Miniaturized biosignature analysis reveals implications for the formation of cold seep carbonates at Hydrate Ridge (off Oregon, USA)

Methane-related carbonates from Hydrate Ridge typically show several macroscopically distinguishable mineral phases, namely whitish aragonite, lucent aragonite, and gray micrite. The relationship of these phases to particular microorganisms or biogeochemical processes is as yet unclear. We used a miniaturized biomarker technique on mg samples, combined with factor analysis and subsequent electron microprobe analysis, to study lipid biomarkers and chemical compositions of the individual phases. This allows us to identify particular mechanisms involved in the formation of the different carbonate precipitates. Our combined analysis of biomarkers and petrographic traits shows that most of the lipids related to the anaerobic oxidation of methane (>90% by weight) are concentrated within only a minor compartment (~20% by volume) of the Hydrate Ridge carbonates, the whitish aragonite. The patterns indicate that the whitish aragonite represents fossilized biofilms of methanotrophic consortia containing mainly archaea of the ANME-2 group and sulfate reducing bacteria, whereas the precipitation of the lucent aragonite may have lacked the immediate proximity of microorganisms during formation. By contrast, the gray micrite formed by incorporation of allochthonous organic and inorganic matter during carbonate precipitation induced by the anaerobic oxidation of methane involving ANME-1 archaea.


20
Specific carbonates occur at cold seep sites, where methane-rich fluids are leaking from the seafloor. These "seep carbonates" typically show highly negative δ 13 C values (Greinert et al., 2001) indicating that they formed from bicarbonate produced by anaerobic oxidation of methane (AOM; Ritger et al., 1987). AOM is mediated by a consortia of methanotrophic archaea and sulphate-reducing bacteria (SRB), which have been 25 characterized by 16S rRNA investigations (Hinrichs et al., 1999;Boetius et al., 2000).

EGU
Two major phylogenetic groups of methanotrophic archaea (ANME-1 and ANME-2, ANME=anaerobic methane oxidizers) were distinguished. While ANME-2 archaea have been observed in tight association with SRB of the Desulfosarcina/Desulfococcus group, ANME-1 archaea sometimes occur with these SRB, but at other times are observed as monospecific aggregations or isolated filaments (Orphan et al., 2002). In 5 anoxic marine sediments, carbonate crusts, and recent microbial mats from cold seep sites, methanotrophic consortia can be traced using specific, strongly 13 C-depleted biomarkers. Different species of methanotrophic archaea are considered to be the sources of characteristic isoprenoids (Hinrichs et al., 1999;Elvert et al, 2005;Blumenberg et al., 2004;Pape et al., 2005). These isoprenoids include C 20 and C 25 irregular 10 isoprenoid hydrocarbons (crocetane and 2,6,10,15,19-pentamethylicosane and unsaturated derivatives), the glycerol diethers archaeol and sn-2-hydroxyarchaeol (2,3-di-Ophytanyl-sn-glycerol and 2-O-3-hydroxyphytanyl-3-O-phytanyl-sn-glycerol), and glycerol dialkyl glycerol tetraethers (GDGT) carrying two C 40 isopranyl moieties. Nonisoprenoid monoalkylglycerolethers (MAGEs), 1,2-dialkylglycerolethers (DAGEs), and 15 C 14 to C 18 n-, i soand antei so-fatty acids, and alcohols found at cold-seeps have commonly been regarded as biomarkers for associated SRB (Pancost et al., 2001a;Hinrichs et al., 2000). The cold seep sites at Hydrate Ridge, located about 90 km off the coast of Oregon (USA) at 600 m to 800 m water depth, have been extensively studied since the 20 mid-1980s. Different seep carbonate lithologies have been the targets of several investigations (Kulm and Suess, 1990;Ritger et al., 1987;Bohrmann et al., 1998;Greinert et al., 2001). Generally, authigenic carbonates from Hydrate Ridge consist primarily of aragonite (Greinert et al., 2001 EGU various components, namely shell fragments, pellets containing pyrite, peloids and detrital quartz, and feldspar grains (Teichert et al., 2005). In order to study the linkage of these phases to particular microorganisms and/or biogeochemical processes, we used a miniaturized biomarker technique, combined with factor analysis, and subsequent electron microprobe analyses. The aim was to iden-5 tify differences in the lipid biomarker patterns and the chemical compositions between the phases that would allow us to understand the mechanisms involved in forming the particular carbonate precipitates.

Material and methods
Sample collection -The samples were obtained from a carbonate block collected dur- Sample preparation -From a 28.5 cm (length) by 50 mm (diameter) core drilled from sample TVG 13, 18 micro drill cores (<2 mm long, 2 mm in diameter, 6-21 mg in weight) were taken using a diamond-studded hollow drill. Microscope observations allowed us to classify the micro-drill cores into three distinct phases according to Teichert et  EGU 2 µL of n-hexane were added. 1 µL of each extract was analyzed in a coupled gas chromatograph mass spectrometer (GC/MS). GC/MS -The GC/MS system used was a Varian CP-3800 GC coupled to a Varian 1200 quadrupole MS operated in electron impact mode at 70 eV. The samples were injected on-column into a fused silica capillary column (Phenomenex ZB-1; 30 m; 5 0.25 mm; 0.1 µm film thickness). In the injector, the samples were heated from 50 • C (0.2 min isothermal) to 290 • C at 150 • C/min (5 min isothermal). The GC-oven was programmed from 50 • C (1 min isothermal) to 300 • C at 10 • C/min, and was held at 300 • C for 15 min. Helium was used as the carrier gas at a flow rate of 1.4 mL. Compounds were identified by comparison with published mass spectral data.

10
Factor analysis -A factor analysis was implemented using Statistica 6.0, developed by Statsoft Inc, Tulsa. The compound concentrations were treated as multivariate to show correlations of the compounds with each other (biomarker families). Absolute concentrations (in µg/g carbonate weighted samples) were used as base data. Factors were extracted by Principle Component Analyses (PCA). The maximum number 15 of factors to be extracted was determined using the scree test (Cattel, 1966a). The rotational strategy was varimax normalized (Kaiser, 1958(Kaiser, , 1959. Electron microprobe analysis -Polished thin sections (250 µm thickness) were prepared from sampled areas of the carbonate. Element distributions of Mg, S, Mn, Fe, Sr, (wavelength dispersive system) and Ca (energy dispersive system) were mapped 20 using a JEOL JXA 8900 RL electron microprobe. The acceleration voltage was set to 15 kV and a beam current of 40 nA, measured by Faraday cup, was used. The acquisition time was set to 70 ms per step. The scan grid was spaced at 20 or 40 µm steps, depending on the dimension of each area, resulting in total dimensions between 7×5 and 10×20 mm. The backscatter signal in composition mode and the cathodoluminescence signal (integrated spectral range from 200 to 900 nm) were acquired simultaneously. Since carbonates have sensitive behavior under electron bombardment, the beam diameter was set to 20 µm.

Results
Biomarkers -The whitish aragonite samples showed the highest lipid biomarker concentrations, containing more than 90% of the total AOM-related lipid signature observed (Table 1, Fig. 1). In all whitish aragonite samples (n=8), archaeol, sn-2hydroxyarchaeol and DAGEs were the most prominent lipid biomarkers (Table 1). In  (Table 1). Unlike the whitish aragonite, the lucent aragonite samples (n=6) contained only trace amounts of lipid biomarkers (Table 1). n-alkanes (n-C 23 to n-C 31 ), n-fatty acids (n-C 14 15 to n-C 18 ), squalene, and sterols, specifically cholesterol and sitosterol, dominate the patterns. In two out of six samples, PMI was detected, whereas phytanol and sn-2hydroxyarchaeol were observed in only one sample each. Archaeol was found in three samples, while crocetane and DAGEs were generally below the detection limit in the lucent aragonite (Table 1) EGU samples analyzed. Factor analysis -The two factors extracted accounted for 44.9% and 24.6% of the total variance. The factor loadings plot revealed a compound group consisting of PMI, crocetane, DAGE IIa, DAGE If, archaeol, sn2-hydroxyarchaeol, and phytanol, which has slightly negative loadings with factor 1 and highly positive loadings with factor 2 5 (Fig. 2). A second group of compounds that loads positive with factor 1 and slightly negative with factor 2 included n-alkanes, n-fatty acids, and sterols. One compound, n-tricosane (n-C 23 ), plotted between the two compound groups.
Electron microprobe analyses -The electron microprobe data showed that the whitish aragonite was considerably enriched in Sr compared to the lucent aragonite ( Fig. 3h). Both aragonite phases nevertheless revealed higher Sr concentrations than the gray micrite (Fig. 3f). Ca was somewhat more abundant in the lucent aragonite than in the whitish aragonite and the gray micrite; Mn was not observed in any of the phases. Fe, Mg, and S were detected in the gray micrite, but they were below detection limit in the aragonites. In the gray micrite, distributions of Mg and Ca were anticorrelating 15 (Fig. 3c, e). S and Fe, on the other hand, spatially correlated (Fig. 3b, d).

Discussion
The strong correlation between the concentrations of PMI, crocetane, DAGE IIa, DAGE If, archaeol, and sn-2-hydroxyarchaeol suggests that these AOM-related biomarkers originate from a closely associated biological source. Blumenberg et al. (2004) pro-20 posed high proportions of sn-2-hydroxyarchaeol vs. archaeol, and the presence of crocetane, as traits to distinguish microbial consortia dominated by ANME-2 vs. ANME-1. Concentrations of these compounds are highest in the whitish aragonite. Here, the sn-2-hydroxyarchaeol/archaeol ratios range from 0.48 to 2.13 (Table 1). This spread can be interpreted in terms of varying contributions of ANME-1 vs. ANME-2 archaea, respectively, which were both observed in sediments from Hydrate Ridge (Elvert et al., 2005;Knittel et al., 2003). However, unlike archaeol, sn-2-hydroxyarchaeol is rarely EGU present in the fossil record (Peckmann and Thiel, 2004), indicating its preferential diagenetic degradation, or even conversion to archaeol by dehydroxylation of the phytyl moiety. Therefore, the ratios of sn-2-hydroxyarchaeol/archaeol of Hydrate Ridge material must be interpreted cautiously, especially when comparing with data from recent microbial consortia. As the studied carbonates are several ten thousand years old, the 5 original abundance of sn-2-hydroxyarchaeol may have been considerably higher than it is now. Furthermore, taking into account the prominent occurrence of crocetane found exclusively in whitish aragonite samples ( ANME-1 dominated while the numbers of SRB-cells were markedly low. Although the exact source organisms in these systems are as yet unclear, the high abundances of DAGEs in the whitish aragonite and the strong correlation with archaeal isoprenoid biomarkers clearly imply an origin from within the consortia involved in AOM. The traces of lipid biomarkers in the lucent aragonite did not show any specific pat-20 tern (Fig. 1). Considering (i) the low sample amounts used, (ii) the low compound concentrations, and (iii) the absence of a characteristic biomarker pattern, contamination from the other carbonate phases during sample preparation is a conceivable source for the lipids observed in the lucent aragonite samples. Thus, it seems unlikely that there is direct involvement of particular AOM-related (and other) microorganisms 25 in the precipitation of the lucent aragonite, as proposed for the whitish aragonite.
In the gray micrite, abundant Mg reflects a partly Mg-calcitic mineralogy, corresponding to micrites described at another SE-Knoll location (Teichert et al., 2005). Furthermore, the similarity of distributions of Fe and S in the gray micrite indicates likely EGU pyrite occurrence in these carbonates (Fig. 3b, d; see also Teichert et al., 2005). The gray micrite contained biomarker compounds from both compound clusters revealed by factor analysis (Fig. 2). The presence of PMI, archaeol, and DAGEs, together with the conspicuous absence of sn-2-hydroxyarchaeol and the very low amounts of crocetane, suggests that ANME-1 archaea are involved in the formation of the gray micrite, rather than ANME-2 archaea. On the other hand, factor analysis suggests that long-chain n-alkanes, conventional sterols (sitosterol, cholesterol), and n-fatty acids represent water-column-sourced contributions rather than AOM-derived compounds.
In this context, it is interesting that the intermediate position of n-tricosane between the two compound clusters (Fig. 2) corresponds with a dual, partly AOM-related origin 10 of this hydrocarbon (Thiel et al., 2001). Perylene, which is thought to originate from both terrestrial and aquatic organic matter during diagenesis (Silliman et al., 2000), is presumably derived from allochthonous sources. These combined findings are interpreted to reflect incorporation of allochthonous organic and inorganic matter during AOM-induced carbonate precipitation resulting in the formation of the gray micrite.

Conclusions
Combining miniaturized lipid biomarker analysis and electron microprobe analysis allowed us to resolve biosignatures of a complex microbialite at the mm-scale, and allowed us to develop a model for the origin of distinct carbonate phases. The results showed a highly localized distribution of lipid biomarkers within the Hydrate Ridge car-20 bonates. More than 90% of the AOM-related lipid signature was concentrated in only about 20% of the total carbonate rock volume, specifically in a whitish aragonite phase. The biomarker and inorganic patterns of the whitish aragonite were highly specific and indicated an association with methanotrophic consortia containing ANME-2 archaea and sulfate-reducing bacteria. We suggest that the whitish aragonite formed during periodic methane-rich fluid pulses that disrupted the sediment and led to the growth of the respective microorganisms along fluid pathways. By contrast, low amounts of EGU Peckmann, J. and Thiel, V.: Carbon cycling at ancient methane-seeps, Chem. Geol., 205, 433-467, 2004. Ritger, S., Carson, B., and Suess, E.: Methane-derived authigenic carbonates formed by subduction-induced pore-water expulsion along the Oregon/Washington margin, Geol. Soc. Am. Bull., 98, 147-156, 1987.