Methanotrophic bacteria occupy benthic microbial mats in shallow marine hydrocarbon seeps, Coal Oil Point, California

Authors


Abstract

[1] Microbial mats composed of giant sulfur bacteria are observed throughout the benthos along continental margins. These communities serve to oxidize dissolved sulfides to sulfate, and are typically associated with the recent exposure of sulfide-rich sediments. Such mats are also ubiquitous in areas of hydrocarbon seepage, where they are thought to consume sulfide generated in underlying sediment. Despite the high abundance of dissolved methane in hydrocarbon seeps, few studies have considered the importance of methanotrophy in mat communities. To assess the importance of methanotrophs in microbial mats from hydrocarbon seeps, an approach involving lipid biomarkers, stable isotopes and enrichment culturing was applied. Microbial mat samples were collected from benthic surfaces at two hydrocarbon seeps located in the Coal Oil Point seep field, offshore from Goleta, California. Both samples display a high abundance of 16:1 fatty acids, including two isomers specific to type I methanotrophic bacteria, 16:1(ω8) and 16:1(ω6). Depleted values of δ13C found in 16:1 fatty acids suggests methane assimilation into biomass, whereas three separate investigations of sulfide-oxidizing bacteria yield fractionation factors too small to account for these values. On the basis of these observations and experiments, an isotope mass balance was applied to fatty acids present in the microbial mat samples which indicates methanotrophs contribute up to 46% of total fatty acids. These results implicate methanotrophy as an important function for microbial mats in seep areas, despite the visual appearance of these mats as being composed of giant sulfur bacteria.

1. Introduction

[2] Methane (C1) is the most abundant organic gas in the atmosphere and is an important contributor to global climate change [Cicerone and Oremland, 1988; Crutzen, 1991]. The bacterial consumption of methane is estimated to remove 32 Tg CH4 per year from the atmosphere [Cicerone and Oremland, 1988], which is only about 5% of the total sink. However, this estimate excludes consumption of methane in the subsurface, which prevents release to the atmosphere; this has been estimated at up to 688 Tg CH4 per year [Reeburgh, 1996]. Methane from the marine environment is thought to contribute only 3–5% of the annual methane flux [IPCC, 1994] to the atmosphere, but this is limited because most of the methane produced in the marine subsurface is consumed prior to entering the atmosphere [Kvenvolden et al., 2001; Reeburgh, 2007].

[3] The term methanotroph refers to a microorganism capable of utilizing methane as primary carbon and energy source. Though this process can occur both aerobically and anaerobically [Hanson and Hanson, 1996; Valentine, 2002], in this work only aerobic bacteria are considered. Most studies of methanotrophy in nature have been conducted in terrestrial ecosystems, such as forest soils, grasslands, rice paddies, wetlands, and sediments in fresh, saline or alkaline lakes [Whalen et al., 1992; Amaral et al., 1998; Costello et al., 2002; Macalady et al., 2002; Carini et al., 2005]. Methanotrophy in the marine environment is less-well characterized with disparate studies in the Gulf of Mexico [Zhang et al., 2002; Joye et al., 2004], the Black Sea [Reeburgh et al., 1991, 2006; Durisch-Kaiser et al., 2005], the Southern California borderland [Ward and Kilpatrick, 1990; Hinrichs et al., 2003], the Cariaco Basin [Ward et al., 1987; Kessler et al., 2005], the Eel River Basin [Valentine et al., 2001; Orphan et al., 2004] as well as estuaries [McDonald et al., 2005] and hydrothermal areas [Pond et al., 1998; Kotelnikova, 2002].

[4] Methanotrophs are classified into two major groups based on their internal membrane arrangement and on their carbon assimilation pathway. Generally, type I methanotrophs possess disc-shaped bundles of intracytoplasmic membranes and assimilate C1 carbon via the ribulose monophosphate pathway. Type II methanotrophs have an unusual paired internal membrane structure arranged along the periphery of the cell and utilize C1 carbon via the serine pathway [Hanson and Hanson, 1996; Holmes et al., 1996; Carini et al., 2005]. In natural ecosystems, type I methanotrophs have been observed as the exclusive or dominant group in marine and hypersaline environments [Holmes et al., 1996; Bourne et al., 2000; Khmelenina et al., 2000; Carini et al., 2005]. Type II methanotrophs have been observed as the dominant group in freshwater sediment, peat bogs and rice paddy soils [Costello and Lidstrom, 1999; Dedysh et al., 2000; Eller and Frenzel, 2001].

[5] Several lipids have been used as biomarkers specific for type I and type II methanotrophs. For example, phospholipid fatty acid 16:1(ω8) is considered a specific biomarker for type I methanotrophs and 18:1(ω8) for type II methanotrophs [Bowman et al., 1991; Guezennec and Fiala-Medioni, 1996; Hanson and Hanson, 1996; Macalady et al., 2002]. The neutral lipid diplopterol has also been used to identify methanotrophs in ancient marine sediments [Hinrichs et al., 2003]. Because the CH4 consumed by methanotrophs tends to be depleted in 13C relative to other forms of organic material, their lipids also tend to be depleted in 13C. As a result the application of compound specific isotope analysis has proven to be a powerful tool to study methanotroph distribution in nature [Zhang et al., 2002; Teske et al., 2002; Hinrichs et al., 2003; Pape et al., 2005].

[6] Offshore Goleta, California, lies one of the world's most spectacular areas of natural hydrocarbon seepage, the Coal Oil Point (COP) seep field (Figure 1) [Allen et al., 1970; LaMontagne et al., 2004]. Both natural gas and oil are emitted from the sea floor at this location with estimates of 1–2 × 105 m3 of methane [Hornafius et al., 1999; Quigley et al., 1999] and 80 barrels of oil emitted daily [Clester et al., 1996]. As with other seep areas and hydrothermal systems [Jannasch, 1985; Jacq et al., 1989; Guezennec et al., 1998; Gilhooly et al., 2007], white filamentous microbial mats cover the rocks and other substrates on the sea floor of this seep field. However, despite the numerous investigations of this environment, and its use as a model system, no studies have investigated the impact of methanotrophy on carbon cycling in and around the seeps.

Figure 1.

Coal Oil Point seep field study area, including sampling locations: Brian Seep and Shane Seep. Modified from LaMontagne et al. [2004].

[7] The goal of this study is to determine whether methanotrophs are active members of benthic microbial communities in and around submarine natural gas seeps located in the COP seep field, California. A lipid biomarker approach is applied, coupling mass spectrometric identification of putatively methanotrophic biomarkers with compound-specific 13C analysis. Isotopic (13C) fractionation associated with autotrophic growth of sulfide-oxidizing bacteria is also quantified to allow mass balance calculations.

2. Materials and Methods

2.1. Study Site and Sample Collection

[8] The study sites, informally named Brian Seep and Shane Seep, are located near the COP seep field (Figure 1). Brian Seep is a small gas seep located at 10 m water depth, which encompasses an area of approximately 400 m2. Fewer than 100 individual gas vents are present at this location. Methane comprises greater than 80% of the seep gas, with carbon dioxide, ethane, propane and hydrogen sulfide present at lower levels [Kinnaman et al., 2007]. Based on more than 50 trips made to this location, oil has not been observed to emanate from Brian Seep. The benthic surface at Brian Seep is a combination of rock and sand, and a large seawater intake pipe extends through the seep area. Microbial mats (Figure 2a) are present near the gas vents, existing primarily on rocky outcropping, as a rim around individual gas vents, and at locations where the vent gas contacts a solid surface, such as the pipe. Shane Seep is located at 21 m water depth, and has a gas flux of 3300 m3 per day [Washburn et al., 2005]. This seep is similar in area to Brian Seep but emits approximately 2 orders of magnitude more gas. Oil is also emitted from Shane Seep, typically as a coating on gas bubbles, but neither the flux of oil nor the composition of the oil are well established. The sediment at Shane Seep consists of finer grained material than at Brian Seep, and substantial amounts of tar are present in the sediment. Microbial mats (Figure 2b) cover much of the benthic surface at Shane Seep, substantially more-so than at Brian Seep.

Figure 2.

Images of microbial mats from (a) Brian Seep and (b) Shane Seep. Scale bar is provided in inches and cm.

[9] To provide a comparison site with mats free from methane exposure, we chose a reference site within the subtidal hydrothermal vents offshore White Point, California (33°42′50″N, 118° 19′00″W; 6 m depth). This site is a known habitat for sulfide oxidizers but lacks a significant input of methane [Jacq et al., 1989; Kalanetra et al., 2004].

[10] Sediment and microbial mat samples were collected by scuba divers on March 23 (Brian Seep sediment), July 7 (Brian Seep microbial mat) and July 14 (Shane Seep microbial mat) of 2005 and March 26, 2007 (White Point microbial mat) for fatty acid analysis. Another dive on February 13, 2006 was conducted at Brian Seep to collect a mat sample for enrichment of sulfide oxidizers. At Brian Seep, surface sediment samples (∼0–2 cm depth) were collected from an active gas vent, from ∼70 cm distal from the vent, and a location ∼30 m outside the seep field. Microbial mat samples were collected from the exterior surface of the seawater intake pipe (Figure 2a) by scraping the mat into a sample bottle in situ, and from rock surfaces by chipping the rocks from the sea floor, sealing them in plastic bags, and scraping the microbial mat from the rock once back in the laboratory. The mat sample from the pipe surface was used for biomarker analysis, whereas the rock samples were used for both biomarker analysis and as inoculum for growth of sulfide oxidizers. At Shane Seep and White Point, samples were collected in the same fashion as the rock samples described above, except that chipping was not required as the mats were growing on loose rock. All samples were frozen upon returning to the lab, except for the sample used for enrichment, which was inoculated immediately. Samples were stored at −20°C prior to analysis. Water samples were collected within the venting areas at all sampling sites with 1 L plastic bottles for analysis of δ13C of DIC. The gas from Brian and Shane Seep was collected with a funnel and sealed in 100 ml serum vials in situ for determination of δ13C in methane.

2.2. In Situ Enrichment of Microbial Mat

[11] To track the early development of a microbial mat community in the seep field, and to quantify associated isotope fractionation factors, an in situ enrichment was conducted at Shane Seep beginning on December 20, 2006. Briefly, a series of modular surfaces (10 cm × 10 cm) were bound to a mounting plate (60 cm × 50 cm) with cable ties. The device was positioned approximately 15 cm over the sediment surface at Shane seep such that the modular surface was consistently bathed with seep gas. A modular surface was collected by divers on January 3, 2007, and the mat partitioned into several aliquots. One aliquot was added to 1 ml sterile water and then a drop of diluted sample was distributed on a microscopic slide and dried by flash heating. After Gram Staining, the slide was visualized by oil emersion microscopy to determine morphologies of microbes in the mat sample. A second aliquot was frozen and stored for later analysis of fatty acids. Additional aliquots were archived.

2.3. Extraction and Analysis of Fatty Acids

[12] Fatty acid extraction from sediment, microbial mat, and cultures was based on published procedures [Sun et al., 1999; Ding and Sun, 2005]. The water content in the sediment was estimated based on the weight difference before and after baking 0.5 g of sediment at 100°C for 48 h. The microbial mat was scraped from the rock prior to extraction. Each sample was first extracted with about 10 ml of methanol, followed by 3 × 10 ml methylene chloride—methanol (2:1 v/v) mixture. During each extraction, the sample was sonicated for 10 min and then centrifuged at 4000 rpm for 10 min. All four extractions were combined into a 250 ml separatory funnel and partitioned into organic and inorganic phases formed by addition of 5% NaCl solution. Methylene chloride was the primary solvent in the organic phase to dissolve the lipids extracted from the sample. The organic phase in the funnels was released into a 250 ml round bottom flask. The inorganic phase was washed by 3 × 20 ml methylene chloride at 6-hour intervals and the organic phase was also released into the flask. The solvent from the extractions was removed with a rotary evaporator at 50°C. The residue in the flask was dissolved in a 3 ml 0.5 M KOH/methanol solution (repeated 3 times) and transferred into a 20 ml glass tube fitted with a Teflon-lined cap. The basic solution was saponified at 100°C for two hours. About 10 ml of hexane and 1ml of 5% NaCl were added to the cooling basic solution with gentle shaking performed by hand. The neutral lipids were thereby transferred into the hexane phase. The hexane phase was transferred to a 100 ml round bottom flask after four hours. This extraction was repeated twice more to remove all neutral lipids. The remaining solution was acidified by concentrated HCl to pH < 2 and extracted by 10 ml hexane three times. The extract was transferred to a 100 ml round bottom flask for rotary evaporation at 50°C. The flask was subsequently rinsed with l ml 14% BF3/methanol solution three times followed by 1 ml methanol three times. All the reagents were transferred to a 20 ml glass tube outfitted with a Teflon-lined cap, and heated at 100°C for two hours to form fatty acid methyl esters (FAMEs). After heating for two hours, about 3 × 6 ml (four hour intervals) of hexane was added into the solution to extract FAMEs into a 50 ml round bottom flask. The hexane was again removed by evaporation. 3 × 1 ml hexane was used to collect the FAMEs into 5 ml glass vials equipped with Teflon-lined caps. The solvent was evaporated under a stream of dry N2 gas. 50 μL of 2 mg/L nonadecanoic acid methyl ester (internal standard) and 50 μL hexane were added into the vial prior to gas chromatography (GC) analysis.

[13] The FAMEs were analyzed with an HP-5890 series II GC system equipped with an on-column split injector and a flame ionization detector. 1 μL of sample was injected for analysis and compound separation was achieved by an Alltech 30 m × 0.25 mm AT™-5MS capillary column. The oven temperature gradient was 70–170°C at 20°C/min, followed by 170–300°C at 4°C/min and held at 300°C for 5 min. An HP 3396 series III integrator was used to quantify the fatty acid concentrations by comparing the area of each fatty acid to an internal standard.

2.4. Identification of Fatty Acid Structure

[14] An HP 5890 GC combined with an HP 5970 GC mass selective detector (GC-MS) was used to identify the structure of each fatty acid. The column and the temperature gradient of the GC-MS were the same as for the FAME analysis described above and helium was used as carrier gas. Operating conditions of the GC-MS were mass range 20–560 amu with a 0.4 s scan interval and 70 eV ionizing energy. To determine the double bond position and geometry of monounsaturated fatty acids, the FAMEs were reacted with dimethyl disulfide (DMDS) following the method of Nichols et al. [1986]. Briefly, the FAMEs were treated with DMDS under 6% iodine solution (w:v in diethyl ether) at 50°C for 48 h. After that, the iodine was removed with 5% Na2S2O3 and the organic product was reanalyzed by GC-MS as before.

2.5. Isotope Analysis of Individual Lipids and CO2

[15] Stable carbon isotope ratios (δ13C) were measured using a Thermo Finnigan GC-Isotope Ratio Mass Spectrometry system (IRMS) at the Marine Science Institute Analytical Lab, University of California, Santa Barbara. This system includes a Trace GC with splitless injector, a GC combustion III interface to oxidize organic compounds to CO2, and a Delta Plus XP mass spectrometer. The GC combustion III unit catalyzes the oxidization of organic matter with Cu/Ni/Pt wire heated to 950°C and water is removed through a selectively permeable membrane. The carrier gas in the system is helium with a flow rate of 1.0 mL/min. For FAMEs, the column and the oven temperature gradient were the same as those used for fatty acid analysis. Peaks eluted from the GC column were converted into CO2 and analyzed by IRMS. Carbon isotope ratios for each compound were measured relative to a CO2 standard (δ13C = −32.6‰) made by Air Liquid. To test the reproducibility, duplicate measurements of each sample were conducted. For 65% of δ13C values in identified fatty acids, the difference between duplicates was less than 0.2‰ and in no case was greater than 0.5‰. The internal standard—nonadecanoic acid methyl ester was used to track the accuracy of the instrument. To obtain actual fatty acid isotope ratios, the δ13C of FAMEs were corrected according to carbon numbers added during their derivitization. The δ13C of the methanol (one carbon added to FAMEs) was determined by direct combustion through an elemental combustion system from Costech Instruments with a Delta Plus Advantage IRMS from Thermo Finnigan. For DIC analysis, 10 ml water was injected into a sealed serum vial and acidified with 6 mol/L HCl to produce CO2. The column used for CO2 measurements was a 0.32 mm × 30 m ID CarbonPlot capillary column (Agilent Technologies). The isothermal temperature was set at 60°C. The software, Isodat NT, was used to calculate the δ13C values.

2.6. Enrichment of Sulfide-Oxidizing Bacteria

[16] The microbial mat sample collected from Brian Seep in February 2006 was used as inoculum to investigate the lipid composition of sulfide oxidizers in the absence of methane. The enrichment of sulfide-oxidizing bacteria from microbial mats was carried out in marine gradient medium, established in 20 ml (16 × 150 mm) screw-capped tubes. Marine gradient medium was constructed in the tube according to Nelson's [1992] method. Briefly, the solidified medium in the bottom was 4 ml J3 medium (pH 8.4; 1.5% agar and the concentration of NaHCO3 was 2.0 mM) supplemented with freshly neutralized Na2S. J3 medium is a Marine Basal Medium [Nelson, 1992]. The semisolid J3 medium without sulfide and thiosulfate (0.25% agar) overlaid the solid medium in the bottom. In this gradient medium, all of the reduced sulfur was fixed in the bottom medium. The sterile phenol red solution (0.5%, Fisher) was used as pH indicator in both top and bottom medium. Gradient media was aged for 2 to 3 days prior to inoculation in order to establish a sulfide-oxygen interface between the two media. A small inoculum of freshly-sampled microbial mat was then added to the medium. The enrichment was conducted in a 20 L gas tight container filled with an 80/20 mixture of air/CO2. To maintain the δ13C of dissolved inorganic carbon (DIC) within a narrow range, the same source of DIC was used for both the medium and the headspace. This was accomplished for the headspace by quantitatively acidifying NaHCO3 to CO2. After 5–7 days the incubation tubes were removed from the container and used for fatty acid and isotope analysis. At the beginning and the end of enrichment, the gradient media (solid and semisolid parts) were sampled and acidified to produce CO2 for δ13C analysis. The δ13C-CO2 in the container was also sampled concurrently.

3. Results

3.1. Fatty Acids From Seep Sediments

[17] Three sediment samples collected from Brian Seep were analyzed for biomarkers related to methanotrophs. Eleven fatty acids were quantified and identified in the three samples, including straight chain saturated fatty acids 12:0 (not found in sediment 30m from the gas vent), 14:0, 15:0, 16:0 and 18:0; branched i-15:0 and a-15:0; and unsaturated 16:1 and 18:1 series (Figure 3). The concentrations of fatty acids in the sediment from the vent were substantially higher than in the samples from 70 cm or 30 m away from the vent. At all three sites, 16:0 and 16:1 were the most abundant and each accounted for more than 20% of total fatty acids in the sample.

Figure 3.

(a) Concentrations and (b) δ13C values of fatty acids extracted from surficial sediments collected at a gas vent, 70 cm removed from the vent, and 30 m removed from the vent–in the area of Brian Seep (the error bars are smaller than the data points and are therefore not shown). The δ13C of DIC for bottom water at Brian Seep was −0.3‰, but was not measured in the pore fluid.

[18] The δ13C values of individual fatty acids ranged from −20‰ to −30‰ (Figure 3b). In general, δ13C values in saturated straight chain fatty acids such as 14:0 and 18:0 were more enriched, and the values in branched and unsaturated fatty acid were more depleted, though exceptions include 12:0 (−30.2‰) and 18:1(ω7) (−24.2‰) at the gas vent. For most of the fatty acids, the lowest δ13C values were found in seep surface sediment, except for 16:0, 18:0 and 18:1(ω7). In each sediment sample, the most depleted δ13C of fatty acids was detected in 16:1(ω5) and the most enriched δ13C was observed in 18:0.

[19] These results show higher levels of all identified fatty acids in the gas vent compared to sediment further removed from seepage (Figure 3a). However, there are no clear indications of methanotrophy apparent in the distributions or isotopic composition of fatty acids present, with the possible exception of 16:1(ω5) which was present only at low concentrations, but was the most 13C depleted fatty acid in each sample.

3.2. Fatty Acids From Microbial Mats

[20] The filamentous microbial mats collected from Brian and Shane Seeps (Figures 2a and 2b.) colonize benthic surfaces throughout their respective seep fields and their visual appearance is consistent with large sulfide-oxidizing bacteria. The habitats of these mats collected from shallow seep areas differ slightly from those found in seeps at greater depth [Mills et al., 2004; Valentine et al., 2005] as these mats are not observed at the sediment-water interface, but rather colonize surfaces of rocks and other hard substrates. Fatty acids were extracted from the microbial mat samples collected from Brian Seep and Shane Seep were then quantified, identified and analyzed for δ13C by compound specific isotope analysis. All eleven identified fatty acids in the sediment samples were detected in these microbial mats (Table 1). Unlike sediment samples, the 16:1 isomers are collectively the most abundant fatty acids in all mat samples. The next most abundant fatty acid was 16:0 in each sample. These two fatty acids accounted for more than 60% of the total in each sample. The δ13C values of individual fatty acids in microbial mat samples were more variable than in the sediment samples (Table 1). Values ranged from −24.4‰ to −40.2‰ in the mat from Brian Seep, and from −27.3‰ to −40.7‰ in the mat from Shane Seep. The most negative δ13C values were found in 16:1 isomers, including 16:1(ω5).

Table 1. Relative Percentage, δ13C (‰) and Apparent equation image (‰) for Fatty Acids From Brian Seep, Shane Seep, White Point, an in Situ Enrichment, and a Laboratory Enrichment
FAaBrian Seep MatShane Seep MatWhite Point MatIn Situ EnrichmentLaboratory Enrichment
%bδ13Cɛc%bδ13Cɛc%bδ13Cɛd%bδ13CɛDICcɛCO2e%bδ13Cɛf
  • a

    Fatty acids.

  • b

    Relative percentage of fatty acids in sample.

  • c

    δ13C in DIC is −0.3‰.

  • d

    δ13C in DIC is −8.2‰.

  • e

    δ13C in CO2 is 16.9‰.

  • f

    δ13C in DIC is −2.9‰.

  • g

    None detected.

12:01.5−24.424.14.3−30.229.92.8−22.013.82.3−28.628.345.52.1−27.524.6
14:06.8−32.332.09.9−37.537.211.6−24.316.114.1−22.121.839.08.8−24.321.4
i-15:01.3−30.730.41.5−32.932.60.8−29.721.51.3−28.828.545.74.7−28.926.0
a-15:00.9−31.130.82.5−35.034.71.4−29.721.52.7−31.030.747.91.7−23.320.4
15:02.2−28.127.85.9−29.228.92.2−27.419.25.0−27.527.244.49.7−24.521.6
16:144.5−39.639.337.2−40.740.438.5−35.327.129.2−9.28.926.124.0−21.019.1
16:1(5)14.0−40.239.96.6−40.640.3NDgNDgNDgNDgNDgNDgNDgNDgNDgNDg
16:018.1−32.532.220.7−31.531.224.9−33.725.829.2−19.118.836.031.8−23.320.4
18:1(9)3.6−27.827.52.3−33.032.72.2−26.418.24.4−24.524.241.43.5−20.817.9
18:1(7)3.7−28.628.36.9−32.632.313.0−38.230.05.9−11.711.428.67.0−21.118.2
18:03.6−24.724.42.5−27.327.32.9−26.318.16.0−22.722.439.66.8−22.219.3

[21] The 16:1 fatty acids extracted from microbial mat samples in the seep area were more abundant than any other fatty acids and also had the most depleted δ13C values. The 16:1 series was initially resolved into two peaks by GC (Figure 4a) for the mat samples collected from Brian and Shane Seeps. The small peak was sharp, narrow and relatively symmetric. The large peak was abnormally wide, asymmetric and had several small shoulders (Figure 4a), indicating coelution of several compounds. The GC-MS of these two peaks were very similar. After reacting with DMDS the two initial peaks were resolved into 6 peaks which were identified as 16:1(ω9), 16:1(ω8), 16:1(ω7), 16:1(ω6), 16:1(ω5c) and 16:1(ω5t) (peak numbers 1–6 in Figure 4b), based on the fragmentation patterns and retention times.

Figure 4.

Chromatogram of fatty acids extracted from a microbial mat sample collected from Brian Seep before and after DMDS adduction: (a) the 16:1 family of fatty acids includes several isomers that could not be resolved and identified; and (b) after DMDS adduction 16:1 series fatty acids were resolved into six peaks identified as: 16:1(ω9), 16:1(ω8), 16:1(ω7) and 16:1(ω6), 16:1(ω5c) and 16:1(ω5t), respectively.

[22] The presence of 16:1(ω6) and 16:1(ω8) in the microbial mat samples is consistent with methanotrophic bacteria as these lipids are known biomarkers for type I methanotrophs and are rare in other organisms. The depleted δ13C values for 16:1 fatty acids further support this interpretation and also indicate that methanotrophic biomass contributes significantly to the microbial mats. However, in order to determine the extent to which methanotrophs contribute to the microbial mat community, it is necessary to know the isotopic composition of lipids arising from other members of the microbial mat community.

3.3. Fatty Acids and Isotope Fractionation From Sulfide-Oxidizing Bacteria

[23] To determine if the 13C-depleted lipids present in the microbial mats from Brian and Shane Seeps can be attributed to methanotrophy, the isotopic contributions from other potential sources, particularly sulfide oxidizers, must be known. To address this uncertainty, three distinct approaches were taken. First, a microbial mat sample was collected from Brian Seep and used as inoculum to enrich sulfide oxidizers in the absence of methane. Fractionation factors for biosynthesis of individual lipids during autotrophic growth were determined by quantifying the δ13C of supplied DIC and the δ13C of resulting lipids. Second, a natural microbial mat sample was collected from a coastal hydrothermal vent (a habitat known to harbor sulfide-oxidizing bacteria, but lacking a significant supply of methane) and the fractionation factors associated with lipid biosynthesis during autotrophic growth were inferred through comparison of lipid δ13C with ambient δ13C-DIC. Third, a short-term enrichment was conducted in situ at Shane Seep, whereby microbial mats were cultivated at the sea floor within the seep on an artificial surface; fractionation factors associated with lipid biosynthesis during autotrophic growth were inferred through comparison of lipid δ13C with both ambient δ13C-DIC and δ13C-CO2 from seep gas. Results from these experiments are described in the following three sections and are summarized in Table 1.

3.3.1. Laboratory Enrichments

[24] A fresh microbial mat sample collected from Shane Seep was used as inoculum to enrich sulfide-oxidizing bacteria on gradient medium. The fatty acids in this mat sample were analyzed in order to define the composition of the inoculum (Figure 5a) and to control for potential carryover. The dominant fatty acids remained 16:1 isomers and 16:0 (about 30% each), with a decrease in the relative abundance of 16:1 fatty acids and an increase in 16:0 fatty acids. The relative percentages and δ13C values of other fatty acids differed somewhat with the sample collected previously, but the lowest δ13C values still occurred in 16:1 isomers in both samples and with similar values.

Figure 5.

(a) Relative percentages and (b) δ13C values of fatty acids extracted from a natural microbial mat sample collected from Brian Seep and a methane-free incubation of the same microbial mat. The δ13C of DIC was −0.3‰ at Brian Seep and −2.9‰ for the enrichment.

[25] After 5–7 days of enrichment substantial amounts of filamentous biomass accumulated on the interface between top and bottom medium, suggesting the dominant microbe in the enrichment needed both reduced sulfur compounds and O2 for growth - an important feature of sulfide oxidizers. The sulfide-oxidizing enrichment contained many of the same fatty acids as the inoculant, with 16:1(ω5) notably absent (Figure 5b). The relative percentages of the other ten fatty acids also varied between inoculant and enrichment (Figure 5b). In the inoculant, 15:0 was a minor component, but after incubation, its relative abundance increased to more than 9%. The relative abundance of 16:0 increased slightly whereas 16:1 (16:1(ω5) included) decreased slightly.

[26] In the microbial mat sample (e.g., inoculant), 16:1 fatty acids were significantly depleted in 13C (Figure 5b). Among them, the δ13C of 16:1(ω5) was −44.6‰ and the average of other 16:1 isomers was −36.1‰. The δ13C values of other fatty acids were more enriched compared to 16:1 series, ranging from −24.4‰ to −31.3‰. In the methane-free enrichment, the 16:1 isomers had a δ13C value of −21‰, and eight fatty acids displayed δ13C values more enriched than −25‰. Only two fatty acids (12:0 and iso-15:0) showed δ13C values more depleted than −25‰, including iso-15:0 with the most negative δ13C (−28.9‰). For the individual fatty acids, the δ13C value in the inoculum was lower relative to that in the culture, with the exception of 15:0 which showed little difference. The largest difference of δ13C was detected in 16:1 isomers. The inoculant was depleted by as much as 15.1‰ relative to the culture. Besides 16:1 isomers, the differences of δ13C in 16:0, 18:1(ω9) and 18:1(ω7) were over 5‰ while other fatty acids differed by less than 4‰.

[27] In order to determine the fractionation associated with autotrophic growth of sulfide oxidizers, the δ13C of the lipids was compared with the δ13C-CO2 in the enrichment medium. The per mille enrichment factor, ɛ, is defined by Hayes [1993] as:

display math

where α represents the fractionation factor for coexisting CO2 and fatty acid. If the δ13C values of CO2 and fatty acid are near zero, the ɛδ13C-CO2δ13C-fatty acid. The δ13C of DIC in the seawater collected from Brian Seep was −0.3‰, and the apparent ɛ ranged from 24.1‰ (15:0) to 44.3‰ (16:1(ω5)) in the natural microbial mat. The enrichment medium was found to have a δ13C-CO2 ranging from −2.1‰ to −3.7‰, with an average of −2.9‰, whereas δ13C in the fatty acids ranged from −20.8‰ to −28.9‰. Apparent isotopic fractionation associated with fatty acid biosynthesis by these autotrophic, sulfide-oxidizing bacteria thus ranges from 17.9‰ to 26.0‰ (Table 1), with a mass weighted average ɛ of ∼20.7‰.

3.3.2. Hydrothermal Communities

[28] In the microbial mat sample collected from White Point hydrothermal vent, the δ13C values of fatty acids ranged from −22.0‰ in 12:0 to −38.2‰ in 18:1 (ω7) (Table 1). The second most depleted fatty acid was 16:1 (δ13C = −35.3‰) which was also the most abundant fatty acid in the mat, and followed by 16:0 (δ13C = −33.7‰), which was the second most abundant fatty acid in the mat. Unlike the mat samples from Brian and Shane seeps, only 16:1(9) and 16:1(7) were found in the mat from White Point hydrothermal vent. The δ13C of the DIC in the adjacent seawater was −8.2‰, and the apparent ɛ values between DIC and the fatty acids range from 13.8‰ to 30.0‰, with a mass weighted average −24.8‰. It is known that sulfide-oxidizing bacteria are the dominant bacteria in White Point hydrothermal vent [Jacq et al., 1989; Kalanetra et al., 2004]. Thus fractionation factors such as these are of insufficient magnitude to account for the depleted δ13C values observed from COP.

3.3.3. In Situ Seep Enrichment

[29] Because the ɛ values determined from enrichment cultures and from White Point may not represent dominant sulfide oxidizers from COP, an in situ enrichment was conducted to estimate these ɛ factors. The in situ enrichment was set for two weeks at the sea floor within Shane Seep where it was continuously bathed with seep gas. This period of time was chosen to be sufficiently long to allow for growth of sulfide oxidizers, but to be too short to favor growth of methanotrophs, which typically grow slowly [Lidstrom, 1988]. Visualization of the resulting mat using microscopy showed that long, filamentous bacteria containing cellular inclusions, presumed to be elemental sulfur, dominated the in situ enrichment. The δ13C values of individual fatty acids ranged from −9.2‰ to −31.0‰ (Table 1). In stark contrast to natural microbial mats, the most enriched 13C occurred in fatty acid 16:1. Since the device for in situ incubation was bathed in seep gas it is likely the sulfide-oxidizing bacteria utilized CO2 from the seep gas, which is isotopically enriched relative to sea water DIC. Because of the two potential carbon sources, we calculated two groups of ɛ values based on the difference between the δ13C values in each fatty acid with the δ13C values of DIC and of the CO2 from the seep gas (Table 1). The δ13C of DIC in Shane Seep was −0.3‰ as measured during the enrichment period and the resulting ɛ values range from 8.9‰ to 30.7‰, with mass weighted average 17.4‰. The δ13C of the CO2 from the seep gas at this time was 17.1‰. The corresponding ɛ values ranged from 26.1‰ to 47.9‰, with mass weighted average 34.6‰. The lack of putative methanotrophic biomarkers (16:1(ω6, and 8)) and of isotopically-depleted lipids (δ < −35 ‰) suggests that methanotrophs did not contribute significantly to this enrichment. The variability in lipid δ13C likely arises from multiple sources of CO2 feeding autotrophy, with sulfide oxidizers utilizing primarily CO2 derived from seep gas, and being the dominant producers of 16:0, 16:1 and 18:1(ω7), with ɛ values of ∼20 to 25 ‰.

4. Discussion

4.1. Aerobic Methanotrophs Inhabit Gas Vents

[30] Aerobic methanotrophy has been studied extensively owing to the importance of methane as a greenhouse gas. The vast majority of studies concern terrestrial but not marine environments. Studies involving marine methanotrophs have involved the quantification of in situ methane oxidation rates [Ward and Kilpatrick, 1990; Reeburgh et al., 1991; Valentine et al., 2001; Krüger et al., 2005], the use of biomarkers to determine past methane emissions [Hinrichs et al., 2003], the isolation of novel strains [Sieburth et al., 1987; Lidstrom, 1988; Holmes et al., 1996], endosymbiotic associations with macrofauna [Cavanaugh et al., 1992; Fisher et al., 1993], and the associations with methane hydrate [Bidle et al., 1999]. The presence of methanotrophs in shallow-water seeps was previously undocumented.

[31] Six distinct 16:1 fatty acid isomers were identified in microbial mat samples collected from areas of active gas venting. While 16:1(ω7) and 16:1(ω9) are found in many organisms, 16:1(ω8) and 16:1(ω6) are used as putative biomarkers for type I methanotrophs [Bowman et al., 1991; Jahnke et al., 1995]. In addition to 16:1(ω8) and 16:1(ω6), type I methanotrophs also produce cis and trans 16:1(ω5), which were also identified in the samples. The presence of putative methanotropic biomarkers in these natural microbial mat samples, coupled with the observed 13C-depletions in the 16:1 fatty acids and the abundance of methane in this environment, strongly support the existence of active methanotrophic bacteria in these microbial mats.

[32] Based on numerous sea floor observations at the COP seep field, including those made at the time of sampling, microbial mats preferentially grow on solid surfaces when in shallow-water (depth < 30 m). At Brian Seep the microbial mat was attached to a large, mineral encrusted pipe overlaying an active gas vent. At Shane Seep mats were located on rocky outcrops adjacent active vents. As a comparison, in seep areas at greater depth [Mills et al., 2004; Valentine et al., 2005], mats are normally found at a physical redox boundary, such as the sediment-water interface. This difference suggests both oxidized and reduced chemicals are supplied to the microbial mats by way of seawater at the COP seep field, and not through interstitial fluids. In fact, this seep field is characterized by oxygen rich waters as well as strong bottom currents, and the seep gas provides the reduced sulfur compounds. Given the high ambient oxygen levels, the physical orientation and attachment of the microbial mats, the presence of putative biomarkers for aerobic methanotrophy, and the notable absence of typical biomarkers for sulfate reducing-bacteria, such as i-17:1 or i-17:0, it is likely that these microbial mats are oxygen rich, and that aerobic metabolism predominates.

[33] Application of microscopy to view various mat samples from COP reveals that filamentous organisms represent the dominant morphotype. While originally we assumed these filaments represented sulfide oxidizers, several works [Völker et al., 1977; Stoecher et al., 2006; Vigliotta et al., 2007] have shown the existence of filamentous, methanotrophic bacteria. While such organisms are currently known only in terrestrial environments we find it likely that similar organisms inhabit the mats described here.

4.2. Estimating the Relative Abundance of Methanotrophs in Microbial Mats

[34] The relative abundance of methanotrophs among the microbial mat community can be estimated in terms of lipid abundance using an isotope mass balance approach. This is achieved by quantifying the percentage of fatty acids derived from methanotrophic bacteria using the following mass balance,

display math

where δ13Ca is the average carbon isotope composition of all fatty acids, δ13Cm, δ13Cs, and δ13Co are the average carbon isotope compositions of fatty acids derived from methanotrophs, sulfide-oxidizing bacteria, and other organisms, respectively, and where fm, fs, and fo are the fractions of fatty acids derived from these same groups. By assuming that methanotrophs and sulfide oxidizers are the two dominant metabolic groups in the microbial mats, equation (2) can be simplified as follows:

display math

and equation (3) can be rearranged to solve for fm

display math

[35] In order to quantify fm, average isotopic compositions of fatty acids must be determined for the entire mat, the methanotrophs, and sulfide oxidizers. Of these only δ13Ca is determined directly, whereas δ13Cm and δ13Cs must be inferred. The following equation was used to calculate δ13Ca,

display math

where Ri and δ13Ci represent the relative abundance and δ13C of individual fatty acids. Identified fatty acids account for more than 98% of total fatty acids in the microbial mat.

[36] A δ13C value of −45‰ was assigned for δ13Cm. This was based on the δ13C of 16:1(ω5), which was the most 13C-depleted fatty acid in each sample. This fatty acid is known to be produced by methanotrophic bacteria and was not observed in any of the three studies of sulfide-oxidizing communities. A δ13C value of −21‰ was assigned for δ13Cs. This was determined by applying a fractionation factor (20.7‰), determined in section 3.3, to the δ13C of seawater at the seep site (−0.3 ± 0.2 ‰; n = 4), which was measured independently.

[37] Based on values of −32‰ for δ13Ca, −45‰ for δ13Cm and −21‰ for δ13Cs, fm is estimated to be 46%. That is, nearly half the lipids present in the microbial mats are attributed to methanotrophic bacteria, clearly indicating an important role for these organisms in the seep environment. Calculations involving more conservative values for δ13Cm still yield a high percentage of methanotrophic lipids. For example, a δ13Cm = −67‰ (seep CH4 with a biosynthetic fractionation of 22‰) yields fm = 24%. More liberal estimates can also be justified if enriched CO2 from the seep gas becomes incorporated into biomass, as described in section 3.3.3.

[38] The mass balance equations above were also used to quantify fm for each fatty acid independently with the same assumptions. Results indicate that the fraction of 16:1 fatty acids attributable to methanotrophs ranges from 79% at Shane Seep to 83% at Brian Seep. For other fatty acids, the contribution of methanotrophy is more than 30% in 12:0, 16:0 and 18:1 series, lower than 25% in 14:0, 18:0 and branched iso- and anteiso-15:0, and less than 11% in 15:0. These results seem reasonable given that each of these fatty acids has been observed previously in both methanotrophic and sulfide-oxidizing bacteria [Bowman et al., 1991; Jacq et al., 1989].

4.3. Coexistence of Methanotrophic and Sulfide-Oxidizing Bacteria

[39] While sulfide- and methane-oxidizing bacteria are known to coexist as endosymbionts [Fisher et al., 1993; Pond et al., 1998], there have been only limited observations regarding the cooccurrence of these groups in other environments [Werne and Damsté, 2005; Roy et al., 2004], and no definitive evidence has previously been presented that these groups coexist in microbial mat communities. The coexistence of the two bacterial groups observed in these seeps is most likely driven by the abundance of methane, sulfide, CO2 and oxygen. Methane and sulfide are supplied from the seep gas whereas oxygen is delivered from the sea water. Dissolved inorganic carbon may be delivered from sea water or via CO2 in seep gas. The availability of these chemicals and the growth potential of the individual organisms are likely primary factors governing the distributions of these microbial groups. The appearance of macrofauna in the mat suggests that resistance of the microbes to predation is another important factor controlling their distributions.

[40] Seep environments harboring methanotrophic mats differ from most benthic settings where mats are present [Paull et al., 2005]. That is, methane is typically oxidized in the subsurface with sulfate as oxidant and appreciable amounts do not reach the oxic-anoxic interface in sediments characterized by diffusion as dominant mixing process [Valentine and Reeburgh, 2000]. Therefore methanotrophs are unlikely to significantly inhabit benthic mats except in seep environments where methane is advected through the underlying sediments. This is likely the single most important factor controlling the global distribution of benthic methanotrophs in marine environments.

5. Conclusions

[41] This study reveals the cooccurrence of methanotrophs and sulfide oxidizers in an oxic, shallow coastal seep area. Such communities likely occur in other seep environments and are likely to harbor more complex ecophysiology than described here. Likely interactions include competition for space and nutrients, genetic and metabolic exchange, differential growth rates and resistivities to predation, and population successions. The development of microbial mats in seep environments and the interactions between the bacterial groups and their surroundings is the focus of ongoing investigation.

Acknowledgments

[42] We thank Luis Busso, Justine Kimball and Shane Anderson for assistance collecting samples. We acknowledge Molly Redmond, Susan Mau, Frank Kinnaman, Monica Heintz, Sankar Marichamy and Douglas Nelson for providing comments on an early draft of the manuscript. We appreciate Frank Kinnaman and Robert Petty's direction on isotope ratio analysis, and Sankar Marichamy's assistance with microscopy. Funding for this study was provided by the National Science Foundation (OCE-0447395) and by the Petroleum Research Fund (40643-G2).

Ancillary