Comparative Genomics of “Dehalococcoides ethenogenes” 195 and an Enrichment Culture Containing Unsequenced “Dehalococcoides” Strains

Dehalococcoides ethenogenes 195 and Dehalobacter restrictus PER-K23 (DSMZ strain number 9455) were grown in 100-ml-batch pure cultures with a defined mineral salts medium containing 5 mM acetate as the carbon source, H₂ as the electron donor (added as H₂:CO₂ gas [80:20, vol/vol]), and 78 μmol TCE as the electron acceptor, as previously described (14, 24).

ANAS was enriched from contaminated soil obtained from Alameda Naval Air Station and has been functionally stable for over seven years, with the ability to reductively dechlorinate TCE completely to ethene. Previous clone library studies have shown that Dehalococcoides bacteria comprise approximately 30% of the ANAS culture (31). Lactate (25 mM final concentration) is supplied as both the carbon source and electron donor, while TCE (0.1 mM final concentration) is supplied as the terminal electron acceptor. The semibatch growth and maintenance procedures have been previously described (24, 31).

A 50-ml sample of the ANAS culture was collected after one 100 μM dose of TCE was degraded to VC and ethene. Cells were collected by centrifugation (12,000 × g for 3 min at 4°C), the supernatants were discarded, and the cell pellets were stored at −80°C until processing. ANAS genomic DNA (gDNA) was isolated from frozen cell pellets by using an UltraClean Mega Prep soil DNA kit (Mo Bio Laboratories, Carlsbad, CA) according to the manufacturer's instructions. DNA was stored at −80°C prior to further use.

D. restrictus and strain 195 cells were harvested from 10 100-ml batches per culture to ensure that a sufficient quantity of DNA could be extracted for application to microarrays. Cells were collected by vacuum filtration with hydrophilic Durapore membrane filters (0.22-μm pore size, 47-mm diameter; Millipores, Billerica, MA), and the filters were stored in 2-ml microcentrifuge tubes at −80°C until further processing. gDNA was isolated from the filters by using a phenol extraction method as follows. The 2-ml microcentrifuge tubes containing frozen filters were amended with 250 μl lysis buffer (200 mM Tris [pH 8.0], 50 mM EDTA, and 200 mM NaCl), 100 μl 10% sodium dodecyl sulfate, 1 g 100-μm-diameter zirconia-silica beads (Biospec Products, Bartlesville, OK), and 1.0 ml buffer-equilibrated phenol (pH 8.0) (Sigma-Aldrich, St. Louis, MO). The cells were lysed by heating the tubes to 65°C for 2 min, bead beating with a mini bead beater (Biospec Products) for 2 min, incubating for 8 min at 65°C, and bead beating again for an additional 2 min. Cellular debris was collected by centrifugation (12,000 × g for 5 min at 4°C), and the aqueous lysate was transferred to a new, DNase-free microcentrifuge tube. The aqueous lysate was extracted twice with 1 volume of phenol (pH 8.0)-chloroform-isoamylalcohol (125:24:1, vol/vol) and once with 1 volume of chloroform-isoamylalcohol (24:1, vol/vol) (Sigma-Aldrich). DNA was precipitated by adding 0.1 volume of 3 M ammonium acetate (pH 5.2) and 1 volume of 100% isopropanol. The precipitate was collected by centrifugation, washed once with 80% ethanol, and resuspended in 40 μl of nuclease-free water. Contaminating RNA was removed by RNase digestion with DNase-free RNase according to the manufacturer's instructions (Roche Applied Science, Indianapolis, IN). The purified DNA was stored at −80°C prior to further use.

The mass of DNA per volume was quantified by using a fluorometer (model TD-700; Turner Designs, Sunnyvale, CA) and a Quant-iT PicoGreen dsDNA assay kit (Invitrogen Molecular Probes, Carlsbad, CA) according to the manufacturer's instructions.

The microarrays targeting genes within the genome of Dehalococcoides ethenogenes 195 were produced by Affymetrix (Santa Clara, CA) as prokaryotic midi format (format 100) photolithographic microarray chips. Each target sequence is represented on the array as a collection of 11 probe pairs (22 total probes) consisting of 25-mer oligonucleotide probes which are distributed along the length of the respective gene and are collectively referred to as a probe set. Each chip contains 1,624 probe sets. To control for nonspecific hybridization, each probe pair consists of a perfect-match probe and a corresponding single-mismatch probe in which the 13th nucleotide consists of a base that would mismatch the target. While the perfect-match probe provides measurable fluorescence when sample binds to it, the paired mismatch probe is used to detect and eliminate any false or contaminating fluorescence within that measurement resulting from nonspecific hybridization (2). Probe sets were designed to target 1,560 nonredundant genes from the genome of strain 195, including 48 of 51 structural-RNA-coding sequences and 1,512 of 1,514 of the predicted protein-coding sequences identified by Seshadri et al. (33). The number of probe sets on the array is less than the total number of genes in the strain 195 genome because the genome contains a 33-gene duplication and a 22-gene triplication. In addition, 45 probe sets, 24 as positive controls and 21 as negative controls, were designed on the array to facilitate calibration and to resolve background signals. The positive-control probe sets targeted Bacillus subtilis genes lys, phe, thr, and dap. To generate control DNA, plasmids containing lys, phe, thr, and dap were purified from strains ATCC 87482, ATCC 87483, ATCC 87484, and ATCC 87486, respectively. The negative-control probe sets were designed to hybridize with specific Escherichia coli and bacteriophage genes that were not expected to be present in the experimental cultures.

gDNA was prepared for application to the microarrays according to the protocols outlined in section 3 of the Affymetrix GeneChip Expression Analysis Technical Manual (Affymetrix, Santa Clara, CA), with minimal modifications. Briefly, purified gDNA (1 μg per array) was mixed with the positive-control plasmids (7.6 nM lys, 15.2 nM phe, 30.4 nM thr, and 114.0 nM dap) described above and fragmented to an average size of 50 to 200 bp by enzymatic digestion with amplification grade DNase I (Invitrogen Life Technologies, Carlsbad, CA). Each 50-μl fragmentation reaction consisted of gDNA and 0.02 units DNase I per μg DNA diluted in 1× One-Phor-All buffer (Amersham Biosciences). The mixture was incubated for 20 min at 25°C, followed by 10 min at 98°C to inactivate the DNase I. The fragmentation products were visualized on a 4% MetaPhor agarose (Cambrex Bio Science Rockland, Inc., Rockland, ME) electrophoresis gel run at 70 V for 2 h. A 25-μl amount of the fragmentation product was then used for further processing. The fragmentation products were biotin end labeled by adding 25 μl of fragmentation product to 10 μl of 5× reaction buffer (Affymetrix, Santa Clara, CA), 2 μl 7.5 mM GeneChip DNA-labeling reagent (Affymetrix), 2 μl terminal deoxynucleotidyl transferase (Promega), and 11 μl nuclease-free water, incubating at 37°C for 60 min, and adding 2 μl 0.5 M EDTA (pH 8.0) (Invitrogen Life Technologies, Carlsbad, CA) to stop the reaction. Microarray chips were prehybridized with 1× hybridization buffer (100 mM morpholineethanesulfonic acid [MES; Sigma-Aldrich], 1 M Na⁺ [Sigma-Aldrich], 20 mM EDTA [Sigma-Aldrich], 0.01% Surfact-Amps Tween 20 [Pierce Chemical]) in a chip-rotisserie oven for at least 20 min at 60 rpm and 45°C. The samples were denatured by being heated to 99°C for 5 min. Hybridization buffer was removed from each microarray chip and replaced with an equal volume of sample hybridization mixture (1× MES hybridization buffer, 7.8% dimethyl sulfoxide [Sigma-Aldrich], 0.1 mg/ml herring sperm DNA [Promega Corporation], 0.5 mg/ml bovine serum albumin [Invitrogen Life Technologies, Carlsbad, CA], 50 pM B2 control oligonucleotide [Affymetrix], and 1 μg fragmented, labeled DNA). The hybridization was carried out by incubation in a chip-rotisserie oven (Affymetrix) for 16 h at 45°C and 60 rpm. Following hybridization, the hybridization mixture was removed, and the microarrays were processed and scanned according to standard Affymetrix protocols.

gDNA from each culture was divided into replicate samples which were independently fragmented, labeled, and hybridized to the arrays. Two microarrays were processed for the positive control (strain 195), two for the negative control (D. restrictus), and five for ANAS (two analyses from one biological sample followed 1 year later by three analyses of a second biological sample).

The hybridization signal intensity for perfect match and mismatches in each probe set from each scan was computed by using Affymetrix GeneChip software and the MAS5 algorithm (3). The data set of each microarray was normalized by scaling the signal intensities of the added positive controls to a target signal intensity of 2,500, thus allowing for comparisons between microarray chips. The GeneChip MAS5 software was used to implement the Wilcoxon signed-rank-based algorithm to evaluate the “presence” or “absence” of individual target sequences and return a statistical P value (3, 25). A gene was considered “present” in a sample if each replicate probe set had a signal intensity greater than the highest signal intensity measured for the negative controls and a P value of less than 0.05. Due to the stringency of the P value criterion, only 8 genes total were designated absent due solely to the signal intensity criterion.

PCR primer sets were designed to target several selected protein-coding genes of strain 195. These primer sets were then used to test for the presence of these genes in the ANAS enrichment. PCR amplification was performed in reaction mixtures containing 20 to 100 ng of gDNA and (as final concentrations) 1× PCR buffer II (Perkin Elmer), 2.5 mM MgCl₂, a 200 μM concentration of each deoxynucleoside triphosphate, a 300 nM concentration of each forward and reverse primer, and 0.025 U of AmpliTaq Gold (Perkin Elmer) per μl of reaction mix. The reaction mixtures were incubated in an Eppendorf Mastercycler gradient thermocycler (Eppendorf, Westbury, NY) at 94°C for 12 min (for initial denaturation and activation of AmpliTaq Gold), followed by 30 cycles at 94°C for 1 min, 50°C for 45 s, and 72°C for 2 min and then by a final extension period of 12 min at 72°C. Amplified DNA was processed by using an Agilent 2100 bioanalyzer. Visual detection of a band corresponding to a DNA fragment of expected size indicated the presence of a gene. The absence of a band of the expected fragment size indicated the absence of the gene. The PCR results were compared with the microarray probe hybridization results to support determinations of presence/absence.

The microarray data analyzed in this study have been deposited in the NCBI Gene Expression Omnibus database under accession number GSE9612 (http://www.ncbi.nlm.nih.gov/projects/geo/index.cgi).

To evaluate the hybridization efficiency of the microarray probe sets and to detect probe sets that might yield false negatives, gDNA isolated from strain 195 was applied to the microarrays. All (1,579 of 1,579) of the probe sets on the array were positively detected. Conversely, the specificity of the microarray was tested by applying gDNA from D. restrictus, a dehalogenating bacterium which shares little sequence homology with strain 195. This hybridization resulted in 1,575 probe sets (99.7%) for which the hybridization results did not meet our criteria for classification as present and were thus deemed absent. The four probe sets that were deemed present with gDNA from D. restrictus were all tRNA-encoding genes. Cross-species hybridization with tRNAs is not surprising since these sequences tend to be highly conserved among bacteria. No significant hybridization was found with the D. restrictus 16S rRNA gene, which shares approximately 75% identity over 1,511 bases with that from strain 195.

The dynamic range of the microarray was examined in a titration experiment in which different masses of gDNA (0.5 μg, 1 μg, 2.5 μg, and 5 μg) from ANAS were applied to the arrays. The results showed that average signal intensities increased in a nonlinear manner corresponding to increasing masses of gDNA applied to the arrays, exhibiting a dynamic response over the mass range of applied DNA (Fig. 1A and 1B). The coefficients of variation (CVs) across the disparate probe sets within each microarray data set are 58%, 53%, 47%, and 48% for 0.5 μg, 1 μg, 2.5 μg, and 5 μg of gDNA applied to the arrays, respectively. The similarity of the CVs between data sets suggests that the mass of DNA had little effect on the variation of hybridization efficiencies. Each array gave a range of signal intensities above the background noise, indicating that even the lowest mass of DNA tested was sufficient to allow the differentiation of signal intensities between probe sets and to produce detectable signals. The data demonstrate a predictable trend of higher masses of applied DNA yielding greater signal intensity with slightly less variation between probe sets. A mass of 1 μg of gDNA was chosen for subsequent application to microarrays in this study because this mass falls within the observed linear range of correspondence between signal intensity and gDNA mass (Fig. 1B).

The microarray data were highly consistent for replicate arrays containing DNA from a single culture. The hybridization signal intensity profiles for replicate arrays of DNA extracted from strain 195 and ANAS demonstrate a high level of analytical reproducibility (Fig. 2). The CV was calculated for each probe set to which ANAS gDNA hybridized across the five replicate ANAS microarrays. The average CV for a probe set across replicates is 10.1%, demonstrating that the data are highly reproducible among both biological and analytical replicates.

Of the 1,579 probe sets on the strain 195 microarray, ANAS gDNA hybridized to 1,369 (86.7%), representing genes that are conserved between strain 195 and the ANAS enrichment (Fig. 3).

Dehalococcoides strains within the ANAS enrichment show extensive gene similarity with the putative housekeeping genes of strain 195. Key conserved genes included gyrA (DET1630), gyrB (DET0004), ftsZ-1 (DET0343), ftsZ-2 (DET0636), dnaG (DET0552), frr (DET0374), infC (DET0752), nusA (DET0985), pgk (DET0744), pyrG (DET1410), rpoD (DET0551), and smpB (DET1506). To verify the microarray results, Dehalococcoides-specific PCR primers were designed for three of these housekeeping genes, the ANAS gDNA was amplified, and the corresponding genes were confirmed to be present in ANAS by visual detection of the amplified fragment (Table 1).

Metabolically important genes, such as hydrogenase genes and genes involved in nitrogen fixation, carbon metabolism, cobalamin processing, and amino acid biosynthesis, have been putatively identified in strain 195 (33). ANAS DNA hybridized to probes for nearly all of these metabolic genes. All of strain 195's five putative hydrogenase complexes (26 genes), nine putative nitrogen fixation genes, 13 genes associated with the Wood-Ljungdahl carbon fixation pathway, and 20 genes implicated in acetate assimilation, gluconeogenesis, and the citric acid cycle were found in the ANAS culture. Genes thought to be involved in cobalamin transport and processing (including DET0245 and -0246, DET0650 to -0652, DET0654, DET0658, DET0684 to -0686, DET0688, DET0692, DET1138, and DT1139) were also detected in the ANAS culture. PCR was performed to corroborate the presence or absence of 22 Dehalococcoides-specific genes in the ANAS culture (Table 1).

Two hundred ten probe sets (13.3%) did not meet the criteria to be regarded as present in response to ANAS gDNA (see Table S1 in the supplemental material). The genes from strain 195 that were not identified in the ANAS culture can be organized by their location in the strain 195 genome and by their predicted function. Most of the genes not detected in the ANAS culture are localized in clusters in the strain 195 genome (Fig. 3). Interestingly, 123 of the 210 probe sets that did not hybridize with ANAS gDNA represent genes located within potential IEs identified by Seshadri et al. (33) in strain 195 that include genes predicted to encode mobile genetic elements (MGEs), such as phage proteins, endonucleases, transposases, recombinases, resolvases, and restriction modification system proteins. In fact, 88% of the genes located in IEs in strain 195 were not detected in ANAS, including five of the nine IEs that were missing entirely. Further, no intact IEs from strain 195 were detected in ANAS (Table 2). Other genes not detected in the ANAS culture, such as DET0315 to -0323 and DET1508 to -1530, are localized in clusters and flanked by recombination-related genes but are not located inside identified IEs.

Genes identified in strain 195 that were not detected in ANAS can be sorted by function (Table 3). One hundred twenty-nine (61.4%) of the 210 probe sets to which ANAS gDNA did not hybridize cluster into a group of genes described as unclassified or poorly classified by an analysis of clusters of orthologous groups (COGs) by DOE JGI. The remaining genes targeted by probe sets to which ANAS gDNA did not hybridize represent 13 COGs as described in the Integrated Microbial Genomes (IMG) database (DOE JGI; http://img.jgi.doe.gov/).

Strain 195 contains 19 genes orthologous to the catalytic A subunit genes of RDs, which may be key functional genes for Dehalococcoides bacteria because halogenated organics serve as the only identified terminal electron acceptors for this bacterium. ANAS gDNA did not hybridize to 13 of the 19 probe sets targeting putative RD genes in strain 195 (Table 4). Ten of these 13 RD genes have been included in the energy production and conversion COG by JGI, and the remaining three nonhybridized RD genes, including one ortholog containing a point mutation (DET1062), are unclassified. ANAS gDNA did not hybridize to probes targeting the pceA gene (DET0318), which codes for a subunit of the enzyme responsible for reducing PCE to TCE in strain 195 (29). The PCR results suggested the absence of two putative RD genes, including pceA in ANAS (Table 1), and correspondingly, ANAS has consistently failed to dechlorinate PCE (31). The gene in strain 195 that codes for subunit A of the protein responsible for TCE reduction to ethene, designated tceA (DET0079) (27), was detected in ANAS gDNA by the microarray. Congruently, ANAS is able to completely dechlorinate TCE, and the presence of tceA in ANAS was confirmed by PCR (Table 1).

The largest functional category of strain 195 genes not detected in the ANAS culture (22 of 210 probe sets) is labeled “transcription” and includes predicted transcriptional regulators. Of these genes, 13 are located in IEs in the strain 195 genome. The nine putative transcriptional genes not located in IEs are all located near RD genes in the strain 195 genome, including DET0300, DET0304, DET0316, DET1515, DET1520, DET1523 to -1525, and DET1536. Many of these genes (DET0300, DET0304, DET0316, DET1520, and DET1523 to -1525) encode transcriptional regulators predicted to regulate the expression of RD genes (33) that also were not detected in the ANAS culture. DET1536 is the only putative transcriptional regulator gene not detected in ANAS that is associated with a putative RD gene (DET1535) that is present in ANAS.

Replication, recombination, and repair genes comprise the next largest COG of strain 195 genes not detected in ANAS and include putative transposase and recombinase genes. It is not surprising that 17 of the 19 probe sets in this group target genes located in IEs. The two genes not located in IEs are DET0323, predicted to encode a phage-related recombinase, and DET1301, encoding a putative transposase.

Of the 210 probe sets undetected in ANAS gDNA, 13 genes were categorized as participating in signal transduction, with 8 of these also being classified as participating in transcription. Of these 13 predicted signal transduction genes, 11 are identified as members of two-component systems regulating RD genes that also were not detected in ANAS. The two signal transduction genes not detected in ANAS that are not predicted to regulate RD genes are located in IEs.

Amino acid transport and metabolism genes comprise the next largest COG of strain 195 genes not detected in ANAS. Most genes identified in amino acid biosynthetic pathways of strain 195 were found in ANAS, with the exceptions of one gene located in an IE (DET1118) and two operons that neighbor each other, the putative tryptophan operon and an operon encoding a predicted ABC transporter. The strain 195 tryptophan biosynthesis operon is thought to be composed of eight genes, six of which were identified as absent in the ANAS culture. The genes in ANAS not detected by the microarray (DET1481 to -1485 and DET1487 and -1488) are predicted to encode the pathway responsible for transforming chlorismate to tryptophan. Although the trpE gene (DET1481) did not meet the criteria for detection in ANAS by microarray analysis (P value of 0.0562, just above the 0.05 threshold, in one of the five replicates), it was detected by PCR and is consequently considered present in ANAS. Because the PCR primers target a different region of DET1481 than the microarray probes, it is probable that the nucleotide sequence within ANAS targeted by the DET1481 PCR probes is identical to that in strain 195; however, there is likely a polymorphism in the region of DET1481 that is targeted by the microarray probes. The PCR analysis also suggested the absence of several other tryptophan genes in ANAS (Table 1). One hypothetical protein and five genes predicted to encode subunits of an ABC transporter comprise an operon in close proximity to the tryptophan operon that is implicated in peptide or nickel transport. Of the six genes in this operon, all five putative ABC transporter subunits (DET1490 to -1494) were not detected in ANAS, and the absence of DET1494 was supported by PCR analysis (Table 1).