Genome Sequence, Proteome Profile, and Identification of a Multiprotein Reductive Dehalogenase Complex in Dehalogenimonas alkenigignens Strain BRE15M.

Bacteria of the genus Dehalogenimonas respire with vicinally halogenated alkanes via dihaloelimination. We aimed to describe involved proteins and their supermolecular organization. Metagenomic sequencing of a Dehalogenimonas-containing culture resulted in a 1.65 Mbp draft genome of Dehalogenimonas alkenigignens strain BRE15M. It contained 31 full-length reductive dehalogenase homologous genes (rdhA), but only eight had cognate rdhB gene coding for membrane-anchoring proteins. Shotgun proteomics of cells grown with 1,2-dichloropropane as an electron acceptor identified 1152 proteins representing more than 60% of the total proteome. Ten RdhA proteins were detected, including a DcpA ortholog, which was the strongest expressed RdhA. Blue native gel electrophoresis (BNE) demonstrating maximum activity was localized in a protein complex of 146-242 kDa. Protein mass spectrometry revealed the presence of DcpA, its membrane-anchoring protein DcpB, two hydrogen uptake hydrogenase subunits (HupL and HupS), an iron-sulfur protein (HupX), and subunits of a redox protein with a molybdopterin-binding motif (OmeA and OmeB) in the complex. BNE after protein solubilization with different detergent concentrations revealed no evidence for an interaction between the putative respiratory electron input module (HupLS) and the OmeA/OmeB/HupX module. All detected RdhAs comigrated with the organohalide respiration complex. Based on genomic and proteomic analysis, we propose quinone-independent respiration in Dehalogenimonas.

Introduction

Organohalides are common soil and groundwater pollutants. Although they can be naturally produced, the high concentrations found at many industrial sites are due to improper storage or disposal practices and represent a threat to human and environmental health because of the toxicity or carcinogenicity of many of them.[1] A potential solution to treat sites impacted with organohalides relies on organohalide-respiring bacteria (OHRB), which can use organohalides as electron acceptors to harness energy and the reduction often results in decreased toxicity of the contaminants.[2]

Reductive dehalogenation is catalyzed by enzymes termed reductive dehalogenases (RDases). Respiratory RDases in OHRB have conserved features in their amino acid sequence including a twin arginine translocation (TAT)-leader peptide that tags the protein for transport to and/or across the cytoplasmic membrane, and two iron–sulfur cluster binding motifs.[3] Almost all characterized RDases contain a corrinoid in the active center, although corrinoid-binding motifs have seldom been found in the amino acid sequences.[4] Typically, RDases are encoded in operons composed by at least two genes: rdhA, encoding for the catalytically active subunit RdhA, and rdhB encoding a putative membrane anchor RdhB.[5] RdhA proteins are usually located attached to the cytoplasmic membrane facing the outer surface, indicating the need for both, the TAT-leader peptide and the putative membrane-anchoring RdhB protein.

In many OHRB, quinones play an important role as electron carrier between electron supplying protein complexes and the RDase (e.g., in Sulfurospirillum or Dehalobacter).[6] In these cases, energy is conserved via the quinone-mediated transport of protons across the membrane.[6] In contrast, the genomes of the obligate OHRB Dehalococcoides spp. lack the genes for complete biosynthesis of quinones.[6−8] In quinone-independent reductive dehalogenation, energy conservation was suggested to be coupled to the transfer of electrons from a hydrogenase to a complex iron–sulfur molybdoenzyme (CISM).[9] Recent studies have described a quinone-independent respiratory chain in Dehalococcoides mccartyi strain CBDB1, catalyzed by an organohalide respiration complex (OHR complex) that directly transfers the electrons derived from H2 oxidation by hydrogen uptake hydrogenase (Hup) via a CISM to the RDase.[10] In this proposed scheme, Fe–S clusters in several of the multiprotein complex subunits act as an electron relay between hydrogenase and RDase.[9,11] According to Seidel et al.,[11] the new nomenclature OmeA/OmeB was proposed for the CISM α- and γ-subunit, standing for organohalide respiration-involved molybdoenzyme. HupX, encoded in the hup operon in Dehalococcoides, interacted strongly with OmeA and OmeB, forming a tight module. A much less strong interaction was found between the OmeA/OmeB/HupX module and a hydrogenase module HupS/HupL.

Bacteria of the genus Dehalogenimonas are obligate OHRB (class Dehalococcoidia), and were known for their potential to use vicinally halogenated alkanes.[12] Recently, isolated Dehalogenimonas strains have been shown to also respire with chlorinated ethenes and benzenes.[13,14] To date, the genomes of five Dehalogenimonas strains have been sequenced: (i) Dehalogenimonas lykanthroporepellens strain BL-DC-9T,[15] (ii) Dehalogenimonas alkenigignens strain IP3-3,[16] (iii) Dehalogenimonas strain WBC-2,[17] (iv) Dehalogenimonas formicexedens strain NSZ-14T,[18] and (v) “Candidatus Dehalogenimonas etheniformans” strain GP.[13] Only three proteins involved in organochlorine respiration have been characterized in Dehalogenimonas spp.: (i) DcpA has been shown to dechlorinate 1,2-dichloropropane (1,2-DCP) to propene,[19] (ii) TdrA was responsible for the hydrogenolysis of trans-dichloroethene to vinyl chloride,[17] and more recently, (iii) CerA was shown to catalyze the dechlorination of vinyl chloride to ethene.[13] The sequenced Dehalogenimonas genomes harbor 22–52 rdhA genes, and in contrast to all other known OHRB, most of them lack colocated genes for membrane-anchoring protein cognates (rdhB), a feature that remains unexplained.

Previous work in our laboratory focused on enriching a nonmethanogenic Dehalogenimonas-containing culture from sediment samples collected from the Besòs River estuary (Sant Adrià de Besòs, Barcelona, Spain), which appeared to exclusively transform vicinally halogenated alkanes via dihaloelimination.[20] The compounds transformed included 1,2-DCP, 1,2,3-trichloropropane, 1,2-dichloroethane, 1,1,2-trichloroethane, 1,1,2,2-tetrachloroethane, and 1,2-dibromoethane.[20−23] Here, we provide biochemical data on the molecular organization of the proteins involved in the respiratory chain of this novel Dehalogenimonas strain. For this, we obtained a high-quality draft genome sequence of strain BRE15M and applied a combination of proteome profiling, blue native polyacrylamide gel electrophoresis (BNE), dehalogenation activity assays and protein mass spectrometry. In line with recent studies with D. mccartyi strain CBDB1, our data suggest the existence of a quinone-independent respiratory dehalogenase protein complex in Dehalogenimonas, thereby extending the phylogenetic occurrence of these complexes outside of the genus Dehalococcoides.

Materials and Methods

Cultivation of the Dehalogenimonas-Containing Culture

A Dehalogenimonas-containing sediment-free culture transforming 1,2-DCP to propene was maintained for more than 5 years in our laboratory as described elsewhere.[20] Each microcosm contained 70 mL of an anaerobic bicarbonate-buffered medium, reduced with Na2S·9H2O and l-cysteine (0.2 mM each), amended with 5 mM sodium acetate as carbon source, and gassed with N2/CO2 (4:1, v/v, 0.2 bar overpressure) and H2 (added to an overpressure of 0.4 bar). All microcosms were cultivated under static conditions in the dark at 25 °C. To enrich the Dehalogenimonas population, we followed a dilution-to-extinction procedure[24] together with the addition of streptomycin (50 mg L–1) and vancomycin (5 mg L–1). Cultures were transferred into a fresh medium with an inoculum of 10% (v/v) after the dechlorination of two to three doses (100–500 μM) of 1,2-DCP.

For genome sequencing, 420 mL of this anaerobic enriched culture was used growing on 1,2-DCP as the only halogenated electron acceptor for 34 transfers (10% v/v). For proteomic studies, cultures growing on 1,2-DCP for >42 transfers (10% v/v) were used after the dechlorination of at least 2 mM 1,2-DCP.

Gas Chromatographic Analysis

1,2-DCP and propene were analyzed from headspace samples with a gas chromatograph (GC) model 6890N (Agilent Technologies; Santa Clara, USA) equipped with an HP-5 column (30 m × 0.32 mm with 0.25 μm film thickness; Agilent Technologies) and a flame ionization detector. Helium was used as the carrier gas (1.5 mL min–1). The injector and detector temperatures were both set at 250 °C. After the injection of the sample (split ratio = 2:1), the initial oven temperature (40 °C), ramped at 10 °C min–1 to 50 °C, and then ramped at 20 °C min–1 to 120 °C. Both compounds were identified and quantified using external chemical standards. Nominal concentrations of 1,2-DCP and propene are expressed as μmol per L of liquid volume or in nanomoles per bottle or vial.

Metagenomic Sequencing and Genomic Analysis

Details regarding cell harvesting, DNA extraction, genome sequencing, and assembling are provided in the Supporting Information Materials and Methods section. A final automatic annotation was performed using the MicroScope annotation pipeline.[25] Locus numbers used in this article are derived from the MicroScope annotation. The locus numbers are prefixed with “BRE15M_v1_” and an accompanying number, whereby the first two digits denote the scaffold number, and the last four digits denote the gene number of each particular scaffold. Predicted proteins were searched for signal peptides mediating the export from the cytoplasm using SignalP-5.0.[26] Transmembrane helices were detected using TMHMM server.[27] Subcellular location predictions were performed using PSORTb server.[28] The genome is available at GenBank under accession QEFQ00000000. The genome and annotations are also publicly accessible in the MicroScope platform (https://mage.genoscope.cns.fr/microscope/home/index.php, accession WGS BRE15M.1).

Protein Profiling

For cell harvesting, triplicate 140 mL samples of a 1,2-DCP dechlorinating Dehalogenimonas culture were separately centrifuged at 8000g and 10 °C to a final volume of 1 mL. A detailed protocol is provided in the Supporting Information Materials and Methods section. Cells were lyzed by freeze–thawing (−20 °C/+20 °C, 6×) and sonication for 30 s. The samples were then concentrated by ultrafiltration with Amicon Ultra 0.5 mL centrifugal filter units with a 10 kDa cutoff (Millipore) at 14,000g for 30 min at 10 °C. After centrifugation, filters were turned upside down and the protein extract was spun off for 2 min at 1000g, 10 °C. Protein concentrations in the obtained samples were quantified by the Bradford method. All triplicate samples were then processed by in-solution digestions for nano-liquid chromatography tandem mass spectrometry (nLC–MS/MS) as described below.

Blue Native Gel Electrophoresis

For each BNE gel, 140 mL samples of a 1,2-DCP dechlorinating Dehalogenimonas culture were centrifuged as described above. One milliliter of aliquoted concentrated cells was centrifuged again at 9000g for 20 min and 800 μL of the supernatant was removed. Then, 1× PBS buffer (pH 7.2) was added to a final volume of 600 μL. Cell density was quantified at this point by direct epifluorescence microscopic cell counting on slides covered with agarose as previously described.[29] Then, cells were lysed by a bead beater (FastPrep FP120, Thermo) at speed 4 m s–1, in six cycles of 40 s beating alternated with 1 min resting in ice. Membrane proteins were solubilized by adding n-dodecyl β-d-maltoside (DDM) to a concentration of 0.05, 0.1, or 1% w/v, and incubating the samples at 10 °C for 1 h with gentle shaking under anaerobic conditions. After DDM incubation, samples were centrifuged at 16,600g for 45 min at 4 °C. The supernatant (DDM-solubilized protein extract) was amended with 0.125% w/v Coomassie G-250 additive and loaded in triplicate wells (25 μL/well) onto a precasted 4–16% gradient Bis-Tris gel (NativePAGE Novex, Invitrogen) and run inside an anaerobic glove box, maintaining the BNE system in contact with ice packs. Light blue cathode buffer and anode buffer were prepared following the manufacturer’s instructions, degassed, and cooled down to 4 °C before the electrophoresis was started. NativeMark Unstained Protein Standard (Invitrogen) was used as protein ladder. Electrophoresis was run at 150 V for 60 min, and then at 200 V for 30 min more. Once the electrophoresis was finished, one of the replicate lanes and the protein ladder lane were cut from the gel using a scalpel and the protein bands were stained for visualization by a short silver staining method.[30] The rest of the gel was stored in anode buffer at 4 °C.

Dehalogenase Activity Assays

To test RDase enzyme activity, specific assays were set-up inside an anoxic glove box.[31] Anoxic 10 mL glass vials were used, containing 2 mL of an assay buffer with 200 mM potassium acetate buffer (pH 5.8), 2 mM methyl viologen, and 2 mM titanium(III) citrate [2 mM in respect to titanium(III)] and amended with 200 μM of 1,2-DCP. For the whole-cell activity assays, 200 μL of a cell concentrate (109 to 1010 cells mL–1) were inoculated to each assay vial. These experiments were performed in triplicates.

To measure activity in BNE gels, a nonstained lane from the BNE was cut into slices using a steel scalpel. Each slice was introduced into a 2 mL assay vial. Slices were smashed into smaller pieces to better expose proteins to the assay buffer. The vials were closed with Teflon-coated rubber septa and aluminum crimps and amended with 1,2-DCP from a 1 mM acetone stock solution with a glass syringe (Hamilton) to a final concentration of 200 μM 1,2-DCP.

For each activity measurement from BNE gels, different controls were included. As a positive control, 25 μL of the crude protein extract were added to the assay buffer instead of a slice. Also, controls with either 25 μL supernatant of the solubilized protein extract or the remaining pellet after solubilization were included to evaluate the efficiency of the membrane protein extraction. To control for abiotic transformation of 1,2-DCP, parallel negative controls containing the assay buffer (i) without cells or gel slices from outside the area where proteins were loaded and (ii) with 25 μL of crude extract but no 1,2-DCP was included to test the production of propene from other sources than 1,2-DCP. All controls were performed in triplicate.

The activity test vials were thoroughly mixed and incubated upside down inside the glove box, at 30 °C without shaking. After 24 h, the headspace of each sample was analyzed for propene production by gas chromatography.

Protein Identification

Silver stained bands were cut from BNE gels, the bands were washed with ddH2O and destained in a 1:1 (v/v) mixture of 30 mM K3[Fe(CN)6] and 100 mM Na2O3S2. Destained BNE bands and in-solution samples were treated as described previously to reduce and alkylate cysteine residues.[11] Protein digestion from both in gel and in-solution samples was done overnight at 37 °C on a shaker after the addition of 0.1 μg of porcine trypsin (Proteomics Sequencing Grade, Promega). The obtained peptides were extracted from the slices by 10 min incubation in a solution containing formic acid 5% (v/v) and acetonitrile 50% (v/v) three times and afterward extracts were combined. Extracts from both in gel and in-solution samples were dried, resolubilized in 0.1% of formic acid solution, and desalted using ZipTip-μC18 material (Merck Millipore).

Peptides were analyzed by nLC–MS/MS on a nanoUPLC system (nanoAcquity, Waters) hyphenated via a TriVersa NanoMate (Advion, Ltd., Harlow, UK) to an Orbitrap Fusion mass spectrometer (Thermo Scientific) as described previously.[11] Glyceraldehyde-3-phosphate dehydrogenase from Staphylococcus aureus was used as internal standard and was added to the gel slices before derivatization.

Peptide identification was conducted by Proteome Discoverer (v2.2, Thermo Fisher Scientific) using SequestHT as a search engine and a FASTA-file of proteins from the draft genome of the Dehalogenimonas strain in the culture as a database. False discovery rates for peptide identification of 1% was set as threshold, calculated versus a decoy database using the percolator node. Label free quantification of the detected proteins was performed with the abundance value obtained from Proteome Discoverer Minora node. The relative protein abundance in this study was defined as the abundance of a protein in one slice relative to the total abundance of this protein across all slices of the BNE gel lane. To detect the selenocysteine residue in OmeA different versions of the protein with a selenocysteine, a cysteine, or a serine instead of the TGA stop codon were added to the database and the search algorithm was adapted to detect selenocysteine as a modification of cysteine. Results were confirmed by manually examining the respective MS2 spectra.

Results

Draft Genome Sequence of D. alkenigignens BRE15M

Metagenome sequencing was performed on DNA extracted from the enrichment culture to investigate the genomic properties of the Dehalogenimonas strain and to allow for further proteomic studies. A high-quality draft genome of a Dehalogenimonas sp. was obtained consisting of a total of 1,654,503 bp within 18 contigs, a GC content of 56.3%, and containing 1795 predicted protein coding genes, all similar to reported Dehalogenimonas genomes (Table S1). CheckM results indicated that the genome was 97.8% complete with no contamination (0%).[32] In line with previous 16S ribosomal RNA (rRNA) gene sequencing of cultures from this enrichment,[20] the 16S rRNA gene of the Dehalogenimonas genome was 100% identical to the 16S rRNA gene of D. alkenigignens strain IP3-3, suggesting the strain in our culture belongs to the species D. alkenigignens. Average nucleotide identity (gANI) of the whole genome to D. alkenigignens strain IP3-3 showed 98.7% ANIb (BLASTN) identities from approximately 92% of the regions that aligned. We denominated our strain D. alkenigignens strain BRE15M. The close similarity with D. alkenigignens was further supported by phylogenetic analysis of concatenated protein sequences derived single copy marker genes (Figure ). Reciprocal BLASTP analyses of protein sequences from D. alkenigignens strains BRE15M and IP3-3 suggested no major functional differences exist between the strains (Supporting Information Results section).

Figure 1

Phylogenetic tree showing the relationship of D. alkenigignens strain BRE15M to other strains of Dehalogenimonas and D. mccartyi strain 195 using a phylogenomic approach. Phylogenetic analysis was performed using an alignment of concatenated protein sequences of 34 single copy marker genes from CheckM analyses.[32] The tree was constructed using default parameters using FastTree (version 2.1.10).[33] The scale bar represents 5% sequence divergence. GenBank accession numbers are in parentheses.

The genome contains 31 full-length rdhA genes that were not truncated by contig breaks, while two additional partial rdhA (BRE15M_v1_070029 and BRE15M_v1_150066; in the following section the MicroScope gene locus tag “BRE15M_v1” will be omitted) were apparently truncated because of being at contig break. Only eight rdhA had cognate genes for membrane-anchoring proteins encoded directly adjacent to the rdhA (Table S2). In total, six rdhB were automatically predicted, and we manually identified three additional possible rdhB that were adjacent to rdhA and encoded proteins with transmembrane helixes that may anchor RdhA proteins into membranes (Table S3, Figure S1). Interestingly, one rdhB gene (_030121) was found at a locus where no rdhA was present, the closest being over 22 kbp away. This brought into question whether the encoded protein may be functional as an RdhB. We therefore searched for closest homologs of this protein in other Dehalococcoidia. Although the closest homologs to this RdhB in other Dehalogenimonas spp. were also not adjacently encoded to an rdhA gene, several homologs with >80% amino acid identity in multiple D. mccartyi strains (CBDB1, DCMB5 and VS) harbored an adjacently encoded RdhA. Together this suggests locus _030121 encodes a functional RdhB.

The genome encoded various uptake and cytoplasmic hydrogenases, formate dehydrogenases and a possible OHR complex, similar to previously described in Dehalococcoides spp.[9,11] No genes for biosynthesis of quinones or cytochromes were found. Further details regarding gene annotations and protein properties are presented in the Supporting Information Results section.

Proteome Profiling

The shotgun proteomic analysis of strain BRE15M grown with 1,2-DCP identified a total of 1152 proteins, which represented the 64% of the annotated proteins in strain BRE15M (Table S12). A total of 10 RDases were expressed in strain BRE15M when grown with 1,2-DCP (Tables and S12). The most abundantly expressed RDase (_160008) was an ortholog of the RDase DcpA, which catalyzes reduction of 1,2-DCP to propene in D. mccartyi, Dehalogenimonas and uncultured bacteria. The RDase _160008 had >90% amino acid sequence identity to DcpA (Table S13).[19] The second most abundant RDase (_160007) is encoded adjacently to the DcpA ortholog and its putative membrane-anchoring protein DcpB (_160009) (Figure S2). Interestingly, _160007 is one of the four RDases which lacks the TAT signal peptide and has a predicted cytoplasmic location according to Phobius[34] (Table S2).

Table 1

List of RdhA Proteins Detected in Full Protein Extracts of D. alkenigignens Strain BRE15M Grown with 1,2-DCP as Only Halogenated Substrate for More Than 35 Transfersa

	replicate 1		replicate 2		replicate 3
accession	abundance	rank	abundance	rank	abundance	rank
BRE15M_v1_160008 (DcpA)	7.34 × 107	88	2.40 × 108	10	3.44 × 108	7
BRE15M_vl_160007	4.55 × 107	134	7.95 × 107	41	1.32 × 108	30
BRE15M_v1_40117	3.46 × 107	162			1.19 × 107	375
BRE15M_v1_30017	2.17 × 107	246	1.27 × 107	276	1.77 × 107	300
BRE15M_v1_10074			5.44 × 106	435	1.32 × 107	348
BRE15M_v1_150026	6.46 × 105	808	2.51 × 106	583	2.99 × 106	654
BRE15M_v1_30093	1.11 × 106	724
BRE15M_v1_30016	2.48 × 106	622	5.84 × 105	823	1.34 × 105	1086
BRE15M_v1_30086	1.22 × 105	896	9.75 × 104	940	1.82 × 106	746
BRE15M_v1_170013					9.49 × 104	1100
protein concentration in the sample in mg/mL	6.9		2.14		5.08
number of detected proteins	901		947		1114
sum of abundances of all detected proteins	2.74 × 1010		1.84 × 1010		2.61 × 1010

The rank columns show the rank of a protein relative to all proteins in that sample, calculated from the abundance values. Empty cells indicate that a protein was not detected in this replicate.

A membrane-bound [NiFe] hydrogenase (HupL/HupS, _070001, _070002), was expressed at high abundance (absolute abundance > 106) as described for D. mccartyi strains.[9,11,35] Also other proteins reported to be part of the OHR complex: OmeA (_040002/_040003), HupX (_040001) and OmeB (_130001)[9,11] were detected in each of the protein extract triplicates of strain BRE15M with >106 abundance. None of the few annotated genes involved in heme or quinone biosynthesis were expressed, further supporting the conclusion that strain BRE15M is independent from cytochromes and quinones. Finally, heterodisulfide reductase A and B subunits (_010277 and _010278) were expressed in strain BRE15M, proteins that are not encoded in D. mccartyi strains.[13]

Detection of the OHR Complex of Strain BRE15M by BNE

When membrane proteins of strain BRE15M were solubilized in 1% DDM without an additional cell disruption step and separated by BNE, dehalogenating activity was detected in gel slices corresponding to a molecular mass range between 146 and 242 kDa (Table S14). The activity produced low amounts of propene from 1,2-DCP by dichloroelimination. No defined bands were observed in the gel after silver staining (not shown), and most of the dehalogenating activity detected in the controls remained in the pellet control (390.14 ± 1.77 nmol propene), compared to the protein extract supernatant (146.93 ± 0.75 nmol propene). Together, this indicated inefficient protein extraction (Table S14).

Next, cell disruption by bead beating combined with solubilization with three different DDM concentrations was tested (0.05, 0.1 and 1% DDM). The detailed conditions for these experiments are shown in Table S15. The resolution of the protein bands in the native gel was different among the three conditions tested (Figure ). With 0.05% DDM, the observed bands were very weak and blurry. On the contrary, in 1.0% DDM bands were defined and intense, however, the bottom part of the gel (lower molecular weight) repeatedly showed a blue Coomassie accumulation that impeded the visualization of the bands located below. With 0.1% DDM bands were well defined along the whole lane. Parallel unstained gel lanes with the three DDM concentrations were cut into six slices each, following the stained molecular mass ruler marks and each section was subjected to a dehalogenase enzyme activity test with 1,2-DCP as electron acceptor. Propene was detected in all gel slices corresponding to a molecular weight of 146–242 kDa in the three gel lanes. Activity was also detected in gel slices at 242–480 kDa in gels with 0.1 and 1% DDM, and at 66–146 kDa in the gel with 0.1% DDM (Figure ). Overall, a slight shift of the reductive dehalogenation activity toward higher molecular weight fractions was observed at higher DDM concentration. Notably, maximum dehalogenating activity was detected in the molecular weight range between 146 and 242 kDa with 0.1% DDM (Figure ). Dehalogenating activity was never detected below 66 kDa in any of the experiments, indicating that monomeric RdhA protein (∼50 kDa) was part of a higher molecular mass respiratory complex or that it lost its activity as a monomeric enzyme.

Figure 2

Distribution of dehalogenating activity in BNE with protein membrane samples of strain BRE15M solubilized with 0.05, 0.1 or 1% (w/v) DDM after lysis by bead beating. Dehalogenating activity was measured in nanomoles of propene produced from 1,2-DCP after 24 h of incubation. Data shows means of duplicate cultures ±SD. Dehalogenating activity in the corresponding controls is shown in Table S14.

To narrow down the molecular range of the maximum dehalogenating activity, two replicate gel lanes with 0.1% DDM were cut into slices of ∼2 mm across the mass range of 63–490 kDa and were subjected to activity assays (Figure ). Maximum activity was detected in slices VI and VII in both gels, corresponding to molecular weights of 179–229 kDa. These results strongly support the presence of a organohalide respiratory complex in strain BRE15M.

Figure 3

Distribution of dehalogenating activity in BNE with protein samples of strain BRE15M solubilized with 0.1% (w/v) DDM after lysis by bead beating. Dehalogenating activity was measured as nmol propene derived from 1,2-DCP dihaloelimination detected in enzyme assays within gel slices. Data shows means of duplicate culture samples ±SD. Molecular weight ranges were estimated from gel images using Image Lab software. “M”: protein molecular mass marker.

Mass Spectrometric Analysis of BNE Slices

Each of the six analyzed regions of the stained gels with 0.05, 0.1 and 1% DDM were subjected to in-gel trypsin digestion and subsequently analyzed by nLC–MS/MS. The number of unambiguously identified proteins for each gel region ranged from 175 to 966 (Figure S3). Thirteen RdhA paralogs were detected in the gel experiment treated with 0.05% DDM, 10 with 0.1% DDM and 9 with 1% DDM (Tables S16–S18). In accordance with the production of propene detected in the in-gel dehalogenating activity tests, the DcpA ortholog predicted to be involved in 1,2-DCP dihaloelimination was abundantly detected in gel regions from the molecular range 146–242 kDa (Figure S4). Interestingly, the RDase with locus tag _160007 was the most abundant RDase expressed in the slices with maximum dechlorinating activity in the gels with 0.1 and 0.05% (Figure S4).

In addition to RDases, we detected several other proteins in the gel slices in which the OHR complex was detected (full lists are given in Tables S16–S18). The DcpA ortholog and the RDase with the locus tag _160007 comigrated with the two hydrogenase subunits HupS and HupL, the OmeA, the putative membrane-integrated anchor OmeB, and the four-FeS cluster protein HupX, in gel regions that showed maximum in-gel dehalogenating activity (regions 3–4 in proteins solubilized with 1% DDM, and region 3 in proteins solubilized with 0.05 or 0.1% DDM) (Figures and 4). These eight proteins showed a clear maximum of relative protein abundance in region 3 in the gel with 0.1% DDM-extracted proteins.

Figure 4

Relative abundance of proteins predicted to be associated with the OHR complex at different molecular weights in BNE gels after solubilization of strain BRE15M cells at different concentrations of DDM (panel A: 1% DDM, panel B: 0.1% DDM, panel C: 0.05% DDM).

Only one RdhB protein was identified with a single peptide in 0.1 and 1% DDM BNE slices being colocalized with the DcpA ortholog (Figure A,B). This RdhB protein with the locus tag _160009 is encoded in the same operon as DcpA in strain BRE15M (Figure S2) and shares 88% sequence identity (e-value 3 × 10–40) with the DcpB (locus tag Dehly_1525) in D. lykanthroporepellens strain BL-DC-9, predicted to function as a membrane anchor for DcpA. Interestingly, the topology of DcpB in strain BRE15M differs from the majority of the RdhB reported so far, with two predicted transmembrane regions instead of three (Figure S1). In summary, the identified proteins constitute the putative subunits of the OHR complex in strain BRE15M, and appear to functionally correspond with the subunits recently described for the OHR complex in D. mccartyi strain CBDB1.

After determining protein composition in broad gel bands, we also determined the presence of proteins within single defined bands (∼2 mm). For that we ran a sample on a BNE gel with proteins solubilized with 1% DDM, stained the gel with a silver protocol and analyzed dominant protein bands by nLC–MS/MS (Figure S5A, Table S19). As shown in Figure A, the RDases (catalytic subunit DcpA, and the RDase with locus tag _160007), the OmeA, the membrane subunit OmeB and the membrane-associated protein HupX were detected in the band VII (242 kDa). The membrane anchor DcpB colocalized with DcpA in band VI (202 kDa). The HupLS complex had its maximum relative abundance in band V (176 kDa). To date, it is unknown how protons are pumped across the membrane in D. mccartyi strains, but Seidel et al. (2018)[11] found that a protein annotated as a K+-insensitive pyrophosphate energized proton pump (HppA) colocalized with the OmeA/OmeB/HupX module across BNE gels in D. mccartyi strain CBDB1. Interestingly, we detected the homologous HppA protein (_080016) with a maximum of relative protein abundance in band VII (Figure A), which coincides with the OmeA/OmeB/HupX module. Potentially supporting this physical association, the hppA gene was encoded within the same gene neighborhood as the omeA and hupX genes in Dehalogenimonas strains BRE15M and IP3-3.

Figure 5

Relative abundance of proteins associated with the OHR complex at different molecular weights in BNE gels after solubilization of strain BRE15M cells at 1% DDM (panel A), 0.1% DDM (panel B), and 0.05% DDM (panel C).

The same procedure was followed for dominant protein bands in a silver-stained gel lane with 0.1% DDM-solubilized proteins focusing on the region 66–480 kDa (Figure S5B). All of the proteins suspected to be part of the respiratory complex were detected in band VII (205–229 kDa) with the exception of DcpB, and in band VI (179–205 kDa) with the exception of DcpB and HupS (Figure B, Table S20). These bands VI and VII corresponded to the unstained parallel gel slices with highest dehalogenating activity (Figure ). HupL had its maximum of relative protein abundance in band VI (179 kDa), yet HupLS are detected together in band VII, which prevented the categorical assumption of an independent hydrogenase module. A remarkable observation was again the colocalization of the DcpB ortholog with its respective DcpA in band IV (139 kDa) in a region where HupX and OmeA were absent.

Finally, a BNE gel lane with proteins extracted with 0.05% DDM was analyzed (Figure S5C, Table S21). No RDase activity was detected in any of the 19 gel slices picked from the parallel unstained gel. All subunits of the complex described above, apart from DcpB, were again identified in the excised silver stained bands. However, the distribution of the relative protein abundance across the gel lane pointed to the presence of three modules (Figure C). The two subunits of the Hup complex exhibited identical migration patterns and shared a clear maximum relative abundance in band V (123 kDa) (Figure C), which coincides well with the predicted molecular weight of the dimer of HupL/HupS (93.9 kDa). This observation strengthens the hypothesis that HupL and HupS form a hydrogenase module also in Dehalogenimonas strains. The DcpA ortholog and the RdhA with locus tag _160007 had a similar migration pattern. Unlike the protein distribution observed in gels with the 1% DDM or 0.1% DDM solubilized proteins, both RdhA were found in all 19 slices, but never more than 20% of relative abundance. Similarly, the putative proton pump protein HppA was broadly distributed across the gel. As noted earlier, RdhB proteins were not detected from 0.05% DDM treatments, suggesting the need of a higher detergent concentration to detach this protein from membrane or from the complex. The OmeA/OmeB/HupX module reached its maximum of relative protein abundance in band XVIII (450–480 kDa), and they comigrated in the molecular range between 320 and 480 kDa (bands XV to XIX). This indicated that they were tightly connected to each other. Together, these results indicate weak interaction between the OmeA/OmeB/HupX module and the HupL/HupS module but strong interactions within the modules. The lack of dehalogenating activity in the different slices of this gel indicates the need of a full multimeric complex to proceed with organohalide respiration, rather than the RdhA catalytic subunit alone.

Based on these results, we propose a respiratory complex for strain BRE15M that includes subunits of the CISM complex, subunits of the Hup hydrogenase, and RDase subunits (Figure ).

Figure 6

Model of the OHR complex of D. alkenigignens strain BRE15M. The proteins found throughout this study to be part of the suspected multimeric respiratory complex are the organohalide respiration involved molybdoenzyme OmeA (locus tag _040002/_040003), and its membrane-anchoring subunit OmeB (_130001), the hydrogenase large subunit HupL (_070001), and its small subunit HupS (_070002), the HupX (_040001), the RDase DcpA (_160008), and its putative membrane-anchor DcpB (_160009).

Discussion

The draft genome sequence of strain BRE15M provides a basis to study its proteome profile and allowed for a characterization of its organohalide respiratory system. Annotation of the genome sequence excluded the possibility that organohalide respiration is quinone-mediated in Dehalogenimonas because no functional cytochromes or biosynthetic pathways for quinones were encoded. The genome encoded 31 full-length rdhA, 22 of which were detected in our proteomic experiments (Table S2), even though 1,2-DCP was the only halogenated electron acceptor in the medium. The two most abundant RdhA proteins (DcpA ortholog and _160007) are encoded together in one operon and were colocalized across the gels regardless of the DDM concentration. To date, DcpA was the only RdhA for which a 1,2-DCP dihaloelimination function was assigned.[19] However, in our study the RDase _160007 was detected at even higher abundances than DcpA in gels with proteins solubilized with 0.1 and 0.05% DDM (Figure S4). Further comparative proteomic studies with strain BRE15M growing with different electron acceptors can elucidate whether the RdhA _160007 is specifically induced by 1,2-DCP or, instead, whether its expression is either broadly induced or even constitutive. Constitutive expression of RdhA in OHRB is common, as illustrated by the dehalogenase CbdbA80 in D. mccartyi strain CBDB1.[36] The colocalization in the genome as well as in gel slices of these two RdhA may suggest either strong interaction, or that they are performing the same function in different copies of the respiratory complex. These results are nevertheless intriguing because RdhA _160007 does not have a TAT leader peptide. Further, no RdhA sequences from strain BRE15M contained predicted multiple transmembrane helix domains that might directly anchor them to membranes, as was recently described in other genomes.[37−39] One possibility is that both RDases are exported together as a heterodimer via the TAT machinery associated with the DcpA, and once in the outer part of the membrane, _160007 is either attached to a different copy of the DcpB or remains attached to DcpA forming a stable heterodimer in vivo. This hypothesis would assume that one RdhB protein can anchor different RdhA proteins to the membrane. If this is true, the four RdhA proteins without TAT leader peptide encoded in the genome might be anchored to one of the nine encoded RdhB, forming heterodimers with other RdhAs. This may occur even if the RdhA protein without TAT leader peptide is not encoded in the same operons as another RdhA protein. Alternatively, a feasible explanation for the lack of TAT leader peptide is that a RdhA protein might function as a cytoplasmatic nonrespiratory dehalogenase, as reported for aerobic bacteria[5,37] and that the association with the OHR complex is an artifact after cell disruption.

In this study, we also provide biochemical evidence that RDases in strain BRE15M are part of a large respiratory complex. The nonionic detergent DDM was suitable to detach membrane associated proteins while maintaining protein–protein interactions. Protein solubilization with 0.1% (w/v) DDM enabled the identification of a putative multiprotein complex composed by subunits of the CISM complex, a dimeric [NiFe] hydrogenase, and the RDase system in one single gel slice that was at a position in the gel with maximum dechlorination activity within the measured gel lane. Our results with lower DDM concentration (0.05%) were surprisingly counterintuitive because it led to partial disintegration of the protein complex. This might be explained by the fact that the ionic protein–protein interactions within the complex are not affected by this nonionic detergent, but they may disrupt lipid–protein interactions of membrane proteins. In contrast, increasing concentrations of DDM might drive the protein subunits to move together, as high salt pushes hydrophobic interaction together.

The analysis of the complexome using 0.05% (w/v) DDM, where the complex partially disintegrates, strongly suggests the presence of three modules. The largest module comprised OmeA, OmeB and HupX. The calculated size of this module of 180.9 kDa did not match with the molecular mass corresponding with the gel slice in which these proteins had its relative maximum abundance (band XVIII, Figure C, 450 kDa), suggesting that the Ome/HupX module polymerizes. As observed from the distribution of 0.05% DDM-solubilized proteins across the gel, HupX had a much stronger interaction with OmeA and OmeB than with the two hydrogenase subunits, which agrees with previous results with D. mccartyi strain CBDB1.[11] However, a noticeable difference between these two strains is that HupX is encoded in an operon together with HupL and HupS in D. mccartyi strain CBDB1,[10] while it forms an operon with OmeA and OmeB in strain BRE15M, further supporting its role as the four iron–sulfur cluster β-subunit of the CISM module. We also biochemically detected a protein annotated as K+-insensitive pyrophosphate energized proton pump (HppA) that migrates similarly to the OmeA/OmeB/HupX module in a BNE gel with 0.1% DDM-solubilized proteins. However, the broad distribution of HppA across the bands in BNE gels with 0.05% DDM questions any strong interaction between HppA and the OmeA/OmeB/HupX module but provides a basis to speculate that HppA could be part of the respiratory complex of strain BRE15M and may translocate protons through the subunits of this complex.

The hydrogenase subunits HupL and HupS constitute another defined module and its facile separation from the other OHR complex subunits has also been observed in D. mccartyi strain CBDB1.[11] The molecular weight range in which the maximum relative protein abundance of these subunits was detected is 159–179 kDa in the 0.1% DDM treatment (band V) and 123–146 kDa in the 0.05% DDM treatment (band V), which slightly differs from the theoretical calculated size of 93.9 kDa for this module, however, size determinations in BNE gels can only considered as rough estimations. Finally, the RDase DcpA, which functioned as terminal reductase in the respiratory complex, is attached to the putative membrane anchor DcpB. We detected comigration of both proteins in BNE gels loaded with proteins solubilized with 1% DDM (band VI, 202 kDa, Figure A) and 0.1% DDM (band IV, 139 kDa, Figure B). These results constitute, together with a recent study with D. mccartyi strain CBDB1, one of the few biochemical evidences of specific interaction between an RdhA and its corresponding RdhB.[10]