Prof. Christof Holliger, Dr. Julien Maillard, Laure Prat, Aurélie Duret, Noam Shani
Swiss National Science Foundation
May 2005 – April 2008
Prof. W. Hagen, TU Delft, Pays-Bas ; Prof. J. Fontecilla, IS/LCCP,Grenoble, France, Prof. B. Kräutler, University of Innsbruck, Austria
The goal of this project was three-fold. First, highly enriched cultures obtained in our laboratory completely dechlorinating chloroethene to ethene were intended to be characterized in detail concerning the bacteria and reductive dehalogenases involved. Second, the dehalorespiration chain of Desulfitobacterium hafniense strain TCE1 was investigated, and third, the biochemical properties of the PCE reductive dehalogenase of D. hafniense strain TCE1 was determined.
In order to identify additional proteins involved specifically in dehalorespiration, 2D proteomic analysis was applied to membrane proteins of strain TCE1. This work was done in collaboration with the Laboratory of Molecular Ecology (LEM) at the University of Pau, France. Protein identification was performed by mass spectroscopy at the Protein Core Facilities (PCF) at the EPFL, Lausanne.
Membrane protein extracts were prepared from cells growing anaerobically on two alternative terminal electron acceptors, fumarate or PCE. The comparison of membrane protein expression in both conditions should point out proteins that are induced by PCE.
The 2D-analysis of PCE-grown culture revealed 310 spots, among which 160 were also found in fumarate-grown culture. According to the quantification, 72 spots showed increased intensity (>4-fold) when grown on PCE. Among the PCE-induced proteins, PceA was clearly identified. PceB which might play the membrane anchor for PceA, is too small (10kDa) to be seen in the type of gel used here (>20 kDa). IscS, a protein involved in Fe/S cluster formation, was also induced by PCE, which might be explained by the need of a larger pool of Fe/S clusters for the biosynthesis of PceA. An additional and unexpected protein was also identified: a rhodanese-like protein, to which no function could be associated yet. Reverse Transcription-PCR is being performed at the moment in order to confirm the proteomic data.
According to the analysis of D. hafniense strain Y51 genome sequence, the rhodanese gene is located in an operon presenting significant homology with the DMSO reductase family. Six genes were identified and named tsrF, tsrA, tsrB, tsrC, tsrD and tsrE due to similarities to the corresponding subunits A, B, C, and D of the membrane-associated thiosulfate reductase (Tsr). Both tsrF and tsrE gene products show clear homology to rhodanese, tsrE encoding for the rhodanese observed on PCE-grown culture during the 2D analysis. Reverse-transcription PCR experiments demonstrated that all genes are co-transcribed, giving a first indication that TsrE and the thiosulfate reductase activity might be linked. However their interplay remains to be established on protein level.
TsrE was investigated on biochemical level in order to understand its actual role in the dehalorespiration process. Rhodanese-like proteins are widely distributed in all kingdoms of life. Rhodanese is the common name for proteins with sulfur transferase activity, and harbors a conserved active site signature (CRXGX[R/T]). The variability around this signature occurs at different levels including the active site loop as well as the protein domains arrangement. Sequence analysis of TsrE revealed a predicted 32 kDa protein composed by two rhodanese domains, one N- and one C-terminal domain (with homology to the Prosite rhodanese motif PS50206), both of which containing a clear cysteine at the active site. Additionally TsrE possesses an N-terminal lipoprotein signal peptide with a clear lipobox signature. While the rhodanese domain is very common in proteins (about 17’000 entries in sequence database), only 7 proteins show this same architecture. Unfortunately no specific function has been identified for any of them.
Recombinant TsrE was heterologously produced in E. coli, purified and subjected to preliminary kinetic analysis. The standard rhodanese assay was performed based on the sulfur transfer between thiosulfate and potassium cyanide. Different TsrE variants were constructed by replacing the active site cysteine by serine: C140S, C259S, and the double variant C140S/C259S. While the double variant was totally inactive as expected, the two active sites showed unequal contributions to the sulfur transferase activity. Indeed, the N-terminal domain of TsrE seems to be accessory in that it shows a lower affinity for thiosulfate, and the C140S mutation has no effect on TsrE activity. The overall low affinity for thiosulfate suggests that another substrate might be physiologically relevant.
TsrE is clearly associated with the cytoplasmic membrane probably by its lipid moiety. The exact topology remains however to be elucidated. In order to define the accurate localization of TsrE, a protocol based on mutanolysin was adapted for D. hafniense strain TCE1 to form protoplasts. These cell-wall free cells were then treated with proteinase K. If TsrE is facing the outside of the cytoplasmic membrane, as most lipoproteins do, it should be degraded by proteinase K. Curiously TsrE seems to be only partially degraded as a band of 25 kDa is still detected. Sequence analysis of TsrE, and the 3D structure of homologous proteins, did not suggest any transmembrane α-helix which could support the idea of a complex topology with domains on both sides of the membrane. One possible explanation could be that part of TsrE is resistant to proteinase K. Further investigations with alternative proteases and a denaturation step prior to protease treatment should give an unequivocal answer about TsrE topology.
The specific function of TsrE in dehalorespiration is at present unclear. Several roles for rhodanese have been already suggested, in particular in Fe/S cluster formation and in sulfur metabolism. To obtain information on the accurate role of TsrE, identification of partners was considered. Different protein interactions tests such as bacterial two-hybrid experiments, cross-link or immunoprecipitation were attempted, but no clear interaction was observed. Moreover, no specific in vivo substrate of TsrE was so far identified.
It seems that TsrE could act as sulfur carrier to deliver sulfur for needy proteins. It could mobilize sulfur and maintain it in an activated form ready to be used for the cell.
It is well established that PceA is associated with the cytoplasmic membrane. A previously applied approach using membrane-impermeable artificial electron donors suggested that PceA faces the cytoplasm. The presence of the N-terminal Tat signal peptide indicates, however, a post-translational translocation to the outside of the cytoplasmic membrane. Moreover, a recent study has clearly established that PceA of Sulfurospirillum multivorans is translocated across the membrane, and that this process is growth-substrate dependent.
Localization of PceA from D. hafniense strain TCE1 followed the following strategy. PceA was over-expressed heterologously in E. coli without the Tat signal peptide because of its high hydrophobicity that might have been limiting both in expression yield and localization. The biomass was harvested and processed for purification. Despite the lack of the signal peptide, large inclusion bodies were observed indicating a poor solubility for PceA, and that required modification of the purification protocol. Finally PceA was partially purified under denaturing condition with high concentrated urea. Polyclonal anti-PceA antibodies could be raised and used in immunoblot detection. However, preliminary western blot analysis showed strong background probably due to the purification mode. Clear improvements were obtained after purifying the antibodies on PceA-bound membranes.
Localization of PceA was analyzed from protoplasts of TCE1 cells treated with mutanolysin (see sub-project B). The results obtained here undoubtedly showed the external localization of PceA, thus confirming data obtained with S. multivorans.
From various studies on other Tat-dependent proteins (TMAO reductase, DMSO reductase) it has been shown that the signal peptide takes part in at least two processes. First, it is recognised by specific chaperones involved in folding and assembly of the Tat-dependent protein, and second, it interacts with the Tat machinery itself. There is clear evidence that the conserved twin-arginine residues are essential for the second process. The interaction with chaperones on the other side requires a much bigger part of the signal peptide, most probably the hydrophobic region.
No such specific chaperone is known for corrinoid-dependent oxidoreductases. However, the size and complexity of the PCE reductive dehalogenase suggests that helper proteins are involved in the assembly and translocation of the mature enzyme across the cytoplasmic membrane.
Tn-Dha1, the transposon containing the pce genes, has been identified in genomic DNA of strain TCE1, and consists of two copies of an insertion sequence (ISDha1), flanking six other ORFs. Besides the pceABCT gene cluster, a putative tatA (or truncated tatB, Tn-tatA/B) gene is also present in Tn-Dha1, and possibly involved in the specific translocation of PceA. Moreover three additional tatA genes and a tatBC operon are located on the genomes of desulfitobacteria.
The transcription of Desulfitobacterium tat genes was investigated by RT-PCR in strain TCE1 growing on fumarate or PCE as alternative electron acceptors, clearly indicating that one of the genomic tatA gene (tatA1, corresponding to DSY0284 locus in D. hafniense strain Y51 genome) is mostly used, while tatA2 (DSY1444) is not transcribed, and tatA3 (DSY2216) only slightly, all of them independently of the electron acceptor used. The tat gene located on the transposon (Tn-tatA/B) showed an increased transcription rate following a pulse of PCE in a fumarate-growing culture, suggesting that the Tn-based tat gene is positively regulated by PCE, and might be specifically involved in the Tat translocation of the PCE reductive dehalogenase.
Preliminary experiments were also done to assess the functionality of tatA and –B genes using a heterologous system in E. coli tat mutant stains. The tatA, and –B genes of strain TCE1 were (or currently are being) cloned into a plasmid dedicated to the constitutive low expression of Tat proteins. So far, while testing the experimental set-up, TatA3 was shown to partially complement an E. coli tatAtatE mutant strain, as it suppresses the sensitivity to SDS, a typical phenotype of tat mutants. Indeed amidases conferring the SDS resistance, are translocated into the periplasmic space in E. coli by the Tat system. A full description of Desulfitobacterium Tat proteins functionality is under progress.
The PCE reductive dehalogenase (PceA) belongs to a family of corrinoid-dependent enzymes. The electron transfer towards PCE is mediated by the corrinoid cofactor. Therefore it is of great interest to investigate which type of corrinoid cofactor is inserted in PceA. This question is also motivated by the discovery of a novel corrinoid cofactor in the reductive dehalogenase of Sulfurospirillum multivorans.
Investigations whether D. hafniense strain TCE1 is able to synthesize its own corrinoid was performed by cultivation of strain TCE1 on medium depleted in vitamin B12. Strain TCE1 was indeed still growing when vitamin B12 was depleted showing that strain TCE1 is able to synthesize its own corrinoid.
Genes that play a role in corrinoid biosynthesis have been identified in the genome of D. hafniense strain Y51. The presence of a selection of five genes among them (cbiX, cbiJ, cbiT, cbiC, cobs) on the genome of D. hafniense strain TCE1 was confirmed by PCR with specific primers for these genes. A qualitative gene expression study was done with strain TCE1 under various growth conditions where specific RNA transcripts were analyzed by an optimized RT-PCR step using a combination of random hexamers and gene-specific primers. Under standard culture conditions with yeast extract and vitamin B12, none of the 5 genes studied was significantly transcribed. When yeast extract was replaced by amino acids, all genes were transcribed in the absence of vitamin B12, whereas in its presence, only weak transcription was observed (10-100 times less). Replacing PCE by fumarate as electron acceptor showed similar results, however, transcription seemed to be overall. Further investigations with various precursors of vitamin B12 added to the growth medium will be performed in order to understand the corrinois biosynthesis pathway (collaborative work with Prof. Kraütler, University of Innsbruck, Austria). The effect of the precursors on the activity of the reductive dehalogenase will be also measured.
Recently, an alternative gene regulatory pathway provided by RNA elements has been discovered, so-called “riboswitches”. Metabolite-induced restructuring of a riboswitch leads to premature termination of transcription and/or inhibition of translation of the downstream genes, which have a close physiological relationship with the bound metabolite. A recent example of a B12 riboswitch has shown that the 5’UTR region of the btuB gene of E. coli interacts specifically with coenzyme B12 and its derivatives thereby leading to changes in the RNA structure and hence to an altered expression of the downstream btuB gene. Bioinformatic analysis of D. hafniense strain Y51 genome clearly revealed riboswitches upstreams of several B12-related genes suggesting that riboswitches might be of great importance for the regulation of the B12 biosynthesis pathway in desulfitobacteria. New lines of investigation will be developed here in order to characterize the actual activity of B12 riboswitches using E. coli and dedicated reporter genes to assess for their activity towards vitamin B12 and its precursors.