From organic-rich black shales of Hole 1258B, typically 30 g of wet sediment was extracted. From the resulting highly concentrated polar fractions, aliquots equivalent to an extract of ~5 g of sediment were analyzed by HPLC-MSn (Table T1). These sample concentrations proved to be insufficient for the detection of IPLs (Fig. F2). The organic-rich black shales are particularly challenging for the detection of miniscule quantities of IPLs because high quantities of fossil refractory organic carbon are difficult to separate from the polar lipids and contribute to a highly complex analytical background matrix that possibly causes ion suppression and dilution of analyte ions in the ion trap of the mass spectrometer. Figure F2 shows a representative density map plot from Sample 207-1258B-52R-2, 80–90 cm. After the dark shading in the density map is normalized to the signal intensity, the overall dark color indicates the lack of any compounds eluting above the relatively high background. Determination of prokaryotic cell numbers by catalyzed reporter deposition–fluorescence in situ hybridization (CARD-FISH) (Schippers and Neretin, this volume) suggests very low concentrations of cells, typically below 105 cells/cm3.
Following evidence for diagenetically active sediments being not only in black shales but also in horizons overlying the black shales (Arndt et al., 2006), we selected samples from the respective organic-lean intervals (Table T1). As a result of much more favorable signal-to-noise ratios, we were able to analyze samples as highly concentrated solutions, equivalent to injection of an extract from 69 and 120 g of sediment. IPLs were detected in these two samples and are described in detail below.
Table T2 shows that IPLs were found in both of the large sediment samples extracted. The more deeply buried, older Sample 207-1258B-39R-5, 130–140 cm, contained an order of magnitude less lipids than Sample 207-1257C-5R-2, 130–140 cm (Table T2). Absolute concentrations of these lipids cannot be computed at this point because of the lack of authentic standards, but we can compare the analytical IPL response to that in other environments that contain similar or identical compounds. For example, in deeply buried sediments off Peru typically containing 2 x 106 to 7 x 106 prokaryotic cells/cm3 (total cells hybridizing to either an archaeal or bacterial FISH probe) (Biddle et al., 2006), archaeal IPLs are on average 20 times more abundant than in the two samples from Demerara Rise in which IPLs were detected. Hence, the IPL abundance is consistent with low prokaryotic cell numbers 105 cells/cm3, which is reasonably consistent with the findings of Schippers and Neretin (this volume).
The archaeal lipids found are exclusively glycolipids, with mainly the GDGT core lipid. Related lipids have been detected in deeply buried sediments recovered during Leg 201 off Peru, where parallel phylogenetic analysis based on extractable 16S ribosomal ribonucleic acid suggests that they derive from sedimentary Crenarchaeota (Biddle et al., 2006). The glycolipids found herein contained no cyclopentyl rings within the GDGT macrocycle, a feature more common to the few Euryarchaeota we have studied that contain the GDGT core lipid, though the Euryarchaeota generally have a mixture of glycolipids and phospholipids in their membranes (H.F. Fredricks and K.-U. Hinrichs, unpubl. data). In Sample 207-1257C-5R-2, 130–140 cm (Table T2; Fig. F3), two unusual core lipids were observed that are 14 Da larger than the usual di-phytanyl archaeol and the di-biphytanyl GDGT core lipid. These so-called "DEG + 14-diglycosyl" and "GDGT + 14-diglycosyl" lipids are clearly identified as archaeol- and GDGT-like IPLs by their MSn signals, but how the extra 14 Da is incorporated into the core lipid is unclear; the slightly shorter retention time (i.e., corresponding to a lower polarity) is consistent with an additional (nonpolar) CH2 unit, though this has not been observed in any archaea to our knowledge. No nonitol-based GDGT core lipids were observed as found in various Leg 201 sediments (Biddle et al., 2006; Sturt et al., 2004) and observed members of the crenarchaeotal order Sulfolobales (Langworthy et al., 1974; Sugai et al., 1995).
The small range of bacterial lipids found is intriguing. Phosphatidylcholine (PC) lipids are relatively rare among bacteria. PC lipid is more common in Eukaryotes, where it is usually the most abundant lipid (e.g., Raetz, 1986). However, it is present in about 10% of bacterial species (Sohlenkamp et al., 2003) and observed in a significant number of proteobacterial isolates from the deep subsurface (Schubotz, 2005). Phosphatidyldimethylethanolamine (PDME) is tentatively identified based on the neutral loss of 169 Da in a fragmentation pattern common to most phospholipids. It is observed in Sample 207-1257C-5R-2, 130–140 cm, and because it contains the same core lipids as the PC species found, we suggest that it is derived from organisms similar to or related to those containing the PC species. PDME is reported as an intermediate in the biosynthesis of PC via methylation of phosphatidylethanolamine (Sohlenkamp et al., 2003) but is relatively rare in bacteria. Fang et al. (2000) reported PDME in barophilic bacteria from the Marianas Trench, along with PC and several other phospholipids.
The pattern and abundance of lipids observed in these Leg 207 Demerara Rise sediments is different from those of the Peru margin (Biddle et al., 2006; Sturt et al., 2004). Although no bacterial lipids were observed in the Peru margin sediments, the seemingly ubiquitous diglycosyl-GDGTs were observed in both locations. The Peru margin showed at least an order of magnitude higher abundance of lipids. This is consistent with higher microbial activity and cell concentration in younger Peru margin sediments containing organic material of higher reactivity.