Stratification of archaeal membrane lipids in the

Environmental Microbiology (2016) 18(12), 4324–4336

doi:10.1111/1462-2920.13289

Stratification of archaeal membrane lipids in the ocean and implications for adaptation and chemotaxonomy of planktonic archaea

Chun Zhu,1,2*† Stuart G. Wakeham,3 Felix J. Elling,1 Andreas Basse,1,4 Gesine Mollenhauer,1,4 € nneke1 and Gerard J. M. Versteegh,1 Martin Ko 1 Kai-Uwe Hinrichs 1 MARUM Center for Marine Environmental Sciences and Department of Geosciences, University of Bremen, D-28359 Bremen, Germany. 2 School of Earth and Ocean Sciences, Cardiff University, Cardiff, CF10 3AT, UK. 3 Skidaway Institute of Oceanography, 10 Ocean Science Circle, Savannah, GA 31411, USA. 4 Alfred-Wegener-Institute for Polar and Marine Research (AWI), Bremerhaven, Germany.

Summary Membrane lipids of marine planktonic archaea have provided unique insights into archaeal ecology and paleoceanography. However, past studies of archaeal lipids in suspended particulate matter (SPM) and sediments mainly focused on a small class of fully saturated glycerol dibiphytanyl glycerol tetraether (GDGT) homologues identified decades ago. The apparent low structural diversity of GDGTs is in strong contrast to the high diversity of metabolism and taxonomy among planktonic archaea. Furthermore, adaptation of archaeal lipids in the deep ocean remains poorly constrained. We report the archaeal lipidome in SPM from diverse oceanic regimes. We extend the known inventory of planktonic archaeal lipids to include numerous unsaturated archaeal ether lipids (uns-AELs). We further reveal (i) different thermal regulations and polar headgroup compositions of membrane lipids between the epipelagic ( 100 m) and deep (>100 m) populations of archaea, (ii) stratification of unsaturated GDGTs with varying redox conditions, and (iii) enrichment of tetraReceived 8 December, 2015; revised 29 February, 2016; accepted 3 March, 2016. *For correspondence. E-mail chun.zhu@exxonmobil. com. Tel: 1 832 728 7787; Fax: 1 832 625 2680. †Present address: ExxonMobil Upstream Research Company, 22777 Springwoods Village Parkway, Spring, TX 77389, USA. C 2016 Society for Applied Microbiology and John Wiley & Sons Ltd V

unsaturated archaeol and fully saturated GDGTs in epipelagic and deep oxygenated waters, respectively. Such stratified lipid patterns are consistent with the typical distribution of archaeal phylotypes in marine environments. We, thus, provide an ecological context for GDGT-based paleoclimatology and bring about the potential use of uns-AELs as biomarkers for planktonic Euryarchaeota.

Introduction Distinct clades of the domain Archaea are ubiquitous in the ocean and form significant fractions of the marine planktonic microbial community. Two major groups of planktonic archaea affiliated with the phyla Thaumarchaeota and Euryarchaeota have been frequently found based on their 16S rRNA gene sequences (Delong, 1992; Fuhrman et al., 1992). Typically, Euryarchaeota are more abundant in epipelagic waters whereas Thaumarchaeota are more important below the photic zone, numerically dominating the prokaryotic group in meso-bathypelagic and bathypelagic waters (Massana et al., 2000; Karner et al., 2001; Herndl et al., 2005). While a few species of marine Thaumarch€nneke aeota have been cultured from epipelagic waters (Ko et al., 2005; Qin et al., 2014; Elling et al., 2015; Bayer et al., 2015), cultivated representatives of the planktonic Euryarchaeota are still lacking. To date, all thaumarchaeal cultures are aerobic ammonia-oxidizers that gain energy by chemoli€nneke et al., 2005) and thotrophic catabolism (Ko autotrophic or obligate mixotrophic anabolism (Qin et al., 2014). Thaumarchaeotal ammonia-oxidizers fall into the ‘shallow’ and ‘deep’ phylogenetic clades with their interface located at the bottom of the euphotic zone (Francis et al., 2005; Hallam et al., 2006; Hu et al., 2011). Single-cell genomic data have revealed that putative DNA photolyase and catalase genes, adaptive mechanisms for reducing light-induced damage, were found exclusively in members of the shallow clade (Luo et al., 2013). On the other hand, single cell- and metagenomic approaches suggest that some Euryarchaeota are photo-heterotrophs and reside exclusively in the euphotic zone (Frigaard et al., 2006; Iverson et al., 2012). In contrast, genes potentially encoding for

Stratification of archaeal membrane lipids 4325

IPLs

HO

HO

O

HO HO

O

GDGTs

OH

O

HO

HO

O

HO

O

OH HO HO

1G-GDGTs

OH

O

O

HO

GDGTs

HO

O

OH

1G-OH-GDGTs

2G-GDGTs HO HO

OH

HO HO

OH

O OP O OH

HO O

GDGTs

HO

O

HO

O

OH-GDGTs

OH

HO

O

OH HO

OH

O

HO

O

OH

OH-GDGTs

2G-GDGTs

HPH-GDGTs

CLs R

HO

O

O

O

O

OH

O

OH

R

O HO

O

O

GDGT0 or OH-GDGT0

GDGT1 or OH-GDGT1

R

OH

O O

O

OH

O

O HO

O

O O

HO

GDGT3

GDGT2 or OH-GDGT2 O

O

OH

O

O HO

O

HO

O O

O

Cren' (GDGT5')

Cren (GDGT5) BP0:1 O O BP0:2 O

HO

OH

GDGT0:3

P0:2 O P0:2 O

O HO

OH

HO

AR0:4

O

R

O

AR or OH-AR

Fig. 1. Structures of archaeal IPLs associated with different headgroups and CLs with different numbers of cycloalkyl, hydroxyl groups, and double bonds on the alkyl moieties. 1G 5 monoglycosyl; 2G 5 diglycosyl; HPH 5 hexosephosphohexose; R 5 –H or –OH; GDGT0-3 5 glycerol dialkyl glycerol tetraethers with 0–3 cyclopentyl rings, respectively; Cren (GDGT5) 5 crenarchaeol; Cren’ (GDGT5’) 5 regioisomer of crenarchaeol; AR 5 archaeol. OH-GDGT0-2 5 monohydroxylated GDGTs with 0-2 rings, respectively; BP0:1 5 acyclic biphytene; BP0:2 5 acyclic biphytadiene; P0:2 5 acyclic phytadiene; saturated GDGTs are expressed as GDGTn, and unsaturated GDGTs and AR are expressed as GDGTm:n and ARm:n, respectively, where m 5 the number of cycloalkyl moieties, and n 5 the number of double bonds. Note that the exact double bond positions remain unknown.

anaerobic respiratory chains have been detected in Euryarchaeota inhabiting deep waters (500 m) (Moreira et al., 2004; Martin-Cuadrado et al., 2008). These physiological capacities may provide Euryarchaeota competitive advantages in occupying the euphotic and in anoxic zones. These genetic surveys implied that redox and light conditions are important for shaping the distribution of archaeal ecophylotypes and phylotypes. In addition to genetic markers, microbial membrane lipids are commonly used to study the distribution and ecology of planktonic microbes. Lipid compositions also encode adaptive strategies of the source organisms. To

date, studies of planktonic archaeal lipids have primarily targeted the fully saturated glycerol dibiphytanyl glycerol tetraethers (GDGTs) with zero to three cyclopentyl moieties (GDGT0–3) as well as GDGTs with four cyclopentyl and one cyclohexyl moieties known as crenarchaeol (Cren) and its isomer (Cren0 ) as putatively Thaumarch et al., aeota-specific lipids (Fig. 1; Sinninghe Damste 2002). However, a recent study suggests that Euryarchaeota are also capable of producing Cren and Cren0 (Lincoln et al., 2014), complicating the chemotaxonomic signature of thaumarchaeotal and euryarchaeotal lipids. Euryarchaeota-specific lipids have not been identified yet

C 2016 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 18, 4324–4336 V

4326 C. Zhu et al. and the relatively low diversity of known GDGT homologues is in strong contrast to the rapidly increasing number of phylotypes (Francis et al., 2005; Hu et al., 2011; Luo et al., 2013) and metabolisms (Moreira et al., 2004; Hallam et al., 2006; Martin-Cuadrado et al., 2008) among planktonic archaea as revealed by genetic surveys. GDGT compositions in the epipelagic ocean (100 m) and in underlying sediments correlate with in situ water temperatures and with annual mean sea surface temperatures (SSTs) respectively (Schouten et al., 2002; Schouten et al., 2013). The correlations suggest thermal adaption of GDGT lipids and have resulted in the establishment of TEX86 (Schouten et al., 2002), a proxy quantifying the relative distribution of fossil GDGTs preserved in marine sediments for reconstruction of past SSTs. However, in oceans below 100 m water depth, relationships between GDGT distribution and in situ temperature have not been found (Schouten et al., 2002; Wuchter et al., 2005b; Turich et al., 2007; Schouten et al., 2012; Basse et al., 2014; Xie et al., 2014). This observation has triggered debates on the extent to which GDGT compositions in deep waters might be regulated by processes (e.g. salinity changes, nutrient fluctuations, variation of major phylotypes with depth, oxygen concentration) different from those in epipelagic waters (Ingalls et al., 2006; Turich et al., 2007; Taylor et al., 2013). However, due to the strong interference by GDGTs descending from overlying waters (so-called “fossil GDGTs”) on the in situ signals generated by archaea living at deeper depths (Schouten et al., 2012), geochemistryassociated trends in adaptation of archaeal lipids below 100 m depth remain poorly characterized. The low diversity of known planktonic archaeal lipids and poorly characteried of lipid compositions of the deep waterdwelling population(s) partially result from the paucity of representative cultures. Furthermore, analytical protocols established more than a decade ago did not yield comprehensive characterization of archaeal lipid distributions (Hopmans et al., 2000; Sturt et al., 2004); as such, important lipid series and distribution patterns were not recognized until recently (e.g. Basse et al., 2014; Zhu et al., 2014). Here, we characterize the lipidome of planktonic archaea from the oceans’ epipelagic to abyssal depths using a recently developed analytical protocol (Zhu et al., 2013). Our study unravels stratified lipid signatures of uncultivated archaea across physicochemically distinct habitats, providing new insights into structural diversity, adaptive patterns, and chemotaxonomic potential of planktonic archaeal lipids. Results and discussion Structural diversity and compositional variability of archaeal lipids in the ocean The inventory of glycerol-based membrane lipids of planktonic archaea from suspended particulate matter (SPM)

samples consists of intact polar lipid (IPLs) and their natural degradation products, core lipids (CLs) derived from hydrolytic cleavage of polar headgroups of IPLs. We detected five other types/classes of CLs in addition to the GDGT0–3,cren,cren0 series (Fig. 1). These lipids are archaeol (AR), monohydroxylated ARs (OH-ARs), monohydroxylated GDGTs (OH-GDGTs; (Liu et al., 2012), unsaturated acyclic ARs (uns-ARs) with 1–8 double bonds (AR0:[1–8]), and unsaturated acyclic GDGTs (uns-GDGTs; Zhu et al., 2014) bearing 1–8 double bonds (GDGT0:[1–8]). Unsaturated archaeal ether lipids (including uns-GDGTs and unsARs) have not been detected previously in SPM samples, although they have been reported in a few cultivated extremophiles and methane seep sediments (Franzmann et al., 1988; Gonthier et al., 2001; Nishihara et al., 2002; Nichols et al., 2004; Dawson et al 2012; Zhu et al., 2013; Yoshinaga et al., 2015). Zhu et al., (2013) found that unsGDGTs coelute with their cycloalkylated counterparts in commonly used normal phase HPLC-MS protocols, and therefore, they may have been overlooked previously in studies of mesophilic planktonic archaeal lipids. GDGT-based IPLs comprise six classes (Fig. 1): monoglycosidic-GDGTs (1G-GDGTs), diglycosidic-GDGTs (2G-GDGTs), 2G-monohydroxylated-GDGTs (2G-OHGDGTs), hexosephosphohexose-GDGTs (HPH-GDGTs), traces of 1G-OH-GDGTs, and traces of triglycosidicGDGTs (3G-GDGTs). The associated headgroup types are in line with those found in Nitrosopumilus maritimus (Schouten et al., 2008; Elling et al., 2014), and cultures of other mesophilic Thaumarchaeota (Pitcher et al., 2011; Elling et al., 2015). The IPLs forms of unsaturated archaeal ether lipids were not detected in SPM. A principal component analysis (PCA) was performed in order to identify relationships between compositional variability of archaeal lipids and environmental conditions (Fig. 2). This analysis shows that the first three principal factors F1, F2, and F3 represent 29.1%, 24.5% and 18.1% of the total variability in the data set respectively, and account for 72% of variance collectively. We discuss lipid distribution, in particular those showing high loadings on F1-3, in context with environmental conditions in the following sections. Patterns of lipid adaptation to temperature and light conditions In the PCA, F1 displays the highest absolute loadings of ring index (quantifying the cyclization degree; see Experimental procedures), hydroxylation index (quantifying the hydroxylation degree; see Experimental procedures), and %OH-GDGTs followed by %1G-GDGTs and %2G-GDGTs for lipid variables whereas the environmental variables temperature, salinity, redox and light are projected to moderate loadings (Fig. 2a). This implies covariation of these lipid-based variables and environmental conditions.



Fig. 2. A PCA of lipid-based (grey) and environmental (dark blue) variables across sampling sites. UI 5 unsaturation index; HI 5 hydroxylation index; RI 5 ring index. For lipid abbreviations see Fig. 1.

Indeed, ring indices of GDGTs in epipelagic waters (100 m depth) correlate with in situ temperature (Fig. 3a). Such correlations are consistent with the regulation of membrane lipid fluidity in epipelagic archaea via adjusting the cyclization degree of the core moieties in response to temperature changes (Schouten et al., 2002; Wuchter et al., 2005; Elling et al., 2015). However, correlations break down in deep (>100 m) waters (Fig. 3a). Previous investigations, which also found such correlations limited to waters shallower than 100 m, concluded that fossil CLs biosynthesized in, and descending from, overlying epipelagic waters interfered with the in situ signals generated by living archaea at greater depths (Schouten et al., 2012, 2013). Our data support the former conclusion and consequently, the specific lipid profiles of deep-dwelling archaea and their adaptive response to temperature in the deep ocean remain unconstrained. Reactive headgroups of IPLs are enzymatically cleaved off after cell death to produce more stable CLs. IPLs,there-

fore, provide a better approach than CL counterparts to study live archaea in situ by minimizing interferences from extracellular fossil CL residues. Regressing ring indices of 2G-GDGTs with in situ temperature yields two types of correlations that are best separated by the 100 m water depth (Fig. 3b) but that are independent of other environmental variables such as geographical location, redox condition (note that 2G-GDGTs are undetectable in deep sulfidic waters of the Black Sea and Cariaco Basin), and salinity (Fig. 4a–c). Consistent with 2G-GDGTs, ring indices of 2G-OH-GDGTs from epipelagic and deep also show two different types of correlations with in situ temperature (Fig. 3c). However, ring indices of 1G-GDGTs and HPHGDGTs only correlate weakly (R2 < 0.3) with temperature throughout the water column and without depth-related pattern differentiations (Supporting Information fig. S1d,e). Despite 1G-GDGTs being the major lipid group of the marine thaumarchaeon N. maritimus (Elling et al., 2014), they may also be formed as diagenetic intermediates of more functionalized IPLs (e.g. 2G-GDGTs and HPHGDGTs), compromising their use as suitable biomarkers for physiologically active archaea. Similar to the regulation of GDGT cyclization degree with temperature, planktonic archaeal communities also appear to adjust the hydroxylation degree of membrane lipids (quantified by the hydroxylation index) with ambient € et al., 2015). While temperature (Huguet et al., 2013; Lu correlation between hydroxylation indices of OH-GDGTs (CLs) and temperature is only found in epipelagic waters, correlation between hydroxylation indices of 2G-OHGDGTs (IPL counterparts) and temperature shows two different patterns separated by the 100 m water depth (Supporting Information fig. S1a,b). Headgroup distribution among GDGTs, that is, the relative proportions of 1G, 2G, and HPH moieties, appears to be independent of temperature, salinity, or redox conditions (Fig. 2) but again, clearly differs between epipelagic and deep lipid populations. For instance, plotting the abundances of 2G-associated IPLs ([2G-GDGTs] 1 [2G-OHGDGTs]) against total IPLs shows two significant correlations (Supporting Information fig. S1c), indicating the deep archaeal population consistently enriches 2G-associated IPLs (ca. 50% of IPLs based on regression) relative to the epipelagic counterpart that has higher fractions of 1Gassociated IPLs ([1G-GDGTs] 1 [1G-OH-GDGTs]; ca. 70% of total IPLs). Whereas archaea are known to regulate their lipid composition in response to ambient temperature, our study suggests that epipelagic (100 m) and deep (>100 m) water-dwelling archaea exhibit distinct patterns of lipid regulation with temperature change. This distinction explains the PCA loadings on F1: both the light condition (epipelagic euphotic vs. deep sub- and aphotic zones) and temperature gradient show significant loadings on F1 with


4328 C. Zhu et al.

Fig. 3. Plots of ring indices (RI) and TEX86 values of different lipid groups against in situ temperatures (a-e; only R2 values > 0.60 with P < 0.01 shown), and a schematic profile of 2G-GDGTs-TEX86 values down a water column in comparison to a hypothetical temperature profile (f). Note, in panel (f) the qualitative dual profiles (blue and red) are based on the dual regressions (blue and red) shown in panel (e), while open circles and dashed black lines represent hypothetical sampling sites and the resulting 2G-GDGTs-TEX86 profile, respectively.

Fig. 4. Plots of ring indices (RI) of different IPLs classes against in situ temperatures. Geographic location (Supporting Information table S1), redox condition, salinity values, and water depth are represented by different colors. C 2016 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 18, 4324–4336 V

Stratification of archaeal membrane lipids 4329 lipid variables such as hydroxylation index, ring index, %OH-GDGTs (comparable to the hydroxylation index; see Experimental procedures), and headgroup composition (i.e., %1G-GDGTs, and %2G-GDGTs; Fig. 2). The positive/negative properties of loading results from their positive/negative correlations among variables or different partitioning of 1G- and 2G-GDGTs between epipelagic and deep waters (Fig. 3 and Supporting Information fig. S1). Redox and salinity show comparable loadings as temperature on F1, and we attribute this to the coincidence of lower temperatures with lower salinity and lower oxygen content in the Black Sea relative to most other settings in this study. Thaumarchaeota have long been considered as the principal producers of GDGT1–3,Cren,Cren0 throughout the oceans except for methane-rich waters in deep anoxic basins where anaerobic oxidation of methane occurs (Wakeham et al., 2004; Schouten et al., 2013). They are phylogenetically grouped into Cluster A and Cluster B dominating the shallow (euphotic) and deep (sub-aphotic and aphotic) zones respectively (Francis et al., 2005; Hallam et al., 2006; Hu et al., 2011). Genes coding for the geranylgeranylglyceryl phosphate synthase, an enzyme catalyzing the formation of the sn23 ether bond of archaeal lipids, also differ between the shallow and deep groups (Villanueva et al., 2014). The widely held notion that Thaumarchaeota dominate the GDGT1–5 biosynthesis has, however, been tempered by a recent study (Lincoln et al., 2014) suggesting that Euryarchaeota provide an additional source of GDGT1–5 in the ocean, especially in epipelagic waters where Thaumarchaeota are barely detectable. Regardless of this debate, GDGT producers in epipelagic vs. deep waters appear to differ genetically and phenotypically, explaining their different lipid patterns. The TEX86 paleothermometer relates the relative distribution of a subgroup of GDGT CLs (GDGT1–3,Cren0 ; structures in Fig. 1) to epipelagic water column temperature, most often taken as the SST, based on an empirical calibrations of core-tops (Schouten et al., 2002; Wuchter et al., 2005, 2006; Kim et al., 2010) and epipelagic particulate matter (Wuchter et al., 2005, 2006). While TEX86 of GDGTs (CL-based TEX86) preserved in sediments is widely used for paleoclimatology, the initial generation of TEX86 signatures from specific IPLs (the biological precursors of CLs; producing IPL-specific TEX86, or IPL-specific TEX86) has not been determined. In this study, we found two types of correlations between 2G-GDGTs-TEX86 values and temperatures (Fig. 3e), consistent with the general dual (epipelagic vs. deep) thermal regulation patterns of IPL-derived GDGTs (Fig. 3b,c). By comparison, CL GDGT-TEX86 shows only one correlation with in situ temperature in waters shallower than 100 m but no relationship at depth (Fig. 3d), as reported previously (Schouten et al., 2013 and refs. therein). If indeed epipela-

gic and deep-dwelling archaeal populations express TEX86 signals differently due to inherently different thermal regulation of IPLs, then temperature may not be the only control on the sedimentary TEX86 values. Caveats and alternative interpretations for the TEX86 proxy are exemplified as follows: i. The relationships between 2G-GDGT-TEX86 and temperature differ significantly above vs. below 100 m depth, resulting in a significantly elevated TEX86 value below ca. 100 m depth for a hypothetical water column setting as shown schematically in Fig. 3f. This depth-related decoupling may help explain the increasing rather than decreasing TEX86 values observed in colder sub-epipelagic waters (Schouten et al., 2012; Basse et al., 2014; Xie et al., 2014). ii. By extension, if in shallow (100 m water depth) settings, such as continental shelves, the deep archaeal component is largely missing, there may be a significant mismatch of IPL-TEX86-derived temperatures between shallow and deep settings with the same actual epipelagic temperatures as shallow settings but harboring a greater amount of deep-dwelling archaea. iii. Likewise, the population of deep ammonia-oxidizing archaea, the principal producers of cyclic tetraether lipids, is greatly diminished in an euxinic basin compared to those thriving in an oxygenated setting. Despite the actual surface water temperatures being the same, the different redox conditions might lead to a mismatch of IPL-TEX86-derived temperatures between euxinic and oxygenated settings. In fact, apparent ‘cold biases’ of TEX86-derived temperatures have been reported in shallow coastal seas (Herfort et al., 2006; Leider et al., 2010; Wei et al., 2011; Zhu et al., 2011), the present-day Black Sea as a typical euxinic basin (Schouten et al., 2002), the Benguela upwelling system (Lee et al., 2008), and the past Mediterranean Sea during the sapropel formation phases (Menzel et al., 2006). Our study adds ecological constraints to the use of fossil TEX86 signals for temperature reconstruction. Patterns of lipid distributions across light and redox gradients The PCA further reveals strong loadings of several unsaturated archaeal ether lipids (including uns-GDGTs and unsARs) with different environmental parameters reflected by F2 and F3 (Fig. 2b), indicating that unsaturated archaeal ether lipids might resolve ecological adaptations of their source producers. Diverse uns-GDGTs and uns-ARs occur as prominent constituents in oxygen-depleted settings such as the oxygen minimum zone of the Eastern


4330 C. Zhu et al. Fig. 5. Partial composite mass chromatograms of saturated and unsaturated archaeal core ether lipids (detected as [M 1 NH4]1) from sites a-d that represent the four main habitat types distinguished in this study, as well as N. maritimus (N. mari.) harvested during the late exponential phase (e) and T. acidophilum (T. acido.) harvested during the stationary phase (f) (cf. Table 1). For AR0:n and GDGT0:n, 0 5 no cycloalkyl moieties, n above the yellow peaks 5 number of double bonds, with isomers of uns-AR0:[2–4] denoted as a and b; GDGT0–5 5 saturated GDGTs with 0–5 cycloalkyl moieties shown above the grey peaks, where the peak 5 5 crenarchaeol (cren) and peak 50 5 the regioisomer of crenarchaeol (cren’); n.d. 5 not detected; 3 n 5 signal response amplified n times. Extracted m/z: 1319.3498, 1317.3342, 1315.3186, 1313.3030, 1311.2874, 1309.2718, 670.7077, 668.6921, 666.6765, 664.6609, 662.6453, 660.6297, 658.6141, 656.5985, 654.5829. The extraction window of individual ion chromatograms is 60.05 m/z units.

Tropical North Pacific (ETNP; Fig. 5a) and the sulfidic Black Sea (BS; Fig. 5b). Uns-GDGTs were generally not found in well-oxygenated waters at all depths (Fig. 6a), except for trace amounts of monounsaturated GDGT (GDGT0:1) detected occasionally. Trace amounts of GDGT0:1 may

derive from dehydration of mono-OH-GDGTs (cf. Liu et al., 2012) that are present in Thaumarchaeota (Elling et al., 2014) and that are abundant in oxic waters in this and previous studies (Huguet et al., 2013). Only mono- and di-unsaturated GDGTs (GDGT0:[1-2]) are detected, as

Fig. 6. Abundances of characteristic archaeal lipids relative to total archaeal CLs. Oxic: O2 10 mM; suboxic: detection limit < O2 < 10 mM; anoxic: O2 below detection limit. Detection limit 1 mM (see Experimental Procedures). C 2016 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 18, 4324–4336 V


Fig. 7. Depth profiles of AR0:4 and crenarchaeol (Cren) concentrations (ng/L) in SPM from representative oxic/suboxic water columns: Station CBi, off Cape Blanc, NW Africa (a), the equatorial Pacific (EQPAC; b), and Station 1, eastern tropical North Pacific (ETNP; c). Details of sampling sites see Supporting Information table S1.

minor lipids (