Isoprenoid Quinones Resolve the Stratification of Redox Processes in a Biogeochemical Continuum from the Photic Zone to Deep Anoxic Sediments of the Black Sea

Geochemical data revealed a strong vertical stratification of the Black Sea water column (Fig. 3). The salinity increased from 17.7 at the surface to 22.3 in deep waters. The steepest increase occurred in a discrete zone between 80 and 150 m below sea level (mbsl). Similarly, temperature increased with depth from 8.5 to 9°C. A fluorescence minimum was detected at 70 mbsl. Dissolved oxygen concentrations decreased in a narrow depth interval between 70 and 150 mbsl from >250 μmol kg⁻¹ to below detection. Hydrogen sulfide (HS⁻) was first detected at 100 m and slightly increased in concentration to 11.5 μmol liter⁻¹ at ca. 1,100 mbsl. Both dissolved phosphate (PO₄³⁻) and ammonium (NH₄⁺) were only detectable below 120 mbsl. PO₄³⁻ showed a distinct peak at 130 m followed by an increase from 300 mbsl into deeper waters, while NH₄⁺ concentrations continuously increased toward the seafloor. Total Fe concentrations were relatively high in the uppermost water sample (0.31 μmol liter⁻¹ at 17 mbsl), decreased rapidly below this depth, showed a distinct peak at 111 mbsl (1.35 μmol liter⁻¹), and were generally high in the deeper waters. Although our analysis on water column samples could not distinguish between different valence states of iron, previous studies demonstrated that Fe³⁺ and Fe²⁺ are predominant in the oxic and anoxic zones, respectively (56, 57).

Based on the geochemistry, we defined the zonation of the water column at site GeoB15105 as an oxic zone from 0 to 70 mbsl, suboxic zone (chemocline; cooccurrence of HS⁻ and O₂) from 70 to 150 mbsl, and anoxic zone from 150 mbsl to the seafloor (∼1,200 m).

The pore water profiles of dissolved PO₄³⁻ and NH₄⁺ followed similar trends, with increasing concentrations within the first 250 cm below seafloor (cmbsf). Whereas NH₄⁺ concentrations stayed fairly constant below 200 cmbsf, PO₄³⁻ slightly decreased toward the deepest investigated sample (815 cmbsf). Dissolved Fe²⁺ concentrations were below detection from 0 to 600 cmbsf and strongly increased from 600 to 800 cmbsf. Dissolved sulfate (SO₄²⁻) concentrations decreased rapidly within the upper 90 cmbsf but remained above the detection limit down to 400 cmbsf. HS⁻ concentrations showed a maximum at approximately 100 cmbsf. Below 400 cmbsf, HS⁻ concentrations were close to zero. Methane concentrations were low (<75 μM) in the top 30 cm of the sediment but increased to almost 15.4 mM at 623 cmbsf. The δ¹³C_CH4 significantly decreased from −53‰ to −88.5‰ within the top 127 cmbsf. Below this depth, values increased to −69‰ at 815 cmbsf. Based on these results, the sulfate-methane transition zone (SMTZ) was assumed to cover the depth range from 80 to 400 cmbsf, with the core of the SMTZ at around 100 cmbsf.

The 8-m-long sediment core contains two major sedimentary units that have been recognized throughout the Black Sea basin (58–60). Unit I comprises laminated coccolith ooze (∼1 to 3 wt% total organic carbon [TOC]) (Fig. 3E) deposited during the last 3 ka, while Unit II consists of organic-lean lacustrine deposits (Unit II, ∼0.5 to 1 wt% TOC) overlain by an organic-rich sapropel (∼4 wt% TOC) deposited from 7.5 to 3 ka before present (58, 59, 61), coeval to the development of water column anoxia in the Black Sea (60, 62).

Quinones detected in Black Sea suspended particulate matter and sediment samples comprised 43 different compounds, including vitamin K₁, PQ_9:9, chlorobiumquinone (ChQ_7:7), polyunsaturated UQs, and MKs with variable chain lengths and degrees of unsaturation, fully saturated and monounsaturated MK₆, as well as the functional quinone analogs MP_5:3 and MP_5:4 (see Fig. S1 in the supplemental material). ChQ_7:7 was identified by accurate molecular mass and isotope pattern match of its proposed elemental formula in full scan mode, MS² fragment spectra (Fig. S2), and comparison of retention time and fragmentation pattern with ChQ_7:7 in extracts of an enrichment culture of the green sulfur bacterium Chlorobium phaeobacteroides BS1. We tentatively identified several novel UQ and MK isomers based on accurate molecular mass in full scan (MS¹) mode and characteristic fragmentation in MS² mode (Fig. S1). These isomers do not represent variable double-bond positions, since they were only detected for fully unsaturated (i.e., one unsaturation per isoprene unit) quinones. The quinones PQ_9:9 and UQ_7:7 to UQ_10:10 each showed one early [Q_m:n(a)] and one late [Q_m:n(b)] eluting isomer with highly similar MS² fragmentation. Based on comparison with the retention time of the commercially available UQ_10:10 (all-trans) standard, the early eluting UQ isomers (a-series) in our samples represent trans-isomers. Thus, the late eluting isomers (b-series) potentially represent compounds with one or more double bonds in cis-configuration. Up to four MK isomers with highly similar MS² fragmentations were detected, labeled consecutively a to d [MK_m:n(a-d)] from early to late eluting. MK and methyl-MK cis- and trans-isomers have previously been detected in the bacterium Mycobacterium phlei (63) and the archaeon Thermoplasma acidophilum (15). Thus, it is possible that the different isomers represent cis and trans isomers. However, with the analytical methods used in this study, configurational isomerism cannot be resolved.

Total quinone concentrations in the water column ranged between 0.21 and 9.4 ng liter⁻¹ (Fig. 4). The highest concentrations in the water column were measured at 40 mbsl within the oxic zone and at 300 mbsl within the anoxic part of the water column. From the oxic zone to the chemocline, quinone concentrations decreased 7-fold. The lowest concentrations occurred in the deepest water column sample at 1,200 mbsl.

Concentration profiles and relative abundances of individual quinones were strongly linked to the geochemical zonation of the Black Sea water column. The photic zone, represented by the 40-mbsl sample, was characterized by high relative abundances and concentration maxima of vitamin K₁, PQ_9:9(a), and UQs but low relative and absolute abundances of MKs (Fig. 4 and Fig. S3). The upper chemocline (90 and 120 mbsl) samples were dominated by saturated MKs (MK_6:0 and MK_6:1; 70% of total quinones) with minor contributions of UQs and methylene-ubiquinone (MQ_8:7; structural identification shown in Fig. S2). The lower contribution of MK_6:0 and MK_6:1 to total quinones compared to previously reported values by Elling et al. (16) results from the detection of MQ_8:7, which was not included in the earlier study. Polyunsaturated MKs first occurred in the photic anoxic zone (150 m), and ChQ_7:7 was only observed in this zone (Fig. 4 and Fig. S3). The deep (aphotic) anoxic zone was characterized by high relative abundances of polyunsaturated MKs (up to ∼50%) and UQs and minor contributions of MQ_8:7 and PQs (Fig. 4). While relative abundances of quinones were relatively uniform in the deep anoxic zone, most individual compounds showed distinct concentration profiles with either shallow maxima at 300 mbsl (e.g., UQ_8:8, UQ_10:10, MK_6:6, and MQ_8:7), deep maxima at 700 to 900 mbsl [e.g., UQ_9:9 and MK_10:10(b)], or at both depths [e.g., UQ_9:9(a), MK_7:6, and MPs; Fig. S3]. MPs were absent from the oxic part of the water column and the chemocline but were detected in the anoxic part of the water column, except at 1,200 mbsl (Fig. S3). Most notably, UQ isomers were detected only in the anoxic zone, and the concentration profiles of different MK and UQ isomers were highly divergent (Fig. S4).

To further constrain the sources of quinones, we compared their depth profiles with selected apolar microbial lipid biomarkers, such as cholesterol, as a general eukaryotic biomarker, alkenones for haptophyte algae, and isorenieratene (structural identification using high-performance liquid chromatography-mass spectrometry [HPLC-MS] as shown in Fig. S2) for anaerobic photosynthetic bacteria. Alkenones and cholesterol showed the highest abundances in the photic zone and followed the trend of vitamin K₁, PQ_9:9(a), and most UQs, while isorenieratene was only detected in the photic anoxic zone where ChQ_7:7 occurred (Fig. S3 and S5).

In the sediments, total quinone concentrations were highly variable and ranged from 2 to 1,000 ng g⁻¹ sediment dry weight (sed. dw.) (Fig. 4). Highest concentrations were observed in the surface sediment and the sapropel layer (S1). Below the sapropel in lithological Unit II, concentrations were lowest and showed little variability.

Relative abundances of quinones differed between Unit I and Unit II sediments but were relatively uniform within each unit. Unit I sediments were dominated by UQs and PQs with minor amounts of polyunsaturated MKs, MK_6:0, MK_6:1, and MQ_8:7 (Fig. 4). Unit II sediments were characterized by higher relative abundances of UQs, ChQ_7:7, polyunsaturated MKs, MK_6:0, and MK_6:1 and lowered abundances of PQs. Absolute abundances of many quinones were dependent on the sedimentary unit, as were TOC and total quinones, showing a strong decrease in the upper 50 cmbsf, peaking at 398 cmbsf [e.g., vitamin K₁, PQ_9:9(a), MK_7:7(b)] or above [e.g., MK_7:7(a), MK_6:0, and MK_6:1], and low abundances in Unit II (Fig. S7). In contrast, concentrations of MQ_8:7, ChQ_7:7, and MPs gradually decreased with sediment depth irrespective of changes in lithological units. Multiple MK and UQ isomers were found only in the sediments and not in the water column [e.g., MK_7:7(a) and UQ_7:7(b)], and the relative abundances of isomers were different in the two sedimentary units and the water column. Additionally, concentration profiles of these isomers often diverged.

Like those of quinones, concentrations of apolar lipid biomarkers decreased strongly within the first 50 cmbsf (Fig. S5). In deeper sediments, concentrations varied distinctly, with alkenones showing no peak in the sapropel, cholesterol peaking only in the sapropel, and isorenieratene showing peaks at 113 cmbsf and in the sapropel at 398 cmbsf.

In the Black Sea water column, different quinones peaked at distinct, geochemically defined zones, indicating that quinones were produced in situ and rapidly turned over. This view agrees with earlier observations indicating high rates and efficiency of biomass turnover in the Black Sea, particularly in the oxic and suboxic zones (64). We used respiratory quinone profiling to examine the stratification of microbial communities and associated metabolic processes in the water column and the sediment. Metabolic stratification is indicated by the distinct clustering of quinone types, geochemical parameters, and sample depths in nonmetric multidimensional scaling (NMDS) analysis (see Fig. S8 in the supplemental material) as well as divergent quinone diversity indices (see the supplemental material). The biogeochemical zonation and the potential sources of quinones in the southern Black Sea are summarized in Table 1 and Fig. 5. Microbial communities and metabolisms in the water column were separated into (i) the oxic (photic) zone supporting oxygenic photosynthesis, (ii) the suboxic zone dominated by thaumarchaeal ammonia oxidation, (iii) the anoxic photic zone inhabited by sulfur-oxidizing photosynthetic bacteria, and (iv) the dark anoxic zone, which supports a variety of bacterial and archaeal metabolisms, such as methane oxidation, anaerobic ammonium oxidation (anammox), and sulfate reduction.

The major quinone types in the oxic water column are associated with aerobic autotrophy and heterotrophy (UQs) as well as oxygenic photosynthesis [UQs, vitamin K₁, and PQ_9:9(a)] (9). Sources for UQ_9:9, UQ_10:10, vitamin K₁, and PQ_9:9(a) are both cyanobacteria and eukaryotic algae (8, 10, 12, 13), substantiated by their covariation with cholesterol and alkenones, while UQs with side chain lengths of 7 to 10 are additionally produced by alpha-, beta-, and gammaproteobacteria (9, 10). Only the trans-isomers (a-series) were detected in the oxic zone, indicating that aerobic organisms predominantly synthesize UQs with this specific stereochemical configuration.

The detection of MK_6:0 and MK_6:1 indicates the occurrence of Thaumarchaeota at 40 mbsl (16), although the concentrations of these quinones and the contribution to the overall quinone pool were comparatively low (Fig. 4). Accordingly, low thaumarchaeal biomass in the oxic zone was previously implicated by low concentrations of intact polar glycerol dibiphytanyl glycerol tetraethers observed previously (16), typical membrane lipids of planktonic Thaumarchaeota (83, 84).

The quinone composition in the suboxic zone of the water column is substantially different from that observed in the oxic zone. The depth profiles of the thaumarchaeal quinones MK_6:0 and MK_6:1 showed a distinct concentration maximum at 120 mbsl (Fig. 5 and Fig. S3), indicating maximum thaumarchaeal abundance in the chemocline, where both ammonia and oxygen are almost depleted (16), which is in agreement with previous observations based on the abundances of thaumarchaeal 16S rRNA and ammonia-monooxygenase subunit A gene biomarkers (29, 38). The predominance of thaumarchaeal quinones (>70% of total quinones) (Fig. 4) and the decrease in dissolved ammonium concentration (Fig. 3) indicate that archaeal ammonia oxidation is the major respiratory process in this zone. Conversely, thaumarchaeal lipids have been found to comprise less than 10% of the combined bacterial and archaeal membrane lipids in this zone (16, 37), suggesting that microbial biomass in the chemocline is dominated by bacteria, whereas respiration is dominated by ammonia-oxidizing Thaumarchaeota. The high ratio of quinones relative to lipids (biomass) in Thaumarchaeota indicates high metabolic activity, as suggested by laboratory experiments (24), but may also reflect the low energy yield of ammonia oxidation that requires nitrifiers to oxidize 25 to 100 mol ammonia for each mole of CO₂ fixed (67).

The absence of quinones involved in photosynthesis indicates that photosynthetic cyanobacteria and eukaryotic algae either do not inhabit the suboxic zone or are metabolically not active. The UQs detected in the suboxic zone therefore likely originate from aerobic ammonia-, iron-, and sulfur-oxidizing alpha-, beta-, and gammaproteobacteria as well as aerobic nitrite- and methane-oxidizing alpha- and gammaproteobacteria (Table 1) (9, 85). The occurrence of these bacteria and biogeochemical processes in the Black Sea chemocline has also been attested by previous geochemical as well as gene and lipid biomarker surveys (34, 35, 86, 87). High relative abundances of a quinone specific for type I methanotrophs, MQ_8:7 (85), corroborate the importance of aerobic methanotrophy mediated by gammaproteobacteria in the Black Sea chemocline.

Penetration of H₂S-containing waters into the photic zone (Fig. 3) enables bacterial anoxygenic photosynthesis. Indeed, the carotenoid isorenieratene (88) and ChQ_7:7, which are both diagnostic for anaerobic phototrophic green sulfur bacteria of the family Chlorobiaceae (72, 89), showed distinct concentration maxima in anoxic, sulfidic waters at 150 mbsl (Fig. 5 and Fig. S3). This habitat depth is close to the maximum depth for anaerobic phototrophic growth calculated from the optical transparency of seawater (90). MK_7:7, which is a major quinone in the chlorosomes of Chlorobiaceae (72, 91), is the only MK that significantly increased in concentration at this depth (Fig. S3, b-isomer), suggesting a major contribution of these organisms to microbial activity in the deep chemocline. The abundance of a single Chlorobium species at depths of up to 145 mbsl has been reported to be a widespread feature of the Black Sea and attributed to extremely low-light adaptation of this particular phylotype (40, 41, 73). The peak of ChQ_7:7, the distinctive covariation with isorenieratene, as well as MK_7:7 indicate the high relative abundance of extremely low-light-adapted Chlorobiaceae in the Black Sea chemocline and validate the use of ChQ_7:7 as a biomarker for anoxygenic photosynthetic Chlorobiaceae.

Within the anoxic zone, quinone diversity indices are highest (Fig. S6) and abundances peak at concentrations more than 3-fold higher than those in the oxic and other layers (Fig. 4), indicating the highest microbial respiratory capacity. The first appearance of polyunsaturated MKs in the anoxic zone (including the photic, deep chemocline at 150 mbsl; Fig. 4 and Fig. S3) reflects a shift from aerobic and microaerobic archaeal and bacterial respiration to mainly bacterial, MK-dependent, anaerobic respiration. Likely sources of polyunsaturated MKs in the anoxic zone are diverse sulfate-reducing deltaproteobacteria, such as Desulfobacter, Desulfococcus, and Desulfosarcina spp., which produce predominantly MK_7:7 (76). Furthermore, Desulfovibrio and Desulfuromonas synthesize MK_6:6 and MK_8:8, respectively, as major quinones (8). Sulfur-oxidizing Epsilonproteobacteria previously detected in the Black Sea (36), e.g., Sulfurimonas and Sulfurovum, might represent additional sources of MK_6:6 (92). Potential sources for MK_7:7 and MK_9:9 are sulfate-reducing Firmicutes related to the genus Desulfotomaculum (8). Several of these bacterial clades have been suggested to be responsible to a large extent for dark carbon fixation below the chemocline (93). The occurrence of quinones associated with sulfate-reducing bacteria in the anoxic zone is consistent with the abundance of bacterial ether lipids reported previously for this zone (37), which are almost exclusively associated with anaerobic bacteria (94) and particularly sulfate reducers (95, 96). An activity and abundance maximum of sulfate reducers and sulfur oxidizers in the upper anoxic zone is consistent with the observations that these bacteria first appear beneath the chemocline (34, 36, 77, 87) and that sulfur oxidation and sulfate reduction rates are highest in the upper anoxic zone of the Black Sea (30, 45).

Bacteria of the phylum Planctomycetes might represent an additional source of MK_6:6 (9, 78, 97). Bacteria affiliated with a deeply branching monophyletic lineage of this phylum perform the reduction of NO₂⁻ to N₂ by NH₄⁺, i.e., anaerobic ammonium oxidation (anammox). Based on the analysis of 16S rRNA gene markers and the anammox-specific ladderane lipids, anammox bacteria have been shown to be present mainly within the suboxic zone of the Black Sea (34, 39). Moreover, previous studies proposed the cooccurrence and coupling of bacterial anammox and archaeal ammonia oxidation within the chemocline of the central Black Sea (29, 38). In contrast, the quinone profiles at our study site in the southern Black Sea suggest a vertical separation of the depth habitats of ammonia-oxidizing Thaumarchaeota and anammox bacteria by up to 200 m (Fig. 5), with Thaumarchaeota being confined to the suboxic zone and anammox bacteria residing within the upper anoxic zone (Table 1), although geochemical coupling of these processes cannot be excluded. High sulfide concentrations appear to inhibit growth of anammox bacteria (98), thus anammox bacteria are likely restricted to the upper part of the anoxic zone.

Actinobacteria are likely sources of fully unsaturated and partially saturated long-chain MKs, specifically of MK₈-MK₁₁ (8, 99). Bacteria affiliated with this phylum have been implicated in denitrification, i.e., the heterotrophic reduction of NO₃⁻ to N₂, within the anoxic zone of the Black Sea (36). Quinones with 9 to 11 isoprenoid units are abundant in the anoxic zone of the Black Sea, especially in the deeper part (Fig. S3), suggesting a significant contribution of Actinobacteria to metabolic activity, possibly denitrification, in the deep anoxic Black Sea.

The occurrence of UQs, PQs, and MQ_8:7 in the anoxic zone is enigmatic, as these quinones are typically associated with aerobic metabolism. It is plausible that some anaerobic bacteria use ubiquinone-dependent pathways, as is the case for the nitrite-reducing methanotroph “Candidatus Methylomirabilis oxyfera” (100–102). Likewise, activity of methanotrophic bacteria (a source of UQs and MQ_8:7) that potentially use nitrate or nitrite instead of oxygen as the electron acceptor has been reported from the cores of marine oxygen minimum zones (103, 104). Methanotrophic bacteria could also be involved in cryptic Fe and methane cycling, as previously implied for other anoxic habitats (105–109). Alternatively, aerobic respiration might be supported by episodic lateral intrusions of modified Bosporus water that provide dissolved oxygen (as well as inorganic electron acceptors and fresh organic matter) to the suboxic and upper anoxic zones of the Black Sea (56, 110, 111). Although lateral intrusions and vertical transport could also deliver fossil quinones to the anoxic zone (e.g., see reference 64), the isomer pattern of UQs suggests in situ production of aerobic-type quinones: late eluting UQ isomers occurred only in the anoxic water column, while they were absent from the oxic and suboxic zones (Fig. S3 and S4).

Water column methane concentrations were not measured at our study site, but methane has been shown to be present at micromolar levels in the anoxic zone at other sites in the Black Sea (34, 46). While it was suggested that most of the methane in the Black Sea water column derives from the sediment (46), the detection of MPs (Fig. 5 and Fig. S3) indicates that methanogenesis also occurs within the anoxic waters. MPs are exclusively found in the archaeal order Methanosarcinales (16, 20), which comprises the metabolically most versatile methanogens utilizing CO₂ + H₂, acetate, and methylated compounds as substrates (112). Methanogens are likely outcompeted by sulfate-reducing bacteria for acetate and H₂ due to thermodynamic constraints (113–115). Thus, utilization of noncompetitive methylated carbon substrates (114, 116) is a likely methanogenic pathway employed by Methanosarcinales in the Black Sea. However, anaerobic methanotrophic archaea (ANME-2), which are phylogenetically associated with the Methanosarcinales, have also been detected in the anoxic water column of the Black Sea (34, 35, 82). While the presence of MP in ANME-2 has not been confirmed, they have been implicated to function as electron carriers in these archaea (117), which may represent an additional source of MPs.

Other quinones specific for archaea, i.e., fully saturated menaquinones and naphthoquinones (16), were not detected in the anoxic zone of the Black Sea. This indicates that most anaerobic planktonic Cren- and Euryarchaeota either do not produce quinones or that these archaea synthesize bacterium-like polyunsaturated quinones similar to those found in extremely halophilic archaea of the order Halobacteriales (14) and thermophilic archaea of the order Thermoplasmatales (16), the latter being the closest cultivated relatives to the uncultivated planktonic marine group II Euryarchaeota, also found in the Black Sea (82, 118).

Quinone concentrations are highest at the sediment surface and in the sapropel (Fig. S8A). The concentration maxima could be related to zones of higher activity, which would be consistent with higher microbial activity typically observed in shallow sediments (119) and organic matter-rich layers (120). On the other hand, these zones would show highest preservation (121, 122) if quinones derived from the water column would accumulate as molecular fossils. This dichotomy between the dependency of metabolic activity (and, thus, potentially quinone abundance) on organic matter availability and the possibly higher preservation of allochthonous quinones in high-TOC intervals poses significant challenges to the interpretation of sedimentary quinone profiles, because the fate of quinones after deposition, i.e., their degradation, transformation, and/or preservation, remains unknown. At least some quinones, such as PQs and vitamin K₁, likely are molecular fossils, as they are typically associated with aerobic metabolisms and/or phototrophy and their abundances are correlated with preserved oxic/photic zone lipid biomarkers, such as alkenones (haptophyte algae) and cholesterol (eukaryotes) (Fig. S5 and S10). However, normalization of quinone concentrations to TOC content (Fig. S9) reveals that quinone abundances are elevated in high-TOC intervals. Maxima in quinone concentrations thus may not be primarily controlled by preservation but could reflect higher standing stocks of metabolically active microbes. Further, the depth profiles of several quinones differ significantly from those of the putative fossil quinones (Fig. S7), and although preferential export and selective preservation could explain the observed signals, we hypothesize that most quinones were produced in situ by benthic microbes. We base this hypothesis on three lines of evidence. (i) Selective degradation of quinones is unlikely due to their highly similar chemical structures. (ii) We further observed distinct compositional changes between the water column and the sediments within different quinone classes, e.g., isomer distributions (Fig. S4). This becomes particularly apparent when the full diversity of quinones is considered (Fig. S11) and is supported by statistical analysis, which revealed distinct clusters of water column and sedimentary quinones in the NMDS space (Fig. S12). (iii) Some quinones, for example, MK_7:7(a), MK_7:7(d), UQ_11:11 (a), and UQ_11:11(b), were detected in the sediments but not in the water column. Together, these findings provide strong evidence that quinone compositions of planktonic and sedimentary communities differ.

The sedimentary quinone composition seems to be dependent on lithology. Relative abundance plots of major quinone groups (Fig. 4) and statistical analysis (NMDS) of individual quinone abundances revealed two clusters that align with lithological Unit I (euxinic marine) and Unit II (lacustrine) and correlate to some extent with the geochemical zonation (Fig. S12 and S13; see the next section for a more detailed discussion). We observed, for example, higher relative abundances of archaeal MKs, ChQ, and MK_8:8-MK_11:11 in Unit II than in Unit I as well as fine-scale compositional changes within different quinone classes (Fig. S11). The quinone profiles thus agree with the hypothesis that community assembly and metabolisms in the deep sedimentary biosphere are determined by the initial community composition, and therefore the environmental conditions during sediment deposition (123–127), modulated by selective persistence (128). Within each lithological unit, the relative quinone compositions do not change with depth, not even in the sapropel. This invariance might be indicative of a high background signal of quinones associated with dormant or dead, but preserved, microbial biomass that likely masks in situ activity of a much smaller pool of living microorganisms. Still, the distributions of individual quinones allow the characterization of metabolic stratification, as discussed below.

Three biogeochemical zones can be distinguished based on geochemical characteristics and quinone profiles: (i) the sulfate reduction zone, including the SMTZ; (ii) sapropel S1; and (iii) the methanogenic zone.

Concentrations of most quinones decrease within the first 50 cmbsf (Fig. 4 and Fig. S5), likely reflecting decreasing metabolic activity concomitant with gradual depletion of electron acceptors (e.g., sulfate) with depth (Fig. 3). Within the broad SMTZ (centered around ca. 80 to 150 cmbsf), concentrations of many quinones either stabilize or increase. Compounds that increase include several polyunsaturated MK isomers, e.g., MK_6:6(b), MK_7:7(d), and MK_8:8(c) (Fig. S7). Sulfate-reducing Deltaproteobacteria, possibly also engaged in AOM, are probable sources for these compounds. Similarly, MPs could originate from ANME archaea, which were previously detected in this depth interval (129) and which were suggested to utilize these quinone analogs (117). Saturated MKs, which are more prevalent than polyunsaturated MKs in cultivated archaea, were not abundant (apart from the thaumarchaeal MK_6:0 and MK_6:1), suggesting that benthic archaea do not produce them. Notably, some UQ isomers show peaks in concentration within the SMTZ (Fig. S7). The sources of and processes associated with these aerobic-type quinones in anoxic sediments remain unclear, but it is plausible that some anaerobic bacteria use ubiquinone-dependent pathways, as discussed above for the anoxic water column. A possible source of UQs could be methanotrophic bacteria involved in AOM coupled to iron reduction, as discussed above for the water column. Indeed, Fe²⁺ accumulates in the pore water of the methanogenic zone concomitant with a decrease in methane concentration (Fig. 3). This release of Fe²⁺ from the organic-poor, lacustrine sediment of Unit II is strikingly similar to lacustrine sediments of the Baltic Sea, where the occurrence of AOM coupled to iron reduction was suggested (105). The lack of MPs in this zone indicates that any archaea involved in this process are not Methanosarcinales. Quinones characteristic for type II methanotrophs (MQs) were not detected in this zone either, while UQ_8:8, the dominant quinone of type I aerobic methanotrophs (85), was a major component of the UQ pool (Fig. S11), potentially indicating activity of this clade.

Saturated and polyunsaturated MK concentrations peak either slightly above or within the sapropel (Fig. S7), indicating niche metabolic segregation. Coinciding concentration maxima of ammonia, phosphate (Fig. 3), and acetate (129) indicate enhanced heterotrophic activity in this zone. Due to a lack of electron acceptors, fermentation is likely to be the dominant heterotrophic process, as previously observed for deeply buried sapropels in the Mediterranean Sea (120). Multiple lineages of uncultivated archaea have been suggested to use heterotrophic and/or fermentative metabolisms (130–132), e.g., acetogenesis (133–135), and were reported to be abundant in the sediments studied here (Bathyarchaeota/MCG and MBG-D) (129). However, their quinone compositions remain unknown. Similarly, members of the candidate phylum “Atribacteria” (formerly OP9 and JS1), for which fermentation appears to be a characteristic metabolism (136, 137), were found in high relative abundances throughout the studied sediment core (129). While most fermenting bacteria do not produce isoprenoid quinones (8, 10), homoacetogens are a likely source for the high concentrations of MKs (e.g., MK_8:8) (138, 139) observed in the sapropel. Potential additional sources of polyunsaturated MKs are Chloroflexi, i.e., green nonsulfur bacteria (9, 140), which were previously found to be abundant and metabolically active in Mediterranean sapropels (120) and other deep biosphere environments (128, 141). Collectively, geochemical and quinone biomarker evidence indicates that the sapropel supports high heterotrophic activity, which in turn may supply acetate and potentially other metabolites (e.g., H₂) to the sulfate reduction and methanogenic zones above and below the sapropel.

Based on methane concentrations and carbon isotope ratios as well as quinone distributions, methanogenesis likely occurs in two modes. The gradual decrease in MP concentration within the upper 400 cmbsf indicates that methanogenesis and methane oxidation by Methanosarcinales (including ANME-2) predominantly occurs within the sulfate reduction zone, where methanogenesis is likely methylotrophic. Further downcore, methane concentrations strongly increase to a maximum below the sapropel at ∼630 cmbsf, indicating activity of methanogens other than Methanosarcinales and thus a vertical segregation of methylotrophic and hydrogenotrophic methanogens in Black Sea sediments. The methanogens inhabiting the deep sediments are likely obligate hydrogenotrophs such as Methanobacteriales or H₂-dependent methylotrophs such as Methanomassiliicoccales, which do not produce quinones or quinone analogs (16, 142, 143), or are affiliated with uncultured, potentially methanogenic lineages such as the Bathyarchaeota (144) or Verstraetearchaeota (65).

By using quinone profiling, a clear zonation of microbial diversity and redox processes could be resolved throughout the Black Sea water column and sediments. Coupled with geochemical data, quinone distributions indicated niche segregation between biogeochemical processes (e.g., photosynthesis, nitrification, aerobic methanotrophy, and anoxygenic photosynthesis) within the multilayered chemocline and anoxic zone (anammox, sulfate reduction, and methanogenesis/AOM) and continuation of anoxic aqueous community composition and respiration modes into shallow subsurface sediments. Within the sediment, segregation of microbial communities and respiration modes appeared to be driven by both lithology and geochemical gradients, with distinct quinone maxima around the sapropel reflecting intense heterotrophic activity.

Future work will need to target the relationship between biomass, metabolic activity (rates) and quinone abundances or distribution patterns, the source assignment of novel isomers, and the preservation potential of quinones in sediments. Cultivation experiments, polyphasic water column and sediment profiling studies, as well as development of methods for the isotopic analysis of quinones (e.g., for distinguishing MPs originating from methanogenic versus methane-oxidizing Methanosarcinales) will be needed to promote the utility of quinones as process biomarkers.

Suspended particulate matter and sediment samples were collected in the southern Black Sea in February 2011 at site GeoB15105 (Fig. 2) during R/V Meteor cruise M84/1 (145). Suspended particulate matter was recovered at nine water depths, ranging between 40 and 1,200 m, by pumping 6 to 204 liters of seawater through two precombusted 0.7-μm-pore-size glass fiber filters in tandem using in situ pumps (see Table S2 in the supplemental material). Recovered filters were immediately wrapped in combusted aluminum foil and stored at −20°C. Due to the use of 0.7-μm-pore-size filters, membrane lipid and quinone concentrations should be regarded as minimum estimates (146). Water column profiles of temperature, chlorophyll a fluorescence, and depth, as well as pressure and dissolved oxygen, were measured with a vertical resolution of 1 m using a CTD rosette (GeoB15105-5). The salinity was derived from conductivity, while the density was calculated from pressure and temperature measurements as well as salinity. Water column samples (10 to 20 ml) were retrieved from these hydrocasts with Niskin bottles and directly filtered (0.2-μm syringe microfilter) after the recovery of the CTD rosette. Sediments were recovered using a multicorer (GeoB15105-4) and a gravity corer (GeoB15105-2). Sampling procedures were described in detail by Zabel et al. (145). In brief, samples for gas analyses were taken immediately after core recovery from the freshly cut surfaces after core extrusion (multicorer) or cutting into 1-m sections (gravity corer). The cores were then sealed with end caps and transferred to a cold room (4°C). In the cold room, pore water was extracted from sediment cores through holes in the core liner with Rhizon microsuction samplers (0.1-μm filter width; Rhizosphere Research Products, Wageningen, the Netherlands) immediately after core recovery and split into subsamples for offshore and onshore analysis (145). After core splitting, sediment for lipid analyses was sampled in the cold room within 24 h after core recovery and stored at −20°C in brown glass bottles until solvent extraction.

Dissolved sulfate was determined by ion chromatography (Metrosep A Supp 5 column, with conductivity detection after chemical suppression; Metrohm Compact IC) in samples diluted 1:100 with Milli-Q-grade H₂O. Concentrations of dissolved phosphate were determined by forming phosphomolybdenum blue complexes (1 ml sample, 50 μl ammonium molybdate solution, 50 μl ascorbic acid solution, and 10 min of incubation) and measuring extinction at 820 nm using a Hach Lange DR 5000 spectrophotometer. Total dissolved iron (Fe_tot) was determined in HNO₃-acidified water samples by inductively coupled plasma optical emission spectroscopy (ICP-OES; axial plasma observation; Agilent 720). Ferrous iron (Fe²⁺) was analyzed photometrically onboard directly after sample collection (147). Dissolved ammonium was detected with a flow injection, Teflon tape gas separator technique after Hall and Aller (148) using a temperature-compensated conductivity meter (no. 1056; Amber Scientific) equipped with a micro-flowthrough cell (529; Amber Scientific) and a strip chart recorder. Dissolved hydrogen sulfide was determined in samples fixed with ZnCl₂ using the photometric methylene blue method (149).

Concentrations of dissolved methane were determined on board according to previously reported protocols (150, 151). In brief, 2 to 3 ml of wet sediment was incubated in a gas-tight 22-ml glass vial with a Teflon septum for 20 min at 60°C. After heating, 100- to 500-μl subsamples were taken from the headspace gas with a gas-tight syringe and analyzed by gas chromatography-flame ionization detection using a Carboxen-1006 PLOT fused-silica capillary column (0.32 mm by 30 m; Supelco, Inc., USA) on a ThermoFinnigan Trace GC. Methane concentrations were calculated from the partial pressure of methane in the headspace gas (calibrated against commercial standards; Scotty Laboratory Gases), headspace and sample volumes, and corresponding porosity data.

The stable carbon isotopic composition of methane was determined from duplicate analyses of the same samples using a Trace GC Ultra coupled to a Delta Plus XP isotope ratio mass spectrometer via a GC Combustion III interface (all ThermoFinnigan), as described previously (152). Carbon isotope ratios are reported in the δ-notation as per milliliter of deviation from the Vienna Pee Dee Belemnite (VPDB) standard. Analytical precision was determined by repeated injections of commercially available standards (methane at 100 ppm; Air Liquide) and was better than 1‰.

For TOC analysis, an aliquot (∼3 g) of homogenized and freeze-dried sediment sample was decalcified by acidification with 10% HCl. After washing with water, centrifugation, and freeze drying, between 10 and 30 mg of decalcified sediment was weighed into tin capsules and analyzed in duplicate on a Thermo Scientific Flash 2000 elemental analyzer connected to a Thermo Delta V Plus IRMS (153). TOC concentrations refer to initial sediment dry weight (corrected for carbonate content determined by weighing before and after decalcification) and are reported in weight percent (wt%) with a limit of detection of 20 μg C.

Isoprenoid quinones and lipids were ultrasonically extracted from filters and sediments by following a modified Bligh & Dyer protocol (95) with dichloromethane-methanol-buffer (1:2:0.8, vol/vol/vol) using phosphate and trichloroacetic acid (CCl₃CO₂H) buffers (each 2×). The total lipid extract (TLE) was dried under a stream of N₂ and stored at −20°C until analysis.

TLEs were analyzed on a Dionex Ultimate 3000RS ultra-high-performance liquid chromatography (UHPLC) system connected to a Bruker maXis Plus ultra-high-resolution quadrupole time of flight tandem mass spectrometer (qToF-MS) equipped with an electrospray ionization (ESI) ion source operating in positive mode (Bruker Daltonik, Bremen, Germany). For quantification, a known amount of C₄₆-GTGT (glycerol trialkyl glycerol tetraether) standard was added to each sample before injection. Analytes were separated using reversed-phase (RP) HPLC on an ACE3 C₁₈ column (2.1 by 150 mm; 3-μm particle size; Advanced Chromatography Technologies, Aberdeen, Scotland) maintained at 45°C as described previously (154).

The mass spectrometer was set to a resolving power of 27,000 at m/z 1,222, and every analysis was mass calibrated by loop injections of a calibration standard and correction by lock mass, leading to a mass accuracy of 1 to 3 ppm (155). Ion source and other MS parameters were optimized by infusion of standards into the eluent flow from the LC system using a T-piece. Quinones were identified by retention time, accurate molecular mass, and MS² fragmentation according to Elling et al. (16). Integration of peaks was performed on extracted ion chromatograms of ±10-mDa width and included the [M+H]⁺, [M+NH₄]⁺, and [M+Na]⁺ ions. Where applicable, double-charged ions were included in the integration.

Additionally, selected samples were analyzed for low-abundance methanophenazines under the same chromatographic conditions on a Dionex Ultimate 3000RS UHPLC system connected to an ABSciEX QTRAP4500 triple-quadrupole/ion trap MS equipped with an ESI ion source operating in positive mode. Target compounds were detected by scheduled multiple-reaction monitoring of diagnostic tandem MS (MS/MS) transitions. Ion source, multiple-reaction monitoring transitions, and other MS parameters were optimized by direct infusion of commercially available standards as well as TLEs of Nitrosopumilus maritimus and Methanosarcina acetivorans.

Quinone abundances were corrected for the relative response of commercially available menaquinone (MK_4:4 for MKs, chlorobiumquinone [ChQ_7:7], and vitamin K₁) and ubiquinone (UQ_10:10, trans-isomer for UQs, and PQs) standards (Sigma-Aldrich, St. Louis, MO, USA) versus the C₄₆-GTGT standard. Due to a lack of an authentic standard, MP concentrations were not corrected for their relative response and thus are not included in total quinone abundance and distribution patterns or calculations of diversity indices. The abundances of isorenieratene were corrected by the relative response of a commercial β-carotene standard (Sigma-Aldrich). Similarly, the abundances of cholesterol and alkenones were corrected by the relative response of a commercial cholesterol standard (Sigma-Aldrich) and synthetic C_37:2 and C_37:3 alkenone standards (156). The detection limits for quinones and lipids, as detected for authentic standards using the qToF and triple-quadrupole MS, were approximately 5 pg and 1 pg on column, respectively, depending on compound class and considering a signal-to-noise ratio of greater than 3.

Nonmetric multidimensional scaling (NMDS) analyses were performed in R (version 3.3.1 [157]) using the vegan package (version 2.4.2 [158]) separately on (i) water column suspended particulate matter, (ii) sediment quinone data sets, and (iii) the combined data sets. Normalized quinone concentrations (percent) were used for all statistical analyses, and geochemical/oceanographic parameters were fitted to the data. Geochemical/oceanographic parameters used for water column samples were temperature, fluorescence, salinity, O₂, NH₄⁺, PO₄³⁻, HS⁻, and SO₄²⁻. Geochemical parameters used for sediment samples were TOC, CH₄, δ¹³C_CH4, NH₄⁺, PO₄³⁻, HS⁻, and SO₄²⁻. For combined analyses, only parameters available for both sediments and water column (NH₄⁺, PO₄³⁻, HS⁻, and SO₄²⁻) were considered.