Direct feeding on chitinous siboglinid worm tubes by an osteopeltid limpet (Vetigastropoda: Lepetelloidea)

Diet switching and subsequent specialization are powerful evolutionary processes that can open up previously underexploited ecological opportunities, like the giant panda evolving from a carnivorous ancestor to the modern species feeding almost exclusively on bamboo. The deep ocean is generally a nutrient-poor environment, leading to many animals having narrow trophic niches associated with unusual habitats or substrate types. Chemosynthesis-based ecosystems such as hydrothermal vents and hydrocarbon seeps are environments that require exceptional adaptations to inhabit (Levin, 2005), while organic substrates fallen from pelagic zones represent a different kind of 'oasis' on the seafloor (McClain et al., 2025; Smith & Baco, 2003). These organic falls range from wood and plant debris like seagrass blades to animal remains such as whale carcasses and cephalopod beaks, and even locally produced material like shark and skate egg cases (Marshall, 1987; Marshall & Walton, 2021). These hard parts made from proteins like keratin or collagen and polysaccharides such as cellulose and chitin are challenging to break down but represent a significant reserve of carbon and energy in the deep sea that may remain unrecycled if not for animals specializing to feed on them. A prime example of animals that have evolved to utilize these rare parcels of energy in the deep are the “cocculiniform” limpets, with different lineages occurring on specific types of organic substrates. Cocculiniforms were once thought to be a monophyletic group but were later revealed to be diphyletic, containing morphologically convergent members from superfamilies Cocculinoidea and Lepetelloidea (Haszprunar et al., 2022). For example, many cocculinids live on sunken wood, addisoniid limpets typically occur on elasmobranch egg cases, cocculinellids on fish bones, some pseudococculinids are restricted to decaying algae and stone crab carapaces, bathysciadiid limpets live on octopus and squid beaks, lepetellids feed on tubes of the onuphid polychaete genus Hyalinoecia, and osteopeltid limpets are known from whale carcasses including one genus restricted to baleens (Haszprunar, 1998; Judge & Haszprunar, 2014; Marshall, 1987). In most cases, these limpets are thought to have symbiotic bacteria inside their gut to help digest these hard materials (Marshall & Walton, 2021). Many annelid worms produce chitinous tubes. Tubeworms in the annelid family Siboglinidae (also known as “vestimentiferans”) produce chemically stable and very robust tubes composed of β-chitin microfibrils arranged in parallel bundles embedded in a proteinaceous matrix (Ogawa et al., 2013). These worms are abundant at hot vents and cold seeps and are important ecosystem engineers that greatly increase habitat complexity, acting as habitats for other animals (Hilário et al., 2011). Despite a number of limpets known to live on their tubes and graze on bacterial mats (Chen et al., 2019), no animals have been reported to feed directly on the tubes themselves—which are a highly nutritious albeit very difficult to digest. During our survey of a cold seep on the Daini Tenryu Knoll (Kobayashi et al., 2015; Chen, Watanabe, et al., 2025), Nankai Trough, Japan (34°4.5144′ N, 137°47.3103′ E, 609 m deep) by the deep-submergence vehicle (DSV) Shinkai 6500 (dive no. 1865; June 21, 2025; R/V Yokosuka cruise YK25-09C), we discovered an unusual cocculiniform limpet living on the siboglinid tubeworm Lamellibrachia columna (Figure 1A). There appeared to be no preference to newer or older parts of the tube. Where the limpets attached, the underlying tube wall was visibly thinned and often only a translucent membrane remained between the limpet and the worm's inner cavity (Figure 1B). As siboglinids are able to secrete and repair the tube from the inside (Hilário et al., 2011), the translucent part could be newly secreted to prevent penetration. In cases where a single limpet was attached in isolation, the attachment site showed concentric ring-like feeding marks, indicating gradual penetration through β-chitin sheets that make up the Lamellibrachia tube (Figure 1C). These point to active feeding upon the substrate itself, and not just microbial biofilms on the surface—given the otherwise pristine condition of tubes lacking limpets, we deduce that the limpets are uniquely responsible for grazing and consuming the tubes. We identified this tube-eating limpet (Figure 1D) as belonging to an undescribed genus of Osteopeltidae, and molecular barcode of the mitochondrial COI (cytochrome c oxidase subunit I) gene obtained following published methods (Ogura et al., 2018) was congruent with this placement (sequence deposited in GenBank under accession number PX251951). The two known osteopeltid genera both occur on whale carcasses, with Osteopelta found on bones and presumably feeding on bacterial mats and Baleenopelta restricted to baleens (Lesicki, 1998). As neither of these are chitinous food sources, it appears that this tube-eating limpet has newly acquired either endogenous enzymatic machinery or symbiotic microbes capable of breaking down chitin. Since molluscs are capable of producing and re-absorbing chitinous structures themselves (e.g., periostracum, operculum), they may also simply be able to repurpose chitinase genes already in their genomes for digestion. The alimentary systems of cocculiniform limpets are known to be convergently modified according to their food preferences (Marshall & Walton, 2021). Those feeding on polysaccharides like chitin and cellulose tend to possess large stomach volumes with expanded esophageal glands (e.g., bathysciadiids eating cephalopod beaks and bathyphytophilids ingesting seagrass rhizomes), while those relying on fibrous proteins having reduced stomach volumes and instead possess hypertrophied intestinal sacs (e.g., cocculinellids feeding on fish bone and addisoniids living on elasmobranch egg cases) (Haszprunar, 1987; Lesicki, 1998). In order to examine anatomical adaptations of the tube-eating limpet, we carried out a 3D reconstruction and volumetric calculations of the digestive system from micro-computed tomography (μCT) data obtained using a commercial μCT scanner (ScanXmate-D160TSS105, Comscantecno; 992 projections of 986 × 780 pixels each at a resolution of 5.922 μm), following published methods (Chen et al., 2023). Indeed, our reconstruction (Figure 2) revealed a sizeable stomach occupying 5.35% of the body volume and a pair of enlarged esophageal gland taking up 2.22% of the body volume. Compared to published anatomical description of genus Osteopelta (Haszprunar, 1988), the stomach of the tube-eating limpet appears to be enlarged and the intestine is substantially longer. These anatomical features also point to this limpet obtaining nourishment from polysaccharides like chitinous worm tubes. Unfortunately, no anatomical data are available from the confamilial and keratin-feeding Baleenopelta for comparison (Marshall & Walton, 2021). On the other hand, lepetellid limpets that feed on onuphid worm tubes have entirely different modifications such as a reduced stomach and multiple digestive glands (Judge & Haszprunar, 2014). Onuphid tubes are made from onuphid acid which is a sugar-phosphate polymer, rather than chitin as in siboglinids (Judge & Haszprunar, 2014), and may require different anatomical modifications for digestion. We do note, however, that true limpets (Patellogastropoda) which are grazers or detritivores on regular non-organic substrates also exhibit very small stomachs and enlarged intestines (Ponder et al., 2020). Similar modifications in lepetellids may or may not be directly associated with onuphic acid feeding. Our discovery of a cocculiniform limpet feeding on siboglinid worm tubes expands the broad spectrum of exotic diets used by these gastropods (Judge & Haszprunar, 2014; Marshall & Walton, 2021). The family Osteopeltidae now consists of three groups each feeding on distinct diets (bacterial mat on whale bones for Osteopelta, fibrous proteins for Baleenopelta, and polysaccharides for the tube-eating limpet), making it a suitable model system for a more detailed investigation of how cocculiniform limpets actually metabolize their unusual food. Given the broad, global occurrences of cold seeps (Levin, 2005) and siboglinid tubeworms (Hilário et al., 2011), tube-eating limpets may play an important role in nutrient cycling of deep-sea ecosystems. These limpets are likely eaten by decapod crustaceans and fish, which would pass the chitin-derived carbon up the food chain. As L. columna tubeworms from seeps have lighter δ15N values than photosynthetic food sources (Nomaki et al., 2024), stable nitrogen isotope ratios may prove to be useful markers in tracking the contribution of tube-feeding. Other limpets from the same gastropod order as osteopeltids such as Pyropelta and Lepetodrilus commonly occur on chitinous substrates such as siboglinid worm tubes and periostracum of provannid snails or mussels (Chen et al., 2025). It is possible that they also partially gain nourishment from chitin-feeding. Our finding highlights the remarkable adaptations of cocculiniform limpets in utilizing apparently indigestible food sources (Haszprunar, 1998), and suggests the impacts of chitin-feeding may be widespread on the deep seafloor. Chong Chen conceived and designed the project. Chong Chen and Hiromi Kayama Watanabe collected and sorted specimens at sea. Chong Chen carried out 3D reconstructions and made the figures. Hiromi Kayama Watanabe obtained the molecular data. Chong Chen drafted the manuscript, which was critically revised by Hiromi Kayama Watanabe. We thank the captain, crew, and the scientific party on-board R/V Yokosuka during the JAMSTEC × Ocean Census collaborative “SHINKAI” expedition (research cruise YK25-09C). We extend our gratitude to the HOV Shinkai 6500 team for their excellent support of our scientific activities. Research cruise YK25-09C on-board R/V Yokosuka was co-funded by JAMSTEC and the Nippon Foundation – Nekton Ocean Census (https://oceancensus.org/) through a Cooperative Research Agreement. We thank Ocean Census for supporting and accelerating species discovery in this work. Akinori Yabuki (JAMSTEC) is gratefully acknowledged for leading the cruise YK25-09C, and Michelle Taylor (Nippon Foundation – Nekton Ocean Census) for leading the Nippon Foundation – Nekton Ocean Census team on-board. Katsunori Kimoto (JAMSTEC) kindly provided assistance with the μCT scanning process, and Hidetaka Nomaki (JAMSTEC) is acknowledged for useful discussion on stable isotopes. We appreciate constructive comments from Lisa Levin (Scripps Institution of Oceanography) and an anonymous reviewer, which improved earlier versions of this paper. The authors declare no conflicts of interest. The μCT image stack (Chen & Watanabe, 2025) is available in Figshare at https://doi.org/10.6084/m9.figshare.30025459. The COI-sequence is available in the National Center for Biotechnology Information (NCBI) GenBank under accession number PX251951 at https://www.ncbi.nlm.nih.gov/nuccore/PX251951.1. The specimens examined have been deposited in the National Museum of Science and Nature, Tsukuba, Japan (NSMT-Mo 82424-82425) and the Senckenberg Research Institute and Natural History Museum Frankfurt, Frankfurt, Germany (SMF 384716-384717).