The Bacteriohopanepolyol Inventory of Novel Aerobic Methane Oxidising Bacteria Reveals New Biomarker Signatures of Aerobic Methanotrophy in Marine Systems

Aerobic methane oxidation (AMO) is one of the primary biologic pathways regulating the amount of methane (CH4) released into the environment. AMO acts as a sink of CH4, converting it into carbon dioxide before it reaches the atmosphere. It is of interest for (paleo)climate and carbon cycling studies to identify lipid biomarkers that can be used to trace AMO events, especially at times when the role of methane in the carbon cycle was more pronounced than today. AMO bacteria are known to synthesise bacteriohopanepolyol (BHP) lipids. Preliminary evidence pointed towards 35-aminobacteriohopane-30,31,32,33,34-pentol (aminopentol) being a characteristic biomarker for Type I methanotrophs. Here, the BHP compositions were examined for species of the recently described novel Type I methanotroph bacterial genera Methylomarinum and Methylomarinovum, as well as for a novel species of a Type I Methylomicrobium. Aminopentol was the most abundant BHP only in Methylomarinovum caldicuralii, while Methylomicrobium did not produce aminopentol at all. In addition to the expected regular aminotriol and aminotetrol BHPs, novel structures tentatively identified as methylcarbamate lipids related to C-35 amino-BHPs (MC-BHPs) were found to be synthesised in significant amounts by some AMO cultures. Subsequently, sediments and authigenic carbonates from methane-influenced marine environments were analysed. Most samples also did not contain significant amounts of aminopentol, indicating that aminopentol is not a useful biomarker for marine aerobic methanotophic bacteria. However, the BHP composition of the marine samples do point toward the novel MC-BHPs components being potential new biomarkers for AMO.


Introduction
Methane (CH 4 ) is a potent greenhouse gas, and its atmospheric concentration has tripled since pre-industrial times (e.g. [1,2]). Global oceans hold large subsurface reservoirs of CH 4 in the form of gas hydrates. These stores are precariously dependent on temperature and pressure. A rapid destabilisation of gas hydrates has been proposed to have caused vast releases of marine CH 4 in the past [3]. Increased input of CH 4 into the atmosphere has been interpreted through records of excursions of significant δ 13 C depletion in the geological record, such as in the Palaeocene-Eocene Thermal Maximum (PETM) [4][5][6][7]. CH 4 release into the atmosphere is regulated by methanotrophy, which converts CH 4 into CO 2 , thereby playing a key role in the carbon biogeochemical cycle [8]. Although traditionally anaerobic archaea have been the most studied methanotrophs (e.g., AMNE-1 and ANME-2; cf. [9]), recent observations have highlighted the importance of bacteria performing aerobic CH 4 oxidation (AMO) in marine, estuarine, and riverine fan environments (e.g., [10][11][12][13][14]). For example, pelagic AMO activity rose significantly after the Macondo oil well blowout in 2010 [15]. However, this activity was short-lived, highlighting the complexity of natural community interactions in response to increased CH 4 [16]. It is thus important to recognise and trace methanotrophy during past extreme events in order to understand its potential to mitigate future CH 4 release.
AMO bacteria belong to two phyla, Proteobacteria and Verrucomicrobia. Most isolates of Verrucomicrobia are thermoacidophilic [17][18][19][20], and have been found primarily in acidic, geothermal environments [21]. Aerobic methanotrophic members of Proteobacteria belong to two distinct classes, separated based on their carbon assimilation pathways [8]. Type II methanotrophs, members of the Alphaproteobacteria, are associated with terrestrial settings ( [8], and references therein), whereas Type I methanotrophs members of the Gammaproteobacteria are widespread in aquatic systems, although they are also found in terrestrial systems. Both Type I and Type II methanotrophs are known to synthesise bacteriohopanepolyol (BHP) lipids [22]. BHPs are precursors to hopanes, which are the most ubiquitous lipids in the geological record [23,24]. Therefore, being able to trace AMO using hopanoid biomarkers is of value to the study of the carbon cycle in the past.

Marine Carbonates
Carbonate, Gulf of Mexico ---+ -+ - [72] Authigenic carbonates, Gulf of Mexico - Seep Carbonate, Arabian Sea ---+ -+ - [106] Other samples . We can therefore discount SRB as the source of aminopentol in an environmental sample with a high ratio of aminopentol:aminotriol. A similar approach was recently used by [49] based on the ratio of aminotetrol:aminotriol which has been found in the range 1:20-100 in some species of Desulfovibrio SRB [47,50,51]. Interestingly, some of the species of SRB cultures analysed by Blumenberg et al. [47,50,51] also synthesised diplopterol and diploptene, which could explain the enhanced presence of these lipids in CH 4 -influenced anoxic sites.
Representatives from only a small number of Type I methanotroph genera have been tested for BHP production (e.g., [22,36,46,52,53]. Many more recently described genera, including novel genera from marine and other (hyper)saline environments, have yet to be explored (e.g., [54][55][56]). Moreover, relatively few studies of BHP distributions have targeted marine environments (Table 1). It is important, therefore, to determine whether aminopentol, which is seen as a diagnostic marker for Type I methanotrophs, is present in methanotrophs isolated from marine and other saline environments, and whether we can find aminopentol in CH 4 -influenced marine sediments. This knowledge will have implications for the use of aminopentol as a biomarker to trace AMO in modern and ancient marine environments.
In this study, our goal was to develop an appropriate biomarker approach for AMO, which will allow high throughput analysis of sediment without the requirement for laborious chemical conversion steps prior to compound specific isotope analysis. To this end, we screened the BHP distributions of three genera of aerobic methanotrophs (i.e., Methylomicrobium, Methylomarinum and Methylomarinovum). Methylomarinum and Methylomarinovum have not previously been investigated for BHPs. Two species of Methylomicrobium have been described previously [52,53], however, we include an additional species Methylomicrobium kenyense. These data are combined with literature BHP distributions of other AMO genera, including the recently reported Type I genus, Methylobacter [28], in order to facilitate interpretation of BHP distributions in six selected modern marine sediments from CH 4 -influenced systems. Two marine sites not influenced by CH 4 were also investigated as controls for background marine BHP signatures.

Methanotroph pure cultures
Cultivation of Methylomarinum vadi IT-4. Previously described Methylomarinum vadi IT-4 was isolated from a microbial mat sample (in-situ temperature 30-40°C) collected at a shallow marine hydrothermal system (depth,~23 m) in a coral reef off Taketomi Island, Okinawa, Japan [56,57]. Cultivation of this strain was performed at JAMSTEC, Japan, using MJmet medium at pH 6.6 at 37°C. A detailed site description and the enrichment and isolation procedures can be found in [57].
Peru Margin (PM). An intense upwelling regimes fertilises surface water productivity on the Peru Margin (PM). This lends to oxygen utilisation in the water column, causing the Eastern South Pacific Oxygen Minimum Zone (ESP OMZ) [74]. Three PM sediments were analysed (10-15, 20-25 and 40-45 cmbsf) from a core taken within the ESP OMZ, at 100 m water depth [73,75].

Lipid extraction
Total lipid extraction. All freeze-dried bacterial cells and marine sediments, except the Barents Sea samples extracted at GFZ Potsdam, were extracted using a modified Bligh-Dyer method [76,77]. Briefly, freeze-dried material was extracted in 19 mL of a 10:5:4 (v:v:v) mixture of MeOH:chloroform:H 2 O in a 50 mLTeflon tube. This mixture was sonicated for 15 min at 40°C, and centrifuged for 10 min. The supernatant was transferred to a second tube, and the residue re-extracted twice more. The chloroform in the supernatant was separated from the aqueous phase by adding water until the H 2 O:MeOH ratio was 1:1 (v:v), and collected. This procedure was repeated for the subsequent extractions. The collected chloroform total lipid extract (TLE) was dried by rotary evaporation in a round-bottom flask. The extraction protocol at GFZ Potsdam was similar but used a mixture of MeOH:DCM:ammonium acetate buffer [78].
Solid Phase Extraction. In-house comparisons have shown that amino-BHPs are better detected after solid phase extraction (SPE). An aliquot of the TLE was separated over a 1 mg NH 2 solid phase extraction cartridge, as described in [79]. Briefly, the aliquot was dissolved and loaded onto a hexane-rinsed cartridge using 200 μL chloroform. Six mL of a 98:2 (v:v) diethyl ether:acetic acid solution was eluted. Residual material was dissolved with 200 μL 2:1 (v:v) chloroform:MeOH and loaded onto the cartridge, followed by elution with 10 mL of MeOH. BHPs were isolated from the MeOH fraction.

Lipid analyses
BHP preparation and HPLC/APCI-MSn analyses. A known amount (ca. 5-10 μg/g dry sediment) of internal standard (5α-pregnane-3β,20β-diol) was added to SPE extracts of the TLE for BHP analysis. Samples were acetylated in 0.5 mL of a 1:1 (v:v) mixture of acetic anhydride and pyridine at 50°C for 1 h, then left to stand overnight at room temperature [80]. Solvent was dried under a stream of N 2 on a 50°C heating block. BHP samples were dissolved in MeOH:propan-2-ol (3:2; v:v), and filtered on 0.2 μm PTFE filters.
BHPs were analysed by high performance liquid chromatography coupled to positive ion atmospheric pressure chemical ionization mass spectrometry (HPLC/APCI-MS), using a datadependent scan mode (2 events) on an HPLC system equipped with an ion trap MS, as described in [46,81]. Further structural information for novel BHPs was obtained by way of MS 3 spectra. BHP concentrations were (semi) quantitatively estimated based on the response factor of authentic standards (M. Rohmer; Strasbourg, France and [46,77]) relative to the internal standard.

Results
In this study we investigated the BHP distributions in species of three AMO marine genera, and of eight marine environments, six of which were CH 4 -influenced.

Novel nitrogen-containing BHP components
In addition to the 'regular amino-BHPs' (e.g., I, II, and III; Figs 1 and 2), a suite of novel compounds were found in the methanotrophs and screened marine samples. Identification of these compounds is described in detail in the Supplementary Information (S1 File). Briefly, these components were related to the 35-amino-BHPs but differ in their terminal groups at C-35, which are tentatively proposed to comprise a methylcarbamate rather than a simple amine on the basis of interpretation of their APCI MS 2 and MS 3 spectra. In each case, the novel compounds (I MC , I MC ', II MC , III MC ; Figs 1 and 2) elute after their 'regular' amino-BHP analogues (i.e., I, I', II, III; Fig 2). This indicates that the tentatively-assigned terminal group structures are less polar than the regular terminal amines (after acetylation). The novel structures include: 35-methylcarbamate-bacteriohopane-32,33,34-triol (MC-triol herein; III MC ), 35-methylcarbamate-bacteriohopane-31,32,33,34-tetrol (MC-tetrol herein; II MC ), 35-methylcarbamate-bacteriohopane-30,31,32,33,34-pentol (MC-pentol herein; I MC ) and an isomer of I MC (I MC ') akin to the early-eluting aminopentol isomer (I'), which was found, based on mass spectra, in a culture of Methylovulum-like strain M200 [46].

Methanotroph BHP signatures
Four previously untested methanotrophs isolated from marine or saline, alkaline lacustrine environments, belonging to the three genera Methylomarinum, Methylomarinovum, and Methylomicrobium, were analysed for their BHP composition. An additional species Methylomicrobium alcaliphilum, the partial BHP composition of which was recently reported in [53], is also shown here in full for comparison with Methylomicrobium kenyense. All of the methanotroph cultures investigated only synthesised BHPs with a nitrogen atom at C-35 position (nitrogen-containing BHPs herein). The relative abundances of BHPs are indicated as the percentage of total BHPs in acetylated extracts, and are presented in BHP inventory of Methylomicrobium spp. The Methylomicrobium alcaliphilum and Methylomicrobium kenyense cultures did not contain aminopentol (I) above detection limit ( Fig 2D and 2E) although M. kenyense was found to contain minor abundance of 3-Me-aminopentol (I 3Me ; 1.0%; Fig 2E). The most abundant BHP in both Methylomicrobium cultures was aminotriol (III), making up ca. 65% of all BHPs in both species. The second most abundant BHP was 3-Me-aminotriol (III 3Me ) at 31.5% in M. alcaliphilum and slightly less in M. kenyense (23.9%). Both species also contained lower levels of aminotetrol (II) and 3-Me-aminotetrol (II 3Me ). M. kenyense also contained unsaturated compounds (ΔIII and ΔIII 3Me ; Fig 2E). The only MC compound identified in either Methylomicrobium sp. was MC-triol and then only at low levels (<3%).

Marine sediment and carbonate BHP signatures
Eight marine settings were studied for their BHP signatures ( Table 2). Six of these were known to be influenced by CH 4 (i.e., HHMV, BSCC, AMV, NZS, GoM cold seeps, GD) and two were used as comparison background marine levels (i.e., GoM sediments, PM).

BHP distributions in aerobic methanotrophs
Previously reported BHP distributions in AMO bacteria. Traditionally, Type I and Type II AMO bacteria had been distinguished by their different BHP signatures (e.g. [52]; see also review in [14]). Prior to the investigation of BHPs in the Methylovulum-like strain M200 [46], most screened Type I methanotrophs synthesised a high percentage of aminopentol (I) and lower contributions from aminotetrol (II) and in some cases aminotriol (III), and clustered in the left-hand corner of the amino-BHP ternary plot (Fig 3A). In contrast, Type II methanotrophs did not contain aminopentol, had varying contributions from II and III, and clustered along the right-hand axis of Fig 3A. The high relative abundance of III observed in Methylovulum-like strain M200 was, therefore, originally seen as an outlier [46]. Similarly, [52] showed that a culture of Methylomicrobium album did not contain aminopentol. At the time this was presumed to be a contaminated culture, however, [53] also did not report I synthesis in cultures of Methylomicrobium alcaliphilum. All of the recently analysed Methylobacter spp.
[28] join the more typical Type I methanotrophs in the left-hand corner of the plot, however, Methylobacter sp. BB5.1 increased the spread of the cluster with almost 40% III content.
Amino-BHP distributions in previously untested Type I AMO bacterial cultures. It was assumed that the screened species of AMO Type I bacteria investigated in this study would display similar BHP distributions as those of previously reported Type I bacteria. All three bacterial genera screened do indeed only contain amino-BHPs (Fig 2). However, the relative distribution of specific nitrogen-containing BHPs varies between genera, as well as between species belonging to the same genus. To allow for a more accurate comparison with data from the literature (Fig 3A, circles), only aminopentol, aminotetrol, aminotriol, and their methylated equivalents were considered when producing the ternary plot of 'regular' nitrogen-containing BHPs of the novel Type I cultures (Fig 3A, diamonds).
Aminopentol is the most abundant BHP in the novel species Methylomarinovum cadicuralii IT-9, which is in agreement with literature BHP compositions of most other Type I methanotrophs (Fig 3A, circles) Fig 2C). Methylomarinum vadi IT-4 shows relatively high proportions of aminopentol, but it is not the most abundant BHP (Fig 2A). Moreover, in our screening of two species of Methylomicrobium spp., aminopentol was not detected, similar to reported cultures of Methylomicrobium album and Methylomicrobium alcaliphilum [52,53]. Our results seem to confirm the near-absence of aminopentol in all screened Methylomicrobium spp., which are the first Type I methanotrophs apparently unable to synthesise aminopentol. However, changes in BHP composition can occur at different growth stages and under different conditions (e.g. [27, 83,84]), so further studies would be required to fully confirm this. It appears that the BHP distributions of Methylomicrobium and Methylovulum, which do not synthesise high amounts of aminopentol, should also no longer be considered outliers given the low levels of aminopentol in M. vadi IT-4 and Methylomarinovum spp. (Fig 2). This suggests a greater variance in the BHPs of Type I methanotrophs than previously thought. Furthermore, as Methylomicrobium has been isolated from a diverse range of marine environments [85][86][87], the absence of aminopentol in this genus might greatly affect its application as a marine aerobic methanotrophy biomarker. This, however, does not invalidate the use of aminopentol as a biomarker for methanotrophy.
There is also significant variation in the relative abundances of the other nitrogen-containing BHPs in Type I methanotrophs. A suite of novel BHPs identified as methylcarbamate (MC) BHPs are detected in all three genera screened in this study (Fig 2). These have not been reported in previous studies. Therefore, data available at Newcastle University from the analy- These results indicate that MC-BHPs are not universally present when their regular homologues are detected, and may be species specific and/or dependent on variations in growth conditions such as pH.
The two species of Methylomarinovum display significant variations in their BHP compositions (Fig 2B and 2C). Methylomarinovum sp. IN45 had a relatively low level of aminopentol. The most abundant BHP in Methylomarinovum sp. IN45 is II MC . Although in lower abundances, III MC and I MC are also higher in comparison to their 'regular' homologues in this species compared to M. caldicuralii IT-9 (Fig 2B and 2C). In contrast, the most abundant BHP in Methylomarinovum caldicuralii IT-9 is aminopentol, followed by almost equal amounts of I MC . The different relative BHP distributions between the Methylomarinovum spp. highlight that there can be significant variations within a genus. Methylomarinovum sp. IN45 was isolated from a deep-sea hydrothermal field and perhaps the high levels of methylcarbamate components observed are the result of a physiological adaptation to higher pressure in this environment. This may explain why the relative abundances of components in Methylomarinovum caldicuralii IT-9, the same genus but isolated from a shallow submarine hydrothermal environment, are quite different. Perhaps the complex functionality of the terminal group of the methylcarbamate components is more effective at stabilising the cell membrane and decreasing fluidity under these conditions.
3-methylaminotriol (III 3Me ) was observed in both Methylomicrobium spp. (23.9-31.5% of total BHPs; Fig 2D and 2E) in agreement with a recent report in [53]. This compound was accompanied by low levels of 3-methylaminotetrol (II 3Me ) in both species and trace amounts of 3-methylaminopentol in M. kenyense (I 3Me ). The absence of C-3 methylated structures in the previously investigated Methylomicrobium album strain BG8 [52] may appear inconsistent with the organisms investigated here; however, genomic investigations have revealed that M. album is separated from halo(alkali)philic representatives of the Methylomicrobium genus such as M. alcaliphium and M. kenyense [58], and perhaps specific environmental conditions influence the BHP composition of Methylomicrobium spp. as they seemingly do within the Methylomarinovum genus.
No C-3 methylated equivalents of aminotriol (III 3Me ), aminotetrol (II 3Me ), nor aminopentol (I 3Me ) were present in Methylomarinum vadi IT-4 or the Methylomarinovum spp., adding to examples of Type I methanotroph species that contain amino-BHPs, but not their C-3 methylated equivalents (e.g., see review in [14]). The most abundant BHPs in Methylomarinum vadi IT-4 were aminotriol (III) and MC-triol (III MC ), which were present in equal amounts (Fig  2A). Similar amounts of II and II MC , and I and I MC were also observed in this culture. A high proportion of III is unusual for a Type I methanotroph, but has been observed before in the Methylovulum-like strain M200. (Fig 4A; [46]). The new data reiterate that aminopentol is not always the most abundant BHP in Type I methanotrophs, nor necessarily the most appropriate biomarker for AMO.

BHPs in marine environments
Lack of BHP diversity in marine environments. The screened marine sediments and authigenic carbonates do not show large diversity in their BHP signatures (S1 Table). The limited BHP distributions are also comparable to other reported marine sediment BHP signatures, all dominated by BHT and BHT isomer, from a number of locations including the Black Sea [88], the Benguela upwelling system [89], and the Arabian Sea [90]. More recently a similar pattern was also seen in water column samples from the California Current system, where the wide diversity observed in the gene responsible for hopane cyclisation (squalene-hopene cyclase) was not reflected by distinct BHP fingerprints related to this potential range of sourceorganisms [91]. However, genetic information is quickly lost, and we must strive to find lipid biomarkers to trace particular metabolisms in the geological record.
Non-nitrogen-containing BHP concentrations in the screened sediments do not show remarkable signatures (S1 Table). BHT and anhydro-BHT, thought to be a degradation product of BHT and other composite BHPs such as BHT cyclitol ether [92], were found at all sites. The presence of soil marker BHPs at some sites, particularly NZS, indicates that these sediments could be influenced by terrestrial input of organic matter (e.g., [77,82,93]). However, as adenosylhopane is an intermediate in the biosynthesis of all other side-chain extended BHPs [94], other sources cannot be entirely excluded. BHT isomer, a biomarker for anaerobic ammonium oxidation [70], was found in high concentrations in GD (previously reported in [70]), as well as in the PM sediments, which underlie the Peruvian OMZ, where anammox is known to be an important process [95], and where BHT isomer has previously been reported from the water column within the OMZ [90]. The most abundant of the three regular amino-BHPs in the CH 4 -influenced marine sediments was aminotriol, which is not source-specific (e.g., [48]).
Aminopentol in marine sediments. Although aminopentol was found in significant abundance in some of the reported and screened Type I methanotroph cultures (Fig 2), it was not found to be abundant in most of the CH 4 -influenced marine sites in this study (Table 2). In fact, it was only detected in AMV, GD surface sediments, and two NZS samples (Fig 3B;  Table 2). The discrepancy between the distinct amino-BHP signatures of isolated Type I AMO bacteria and signatures of CH 4 -influenced marine sites is highlighted in the ternary plots of the relative composition of aminopentol, aminotetrol, and aminotriol (Fig 3A cf. Fig 3B). These differences could be due to the particular methanotrophic bacterial community responsible for methanotrophy in the CH 4 -influenced marine sediments. Ruff et al. [96] found that diversity in the global CH 4 seep microbiome was controlled by environmental factors such as temperature and electron acceptor availability. Considering their findings, it is possible that the environmental conditions in most marine CH 4 -influenced sediments favour specific methanotroph communities. For example Methylomicrobium spp., found in saline environments [85][86][87] and saline, highly alkaline environments [58,97], and which do not produce aminopentol in significant amounts (Fig 2), could be present. However, the absence of C-3 methylated compounds is confounding for a Methylomicrobium source, pointing towards other methanotrophs that do not synthesise aminopentol. Yan et al. [98] found that 85% of the operational taxonomic units (OTUs) from the same sites as our GoM cold seeps did not group with known sequences of a subunit of particulate methane monooxygenase (pmoA). This would suggest the presence of novel methanotrophic species in GoM. In the same way, significant pmoA diversity has been observed in sediments from the North American margin [99], a shallow CH 4 seep [100], a marine estuary [101], and hydrocarbon seeps [102]. pmoA OTUs from the NZS sediments grouped with methanotrophic endosymbionts [68], including Bathymodiolus spp., which have been shown to contain neither aminopentol nor methylated BHPs [103]. Nevertheless, the absence of methylated amino-BHPs in the screened marine sediments ( Table 2) may suggest Methylomicrobium album, or a related species that also does not synthesise methylated amino-BHPs, being the dominant methanotroph in CH 4 -influenced marine environments.
These are not the first reports of marine CH 4 -influenced environments not containing aminopentol (Table 1). For example, using methods targeting the functional gene pmoA, which is produced by most methanotrophs, Type I methanotrophs were detected in all three units of Ace Lake sediments. However, aminopentol was only detected in sediments deposited under freshwater conditions (unit III) despite the fact that the modern meromictic water column, containing relict seawater left behind after the sea level fell around 9000 years ago, hosts the Type I methanotroph Methylosphaera hansonii [104]. No aminopentol was detected in the methanotrophic symbionts in the gill tissue of a cold-seep mussel, despite other lipid-based evidence suggesting the presence of a Type I methanotroph [103,105]. Similarly, CH 4 seep carbonates from Alaminos Canyon, northern Gulf of Mexico [41] and the Northern Arabian Sea [106] were found to lack aminopentol. Conversely, aminopentol was detected in the water column of the Baltic Sea with supporting evidence for the presence of Type I methanotrophs from 13 C-depleted PLFAs [107]. Aminopentol was also detected in the water column of the Black Sea in the oxic-anoxic water transition, but not in the underlying sediment [12,88,108].
The presence of aminopentol in sediments from the AMV, located on the Nile deep-sea fan, in the Eastern Basin of the Mediterranean Sea (Table 2) may be explained by Nile River outflow carrying terrestrial wetland methanotrophy signatures into the Mediterranean, as seen in the Amazon and Congo River fans [14,109]. This would appear to indicate that aminopentol is still an excellent biomarker for terrestrial AMO. The near-absence of soil-marker BHPs in AMV ( Table 2) may still point towards in-situ marine production of aminopentol. However, the relative abundance of soil-markers in terrestrial settings has recently been found to be strongly influenced by environmental factors; higher temperatures and low pH (in peatlands) can both strongly reduce the relative proportion of soil marker BHPs as a proportion of total BHPs [40, 110,111]. Aminopentol was found in NZS sediments that also contained soil marker BHPs (Table 2). Therefore, aminopentol in sediments from NZS may have originated from terrestrial sources. Aminopentol in GD surface sediments may be the result of a distinct AMO community living in the specific environment prone to carbonate formation in GD. Unfortunately, samples were not properly preserved to be able to determine AMO diversity using geneticbased analyses of the pmoA gene in these sediments. The cumulative results of the studied marine sites do, however, indicate that an absence of aminopentol is not necessarily evidence for the absence of methanotrophs or aerobic methane oxidation.

Alternative BHP biomarkers for AMO and implications for the marine sedimentary record of methanotrophy
Regular amino-BHPs. Screened Type I methanotrophs also produced varying amounts, depending on the genera, of aminotetrol (II) and aminotriol (III) (Fig 3A), both of which were found in CH 4 -influenced marine sediments (Fig 3B; Table 2). However, these two amino-BHPs are less source-specific to methanotrophic bacteria than aminopentol, and do not make ideal biomarker lipids for methanotrophy. Given that 3-Me-aminotriol (III 3Me ) made a significant contribution to the amino-BHP abundance in screened cultures of Methylomicrobium spp. (23.9 and 31.5% of total amino-BHPs; Fig 2D and 2E) and 9.8% in Methylobacter sp. BB5. 1 [28], it was expected that III 3Me would be an important amino-BHP in CH 4 -influenced marine sediments. However, III 3Me was not found in any of the screened sediments ( Table 2). Methylomicrobium alcaliphilum and Methylomicrobium kenyense are adapted to high alkalinity, but not necessarily to high salinity [58]. This distinct lack of III 3Me in marine sediment samples would seem to indicate that the Methylomicrobium species we investigated are not the primary source of amino-BHPs in CH 4 -influenced marine environments.
III 3Me has only occasionally been reported from environmental samples including some soils [82,112] and most recently in a peat core from Germany [111], but only at very low levels ( Table 1). Other C-3 methylated amino-BHPs are even less common (Table 1). I 3Me was first reported from a neo-volcanic, eutrophic and saline lake sediment (La Piscina de Yuriria, Mexico; [45]), and subsequently from a geothermal silica sinter (Opaheke Pool hot spring, New Zealand; [113]). The pentafunctionalised II 3Me was also present in the Mexican lake sediment. Both of these compounds were reported in one study on the Black Sea water column [12], but were absent at another site [108]. The apparent discrepancy between the very limited occurrence of C-3 methylated BHPs (as measured using the periodic acid cleavage technique which converts polyfunctionalised BHPs into GC-amenable primary alcohols; e.g., [36,75]) and their wider occurrence in the form of 3-Me hopanes in ancient rocks and oils was first identified in [114]. These authors found 3-Me-BHPs to be abundant only in a very limited number of settings, under quite specific conditions (i.e., some alkaline lakes). The occurrence of 3-Me-hopanes in marine authigenic carbonates [31,115], which form under highly alkaline conditions are also consistent with a Methylomicrobium source ( [58], and references therein). It was further suggested that 3-Me-BHPs and hexafunctionalized BHPs appear to have different sources (possibly, but not necessarily restricted to, only Type I methanotrophs; [114]). Culture studies (on the moderately thermophilic Type I methanotroph Methylococcus capsulatus) have shown that production of C-3 methylated compounds may be related to growth stage. Higher relative proportions of methylated BHPs replaced the non-methylated equivalents during stationary phase growth [83], and appear to be necessary for maintaining intra-cellular membrane structures [27]. These important physiological roles for methylated BHPs are at odds with the very sparse occurrence of these compounds in modern settings (Table 1), and clearly our understanding of the factors controlling their biosynthesis and subsequent preservation in sediments is still limited, hampering interpretation of certain BHP signatures.
Methylcarbamate-BHPs. Most of the marine sediments influenced by CH 4 contained at least MC-triol, albeit at relatively low abundances (Fig 4B; Table 2). The fact that the MC-BHPs were found in all strains of methanotrophs analysed, though not all components in the suite were present in every strain, shows the biomarker potential of these BHPs for AMO (Fig 3A). MC-tetrol (II MC ) was the most abundant component in Methylomarinovum sp. IN45, and MC-BHPs were found in higher abundance than the 'regular' 35-amino-BHP homologues, which may allow this particular hydrothermal vent species to be identified in environmental settings. Unsaturated MC-triol (ΔIII CME ) was found in high abundance in AMV, HMMV, and NZS, but was not found in any of the methanotroph cultures. This is possibly because the BHP signatures in most CH 4 -influenced marine sediments are sourced from AMO bacteria that have no cultured relatives or at least none which have been tested for BHP production.
Given the small diversity in BHPs found in marine sediments and the need for an AMO biomarker, there appear to be few BHPs that meet the criteria of being source-specific and abundant. This has significant implications for the development of a proxy using aminopentol to trace AMO in marine settings. Applying MC-BHPs combined with the traditional suite of amino-BHPs (e.g. aminopentol, aminotetrol, and aminotriol) seems to be the most appropriate biomarker course for AMO.

Conclusions
Isolated methanotrophs from previously unexamined genera and species displayed marked differences in their relative abundances of amino-bacteriohopanepolyols (BHPs). Aminopentol (I) was the most abundant BHP in Methylomarinovum caldicuralii IT-9, which fits with the typical BHP signature of known Type I methanotrophs. However, the BHP signatures of Methylomarinovum sp. IN45 and Methylomarinum vadi IT-4 both did not show aminopentol as the most abundant BHP. Moreover, neither of the Methylomicrobium spp. contained aminopentol and only one contained a low level of 3-methyl-aminopentol showing that not all Type I methanotrophs synthesise aminopentol, agreeing with previous environmental studies. Considering Methylomicrobium can be prevalent in marine environments, this has implications for the use of aminopentol as a biomarker for marine methanotrophy. A suite of components related to amino-BHPs, but with methylcarbamate (MC) terminal groups, were detected for the first time, and were present in all Type I methanotroph strains tested. Marine sediments influenced by CH 4 did not contain significant amount of aminopentol, but did contain MC-BHPs. This study highlights the relatively low BHP diversity within marine sediments, and indicates that the combined use of MC-BHPs and amino-BHPs might be preferential to trace aerobic methane oxidation (AMO) in marine settings.