Crushed core samples stored in falcon tubes at -80 °C were first milled for 10 min to a fine powder and subsequently extracted with a modified Bligh and Dyer method after Sturt et al. (2004). Prior to milling and extraction of each sample a procedure blank was performed. First a milling blank was performed using combusted sea sand (fired at 450 °C for 5 hrs) to clean the mill and to limit cross-contamination of samples. Subsequently, this sea sand was then transferred to geo-cleaned (rinsed three times with a mixture of methanol, MeOH and dichloromethane, DCM) Teflon® containers used for extraction of the samples and solvent-extracted in the same manner as the samples. For this, 100 ng of an internal standard (C46 GTGT) and ca. 50 mL of a solvent mixture of DCM:MeOH:buffer (2:1:0.8, v/v) was added to the sample in the Teflon® container and ultrasonicated for 10 mins using a geo-cleaned ultrasonic stick.

After ultrasonication, the samples were centrifuged (1750 rpm at 10 min) and the supernatant was transferred to a fired separatory funnel. The samples were extracted in four steps, for the first two steps a phosphate buffer (K2HPO4, 50 mM at pH 7.4) was used, in the second step the phosphate buffer was replaced by 5 was used, in the second step the phosphate buffer was replaced by 5 % trichloroacetic acid (50 g L-1 at pH 2), and in the last step only DCM:MeOH (9:1, v/v) was used. Equal amounts of DCM and deionized MilliQ water were added to the extract collected in the separatory funnel, the mixture was shaken, and the organic phase was collected as the total lipid extract (TLE) and blown to dryness under a gentle stream of nitrogen.

An aliquot of the TLE was analyzed via ultra-high-pressure liquid chromatography (UHPLC) coupled to mass spectrometry (MS) on a Dionex Ultimate 3000RS UHPLC connected to an ABSciEX QTRAP4500 Triple Quadrupole/Ion Trap MS (UHPLC-Triple Quad-MS) via a Turbolon electrospray ion source (ESI). Separation of compounds was achieved on a Waters Acquity BEH C18 column (1.7 μm, 2.1x150 mm) equipped with a guard column of the same material following the protocol described in Klein et al. (2015). Compounds of interest were screened for by using multiple reaction monitoring (MRM) and selected ion monitoring (SIM) techniques after Klein et al. (2015). Concentrations of lipids were determined relative to the internal C46 GTGT standard and were corrected for individual response factors using commercially available standards (diC16-DEG, archaeol) and isolated standards from cultures (GDGT-0, 1G-AR, 2G522AR, 1G-GDGT-0, 2G-GDGT-0). The presence of crenarchaeol was confirmed by core GDGT analysis after Becker et al. (2013).

Briefly an aliquot of the TLE was analysed on Dionex Ultimate 3000RS UHPLC connected to a Bruker maXis ultra-high resolution quadrupole time-of-flight mass spectrometer, equipped with an APCI II source. Compounds were separated using two aquity BEH HILIC amide columns (1.7 μm, 2.1x300 mm) in tandem maintained at 50 °C, and n-hexane as eluent A and n528 hexane:isopropanol, 90:10, v:v as eluent B (REF). Drilling mud and extraction blank contamination controls were also run for lipid biomarker analyses.