Impact of Vitamin B₁₂ on Formation of the Tetrachloroethene Reductive Dehalogenase in Desulfitobacterium hafniense Strain Y51

Abstract

Corrinoids are essential cofactors of reductive dehalogenases in anaerobic bacteria. Microorganisms mediating reductive dechlorination as part of their energy metabolism are either capable of de novo corrinoid biosynthesis (e.g., Desulfitobacterium spp.) or dependent on exogenous vitamin B₁₂ (e.g., Dehalococcoides spp.). In this study, the impact of exogenous vitamin B₁₂ (cyanocobalamin) and of tetrachloroethene (PCE) on the synthesis and the subcellular localization of the reductive PCE dehalogenase was investigated in the Gram-positive Desulfitobacterium hafniense strain Y51, a bacterium able to synthesize corrinoids de novo. PCE-depleted cells grown for several subcultivation steps on fumarate as an alternative electron acceptor lost the tetrachloroethene-reductive dehalogenase (PceA) activity by the transposition of the pce gene cluster. In the absence of vitamin B₁₂, a gradual decrease of the PceA activity and protein amount was observed; after 5 subcultivation steps with 10% inoculum, more than 90% of the enzyme activity and of the PceA protein was lost. In the presence of vitamin B₁₂, a significant delay in the decrease of the PceA activity with an ∼90% loss after 20 subcultivation steps was observed. This corresponded to the decrease in the pceA gene level, indicating that exogenous vitamin B₁₂ hampered the transposition of the pce gene cluster. In the absence or presence of exogenous vitamin B₁₂, the intracellular corrinoid level decreased in fumarate-grown cells and the PceA precursor formed catalytically inactive, corrinoid-free multiprotein aggregates. The data indicate that exogenous vitamin B₁₂ is not incorporated into the PceA precursor, even though it affects the transposition of the pce gene cluster.

INTRODUCTION

The anaerobic reductive dehalogenation of organohalides is a metabolic feature widespread among the genus Desulfitobacterium (35). Aliphatic and also aromatic halogenated hydrocarbons (e.g., chlorinated or brominated ethenes and polychlorinated phenols) are reductively dehalogenated by different Desulfitobacterium strains. These strains possess different reductive dehalogenases mediating the anaerobic dehalogenation. Almost all reductive dehalogenases isolated so far harbor a corrinoid cofactor at the active site (11).

The Gram-positive Desulfitobacterium hafniense strain Y51 was shown to have a corrinoid-dependent reductive dehalogenase (PceA) that dechlorinates tetrachloroethene (PCE) to cis-1,2-dichloroethene (33, 34). The gene encoding the PceA enzyme is located in the pceABCT gene cluster that is flanked by insertion sequences including transposase genes. The pceB gene product was proposed to serve as a membrane anchor for PceA (24), although this role has never been confirmed so far. The pceC gene shows homology to open reading frames (ORFs) encoding transmembrane transcriptional regulators of the NirI/NosR-type involved in nitrite or nitrous oxide reduction (28, 39). The pceT gene carries the genetic information for a peptidyl-prolyl cis/trans isomerase. Recently, a role of the PceT protein in the maturation of PceA was proposed (22), and its interaction with the Tat (twin arginine translocation) signal peptide of the intracellular precursor of PceA was shown (20).

When D. hafniense strain Y51 is cultivated in the absence of PCE, the pce gene cluster in whole or in part is irreversibly lost by transposition events (8). Circular intermediates formed after the excision of the transposable elements were identified in D. hafniense strain Y51 (8) and earlier in the closely related D. hafniense strain TCE-1 (4, 19). An acquisition of the gene cluster by horizontal gene transfer has been discussed (19).

For PCE-dependent growth of D. hafniense Y51, no exogenous corrinoid has to be added (33). This can be explained by the presence of corrinoid biosynthetic genes in the genome of the organism (25) indicating the de novo formation of the corrinoid cofactor of PceA. In addition, corrinoid salvaging might occur via the functional expression of a genome-encoded vitamin B₁₂-specific ATP-binding cassette (ABC) transporter (btuFCD). The structure of the PceA corrinoid cofactor in D. hafniense strain Y51 has not been identified so far. The only corrinoid cofactor of reductive dehalogenases identified as yet is that of the tetrachloroethene-reductive dehalogenase in the Gram-negative Sulfurospirillum multivorans, which is a unique norpseudo-B₁₂ (16). This cofactor is synthesized de novo by the organism (unpublished results). Therefore, S. multivorans is able to grow with PCE in the absence of exogenous corrinoids. In contrast, the organohalide-respiring Dehalococcoides mccartyi (phylum Chloroflexi) is strictly dependent on the addition of vitamin B₁₂ to the growth medium (18).

The necessity of the corrinoid cofactor for reductive dehalogenase function became eminently evident when the heterologous production of the enzyme in Escherichia coli was tested, an organism lacking de novo corrinoid biosynthesis (1). In such experiments, the nonactive reductive dehalogenase apoprotein formed intracellular protein aggregates (15, 24, 34). Recently, it has been shown that the solubility of the heterologously produced PceA in E. coli can be increased by the coproduction of its dedicated chaperone, PceT (20). However, no reductive dehalogenase enzyme activity was reported, most probably due to the absence of the corrinoid cofactor.

The anaerobic reductively dehalogenating bacteria (11, 21, 35) were isolated from different environments including soil and sediment and were envisaged as a tool in bioremediation of contaminated sites (5). Hence, the availability of corrinoids in natural environments or at polluted sites and their effect on the reductive dehalogenation may be relevant for the dechlorination potential in soil contaminated with organohalides. Other anaerobic prokaryotes in these environments such as acetogens or methanogens contain corrinoids, which, upon occasional lysis of the organisms, are released and therefore available for dehalogenating bacteria.

The study presented here sheds light on the interplay between de novo corrinoid biosynthesis and the formation of a catalytically active reductive dehalogenase. The effect of exogenous corrinoids on the stability of the pceA gene, its transcription, and its translation as well as the maturation of the PceA protein was investigated. A model for the maturation of the protein depending on the presence of corrinoids and of PCE was developed.

MATERIALS AND METHODS

Cultivation of the organism.

Desulfitobacterium hafniense strain Y51 (33) was cultivated under anoxic conditions in a defined medium described by Scholz-Muramatsu et al. (30). Pyruvate (40 mM) was added as the electron donor and either PCE (10 mM) or fumarate (40 mM) as the electron acceptor. Per 50 ml medium, 1 ml of 0.5 M PCE dissolved in hexadecane was applied. The cultivation was carried out at 28°C in serum glass bottles closed with butyl rubber stoppers or Teflon septa and anaerobized with nitrogen. Unless stated otherwise, vitamin B₁₂ (cyanocobalamin) was not added to the growth medium.

PceA enzyme activity measurements.

The activity of the tetrachloroethene-reductive dehalogenase (PceA) was measured according to the method of Neumann et al. (23). Cells from a 100-ml culture were sedimented and resuspended in 1 ml anoxic buffer (50 mM Tris-HCl, pH 7.5). The cell suspension was mixed with an equal volume of glass beads (0.25- to 0.5-mm diameter; Carl Roth GmbH, Karlsruhe, Germany). For cell disruption, a bead mill was used (5 min 25 Hz; Mixer Mill MM400; Retsch, Haan, Germany). Cell debris was removed by centrifugation (1 min, 5,250 × g). The protein concentration was determined according to the method of Bradford (2) using the Roti-Nanoquant reagent (Carl Roth GmbH, Karlsruhe, Germany).

Immunoblot analysis.

For the detection of PceA protein, crude extracts of D. hafniense Y51 (5 μg protein/lane) were subjected to SDS-PAGE (13.5% polyacrylamide gels). The immunological analysis was done as described earlier (13, 14). PceA antibodies were generated by immunization of rabbits with heterologously produced PceA from D. hafniense PCE-S. PceT antibodies were kindly provided by T. Futagami (Kyushu University, Fukuoka, Japan). The PceA antiserum was diluted 5,000-fold, the PceT antiserum 50,000-fold, and the antibodies were detected via a secondary antibody coupled to alkaline phosphatase.

FRIL.

For freeze fracture replica immunogold labeling (FRIL), D. hafniense Y51 cells were harvested in the late exponential growth phase by centrifugation (4,600 × g, 5 min, 28°C) and washed with substrate-free medium. Resuspended cells were rested in anaerobic glass vials for at most 30 min at 28°C. Freeze fracture of the cells was done according to the protocol described for Sulfurospirillum multivorans (13). A normal replication procedure with platinum (2-nm thickness) as the first layer and carbon (20-nm thickness) as the second layer was used. Replicas were treated with 5 mg/ml lysozyme for 30 min at 37°C. The replica immunolabeling was carried out as described earlier (7, 13). The PceA antibody was diluted 1:50 in labeling blocking buffer (LBB) consisting of 1% (wt/vol) bovine serum albumin, 0.5% gelatin, and 0.005% Tween 20. Images were taken as digital pictures with a Zeiss EM 902A electron microscope (Zeiss, Oberkochen, Germany) operated at 80 kV using a FastScan TVIPS CCD-camera 1k × 1k (TVIPS, Munich, Germany). The digital camera was operated by the E-Menu4 software (TVIPS, Munich, Germany). The number of immunogold signals/μm² was determined as the average of the fractured areas of 10 cells.

DNA isolation.

D. hafniense Y51 was subcultivated with either PCE or fumarate; the subcultivation experiments were performed twice for each electron acceptor. In the late exponential growth phase (optical density at 578 nm [OD₅₇₈], ∼0.6), samples for DNA isolation were taken from the cultures and harvested by centrifugation (16,000 × g, 5 min, 10°C). The cell pellets were stored at −20°C. DNA was extracted using the InnuPREP Bacteria DNA kit (Analytik Jena, Jena, Germany). An additional RNA digestion was carried out by RNase I treatment (Sigma-Aldrich, Hamburg, Germany). The quality of isolated DNA was determined via agarose gel electrophoresis and photometric analysis (UV/Vis Cary 100 spectrophotometer; Agilent Technologies, Böblingen, Germany). Isolated DNA was stored at −20°C.

RNA isolation and reverse transcription.

Sample preparation was done according to the DNA isolation protocol (see above) in the exponential growth phase (OD₅₇₈, ∼0.4). Cell pellets were stored at −80°C. RNA was isolated using the RNeasy minikit (Qiagen, Hilden, Germany). An additional DNA digestion was carried out by treatment with DNase I (RNase free; Roche, Mannheim, Germany). The quality of isolated RNA was determined via agarose gel electrophoresis and photometric analysis (UV/Vis Cary 100 spectrophotometer; Agilent Technologies, Böblingen, Germany). Isolated RNA was stored at −80°C. Reverse transcription of 1.25 μg total RNA was carried out for pceA using the RevertAid First Strand cDNA synthesis kit (Fermentas, St. Leon-Rot, Germany). As reference, the transcript of the rpoB gene was used (for primer sequences, see “qPCR” below). To exclude DNA contamination, RNA without reverse transcriptase treatment was applied to quantitative real-time PCR (qPCR). Isolated cDNA was stored at −20°C.

qPCR.

Quantification of gene and transcript levels was implemented by qPCR in a CFX96 Real-Time PCR system (Bio-Rad, Munich, Germany) for pceA (GenBank accession no. AAW80323.1; forward primer, 5′-GGA GTG TAA TCC CGC TTT ATC-3′; reverse primer, 5′-AAT TTC CAC TGT TGG CCT TGT-3′; 136 bp) and for rpoB as the reference gene (NCBI accession no. YP_516696.1; forward primer, 5′-GAT TCG GGC TTT GGG TTA TGC-3′; reverse primer, 5′-CGC AGA CGC TTG TAG ATT TCC-3′; 138 bp). The rpoB gene and transcript level were chosen because they remain stable during subcultivation. Reaction mixtures contained 5 μl 2× MaximaTM SYBR green qPCR Master Mix (Fermentas, St. Leon-Rot, Germany), 0.4 μl of 10 pmol/μl of each primer, and 25 ng DNA or 0.8 μl cDNA (final volume, 10 μl). Each reaction was carried out three times. A three-step cycling program with an initial denaturation for 10 min at 95°C followed by 40 cycles of 15 s at 95°C, 30 s at 60°C, and 30 s at 72°C was chosen. Data acquisition was performed during the extension step. After the program was finished, a melting curve analysis to specify the PCR product was carried out. Data analysis was carried out via the 2^−ΔΔCt method (17). As calibrator, DNA or cDNA of D. hafniense Y51 cultivated on pyruvate and PCE was used. In each experiment, the threshold cycle (C_T) value of no-template controls (NTC) was about 30. The difference between the C_T values of the samples with the lowest template concentrations and of the NTCs never did fall below a minimum of 7 cycles.

Isolation and analysis of corrinoids.

Cells of D. hafniense Y51 were harvested by centrifugation (10 min, 6,700 × g, 10°C); 5 g of wet cells was resuspended in 30 ml 50 mM Tris-HCl (pH 7.5). The cells were disrupted in a French pressure cell at 2,000 lb/in² (French pressure cell press; Sim-Aminco, Spectronic Instruments, New York, NY). The corrinoid extraction protocol was based on the method described by Stupperich et al. (32). The final concentration of KCN in the crude extract was raised to 100 mM. The Amberlite XAD4 (Sigma-Aldrich, Munich, Germany) was washed with methanol and equilibrated with 0.1% acetic acid prior to use. Each milliliter of cyanide extract was mixed with 0.25 g XAD4 material (16 h, 22°C, 400 rpm). The XAD4 material was subsequently washed with distilled water. The corrinoids were eluted with methanol, and the eluate was evaporated to dryness in a Speed Vac Concentrator (Speed Vac Concentrator 100H; Savant, Midland, MI). Concentrates were dissolved in distilled water and applied to a column filled with 3 g neutral aluminum oxide. Elution was conducted with distilled water. The first 20 ml of the eluate was concentrated to 100 μl in a vacuum concentrator. For analysis of isolated corrinoids, a UV/visible light (UV/Vis) absorption spectrum was recorded (UV/Vis Cary 100 spectrophotometer; Agilent Technologies, Böblingen, Germany).

Isolation and purification of PceA aggregates.

Cells of D. hafniense Y51 of subculture 15 were harvested by centrifugation in the late exponential growth phase (OD₅₇₈, ∼0.6). Cells resuspended in 50 mM Tris-HCl (pH 7.5) were disrupted in a French pressure cell at 2,000 lb/in² (French pressure cell press; Sim-Aminco, Spectronic Instruments, New York, NY). Cell debris and large membrane fragments were sedimented in two centrifugation steps (5,250 × g and 12,000 × g, 20 min, 4°C). The supernatant of the second centrifugation step was subjected to ultracentrifugation (100,000 × g, 45 min, 4°C) and thus separated into soluble extract and particulate fraction. The latter fraction was washed twice with 50 mM Tris-HCl (pH 7.5) and analyzed via negative-staining electron microscopy (see below).

The extracted membrane fraction was resuspended in solubilization buffer (20 mM Tris-HCl [pH 8.0], 0.2 M NaCl, 1% sodium deoxycholate). After 30 min of incubation at room temperature, the suspension was centrifuged, and the pellet was washed four times with 0.25% sodium deoxycholate in 20 mM Tris-HCl (pH 8.0). After repeated electron microscopic control, the pellet was resuspended in distilled water and subjected to density gradient centrifugation. A discontinuous sucrose gradient (70, 72, 74, 76, 78, 80, 82, 84, 86, and 100% [wt/vol]; volume, 2 ml) was created. Immediately after centrifugation (3 h; 100,000 × g; 4°C) each layer (2 ml) of the gradient was carefully removed and protein precipitation according to the method of Wessel and Flügge (37) was conducted. The precipitated protein was resuspended in 30 μl 50 mM Tris-HCl (pH 7.5). Equal volumes of the fractions were analyzed via SDS-PAGE (5 μl for immunoblot analysis, 2 μl for silver staining). Silver staining was carried out according to the method of Schägger (29). From the lane of the aggregate-containing fraction, major bands were excised. As a control, the same areas were excised from the lane of the aggregate-free fraction of PCE-grown cells.

Analysis of the PceA aggregates.

The silver was removed by adding potassium ferricyanide and sodium thiosulfate (9). Subsequently, the sample was proteolytically digested as described earlier (12). Peptides were reconstituted in 0.1% formic acid, injected by an autosampler, and concentrated on a trapping column (nanoAcquity ultrahigh pressure liquid chromatography [UPLC] column, C₁₈, 180 μm by 2 cm by 5 μm; Waters, Eschborn, Germany) with water containing 0.1% formic acid at flow rates of 15 μl/min. After 4 min, the peptides were eluted onto the separation column (nanoAcquity UPLC column, C₁₈, 75 μm by 100 mm by 1.7 μm; Waters, Eschborn, Germany). Chromatography was performed by using 0.1% formic acid in solvents A (100% water) and B (100% acetonitrile), with peptides eluted over 30 min with an 8 to 40% solvent B gradient using a nano-high-pressure liquid chromatography (nano-HPLC) system (nanoAcquity; Waters) coupled to an LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific). Continuous scanning of eluted peptide ions was carried out between 400 and 2,000 m/z, automatically switching to tandem mass spectrometry collision-induced dissociation (MS/MS CID) mode on ions exceeding an intensity of 3,000. For MS/MS CID measurements, a dynamic precursor exclusion of 3 min was enabled.

Raw data were applied to a database search using the Thermo Proteome Discoverer software (v1.0 build 43) to carry out a tandem ion search algorithm from the Mascot house server (v2.2.1) by database comparison against the genome project entry of D. hafniense Y51 in the National Center for Biotechnology Information (NCBInr database, September 2010) with a 10-ppm tolerance for the precursor and 0.8 Da for MS-2 fragments. Furthermore, trypsin with a maximum of two missed cleavage sites was selected, and variable modifications, such as methionine oxidation and carbamidomethylation of cysteine, were allowed. Peptides were considered to be identified by Mascot when a probability of <0.05 (probability-based ion scores threshold, >40) was achieved. Protein identification was positive when at least two peptides were identified.

Negative-staining electron microscopy.

The aggregate-containing fraction was adsorbed to carbon-coated 400-mesh copper grids for 10 min, washed in distilled water, and negatively stained with 2% uranyl acetate for 1 min. For microscope and imaging parameters, see above.

RESULTS AND DISCUSSION

PceA formation in fumarate-grown cells is influenced by vitamin B₁₂.

D. hafniense Y51 was routinely cultivated on a pyruvate/PCE-containing medium (30) to select for cells functionally expressing the pce genes. For strain maintenance, the organism was grown on medium void of vitamin B₁₂ (cyanocobalamin). D. hafniense Y51 is able to synthesize corrinoids de novo. Pyruvate was used as the electron donor and PCE as the electron acceptor in the energy metabolism of the organism. The addition of vitamin B₁₂ (50 μg/liter cyanocobalamin) to the cultures had no influence on the PCE-reductive dehalogenase (PceA) activity in crude extracts or on the formation of the PceA protein as tested via immunoblotting (Fig. 1A and B, lane 1, “with PCE”). This result pointed to an adequate supply of the PceA enzyme with corrinoid cofactor produced by the de novo corrinoid biosynthesis in D. hafniense Y51. PCE-grown cultures (with and without vitamin B₁₂) were used to inoculate media containing fumarate rather than PCE as the electron acceptor. For the subsequent long-term cultivation, an inoculum of 10% was used for each transfer. The experiment was conducted in media either with or without vitamin B₁₂. During the subcultivation, PceA activity and the PceA protein level were determined in the late exponential growth phase. It should be noted that two forms of PceA were detected in the crude extracts (prePceA with the Tat signal peptide and PceA without the Tat signal peptide) (Fig. 1A and B).

Fig 1.

Protein level and specific activity of PceA during subcultivation of D. hafniense Y51 on pyruvate and fumarate in the absence of PCE. The cultivation was carried out without vitamin B₁₂ (A) or with vitamin B₁₂ (50 μg/liter) (B) in the growth medium. The first lane represents the preculture cultivated on pyruvate and PCE. Crude extracts (5 μg protein for each lane) were separated by SDS-PAGE and analyzed via immunoblotting with PceA-specific antibodies. pre, precursor form of PceA with the Tat signal peptide; mat, processed form of PceA without the Tat signal peptide.

After one cultivation step in the absence of PCE and vitamin B₁₂, one-half of the PceA activity was lost, and after six transfers almost no activity was detectable (Fig. 1A). The same result was observed for the PceA protein. The loss of the PceA activity was significantly delayed when vitamin B₁₂ was present in the medium (Fig. 1B). After 10 transfers, the enzyme activity was decreased by almost 50%. Until subcultivation step 20, the PceA protein level was comparable to that of PCE-grown cells; however, less than 10% of the enzyme activity remained. This experiment was repeated several times and was reproducible for the subcultivation in the absence of vitamin B₁₂. In the presence of vitamin B₁₂, the enzyme activity was reduced by 90% after a minimum of 15 to 20 subcultivation steps; however, different results were obtained for the number of transfers required for a complete loss of the PceA protein (see Fig. S1 in the supplemental material). The minimal transfer number was about 30 as shown in Fig. 1B. In other experiments, the PceA level remained constant as determined by immunoblotting. Nonetheless, all experiments lead to the conclusion that inactive PceA protein is formed in the course of the subcultivation in the absence of PCE and in the presence of vitamin B₁₂. Our results are in accordance with a recent study with D. hafniense strain TCE-1, which showed a very slow decrease within 30 subcultivations in the presence of vitamin B₁₂ (250 μg/liter) (4). An earlier study with D. hafniense Y51, however, showed a fast decrease of the pceA gene level within only a few subcultivation steps (8). In this study, a vitamin B₁₂-free medium was used. The discrepancy between the two preceding studies can now be explained by our finding that vitamin B₁₂ has an impact on the rate of PceA decrease.

The pceA gene level is affected by vitamin B₁₂.

The loss of PceA upon subcultivation of D. hafniense Y51 has earlier been described and attributed to the excision of the pce gene cluster (8). To monitor the loss of the pceA gene during the long-term subcultivation experiment depicted in Fig. 1, qPCR was used. After 2 subcultivation steps in the absence of PCE and vitamin B₁₂, a decrease of almost 50% of the pceA gene level was observed (Fig. 2A). No pceA gene was detected after 8 transfers (corresponding to ∼27 generations). Concomitantly with the loss of the pceA gene, the pceA transcript level was reduced as determined by reverse transcription-qPCR (Fig. 2B). In the presence of vitamin B₁₂, a slow decrease of the gene number was observed (Fig. 2A). In subculture 15, the pceA level was still about 50%. The complete loss of the pceA gene in cultures supplemented with vitamin B₁₂ occurred after 60 subcultivations steps. Again, the pceA transcript level decreased almost in parallel to the pceA gene level (Fig. 2B). These results point to a stabilizing effect of vitamin B₁₂ on the pce gene cluster in D. hafniense Y51, which also explains the effect of vitamin B₁₂ on PceA in the course of the long-term cultivation described above. This result was unexpected and cannot be explained so far.

Fig 2.

Relative pceA gene level (A) and transcript level (B) in cultures of D. hafniense Y51 determined by quantitative PCR. Cells were subcultivated in the absence of PCE with or without cyanocobalamin (vitamin B₁₂, 50 μg/liter) in the growth medium.

Localization of PceA in the presence or absence of PCE and vitamin B₁₂.

The finding that the PceA activity decreased faster than the level of the PceA protein in cells grown in the presence of vitamin B₁₂ raised the question of the fate of the exogenous corrinoid on the one hand and of the PceA protein on the other hand. Therefore, cells subcultivated for 3 and 6 steps in the absence, or for 30 and 60 steps in the presence, of vitamin B₁₂ were subjected to corrinoid extraction. In both cultivations, the corrinoid level decreased significantly (Fig. 3). With exogenous vitamin B₁₂, less than ∼20% of the initial level was recovered in cells of subcultivation step 30 (Fig. 3A). In vitamin B₁₂-depleted cells, only a minor decrease of the corrinoid level was observed after 10 subcultivation steps (Fig. 3B). It is assumed that a cofactor-free PceA apoprotein is formed when the cells are subcultivated in the long term with fumarate instead of PCE and in the presence of vitamin B₁₂.

Fig 3.

UV/Vis absorption spectra of isolated corrinoids from cells of D. hafniense Y51 subcultivated on pyruvate and fumarate in the absence of PCE. Cells were cultivated with vitamin B₁₂ (50 μg/liter) (A) or without vitamin B₁₂ (B) in the growth medium. sc, subculture.

To investigate the subcellular localization of PceA under the different growth conditions (pyruvate/PCE and pyruvate/fumarate either with or without vitamin B₁₂) by electron microscopy, we used PceA-directed antibodies for freeze fracture replica immunogold labeling (FRIL). This method was shown earlier to be useful for the detection of membrane integral proteins or proteins attached either to the protoplasmic or the exoplasmic face of the cytoplasmic membrane (6, 7, 13, 26, 31, 38). As an example, the results obtained for subculture 15 in the presence of vitamin B₁₂ are shown in Fig. 4.

Fig 4.

Subcellular localization of PceA in D. hafniense Y51 determined by freeze fracture replica immunogold labeling (FRIL). Cells were cultivated on pyruvate and PCE (A to C) or on pyruvate and fumarate in the absence of PCE (D to F) (15 subcultivations). Cultivation was carried out with vitamin B₁₂ (50 μg/liter) in the growth medium. The electron micrographs show freeze fractures through the cytoplasm (Cy), the protoplasmic fracture face (PF), and the exoplasmic fracture face (EF). The PceA protein was detected using PceA-specific antibodies and labeled with 10 nm gold-conjugated secondary antibodies (black dots). The scale bar in panel A represents 100 nm and is valid for all the panels.

To compare the results obtained under different growth conditions, the PceA signals per area were counted (Fig. 4). The PceA signal distribution shown in Fig. 4A to C for cells grown on pyruvate/PCE revealed the presence of PceA proteins in the cytoplasm, at the protoplasmic fracture face, and predominantly at the exoplasmic fracture face of the cell membrane. The PceA signals in all three detection areas were almost evenly distributed and did not accumulate in a certain cell section. In contrast, in cells grown in the absence of PCE, the PceA protein was detected in large aggregates localized in the cytoplasm (Fig. 4D). In the majority of the analyzed cells, just one cytoplasmic PceA patch that seemed to be associated with the membrane in each case was discovered. Under these conditions, PceA signals were found neither on the protoplasmic face nor on the exoplasmic face of the cytoplasmic membrane (Fig. 4E and F). The absence of exoplasmic PceA in the freeze fracture of whole cells and the presence of both forms of PceA in the crude extract (Fig. 1B) seem contradictory at first glance. These may be explained by a partial processing of prePceA by the cells' signal peptidase, which may occur upon lysis of the cells. The results depicted in Fig. 4 point to a predominantly exoplasmic and membrane-bound state of PceA in PCE-grown cells and an intracellular and aggregated state of PceA in fumarate-grown cells. The aggregation of PceA in the cytoplasm was observed in PCE-depleted cultures independent of the absence or presence of exogenous vitamin B₁₂ but never in cells cultivated with PCE. The aggregation was already visible after one transfer to medium void of PCE (see Fig. S2 in the supplemental material). When the closely related strain D. hafniense PCE-S (35) was subjected to the same experimental procedure and pyruvate/fumarate-grown cells were investigated using FRIL, similar PceA-containing aggregates were detected (see Fig. S3 in the supplemental material). An aggregation of PceA was also observed upon long-term subcultivation of D. hafniense Y51 in the absence of vitamin B₁₂.

It is conceivable that the aggregates formed in the absence of PCE serve as a “reservoir” for prePceA, which may be maturated as soon as PCE becomes available. In this case, PceT might be required as a chaperone for prePceA maturation. Therefore, we investigated the localization of PceT by FRIL in the absence or presence of PCE (see Fig. S4 in the supplemental material) using a specific PceT antibody (22). The experiments showed that in PCE-grown cells the PceT protein was evenly distributed either in the cytoplasm (see Fig. S4A in the supplemental material) or on the protoplasmic face of the cytoplasmic membrane (see Fig. S4B). Almost no PceT signals were detected on the exoplasmic face (see Fig. S4C). In fumarate-grown cells of D. hafniense Y51, the PceT protein showed aggregate formation (see Fig. S4D) as detected for prePceA inside the cell (Fig. 4D and E). While the prePceA protein was detected exclusively in the aggregates, still single PceT signals were found in the cytoplasm or at the protoplasmic face of the cytoplasmic membrane. No PceT signals were detected on the exoplasmic fracture face of the cytoplasmic membrane (see Fig. S4F in the supplemental material).

It was proposed earlier that PceT keeps the prePceA protein in an open conformation for the incorporation of the metal-containing cofactors (20). Such an incomplete folding might support the protein aggregation in D. hafniense Y51 cells in the absence of the corrinoid cofactor. The earlier finding that PceT binding increased the solubility of heterologously formed cofactor-free prePceA in E. coli (20) is not necessarily contradictory to our observation. Other proteins or factors present in D. hafniense Y51 might favor the aggregate formation.

Analysis of the PceA aggregates.

Upon subcellular fractionation of D. hafniense Y51 cells grown for 15 subcultivations on pyruvate/fumarate in the presence of vitamin B₁₂, most of the PceA protein sedimented with the membrane fraction, probably due to sedimentation of the PceA-containing aggregates. Immunoblot analysis of the membrane fraction revealed that exclusively the precursor form of PceA was present. The isolated membrane fraction containing the protein aggregates showed no significant PceA enzyme activity and level of corrinoids (data not shown). Resuspended membrane fractions derived from pyruvate/fumarate-grown cells were subjected to electron microscopy to examine the size and shape of the prePceA-containing aggregates. The enriched particles (Fig. 5A) exhibited an almost globular shape with a diameter of about 100 nm. In some of the particles, a groove was visible with a central cavity the origin of which is unknown. No aggregates were detected when the aggregate enrichment procedure was conducted with cells grown on pyruvate/PCE. When the large-deletion (LD) variant (8), which lacks the pce genes, was applied to FRIL, no protein patches were found inside the cells (see Fig. S5 in the supplemental material). These results indicate that the protein aggregation in the D. hafniense Y51 wild-type cells is a PceA-dependent process.

Fig 5.

Analysis of PceA aggregates enriched from cells of D. hafniense Y51 cultivated on pyruvate and fumarate in the absence of PCE. Cultivation was carried out with vitamin B₁₂ (50 μg/liter) in the growth medium. The enriched PceA aggregates were subjected to electron microscopy (A, arrows) and SDS-PAGE (B). Lane 1, silver stain; lanes 2 and 3, immunoblot analysis of the prePceA and the PceT protein using specific antibodies. prePceA (DSY2839), precursor form of the PCE reductive dehalogenase; EF-Tu (DSY0469), elongation factor Tu; PceT (DSY2836), peptidyl-prolyl cis/trans isomerase; CobT (DSY2114), nicotinate-nucleotide dimethylbenzimidazole phosphoribosyltransferase.

To elucidate the major components of the intracellular protein aggregates, the particles were extracted from the membrane fraction by stepwise washing with detergent and purified via sucrose density gradient centrifugation. The fraction containing the aggregates (86% [wt/vol] sucrose) was subjected to electron microscopy. The size and shape of the particles seemed to be unaffected by the purification procedure. Subsequently, the aggregate fraction was applied to one-dimensional SDS-PAGE (Fig. 5B, lane 1). The most salient protein bands, visible after silver staining, were excised and analyzed using LC-MS/MS measurements. In addition to the precursor of PceA and the PceT protein, the presence of which was also proven via immunoblot analysis (Fig. 5B, lanes 2 and 3), other proteins were found. Besides elongation factor Tu (EF-Tu/DSY0469; Fig. 5B, lane 1), a nicotinate-nucleotide dimethylbenzimidazole phosphoribosyltransferase (CobT/DSY2114; Fig. 5B, lane 1) was unambiguously identified as a component of the prePceA aggregates. This enzyme is involved in the late steps of corrinoid biosynthesis (for a review, see reference 36). When PCE-grown cells were subjected to the same fractionation procedure, no bands were detected on the SDS polyacrylamide gels.

The aggregate composition implicates an impeded biosynthesis and maturation of the prePceA protein. The aggregation also indicates irregular folding of the cofactorless apoprotein and may precede protein degradation. The presence of elongation factor Tu, a component of the translation machinery (27), and the absence of corrinoids point to the formation of the protein aggregates either in parallel with or immediately after the translation of the reductive dehalogenase transcript and prior to cofactor incorporation. The incorporation of the corrinoid cofactor into the enzyme is expected to require the interaction of different proteins from the corrinoid cofactor biosynthesis machinery with the prePceA apoprotein. One of these proteins might be the nicotinate-nucleotide dimethylbenzimidazole phosphoribosyltransferase (CobT), which catalyzes the activation of the lower ligand base in the late steps of corrinoid biosynthesis (3). It is not yet known, however, if CobT is involved in corrinoid cofactor incorporation into corrinoid-containing proteins.

From the data presented here and in previous publications (20, 22, 24) a model for the maturation of PceA may be derived (Fig. 6). According to this model, the pce genes are transcribed in PCE-grown cells and the PceA cofactor-free precursor is formed and binds to the PceT chaperone (22). When corrinoid cofactor is provided by de novo biosynthesis, it is incorporated into prePceA. After incorporation of this cofactor and assembly and incorporation of the iron-sulfur clusters, the precursor protein is correctly folded and exported to the exoplasm by the Tat machinery. After cleavage of the signal peptide, the protein is bound to PceB, which has been suggested to serve as a membrane anchor for PceA (10, 24, 35). In cells subcultivated for a few steps with fumarate instead of PCE, corrinoid biosynthesis is impeded; hence, this cofactor is not available for incorporation into prePceA. This causes aggregation of the prePceA together with PceT and other proteins inside the cells. Excision of the pce gene cluster occurs upon long-term cultivation with fumarate. The loss of the gene cluster is delayed in the presence of exogenous vitamin B₁₂. To verify this model, further experiments are required, which are under way in our laboratory.

Fig 6.

Tentative model of PceA maturation depending on vitamin B₁₂ and on the electron acceptor. For a detailed explanation, see the text.