Comparative Lipidomics of Oral Commensal and Opportunistic Bacteria

The oral cavity contains a vast array of microbes that contribute to the balance between oral health and disease. In addition, oral bacteria can gain access to the circulation and contribute to other diseases and chronic conditions. There are a limited number of publications available regarding the comparative lipidomics of oral bacteria and fungi involved in the construction of oral biofilms, hence our decision to study the lipidomics of representative oral bacteria and a fungus. We performed high-resolution mass spectrometric analyses (<2.0 ppm mass error) of the lipidomes from five Gram-positive commensal bacteria: Streptococcus oralis, Streptococcus intermedius, Streptococcus mitis, Streptococcus sanguinis, and Streptococcus gordonii; five Gram-positive opportunistic bacteria: Streptococcus mutans, Staphylococcus epidermis, Streptococcus acidominimus, Actinomyces viscosus, and Nanosynbacter lyticus; seven Gram-negative opportunistic bacteria: Porphyromonas gingivalis. Prevotella brevis, Proteus vulgaris, Fusobacterium nucleatum, Veillonella parvula, Treponema denticola, and Alkermansia muciniphila; and one fungus: Candida albicans. Our mass spectrometric analytical platform allowed for a detailed evaluation of the many structural modifications made by microbes for the three major lipid scaffolds: glycerol, sphingosine and fatty acyls of hydroxy fatty acids (FAHFAs).


Introduction
Omics technologies, particularly genomics, are increasing our knowledge base of microbial communities resident in oral and gastrointestinal compartments.This information increases our understanding of the pathogenic processes that contribute to a diversity of oral diseases.Within the oral cavity, microbes are prevalent on the tongue, buccal mucosa, palate, gingiva, teeth, salivary glands, and saliva.Diverse microbial populations of over 600 bacterial species and 100 fungal species have been identified [1], encompassing both commensal and opportunistic bacteria and fungi.
Common commensal bacteria, such as Streptococcus intermedius, Streptococcus mitis, Streptococcus sanguinis and Streptococcus gordonii, are known as early colonizers that adhere to the underlying epithelial cells and function as a protective barrier [2].They serve as a scaffold for other oral bacteria, ultimately leading to the formation of multi-species biofilms [3].These commensal species are in symbiosis with their human hosts, antagonizing the growth of opportunistic bacteria [4], thereby aiding in the prevention of dental carries and periodontal disease [5].
With regard to oral diseases, tooth decay (enamel destruction) and periodontitis (plaque formation and gum weakening) are caused by bacteria.Streptococcus mutans is a key player in tooth decay while Porphyromonas gingivalis and Fusobacterium nucleatum are involved in periodontitis.F. nucleatum is unique in that it is a bridge that links early and late bacterial colonizers involved in plaque biofilm formation [6].The mechanisms of polymicrobial biofilm formation and pathogenicity involve complex membrane proteinlipid-carbohydrate complexes [7] generated by multiple interacting bacterial and fungal species [8][9][10].For example, in bacterial interactions, the growth of Veillonella parvula is augmented by the supply of lactic acid provided by Streptococcus species [9], while in fungalbacterial interactions, polysaccharides secreted by Candida albicans augment Streptococcus mutans' contribution to biofilm formation [10,11].Fungal-bacterial interactions result in increased virulence, as evidenced by invasive candidiasis and early childhood caries with C. albicans combined with Streptococcus spp.[10,11].
Oral health and overall systemic health are intrinsically linked.Oral opportunistic bacteria can also enter the bloodstream and result in systemic infections and infections of other tissues.Current thoughts are that oral bacteria can augment inflammatory processes involved in cardiovascular disease [6], respiratory disease [12], rheumatoid arthritis [13], cancer [1,14,15], Alzheimer's disease [16], multiple sclerosis [8], and bacterial vaginosis associated with poor pregnancy outcomes [17][18][19].For example, P. gingivalis is linked to diabetes, cardiovascular diseases and Alzheimer's disease [16].Similarly, several species of oral commensal streptococci, including S. gordonii, S. mitis, S. sanguinis and S. oralis, have also been implicated in infective endocarditis [7].From these reports, it is evident that a healthy balance of both commensal and opportunistic microbes in biofilms of multispecies communities is required to maintain not only good oral health, but ultimately systemic health.The lipodomic profiles associated with these microbes provide for proper construction of oral biofilms [9,10].
These lipodomic profiles associated with oral microbes are intimately involved with the formation of these biofilms.To better understand how they are constructed, it is imperative that lipid profiles for oral bacteria and fungi be defined.However, at this time lipidomics evaluations of oral bacteria and fungi have been limited to T. denticola [20], S. mutans [21][22][23], and C. albicans [24,25].We undertook a larger comparative study utilizing electrospray high resolution mass spectrometry (ESI-HRMS) to monitor lipids and to validate their identity by MS 2 .ESI-HRMS provides the sensitivity and specificity required for the analyses of diverse complex bacterial lipids that possess aminoacyl, peptidyl, and glycosyl modifications [26,27].
Between sample injections, the syringe and tubing were flushed with 1 mL of methanol followed by 1 mL of hexane/acetate/chloroform/water (3:2:1:0.1 ratio).FIA has the advantages of high sample throughput with a short analysis time and data acquisition, with a constant concentration of the lipid matrix.
Based on our infusion solvent, the predominant ions were [M+H] + , [M-H 2 O+H] + or [M+NH 4 ] + in PESI, while they were [M-H] − or [M+Cl] − in NESI, where M is the exact mass of each lipid and +H is addition of a proton while -H is loss of a proton.For MS 2  analyses, precursor ions were selected with a 0.4 amu window and collision energies of 15, 30 and 50 arbitrary units.Product ions were monitored with a resolution of 240,000 [28,29].

Data Reduction
Mass spectrometric data were imported into an Excel spreadsheet containing our inhouse master lipidomics database of over 12,000 individual lipids from 193 different lipid families.The imported data included individual scanned masses and their associated peak intensities, which were then matched to lipids in the master database provided that the error was <2.0 ppm.For positive hits, the extracted mass and the associated peak intensity were imported into a new active spreadsheet if the peak intensity was >100,000 integrated counts (signal/noise > 3).For positive hits, the extracted mass and the associated peak intensity were imported into a new active spreadsheet if the peak intensity was >100,000 integrated counts (signal/noise > 3).Data are presented as a rank order with the most intense peak being assigned a value of 1.0.For the most intense peak, its relative levels (Relative level = endogenous lipid peak area/peak area of 2 nmoles [ 13 C 3 ]DG 36:2) are included in brackets in the Supplementary Tables.
Since there is no common "housekeeping" lipid for all the microbial strains we examined, to assess potential ion suppression, we calculated the ratio of the 2 internal standards in each extraction.For the PESI analyses, the [ 31 H 2 ]PE 34:1/[ 13 C 3 ]DG 36:2 ratio was 2.2 ± 10% while for the NESI analyses, the ratio was 1.6 ± 10%.

Consideration of Targeted vs. Non-Targeted Lipidomics Analyses
The ultimate goal of our research program is to establish a number of absolute quantitation assays for key lipid biomarkers of microbial infections, at the sub-threshold level, relative to clinical signs.This will be a laborious and expensive undertaking, since the analytical standards and stable isotope internal standards will generally require synthesis.Therefore, it is critical to identify and validate which lipid biomarkers represent a worthwhile investment of these resources.Non-targeted lipidomics, utilizing FIA-ESI, allows a broader analysis of the wide structural diversity of bacterial lipids that is not achieved with chromatographic methods.In addition, membrane adaptations that bacteria invoke with environmental changes can be monitored [30].
The utility of non-targeted analysis to identify potential bacterial lipid biomarkers has already been demonstrated, e.g., unique lipid biomarkers in the serum of cattle with paratuberculosis [26] and mycolic acid biomarkers useful for the characterization of Gordonia spp. in human sputum samples [31].While a number of studies have focused on some oral bacterial lipid families, there is currently no comparative lipidomics studies of oral microbiota.This study represents the first step in this effort.Optimistically, unique lipid biomarkers associated with oral microbial dysbiosis can lead to potential diagnostic tests while increasing our understanding of the interactions between oral microbial species [6,7,9,10].
The diversity of microbial lipids is discussed next.Detailed information on each lipid is presented in the Supplementary Tables.This includes the lipid exact mass, monitored ions, and levels for each sample.

Modified Fatty Acyls: Aminoacyl Hydroxy-Fatty Acids (HFAs) in Gram-Negative Bacteria
Gly-HFA 16:0 (commendamide) was the predominant Gly-HFA family member across a number of Gram-negative oral bacteria investigated in this study.Commendamide was first reported for Bacteroides spp.isolated from GI microflora [32][33][34].This is the first report of commendamide in oral microflora (Table 2), with MS 2 analyses confirming the structure via generation of the product ions [Gly = 74.0248]− and [Gly-CO-CH 2 = 116.0353]− .The roles of this bacterial endocannabinoid agonist in oral and GI function remain to be established.
Table 3. Gly-Ser-FAHFAs and the Gly-Ser-FAHFA-phospho-diacyl-glycerols (Gly-Ser-FAHFA-P-DG). Data are presented as Rank orders with the relative levels of the most abundant family member to 2 nmoles of [ 13  The Gly-Ser lipids have been extensively studied in the laboratory of Dr. F. C. Nichols, including the synthesis of analytical standards and stable isotope internal standards for absolute quantitation [42].These dipeptide lipids have been recovered from periodontitis samples [43][44][45][46], arteries [44] and human serum [43].In these clinical cases, the Gly-Ser lipids are virulence factors present in outer membrane vesicles [47][48][49] and are TLR2 ligands [45].Lower serum levels of the predominant family member, Gly-Ser-FAHFA 32:0, have been monitored in both multiple sclerosis and Alzheimer's patients [43].These researchers have suggested that this results in altered brain microglial function [50], a cellular pathway in neurodegeneration.In addition to Gly-Ser FAHFAs in some Gramnegative bacteria, phosphor-glycerol serine-glycine lipo-dipeptides (Gly-Ser-FAHFA-P-DG) have also recently been discovered [46].In our study we only monitored these complex lipids in P. gingivalis and F. nucleatum (Table 3).
Gln-and Orn-FAHFAs have been reported previously in Gram-negative bacteria [33].In our study we only monitored Gln-FAHFAs in P. gingivalis, representing a unique lipid biomarker for this microbe.The product ions (Figure 1) for both asparagine-and glutamine-FAHFAs were monitored in the MS 2 analysis of Gly-Ser-FAHFA 32:0 (653.Orn-FAHFAs were also monitored in P. ginivalis, T. denticola, and A. muciniphilia (Supplementary Tables).The MS 2 product ions [133.09715/115.08659]+ validated the amino acid component as ornithine.
Recently, a new family of phosphor-glycerol Gly-Ser-FAHFAs in P. gingivalis has been reported, which are also TLR2 ligands [46].We also monitored these lipids in P. gingivalis and in F. nucleatum (Table 3) with the dominant family member Gly-Ser-FAHFA-P-DG (62:0).Lipid identities of Gly-Ser-FAHFA-P-DGs were validated with the MS 2 product ions listed in Table 4.
Gly-Ser-hydroxy-fatty acids (Gly-Ser-HFA), which may be metabolites and/or precursors of Gly-Ser FAHFAs, have also been reported in Gram-negative bacteria [48,52].We only detected these lipids in P. gingivalis with Gly-Ser-HFA 17:0 being the dominant family member.Structural validation was obtained with the product ions for glycine, serine, Gly-Ser, and HFA 17:0, all with <1 ppm mass error.Of significant relevance to these findings are the early astute observations of significant levels of 3-hydroxy fatty acid 17:0 in Gly-Ser-hydroxy-fatty acids (Gly-Ser-HFA), which may be metabolites and/or precursors of Gly-Ser FAHFAs, have also been reported in Gram-negative bacteria [48,52].We only detected these lipids in P. gingivalis with Gly-Ser-HFA 17:0 being the dominant family member.Structural validation was obtained with the product ions for glycine, serine, Gly-Ser, and HFA 17:0, all with <1 ppm mass error.Of significant relevance to these findings are the early astute observations of significant levels of 3-hydroxy fatty acid 17:0 in periodontitis [56], again, a possible metabolite and/or precursor of Gly-Ser FAHFAs, since the HFA in these lipids is 3-HFA 17:0.

Glycerolipids (GL) and Modified-GLs
The monoacylglycerols monitored were mainly 16:0, 18:0, and 18:1, with the highest levels in C. albicans.In contrast, diacylglycerol (DG) levels were highest in F. nucleatum, which also expressed the most diverse array of DGs.Alanyl-DGs were monitored in all Gram-positive bacteria except for S. oralis and S. gordinii.TGs were highest in S. acidominus, P. brevis and C. albicans (Supplementary Tables).

Glycerphospholipids (GPLs)
In general, while phosphatidylcholines (PCs) were detected across Gram-negative and Gram-positive bacteria, phosphatidylethanolamines were more prevalent in Gram-negative bacteria.Plasmalogens, phosphatidylethanolamines, and phosphatidic acids were only monitored at very low levels.Lysophosphatidic acid 16:0 was uniquely monitored at high levels in C. albicans (Supplementary Tables).
The dominant phosphatidylcholine (PC) in oral microflora was PC 34:1 with the highest levels in C. albicans, S. acididominimus, and C. albicans (Table 5).These microflora, along with P. vulgaris (high levels of PC 30:1) and N. lyticus (high levels of PC 30:0), have lysophosphocholine levels that are fractions of the PC levels (Table 5), similar to observations in eukaryotes.In sharp contrast to eukaryotes, a larger number of oral microflora have atypical LPC levels that are multiples of the endogenous PC levels (Table 5).These data suggest that LPCs may play unique roles in the membranes of these microflora and are not just degradation products or precursors of PCs.Another unique feature of bacterial PCs, compared to eukaryotes, is the absence of PCs with polyunsaturated fatty acids.This more limited variation in PC lipids in bacteria may contribute to a more rigid cell membrane and altered lipid raft function.
Structural identities of PCs and LPCs were validated in PESI with the product ion for phosphocholine (184.0738).This is important, since odd carbon phosphatidylethanolamines would be monitored at PC masses in PESI (e.g., PC 32:0 = PE 35:0 = 733.5622).
Nanobacter lyticus (TM7x) was included in these analyses since this ultrasmall bacterium (200 to 300 nm), which possesses a Gram-positive cell envelope, survives as an epibiont on the surfaces of larger oral bacteria [67,68] and is present in human saliva [69,70].Our data are the first lipidomics characterization of this important oral bacteria and demonstrate the major lipid family produced by the more compact genome of this bacteria is glycerol-phosphocholines (Figure 4; Supplementary Tables).
Phosphatidylglycerols (PGs) were monitored at low levels in all microflorae with no one family member being dominant (Supplementary Tables).In all cases, the MS 2 products were [DG-H 2 O] + , indicating that the lipids were PGs and not the isobars acyl-lyso-PG, also termed semi-lysobisphosphatidic acid (SLBPA) and bis(monoacylglycerol)phosphate (BMP) in the literature.
one family member being dominant (Supplementary Tables).In all cases, the MS 2 p ucts were [DG-H2O] + , indicating that the lipids were PGs and not the isobars acyl-lysoalso termed semi-lysobisphosphatidic acid (SLBPA) and bis(monoacylglycerol)phosp (BMP) in the literature.

Sphingolipids: Ceramides
While ceramides and GroP-ceramides were more predominant in Gram-nega bacteria and C. albicans, deoxy-ceramides were monitored across oral microflora (Sup mentary Tables).High levels of galactosyl dihydroceramides were monitored in C. cans.
Deoxy-ceramide sphingolipids lack the 1-hydroxy group of the sphingolipid h group.serine palmitoyl transferase is promiscuous and can utilize alanine rather than ine in the condensation reaction with a fatty acyl-CoA to generate deoxy-ceramides ra than a ceramide.Deoxy-ceramide 34:2 (Cer 34:2;O) was the dominant family member w highest levels in C. albicans, S. acidominimus, and S. mutans (Figure 5).Lipid identities w confirmed by the dehydrated deoxy-sphingosine bases as MS 2 products in PESI.

Sphingolipids: Sphingomyelins
Sphingomyelin levels were found to be low except for C. albicans and S. acidomini (Figure 6).In all cases, the MS 2 product was (Phosphocholine = 184.0733)+ , indicating

Sphingolipids: Ceramides
While ceramides and GroP-ceramides were more predominant in Gram-negative bacteria and C. albicans, deoxy-ceramides were monitored across oral microflora (Supplementary Tables).High levels of galactosyl dihydroceramides were monitored in C. albicans.
Deoxy-ceramide sphingolipids lack the 1-hydroxy group of the sphingolipid headgroup.serine palmitoyl transferase is promiscuous and can utilize alanine rather than serine in the condensation reaction with a fatty acyl-CoA to generate deoxy-ceramides rather than a ceramide.Deoxy-ceramide 34:2 (Cer 34:2;O) was the dominant family member with highest levels in C. albicans, S. acidominimus, and S. mutans (Figure 5).Lipid identities were confirmed by the dehydrated deoxy-sphingosine bases as MS 2 products in PESI.

Sphingolipids: Ceramides
While ceramides and GroP-ceramides were more predominant in Gram-negat bacteria and C. albicans, deoxy-ceramides were monitored across oral microflora (Supp mentary Tables).High levels of galactosyl dihydroceramides were monitored in C. a cans.
Deoxy-ceramide sphingolipids lack the 1-hydroxy group of the sphingolipid he group.serine palmitoyl transferase is promiscuous and can utilize alanine rather than s ine in the condensation reaction with a fatty acyl-CoA to generate deoxy-ceramides rat than a ceramide.Deoxy-ceramide 34:2 (Cer 34:2;O) was the dominant family member w highest levels in C. albicans, S. acidominimus, and S. mutans (Figure 5).Lipid identities w confirmed by the dehydrated deoxy-sphingosine bases as MS 2 products in PESI.

Sphingolipids: Sphingomyelins
Sphingomyelin levels were found to be low except for C. albicans and S. acidominim (Figure 6).In all cases, the MS 2
The roles of these unique highly charged lipids in oral and gut microbes remain t established.However, the location of sulfonolipids in the cell envelope of Cytophags suggests that they may contribute to membrane charge in some Gram-negative bac [84].
The roles of these unique highly charged lipids in oral and gut microbes remain to be established.However, the location of sulfonolipids in the cell envelope of Cytophags spp.suggests that they may contribute to membrane charge in some Gram-negative bacteria [84].

Glycopeptidolipids (GPL)
GPLs have been reported for a number of non-tuberculosis-causing Mycobacteria.These large molecular weight GPLs have a tripeptide-amino-alcohol core (Phe-Thr-Ala-Alaninol) with a 3-hydroxy or a 3-methoxy C26-C33 fatty acyl chain N-linked to the Phe.
GPLs are thought to be involved in virulence, biofilm formation, and sliding behavior [85].In this regard, GPLs may contribute to secretions involved in the sliding behavior of A. viscosus [86].

Mutanamides: Lipopeptides
The Gram-positive bacterium S. mutans acylates dipeptide products of non-ribosomal origin.Examples of this are the mutanamides, where Leu-Ala is acylated with keto fatty acids of various carbon lengths [87].In our analyses, we detected a number of mutanamides in Gram-positive but not Gram-negative bacteria or in C. albicans (Figure 7).Highest levels were monitored in S. intermedius, A. viscosus, and S mutans.The extent of the biological activity of these novel lipopeptides remains to be explored, but we do know that they inhibit fungal hyphal formation [87].
GPLs are thought to be involved in virulence, biofilm formation, and sliding behavior [85].In this regard, GPLs may contribute to secretions involved in the sliding behavior of A. viscosus [86].

Mutanamides: Lipopeptides
The Gram-positive bacterium S. mutans acylates dipeptide products of non-ribosomal origin.Examples of this are the mutanamides, where Leu-Ala is acylated with keto fatty acids of various carbon lengths [87].In our analyses, we detected a number of mutanamides in Gram-positive but not Gram-negative bacteria or in C. albicans (Figure 7).Highest levels were monitored in S. intermedius, A. viscosus, and S mutans.The extent of the biological activity of these novel lipopeptides remains to be explored, but we do know that they inhibit fungal hyphal formation [87].

Unique Fungal Lipid Biomarkers
In our study, four lipids distinguished C. albicans from all monitored bacteria were ergosterol, inositol phospho-ceramides, sulfo-phosphatidylglycerols, and lys phatidic acids.
The fungal inositol phosphoceramides (IPC) again were monitored in C. albic not bacteria (Supplementary Tables).IPC 38:0;O was the dominant member of th family.
Unexpectedly, sulfo-phosphatidylglycerols (Sulfo-PGs) were detected in C. but not bacteria (Supplementary Tables).The dominant family member was sulfo-P Previously, these sulfated lipids have only been reported for archaebacteria [92].Th alence of sulfo-PGs in other fungi and their functions remain to be explored.Wh sufonolipid sulfobacins (N-acylated capnine) have also been reported for bacteria and fungi (see Section 3.9), this is the first report of sulfo-PGs in a fungus.
While lysophosphatidic acids (LPA) are also present in bacteria, the only robu levels monitored in this study were in C. albicans (Supplementary Tables).

Unique Fungal Lipid Biomarkers
In our study, four lipids distinguished C. albicans from all monitored bacteria.These were ergosterol, inositol phospho-ceramides, sulfo-phosphatidylglycerols, and lysophosphatidic acids.
The fungal inositol phosphoceramides (IPC) again were monitored in C. albicans but not bacteria (Supplementary Tables).IPC 38:0;O was the dominant member of this lipid family.
Unexpectedly, sulfo-phosphatidylglycerols (Sulfo-PGs) were detected in C. albicans but not bacteria (Supplementary Tables).The dominant family member was sulfo-PG 38:0.Previously, these sulfated lipids have only been reported for archaebacteria [91].The prevalence of sulfo-PGs in other fungi and their functions remain to be explored.While the sufonolipid sulfobacins (N-acylated capnine) have also been reported for bacteria [83,84] and fungi (see Section 3.9), this is the first report of sulfo-PGs in a fungus.
While lysophosphatidic acids (LPA) are also present in bacteria, the only robust LPA levels monitored in this study were in C. albicans (Supplementary Tables).

Summary
Microbial lipids have a broader range of structural diversity and complexity, compared to the mammalian lipidome.Species-specific lipid modifications (e.g., glycosylation, incorporation of amino acids and peptides) provide the potential to identify lipid biomarkers for bacteria and fungi.Lipid biomarkers can serve as research tools in the study of biofilm production and functions, pathogenic lipids, cell wall/envelope molecular adaptation, and detection of sub-clinical microbial infections.
The extensive diversity of lipid modifications was clearly demonstrated in our study.The lipid scaffolds of glycerol (glycerolipids, glycoglycerolipids, glycerophospholipids, and glycerophospholipid modified ceramides), sphingosine (ceramides, sphingomyelins), and FAHFAs (Gly-, Gly-Ser-, Orn-, Asn-, Gln-FAHFA) were all found to possess a significantly greater molecular diversity than in mammals, making them valuable biomarkers.A summary of the specific lipid findings is presented in Table 7.

Limitations
Biological Limitations: Bacteria and fungi have complex life cycles and their lipidomes will vary with those cycles.We only take a snapshot of one point in time with our commercial samples.However, these analyses will allow us to define major unique lipid families in each bacterial strain.Bacterial expression in their hosts will be complex but a range of members of each unique lipid family has the possibility of being monitored and can be optimized by longitudinal sample collections [9].In addition, while we obtained detailed lipidomics data for 17 oral bacteria and 1 oral fungus, the oral cavity has over 600 bacterial species and 100 fungal species.This necessitates efforts to continue to expand this first lipidomics database for oral microbes.Such a database is essential for studies of bacterial adaptation to environmental changes, including the development of resistance to antibiotics.
Technical Limitations: Our HR-MS analytical platform (≤2 ppm mass error), which utilizes both PESI and NESI, significantly reduces the risk of lipid misassignments.However, there are a number of lipid structural isobars that require MS 2 and/or TLC evaluations for full structural validation.Over the last 10 years, our Metabolomics Unit has built a database of a number of these specific issues and optimal technical solutions.Specific issues include our inability to distinguish between: (i) a cyclopropyl group and a double bond in a fatty acid chain, and (ii) an added methyl group vs. addition of a CH 2 in a fatty acid.Again, MS 2 and/or TLC evaluations will be our first strategies with lipids of high interest.NMR may be considered if required, but this involves significant scale-up and purification methods due to the lower sensitivity of NMR.

Table 1 .
List of the commercial microflora utilized in this lipidomics study.

Table 2 .
Glycine hydroxy fatty acids (HFA), including commendamide (Gly-3-HFA 16:0), found in only 3 of the 8 Gram-negative bacteria.Data are presented as Rank orders with the relative levels of the most abundant family member to 2 nmoles of [ 13 C 3 ]DG 36:2 in parentheses.Blanks represent Not Detected.

Table 5 .
Relative levels of the most abundant PC family members to 2 nmoles of [ 13 C 3 ]DG 36:2.Relative LPC levels and the ratios of these to the PCs are presented.Blanks represent Not Detected.

Rank Orders of PC and LPC in Nanobacter lyticus.
product was (Phosphocholine = 184.0733)+ , indicating t T.

Table 6 .
Relative levels of MGCCs to 2 nmoles of [ 13 C 3 ]DG 36:2 in N. lyticus.Relative LPC levels and the ratios of these to the PCs are presented.

Table 7 .
A summary of the presence or absence of specific lipid families in the microbes of this study.Since we set a rigorous threshold for the peak intensity for individual lipids, this summary does not include potential lipids present at lower concentrations.Green = present; Red = not detected.