A multi-isotope study of deep-sea mussels at three different hydrothermal vent sites in the northwestern Pacific

Abstract

To investigate symbiotic bacterial ecosystems at different deep-sea hydrothermal systems in the northwestern Pacific, compound-specific carbon and hydrogen isotope analyses of lipid biomarkers have been performed in addition to bulk C, H, N, S and O isotope analyses on Bathymodiolus mussels from the Hatoma seamount (B. platifrons), the Daiyon-Yonaguni (Yonaguni) knoll (B. platifrons), and the Suiyo seamount (B. septemdierum). The two B. platifrons contain large amounts of diploptene, while no hopanoid is detected in B. septemdierum, suggesting that B. platifrons and B. septemdierum harbors methanotrophic and thiotrophic bacteria, respectively. In spite of the same symbiont, the large bulk δ¹³C difference between the Hatoma (− 44.8‰) and Yonaguni (− 24.5‰) mussels reflects isotopically distinct hydrothermal CH₄ (Hatoma: ~ − 48‰, Yonaguni: ~ −26‰) as a carbon source. Fatty acids of the Hatoma and Yonaguni mussels are more enriched in D (− 144 to − 101‰) than the Suiyo mussel (− 265 to − 162‰), suggesting that D-depleted magmatic water or D-enriched hydrogen derived from CH₄ could be a partial hydrogen source for methanotrophy or thiotrophy, respectively. Apparent positive correlations are observed between δ¹³C and δD of the bulk and biomarkers for each mussel due to similar biochemical processes during de novo synthesis. The compound-specific δ¹³C and δD variations have provided much information on not only distinct carbon and hydrogen sources but also the lipid synthesis with respect to different symbiotic bacterial ecosystems.

Introduction

Various faunal communities have been observed around hydrothermal vents on the deep-sea floor. The faunal communities depend primarily on organic matter produced by chemoautotrophs such as sulfur-oxidizing bacteria (thiotrophy), or methane-oxidizing bacteria (methanotrophy). To date, many isotope studies have been performed using bulk tissues of fauna harboring bacterial symbionts in order to elucidate their metabolism and ecology at the hydrothermal vents (e.g. Kennicutt et al., 1992, McKiness et al., 2005). The bulk isotope studies provide valuable information on the source material utilized and kinetic fractionations as well as food webs (e.g. Conway et al., 1994). Also the bulk isotope study sometimes has an overprinting signal not only of fauna and bacteria but also from both thiotrophy and methanotrophy (Pond et al., 1998, Yamanaka et al., 2003).

A molecular biomarker study is one approach to deconvolute the bacterial community, using characteristic fatty acids and cyclic isoprenoids as biomarkers (e.g. Ben-Mlih et al., 1992, Phleger et al., 2005). In particular, carbon isotope analysis of individual biomarkers is a useful means to infer a specific source of bacteria (e.g. Freeman et al., 1990, Kohnen et al., 1992). Carbon isotope studies of fatty acids, as well as steroids and hopanoids in deep-sea mussels, have been carried out to distinguish bacterial metabolism pathways as well as source carbon (Fang et al., 1993, Abrajano et al., 1994, Jahnke et al., 1995, MacAvoy et al., 2002). The importance of faunal usage of organic matter to chemotrophic bacterial production is revealed using the δ¹³C of specific fatty acids (MacAvoy et al., 2003). A recent study has also revealed a strong relationship in δ¹³C between methanotrophic steroid and faunal fatty acid, implying the carbon transfer between symbionts and hosts (Duperron et al., 2007). All these studies are performed on deep-sea mussels in cold hydrocarbon seeps of the Gulf of Mexico. Although a few compound-specific δ¹³C analyses are reported from deep-sea mussels around hydrothermal vents in the Atlantic (Pond et al., 1998), few biomarker occurrences have been reported from the Pacific Ocean.

In addition to carbon isotopes, hydrogen isotopic composition of lipid biomarkers is expected to provide information on biological activities with respect to water environment and lipid biosynthesis, because carbon-bound hydrogen in lipids such as hydrocarbon and fatty acid is stable and does not readily exchange with hydrogen of the surrounding water (Schimmelmann et al., 2006). Furthermore, hydrogen-containing reduced chemical species such as H₂, CH₄ and H₂S play important roles in bacterial metabolism as an electron donor around deep-sea hydrothermal environment. Although organic hydrogen of biomass is usually derived from water, a compound-specific hydrogen isotope study has demonstrated that about 30% of hydrogen from CH₄ utilized by methanotrophs are indirectly incorporated into the biomass in culture (Sessions et al., 2002), suggesting that the lipid hydrogen has a potential to discriminate bacterial metabolisms and hydrogen sources. Moreover, magmatic water is more depleted in D relative to normal seawater (e.g. Hoefs, 1997). The isotopically mixed water between magmatic water and seawater at hydrothermal vents could be utilized by bacteria to affect the hydrogen isotopic composition of their lipid biomarkers. However, such a molecular hydrogen isotope study has never been reported from bacterial communities at natural hydrothermal vents.

In the northwestern Pacific, a previous microscopic and phylogenic study on hydrothermal vent mussels (Bathymodiolus) clarified that B. platifrons at the North Knoll of the Iheya Ridge in the Okinawa Trough and B. septemdierum at the Myojin Knoll in the Izu–Bonin Arc harbored methanotrophic and thiotrophic endosymbiont, respectively (Fujiwara et al., 2000). In this study, we have performed compound-specific δ¹³C and δD analyses of lipid biomarkers including fatty acids and diploptene from three Bathymodiolus mussels around hydrothermal vents in the Pacific for the first time, in addition to conventional bulk CHNS isotope analyses. The goal of this study is to reveal δ¹³C-δD variations of biomarkers at different hydrothermal vent sites to clarify carbon and hydrogen sources as well as biochemical processes including lipid syntheses with respect to different symbiotic bacterial ecosystems.