive composition, P. lividus with three sympatric species occurring in Madeira Island (NE Atlantic) that belong to different taxa and bear tube feet with different morphologies. P. lividus Nectin sequence was used to identify homologous sequences in publicly available echinoid proteomes and transcriptomes. In addition, an immunohistochemical analysis using antibodies against Nectin was performed on tube foot histological sections and adhesive footprints. These antibodies were also used on western blots to detect these proteins in tube foot disc and stem extracts. The conservation of the glycosidic fraction was also approached using the five lectins that detected adhesion-specific glycoproteins in P. lividus, by looking for cross-reactivity on histological sections and in blotting assays with the remaining species.

Increasing interest in biological adhesives has been partly driven by the demand for novel biomimetic adhesives with capabilities beyond the synthetic glues currently available to consumers (Davey et al., 2021). However, the precise mechanisms responsible for the superiority of natural bioadhesives remain largely unknown. Adhesives secreted by aquatic invertebrates contain proteins, glycans and lipids in varying proportions, as well as metals involved in crosslinking (Richter et al., 2018). Although adhesives from a growing number of organisms belonging to different taxa have been characterized, these studies are often based on a single species (Lengerer et al., 2019). More interspecific comparisons between closely related and unrelated species are needed to identify shared features, such as biased amino acid distribution, repetitive regions, and recurrent functional domains in putative adhesive and cohesive proteins. This information can provide clues on the key components of future biomimetic adhesives based on a particular amino acid (e.g., catechol-containing polymers inspired by DOPA in mussels; North et al., 2017), or on protein sequence repetitive regions and recurrent functional domains (e.g., functional domain-containing recombinant proteins based on the sequence of sea star Sfp1, Lefevre et al., 2020).

Recent comparative inter-phylum analyses of adhesive proteins revealed conserved blocks of different domains indicative of common evolutionary origin (Davey et al., 2021). It is currently known that the association of the domains vWD–C8–TIL, typical of vertebrate gel-forming secreted proteins like mucins, is recurrent in aquatic putative adhesive and cohesive protein such as sea urchin alpha-tectorin like protein (TR63383_c2_g1_i1), sea star Sfp1 (echinoderm), flatworm Mlig-ap1 and −2 (platyhelminth), and limpet P-vulgata_3 (mollusc) (Hennebert et al., 2014; Wunderer et al., 2019; Kang et al., 2020; Pjeta et al., 2020). This might indicate that these proteins evolved from a common mucin-like ancestor (Davey et al., 2021).

To find sequence conservation at the amino acid level, interspecific comparisons within the same phylum or lower taxonomic levels must be performed. In sea stars, Sfp1-like sequences were found in 17 species, representative of 10 families from four orders (Lengerer et al., 2019). However, high sequence variability between the species, prevented the use of an antibody directed against a specific peptide of Sfp1, thus restricting cross-immunoreactivity within the disc adhesive epidermis to two out of the 24 tested species (Lengerer et al., 2019). In sea urchins, no cross-reactivity was found in the adhesive disc epidermis of seven species belonging to three orders and five families of the Class Echinoidea using an antibody raised against the bulk adhesive of one species (Santos and Flammang, 2012).

In the present study, we demonstrated that an antibody against P. lividus Nectin whole protein produces cross-immunoreactivity in the tube foot epidermis and/or in the adhesive footprints of the four tested species belonging to four families (Diadematidae, Arbaciidae, Toxopneustidae, and Echinidae) and three orders (Diadematoida, Arbacioida, and Camarodonta) (Figure 10). More intense labeling, possibly indicative of a higher protein sequence homology and consequently higher antibody affinity, was observed in S. granularis which is phylogenetically closer to P. lividus (both belong to order Camarodonta). A similar result was obtained when we compared Nectin-like protein sequences from ten species belonging to six families (Cidaridae, Arbaciidae, Toxopneustidae, Strongylocentrotidae, Echinometridae, and Echinidae) and three orders (Cidaroida, Arbacioida, and Camarodonta). Closely related species belonging to the same family have Nectins with higher sequence homology (Supplementary Table 2).

Our immunohistochemical assays show that Nectin-like proteins are present both in the non-adhesive stem and in the adhesive disc of P. lividus, D. africanum, A. lixula and S. granularis, indicating that, in adult sea urchins, it might maintain an important cell adhesion role. However, a role in tube foot adhesion cannot be discarded because Nectin is consistently present in adhesive footprints. Moreover, like sea star Sfp1, it possesses several tandemly repeated discoidin-like (or F5/8 type C) domains, which allow protein-carbohydrate interactions. However, Sfp1 is more clearly associated with sea star tube foot adhesion because it is present exclusively in the adhesive granules of all the species studied (Hennebert et al., 2014; Lengerer et al., 2019).

The use of five lectins to detect N-acetylglucosamine (GSL II), chitobiose (WGA, STL, LEL) and N-acetylgalactosamine (SBA), confirmed that LEL produces a strong specific labeling of P. lividus tube foot adhesive epidermis and footprints indicating that the adhesive in this species contains N-acetyl-D-glucosamine β (1,4)N-acetyl-D-glucosamine oligomers with up to 4 carbohydrate units (Simão et al., 2020). Interestingly, in S. granularis, also from the order Camarodonta, it was WGA (and GSL II to a minor extent) which produced an intense labeling of the footprints and the cuticle covering the adhesive epidermis. This indicates that the adhesive in S. granularis also contains chitobiose but with a lower number of units than in P. lividus. These glycans seem to be conjugated to proteins since LEL in P. lividus and WGA in S. granularis pinpointed strongly labeled bands at 35 and > 100kDa (Figure 10). Whether these glycoproteins are homologous remains unanswered, but similar glycans in phylogenetically related species are demonstrated here. For A. lixula, no specific labeling was obtained with the tested lectins. The tube foot adhesive epidermis in D. africanum was stained with SBA, suggesting the presence of N-acetylgalactosamine, but lectin-blots did not corroborate this. The lectin-blots also revealed the presence of two glycoproteins (around 75 and 135 kDa) containing N-acetylglucosamine and N-acetylgalactosamine residues that are conserved in all the species, being present both in disc and stem extracts. These two proteins had been previously reported in P. lividus, being present in the cytoplasm and microvilli of epidermal support cells (Simão et al., 2020). Torn microvilli might explain the fluorescent labeling observed in the footprints of all the species with the five tested lectins. Future studies should perform a full lectin screening for each sea urchin since the composition of the glycosidic fraction of their adhesives seems to be quite variable between species. It should also be stressed that lectins produced much higher corrected fluorescence of the footprints (Figures 4, 9) and a more precise detection of disc-specific proteins (Figure 5) than antibodies. One explanation could be that Nectin detected by the antibody used, is not so relevant for sea urchin adhesion, compared to other glycoproteins detected by the lectins. Glycoproteins are ubiquitous in aquatic adhesives (see Introduction) and glycosylation is pointed to increase conformational stability, enhance protein-binding ability, and make proteins more resistant to degradation (Rzepecki and Waite, 1993; Smith et al., 1999; Smith and Morin, 2002; Ohkawa et al., 2004; Urushida et al., 2007; Zhao et al., 2009; Hennebert et al., 2011, 2014; Pagett et al., 2012; Roth et al., 2012; Wunderer et al., 2019).

Our study demonstrates that although the external morphology of sea urchin tube feet is quite alike, their histology and secretory granule ultrastructure vary between species. As hypothesized by Santos and Flammang (2006), species inhabiting hard substrata in areas with high hydrodynamic forces possess more robust tube foot discs (with denser skeletal elements and thicker muscle and connective tissue layers) than species typical of soft substrata in less exposed habitats. Disc morphology appears independent of tube foot tenacity, since the adhesive force per unit area of individual sea urchins from species with contrasted morphology, taxonomy and ecology is not markedly different (Santos and Flammang, 2006, 2008). What seems to be a significant selective pressure to determine habitat distribution, is the size and shape of the sea urchins. D. africanum with its long thin spines and S. granularis with its big, rounded test are morphologically less adapted to cope with hydrodynamism, being dislodged at lower water velocities than A. lixula and P. lividus (Santos and Flammang, 2007; Tuya et al., 2007). Thus, interspecific tenacity differences seem to be related to dissimilarities in the adhesive composition. A considerable variation in the internal organization of adhesive secretory granules has been observed, but no correlation could be established with taxonomy, habitat, or tube foot morphology (Santos and Flammang, 2006; present study). Although we analyzed tube feet from at least three animals with different test size per species, no influence of age on secretory granule ultrastructure was found either. However, accurate estimation of sea urchin age is still a subject that remains open to discussion (Russell and Meredith, 2000; Narvaez et al., 2016). The present study revealed conservation of Nectin-like proteins among the eleven studied species, but a significant variation of the glycan residues that compose their footprints. However, taxonomically closer species, like P. lividus and S. granularis, seem to possess putative adhesive glycoproteins with similar molecular weights (35 and > 100kDa) and glycans (chitobiose - disaccharides of β-1,4-linked glucosamine units) although with a different number of repetitive oligomers. Nectin, via its discoidin domains, can bind the N-acetylglucosamine carbohydrate moieties (Costa et al., 2010) present in the adhesive glycoproteins, contributing to connect the disc epidermis to the adhesive secretion and thus increasing the cohesion of this interface. The adhesive footprint interspecific glycan variability might also have implications for the enzymatic de-adhesion of temporary attaching marine animals (Lengerer and Ladurner, 2018). Indeed, proteases and glycosidases have been detected in the footprint proteome of the sea star A. rubens (Hennebert et al., 2015), and are also highly over-expressed in sea urchin P. lividus adhesive discs relatively to non-adhesive stems (Lebesgue et al., 2016). Provided that a de-adhesive enzyme-based secretion would cleave the bond between the tube foot cuticle and the adhesive material (Lengerer and Ladurner, 2018), it should be specific to the protein and glycan composition of each species. This hypothesis should be investigated in future studies.

Finally, our findings support data reported on sea stars and barnacles showing that in large structural proteins (like Sfp1 in sea stars, cp-100 k in barnacles and Nectin in sea urchin), the selection pressure is high for the conservation of functional domains (He et al., 2018; Lengerer et al., 2019). The same authors suggested that in small surface-binding proteins, the relative amino acid composition is more variable, being potentially more influenced by adaptations to the habitat and mode of living. More genomic data and tube foot-specific transcriptomes would be required for sea urchins to allow further comparisons of full-length protein sequences. This study shows that post-translational modifications like glycosylation must be taken in the equation since we found evidence of large variation in terms of the conjugated glycans, but with indications of taxonomy-related conservation.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

RJ and RL collected the specimens. LG, PF, PR, and NN performed the experiments. LG and RS performed data and statistical analysis. RS, NN, JM, and JC-C contributed to the conception and design of the study. RS and PF wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.