Sulfur isotopes of hydrothermal vent fossils and insights into microbial sulfur cycling within a lower Paleozoic (Ordovician‐early Silurian) vent community

Abstract Symbioses between metazoans and microbes involved in sulfur cycling are integral to the ability of animals to thrive within deep‐sea hydrothermal vent environments; the development of such interactions is regarded as a key adaptation in enabling animals to successfully colonize vents. Microbes often colonize the surfaces of vent animals and, remarkably, these associations can also be observed intricately preserved by pyrite in the fossil record of vent environments, stretching back to the lower Paleozoic (Ordovician‐early Silurian). In non‐vent environments, sulfur isotopes are often employed to investigate the metabolic strategies of both modern and fossil organisms, as certain metabolic pathways of microbes, notably sulfate reduction, can produce large sulfur isotope fractionations. However, the sulfur isotopes of vent fossils, both ancient and recently mineralized, have seldom been explored, and it is not known if the pyrite‐preserved vent organisms might also preserve potential signatures of their metabolisms. Here, we use high‐resolution secondary ion mass spectrometry (SIMS) to investigate the sulfur isotopes of pyrites from recently mineralized and Ordovician‐early Silurian tubeworm fossils with associated microbial fossils. Our results demonstrate that pyrites containing microbial fossils consistently have significantly more negative δ34S values compared with nearby non‐fossiliferous pyrites, and thus represent the first indication that the presence of microbial sulfur‐cycling communities active at the time of pyrite formation influenced the sulfur isotope signatures of pyrite at hydrothermal vents. The observed depletions in δ34S are generally small in magnitude and are perhaps best explained by sulfur isotope fractionation through a combination of sulfur‐cycling processes carried out by vent microbes. These results highlight the potential for using sulfur isotopes to explore biological functional relationships within fossil vent communities, and to enhance understanding of how microbial and animal life has co‐evolved to colonize vents throughout geological time.


| INTRODUC TI ON
Hydrothermal vents in the modern deep ocean are populated by remarkable communities of highly specialized animals, dependent on symbioses with chemosynthetic prokaryotes for their nutrition.
These environments, which have existed on Earth since the Hadean (Martin et al., 2008;Russell & Hall, 1997), are also known to have been important habitats throughout geological history as evidenced by their fossil record (Campbell, 2006;Georgieva et al., 2021;Little et al., 1998). Microbes capable of oxidizing sulfide are particularly crucial to primary production within modern vent environments (Sievert et al., 2007;Sievert & Vetriani, 2012), and symbioses between metazoans and microbes involved in sulfur cycling are integral to the ability of animals to colonize and thrive at vents (Dubilier et al., 2008). For example, the tube-dwelling annelid worm Ridgeia piscesae (Annelida: Siboglinidae) obtains all of its nutrition from sulfideoxidizing endosymbionts (Chao et al., 2007), while another vent annelid tubeworm, Alvinella pompejana (Annelida: Alvinellidae) associates with and grazes on diverse bacteria involved in sulfur cycling that attach to its outer body wall and tube surfaces (Cottrell & Cary, 1999;Gaudron et al., 2012;Prieur et al., 1990). Furthermore, the microbial communities associated with the dwelling tubes of modern vent annelids can also be rapidly mineralized (and hence fossilized), with a range of microbial morphotypes preserved (Buschmann & Maslennikov, 2006;Georgieva et al., 2015;Maginn et al., 2002).
While it may be supposed that ancient vent animals would have also relied on microbes for their nutrition, the metabolic pathways of ancient hydrothermal vent microbes have not been explored, and it is not known if functional signals can fossilize within the sulfide minerals formed in hydrothermal vent settings.
One of the oldest hydrothermal vent communities that includes metazoans occurs within the Ordovician-early Silurian (~440 Ma) Yaman Kasy volcanogenic massive sulfide (VMS) deposit, Ural Mountains, Russia (Little et al., 1997(Little et al., , 1999. Yaman Kasy VMS deposit hosts the most diverse ancient vent community known, and also demonstrates exceptional fossil preservation by pyrite including micron-scale microbial fossils associated with the dwelling tubes of two fossil tubeworm species, Yamankasia rifeia and Eoalvinellodes annulatus (Georgieva et al., 2018). Sulfide minerals from the Yaman Kasy VMS deposit also contain elevated amounts of microbial hydrocarbons (Blumenberg et al., 2012). In non-vent settings in which fine-grained pyrite formation occurs, stable isotopes of sulfur are frequently employed to investigate the metabolic strategies of both modern and fossil organisms, such as to detect the sulfate-reducing metabolisms of Precambrian fossil microbial filaments (Schopf et al., 2015;Wacey et al., 2011). Microbial sulfate reduction and microbial sulfur disproportionation can result in large δ 34 S fractionations of 30‰-40‰ (Canfield, 2001;Detmers et al., 2001), while the effects of sulfide oxidation on sulfur isotope fractionation are smaller (Glynn et al., 2006). Phototrophic oxidation of hydrogen sulfide to elemental sulfur typically results in a small inverse isotope effect (of +1.8 ± 0.5‰), while oxidation of elemental sulfur to sulfate produces a small normal isotope effect (of −1.9 ± 0.8‰) (Zerkle et al., 2009). Sulfur isotope fractionations that result from chemotrophic sulfide oxidation (up to +8‰) are also distinct from those produced during abiotic oxidation of sulfide with oxygen (∼−5‰) (Zerkle et al., 2016), and can in some cases produce much larger effects (of up to +12.5‰) (Pellerin et al., 2019), suggesting it may therefore also be possible to detect signatures of sulfide oxidation in the fossil record.
Sulfur isotopes of pyrite have also been explored within both modern and ancient hydrothermal vent environments from a range of tectonic settings. δ 34 S values of 0 to +6‰ are typical for sulfide minerals from non-sedimented, Phanerozoic hydrothermal vents (Franklin et al., 2005;Huston, 1999;Seal, 2006). δ 34 S values for sulfides in modern systems generally reflect that of vent fluid hydrogen sulfide (H 2 S) from which the minerals precipitated; sulfide minerals that are in equilibrium with vent fluid H 2 S typically have δ 34 S values within −2‰ to +3‰ of vent fluid H 2 S, at temperatures of 100-400°C (Ohmoto & Rye, 1979;Shanks et al., 1995). At equilibrium conditions, pyrite should have δ 34 S values greater than that of associated vent fluid, but in practice, the δ 34 S values are often lighter, possibly as a result of precipitation from lower temperature fluids (Rouxel et al., 2004) or via thiosulfate intermediates (Ono et al., 2007). δ 34 S values of hydrothermal vent sulfide minerals below approximately −4‰ have been interpreted to indicate contributions from microbial sulfate reduction (MSR) (Ding et al., 2021;Eickmann et al., 2020;Peters et al., 2010). MSR is considered to be a particularly important process at sedimented hydrothermal vents (Fowler et al., 2019;McDermott et al., 2015) and in the hydrothermal vent subsurface, where it occurs during the low-temperature alteration of oceanic crustal rocks (Ono et al., 2012;Rouxel et al., 2004). Apart from several δ 34 S values collected through isotope ratio mass spectrometry for tubeworm fossils from the Yaman Kasy VMS deposit (−2.5‰ to +0.5‰; Herrington et al., 1998) et al., 2003), the sulfur isotope values of hydrothermal vent fossils, both ancient types and those forming in the modern ocean, have been largely unexplored. In addition, sulfur isotopes of seafloor hydrothermal vent sulfide minerals are rarely explored in high spatial resolution (at scales of ≤5 μm), but at this scale, sulfur isotope measurements have the power to resolve sulfur cycling processes related to distinct mineral texture types.
Here, we use high-resolution secondary ion mass spectrometry (SIMS) to explore for the first time the fine-scale δ 34 S signatures of pyrite delineating fossilized vent animals and associated microbes from both modern and Phanerozoic sediment-free hydrothermal K E Y W O R D S chemosynthesis, evolution, microbiology, paleobiology, pyrite preservation vent environments. We discuss our observations in the context of pyrite preservation and the nature of sulfur-based metabolic pathways that contributed to pyrite formation. We suggest that δ 34 S values of pyrite containing microbial fossils are indicative of sulfurcycling metabolisms of microbes from both recent and Ordovicianearly Silurian hydrothermal vent communities.  (Fornari et al., 2012). Both high-temperature vents (with fluids of over 50 to ~410°C) and moderate-to low-temperature vents (with fluids <30°C) occur on the 9°50'N segment (Hessler et al., 1985), while hydrogen sulfide concentrations of high-temperature vents are typically up to 1500 μM, and 100-300 μM at diffuse vents (Le Bris et al., 2006;Le Bris & Gaill, 2007). Alvinella spp. worms typically colonize the surfaces of high-temperature vents, where the dwelling tubes of these annelids are exposed to high rates of mineralization (Georgieva et al., 2015). The Endeavour segment of the Juan de Fuca Ridge is an intermediate-spreading mid-ocean ridge, with less-frequent eruptions and a magma-driven system that distinctly varies from that which occurs beneath the East Pacific Rise (Kelley et al., 2012). Large hydrothermal chimneys of over 30 m height are common at this site, and in general, hydrothermal fluids are enriched in methane and ammonia, unusual for a mid-ocean ridge vent. Vent fluid temperatures at the Main Endeavour Field can be up to 402°C, while proximal diffuse-flow sites are common. R. piscesae occurs at high densities at Endeavour and can colonize a range of vent conditions and exhibit a diversity of morphotypes (Tunnicliffe & Kim Juniper, 1990). The mineralization of the tubes of annelids has also been documented at this site (Cook & Stakes, 1995).

| G EOLOG IC AL BACKG ROUND
Ancient fossil samples used in this study consisted of two tubes from the late Ordovician-early Silurian Yaman Kasy VMS deposit (Figures S1, S3 in File S1), described as the fossil species Yamankasia rifeia and Eoalvinellodes annulatus (Little et al., 1999). The Yaman Kasy VMS deposit occurs in the southern Ural Mountains, Russia and is considered to be early Silurian to late Ordovician in age (Buschmann & Maslennikov, 2006). This deposit exhibits intricate preservation of hydrothermal vents that likely formed within a backarc basin . Vent chimneys and the most diverse ancient vent community are preserved as part of this deposit (Georgieva et al., 2021;Herrington et al., 1998), which has been extensively discussed in scientific literature due to the important insights it provides into ancient hydrothermal vent environments Buschmann & Maslennikov, 2006;Georgieva et al., 2021;Herrington et al., 1998;Herrington, Maslennikov, et al., 2005;Kuznetsov et al., 1993;Little et al., 1997Little et al., , 1999Maslennikov et al., 2017).

| ME THODS
Details of the four fossil tube samples selected for this study are listed in Table 1. Transverse and longitudinal sections of tube walls were prepared into polished blocks. Polished block sample images were collected using both reflected light (RL) and scanning electron microscopy (SEM) in backscatter electron mode to identify mineral phases and select targets for sulfur isotope measurements. RL microscopy was performed using a Zeiss AxioImager M2 microscope, and SEM using a FEI Quanta 650 FEG-SEM, both at the Natural History Museum, UK. Polished blocks were coated with carbon (approximately 10 nm thickness) prior to SEM. Pyrite was the only mineral phase selected for isotopic analysis, as it is the main mineral phase found to be preserving hydrothermal vent fossils (aside from silica and occasionally marcasite, zinc sulfides (sphalerite and/or wurtzite), and minor quantities of copper containing sulfides (chalcopyrite and isocubanite) observed in Georgieva et al., 2015Georgieva et al., , 2018.
Pyrite is also the main mineral phase found to preserve especially fine structures such as microbes (Georgieva et al., 2015(Georgieva et al., , 2018Little et al., 1997). Pyrite texture types targeted for analysis were based on the main textures observed within the specimens: pyrite containing microbial fossils, and pyrite without microbial fossils and exhibiting either a predominantly colloform, porous, or smooth texture.
Microbial fossils in hydrothermal vent pyrites are usually apparent as hollow filaments approximately 1 μm in diameter, present in a range of orientations, and exhibiting curved morphologies and occasionally "cell" shapes typical of bacteria such as septate divisions within filaments and rod-like morphologies (Georgieva et al., 2015(Georgieva et al., , 2018.
High-resolution images of pyrite containing microbial fossils in our samples are presented in Figures S2 and S3 in File S1.
For sulfur isotope analyses, samples were re-mounted into pol-    Figures 4c-e and 5b).

| DISCUSS ION
Recent research has shown that that hydrothermal vents are important preservational settings where mineral precipitation can readily fossilize resistant biological structures (from those made by macrofauna to microbes), at times with very fine detail and at sub-micron scales (Georgieva et al., 2015(Georgieva et al., , 2018(Georgieva et al., , 2021Maginn et al., 2002;Maslennikov et al., 2017). In addition, hydrocarbons of microbial origin may be preserved alongside vent sulfides for hundreds of millions of years (Blumenberg et al., 2012). As such, these environments provide a crucial glimpse into ancient deep-sea communities driven by chemosynthesis. In our study, we are confident that the pyrite we analyzed formed rapidly, preserving the hydrothermal vent fossils without later recrystallization or mineral replacement. As a result, the pyrite δ 34 S values that we have recorded here have not been affected by diagenesis and thus are likely to reflect primary formation conditions (Georgieva et al., 2015(Georgieva et al., , 2021. While there are emerging new studies on pyrite formation and precipitation pathways and the various controls on its sulfur isotope composition at vents (e.g., Findlay et al., 2019), studies of how these processes interact with biological structures are rare. It has been suggested that organisms can promote the precipitation of sulfide minerals at vents (Maginn et al., 2002;Peng et al., 2007;Zbinden et al., 2001) and affect pyrite texture (Georgieva et al., 2015), but apart from observations of increased phosphorus content in a mineralized annelid tube wall (Maginn et al., 2002) as whilst a limited overlap of datasets is seen, there is clearly a significant difference in δ 34 S between pyrite containing microbial fossils, which are always isotopically depleted in 34 S compared to nearby non-fossiliferous pyrite for all samples, whether ancient or modern ( Figure 5). Clearly, the presence or absence of microbial fossils influences the sulfur isotope signature of pyrites with different textures.
The range of pyrite sulfur isotope values reported here for mineralized Alvinella sp. and Ridgeia piscesae tubes are also quite different than those reported from most other modern vent sulfide samples, such as those described by Ono et al. (2007; +0.4‰ to +3.4‰) from the East Pacific Rise and Rouxel et al. (2004;−0.5‰ to +3.9‰) from Lucky Strike on the Mid-Atlantic Ridge. In the latter study, δ 34 S values of −0.5‰ were deemed to be among the lowest reported for seafloor hydrothermal deposits in non-sedimented mid oceanic ridges, and to possibly result from microbial sulfur cycling (Rouxel et al., 2004). In contrast, our δ 34 S results are as low as −7‰ (Table 1) There are several possible abiotic mechanisms through which the intriguing sulfur isotope signatures recorded in this study could have arisen. It is conceivable that the differences between δ 34 S of microbial-textured pyrite and pyrite without microfossils could be due to different pyrite generations with different sulfur isotope characteristics having been measured (for example, resulting from changes in the δ 34 S value of vent fluid). While this is plausible for δ 34 S differences between smooth and porous pyrite in the Ridgeia piscesae sample (Figures 1 and 5b) in which the smooth pyrite presents a later stage overgrowth, this is unlikely for the majority of measurements as the pyrite generation either appears to be the same (as in Figure 1b,d), or were taken from very small areas of sample. Therefore, given the rapid mineralization that would have been needed for preservation (e.g., Georgieva et al., 2015), it is more  (Georgieva et al., 2015;Maginn et al., 2002). Any SIMS analysis spots that appeared to be on a phase other than pyrite were filtered from analyses, on the basis of SEM observations of SIMS spot locations. Thermochemical sulfate reduction, during which sulfate is reduced to sulfide (Machel, 2001), may also result in 34 S-depleted sulfur phases without microbial involvement, resulting in kinetic isotopic fractionations up to 17‰ (Meshoulam et al., 2016). Studies of thermochemical sulfate reduction in modern hydrothermal systems have shown that hydrothermal pyrites formed from entrained seawater sulfate are actually slightly heavier in 34 S (e.g. Petersen et al., 2020). Given the association between negative δ 34 S values and microbial fossils in our samples, we suggest that the above process is a less likely mechanism to explain our observations.
A more reasonable explanation for the data is that the 34 Sdepleted sulfur isotope signatures of pyrite containing microbial fossils in our samples are due to microbial fractionation effects, given the microbial fossils that it has entombed. The occurrence of sulfide minerals even in areas without observable microbial fossils strongly suggests that hydrothermal H 2 S provided the dominant source of sulfur for pyrite formation processes. The δ 34 S of the original vent fluid supplying hydrogen sulfide to the study sites is unknown, but it would be reasonable to assume a value of 0‰ to +6‰ in line with values reported for a range of modern vent sites (Seal, 2006). For 9-10°N on the East Pacific Rise from which our Alvinella sp. sample was collected, δ 34 S values of +4.4‰ to +5.8‰ for hydrogen sulfide were reported by Ono et al. (2007). These values are also compatible with a range of ancient vent sites from throughout the Phanerozoic (Seal, 2006). Alternatively, the δ 34 S values of non-tube wall smooth pyrite (ranging from +0.8‰ to +2.0‰ in the Alvinella sp. and Ridgeia piscesae samples; Table S1) could be taken to represent end-member vent fluids that have not been biologically influenced at the time of deposition. In either case, pyrite containing microbial fossils (which ranges in δ 34 S from −9.3‰ to −0.1‰) in all of our samples ( Figure 5;   (Ding, Seyfried, et al., 2001) host microbes that can metabolize hydrogen (Adam & Perner, 2018).
Hydrogenotrophic sulfate reducers generally produce very small sulfur isotope fractionations between sulfate and sulfide (of +1‰ to +6‰; Hoek et al., 2006), and thus of a similar magnitude to our results. Alternatively, heterotrophic MSR by (hyper)thermophiles at high/optimal temperatures and correspondingly high cell-specific sulfate reduction rates also results in very small fractionations between sulfate and product sulfide (e.g., Canfield et al., 2006 Sulfur-oxidizing bacteria are particularly common at vents (Dick, 2019), and often colonize the tube surfaces of vent tubeworms such as Alvinella sp. and Ridgeia piscesae (Campbell et al., 2003;Kalanetra & Nelson, 2010). Sulfide oxidizers can produce small isotopic depletions in the resulting product (Balci et al., 2007;Lewis & Roy Krouse, 1968;Nakai & Jensen, 1964). Sulfide oxidizers can also produce sulfur products that are enriched in 34 S relative to the reactant (Brunner et al., 2008;Pisapia et al., 2007;Zerkle et al., 2009Zerkle et al., , 2016. Given the latter, sulfide-oxidizing bacteria colonizing the tubes of modern and ancient vent tubeworms could be producing S 0 that is heavier in δ 34 S, which would leave pyrite forming from vent fluid with a more depleted δ 34 S signature and thus may produce the sulfur isotope fractionations observed in our samples ( Figure 5). S 0 is a by-product of microbial sulfide oxidation at vents (Stein et al., 1988;Vetter, 1985) which has also been observed in association with the tubes of Alvinella sp., but does not appear to fossilize well (Georgieva et al., 2015). This suggests that the S 0 left by microbial sulfide oxidation was lost to the system, perhaps being further oxidized to  Pellerin et al., 2019). However, disproportionation is generally associated with much larger sulfur isotope fractionation effects (in the region of 20‰-30‰) than those observed in this study. As hydrothermal vents and particularly the surfaces of animal structures in these settings often host complex microbial communities capable of diverse metabolisms (Campbell et al., 2003;Lopez-Garcia et al., 2002), the δ 34 S values of microbially textured pyrite in this study likely represent a combination of all of the above complex microbial sulfur cycling processes.
It is also possible that negative δ 34 S values of porous pyrite, such as those observed in two of our samples (Figure 5b,d) also represent a contribution from microbial metabolic processes; however, given the absence of fossils in this pyrite, it is difficult to be sure. As microbial fossils are not always preserved alongside remnants of vent animals, further analysis of porous pyrite to establish its mode of formation would be highly beneficial as it may help to recognize microbial processing even in the absence of microbial fossils.

| CON CLUS ION
In this study, we were able to observe consistent δ 34 S signatures in pyrites from the recently-mineralized tubes of two annelid species from two distinct modern vent sites, as well as from two tubeworm fossils from the Ordovician-early Silurian Yaman Kasy VMS deposit.
In all cases, ancient or modern, we observed 34 S depletion of pyrites containing microbial fossils in comparison with adjacent pyrites not containing microbial fossils averaging approximately 3‰. Our results represent the first indication that the presence of microbial sulfur cycling communities active at the time of pyrite formation influence the sulfur isotope signature of pyrite, imparting a distinct 34 S depletion, and suggesting that associated pyrite may have also preserved traces of their metabolisms. We interpret these signatures to have resulted from a combination of sulfur cycling processes, including sulfide or elemental sulfur oxidation, hydrogenotrophic sulfate reduction, and/or heterotrophic sulfate reduction by hyperthermophiles. While these findings require further investigation to elucidate the microbial metabolic pathways responsible for the observed sulfur isotope signatures, our findings unearth new possibilities to explore the functioning of recent (e.g., at inactive vent fields) and ancient vent communities, such as the nature of associations between ancient vent animals and microbes. also very grateful to Tony Wighton and Callum Hatch for specimen preparation. Additionally, we thank the editor and three anonymous reviewers for their comments, which greatly helped to improve this manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data discussed in the paper are present in the main text and Files S1-S4.