|Home | About | Journals | Submit | Contact Us | Français|
Monitoring the activity of target microorganisms during stimulated bioremediation is a key problem for the development of effective remediation strategies. At the US Department of Energy's Integrated Field Research Challenge (IFRC) site in Rifle, CO, the stimulation of Geobacter growth and activity via subsurface acetate addition leads to precipitation of U(VI) from groundwater as U(IV). Citrate synthase (gltA) is a key enzyme in Geobacter central metabolism that controls flux into the TCA cycle. Here, we utilize shotgun proteomic methods to demonstrate that the measurement of gltA peptides can be used to track Geobacter activity and strain evolution during in situ biostimulation. Abundances of conserved gltA peptides tracked Fe(III) reduction and changes in U(VI) concentrations during biostimulation, whereas changing patterns of unique peptide abundances between samples suggested sample‐specific strain shifts within the Geobacter population. Abundances of unique peptides indicated potential differences at the strain level between Fe(III)‐reducing populations stimulated during in situ biostimulation experiments conducted a year apart at the Rifle IFRC. These results offer a novel technique for the rapid screening of large numbers of proteomic samples for Geobacter species and will aid monitoring of subsurface bioremediation efforts that rely on metal reduction for desired outcomes.
The enrichment of Fe(III)‐reducing Geobacter strains has been reported in a variety of environments undergoing bioremediation, ranging from sediments contaminated with de‐icing run‐off (Holmes etal., 2007), landfill leachate (Roling etal., 2001) and petroleum (Rooney‐Varga etal., 1999) to uranium (Anderson etal., 2003), nitrate and technetium (Istok etal., 2004; North etal., 2004). At the Department of Energy (DOE) Integrated Field Research Challenge (IFRC) site in Rifle, CO, the enzymatic reduction of soluble U(VI) to insoluble U(IV) by stimulated indigenous Geobacter species has emerged as a promising remediation strategy. At the site, acetate addition to the subsurface results in a ‘bloom’ of Geobacter microorganisms that is correlated with a decrease in U(VI) concentrations to below the maximum contaminant level (Anderson etal., 2003; Vrionis etal., 2005). As biostimulation progresses, the subsurface microbiology shifts from this Fe(III)‐reducing community to a sulfate reducing community, whose activity results in elevated sulfide production. In some field experiments, this transition is associated with a decrease in the efficiency of U(VI) removal from groundwater. Thus, to ensure the efficacy of this process and better understand how it can be improved across the treatment area, previous research has focused on detecting the activity of Geobacter species and the evolution of the subsurface community as biostimulation progresses (Holmes etal., 2004; Vrionis etal., 2005; Mouser etal., 2009; Wilkins etal., 2009).
Proteogenomic analysis to date at the Rifle site has studied a ‘global’ view of the Geobacter community during the biostimulation campaigns conducted during both 2007 and 2008 (Wilkins etal., 2009; S.J. Callister, M.J. Wilkins, C.D. Nicora, K.H. Williams, J.F. Banfield, N.C. VerBerkmoes et al., submitted). While these analyses revealed a significant amount of data about the community physiology and strain make‐up during acetate addition, time‐intensive computational searching of the data means that this technique is excessively detailed for rapid screening of Geobacter species in large numbers of samples. Instead, a biomarker that quickly allows the identification of Geobacter activity in a sample is needed. Biomarker studies of these populations at the mRNA level have focused on the potential for a range of genes to track Geobacter activity in the subsurface. The nitrogen fixation gene, nifD, has shown promise in this role (Holmes etal., 2004), while Holmes and colleagues (2005) demonstrated that the citrate synthase gene, gltA, could be used to estimate the rates of Geobacter metabolism in a number of environments. However, the extraction of mRNA from large numbers of samples is a difficult and time‐intensive process, and has associated problems of rapid mRNA degradation. In contrast, shotgun proteomic techniques have the potential to offer a faster method of analysis due to high‐throughput pipelines in place at a number of institutions (Lipton etal., 2002; Callister etal., 2006a; VerBerkmoes etal., 2008; Wilkins etal., 2009).
The citrate synthase protein (subsequently referred to as CS in the text), which is responsible for controlling flux into the TCA cycle by catalysing the condensation of acetyl‐CoA and oxaloacetate to produce citric acid, has a number of characteristics that make it a suitable candidate as a Geobacter‐specific peptide‐based biomarker. The amino acid sequence in members of the Geobacteraceae is more closely related to a eukaryotic CS than other prokaryotic sequences (Bond etal., 2005), limiting the potential for false positive identifications from other subsurface species. Within the Geobacteraceae however, certain regions of the protein are highly conserved (Butler etal., 2010) and have the potential to act as biomarkers for general Geobacter abundance and activity (Fig.1). While other proteins contain highly conserved regions that could potentially act as Geobacter biomarkers, these regions frequently match other closely related species, such as Pelobacter and Desulfuromonas. In addition, the analysis of more divergent regions of the CS protein sequence shows potential as a technique to track strain‐level changes within the Geobacter community. Clear shifts in CS unique peptide abundances and diversity are observed over the duration of biostimulation, and therefore may be used to ‘fingerprint’ the microbial community at a specific period during the biostimulation process. As these community ‘fingerprints’ may be characteristic of certain time points during the biostimulation process, they are potentially useful indicators of changes in the biogeochemistry of the system (Wilkins etal., 2009). Finally, the stimulation of microbial growth in the Rifle subsurface via acetate amendment ensures that proteins involved in the efficient utilization of this substrate (e.g. TCA cycle proteins) are abundant within proteomic samples. Citrate synthase peptides are therefore suitable for roles as biomarkers given that they are easily detected where Geobacter species are active. It is likely that the reduction of contaminants such as U(VI) is tightly linked to respiratory processes (Lovley etal., 1991; Gorby and Lovley, 1992), and therefore, the fluctuating abundance of a protein involved in key metabolic processes such as citrate synthase will act as an effective proxy for Geobacter activity.
Here, we have applied these principles to field data obtained during in situ biostimulation experiments at Rifle during two successive field seasons (2007 and 2008). Results indicate that certain highly abundant conserved peptides can act as indicators of Geobacter activity, while abundance patterns for unique peptides matching CS can act as a community ‘fingerprint’ that can then be linked to specific periods of the biostimulation process. The development of a mass‐tag search database containing >200 Geobacter CS amino acid sequences will allow future proteomic samples to be rapidly screened for Geobacter activity and community composition, two factors that have a significant impact on U(VI) removal efficiency.
For this study, a proteomic search database consisting of CS sequences from metagenomic sequence, isolate Geobacter genomes and over 200 cloned Geobacter copies of the protein, was generated. Within such a proteomic search database, unique peptides are defined as being present in only one species and only one protein. In contrast, non‐unique peptides are defined as being present two times or greater, either in different proteins within the same organism, or within copies of a protein in multiple organisms. Alignment of the CS sequences from this search database and other bacterial genomes, together with in silico tryptic digests, identified multiple peptides that were highly conserved within Geobacter species but absent from other prokaryotic and eukaryotic gltA sequences (Fig.1A and B). These sequences were the candidates for general Geobacter biomarker peptides. Analysis of the ‘global’ proteomic datasets from the 2007 and 2008 biostimulation campaigns revealed that a subset of these conserved peptides were the most abundantly detected where Geobacter species were present. Peptides with sequences TIPETFEALPK, SLVTDISYLDPQEGIR and QVVPEYVYTAVR were selected on this basis as the conserved ‘Geobacter’ peptide markers. BlastP analysis of these sequences against the NCBI database revealed that the only significant matches to these sequences were peptides from Geobacter species. From a total of 232 citrate synthase sequences in the mass tag database, the ‘TIPETFEALPK’ peptide was present 171 times, the ‘SLVTDISYLDPQEGIR’ peptide was present 166 times, and the ‘QVVPEYVYTAVR’ peptide was present 124 times.
In silico tryptic digests also revealed a large number of unique peptides present within the mass‐tag database constructed from CS sequences. Because these peptides are unique to one sequence in the database they can be used to constrain the genotype of the strains present in the subsurface at any time point. These peptides come from more divergent regions of the CS protein (Fig.1).
Three proteogenomic datasets were collected during a 13‐day period of the 2007 biostimulation field campaign. As explained in Wilkins and colleagues (2009), all were taken during the period of dominant Fe(III) reduction in the subsurface. The heat map in Fig.2 illustrates the presence of Geobacter species in all samples, as identified by high abundances of conserved CS peptides (TIPETFEALPK, SLVTDISYLDPQEGIR and QVVPEYVYTAVR). Evidence for enzymatic Fe(III) reduction is present in all three samples in the form of aqueous Fe(II), while a rapid decrease in U(VI) concentrations is observed in the later D07 sample (Fig.2A). Acetate amendment was stopped in this biostimulation experiment before the Fe(III)‐reducing system transitioned into sulfate‐reducing conditions.
During the 2008 biostimulation experiment, acetate addition to the same region of the subsurface stimulated a period of dominant Fe(III) reduction, followed by a transition period where the subsurface geochemistry shifted towards dominant sulfate‐reducing conditions. Nine planktonic biomass samples were collected from downgradient well D04 during the experiment and therefore represent the microbial community changes that accompany the geochemical shifts. Previous research has shown that a microbial community shift from Fe(III)‐reducing microorganisms to sulfate‐reducing bacteria occurs over this transition period (Vrionis etal., 2005). Additionally, U(VI) removal from groundwater is typically most effective during this period of dominant Fe(III) reduction, with rates decreasing during the transition period (Anderson etal., 2003). However, geochemical measurements alone cannot accurately predict key process changes, due to the complex cycling that occurs between reduced Fe and S species. During periods of dominant Fe(III) reduction, the precipitation of FeS via the reaction of biogenic sulfide with Fe(II) can prohibit the detection of low‐level sulfate reduction. Conversely, during periods of dominant sulfate reduction in the aquifer, low‐level Fe(III) reduction is difficult to detect due to the titration of Fe(II) from solution by S2−. It is therefore of key importance to be able to quickly screen microbial biomass for organisms of interest, which in this case are the dominant Geobacter communities that develop following acetate amendment.
Conserved peptide abundances indicate planktonic Geobacter activity during the first half of the biostimulation experiment, when the system is apparently dominated by Fe(III) reduction and U(VI) concentrations are lowest (samples D04_day5 through D04_day20) (Fig.3A). Analysis of the sample D04_day23, taken shortly after the cessation of acetate addition, suggests that planktonic Geobacter populations are starting to decrease, and evidence of the biomarker peptides has mainly disappeared in the following sample, D04_day27. As the system geochemistry begins to transition into sulfate reducing conditions (identified by rising S2− concentrations and decreasing Fe(II) concentrations), the remaining samples (D04_day41 and D04_day47) show no evidence for the presence of planktonic Geobacter activity (Fig.3A). During the period where Geobacter abundances decrease (as inferred by conserved peptide abundances), U(VI) concentrations rebound to 3mgl−1. This is likely due to both the temporary cessation of acetate addition and the beginning of dominance by sulfate reduction towards the end of the experiment (Fig.3B).
Although peptide biomarker abundances track Fe(II) production during the initial stages of the 2008 biostimulation experiment, later samples show a disconnect between elevated Fe(II) concentrations that persist in the environment, and low (or absent) biomarker abundances (e.g. D04_day27). As biostimulation progresses towards the transition period, these data suggest that rapid Geobacter growth that characterizes some earlier time points during the most efficient U(VI) removal has ceased. Less active cell growth results in fewer planktonic cells for biomass sampling, and we can infer that any low‐level Fe(III) reduction that occurs during the remainder of biostimulation is likely carried out by mineral‐attached cells (S. Dar, unpublished). There are multiple possible factors responsible for the slowdown of Geobacter cell growth such as the decreasing concentrations of ‘bioavailable’ Fe(III) oxides or other nutrient limitations. The clustering of 2008 samples via non‐metric multidimensional scaling (NMDS) using both conserved ‘biomarker’ and unique CS peptide abundances and the projection of environmental variables onto these axes suggest that acetate concentrations best explain the shifts in Geobacter populations (Fig.4A). This further confirms the importance of ensuring that the key microbial communities during biostimulation receive excess carbon concentrations wherever possible. This plot additionally shows a negative correlation between U(VI) and the samples containing abundant Geobacter, and the visual correlation between low U(VI) concentrations and high biomarker abundance is clear in Fig.3A. Thus, these biomarker abundances are useful as indicators of efficient U(VI) removal from groundwater. During the 2008 biostimulation experiment the highest peptide abundances (D04_day7 through D04_day20) are correlated with the lowest U(VI) concentrations (≤1mgl−1) (Fig.3A), and where peptide abundances begin to decrease (sample D04_day23 onwards), U(VI) concentrations start to rebound within the sampling well. Within the 2007 data, this pattern is repeated with the D07_day21 sample, while the biomass for the D07_day9 sample was sampled just prior to a rapid decrease in U(VI) concentrations (Fig.2A).
While conserved peptides can be used to assess Geobacter abundance within a community, fine‐scale strain‐level shifts between samples can be assessed using unique peptide abundances. Global analysis of unique peptide abundances in 2007 datasets indicated strain shifts and increasing community diversity within the Geobacter population over the duration of the biostimulation period. While this was characterized by an increase in peptides matching G. lovleyi and a decrease in peptides matching G. bemidjiensis and strain M21 as biostimulation progressed, total peptide data revealed that the community was still dominated by strains most closely related to G. bemidjiensis and strain M21, but with increased similarity to G. lovleyi at certain loci (Wilkins etal., 2009). While the abundance of unique peptides matching CS in these global datasets suggested that this protein would be a good candidate for this study, the addition of cloned gltA sequences to the search database for this study increased the strain‐level resolution of the analysis (Fig.5A and B).
Changes in the presence and abundance of unique CS peptides over the course of the 2007 experiment are clearly visible between the two D07 samples (Fig.2B). Phylogenetic analysis of CS sequences matching these unique peptides reveals that the majority fall within clusters ‘1, 2 and 3’ in Fig.4A, and are most closely related to the G. bemidjiensis and strain M21 copies of CS. One exception is a cloned sequence (cluster ‘4’; C0814‐40) that is more closely related to G. uraniireducens and G. lovleyi. Those sequences that match peptides only detected in sample D07_day9 are generally associated with the G. bemidjiensis/strain M21 region of the tree (red labels, grey shaded region, Fig.5A), whereas increases in the diversity of sequences matching CS peptides only detected in D07_day21 are apparent (blue labels, Fig.5A). In this later sample, unique peptides matching five isolate copies of CS and five cloned copies of CS are detected, more than double the number of sequences matching unique peptides in D07_day9. The sample specific presence‐or‐absence nature of these peptides is illustrated in Fig.4B, where peptides unique to one sample constrain the clustering distances between samples.
Shifts in the abundance of these peptides are also apparent between the 2007 D07 samples, as illustrated in Fig.5B. Abundance increases in unique peptides matching G. lovleyi, G. uraniireducens and clone C0814‐40 (‘cluster 4’) are coupled to decreases in G. bemidjiensis/metagenomic sequence peptides (Fig.5B) between D07_day9 and D07_day21 (both shifts P<0.05). Less significant decreases are also observed within clusters ‘1, 2 and 3’ between the same time points. These patterns amongst CS unique peptides are indicative of the Geobacter community shifting from a population dominated by a few strains closely related to G. bemidjiensis and strain M21 during the early period of Fe(III) reduction, to increasing strain diversity later in Fe(III) reduction. Mapping these unique peptides on CS protein alignments, and identifying regions of multiple coexisting peptides previously demonstrated this effect (Wilkins etal., 2009). However, given that the majority of cloned sequences are located within clusters ‘1, 2 and 3’ (Fig.5A), we can infer that despite fine‐scale strain‐level shifts occurring within the Geobacter community during this experiment, the dominant strains still have greatest similarity to G. bemidjiensis and strain M21.
During the 2008 biostimulation experiment, approximately five samples (D04_day5 through D04_day20) were recovered during the period of dominant Fe(III) reduction. Phylogenetic analysis of sequences matching unique peptides detected in these samples reveals differences relative to both D07 samples from 2007. Whereas sequences from ‘cluster 1’ are absent in 2008 samples, there are clear increases in the number of sequences from ‘cluster 2’ relative to the 2007 data (Fig.5C). Unique peptide abundance patterns confirm these differences (Fig.5D); abundances of unique peptides matching G. bemidjiensis/metagenomic sequence suggest the presence of strains that were dominant at the beginning of the 2007 experiment (sample D07_day9) when acetate was added to the pristine aquifer. However, the abundance of G. lovleyi unique peptides indicates some similarity to strains that were present towards the end of biostimulation in 2007 (sample D07_day21). In addition, relatively similar abundance patterns between these five samples (Fig.5D) may suggest the presence of a more stable community of Fe(III)‐reducing bacteria than was present during the 2007 biostimulation experiment. Indeed, the NMDS plot in Fig.4A reveals the tight clustering of samples D04_day5 through D04_day20, indicating significant similarity between these samples.
These apparent differences between Geobacter populations recovered from the subsurface over two years may potentially be linked to the use of the same flow cell for both experiments. A range of analyses have demonstrated that following acetate addition to a pristine aquifer system, a bloom of Geobacter growth is stimulated (Vrionis etal., 2005; Mouser etal., 2009; Wilkins etal., 2009). This initial enrichment of Geobacter may be dominated by only a few strains that couple the highest growth rates to the most efficient utilization of acetate and Fe(III) oxides. As the duration of biostimulation progresses however, strain diversity within the Fe(III)‐reducing microbial community has been seen to increase (Wilkins etal., 2009), and this is reflected in the increased diversity of unique CS peptides observed in sample D07_day21 (Fig.5A and B). Whether this is due to the emergence of slower growing bacteria, the initial effects of sulfate reducers, or changes in the availability of different Fe(III) oxides as terminal electron acceptors, is currently unknown. The 2007 biostimulation experiment was ceased at this point, towards the end of Fe(III) reduction but prior to the transition into sulfate reduction. Potentially, carbon concentrations arising from residual acetate and breakdown products from the initial biomass ‘bloom’ may have allowed a relatively stable and diverse community of Fe(III)‐reducing microorganisms to persist at low levels in the subsurface between experiments. Initial Fe(II) concentrations in well D04 prior to acetate amendment in 2007 (1.52mgl−1) are significantly lower than those seen the following year in 2008 (3.13mgl−1), suggesting Fe(III)‐reducing activity during the period in‐between biostimulation experiments. Following the subsequent addition of acetate at the start of the 2008 campaign, some members of this pre‐existing community may be enriched together with the fast‐growing G. bemidjiensis and M21‐like strains, resulting in a community structure that is different from that seen at the start of acetate enrichment in a pristine aquifer. This ‘legacy effect’ may have implications for biostimulation projects aiming to repeatedly amend an area with carbon and achieve reproducible outcomes over multiple experiments (S.J. Callister, M.J. Wilkins, C.D. Nicora, K.H. Williams, J.F. Banfield, N.C. VerBerkmoes et al., submitted).
This demonstrated ability to detect strain‐level differences between Geobacter‐dominated communities using just CS unique peptides will enable the analysis of biostimulated Geobacter communities during future experiments. Comparisons between new data and patterns observed during the 2007 and 2008 experiments may allow us to infer biogeochemical processes occurring in the subsurface, providing a useful tool to complement geochemical measurements. This rapid screening will be carried out in conjunction with the analysis of conserved CS peptides, which we have demonstrated may be used as indicators of planktonic Geobacter abundance and activity during biostimulation. Given the strong correlation between high ‘biomarker’ abundance and low U(VI) concentrations, these abundances will also be used in conjunction with geochemical measurements for improved prediction of U(VI) removal rates in the subsurface. The availability of a dedicated Geobacter CS search database will further ensure that these analyses can be carried out in an efficient manner.
Samples for groundwater geochemistry and planktonic biomass for proteomics was collected over two consecutive in situ biostimulation projects at the Rifle IFRC site located in Rifle, CO. The in situ biostimulation experiments were carried out in the same flow‐cell during August and September 2007, and July, August and October 2008. During the 2008 biostimulation experiment, acetate addition was ceased for a nine‐day period to allow a ‘groundwater flush’ to take place within the aquifer. For other details on flow‐cell size, operating conditions and biomass collection, see Wilkins and colleagues (2009). Analysis focused on the three samples collected from wells D05 and D07 during the 2007 experiment, and nine planktonic biomass samples collected from well D04 during the 2008 experiment. DNA for metagenomic analysis was extracted from biomass collected from well D05 during a period of dominant Fe(III) reduction during the 2007 biostimulation experiment (B. Methé, pers. comm.). Groundwater samples for DNA analysis of gltA clones were collected during the 2008 biostimulation experiment in well D04. Per sample, 10l of groundwater was concentrated on a Supor‐200 membrane filter (ø=293mm, pore size=0.2µm; Pall Life Sciences, Ann Arbor, MI, USA). Filters were quickly sealed into Mylar bags, flash frozen in an ethanol‐dry ice bath, and stored at −80°C until nucleic acids extraction. These DNA samples for gltA gene analysis were collected on the following dates: 07/19/08, 07/29/08, 08/02/08, 08/14/08 and 08/22/08. Geochemical measurements were carried out as previously described (Wilkins etal., 2009).
DNA was extracted from a portion of the filter, crushed with liquid nitrogen, using the FastDNA SPIN Kit for soil (MP Biomedicals, Solon, OH, USA). DNA was quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). An approximately 803bp DNA fragment was amplified using the primers CS18F (5′‐CTCGCGACATCCGGAGTCT‐3′) and CS821R (5′‐TGTCCGGCGTTCAGGGTAT‐3′) (Holmes etal., 2005) targeting the Geobacteraceae citrate synthase‐encoding gene (gltA), and the following PCR protocol: initial denaturation at 95°C for 5min, 30 cycles of 1min denaturation at 95°C, 1.5min annealing at 55°C–50.5°C (0.5°C decrease per cycle during the first 10 cycles), 1.5min elongation at 72°C, and final elongation at 72°C for 10min. Positive controls, i.e. purified gltA PCR product from Geobacter metallireducens, and negative controls without DNA were always included in PCR amplification experiments. The reaction was carried in a PTC200 Peltier Thermal Cycler (MJ Research, Waltham, MA, USA). The 50µl reaction mixture contained 100ng of DNA, 1× Q‐Solution (Qiagen), 1× PCR Buffer (Qiagen), 1.5mM MgCl2 (Qiagen), 200µM concentrations of each deoxynucleotide (Sigma, Saint Louis, MO, USA), 0.5µM concentrations of each primer, 0.5× BSA (New England Biolabs, Beverly, MA, USA), and 1.25U of Taq DNA polymerase (Qiagen). The presence and size of the amplification products were determined by agarose [1% (w/v)] gel electrophoresis. Bands of the expected size were purified from the gel by excision with a sterile surgical blade and purified with the QIAquick Gel Extraction Kit as recommended by the manufacturer (Qiagen). Four microlitres of the agarose gel‐purified DNA mixture was immediately ligated into the pCR2.1‐TOPO vector (TOPO TA Cloning Kit, Invitrogen, Carlsbad, CA, USA). The gltA sequence was determined for E. coli recombinant vector‐containing colonies with the primers M13F and M13R, in an ABI 3730xl DNA Analyzer using the Sanger chain‐terminator method with fluorescently labelled nucleotides. Chromatograms were visually inspected using the software 4Peaks v1.7 (http://www.mekentosj.com). Recovered gltA sequences were initially compared with GenBank database (Benson etal., 2005) for preliminary identification using the programs BLASTN and BLASTX (http://www.ncbi.nlm.nih.gov/BLAST).
Global as well as soluble and insoluble protein fractions were extracted from cell pellets by using established protocols (Lipton etal., 2002; Adkins etal., 2006). Briefly, frozen cells were thawed on ice, washed using 100mM NH4HCO3, pH 8.4, buffer, and then suspended in a new aliquot of this buffer. Cells were lysed via pressure cycling technology using a barocycler (Pressure BioSciences, South Easton, MA, USA). The suspended cells were subjected to 20s of high pressure at 35 kilopounds per square inch, followed by 10s of ambient pressure for 10 cycles. The protein concentration was determined by a Coomassie assay (Thermo Scientific, Rockford, IL, USA). Protein extraction, digestion and high‐performance LC fractionation were performed as previously described (Callister etal., 2006a). From each collected fraction, peptides were analysed by reversed phase high‐performance LC separation coupled with the use of an LTQ ion trap mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA) operated in a data‐dependent MS‐MS mode. Spectra generated by LC‐MS/MS were analysed using the SEQUEST algorithm in conjunction with predicted protein annotations concatenated from CS sequences described below. CS sequences were obtained from seven isolate Geobacter genomes, metagenomic sequence and translated gltA clones amplified from DNA from well D04 during the 2008 biostimulation experiment. SEQUEST results were preliminarily filtered (Xcorr values of ≥1.9, ≥2.2 or ≥3.5 for 1+, 2+ or ≥3+ if seen once; Xcorr values of ≥1.9 if seen two or more times; no cleavage rules; minimum length of 6 residues), extracted and processed using the PRISM proteomics pipeline developed in‐house (Kiebel etal., 2006).
A Mass‐Tag database was constructed from SEQUEST search results that allowed the rapid screening of proteomic data for CS peptide matches (Smith etal., 2002). Label‐free arbitrary abundance measurements of peptides from the three 2007 and nine 2008 sampling events were obtained from LC‐MS measurements. Triplicate (36 total analyses) measurements were made on a custom‐built HPLC system coupled via ESI to an LTQ‐Orbitrap mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA). HPLC conditions were the same as reported above. Mass measurement accuracy and elution time accuracy cut‐offs of 5ppm and 1%, respectively, were applied to mass and elution time features prior to matching to the reference peptide database.
Peptide abundances were calculated by integrating the signal strength under each peak of the LC‐MS spectra (Smith etal., 2004). Peptide abundances were subsequently log transformed and normalized to a common baseline using central tendency normalization (Callister etal., 2006b). Manipulation of these data, including the generation of heat maps, was carried out using DANTE (Polpitiya etal., 2008), a software tool publically available at http://ncrr.pnl.gov/. Significance of changes in abundances between groups of peptides was assessed using paired t‐tests. Where necessary, significance cut‐offs (0.05) are displayed in the text. Citrate synthase amino acid sequences were aligned using the ClustalW algorithm at the EBI website (http://www.ebi.ac.uk/Tools/clustalw2/index.html). Alignment results were imported into the MEGA software, which was used to construct neighbour‐joining phylogenetic trees (Tamura etal., 2007). Non‐metric multidimensional scaling, principal components analysis and environmental vector plotting were carried out using the vegan package (Oksanen etal., 2009) in the statistical software R.
Barbara Methé at the J. Craig Venter Institute is thanked for providing the metagenomic sequence used in this study. We also thank the city of Rifle, CO, the Colorado Department of Public Health and Environment, and the US Environmental Protection Agency, Region 8, for their cooperation in this study. Pacific Northwest National Laboratory is managed under contract DE‐AC05‐76RL01830 with Battelle Memorial Institute. Portions of this work were performed at the Environmental Molecular Sciences Laboratory, a DOE national scientific user facility located at the Pacific Northwest National Laboratory. This research was sponsored by the Environmental and Remediation Sciences Program, Biological and Environmental Research, Office of Science, US Department of Energy.