1 Introduction

Deep-sea benthic ecosystems mostly rely on organic matter derived from the ocean surface in the form of phytodetritus or ‘marine snow’ showering down to the deep for energy, due to a lack of photosynthetic primary production (Gage and Tyler 1991). Chemosynthetic communities on the deep-sea floor, such as those associated with hydrothermal vents or hydrocarbon seeps, are exceptions where primary production is enabled by microbes utilizing chemical energy from the oxidation of molecules contained in the geofluid (e.g., Fisher et al. 1989; Markert et al. 2007) to produce organic matter (OM). Many animals endemic to these ecosystems and living in close proximity to geofluid flows are symbiotic with such bacteria and depend on them for nutrition (Dubilier et al. 2008). Due to the high productivity of OM in these chemosynthetic communities, they tend to exhibit total biomasses three to four orders of magnitude higher than the surrounding deep-sea floor sustained by phytodetritus (Levin 2005; Van Dover 2000).

The Japanese archipelago is located at boundaries of four plates including the Philippine Sea plate, the Pacific plate, the Eurasian (Amurian) plate, and the North American (Okhotsk) plate. Subduction of the Philippine plate and the Pacific plate beneath the Eurasian plate and the North American Plate, respectively, produce pore fluid seepage to the seafloor in the subducted accretionary prism (e.g., Boulègue et al. 1987, Gamo et al. 1992). These pore fluids contain high concentrations of hydrocarbons, such as methane, either of microbial origin (Toki et al. 2012; Riedinger et al. 2015) or thermogenic origin (Toki et al. 2014). There are numerous chemosynthetic communities around Japan associated with these hydrocarbon seeps from bathyal to hadal depths, especially in Japan Trench and Nankai Trough (Nakajima et al. 2014). These seeps are mainly dominated by vesicomyid clams and siboglinid tubeworms, but the species composition at each site is influenced by factors such as water depth and geographic location (Fujikura et al. 2000; Nakajima et al. 2014). Vesicomyid clams in the subfamily Pliocardiinae gain nutrition from thioautotrophic bacteria living in bacteriocytes within their gills (Arp et al. 1984; Krylova and Sahling 2010). The symbionts produce OM via the oxidation of hydrogen sulfide taken up by the clam’s foot into the blood, using oxygen from the ambient seawater which is drawn in from the siphon and incorporated directly in the gill (Childress and Girguis 2011; Ip et al. 2021). Siboglinid tubeworms possess a unique organ filled with symbiotic thioautotrophic bacteria called the trophosome and have completely lost the gut as adults (Cavanaugh et al. 1981; Nussbaumer et al. 2006). Depending on the genus, tubeworms either conduct all gas exchanges through the plume (gill) or take up oxygen from the plume but sulfides from an extended posterior ‘root’ system (Childress and Girguis 2011; Dattagupta et al. 2006).

While energy sources, i.e., the coupling of reducing and oxidizing compounds in the pore fluid and bottom seawater, respectively, for chemoautotrophic carbon production has been a focus of study in chemosynthetic ecosystems, their carbon and nitrogen sources have rarely been investigated in detail. Symbiotic vesicomyid clams are thought to incorporate ammonium or nitrate for chemosynthesis from ambient seawater bathing their gills, which diffuses directly into the symbionts (Childress and Girguis 2011). The symbionts also mostly take up dissolved inorganic carbon (DIC) directly in the same way (Kochevar and Childress 1996), though the possibility that the host clam assists this by taking up DIC through the foot and transporting this through blood circulation to the symbionts has been suggested (Childress and Girguis 2011). For siboglinid tubeworms, some (e.g., Riftia and Ridgeia) are thought to take up DIC and ammonium/nitrate exclusively from the plume (Liao et al. 2014), while for others (e.g., Lamellibrachia) the capacity to do so from either the plume or ‘root’ remains little-studied. Stable isotope ratio of carbon (δ13C) can be used to estimate carbon sources of organisms (e.g., Nomaki et al. 2008, 2021). However, in the case of chemosynthetic communities, different pathways of chemosynthesis, such as reductive tricarboxylic acid cycle (rTCA), the Calvin–Benson–Bassham (CBB) cycle, and methane oxidation pathways (House et al. 2003; Pearson 2010), lead to varying δ13C values of OM due to the different degrees of isotope fractionation from carbon dioxide to biomass (c.a. 2–13‰ for rTCA, c.a. 5–26‰ for CBB) or from methane to biomass (16–31‰ for aerobic methanotrophs, and 0–40‰ for anaerobic methanotrophs, Pearson 2010). This makes the interpretation of vent and seep animals’ carbon sources challenging from analysis of δ13C alone.

Natural-abundance radiocarbon (Δ14C) has recently become an increasingly used tool for investigating carbon flows in marine ecosystems (Ishikawa et al. 2021). In deep-sea ecosystems, ingestion of freshly deposited organic matters derived from the photic zone (rich in 14C) could be confirmed by analyzing 14C concentrations in the gut contents, body tissue, plus surface sediments and sediment trap samples as potential food sources (Purinton et al. 2008). In chemosynthetic ecosystems such as hydrothermal vent ecosystems, Δ14C analyses can clearly distinguish OM derived from surface primary production (c.a. 40‰) from chemosynthetic OM (below −800‰)—(Nomaki et al. 2019, 2024; Williams et al. 1981). Chemoautotrophic microbes at vents exhibit depleted 14C concentrations because they utilize either 14C-depleted bottom water (c.a. −200‰ in the Pacific Ocean, Stuiver et al. 1983) or 14C-dead DIC from the vent fluid (−1000‰) as their inorganic carbon source (Williams et al. 1981). Hydrothermal vent animals living in close vicinity to vent orifices exhibit highly depleted Δ14C values while those at periphery habitat tended to show Δ14C values closer to that of the bottom water (Nomaki et al. 2024). An exception is that even at the periphery, Lamellibrachia tubeworms exhibited more depleted Δ14C values than co-occurring vesicomyid clams, suggesting that the carbon source of Lamellibrachia endosymbionts includes vent fluid DIC taken up via roots of their hosts deeply embedded in the sediments (Nomaki et al. 2024). Similar to hydrothermal vent areas, Δ14C values of pore fluid at cold seeps is also very low (down to −938‰: Masuzawa et al. 1995), making it possible to discriminate carbon sources for chemolithoautotrophic bacteria and their host animals between DIC in bottom water and pore fluid. To accurately interpret the results, however, it is important to make a clear distinction between these purely subduction-driven cold seeps (sensu Levin 2005) from ‘hydrothermal seeps’ fueled partly by hydrothermalism (Levin et al. 2012) or those supported by warm groundwater such as the Off Hatsushima seep in Sagami Bay, Japan (Tsunogai et al. 1996). In the present study, we focus on porewater-driven seeps and call them ‘cold seeps’ hereafter.

Here, we measured Δ14C, δ13C, and δ15N values of siboglinid tubeworms, vesicomyid clams, and polychaetes living inside their mantle cavities collected from four different cold seep sites around Japan. Their Δ14C values were compared with those of bottom water Δ14C values to evaluate DIC sources of the symbiotic chemoautotrophic bacteria in these animals. Both δ13C and δ15N values were further evaluated to infer metabolism and ecology of animals.

2 Materials and Methods

2.1 Animal sample collections

Animal samples for isotopic measurements were collected between 1989 to 2004 using various research vessels and submersibles (Table 1), including vesicomyid clams, polychaetes living inside their mantle cavities, and siboglinid tubeworms (for a list see Table 2 and Supplementary Table 1). We used samples from four different cold seep sites: landward slope of Japan Trench (39°06.4′ N, 143°53.6′ E, 5343–5345 m deep), Yukie Ridge in the Nankai Trough (33°50′ N, 137°54–55′ E, 1900–2200 m deep), Daisan Tenryu Submarine Canyon in the Nankai Trough (33°38.7′ N, 137°54.498′ E, 3851 m deep), and Kuroshima Knoll on the Ryukyu Arc (24°07–08′ N, 124°12–13′ E, 643–785 m deep) (Fig. 1). Animals were collected using either a suction sampler or a scoop sampler on the remotely operated vehicles (ROVs) Hyper-Dolphin and Kaiko or the human occupied vehicles (HOVs) Shinkai 2000, Shinkai 6500, and Nautile. Once recovered on-board, the animals were identified, dissected into different body parts in some cases, and then frozen in −80 °C freezer until use.

Table 1 Investigated cold seep sites and sampling details. Natural-radiocarbon abundances (D14C) of the bottom water were estimated from reported values
Full size table
Table 2 Cold seep animals measured in this study and their averaged isotopic compositions
Full size table
Fig. 1
figure 1

Map of the relevant cold seep sites (open yellow diamonds) and the site where natural-abundance radiocarbon (D14C) data of dissolved inorganic carbon in the seawater were taken from in this study (red lines: WOCE P09 lines, triangles: Ding et al. 2020, C5 and C8; reversed triangle: WOCE P10, 74 and 80)

Full size image

2.2 14C analyses of organisms

The frozen animal samples were freeze-dried and ground into fine powders. We placed 0.71–1.73 mg dry weight of samples in pre-combusted glass cups and then decalcified them with 0.1 M HCl, and completely dried again. The samples were graphitized by methods modified from Yokoyama et al. (2007, 2010). Briefly, dried samples were combusted in an evacuated quartz tube with copper oxide at 500 °C for 30 min and at 850 °C for 2 h. The CO2 gas was cryogenically purified in a vacuum line and reduced to graphite with hydrogen and an iron catalyst at 630 °C for 6 h. We measured the Δ14C values with a single-stage accelerator mass spectrometer (AMS) at the Atmosphere and Ocean Research Institute, the University of Tokyo (Chiba, Japan; AMS lab code YAUT; Yokoyama et al. 2019). Blank tests for sample preparations and processing found no contamination of modern carbon (Yamane et al. 2019; Yokoyama et al. 2022).

Radiocarbon data reported here are corrected for radioactive decay between 1950 and the year when the samples were measured. The equation used is as follows:

$${{\Delta }^{14}}{\text{C}} = \left\{ {\frac{{{{\left( {^{14}{\text{C}}{/^{12}}{\text{C}}} \right)}_{{\text{sample}}}}}}{{{{\left( {^{14}{\text{C}}{/^{12}}{\text{C}}} \right)}_{\text{OX - 1}}}{e^{\frac{y - 1950}{{8267}}}}}} - 1} \right\}$$
(1)

The Δ14C value of the international standard (Ox-1: oxalic acid-1) takes into account of radioactive decay since 1950 (Stuiver and Polach 1977).

Radiocarbon measurements were conducted with 6 to 9 sets of 6 min measurements. Numbers of measurements depended on the samples, since we aimed to collect more than 20,000 counts to achieve accurate measurements. We also took into account all possible mass fractionations by measuring δ13C values using the AMS instead of an offline isotope ratio mass spectrometer (IRMS). We used four international standards, namely IAEA-C1, IAEA-C3, IAEA-C6, and NIST SRM 4990C, with each standard having different radiocarbon concentrations. Our measured values were calibrated with the calibration curve obtained from the measurements of these standards.

2.3 Analyses of total organic carbon (TOC), total nitrogen (TN) contents, and their stable isotopic compositions

The freeze-dried and powdered samples (111 to 213 µg dry weight) were placed into pre-cleaned silver capsules. The samples were decalcified with 0.1 M HCl and dried on a hot plate at 60 °C, and the capsules were sealed using pre-cleaned forceps. Stable carbon and nitrogen isotopic composition were determined using an isotope ratio mass spectrometer coupled to an elemental analyzer (Flash EEA 1112-DELTA V Advantage ConFloIV System, Thermo Fisher Scientific). Isotope ratios were conventionally expressed as δ values in ‰ (Coplen 2011) relative to carbon and nitrogen international standards (Vienna PeeDee Belemnite and atmospheric N2, respectively). After measuring every 10 to 15 samples, isotopic compositions of internal working standards (Glycine: −33.8‰, Alanine: −19.6‰, Histidine: −10.7‰ for δ13C measurements, and Alanine: 1.58, 9.97, and 20.6‰ for δ15N measurements) were measured and analytical errors for δ13C and δ15N working standards were both <  ±0.15‰.

2.4 14C abundances of bottom water DIC

We estimated Δ14C values of ambient bottom water at each sampling site from previously reported values (Table 1). At both Yukie Ridge and Daisan Tenryu Submarine Canyon at Nankai Trough, the P09 line data obtained during the World Ocean Circulation Experiment (WOCE) sampling campaign in 1994, in close proximity to the animal collection sites (Fig. 1), were used to estimate Δ14C values at 1900 to 2200 m depth (−230 to −220‰: Yukie Ridge) and 3850 m (−230 to −210‰: Daisan Tenryu Submarine Canyon). The P10 stations data of the WOCE sampling campaign in 1993, which were hundreds of kms south from the landward slope of Japan Trench site, were used to estimate bottom water Δ14C values (Fig. 1). Despite the distance, since Δ14C values deeper than 3000 m were almost constant across latitudes along the P10 line (Key et al. 2002), we consider our estimated range (−209 to −199‰) to be not far off from ambient bottom water value at the time of the animal sampling event at Japan Trench.

At Kuroshima Knoll, we referred Δ14C values from published data obtained off Taiwan in 2019, which is 200 to 300 km south/southwest of the animal collection sites (C05 and C08 of Ding et al. 2020). The Kuroshio current is running through these stations of Ding et al. (2020) and also around Kuroshima Knoll, thus it is reasonable to postulate that those stations exhibit similar Δ14C profiles to those at Kuroshima Knoll. As Δ14C profiles of these stations (Ding et al. 2020) as well as other areas where Kuroshio current runs through (Lan et al. 2023) showed sharp decreases in Δ14C values from 500 to 1000 m deep where our sampling stations fall into (643 and 785 m), a small change in the water mass may cause some 10 s of ‰ of Δ14C values in this case. Also, the atmospheric 14C concentrations in northern hemisphere is decreasing with time due to the anthropogenic emission of fossil fuel (Graven et al. 2022), from around 80 to 90‰ during 1997 to 2004 when animals were collected at Kuroshima Knoll to almost 0‰ in 2019 when the Δ14C profiles in Ding et al. (2020) were obtained. The changes in surface water DIC Δ14C values have been reported off California, from 35‰ in 2004 to −6‰ in 2022, where Δ14C values of atmospheric CO2 decreased from 60‰ to −5‰ during the same period (Hauksson et al. 2023). We note that the surface water DIC Δ14C values near Kuroshima Knoll may also have changed to a similar extent, though the subsurface DIC Δ14C values at depths over 500 m are expected to be more stable as shown at Nankai Trough area between 1994 and 2022 (Lan et al. 2024). As such, our estimates of the Δ14C values of the ambient bottom water at Kuroshima Knoll must be treated with caution.

Based on the measured Δ14C values of animals (Table 2 and Supplementary Table 1) and above mentioned estimations of bottom water DIC Δ14C values, we calculated the contributions of seep fluid DIC to the chemoautotrophic productions in the endosymbionts of each animal using the following equation:

$${\text{Contribution of geofluid C}} = {{\left( {{\Delta^{14}}{{\text{C}}_{{\text{animal}}}}--{\Delta^{14}}{{\text{C}}_{{\text{bw}}}}} \right)} \mathord{\left/ {\vphantom {{\left( {{\Delta^{14}}{{\text{C}}_{{\text{animal}}}}--{\Delta^{14}}{{\text{C}}_{{\text{bw}}}}} \right)} {\left( {{\Delta^{14}}{{\text{C}}_{{\text{sf}}}}--{\Delta^{14}}{{\text{C}}_{{\text{bw}}}}} \right)}}} \right. \kern-0pt} {\left( {{\Delta^{14}}{{\text{C}}_{{\text{sf}}}}--{\Delta^{14}}{{\text{C}}_{{\text{bw}}}}} \right)}}$$
(2)

where Δ14C animal is the averaged Δ14C value for each animal species (Table 2), Δ14Cbw is a median of the estimated bottom water DIC value range (Table 2), and Δ14Csf is −1000‰ assuming DIC in seepage fluid contained DIC originated from old (i.e., older than 50,000 years) organic matter preserved in the sediments. Although reported Δ14C values of pore fluid DIC at seep sites are limited, Masuzawa et al. (1995) found that the seep fluid collected in vesicomyid clam colonies indeed exhibited decreasing Δ14C profiles with increasing sediment depths and showed highly depleted values (−938 ± 20‰ and −938 ± 4‰) at 18 cm and 36 cm depths in the sediments.

3 Results

3.1 Natural-abundance radiocarbon

Natural-abundance radiocarbon concentrations (Δ14C) of vesicomyid clams ranged from −275 ± 9‰ (Abyssogena phaseoliformis at landward slope of Japan Trench) to −96 ± 2‰ (Akebiconcha kawamurai at Kuroshima Knoll), and varying with the sampling locality (Fig. 2). Clams from the same sampling locality exhibited almost identical Δ14C values: −247 ± 3‰ (Calyptogena fausta), −245 ± 3‰ (Archivesica magnocultellus), and −237 ± 6‰ (Archivesica tsubasa) at Yukie Ridge; −245 ± 1‰ (Archivesica tsubasa), −243 ± 5‰ (Ectenagena laubieri), and −242 ± 5‰ (Abyssogena kaikoi) at Daisan Tenryu Canyon, which were 10 to 20‰ (Yukie Ridge) and ca. 30 ‰ (Daisan Tenryu Canyon) lower than ambient bottom water values (Fig. 2 Table 2). At Kuroshima Knoll, Akebiconcha kawamurai collected at 643 m water depth exhibited higher Δ14C values (−96 ± 2‰) than Pliocardia kuroshimana collected at 785 m water depth (−169 ± 2‰). Those values were mostly the same (Akebiconcha kawamurai at Kuroshima Knoll) to ca. 30‰ lower than those of the estimated Δ14C values of ambient bottom water (Fig. 2., Table 1). Abyssogena phaseoliformis from Japan Trench exhibited even lowered Δ14C values, ca. 70‰ lower than ambient bottom water values.

Fig. 2
figure 2

Natural-abundance radiocarbon (D14C) of animals collected from cold seep sites around Japan. The gray bars indicate the estimated range of natural-abundance radiocarbon (Δ14C) values of bottom water dissolved inorganic carbon

Full size image

Nautiliniella calyptogenicola, a calamyzine (Chrysopetalidae: Calamyzinae) annelid worm living inside the mantle cavity of Abyssogena phaseoliformis (Miura and Laubier 1989), exhibited Δ14C values of −271 ± 1‰, which was almost identical to the host clam values (−275 ± 9‰). For the siboglinid tubeworm Lamellibrachia columna (‘Lamellibrachia sp. L2’ of Kojima et al. (2001), shown to be conspecific with L. columna in McCowin et al. (2019)) exhibited greatly varying Δ14C values among individuals, from −548 to −329‰ (average −424 ± 101‰), corresponding on average 200‰ lower than those of ambient bottom water (Table 2).

3.2 Stable carbon and nitrogen isotopic compositions

Stable carbon isotopic compositions (δ13C) of cold seep vesicomyid calms differed slightly among sites. At Yukie Ridge and Daisan Tenryu Submarine Canyon in the Nankai Trough, the δ13C values of clams were between −39.4 ± 0.2‰ and −36.9 ± 0.2‰ (Fig. 3). The δ13C value of Abyssogena phaseoliformis collected from the landward slope of Japan Trench (−36.1 ± 0.2‰) were slightly higher than those of the Nankai Trough clams. At Kuroshima Knoll, Akebiconcha kawamurai (−34.1 ± 0.6‰) and Pliocardia kuroshimana (−32.5 ± 0.4‰) displayed even higher δ13C values.

Fig. 3
figure 3

Plots showing A stable carbon isotopic compositions (δ13C) versus stable nitrogen isotopic compositions (δ15N), B δ15N versus natural-abundance radiocarbon (Δ14C), and C δ13C versus Δ14C, of animals collected from cold seep sites around Japan.

Full size image

The δ13C value of the calamyzine polychaete Nautiliniella calyptogenicola (−36.3 ± 0.3‰) was almost identical to its host clam Abyssogena phaseoliformis (−36.1 ± 0.2‰). The δ13C values of the tubeworm Lamellibrachia columna varied greatly among individuals (−44.0 to −23.7‰). These trends in δ13C values were similar to those observed in Δ14C values.

Stable nitrogen isotopic compositions (δ15N) varied greatly across species, even from within the same cold seep site (Fig. 3). At Yukie Ridge, Archivesica tsubasa exhibited very low δ15N values (−6.2 ± 2.1‰), ~ 10‰ lower than those of Archivesica magnocultellus (4.2 ± 1.8‰). At Daisan Tenryu Canyon, the δ15N values were highly variable across individuals (−5.2 to 1.8), although the interspecific variations were small (Ectenagena laubieri: −2.2 ± 2.0‰, Archivesica tsubasa: −2.1 ± 2.8‰, Abyssogena kaikoi: −2.0 ± 2.8‰). Intraspecific variations were generally low for Abyssogena phaseoliformis (−4.7 ± 0.6‰) at the landward slope of Japan Trench, as well as Akebiconcha kawamurai (6.8 ± 1.1‰) and Pliocardia kuroshimana (7.4 ± 0.3‰) from Kuroshima Knoll.

The δ15N value of the calamyzine worm Nautiliniella calyptogenicola was −1.8 ± 0.6‰, ~3 ‰ higher than its host clam Abyssogena phaseoliformis. The siboglinid tubeworm Lamellibrachia columna exhibited intermediate δ15N values (−0.4 ± 1.0‰) among the measured cold seep animals, again with a relatively small intraspecific variation.

4 Discussion

4.1 C sources of hydrocarbon seep clams revealed by Δ14C values

The natural-abundance radiocarbon concentrations in the cold seep vesicomyid clams were generally similar or slightly lower than those of bottom water concentrations. Symbiotic vesicomyid clams burrow into the sediment so they can reach the sulfide-rich layer with their foot to take up hydrogen sulfide, which they then circulate to the gill endosymbionts via the hemolymph (Zal et al. 2000). As their symbionts inhabit cells on the gill surface in contact with bottom seawater in the mantle cavity drawn in through the siphon, the symbionts can take up oxygen and CO2 directly with the aid of carbonic anhydrases (Kochevar and Childress 1996). Our results in general agree with this, given the similar Δ14C values between animals and ambient bottom water Δ14C values. However, there were minor discrepancies between animal and bottom water Δ14C values except in Akebiconcha kawamurai, typically 20 to 30‰ and up to 70‰ which was observed in Abyssogena phaseoliformis (Fig. 2). Those small discrepancies can be attributed to the uptake of geofluid-originated CO2 which is depleted in 14C (Masuzawa et al. 1995), either from their foot or from their siphon. Since vesicomyid clams require oxygen for both respiration and oxidation of hydrogen sulfide, the water they circulate to the gill and mantle cavity is oxygen-rich ambient bottom water and unlikely to contain a high proportion of geofluid containing 14C-depleted CO2 and other reduced compounds. Therefore, it seems likely that vesicomyid clam take up some DIC through the foot and transport it through blood circulation to the symbionts, a possibility that has been suggested previously (Childress and Girguis 2011). The largest discrepancy between clam and ambient bottom water Δ14C values was found in Abyssogena phaseoliformis collected from the landward side of the Japan Trench (Fig. 2). This may be caused by a greater dependency on foot-derived CO2 or higher CO2 fluxes through subsurface fluids at this site. Further geochemical studies across different seepage sites will shed more light on these discrepancies of the dependency on geofluid derived CO2 among sites and species.

4.2 Variations in δ15N values and potential N sources of vesicomyid clams

Some vesicomyid clam species, namely Archivesica tsubasa, Abyssogena kaikoi, Archivesica magnocultellus, and Ectenagena laubieri exhibited large (4.0 to 6.4‰) variations in their δ15N values among individuals, whereas Abyssogena phaseoliformis, Calyptogena fausta, and Pliocardia kuroshimana exhibited similar δ15N values across individuals (differences of 0.4 to 1.3‰) (Fig. 3). Animals endemic to deep-sea chemosynthetic ecosystems have been reported to exhibit a wide variety of δ15N values, potentially reflecting different inorganic nitrogen sources assimilated by chemoautotrophic bacteria (Van Dover 2002). It has also been reported that the isotope effects between microbial biomass and inorganic N source tended to be large when environmental inorganic N sources were high (Liao et al. 2014; Sigman and Fripiat 2019). At a hydrothermal vent field in the Okinawa Trough, endemic animals living in close proximity to the vent orifice tended to have lower δ15N values where the NH4+ concentration was high (Nomaki et al. 2024). As pliocardiine vesicomyid clams rely almost entirely on their endosymbiotic microbes for nutrition (Ip et al. 2021), δ15N values of the host clams are related to those of their symbionts (He et al. 2023; Ishikawa et al. 2024), which is in turn affected by the NO3− or NH4+ concentrations in the environment.

Previously, inorganic nitrogen sources for vesicomyid clam symbionts such as NH4+ and NO3− were thought to be drawn in through the gill from bottom water like O2 and CO2 (Childress and Girguis 2011). Since NH4+ concentrations in the bottom water at vesicomyid clam colonies were very low (Ohnishi et al. 2018, E. Tasumi, unpublished data) as is typical for the deep-sea bottom water, the available inorganic nitrogen sources for vesicomyid clam in the bottom water would appear to be NO3−. The high variability of δ15N values seen in some species of vesicomyids analyzed herein (Fig. 4), however, can be explained by the hypothesis that their inorganic nitrogen source is not solely the bottom water but also the sediment porewater.

Fig. 4
figure 4

Stable nitrogen isotopic compositions (δ15N) of vesicomyid clams and their maximum sizes (shell length), reflecting their maximum burrowing depth in sediments

Full size image

It is notable that the two largest (Fujikura et al. 2000; Fujikura et al. 2008) species (Archivesica magnocultellus, Archivesica tsubasa) exhibited large intraspecific δ15N variations, whereas the two smallest species (Calyptogena fausta, Pliocardia kuroshimana) exhibited small δ15N intraspecific variations (Fig. 4). Vesicomyid clam species with large body size and shell length can burrow deeper into the sediment, where the NH4+ concentration is higher (e.g., Toki et al. 2014). The NH4+ concentration in the sediment at Oomine Ridge, Nankai Trough, which is less than 100 km west from Yukie Ridge and Daisan Tenryu Canyon, showed sharp increase from sediment surface (0 to 17 µM) to 5 or 10 cm depths in sediments (up to 300 µM) (Toki et al. 2014). In the case of small individuals of the same species, however, they can likely only reach the shallower part of the sediment (e.g., shallower than 5 cm), where NH4+ concentration is low. Since the availability of NH4+ induces different isotopic fractionation (Sigman and Fripiat 2019), the differences in δ15N values of the same species of vesicomyid clam potentially reflect their burrowing depths. Two size factors may be at play here, including the maximum body size which dictates how deep a clam species can reach (i.e., the range of NH4+ concentrations), and the individual body size of the specimen analyzed (i.e., how deep the exact individual burrowed). In the case of a large species, while large individuals can burrow deeper into sediments, smaller individuals would reside at relatively shallower depths, and thus measuring δ15N values across a size range may lead to a large intraspecific variation. The possible range of burrowing depth depends on the maximum body size of the species (the larger the species, the larger the difference between juveniles and adults). This may be one reason contributing to the overall pattern seen where large-bodied species tend to exhibit larger variability. Conversely, two smallest species can only burrow in a narrow depth range in the sediment across growth, resulting distributions at similar lower NH4+ concentrations.

The apparent outlier in our data was Abyssogena phaseoliformis which has a large maximum shell length (202 mm: Okutani et al. 2009) but showing low intraspecific variability in δ15N values. This was also the only species for which shell lengths of the analyzed specimens was available (93.8 to 121.0 mm; Supplementary Table 1). There appeared to be no clear trend in this species between shell length and δ15N values, though we note the size range of these specimens was small.

It was unfortunately not possible to measure individual sizes of other species used in this study (only pieces of dissected tissue were available in most cases), the relationships between body size (availability of inorganic nitrogen source in sediments) and their δ15N values would be a hypothesis warranting future study. Measuring the shell length of individuals clams and their burrowing depth in situ is now within reach using acoustic imaging devices (Mizuno et al. 2022). Investigating the δ15N values of clams with such detailed in situ data, together with measuring pore water NH4+ profiles in vesicomyid clam colony, will be able to provide key evidences required to test this hypothesis.

4.3 Ecology of the calamyzine polychaete and the siboglinid tubeworm

The calamyzine worm Nautiliniella calyptogenicola was first reported from the mantle cavity of the vesicomyid clam Abyssogena phaseoliformis (Miura and Laubier 1989), and recently also discovered from Abyssogena mariana, implying its capacity to colonize multiple hosts in the genus Abyssogena (Jimi et al. 2022). Although assumed to be a parasite (Miura and Laubier 1989), its actual relationships with the host have not been closely examined to date. Calamyzine worms are either free-living or symbiotic/parasitic with molluscs or other annelids (Rouse et al. 2022). Using worms collected from Abyssogena phaseoliformis in Japan Trench, we found that N. calyptogenicola exhibited almost identical Δ14C and δ13C values with the host clam, while its δ15N value was 2.9‰ higher the host, on average. The observed ~3.0‰ increase in δ15N value relative to the host clam agrees well with the trophic level increment between consumer and prey organisms (Minagawa and Wada 1984). Together with the identical Δ14C and δ13C values between them, these serve as strong evidence that Nautiliniella calyptogenicola is indeed parasitic on its vesicomyid host.

The siboglinid tubeworm Lamellibrachia columna was the only animal exhibiting clearly 14C-depleted tissue relative to the bottom water, indicating OM production using DIC from the seeping fluid. Unlike the giant tubeworm Riftia pachyptila for which physiology and metabolism is best known among siboglinid tubeworms (Childress and Girguis 2011), how exactly Lamellibrachia incorporates DIC has remained unclear. As tubeworms host symbionts in the completely internal trophosome, the host animal must supply all resources required by the endosymbiont via blood. Rifita and some other genera carry out all gas exchange exclusively across the plume (Childress and Fisher 1992), including taking in CO2 from the plume and transporting it to the symbionts (Williams et al. 1981). Lamellibrachia, however, is capable of gas exchange at both plume and the posteriorly extended ‘root’ which reaches deep into the sediment or fissures to take up sulfides (Cordes et al. 2005; Dattagupta et al. 2006; Li et al. 2019). Recently, a study of Lamellibrachia columna (Lamellibrachia sp. L1 in Kojima (2000)) at a hydrothermal vent in the Okinawa Trough using Δ14C revealed much depleted 14C values compared to the bottom water, suggesting the tubeworm also utilizes DIC from the vent fluid deep inside the sediment (Nomaki et al. 2024). Here, we found cold seep individuals of L. columna displayed a very variable range of Δ14C values (− 548 to − 329‰). This could indicate one of two scenarios: one where the tubeworm flexibly takes in CO2 from both the plume and the root, depending on the environmental condition at its microhabitat; and another where it only uses CO2 diffused in from the root but different intrusion depth of the root leads to variable porewater contribution. Since porewater rises from below, a shallower extension into the sediment would mean more bottom water at that sediment layer, leading to less depleted Δ14C values. This latter scenario is more consistent with previous results from hydrothermal vents where little variability was seen (Nomaki et al. 2024), suggesting that root intrusion (and potentially body size) has a strong effect on how much carbon from seep porewater is fixed by the symbionts. As for vesicomyid clams, body size and in situ distributions in the sediments would provide detailed insights on the controlling factors of carbon and nitrogen isotopic compositions, which reflect the uptake mechanisms.

5 Conclusions

We applied the analyses of natural-abundance radiocarbon to cold seep ecosystems for the first time, in order to elucidate the carbon sources of animals inhabiting these chemosynthetic ecosystems in terms of geofluid vs ambient bottom water. Previously, the stable isotope ratio of carbon (δ13C) has been widely used to estimate carbon sources of seep animals, but differing pathways of chemosynthesis with varying degrees of isotope fractionation made these inconclusive. Based on the measurements of natural radiocarbon (14C) abundances, we found a minor contribution (0 to 9%) of carbon from geofluids in vesicomyid clams, previously thought to only use carbon from the ambient seawater. Furthermore, our stable nitrogen isotope (δ15N) analyses suggested that the nitrogen source of vesicomyid clams also contains the geofluid, which further pointed toward to a novel hypothesis where the δ15N values may reflect burrowing depths with varying NH4+ concentrations. In the siboglinid tubeworm Lamellibrachia we also found a strong carbon contribution from the geofluid but with high variability among individuals, likely reflecting flexibilities in uptake strategies or different intrusion depths of the ‘root’. Cold seeps are widespread systems in the ocean floor that play key roles in the elemental and energy flux between sediment and seawater. Our results show that radiocarbon is a useful probe for pinpointing nutritional sources for seep organisms, opening a new and additional route to understanding these fluxes. These promising results from radiocarbon analyses follow previous successes at hydrothermal vent ecosystems (Nomaki et al. 2019, 2024), and further confirm its usefulness across different chemosynthetic systems.