Present addresses: Sven Jechalke, Federal Research Center for Cultivated Plants, Institute for Epidemiology and Pathogen Diagnostics, Julius Kühn-Institute (JKI), Braunschweig, Germany. A.G. Franchini, Institute of Biogeochemistry and Pollutant Dynamics (IBP), Federal Institute of Technology (ETH), Zürich, Switzerland.
Correspondence: Present address: Felipe Bastida, Department of Soil and Water Conservation and Waste Management, Centro de Edafología y Biología Aplicada del Segura (CEBAS-CSIC), Campus Universitario de Espinardo, 30100 Espinardo, Murcia, Spain. Tel.: +34 968 396 106; fax: +34 968 396 213; e-mail: email@example.com
The flow of benzene carbon along a food chain consisting of bacteria and eukaryotes, including larvae (Diptera: Chironomidae), was evaluated by total lipid fatty acids (TLFAs)-, amino acid- and protein-stable isotope probing (SIP). A coconut-fibre textile, colonized by a benzene-degrading biofilm, was sampled in a system established for the remediation of benzene, toluene, ethylbenzene and xylenes (BTEX)-polluted groundwater and incubated with 12C- and [13C6]-benzene (>99 at.%) in a batch-scale experiment for 2–8 days. After 8 days, Chironomus sp. larvae were added to study carbon flow to higher trophic levels. Gas chromatography-combustion-isotope ratio monitoring mass spectrometry of TLFA showed increased isotope ratios in the 13C-benzene-incubated biofilm. A higher 13C-enrichment was observed in TLFAs, indicative of Gram-negative bacteria than for Gram-positive. Fatty acid indicators of eukaryotes showed significant 13C-incorporation, but to a lower extent than bacterial indicators. Fatty acids extracted from larvae feeding on 13C-biofilm reached an isotopic ratio of 1.55 at.%, illustrating that the larvae feed, to some extent, on labelled biomass. No 13C-incorporation was detectable in larval proteins after their separation by sodium-dodecyl sulphate-polyacrylamide gel electrophoresis and analysis by nano-liquid-chromatography-mass spectrometry. The flow of benzene-derived carbon could be traced in a food web consisting of bacteria and eukaryotes.
Stable isotope probing (SIP) methodologies are mainly based on the incubation of an environmental sample with a 13C-labelled substrate, followed by the analysis of 13C-incorporation in a biomarker. These approaches allow to link a particular biochemical process to the respective organism carrying it out (for a review, see Neufeld et al., 2007). For example, the use of 13C-labelled substrates in combination with fatty acid analysis (Boschker & Middelburg, 2002) allows the identification of the main microbial groups involved in the assimilation of toluene (Pelz et al., 2001) or benzene (Geyer et al., 2005). In addition, amino acid-SIP might provide quantitative information about the assimilation of 13C by different organisms (Pelz et al., 1997), but without taxonomical information. At the same time that microorganisms play an important role in the biodegradation of organic molecules, they also constitute a major carbon source for organisms of higher trophic levels. The fate of organic contaminants is therefore the result of interactions between microbial communities metabolizing organic compounds and grazers feeding on microbial biofilms. However, the flow of carbon channelled through a microbial community into a food web including eukaryotic organisms has not been sufficiently studied in aquatic habitats, while in terrestrial environments, there are a few examples using nucleic acids-SIP (Lüders et al., 2004, 2006) and fatty acid-SIP (Haubert et al., 2009; Ruess & Chamberlain, 2010) for the investigation of food webs. Fatty acid-SIP approaches have been used to elucidate the flow of methane derived from methane-oxidizing bacteria to chironomid larvae (Deines et al., 2007), the carbon transfer from benthic algae to bacteria (Middelburg et al., 2000) and the assimilation of toluene carbon along an artificial bacterial–protist food chain (Mauclaire et al., 2003).
Recently, the SIP of proteins (protein-SIP) has been developed and provides both phylogenetic and functional information of pure and mixed microbial cultures by analysis of labelled and nonlabelled peptides (Jehmlich et al., 2008a, 2010; Bastida et al., 2010). However, the potential of protein-SIP to study the carbon flow from microorganisms to higher organisms has not been explored so far.
The objective of this paper is to elucidate the flow of benzene carbon through a food web consisting of bacteria and eukaryotic organisms using different SIP techniques. For this purpose, a coconut-fibre geotextile colonized by an aerobic benzene degrading biofilm was sampled in an aerated trench system for remediation of BTEX-polluted groundwater (Jechalke et al., 2010) and incubated with 12C- or 13C-benzene in a laboratory-scale batch culture experiment. Chironomus sp. larvae naturally feeding on the biofilm in the trench were incubated in culture experiments to study carbon flow over multiple trophic levels. 13C-enrichment in the total lipid fatty acids (TLFAs) of textile and larvae as well as enrichment in amino acids and proteins of the larvae were analysed to obtain quantitative information on the carbon flow between microorganism and higher trophic levels. In a parallel study, DNA- and RNA-based molecular methods as well as proteomic approaches were applied for (1) analysis of the microbial community structure, (2) identification of microorganisms playing a key role in benzene biodegradation and (3) functional characterization of microbial degradation processes (S. Jechalke et al., unpublished data).
Materials and methods
Chemicals and solvents were obtained in p.A. quality from Merck unless otherwise stated. Nonlabelled benzene (referred as 12C-benzene throughout the manuscript) was obtained from Merck and had an isotope composition of −25.16‰, which corresponds to 1.08 at.%. [13C6]-benzene (referred to as 13C-benzene throughout the manuscript) was obtained from Sigma-Aldrich (St. Louis, MO) with a chemical purity >99% and an isotope purity >99 at.%.
Inflow water and textile for microcosms were obtained from an aerated treatment pond system located in Leuna, Germany. The trench system is percolated with anoxic groundwater loaded with benzene, methyl-tert-butyl ether, ammonium and Fe(II) in average concentrations of 0.3, 0.04, 3 and 0.1 mM, respectively (Jechalke et al., 2010). A detailed description of the field site is given elsewhere (Jechalke et al., 2010).
Fourteen sterilized 1 L Duran® bottles (Schott, Germany) were filled with 300 mL filtered (0.2 μm) inflow water and 9 g coconut textile with biofilm (wet weight, 64 ± 6% moisture) sampled in May 2009. Textile was taken from 60 cm depth of the textile carrier of the treatment pond system. Before sampling, this zone in the pond was characterized by average low oxygen concentrations of 0.3 ± 1 mg L−1. Bottles were closed with Teflon-coated butyl-stoppers (VWR International, Darmstadt, Germany) and screw caps. The residual benzene concentrations were between 0.037 and 0.044 mmol per bottle. Subsequently, 12C- or 13C-benzene were additionally added from 20 mM aqueous stock solutions to obtain final initial benzene concentrations between 0.069 and 0.101 mmol per bottle. Benzene was analysed by headspace gas chromatography according to Kleinsteuber et al. (2008). Six bottles were spiked with 13C-benzene and six bottles with 12C-benzene. Two additional bottles spiked with 13C-benzene were autoclaved as a control to check for potential abiotic benzene losses. Because of the remaining 12C-benzene stemming from inflow groundwater, the 13C-benzene microcosms contained approximately 50 at.% benzene at the beginning of the experiment (Supporting Information, Fig. S1). To mimic field conditions as far as possible, the microcosms received daylight and were incubated at 15 °C, which is similar to the average annual temperature of the pond water. The bottles were shaken at a low rotation speed (85 r.p.m.) on a horizontal shaker to enhance the oxygen transfer into the liquid phase while preserving the integrity of the biofilms. After the complete degradation of benzene, two 12C- and 13C-benzene-spiked microcosms were sacrificed for further analysis (t1, 2 days of incubation). The analysis comprised elemental analysis, TLFA analysis and TLFA-SIP analysis of the textile biofilm (Fig. 1). The remaining bottles were opened under a sterile laminar box for 10 min to exchange the air in the headspace. After closing the bottles again, the spiking was repeated four times more with intermittent degradation of benzene to <0.3 μmol per bottle. Hence, beginning with the second spiking, the 13C-benzene microcosms always received benzene with 99 at.%13C (Fig. S1). After 8 days of incubation (t2), two more 12C- and 13C-benzene microcosms were sacrificed, and subjected to the same analysis as the t1 samples (Fig. 1).
To each of the remaining two 12C- and 13C-benzene microcosms, 30 Chironomus sp. larvae were added and the microcosms were further incubated for 5 days. Then, larvae were harvested and subjected to elemental and TLFA-SIP analysis. In addition, elemental analysis, protein-SIP and amino acid-SIP analysis were performed with the proteins extracted from larvae (Fig. 1), as described below.
Protein extraction from larvae and elemental analysis of protein pellets
Larvae from duplicate bottles were carefully picked with tweezers and washed in a Petri dish with MilliQ water for gut clearance (Feuchtmayr & Grey, 2003) and removal of residual textile. For larval protein extraction, 0.15 g of larvae (about three larvae) were homogenized with a mortar. The suspension was transferred to a 1.5 mL tube and 100 μL of sodium dodecyl sulphate (SDS) buffer (20 mM Tris-HCl pH 7.5, 2% SDS) was added. The mixture was mixed in a vortex for 1 min and sonicated (cooled on ice) for 1 min (Bandelin Sonoplus HD70, cycle 50%). Subsequently, 600 μL of 20 mM MgCl2 solved in TRIS-HCl buffer, 1 μL of 100 mM phenylmethylsulphonyl fluoride (PMSF) and 1 μL of benzonase were added to the crude protein extract. The mixture was incubated for 10 min at 37 °C and then centrifuged for 10 min at 16 100 g and 4 °C. The supernatant containing the protein fraction was transferred to a microtube and proteins were precipitated by the addition of the same volume of 20% TCA and centrifugation for 10 min at 16 100 g and 4 °C. The resulting pellet was dissolved in 100 μL MilliQ-H2O and washed twice with 500 μL of 100% acetone. The protein concentration was determined according to Bradford (1976). From each sample, 80 μg of proteins were used for separation by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) as described elsewhere (Santos et al., 2007).
In order to analyse 13C-incoporation into the bulk pool of larval proteins, bulk protein extracts (200 μg) of larvae were precipitated with 20% trichloroacetic acid in water (v/v) by incubation at 4 °C overnight. After centrifugation at 16 100 g for 15 min, pellets were dissolved with 200 μL of bi-distilled H2O and 1 mL of cold acetone was added. After incubation for 15 min at −20 °C, a protein pellet was obtained by centrifugation as described above. Acetone was discarded by pipetting and acetone remains were evaporated for 1 h at 37 °C. Protein pellets were gently dissolved in 40 μL of bi-distilled H2O and placed in tin capsules. Then, 10 μL of acetone was added and evaporation of liquids was performed for 3 h at 60 °C. The dry protein larval pellets as well as the entire larval tissue and textile pieces were analysed for the carbon isotope composition using a Euro EA Elemental Analyzer (Eurovector, Germany) coupled to an IRMS (MAT 253, Thermo Fisher Scientific, Germany).
Protein preparation and nano-liquid chromatography-mass spectrometry (LC-MS/MS) analysis
Ten bands were excised from a SDS-PAGE performed with proteins extracted from larvae sampled at the end of incubation. Bands were de-stained and subsequently proteolytically digested with trypsin (Sigma-Aldrich, Munich, Germany) overnight at 37 °C according to Jehmlich et al. (2008b). The peptide mixture was extracted twice with 100% acetonitrile and afterwards concentrated by vacuum centrifugation for 15 min.
Peptides were analysed by nano-LC-MS (LTQ-Orbitrap mass spectrometer, Thermo) as described by Bastida et al. (2010). Raw data were applied to a database search using thermo proteome discoverer software (v1.0 build 43) to carry out a tandem ion search algorithm from the Mascot house server (v2.2.1) by database comparison against Diptera in the National Center for Biotechnology Information (NCBInr database 2009-08-24) with 10 p.p.m. tolerance for the precursor and 0.8 Da for MS2 fragments. Further, trypsin with a maximum of two missed cleavages was selected and variable modifications, such as methionine oxidation and carbamidomethylation of cysteine, were allowed. Larval peptides were considered to be identified by Mascot when a probability <0.05 (probability-based ion scores threshold >40) was achieved and at least two peptides were identified. Identification was confirmed in peptides from duplicated gels. For the calculation of the incorporation efficiency, we used the definition given by Snijders et al. (2005), which is based on the isotope pattern of the mass peaks. In addition, a perl script provided by Choudhary et al. (2006) was used. The calculation of the 13C incorporation was described by Jehmlich et al. (2008b).
Total lipids were extracted from 1 g textile using a modified Bligh and Dyer extraction method as described previously (Geyer et al., 2005). The lipids were then transesterified to fatty acid methyl esters (FAMEs) by a mild alkaline methanolysis (Guckert et al., 1985). The complete dried FAME fraction was dissolved in n-hexane containing 20.06 μg mL−1 of 21 : 0 FAME as an internal standard. For the identification and quantification of the FAMEs, a 7890A gas chromatograph coupled to a 5975C mass spectrometer (Agilent Technologies, Waldbronn, Germany) was used. FAMEs were identified by comparing mass spectra and retention times with authentic reference compounds (Bacterial Acid Methyl Ester, BAME Mix Cat. No. 47080-U, Supelco, Sigma Aldrich; and Supelco™ 37 Component FAME Mix Cat. No. 47885-U, Supelco, Sigma Aldrich). The carboxylic acid fractions were separated on an HP-5ms column (30 m × 0.25 mm × 0.25 μm) with helium as a carrier gas at a flow rate of 0.8 mL min−1. One-microlitre sample was injected with a split ratio of 1 : 3 using the following temperature programme: initial temperature of 70 °C held for 1 min, then heated to 150 °C at 20 °C min−1, heated to 280 °C at 4 °C min−1 and finally heated to 300 °C at 20 °C min−1 and held at 200 °C for 5 min. The fatty acids are designated in the form of A:BωC, where A is the number of carbon atoms, B is the number of double bonds and C is the distance of the closest double bond (unsaturation) from the aliphatic end (ω-nomenclature) of the molecule. The prefix ‘10Me’ refers to a methyl branch at the 10th carbon atom, where ‘i’ and ‘a’ refer to iso and anteiso methyl branching; ‘cy’ indicates a cyclopropyl fatty acid. The absolute and relative amounts of TLFAs FAMEs in the samples were determined according to the concentration of the internal standard added.
The Gram-positive indicator fatty acids i14:0, i15:0, a15:0, i16:0, i17:0 and a17:0 and the Gram-negative indicator fatty acids cy17:0, 18:1ω9c, 16:1ω7c and 16:1ω7t were used as relative indications of their respective biomass and for the calculation of the ratio between Gram-positive and Gram-negative bacterial biomass that was used for the characterization of the microbial community composition (Ringelberg et al., 1989; Zelles, 1997; Tscherko et al., 2004; Leckie, 2005). We are aware that the separation of Gram-negative and Gram-positive bacteria by fatty acids should be carried out with caution because new literature demonstrates that some fatty acids are not always specific, but may be used for our relatively simple system. For example, recently, i15:0 fatty acid has been found to be a dominant fatty acid in Acidobacteria and other Gram-negative bacteria (Fukunaga et al., 2008; Kulichevskaya et al., 2010). Nevertheless, considering that the isotopic ratio found in this fatty acid was similar to those of typically Gram-positive fatty acids, we included the i15:0 fatty acid in the group of Gram-positive bacteria. Similarly, 20:5ω3 and 20:4ω6 fatty acids were taken mainly as a eukaryotic origin. In the literature, 20:4ω6 has been used as an indicator of protists (Mauclaire et al., 2003) and animals (Chen et al., 2001; Ruess & Chamberlain, 2010). The 20: 5w3 fatty acid was taken as an indicator of algae (Ruess & Chamberlain, 2010), although it has also been found in different bacterial groups (Freese et al., 2009). The visible abundance of algae in the mats and the 13C-labelling pattern support the use of this fatty acid as an indicator of algae in our system.
Amino acid analysis
The amino acids in the larval protein extract were determined according to the method of Macko et al. (1997) and Silfer et al. (1991). From each sample, a pellet of 100 μg of proteins (obtained as described above) was hydrolysed with 0.5 mL of 6 M HCl at 110 °C for 22 h. Hydrolysates were evaporated to dryness under an N2 atmosphere and then cooled in an ice bath for 5 min. For derivatization, samples were incubated with 1 mL of isopropanol and 0.25 mL of acetyl chloride overnight at 60 °C to obtain isopropylesters. The esterified samples were evaporated to dryness under N2. Amino acid isopropyl esters were acetylated with 500 μL of dichloromethane (CH2Cl2) and 500 μL of trifluoroacetic anhydride (TFAA) for 2 h at 60 °C. After removing residual CH2Cl2 and TFAA by evaporation under N2, the derivatives were dissolved in 100 μL dichloromethane. An external standard (50 μL) containing 50 mg mL−1 of each amino acid was treated in an identical manner as the samples for the identification of amino acids.
The amino acid-trifluoroacetyl isopropyl esters (derivatized amino acids) were analysed by gas chromatography–mass spectrometry using the same GC-MS system as that described for fatty acid analysis. The derivatized amino acids were separated on an HP-5ms column (30 m × 0.25 mm × 0.25 μm) by the injection of 1 μL sample with a split ratio of 15 : 1 using the following temperature programme: initial temperature of 50 °C held for 1 min, then heated from 50 to 100 °C at 30 °C min−1 and held for 5 min, then heated to 200 °C at 10 °C min−1 and held for 5 min, and finally heated to 300 °C at 30 °C min−1 and held for 12 min. Helium was used as a carrier gas at a flow rate of 0.9 mL min−1. The individual amino acids were identified by comparison of the retention times and mass spectra with those of the standard mixture.
Isotope analysis of fatty acids and amino acids
Carbon isotope compositions of FAMEs and amino acids derivatives were analysed using a gas chromatography-combustion-isotope ratio monitoring mass spectrometry (GC-C-IRMS) system. The system consisted of a gas chromatograph (6890 series, Agilent Technology, Palo Alto, CA) coupled via a Conflow III interface (Thermo Electron, Germany) to a MAT 253 mass spectrometer (Thermo Electron). A BPX5 column (50 m × 0.32 mm × 0.5 μm; SGE, Darmstadt, Germany) was used for chromatographic separation with helium as a carrier gas at a flow rate of 1.5 mL min−1 for fatty acids and of 2 mL min−1 for amino acids. For fatty acid analysis, the following temperature programme was used: initial temperature of 70 °C held for 1 min, then heated to 160 °C at 20 °C min−1, then heated to 200 °C at 2 °C min−1 and held for 5 min, then heated to 220 °C at 2 °C min−1 and held for 5 min, then heated at 2 °C min−1 to 270 °C, and finally heated to 300 °C at 20 °C min−1 and held for 10 min. Amino acids were analysed using the following temperature programme: initial temperature of 50 °C held for 1 min, heated to 100 °C at 15 °C min−1 and held for 5 min, then heated to 175 °C at 5 °C min−1 and held for 5 min, then heated to 250 °C at 10 °C min−1 and held for 5 min, and finally heated to 300 °C at 30 °C min−1 and held for 12 min. Samples were injected using either the splitless mode or with a split ratio of 1 : 5.
The carbon isotope ratio of fatty acids and amino acids are reported in the δ notation (per mil) relative to the Vienna Pee Dee Belemnite standard (V-PDB) (Coplen et al., 2006) according to
where Rsample and Rreference are the ratios of the heavy isotope to the light isotope (13C/12C) in the sample and in the international standard, respectively. Samples were measured at least in triplicate, with an analytical error smaller than ± 0.5‰ SD. The δ13C values of the fatty acids and amino acids reported were corrected for the carbon introduced during derivatization (Silfer et al., 1991; Abraham et al., 1998). Because the isotope compositions of low abundant FAMEs could not be determined by GC-C-IRMS, the isotope enrichments of those fatty acids were calculated based on mass spectrometric analysis as described recently (Bombach et al., 2010).
To convert δ13C to atom%13C, the following equation was used:
where δ is the measured δ13C (‰) of the sample using the δ notation and RV-PDB is the isotope ratio of V-PDB=0.0112372 (Slater et al., 2001).
The total amount of 13C in fatty acids was calculated as described by Boschker (2004). The percentage of 13C incorporated into specific Gram-positive (i15:0, a15:0, i16:0), Gram-negative (cy17:0, 18:1ω9c, 16:1ω7c and 16:1ω7t) and eukaryotic (20:4ω6 and 20:5ω3) fatty acids identified by GC-C-IRMS compared with the total amount of 13C in fatty acids was used as an indication of the flow of 13C in the different biomass compartments.
Carbon isotope analysis of textile biofilm and larvae
To determine the isotope ratio of the total organic carbon of textile biofilm and larvae, textile pieces (around 200 mg) and two larvae were air-dried for 25 h at 105 °C and subsequently subjected to elemental analysis in a Euro EA Elemental Analyzer (Eurovector) coupled to an IRMS (MAT 253, Thermo Fisher Scientific). Measured values were expressed in at.% relative to V-PDB as described elsewhere (Coplen et al., 2006).
In order to determine pairwise differences by post hoc tests, the data of relative abundance of fatty acids and bulk analysis were subjected to one-way anova. The post hoc test applied was Fisher's least significant difference (LSD) method at the 95% confidence level. The software used for the statistical analysis was statgraphics plus 2.1.
13C incorporation into microbial biofilm and larvae tissue revealed by elemental analysis and SIP of amino acids
The bulk analysis of the δ13C values of the biofilm grown on textile showed significant enrichment (P<0.05) compared with the textile incubated with 12C-benzene (Table 1). The higher δ13C enrichment found at t2 compared with t1 indicates a continuous incorporation of 13C into the biofilm. Larvae harvested from the 13C-labelled textiles after five days of incubation were 13C-enriched. On the contrary, larvae not incubated or larvae harvested from textiles incubated with 12C-benzene were not enriched in 13C (Table 1).
Table 1. Carbon stable isotope composition (atom%) of biomass from textile and chironomid larvae before and after incubation
Data are average of duplicates and the standard deviation of the mean.
Sample identification: TC, textile control; T1-12C, textile time-point 1 incubated with 12C-benzene; T1-13C, textile time-point 1 incubated with 13C-benzene; T2-12C, textile time-point 2 incubated with 12C-benzene; T2-13C, textile time-point 2 incubated with 13C-benzene; L, larvae control; L-12C, larvae incubated with non-enriched textile; L-13C, larvae incubated with enriched textile.
Data followed by the same letter are not significantly different according to the LSD test (P≤0.05).
1.09 ± 0.00023a
1.08 ± 0.0029a
2.3 ± 0.273c
1.08 ± 0.0055a
6.4 ± 0.068d
1.09 ± 0.00023a
1.09 ± 0.00062a
1.78 ± 0.1276b
As mass spectrometry analysis using LTQ-Orbitrap indicated that the13C-enrichment within larval-specific peptides was below the detection limits of incorporation (2 at.%, Jehmlich et al., 2010), we aimed to analyse the carbon isotope composition of the bulk protein pool and the individual amino acids in the hydrolysate of the larvae protein pellet by GC-C-IRMS. The analysis of the larval protein pellet by elemental analysis revealed a significant enrichment compared with the one of 12C-benzene-incubated larvae (Table 1), but lower than the entire larvae tissue (1.8 at.%). In parallel, several amino acids (Gly, Ser, Val, Leu, Ile, Pro, Asp, Gln, Phe and Tyr) from larval protein pellet hydrolysates were analysed by GC-C-IRMS. Significant (P<0.05) 13C-enrichments in amino acids of larvae incubated with 13C-enriched textile were found in all amino acids analysed (Table 2). In contrast, larvae not incubated in benzene-amended microcosms as well as larvae incubated with 12C-benzene-amended textiles showed δ13C values of 1.08 at.% in both cases.
Table 2. Carbon stable isotope composition (atom%) of the larval protein pellet and the individual larval amino acids
The data are means of duplicate extractions measured three times and the standard deviation of the mean.
Data followed by the same letter are not significantly different according to the LSD test (P≤0.05).
L, larvae control; L-12C, larvae incubated with textile biofilms from the 12C experiment; L-13C, larvae incubated with textile biofilms from the 13C experiment.
1.09 ± 0.00042
1.09 ± 0.00025
1.52 ± 0.08173
1.08 ± 0.00054a
1.08 ± 0.00151a
1.14 ± 0.00210b
1.08 ± 0.00017a
1.08 ± 0.00062a
1.22 ± 0.00311b
1.08 ± 0.00048a
1.08 ± 0.00071a
1.27 ± 0.02337b
1.07 ± 0.00064a
1.09 ± 0.00084a
1.18 ± 0.07550a
1.07 ± 0.00026a
1.09 ± 0.0194a
1.38 ± 0.09094b
1.08 ± 0.00082a
1.08 ± 0.00209a
1.25 ± 0.01450b
1.08 ± 0.00022a
1.08 ± 0.00060a
1.16 ± 0.00430b
1.08 ± 0.00016a
1.08 ± 0.00021a
1.33 ± 0.00699b
1.07 ± 0.00079a
1.07 ± 0.00068a
1.31 ± 0.00484b
1.07 ± 0.00050a
1.07 ± 0.00018a
1.56 ± 0.10815b
Pattern and carbon isotope signatures of TLFA in textile and larvae samples
The TLFAs were analysed to trace the assimilation of benzene-derived carbon through multiple trophic levels. TLFA comprises fatty acids deriving from cellular membranes and storage compounds from living and dead microbial cells, as well as from eukaryotic tissues, thus allowing to trace the carbon flow between different trophic levels. Hence, we use TLFA because it may show the carbon fluxes between the trophic levels more quantitatively than phospholipid fatty acids that are characteristic for living biomass.
The TLFA profile of the textile biofilm comprised 37 fatty acids (Table 3). The major fatty acids identified in the textile include the fatty acids 16:1ω7c, 16:0 and 18:1ω9. Gram-negative (cy17:0, 18:1ω9c, 16:1ω7c and 16:1ω7t) and Gram-positive fatty acids (Tscherko et al., 2004; Leckie, 2005) represented 25.3% and 17.6% of the amount of fatty acids in the initial community, respectively. Polyunsaturated fatty acids (20:4ω6 and 20:5ω3), which mainly derive from eukaryotes, were detectable in a small amount compared with the total amount of fatty acids (1.5%). Significant differences in the community composition appeared when the TLFA profiles of the incubation experiment were compared with those of the initial textile biofilm, indicating a strong influence of benzene on the community composition. For instance, the ratio of Gram-negative to Gram-positive fatty acids was significantly higher (P<0.05) in the microcosms than in the original textile biofilm (Table 3). No significant differences in this ratio were observed between t1 and t2, indicating a certain stability of the community during the incubation with benzene.
Table 3. Average relative abundance (%) of total lipid fatty acids derived from textile biofilm and larvae before and after incubation
The TLFA profile of the larvae comprised 30 fatty acids, but no specific biomarker for larvae was found. The most abundant fatty acids comprised the saturated fatty acid 16:0, the monounsaturated fatty acids 16:1ω7c and 18:1ω9c/t and the polyunsaturated fatty acid 18 : 2ω6,9. In comparison with the textile community, the proportions of the fatty acids with 18 carbons to the total fatty acids were significantly higher in the larvae (37.5% vs. 16.8%) (Table 3).
All fatty acids extracted from non-benzene-incubated control samples and samples from microcosms incubated with 12C-benzene showed natural isotope signatures in the range of 1.06–1.08 at.%. In microcosms amended with 13C-benzene, all fatty acids with the exception of 20:5ω3 showed isotope enrichments in the range of 1.1–4.8 at.% after 2 days of incubation (t1) (Fig. 2). Significantly higher isotope ratios with values between 1.4 and 19.3 at.% were reached after 8 days of incubation (t2) (Fig. 2). Within the microbial community, fatty acids representative of Gram-negative bacteria were around two times higher labelled at t1 and 10 times higher labelled at t2 than those of the Gram-positive bacteria. At t2, fatty acids such as 16:1ω7c, 16:1ω7t, cy17:0 and 18:1ω9c indicative of Gram-negative bacteria showed an enrichment of 17.8, 17.7, 16.6 and 16.4 at.%, respectively, while fatty acids of Gram-positive bacteria showed enrichments up to 4.8 at.% (Fig. 2). At least 57% and 67% of the total 13C in fatty acids was found in those Gram-negative fatty acids (cy17:0, 18:1ω9c, 16:1ω7c and 16:1ω7t) at t1 and t2, respectively. However, only 4% and 3% of the total 13C in fatty acids were found in Gram-positive fatty acids (i15:0, a15:0, i16:0) at t1 and t2, respectively. In contrast to bacterial fatty acids, long straight-chain fatty acids and polyunsaturated fatty acids, indicative of eukaryotes, were labelled to a much lower extent. For instance, at t2, the fatty acids 20:4ω6 and 20:5ω3, typical of eukaryotes, showed 13C-enrichments of 2.2 and 1.5 at.%, respectively. When considering exclusively the fatty acids 20:4ω6 and 20:5ω3, up to 0.13% of the total 13C in all fatty acids was found in these eukaryotic biomarkers.
The natural abundance of 13C in the fatty acids extracted from larvae ranged between 1.06 and 1.08 at.%. When larvae were incubated with 13C-enriched textile, several fatty acids extracted from the larvae showed significant 13C-enrichment, but to a smaller extent than in the textile samples. A maximum 13C-enrichment equivalent to 1.5 at.% was reached in the fatty acid 16:1ω7c (Fig. 2). No significant isotope enrichment was detected in 24:1, 20:5ω3, 18:2ω6,9 and 15:0, suggesting that these fatty acids were not synthesized using 13C-labelled microbial biomass or 13C-labelled benzene.
SIP of proteins
Structural proteins, translation factors, ribosomal proteins and proteins involved in carbohydrate and energy metabolism could be identified using LC-MS/MS analysis of larvae peptides from 10 SDS-PAGE bands and using a Diptera NCBInr database search (Fig. S2, Table S2). The 32 identified larval proteins did not reveal 13C enrichments above 2 at.% in the 95 peptides that have been examined.
In order to elucidate the flow of benzene carbon through a food web consisting of bacteria and eukaryotic organisms, the assimilation of benzene-derived 13C-carbon was traced using different SIP methodologies. Bacterial fatty acids were significantly enriched in 13C after 2 and 8 days of incubation, revealing an assimilation of benzene-derived carbon into the microbial biomass. Several authors have proposed different fatty acids as specific or dominant in Gram-positive and Gram-negative bacteria (Ringelberg et al., 1989; Zelles, 1997; Tscherko et al., 2004; Leckie, 2005). Although 18:1ω9c has been conventionally used as an indicator for fungi (Baath, 2003; Waldrop & Firestone, 2004), several researchers have found it also in bacteria (Ringelberg et al., 1989) and have used this fatty acid as an indicator of Gram-negative bacteria (Tscherko et al., 2004). 18:1ω9c showed a high enrichment that fits with the enrichment found in fatty acids typical for Gram-negative bacteria (Fig. 2), for example the cy17:0, 16:1ω7c and 16:1ω7t.
The higher enrichment of fatty acids indicative of Gram-negative up to 17.7 at.% compared with Gram-positive bacteria up to 4.8 at.% suggests that Gram-negative bacteria played a major role in the benzene metabolism. In fact, at t2, 67% of the total 13C in fatty acids was found in Gram-negative fatty acid indicators, which suggests a predominant use of 13C-benzene by this microbial group and a minor use of other carbon compounds from the original inoculum. In contrast, 3% of the total 13C amount in fatty acids was found in Gram-positive indicators. This indicates that Gram-positive bacteria did not play an important role in the biodegradation of benzene and might predominantly use other carbon sources of the original inoculum.
The high labelling of typical Gram-negative fatty acids, reaching the highest δ13C values of 17.8 at.%, is in agreement with the RNA-SIP results described by S. Jechalke et al. (unpublished data) that allowed the identification of microbial key players of benzene biodegradation within the complex microbial community of the trench system. Out of a clone library of 131 distinct sequences, S. Jechalke et al. (unpublished data) identified 18 sequences in the heavy 13C-labelled fraction of RNA that belonged to Gram-negative organisms: Zooglea, Aquabacterium, Hydrogenophaga, Leptothrix, Dechloromonas, Rhodoferax and unclassified members of the Comamonadaceae family. However, Gram-positive organisms were not detected in the heavy RNA fraction.
Heterotrophic protists and other eukaryotic organisms in the system can obviously utilize a wide variety of food sources including algae, bacteria, detritus and dissolved organic matter. As intermediates in the food web, these organisms provide a critical link in the flow of carbon and nutrients between bacteria and higher trophic levels (Lund et al., 2008). The labelling of specific fatty acids may show to what extent eukaryotes are feeding on benzene-derived biomass, because, until now, benzene degradation capacity has not been described for any of them. The polyunsaturated fatty acid 20:4ω6 is frequently used as a biomarker of protists (Vestal & White, 1989; Mauclaire et al., 2003) and it is also widespread in animals (Ruess & Chamberlain, 2010). It was significantly enriched (2.2 at.%) compared with its natural abundance of 13C (1.07 at.%), but to a much smaller extent than bacterial fatty acids. Similarly, Mauclaire et al. (2003) traced the flux of toluene carbon through an artificial food chain consisting of Pseudomonas sp. and the protist Vahlkampfia sp. by using PLFA-SIP. These authors found δ13C values of about 200‰ in 20:4ω6 (equivalent to 1.3 at.%), which is a lower enrichment than that found in the same fatty acid (2.2 at.%) in our experiment. Recently, an average 13C-incorporation of 51 at.% was found in fatty acids of the ciliate Uronema sp. when the grazing rates of 13C-labelled Eschericha coli were analysed in a waste water system. The significantly higher 13C enrichment compared with that in our study arises from the fact that almost fully 13C-labelled E. coli was the unique carbon source for the Uronema sp. growth while in our study 13C-labelled, but also nonlabelled bacteria served also as carbon sources for protists (Kuppardt et al., 2010).
The polyunsaturated fatty acid 20:5w3 is generally common in algae (Ruess & Chamberlain, 2010), but it has also been found in some bacteria from polar regions, deep sea and anoxic sediments (Freese et al., 2009). The lack of incorporation of 13C carbon atoms in this fatty acid after 2 days of incubation is likely due to its algal origin. After 8 days of incubation, the low enrichment found in this fatty acid is indicative of certain 13C-transfer to eukaryotic trophic levels as demonstrated for 20:4ω6.
Only a minor part of the 13C found in all fatty acids was recovered in specific eukaryotic fatty acids (0.13%). This implies that higher trophic levels such as algae, protists and other microfauna may use labelled bacterial biomass as a food source, but dominantly graze on nonlabelled biomass and carbon sources from the original inoculum or from bacteria that were not involved in benzene degradation during the SIP incubation.
Chironomid larvae have been found quite often in water environments, such as ponds and lakes (Bazzanti et al., 2008; Campbell et al., 2009), and were also found in large numbers in the treatment pond, suggesting a significant role within the carbon flow of the system. Hence, the third trophic level under study in our system is constituted by this invertebrate group. Recently, Deines et al. (2007) provided evidence for the assimilation of methane-derived carbon by chironomid larvae, probably via assimilation of methane-oxidizing bacteria. In our case, the analysis of larval tissue by an elemental analyser indicated a significant enrichment (1.78 at.%), pointing to a carbon flow from the textile community to larvae. In addition, several fatty acids were significantly enriched when larvae were incubated together with the 13C-enriched community. However, similar to previous work (Deines et al., 2007), no specific fatty acids for chironomid larvae were found and we cannot totally discard that labelling was due to symbionts living on the larval surface or bacteria or protozoa only ingested, but not digested by the larvae. Nevertheless, there is evidence that supports the hypothesis of a carbon flow from microorganisms to larvae. For instance, larval protein pellet and individual amino acids were 13C-enriched. In addition, the pattern of enrichment in fatty acids was not the same for larvae and biofilm and further supports an enrichment in larvae that is not due to living symbionts, but the metabolism of carbon sources. For example, 14:0 and 15:0 fatty acids were partially enriched in the biofilm, but not in the chironomid larvae, even when these fatty acids also occurred in bacteria. However, one should keep in mind that dietary routing of fatty acids from resource to consumer can occur. Dietary routed fatty acid are incorporated into body tissues without modification, thus having a similar isotopic ratio as the one in the diet, while a de novo synthesized fatty acid is usually depleted in 13C relative to its precursor (Ruess & Chamberlain, 2010). Moreover, 20:0 fatty acid was enriched (1.3 at.%) in larvae, but their enrichment in the biofilm was not detected by GC-C-IRMS due to its low amount (Table 3). These results suggest that fatty acid enrichment in larvae occurs by assimilation of labelled biomass from the biofilm and de novo biosynthesis of lipids. Fatty acids extracted from larvae were 13C-enriched up to 1.5 at.% and this could indicate that larvae feed mostly on nonlabelled biomass that has been transferred to our experiment from the original inoculums (including dead cells, extracellular carbohydrates, etc.), but still some 13C-enriched biomass consisting of benzene degraders or cross-feeding species have acted as carbon sources for larvae. In addition, the relatively low labelling of the original larval fatty acids may be explained by the longer generation time of eukaryotic organisms compared with bacterial cells.
In addition to fatty acid analyses, protein-SIP approaches may be used to determine carbon flow into biomass, providing functional information about the various species within a biofilm. 13C-incorporation into some microbial proteins reached 96.8 at.% (S. Jechalke et al., unpublished data) and was similar to previous studies using pure cultures (Jehmlich et al., 2008a, b). However, the incorporation of 13C isotopes into larval proteins could not be quantified by protein mass spectrometry (protein-SIP). An incorporation of 13C into proteins lower than 2 at.% is difficult to detect by protein-SIP (Jehmlich et al., 2010). Nevertheless, this could be overcome by isolation of the protein pool of larvae and analysis of the amino acid released by acid digestion. The enrichment in those amino acids was confirmed by isotope mass spectrometry (GC-C-IRMS) because an incorporation of <0.1 at.% can be resolved by this technique. In this work, labelling in single larval amino acids ranged from 1.14 to 1.56 at.% for glycine and tyrosine, respectively. These results support a carbon transfer from biofilm biomass to larvae.
In summary, the use of different complementary SIP methodologies allowed tracing the flow of benzene carbon along a complex food web. Fatty acid and amino acid-SIP provides quantitative evidence of a carbon flow between different trophic levels. Although carbon flow from bacteria to eukaryotes could not be demonstrated by protein-SIP, the key degraders could be identified as Gram-negative bacteria by TLFA-SIP.
This work was supported by the Helmholtz Centre for Environmental Research-UFZ in the scope of the SAFIRA II Research Programme: Revitalization of contaminated Land and Groundwater at Megasites, subproject ‘Compartment transfer’ (Cotra). We are grateful to M. Gehre and U. Günther for their technical support. We are grateful for support from Marie Curie Mobility Actions of the European Commission: Host Fellowship for the Transfer of Knowledge (ToK) ISOTONIC Project (MTKD-CT-2006-042758) (F.B. and A.G.F.).