OUP user menu

The Desulfitobacterium genus

Richard Villemur, Martin Lanthier, Réjean Beaudet, François Lépine
DOI: http://dx.doi.org/10.1111/j.1574-6976.2006.00029.x 706-733 First published online: 1 September 2006


Desulfitobacterium spp. are strictly anaerobic bacteria that were first isolated from environments contaminated by halogenated organic compounds. They are very versatile microorganisms that can use a wide variety of electron acceptors, such as nitrate, sulfite, metals, humic acids, and man-made or naturally occurring halogenated organic compounds. Most of the Desulfitobacterium strains can dehalogenate halogenated organic compounds by mechanisms of reductive dehalogenation, although the substrate spectrum of halogenated organic compounds varies substantially from one strain to another, even with strains belonging to the same species. A number of reductive dehalogenases and their corresponding gene loci have been isolated from these strains. Some of these loci are flanked by transposition sequences, suggesting that they can be transmitted by horizontal transfer via a catabolic transposon. Desulfitobacterium spp. can use H2 as electron donor below the threshold concentration that would allow sulfate reduction and methanogenesis. Furthermore, there is some evidence that syntrophic relationships occur between Desulfitobacterium spp. and sulfate-reducing bacteria, from which the Desulfitobacterium cells acquire their electrons by interspecies hydrogen transfer, and it is believed that this relationship also occurs in a methanogenic consortium. Because of their versatility, desulfitobacteria can be excellent candidates for the development of anaerobic bioremediation processes. The release of the complete genome of Desulfitobacterium hafniense strain Y51 and information from the partial genome sequence of D. hafniense strain DCB-2 will certainly help in predicting how desulfitobacteria interact with their environments and other microorganisms, and the mechanisms of actions related to reductive dehalogenation.

  • Desulfitobacterium
  • reductive dehalogenation
  • dehalogenase
  • bioremediation


Desulfitobacterium spp. are strictly anaerobic bacteria belonging to the Firmicutes, Clostridia, Clostridiales and Peptococcaceae. They were first isolated from environments contaminated by halogenated organic compounds such as soil, wastewater sludges and freshwater sediments while investigators were trying to obtain isolates that could degrade highly chlorinated organic compounds. Several of these compounds are highly resistant to aerobic biodegradation and often toxic to bacteria. Most of the Desulfitobacterium strains can dehalogenate halogenated organic compounds by mechanisms of reductive dehalogenation. They are phylogenetically distinct from other reductive dehalogenating bacteria (Fig. 1) such as Sulfurospirillum halorespirans, Sulfurospirillum multivorans and Desulfomonile tiedjei, which belong to the proteobacteria family, and Dehalococcoides ethenogenes, which is deeply branched in the bacterial kingdom. Only Dehalobacter restrictus is closely related to the Desulfitobacterium genus. The first anaerobic dehalogenating bacteria isolated was Desulfomonile tiedjei strain DCB-1 (Shelton & Tiedje, 1984; DeWeerd, 1990), and the first Desulfitobacterium strain to be isolated was strain DCB-2 (Madsen & Licht, 1992). It was in 1994, with the isolation of Desulfitobacterium dehalogenans, that the Desulfitobacterium genus was created (Utkin, 1994). In 1996, strain DCB-2 was classified as Desulfitobacterium hafniense (Christiansen & Ahring, 1996a). It was also in this same year that Desulfitobacterium chlororespirans strain Co23 (Sanford, 1996), Desulfitobacterium frappieri strain PCP-1 (Bouchard, 1996) and Desulfitobacterium sp. strain PCE1 (Gerritse, 1996) were reported. In subsequent years, further Desulfitobacterium strains were either isolated or detected in bioprocesses or in various environments. Since 2001, a significant portion of the genome of D. hafniense strain DCB-2 has been released, and, more recently, the whole genome of D. hafniense strain Y51 has been published (Nonaka, 2006). These discoveries have enabled the uncovering of some aspects of the functionality of this genus.


Phylogenetic tree of Desulfitobacterium spp. based on 16S rRNA gene sequences. 16S rRNA sequences available from Desulfitobacterium species and strains were aligned with sequences of representative species of different genera belonging to Peptococcaceae, and to Firmicutes (Clostridium, Bacillus and Lactobacillus). GenBank accession numbers are indicated beside the bacterial name. Sequences were aligned by the clustalw 1.82 program (Chenna, 2003). The phylip package (version 3.5; http://evolution.gs.washington.edu/phylip.html) was used for phylogenetic analyses with a bootstrap analysis of 1000 replicates. The trees were inferred from a matrix of pairwise distances by using 1266 aligned positions (dnadist). The 5′ end of the genes that includes the 100−200 nt insertion was excluded. The neighbor-joining program was used to derive the best phylogenetic tree for each replicate. Lastly, the consense program was used to derive the consensus tree. The numbers at the forks indicate the percentage of times a species or a group of species originated at each fork. Bar: 0.02 nucleotide substitutions per site.

Various reviews have been published regarding bacterial dehalogenation and the molecular mechanisms of dehalogenation (El Fantroussi, 1998; Fetzner, 1998; Holliger, 1998; Smidt, 2000a; Häggblom & Bossert, 2003; Smidt & de Vos, 2004). This review covers findings regarding the physiology, genetics and enzymology of members of the genus Desulfitobacterium, especially regarding the reductive dehalogenation of halogenated organic compounds. A section is dedicated to potential applications of these bacteria in bioremediation processes aimed at treating environments polluted with halogenated organic compounds.

Phylogeny and heterogeneity of 16S rRNA genes

The phylogenetic relationship based on 16S rRNA gene sequences of members of the genus Desulfitobacterium is illustrated in Fig. 1. The genus is composed of at least 5 species: Desulfitobacterium metallireducens strain 853–15AT, Desulfitobacterium dichloroeliminans strain DCA1T, Desulfitobacterium dehalogenans strain JW/IU-DCIT, D. chlororespirans strain Co23T and D. hafniense DCB-2T. Strain PCP-1 was first described as D. frappieri but further taxonomy analysis showed that D. frappieri belongs to one species, D. hafniense. In fact, if the insertional sequences in the 16S rRNA genes (see below) are excluded, strains PCP-1 and DCB-2 and seven other strains (strains DP7, TCE1, Y51, GBFH, TCP-A, G2, PCE-S) are very closely related (>99%), suggesting that they belong to the same species. Indeed, DNA—DNA reassociation experiments between the genomes of strains PCP-1, DCB-2, TCE1, DP7 and GBFH showed more than 80% homology (Table 1), confirming that they belong to the same species (Niggemyer, 2001; van de Pas, 2001b). By convention, if two strains have more than 70% homology between their genome, based on DNA−DNA reassociation, they belong to the same species. As strain DCB-2 was published a few months before strain PCP-1, this latter and all other strains named after D. frappieri (strains TCE1, DP7, TCP-A and G2) have to be renamed as D. hafniense with DCB-2 as type strain (Niggemyer, 2001). In this regard, results from genome reassociation experiments are similar to the phylogeny tree of 16S rRNA gene sequences, where the D. chlororespirans genome is more homologous to the D. hafniense genomes than to the D. dehalogenans genome, with 64–71% genome reassociation, which is near the limit of species demarcation.

View this table:

Percentage of DNA—DNA reassociation between genomes of various Desulfitobacterium strains

The confusion in naming strains PCP-1, DP7, TCP-A, G2 and TCE1 as frappieri species instead of hafniense comes from the comparison of their 16S rRNA gene sequences. Originally, the 16S rRNA gene sequence of strain DCB-2 was reported as a 1530-nt sequence, in contrast to strain PCP-1, which has a 1655-nt sequence. A 128-nt insertion near the 5′ end of the gene was found in the 16S rRNA gene of strain PCP-1 but not in the 16S rRNA gene of strain DCB-2 (Bouchard, 1996). Such an insertion is also present in the 16S rRNA gene of strains DP7, TCE1, and TCP-A. The Joint Genomic Institute (http://www.jgi.doe.gov/) released part of the D. hafniense DCB-2 genome, which shows sequence heterogeneity in its 16S rRNA gene copies. Among the six copies of the 16S rRNA gene present in the strain DCB-2 partial genome, an insertion is present in at least two copies (homologous, but not identical), as observed in the other strains. Six 16S rRNA gene copies are also present in the genome of D. hafniense strain Y51 (Nonaka, 2006), and each of them contains an insertion. This form of macroheterogenity was also found in different copies of the 16S rRNA gene of other Desulfitobacterium strains such as D. hafniense strains TCE1, TCP-A, and DP7, D. dehalogenans, D. chlororespirans and Desulfitobacterium sp. strain PCE-1. However, reverse transcriptase polymerase chain reaction (RT-PCR) experiments on total RNA showed that these insertions are probably lost during the rRNA maturation for most of these strains (Villemur, unpublished results). While secondary structure prediction indicated that these insertions should form a loop (Bouchard, 1996), the function of this loop is as yet unknown.

Physiology and dehalogenation spectrum

General physiological characteristics

Desulfitobacterium spp. are slightly curved rods, generally motile and with a cell size varying from 2 to 7 μm (Table 2). When present, 2–6 lateral flagella are observed. Interestingly, sporulation is not a common feature for all strains. Seven strains were reported to form terminal endospores, but such spores were not observed in seven other strains. Strain PCE1 and D. metallireducens do not sporulate but are heat-sensitive, in contrast to D. dehalogenans, which is heat-resistant but has not been reported to sporulate. The optimal temperature and pH for growth of Desulfitobacterium spp. range from 25 to 38°C and from pH 6.5 to 7.8. Their G+C content varies from 45 to 49 mol% (Table 2). They generally stain Gram-positive, although some strains such as strains DCB-2, PCP-1, Y51, and D. cholorespirans stain Gram-negative (Bouchard, 1996; Christiansen & Ahring, 1996a; Sanford, 1996; Suyama, 2001). Phylogeny studies classified Desulfitobacterium spp. in the Firmicutes, which are Gram-positive bacteria (Fig. 1).

View this table:

Physiological characteristics of members of the Desulfitobacterium genus

Terminal electron acceptors

The free-energy value (G°′) or the redox potential (Embedded Image) can give an estimate of the growing conditions that should prevail in a particular environment upon the presence of certain molecules. For example, the redox potential of O2/H2O (Embedded Image) is higher than Embedded Image, (Embedded Image), Embedded Image (Embedded Image) and HCO3/CH4, (Embedded Image) (Table 3). Aerobic microorganisms obviously prevail in aerobic environments. Upon exhaustion of O2, and in the presence of nitrate, sulfate and carbonate, denitrifying bacteria should prevail against sulfate-reducing bacteria and methanogens. However, other elements such as iron(III) and manganese(VI) or organic compounds such as humic acids can be abundant in anaerobic environments and their reduction can be used by microorganisms as terminal electron acceptors. In fact, any couple of molecules that can provide a sufficiently high redox potential can theoretically be used by a microorganism, if this microorganism has developed an enzymatic system to channel the electrons given by these molecules. From this perspective, desulfitobacteria seem to be very versatile bacteria, as they can use a wide range of electron acceptors, such as nitrate, various metals, sulfur derivatives, humic acids and halogenated organic compounds (Table 4).

View this table:

Half-reaction redox potentials of chlorinated compound couples involved in chloridogenesis

View this table:

Some of electron donors and electron acceptors used by members of the Desulfitobacterium genus

Halorespiration, dehalorespiration or chloridogenesis

The terms halorespiration and dehalorespiration are used in the literature for describing reductive dehalogenation of halogenated organic compounds that are linked to energy conservation via electron-transport-coupled phosphorylation. Those who use the term halorespiration suggest that halogenated organic compounds serve as terminal electron acceptors as in other respiratory processes such as nitrate or iron respiration (Smidt, 2000a). Those who use the term dehalorespiration claim that such a respiratory process has not been fully demonstrated and the halogenated organic compounds may only serve as, to quote Holliger (1998), a “necessary electron sink for reducing equivalents generated upon oxidation of the electron donor”. Löffler (2003) suggested using instead the term chloridogenesis, as, to quote these authors, the “term respiration implies a change of redox state, however, the chloride ion that is liberated in (de)chlororespiration is neither oxidized nor reduced, and therefore both terms are misleading”. In this review, we used the term chloridogenesis.

One type of chloridogenesis is hydrogenolytic reductive dehalogenation (Table 3), which catalyzes the replacement of one halogen atom by a hydrogen atom (Holliger, 1998; Smidt & de Vos, 2004). The estimation of the redox potentials of halogenated organic compound couples (R-Cl/R-H) involved in hydrogenolytic reductive dehalogenation ranges between +260 and +660 mV (Vogel, 1987; Dolfing & Harrison, 1992; De Wildeman & Verstraete, 2003; Dolfing, 2003) (Table 3). Hydrogenolytic reductive dehalogenation should prevail over sulfate reduction and is in the same range as nitrate reduction. A second type of chloridogenesis, dihaloelimination, involves halogenated aliphatic compounds in which two adjacent halogen atoms are simultaneously removed and the carbon−carbon bond is converted into a double bond (De Wildeman & Verstraete, 2003; Smidt & de Vos, 2004) (Table 3). Only one molecule of H2 is required for the removal of the two adjacent halogen atoms, as opposed to two molecules for hydrogenolysis. The redox potentials for halogenated aliphatic compound couples involved in dihaloelimination are much higher than those involved in hydrogenolysis−up to 1192 mV for hexachloroethane/tetrachloroethane (Vogel, 1987; De Wildeman & Verstraete, 2003; Dolfing, 2003), equivalent to the redox potential of O2/H2O.

Mackiewicz & Wiegel (1998) were among the first to demonstrate that reductive dehalogenation sustains growth of a Desulfitobacterium strain via electron-transport-coupled phosphorylation when formate was used as electron donor and 3-chloro-4-hydroxyphenylacetic acid (3Cl4OHPA) as electron acceptor in D. dehalogenans cultures. These same authors also showed that, in competition experiments where nitrate, 3Cl4OHPA, fumarate and sulfite were present in the same medium, D. dehalogenans reduces nitrate and 3Cl4OHPA simultaneously for growth, then fumarate and sulfite. Van de Pas (2001c) also showed electron-transport-coupled phosphorylation when D. dehalogenans was cultured in the presence of formate or H2 as electron donors and 3Cl4OHPA. Interestingly, these authors observed that, during pyruvate fermentation in a bicarbonate buffer, twice the amount of acetate was generated compared with lactate. It was shown that CO2 was used by D. dehalogenans as electron acceptor to produce acetate for generating ATP. Activity is though to be present in other strains because pyruvate is fermented by most Desulfitobacterium spp.

Substrate spectrum of dehalogenation

The ranges of molecules that can be dehalogenated by the various Desulfitobacterium strains cannot be compared, as each strain was isolated in a different laboratory and has not necessarily been tested against the same molecules. However, important differences have been observed between Desulfitobacterium strains, as summarized in Table 5. These differences are probably explained by the diversity of the reductive dehalogenases present in these strains (see below). This enzymatic diversity might reflect the variety of naturally occurring halogenated organic compounds found in nature.

View this table:

Substrate spectrum of dehalogenation of members of the Desulfitobacterium genus

The first strain of the Desulfitobacterium genus to be identified as such was D. dehalogenans strain JW/IU-DC1 (Utkin, 1994). Strain JW/IU-DC1 is able to dehalogenate several chlorophenolic compounds when the chlorine is at the ortho position relative to the hydroxyl group (Utkin, 1995). Wiegel (1999) observed dechlorination of para-hydroxylated mono- or polychlorinated biphenyls (PCB), which are mammalian PCB metabolites and exhibit both estrogenic and antiestrogenic effects (for references see Wiegel, 1999). Utkin (1995) also noticed that the presence of a chlorine atom at positions 3, 4 and 6 relative to the hydroxyl group of these phenolic compounds facilitated dehalogenation at position 2 (ortho). On the other hand, the presence of a chlorine atom at position 5 (meta) had a negative effect on the dehalogenation yield of these molecules, especially on dehalogenation at position 2. Desulfitobacterium sp. strain PCE1 can dehalogenate tetrachloroethene (PCE) mostly into trichloroethene (TCE), but small amounts of cis- and trans-dichloroethene (DCE) were observed. It can dehalogenate some chlorophenols (CPs) at the ortho position only. It is one of the few Desulfitobacterium strains that can dechlorinate monoCP (MCP) to phenol, and dehalogenate chlorinated aliphatic and aromatic compounds (Gerritse, 1996).

Desulfitobacterium hafniense strain DCB-2 can dehalogenate several chlorophenols at the ortho- and meta-positions (Madsen & Licht, 1992; Christiansen & Ahring, 1996a). Gerritse (1999) showed that strain DCB-2 dehalogenates PCE slowly (transformed to TCE), and only when it is precultured with pyruvate, or l-lactate and 3Cl4OHPA. Desulfitobacterium hafniense strain PCP-1 is, up to now, the only isolated anaerobic bacterium that can dehalogenate pentachlorophenol (PCP) to 3-CP. It can dehalogenate at the ortho-, meta- and para-positions a large variety of compounds containing a phenyl or pyridine ring substituted with hydroxyl, nitro, methoxy or amino groups. In the case of polychlorinated nitrobenzene, the nitro group is first reduced to an amino group before dehalogenation (Bouchard, 1996; Dennie, 1998). It was the first isolated bacterium reported to dechlorinate a hydroxylated PCB (3,5-dichloro-4-hydroxybiphenyl) under anaerobic conditions. However, it cannot dehalogenate 2-hydroxy-3,5-dichlorobiphenyl, 2-hydroxy-2′,5′-dichlorobiphenyl or 4-hydroxy-2,2′,5′-trichlorobiphenyl, suggesting that the presence and the position of the hydroxyl group are important. A polar substitutent (hydroxyl, amino) or the nitrogen of pyridine is therefore required for the dehalogenation to occur. It remains to be determined if this polar group acts by interacting with the active site of the enzyme or by favouring the dechlorination reaction through an electron donating effect. Dehalogenation of several of these compounds had to be first induced in strain PCP-1 cultures. Two inducers, 2,4,6-trichlorophenol (TCP) and 3,5-dichlorophenol (DCP), have so far been identified−the former inducing ortho-dechlorination and the second inducing meta- and para-dechlorination. For instance, induction with 2,4,6-TCP and 3,5-DCP is required for dehalogenation of polychlorinated veratrole and anisole, but not for guaiacol. Desulfitobacterium hafniense strain DP7 is the only Desulfitobacterium strain not able to dehalogenate chlorophenols or tetrachloroethene. Its origin may explain this lack of dehalogenation activity: it was isolated from healthy human fecal samples, an environment not known to contain halogenated organic compounds. Incubations for 4 months at 37°C with 1 mM of MCPs, 2,3-DCP, 2,4-DCP, 2,5-DCP, 2,6-DCP, PCP, 3Cl4OHPA, and TCE resulted in no dehalogenation (van de Pas, 2001b). Experiments performed in our laboratories with some of these compounds but at lower concentrations confirmed that strain DP7 does not dehalogenate these chlorophenols, but also showed that they do not inhibit its growth (Gauthier, 2006).

Desulfitobacterium hafniense strain TCE1 can dehalogenate PCE and TCE into cis-DCE, but small amounts of 1,1-DCE were also detected. This bacterium is also capable of dehalogenating low concentrations of CCl4 into CHCl3 and CH2Cl2, when first grown with PCE. Gerritse (1999) suggested that this was a cometabolic process, since no dehalogenation was observed when strain TCE1 was given CCl4 without PCE. The PCE dechlorination is an inducible process, as strain TCE1 cells that were grown with fumarate, sulfite or nitrate as electron acceptors dechlorinated PCE at less than 1% of the rate of cells that were grown in the presence of PCE. However, strain TCE1 cells pregrown fermentatively on pyruvate also showed substantial dechlorinating activity, suggesting that PCE is not required for the induction (Gerritse, 1999). Desulfitobacterium hafniense strain Y51 can dehalogenate PCE, and TCE into cis-1,2-DCE. PCE dehalogenation can occur up to 960 μM, which is its maximum water solubility. Strain Y51 can also dehalogenate halogenated ethanes into DCE and 1,1,1,2,2,3,3-heptachloropropane into pentachloropropene, presumably by a dihaloelimination mechanism (Suyama, 2001). Desulfitobacterium hafniense strain TCP-A can partially dechlorinate several CPs at ortho- and meta-positions (Breitenstein, 2001). It dehalogenates 2-MCP slowly into phenol. 2,4,6-TCP acts as an inducer for the dehalogenation of 2,3,4,5-tetrachlorophenol into 3,4,5-TCP and 2,3-DCP into 3-CP.

Desulfitobacterium chlororespirans strain Co23 can dehalogenate at the ortho-position some halogenated aromatic compounds, among them the herbicide Bromoxynil (3,5-dibromo-4-hydroxybenzonitrile) and its metabolite 3,5-dibromo-4-hydroxybenzoic acid, and Ioxynil (3,5-diiodo-4-hydroxybenzonitrile) (Sanford, 1996; Cupples, 2005). However, Ioxynil was not deiodinated when added alone or when added with Bromoxynil. However, Ioxynil dehalogenation was observed when the culture was amended with 3-chloro-4-hydroxybenzoate, suggesting that this system is inducible. Desulfitobacterium dichloroeliminans strain DCA1 can completely dehalogenate several dichloroalkanes into their corresponding alkenes. This is the first bacterial species to show the use of the dichloroelimination mechanism to gain energy. Vinyl chloride and monochloroethane are not produced by strain DCA1 as free intermediates of dehalogenation. Analysis of the dehalogenation products of the stereoisomers of 2,3-dichlorobutane suggests a stereoselective anti dichloroelimination (i.e. with the departing chlorine atoms opposite to each other) (De Wildeman, 2003).


Desulfitobacterium hafniense strains DCB-2 and PCP-1, D. chlororespirans, and D. dehalogenans were showed to O-demethylate and completely dechlorinate chlorinated hydroquinone metabolites, which are naturally occurring molecules synthesized by basidiomycete fungi. No such activity was present in D. hafniense strains DP7 and TCE1 (Milliken, 2004b). However, Desulfitobacterium strain PCE1 transforms 2,3,5,6-tetrachloro-4-methoxyphenol into 2,3,5-TCP through O-demethylation, dechlorination and dehydroxylation (Milliken, 2004a). These are antimicrobial compounds that can serve as substrates for lignin peroxidases in the degradation of lignin by these fungi.

In the presence of fumarate as electron acceptor, D. hafniense strain DCB-2 used a variety of phenyl ethers, which are known to be components of lignin, as electron donors. O-demethylation was responsible for transforming phenyl ethers into phenolic compounds. When phenyl ethers were present as electron donors, dehalogenation of 3Cl4OHPA was observed (Neumann, 2004). Like strain DCB-2, Desulfitobacterium strain PCE-S uses a variety of phenyl ethers as electron donors. Energy is provided by the O-demethylation of phenyl ethers. However, the presence of PCE as electron acceptor and phenyl ethers as electron donors did not sustain the growth of strain PCE-S, suggesting inhibition of O-demethylase. It was hypothesized that PCE could react with the corrinoid cofactors putatively involved in O-demethylation (Neumann, 2004). Finally, O-demethylation was observed with D. hafniense strain PCP-1 with tetrachloroguaiacol, tetrachloroveratrol and pentachloroanisol (Dennie, 1998).

Whole genome sequencing projects revealed that 15 similar O-demethylation operons were found in the strain DCB-2 genome, and 15 vanillate-specific O-demethylase corrinoid protein (odmA) homologs and two genes similar to vanillate:corrinoid protein methyltransferase gene (odmB) were found in the D. hafniense strain Y51 genome (Nonaka, 2006). O-demethylation activity is probably widespread in the genus, and the presence of multiple paralogs in strains Y51 and DCB-2 suggests that they are important for Desulfitobacterium biology, as stated by Nonaka (2006).

Desulfitobacteria in the environment

Most Desulfitobacterium strains were isolated from sites contaminated with anthropogenic halogenated organic compounds in order to study their capacity to dehalogenate these molecules. However, the source of halogenated molecules is not only anthropogenic, because natural halogenated molecules are synthesized by a variety of organisms or have a geogenic origin (Bossert, 2003). These natural molecules can be toxic, have an antimicrobial activity, and probably provide an advantage to the producing organisms for the invasion of an ecological niche. In this regard, desulfitobacteria have probably developed enzymatic systems to use these toxic molecules as electron acceptors, allowing them to survive and colonize a niche. Furthermore, their presence allows the niche to be “detoxified” and may permit subsequent colonization by other microorganisms.


Desulfitobacterium strains were isolated in North America (D. dehalogenans, D. chlororespirans, and D. metallireducens, and D. hafniense strains G2, PCP-1 and GBFH) and in Europe (D. hafniense strains DCB-2, DP7, TCP-A, TCE1 and PCE-S, D. dichloroeliminans, Desulfitobacterium sp. strain PCE1 and Desulfitobacterium sp. strain RPf35Ei). Desulfitobacterium hafniense strain Y51 and Desulfitobacterium sp. strain KBC1 were isolated in Japan. Except for strain DP7, these strains were isolated from cultures enriched from soil, wastewater sludges and/or freshwater sediments. The variety of places from which these strains were isolated suggests that they are ubiquitous in the environment.

Because monitoring strict anaerobic bacteria in heterogeneous environments is very difficult with culture methods, molecular tools have been developed. The most abundant bacterial species in complex microbial communities can be determined by deriving 16S rRNA gene libraries. Total DNA is extracted from an environmental sample and 16S rRNA gene sequences are PCR-amplified and cloned to generate a library representative of most 16S rRNA gene sequences present in the sample. For example, Davis (2002) used 16S rRNA gene libraries to characterize a microbial community in subsurface aquifer at a site contaminated by chlorinated solvents such as PCE. They detected phylotypes closely related to Desulfitobacterium/Dehalobacter spp. From another 16S rRNA gene library, Lecouturier (2003) found a high proportion of clones related to Desulfitobacterium spp. in an anaerobic 5-amino-2,4,6-triiodoisophthalic acid-deiodonating fixed-bed reactor.

Another approach is to use PCR with primers targeting specific Desulfitobacterium DNA sequences. Lévesque (1997, 1998) used nested PCR and competitive PCR with primers targeting specific 16S rRNA gene sequences to detect and quantify D. hafniense strain PCP-1 in soil. When strain PCP-1 was added to soil samples, the PCR detection limit was 800 PCP-1 cells/g of dry soil. Lanthier (2001) developed other primers targeting most Desulfitobacterium species, and examined soil samples collected around the St Lawrence River and estuary in Quebec (Canada). Initially, no PCR signal specific to desulfitobacteria was found when DNA was extracted directly from the soil samples, suggesting that they were either absent, or below the limit threshold of the method, or in their sporulated form and difficult to break. However, incubation of these soils under reductive conditions and with 2,4,6-TCP showed that these bacteria were present in about half of the soil samples tested. 2,4,6-TCP probably stimulated Desulfitobacterium growth and may have been toxic to other competing microorganisms. Some of these soils contained a variety of xenobiotics, such as polycyclic aromatic hydrocarbons, PCBs, and chlorinated phenols. There was, however, no correlation between the presence of Desulfitobacterium-specific PCR signals and the presence of pollutants in soils. Smits (2004) have also developed a quantitative PCR method (real-time PCR) to determine the concentration of desulfitobacteria using 16S rRNA gene sequences. They were able to determine the abundance of D. hafniense strain TCE1 in a mixed culture with Dehalobacter restrictus.

Metals and humic acids

There are several elements in anaerobic environments, such as metals and humic acids, that could act as electron acceptors and support the growth of desulfitobacteria (Table 4). This might explain their recovery from organic-rich environments in which the presence of chlorinated pollutants or sulfite is not expected.

Niggemyer (2001) isolated D. hafniense strain GBFH from an enrichment culture inoculated with sediments from the Coeur d'Alene river delta, which is known to be rich in iron. Strain GBFH reduced As(V) in As(III) in the presence of formate as electron donor and carbon source. It also reduced Fe(III), sulfur, Mn(IV), and Se(VI). With the exception of sulfur (not tested), these metals were also reduced by D. hafniense strains PCP-1, DCB-2 and TCE1. As(V) was not reduced by D. hafniense strain DP7. Fe(III) reduction occurred in initial incubations of strain DP7 but the reducing activity was lost in sequential transfers (Luijten, 2004). Niggemyer (2001) also showed that both D. dehalogenans and D. chlororespirans grew by reducing Fe(III) pyrophosphate, Mn(IV), and Se(VI) but not As(V). However, Luijten (2004) did not observe reduction of Fe(III) citrate and Mn(IV). The authors explained these results by the loss of genes encoding specific processes during adaptation of the bacteria to their growth conditions or the use of citrate as the iron chelating agent instead of pyrophosphate.

Desulfitobacterium hafniense strain G2 can reduce metals such as Fe(III) and U(VI), and use anthraquinone-2,6-disulfonate (AQDS, a humic acid analog) as electron acceptors (Shelobolina, 2003a). In contrast to other Desulfitobacterium strains, which are capable of reducing chelated Fe(III), strain G2 is the first Desulfitobacterium strain shown to reduce insoluble Fe(III) ions such as in poorly crystalline ferric iron oxides and Fe(III) bound to phyllosilicate minerals. These two forms are the main source of Fe(III) in sedimentary environments and the most abundant potential electron acceptors in anaerobic environments. Strain G2 can also oxidize Fe(II) (both chelated and insoluble) in Fe(III) in the presence of nitrate, suggesting a role of strain G2 in iron cycling. Desulfitobacterium metallireducens strain 853-15A was isolated from an enrichment performed with lactate and AQDS as electron acceptor (Finneran, 2002). Some of the electron acceptors used by strain 853-15A are chelated Fe(III) [not crystalline Fe(III)oxide], humic acids, Mn(IV), colloidal sulfur, Se(IV) and Cr(VI).

Indirect evidence that desulfitobacteria are involved in metal-reducing activities has been observed in various systems. Desulfitobacterium sp. strain RPf35Ei was isolated from a sulfate-reducing ethanol-fed fluidized-bed reactor originally inoculated with sludge from a sulfate-reducing upflow anaerobic sludge blanket reactor, and with enrichment cultures from a methanogenic granular sludge and sediments from a mine in Finland (Kaksonen, 2004). No data are available on whether or not strain RPf35Ei can reduce metal, but the fluidized-bed reactor from which it was isolated was used to treat acidic metal-containing wastewater. Shelobolina (2003b) derived an U(VI)-reducing culture from uranium-contaminated sediment. Upon further enrichment with U(VI) and lactate, a 16S rRNA gene sequence related to D. chlororespirans was detected. Petrie (2003) found 16S rRNA gene sequences related to Desulfitobacterium spp. in Fe(III)-enrichment cultures taken from a subsurface sediment contaminated with uranium and nitrate. Kostka (2002) isolated phylotypes closely affiliated to Desulfitobacterium spp. in their 16S rRNA gene libraries derived from an Fe(III)-reducing enriched consortium that originated from subsurface sediments of a site containing iron-rich clay minerals. Reduction of Fe(III) to Fe(II) was observed in a swine manure microcosm amended with FeCl3 as a floculant (Castillo-Gonzalez & Bruns, 2005). This reduction was showed to have a biological origin, and 16S rRNA gene analysis produced sequences related to D. hafniense strain PCP-1 and D. metallireducens.

Desulfitobacterium dehalogenans used humic acids and AQDS as electron acceptors when lactate or hydrogen were provided as electron donor (Lovley, 1998; Cervantes, 2002). Desulfitobacterium sp. strain PCE1 also used AQDS. Microbially reduced humus can transfer electrons abiotically to Fe(III) and Mn(IV) oxides, allowing for its regeneration into the oxidized form (Stone & Morgan, 1984; Lovley & Woodward, 1996). Reoxidized humus can then serve again as electron acceptors for these microorganisms. These reactions may be useful for in situ bioremediation approaches in a polluted ecosystem containing low humic acid concentration but concentrated in iron.


The versatile Desulfitobacterium spp. can also use sulfonates as terminal electron acceptors (Table 4). Sulfonates are organosulfur compounds that are found in natural habitats, but they also have an anthropologic origin in detergents, lubricants or intermediates in chemical synthesis (Lie, 1998). Desulfitobacterium hafniense strain DCB-2 grows on 2-hydroxyethanesulfonate as electron acceptor and generates acetate and sulfide. Desulfitobacterium strain PCE1 and D. dehalogenans can use 2-hydroxyethanesulfonate and alanine-3-sulfonate. Not all Desulfitobacterium strains share this activity, because D. chlororespirans and Desulfitobacterium strain Viet-1 do not grow on tested sulfonates (Lie, 1999).


In anoxic environments, where desulfitobacteria compete against homoacetogens, sulphidogens and methanogens, the availability of electron acceptor species and the concentration of H2 (or other hydrogen donor) determines which group of microorganisms will be favoured. Desulfitobacteria have the advantage of using H2 below the threshold H2 concentration that would allow sulfate reduction and methanogenesis, as long as the availability of halogenated organic compounds is not limited (Löffler, 1999; Bossert, 2003). However, the organic constituents that can be found for instance in soil and sediments can be very complex and may not be used directly by desulfitobacteria. Through a syntrophic relation with microorganisms that can use these complex molecules, desulfitobacteria may acquire their electrons by, for instance, interspecies hydrogen and acetate transfer. Such a syntrophic relation was studied with sulfate-reducing bacteria (SRB) and D. hafniense strain TCE1 (Drzyzga, 2001; Drzyzga & Gottschal, 2002). Strain TCE1 was cocultured with Desulfovibrio fructosivorans growing with fructose and PCE in low sulfate concentrations. Strain TCE1 depended on Desulfovibrio fructosivorans to provide electron donors, because it does not use fructose, or sulfate as electron acceptor. Under these conditions, it was suggested that Desulfovibrio fructosivorans fermented fructose, and produced H2 used by strain TCE1 as an electron donor, keeping the hydrogen partial pressure low, which favours sulfate reduction in low sulfate concentrations (Drzyzga & Gottschal, 2002). However, in high sulfate concentrations, strain TCE1 was outnumbered by Desulfovibrio fructosivorans, and PCE dehalogenation was not occurring. Low sulfate concentrations are often encountered in freshwater and soil environments, suggesting that this syntrophy could occur between desulfitobacteria and SRB. On the other hand, these results predict that desulfitobacteria will not be found in marine environments, as sulfate concentration is high.

Reductive dehalogenases

Tables 6 and 7 summarize the characteristics of the chlorophenol and chloroethene reductive dehalogenases (Rdhases) isolated from Desulfitobacterium strains and their comparison with other Rdhases. So far, six chlorophenol Rdhases and four PCE/TCE Rdhases have been purified from different Desulfitobacterium strains, all from fractions associated with the membrane or the cell wall. All the Rdhases that have been tested for the presence of a corrinoid prosthetic group contained such a group.

View this table:

Properties of chlorophenol reductive dehalogenases

View this table:

Properties of chloroalkyl reductive dehalogenases

Several genes corresponding to Desulfitobacterium Rdhases have been identified and, based on their deduced amino acid sequences, five of the chloroaromatic Rdhases are related to the CprA Rdhase family, and three PCE/TCE Rdhases are related to the PceA family (Fig. 2). The 2,4,6-TCP Rdhase of D. hafniense strain PCP-1 does not belong to these families (see below), and the PCE/TCE Rdhase gene of Desulfitobacterium strain PCE1 has not been isolated yet. The N-terminal amino acid sequences of these Rdhases suggest the presence of a peptide signal, as the cprA/pceA genes encode for a sequence motif RRXFXK, which is characteristic of the TAT secretion pathway (Fig. 3). In most cases, this system involves proteins that bind to various cofactors in the cytoplasm and are thus folded before being exported. Such proteins are predominantly encountered in respiratory electron transport chains (Berks, 2000; Smidt, 2000b), as is to be expected for proteins involved in chloridogenesis. Gene prediction showed that CprA/PceA Rdhases possess cysteine residue-containing motifs characteristic of two Fe-S centres (Fig. 3). Such Fe-S centres have been shown to contain one [4Fe-4S] cluster and one [3Fe-4S] cluster in D. dehalogenans CprA (van de Pas, 1999), and two [4Fe-4S] clusters in Dehalobacter restrictus PceA (Maillard, 2003). The two cystein motifs differ in most cases by the first cystein residue that was proposed to be replaced by a glycine residue in the second cystein motif (von Wintzingerode, 2001). However, Maillard (2003) proposed that the cystein residue located further upstream (10–12 amino acids) could be involved with the other cysteins in binding a Fe-S centre in Dehalobacter restrictus PceA.


Phylogenetic analysis of reductive dehalogenases. Deduced amino acid sequences of characterized and putative reductive dehalogenase genes were aligned with clustalw. Most common gaps were deleted. Phylogenic analysis was made with programs in the phylip package as described in Fig. 1. Trees were inferred from a matrix of pairwise distances (432 aligned positions). GenBank accession numbers are indicated beside the protein name. The scale bar represents 0.20 changes per amino acid position. D.: Desulfitobacterium; De.: Dehalococcoides; Deh.: Dehalobacter; S.: Sulfurospirillum. The strain DCB-2 RdA2 sequence was taken from the D. hafniense strain DCB-2 genome at the Joint Genomic Institute (contig 1057).


Alignment of the N-terminal and the iron–sulfur cluster binding (ISB) region from reductive dehalogenases. The deduced amino acid sequences from characterized and putative reductive dehalogenase genes were initially aligned using the clustalw program and then manually refined. Shaded areas are aligned amino acids in most sequences. The TAT motif (RRXFXK), and the cysteine and proline residues in the two ISB regions are shown under the aligned sequences. A glycine residue was proposed to replace the first cysteine residue in the second ISB region (von Wintzingerode, 2001). CprA and PceA sequences that were almost identical are illustrated by one representative sequence (see Fig. 2 for phylogenic relation). deh: D. dehalogenans; haf: sequences from D. hafniense DCB-2 genome contigs; PCP, KBC1 and Y51: D. hafniense strains PCP-1, KBC1 and Y51; chl: D. chlororespirans; mul: Sulfurospirillum multivorans; Cocc: Dehalococcoides ethenogenes.

Chlorophenol reductive dehalogenases

Chlorophenol Rdhases from D. dehalogenans, D. hafniense strain DCB-2 and Desulfitobacterium sp. strain PCE1 were purified from cultures grown with 3Cl4OHPA (Christiansen, 1998; van de Pas, 1999, 2001a), and from D. chlororespirans with 3-chloro-4-hydroxybenzoate (3Cl4OHBA) (Krasotkina, 2001). These four Rdhases have a similar molecular weight of 47−50 kDa and catalyze the dehalogenation of chlorophenols with chlorine residues at the ortho-position with respect to the hydroxyl group (Table 6). Analysis of the corresponding genes showed that these Rdhases belong to the CprA family. cprA gene products from D. hafniense strain DCB-2, D. dehalogenans and Desulfitobacterium sp. strain PCE1 have more than 95% similarity and have around 76% similarity with the D. chlororespirans CprA.

The Rdhases of D. dehalogenans and D. chlororespirans can dehalogenate 3Cl4OHPA, and several polychlorinated phenols (van de Pas, 1999). The D. chlororespirans Rdhase can also dehalogenate 3,3′,5,5′-tetrachloro-4,4′-dihydroxybiphenyl into trichloro- and dichlorodihydroxybiphenyl at 0.3% of the maximal rate of 3Cl4OHBA. However, this enzyme is only active against phenolic compounds containing the chloro group in ortho to the hydroxyl group. The enzyme isolated from D. hafniense strain DCB-2 was only tested against 3Cl4OHPA and TCE, and the latter was not significantly dehalogenated (Christiansen, 1998).

Dehalogenation with D. hafniense strain PCP-1 against chlorinated molecules must be induced by 3,5-DCP or 2,4,6-TCP (Bouchard, 1996; Dennie, 1998). These two molecules induce the expression of two distinct dehalogenation systems. One is induced by 2,4,6-TCP and is involved mainly in the ortho dechlorination. The other is induced by 3,5-DCP and is responsible for the meta and para dechlorination. A membrane-associated Rdhase was isolated from strain PCP-1 that is responsible for the ortho dechlorination of several polychlorophenols but not of 3Cl4OHPA, in contrast to the other o-CP Rdhases (Boyer, 2003). The crude solubilized dehalogenase preparation performed ortho-dechlorination efficiently against many chlorophenols, but also para-dechlorination against 3,4,5-TCP and 3,5-DCP, although at a lower rate. However, as a crude preparation was used for these assays, the presence of more than one dehalogenase cannot be excluded. The strongest activity was observed with PCP and highly chlorinated phenols as substrates, in contrast to the o-chlorophenol dehalogenases of D. dehalogenans (van de Pas, 1999) and of D. chlororespirans (Krasotkina, 2001), which dechlorinated PCP at a relatively low rate into lesser chlorinated phenols. The apparent molecular weight of this enzyme is 37 kDa as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and 33 800 Da by mass spectrometry analysis. Using the strain DCB-2 genome and with knowledge of the N-terminal amino acid sequence and an internal tryptic peptide sequence, the corresponding gene was isolated. It was designated crdA and showed no homology with any known dehalogenases, indicating a new type of Rdhase. The deduced amino acid sequence revealed a LysM domain, suggesting that the protein is located in the cell wall attached to the peptidoglycan. The sequence does not contain any iron−sulfur centre motif, suggesting that the electron transfer to the corrinoid cofactor is different from the CprA/PceA Rdhases. The corresponding crdA gene in D. hafniense strain DCB-2 encodes for an identical protein except for two small insertions of 3 and 4 amino acids. Protein analysis by two-dimensional gel electrophoresis and RNA analysis by Northern blot and RT-PCR showed that crdA is expressed in the presence or absence of 3,5-DCP or 2,4,6-TCP, the two inducers in strain PCP-1. This suggests that other factors have to be induced by 2,4,6-TCP for CrdA to carry dehalogenation. The crdA gene was shown to be present in D. hafniense strains TCP-A, DP7, DCB-2, and TCE1, D. dehalogenans, D. chlororespirans and Desulfitobacterium sp. PCE-1 as shown by Southern blot experiments. The other Desulfitobacterium strains were not tested. However, crdA transcripts were only detected in strains DCB-2, TCE1 and PCP-1 (Gauthier, 2006). crdA is also found in the strain Y51 genome (Nonaka, 2006).

A membrane-associated dehalogenase with meta- and para-dehalogenation activities and an apparent molecular weight of 57 kDa was isolated from D. hafniense strain PCP-1 (Thibodeau, 2004). It performed meta- and para-dechlorination against several polychlorinated phenols, but also o-dechlorination against 2,4-DCP, 2,4,6-TCP and 2,4,5-TCP. The corresponding gene, named cprA5, encodes for a 548 amino acid protein that presents similarities to CprA/PceA Rdhases (Fig. 2). The cprA5 gene product contains a twin-arginine type (TAT) signal for secretion, a cobalamine-binding site motif, and two iron−sulfur binding motifs. It also shows 66% identity (76−77% similarity) with PceA of D. hafniense strain Y51, Desulfitobacterium sp. strain PCE-S and Dehalobacter restrictus, and less than 40% similarity with the other CprAs (Thibodeau, 2004). In addition to strain PCP-1, cprA5 was found in D. hafniense strains TCP-A and DCB-2, but not in D. hafniense strains DP7 and TCE1, D. dehalogenans, D. chlororespirans and Desulfitobacterium sp. strain PCE1. Gene expression analysis showed that cprA5 is expressed in these three Desulfitobacterium strains only when they were cultured in the presence of chlorophenols (Gauthier, 2006).

PCE/TCE reductive dehalogenases

The four known Desulfitobacterium PCE/TCE Rdhases were isolated from D. hafniense strains PCE-S, TCE1 and Y51, and Desulfitobacterium sp. strain PCE1. Strain Y51 cultures showed a 25-fold increase of PCE/TCE Rdhase activity induced by the addition of TCE (Suyama, 2002). This purified PCE/TCE Rdhase can dehalogenate PCE and TCE into cis-1,2-DCE, and can dehalogenate various chloroethanes such as hexachloroethane into cis-DCE via PCE and TCE. The range of substrates dehalogenated by the Rdhase of strain Y51 is similar to the PCE Rdhase of Dehalobacter restrictus. Dehalogenation of chloroalkanes into lesser chlorinated alkenes is characteristic of the mechanism of dihaloelimination rather than of hydrogenolytic reductive dehalogenation. Results with the Rdhases of D. hafniense strain Y51 and Dehalobacter restrictus PCE suggest that both mechanisms of reductive dehalogenation can be utilized by the same enzymes. The dehalogenation of PCE and TCE into DCE was only demonstrated with strains PCE1, PCE-S and TCE1. The four Desulfitobacterium PCE/TCE Rdhases cannot dehalogenate DCE. None of the chloroaromatic compounds tested with the strains PCE-S and Y51 PCE/TCE Rdhases was dehalogenated.

The N-terminals of D. hafniense strains PCE-S, TCE1 and Y51 PCE/TCE Rdhases are identical to the N-terminal of Dehalobacter restrictus PCE Rdhase (ADIVAPITETSEFPYKVDAK). Cloning and sequencing of the corresponding pceA gene revealed that the deduced amino acid sequences are almost identical between D. hafniense strains PCE-S, TCE1, and Y51, and D. restrictus PceA (more than 96% identity; Fig. 2). In each case, these pceA encode for a 551 amino acid native protein with a 39 amino acid peptide TAT signal, and a 512 amino acid mature protein of 57−58 kDa. The PCE/TCE Rdhase N-terminal of Desulfitobacterium sp. strain PCE1 is completely different (GQESESAIVXFAVQXV) from the others. Recently, Tsukagoshi (2006) isolated Desulfitobacterium sp. strain KBC1, a close relative of strain PCE1, for its ability to dehalogenate PCE to TCE (Fig. 1). The putative pceA gene (named prdA) was cloned and its sequence revealed the same N-terminal as that of strain PCE1 PCE/TCE Rdhase. It encodes a PceA/CprA-type protein of 463 amino acids with a predicted 49 amino acid peptide signal. The predicted mature protein contain 414 amino acid residues with a molecular mass of 45 843 Da.

1,2-dichloroethandichloroelimination system

Dechlorination of 1,2-dichloroethane (1,2-DCA) to ethene with D. dichloroeliminans strain DCA1 is constitutive, because there is no requirement for preculturing strain DCA1 with 1,2-DCA for induction of this activity. Characterization of the 1,2-DCA dichloroelimination system showed that a corrinoid cofactor is involved because vitamin B12 is required for growth and propyl iodide caused a light-reversible inhibition of the dechlorinating activity (De Wildeman, 2003).

Gene arrangement

All the known genes encoding CprA and PceA or putative Rdhases are linked to a gene encoding for a small hydrophobic protein (CprB and PceB; around 100 amino acids), which possesses three membrane-spanning helices (Smidt, 2000b; Villemur, 2002). This protein was proposed to be involved in anchoring the Rdhases to the membrane. Interestingly, the cprB genes are always found upstream of the cprA genes, and the pceB genes downstream of the pceA genes (Fig. 4). The only exception is the cprA5 gene encoding the 3,5-DCP Rdhase of strain PCP-1, which has its cprB downstream (Thibodeau, 2004). Surprisingly, CprA5 is more related to the PceA family than the CprA family (Fig. 2). Preliminary results indicated that CprA5 has very little dehalogenating activity against PCE (Beaudet, unpublished results).


Gene arrangements of reductive dehalogenase related loci. The gene arrangements are based on gene sequences and annotations available in GenBank, and at the Joint Genomic Institute (JGI) for the Desulfitobacterium hafniense strain DCB-2 genome. Numbers on the right side are GenBank accession numbers. tnp: transposition related sequences. The C, D, K, T and Z genes (either associated with pce, cpr, rdh or prd) were named because of their homology with cprC, cprD, cprK, cprT and cprZ characterized in Desulfitobacterium dehalogenans. Tsukagoshi (2006) mentioned that the prdAB sequence was found adjacent to a cprBA sequence in Desulfitobacterium sp. strain KBC1.

Smidt (2000b) isolated and sequenced an 11.5-kbp genomic fragment from the D. dehalogenans containing the cprBA genes. Six transcriptional genes were identified and named cprC, cprD, cprE, cprK, cprT and cprZ (Fig. 4). Except for the cprZ gene product, where no function has been assigned, the predicted proteins encoded by these genes seem to be involved in regulation, maturation, or action of CprA. CprC contains six transmembrane helices and has similarities with membrane-bound regulators of the NosR/NirI type. This type of regulator may play a role in a signal transduction pathway and the control of transcription. CprK has similarities with transcriptional regulators of the CRP-FNR family. CprC and CprK could then be involved in regulation of the expression of cprA. Indeed, characterization of CprK from D. dehalogenans and D. hafniense strain DCB-2 showed that CprK binds in vitro 3Cl4OHPA, and this complex in turn promotes a specific interaction with a DNA sequence upstream of several cpr genes (Fig. 5), and can stimulate in vivo transcription. Furthermore, the oxidized form of CprK does not bind in vitro its target DNA, suggesting that CprK is redox-regulated (Pop, 2004; Gabor, 2006). This regulation does not, however, appear to be located at the Cys11-Cys200 disulfide bridge (Gabor, 2006). The cprD and cprE gene products are similar to chaperonins of the GroEL-type. CprT has similarity with a trigger factor, a peptidyl propyl isomerase, considered to be involved in protein folding. CprD, CprE and CprT could then be involved in the correct folding, processing and assembly of CprA and insertion of its cofactors. Finally, CprZ has no similarity with proteins with known function. Northern blot experiments showed that transcripts of these genes were only observed when D. dehalogenans was grown by chloridogenesis, in the presence of 3Cl4OHPA, but not in the presence of fumarate, nitrate or pyruvate as electron acceptor. cprBA transcripts were shown to be induced 18-fold after 3 h under chloridogenesis conditions. Such transcriptional stimulation was also observed with Desulfitobacterium sp. strain KBC1. The transcripts of the putative PCE Rdhase genes, prdAB, were only detected when strain KBC1 was cultured in the presence of PCE, but not in the presence of TCE or 3Cl4OHPA (Tsukagoshi, 2006). These results suggest that the expression of Rdhase genes are tightly regulated by their substrates.


Model of the chain of reactions associated with reductive dehalogenation. HK: histidine kinase with response regulator (ReRe). TAT: twin arginine-specific complex. Me: metal ions that may be involved in the control of gene expression and are incorporated in mature reductive dehalogenase. Reductive dehalogenases are represented facing the outside of the cytoplasmic membrane but could also be facing the inside. H2-ase: hydrogenase. Co: corrinoid. FeS: iron−sulfur centres (based on Holliger, 2003; Smidt & de Vos, 2004).

The gene arrangement pceABCT was found in D. hafniense strains Y51 and TCE1, and in Dehalobacter restrictus (Maillard, 2003, 2005; Furukawa, 2005; Futagami et al., 2005) (Fig. 4). pceC and pceT encode for proteins related to CprC and CprT. Interestingly, sequences related to transposable elements homologous to the IS256 family were found flanking these three loci, suggesting horizontal transfer between Dehalobacter restrictus and D. hafniense. Indeed, the three loci are more than 99% identical in their DNA sequence. The addition of cis-DCE to strain Y51 culture inhibited its growth. Inhibition was also observed with PCE but to a lesser extent. When strain Y51 was cultured in serial subcultures in the presence of cis-DCE or PCE, the majority of the recovered colonies lost their PCE-degrading activity after 3–5 subcultures (Furukawa, 2005; Futagami et al., 2005). This also occurred when Y51 was cultured in the absence of cis-DCE or PCE but not in the presence of TCE. Gene analysis showed that these non PCE-dechlorinating colonies had lost the pceABCT gene locus or a sequence upstream of pceA, where a potential promoter is located. Futagami et al. (2005) and Maillard (2005) observed circular forms containing the transposable elements and the pceABCT genes in strains Y51 and TCE1, suggesting that these genes are part of a catabolic transposon.

The available sequence of the D. hafniense strain DCB-2 genome revealed at least 5 chromosomic regions containing rdh-related genes (Villemur, 2002) (Fig. 4). (Here rdh will refer to putative cpr or pce genes.) Villemur (2002) proposed to name the six related rdhA genes found in these chromosomic regions cprA1 to cprA4, and rdA1 and rdA2. The contig 1065, which spans approximatively 55 kbp, contains three rdh gene clusters. The first one has a cpr gene arrangement identical to the one of D. dehalogenans. Nucleic acid sequence showed 83% identity between the two loci, and 80 to 94% similarities between their respective gene products. It contains the cprA1 gene, which encodes the 3Cl4OHPA Rdhase of D. hafniense strain DCB-2 (Christiansen, 1998). The second rdh gene cluster is again related to cpr in sequences and gene arrangement, and contains two adjacent cprBA operons (cprA2 and cprA3). Surrounding these operons are genes related to cprK, cprT, cprC, cprZ, and cprD. Twin cprBA genes were observed in D. hafniense strain PCP-1 and in D. chlororespirans. The third rdh gene cluster on contig 1065 is more related to the pce gene arrangement found in D. hafniense strains Y51 and TCE1, and Dehalobacter restrictus. Villemur (2002) named RdA1 the putative reductive dehalogenase encoded by this gene cluster. This third rdh gene cluster, which includes the rdhABCT open reading frames, has 65% identity with the corresponding cluster in D. hafniense strains Y51 and TCE1, and Dehalobacter restrictus, and their corresponding gene products are between 68 and 77% similar. This rdhA gene in strain DCB-2 encodes the m-, p-CP Rdhase (CprA5) in strain PCP-1 (Fig. 2). Flanking the rdhABCT genes in contig 1065 are two genes (rdhK and rdhC) that encode for related CprK and CprC. Furthermore, as observed in D. hafniense strains Y51 and TCE1 and Dehalobacter restrictus, a transposable element was found upstream of this gene arrangement. This suggests that this locus was acquired by horizontal transfer. Rhee (2003) also observed a similar Rdhase-like gene arrangement in DNA extracted from a 2-bromophenol-degrading consortium. The genes flanking this arrangement were also related to transposases.

Contig 1022 contains the fourth gene cluster in the strain DCB-2 genome (cprKBA) encoding for cprA4. Finally, rdA2 was found in contig 1057. The gene arrangement showed that a rdhB-type gene is present downstream of this rdhA gene, suggesting a pce gene arrangement. These two genes are flanked by cprK- and cprC-related genes. As contigs 1022 and 1057 are relatively short compared with contig 1065, other cpr/pce-related genes could be present upstream or downstream in these chromosomal regions. The fact that several potential Rdhase genes are present in D. hafniense DCB-2 and probably in other Desulfitobacterium species may be related to the spectrum of substrates that they can dehalogenate. For instance, PCR analyses showed that cprA1 is not present in D. hafniense strain PCP-1, correlating with the absence of such activity in this strain (Villemur, 2002). Furthermore, Gauthier (2006) showed that three putative cprA-related genes (cprA2, cprA3 and cprA4) were expressed in D. hafniense strains DCB-2, TCP-A and PCP-1 when cultured in the presence or absence of 3,5-DCP or 2,4,6-TCP, suggesting that they are constitutively expressed in the culture conditions used to grow these three strains.

Heterologous expression of Rdhase genes

The pceAB operon of D. hafniense strain Y51 was overexpressed in Escherichia coli, but no dehalogenating activity was observed (Suyama, 2002). It was suggested that the enzyme might not fold properly and/or that the proper cofactor was not present in E. coli. Despite its lack of dechlorinating activity, this resulting PceA was used to raise antibodies, which were used to show that the mature and active PceA was located in the strain Y51 periplasmic fraction. This result is odd because Gram-positive bacteria such as Desulfitovibrium spp. should not have a periplasm. Suyama (2002) used fractionation techniques developed for Gram-negative bacteria, and the proteins released upon digestion by lysosyme would probably come from cell wall. Overexpression of S. multivorans pceAB genes in E. coli was also achieved, but no dehalogenating activity was observed (Neumann, 1998).

Mechanisms of electron transfer of CprA/PceA Rdhases

Smidt (1999) used the conjugative transposon Tn916 to derive chloridogenesis-deficient mutants in D. dehalogenans. The mutants were unable to dehalogenate 3Cl4OHPA and differed in their ability to use nitrate and fumarate. The genes that were affected by Tn916 encoded for proteins related to structural or regulatory components shared by different respiration chains present in D. dehalogenans such as a sensory transduction histidine kinase, an oxygen-independent coproporphyrinogen III oxidase (porphyrin biosynthesis), and formate-hydrogen lyases. One of the transposons inserted upstream of a gene cluster encoding for proteins related to the small and large subunits of periplasmic Ni-hydrogenases (hydA, hydB), and to cytochrome B (hydC), which serve as mediators in the electron transfer from H2 to the quinone. It was also shown with other chloridogenesis bacteria, such as Desulfomonile tiedjei DCB-1, Dehalococcoides ethenogenes, Desulfitobacterium sp. strain PCE-S, Dehalobacter restrictus and Sulfurospirullum multivorans, that hydrogenases or formate dehydrogenases, facing the outside of the cytoplasmic membrane, were associated as components of an electron-donating enzymatic system (Schumacher & Holliger, 1996; Miller, 1997a, 1998; Magnuson, 1998; Louie & Mohn, 1999). Furthermore, cytochromes and quinones have been proposed to be involved in the electron transfer from the hydrogenases to Rdhases (Holliger, 2003) and in the generation of the proton gradient (Fig. 5). All the CprA/PceA Rdhases have been associated with the membrane fraction, except the PCE Rdhase of S. multivorans. Although no obvious transmembranous domain was observed, the CprA/PceA Rdhase may be anchored to the membrane by the hydrophobic CprB/PceB protein. These two genes are probably coexpressed, as observed with the cprBA genes in D. dehalogenans. Because all the CprA/PceA Rdhases have a TAT signal, it is conceivable that these Rdhases may face the cell-wall region of the membrane, as observed with D. hafniense strain Y51 PceA. Furthermore, the non-CprA, 2,4,6-TCP Rdhase of D. hafniense strain PCP-1 contains a LysM domain, which is present in a large number of proteins involved in binding peptidoglycan (Boyer, 2003), one of the major cell-wall constituents.

The detailed mechanisms of the dehalogenation process by CprA/PceA Rdhases are still to be elucidated. Two models have been proposed, based on experiments on PCE/TCE Rdhases of S. multivorans and Dehalobacter restrictus (Fig. 5). One model proposed from the characterization of the PCE Rdhase of S. multivorans involves the reduction of the halogenated molecule by the Co(I) corrinoid and the formation of an alkyl-Co(III) corrinoid. Upon protonation, the alkyl moiety is cleaved, and the Co(III) corrinoid is reduced by the two iron−sulfur centres of the enzyme to the Co(I) state (Neumann, 1996; Holliger, 2003). The second model is based on the characterization of Dehalobacter restrictus PCE/TCE Rdhase, where reduction of the halogenated molecule by the Co(I) corrinoid first produces a radical that is further reduced by one of the iron–sulfur centres. Only the Co(II) corrinoid is formed, which is then reduced back to the Co(I) state by the other iron-sulfur centre (Schumacher, 1997; Banerjee & Ragsdale, 2003; Holliger, 2003). The only study on Desulfitobacterium Rdhase, the D. dehalogenans o-CP Rdhase, showed no radical or Cob(III)alamine intermediate (van de Pas, 1999).

In order to characterize all the mechanisms involved in chloridogenesis better, a gene transfer system is needed. Smidt (2001) developed a cloning vector for replication in D. dehalogenans and for possible specific chromosomal insertions. They disrupted frdA that encodes for a putative fumarate dehydrogenase and observed a partially imparted fumarate reductase activity. This approach could be used to specifically target genes putatively involved in chloridogenesis.

The D. hafniense genome

The whole genome of D. hafniense strain Y51 has recently been completed (Nonaka, 2006). Its genome is 5 727 534 bp, with 5060 predicted protein-coding sequences; among them, 25% are not related to any known function. The partial genome sequence of strain DCB-2 has been available since 2001 (last update in November 2003) at the Joint Genome Institute (http://www.jgi.doe.gov/). Based on the GenBank database, more than 5.1 million bp were determined on 337 contigs, and 4389 protein-encoding gene candidates were detected.

These genome sequences are of great importance because knowledge of them will allow better understanding of the various mechanisms of dehalogenation by desulfitobacteria. For example, they allowed the identification of the genes crdA and cprA5 encoding for two Rdhases in D. hafniense strain PCP-1 (Boyer, 2003; Thibodeau, 2004). Furthermore, genes encoding for putative Rdhases were observed, suggesting that strain DCB-2 has the potential to dehalogenate other halogenated organic compounds (Villemur, 2002).

The sequences can also predict the metabolic pathway used by desulfitobacteria. For instance, several coding sequences have been found in the strain Y51 genome that are related to dimethyl sulfoxide reductase or polysulfide reductase complexes, which are thought to participate in the electron transfer process. This suggests that strain Y51 has the potential to reduce a wide variety of molecules, most of which are still unknown. The sequences also allow comparisons to be made between specific genes and those of other microbial genomes (Klein, 2001). Ziegler (2002) found in the strain DCB-2 genome a gene encoding for cyanophycin synthetase-like enzyme. This enzyme is usually found in cyanobacteria and catalyses the formation of cyanophycin, a protein-like reserve polymer consisting of aspartic acid and arginine. This gene was expressed in E. coli and produced cyanophycin. The authors suggest that cyanophycin is more widely distributed in prokaryotes than initially thought.


One of the ultimate practical goals of studying Desulfitobacterium spp. is the development of efficient approaches to bioremediating soil/sediments or water contaminated with halogenated organic compounds. The ability of Desulfitobacterium spp. to use halogenated organic compounds and metals as electron acceptors could be applied to the development of bioremediation strategies to decontaminate areas co-contaminated with both of these contaminants. For example, soils cocontaminated with As(V) and halogenated organic compounds could be bioremediated. Parallel to dehalogenation, the insoluble As(V) can be reduced and converted into the more soluble As(III). Although this is the toxic form, As(III) can than be recovered by pumping the lixiviat and treated off-site, providing a means of decontamination.


In situ soil treatments could be improved by adding the appropriate nutrients to stimulate Desulfitobacterium growth (biostimulation) or to favour syntrophic relationships with other microorganisms. Monitoring the presence of desulfitobacteria in polluted environments could be used as a mean to estimate the potential of this environment to be decontaminated by biostimulation. As mentioned before, this can be achieved with molecular tools targeting 16S rRNA gene-specific sequences of Desulfitobacterium spp., or any other reductive dehalogenating bacteria. Alternatively, PCR primers can target genes involved in reductive dehalogenation. Degenerated oligonucleotides designed to target specific sequences of cprA/pceA genes have been tested in a number of systems (von Wintzingerode, 2001; Ahn, 2003; Rhee, 2003; Regeard, 2004). Von Wintzingerode (2001) were able to detect cprA/pceA-like gene fragments from both trichlorobenzene- and 1,2-dichloropropane-dechlorinating microbial consortia. Rhee (2003) detected a pceA-like sequence from a sulfidogenic 2-bromophenol-degrading consortium enriched from contaminated estuarine sediments. When bacteria or genes involved in reductive dehalogenation are detected, the dehalogenating capacity of the microbiota of the soil would have to be tested with microcosm assays because these bacteria or genes differ in their capacity to dehalogenate halogenated organic compounds.

The detection of desulfitobacteria in a site using molecular probes targeting specific 16S rRNA gene sequences is no warrant of the potential for reductive dehalogenation because, as revealed by D. hafniense strain DP7, these activities may be lacking (van de Pas, 2001b). Targetting Rdhase signature is certainly a better approach, but even this criterion is no warrant of active reductive dehalogenating activity. The crdA gene was detected in strain DP7, but RT-PCR experiments showed that it was not expressed under the conditions used (Gauthier, 2006). Another example is S. multivorans strains N and K, both containing identical pceA gene. Only strain K, however, is capable of PCE dechlorination. Siebert (2002) showed by RT-PCR that in both strains pceA transcripts were present. However, a corrinoid factor was not synthesized by strain N, and the apoprotein of the dehalogenase was not formed in detectable amounts. It was suggested that the inability of strain N to dechlorinate PCE is a result of the lack of the specific PCE dehalogenase corrinoid cofactor.


Another approach to bioremediating sites polluted with halogenated organic compounds is to inoculate Desulfitobacterium cultures or a halogenated compound-degrading consortium containing desulfitobacteria in these sites. Bioaugmentation experiments were carried out with D. hafniense strain PCP-1 to develop a highly efficient bioremediation treatment of PCP-contaminated soils (Beaudet, 1998; Lanthier, 2000). Strain PCP-1 was added (106-107 cells per g soil) to highly contaminated soil slurry microcosms amended with 100–750 mgPCP −1 kgsoil −1 of soil, and the microcosms were incubated under conditions favouring growth of strain PCP-1. More than 90% of the initial PCP was dechlorinated in microcosms containing 100 to 200 mgPCP −1 kgsoil −1. Higher PCP concentration (300–500 mgPCP −1 kgsoil −1) required reinoculation of strain PCP-1 after 12 days to achieve the same level of dechlorination. No PCP degradation was observed at 750 mgPCP −1 kgsoil −1. Strain PCP-1 was detected by PCR up to 500 mgPCP −1 kgsoil −1. At 750 mgPCP −1 kgsoil −1, strain PCP-1 was no longer detected after a few days, suggesting a toxic effect of PCP at this concentration. Concentration of strain PCP-1 in this system was determined by competitive PCR (Lévesque, 1998), and shown to be stable at around 108 cells mL−1 of slurry during the treatments, except in the 750 mgPCP −1 kgsoil −1 microcosms, where the cell concentration dropped below the limit of detection after 12 days.

El Fantroussi (1997) used a similar approach in which D. dehalogenans was inoculated in soil slurry microcosms artificially contaminated with 3Cl4OHPA. Complete degradation of 3Cl4OHPA was achieved in 4 days instead of 11 days with non-inoculated microcosms. Monitoring by nested PCR showed that D. dehalogenans was present during the 4 days and after.

Treatment of PCE-contaminated soils could be possible with the combination of D. hafniense strain Y51 and Fe0. Inoculation of strain Y51 in soil slurry microcosm amended with 60 μmol PCE kg−1 soil led to complete transformation of PCE into DCE in 40 days. Combined with 1% Fe0, PCE was dehalogenated into DCE in 6 days. DCE was then completely dehalogenated by a subsequent abiotic chemical process with Fe0. The Fe0 corrosion generated by reaction with water can produced H2, which can be used by strain Y51 as electron donors to dehalogenate PCE (Lee, 2001).

De Wildeman (2004) experimented with the use of D. dichloroeliminans strain DCA1 to treat groundwater contaminated with 1,2-DCA. Strain DCA1 was inoculated in non-sterile groundwater amended with 1,2-DCA and lactate as carbon source and electron donor. Strain DCA1 was capable of competing with the endogenous microbiota and achieved complete transformation of 400 μM 1,2-DCA into ethene in 3 days at 28°C. Strain DCA1 tolerated 10 g L−1 NaCl, and was still capable of dehalogenation at pH 6.0 and 12°C. Strain DCA1 is more advantageous to use compared with Dehalococcoides ethenogenes strain 195, which also dechlorinates 1,2-DCA into ethane but produced up to 1% of vinyl chloride, a very toxic product. Strain 195 is also very sensitive to oxygen, while strain DCA1 survived to 20% oxygen, although dehalogenation was inhibited in air. However, as soon as the oxygen was consumed by facultative microorganisms, strain DCA1 resumed dechlorination of 1,2-DCA.

Successful accomplishment of dehalogenation by the bioaugmentation approach will certainly be influenced by the soil composition. Our group succeeded in dechlorinating PCP in PCP-amended soil rich in organic matter and inoculated with strain PCP-1 (Lanthier, 2000). However, such dehalogenation activity was not observed in a sandy soil amended with PCP, with a decline of strain PCP-1 cell concentration. We also tried this bioaugmentation approach with a soil sample taken from a contaminated site which contained a high concentration of PCP (mgPCP −1 kgsoil −1), creosote and organic matter. No dehalogenation occurred, and the strain PCP-1 cell concentration declined, even when the soil was diluted with a non-contaminated soil rich in organic matter. This suggests that factors such as the organic matter concentration and the toxicity of other pollutants have to be taken into account for the survival of reductive dehalogenating bacteria.


One of the problems associated with in situ bioremediation is that halogenated organic compounds contaminating a soil can diffuse in groundwater. Therefore, these pollutants have to be extracted from soil/sediments or wastewater, collected, and properly treated. Wastewaters contaminated with halogenated organic compounds generated by industrial activity can also be treated before being released. Bioreactors with halogenated compound-degrading consortium containing desulfitobacteria can be developed for that purpose.

Lecouturier (2003) developed a laboratory-scale anaerobic continuous fixed-bed reactor capable of completely dehalogenating 5-amino-2,4,6-triiodoisophthalic acid (ATIA). This reactor was inoculated previously with an ATIA-enriched culture from an industrial wastewater treatment plant with a history of contamination with triiodinated compounds. ATIA is a component of X-ray contrast agents in radiology that contaminates hospital wastewaters. The specific deiodination rate reached 13.7 mmolATIA gVSS −1 d−1 (VSS, volatile suspended solid). As mentioned before, 16S rRNA gene sequences related to Desulfitobacterium spp. were found in the microbial community reactor. Beaudet (1997) used the original PCP-degrading methanogenic consortium from which D. hafniense strain PCP-1 was isolated (Juteau, 1995a,b) to develop an upflow anaerobic fixed-film reactor. This reactor was used to treat PCP extracted from contaminated wood chip and wood powder. Complete dechlorination of PCP and phenol removal were achieved for a PCP loading rate of 50−60 μmol L−1 of reactor volume per day (Embedded Image). Yang (2005) enriched desulfitobacteria in subcultures derived from a PCE-degrading anaerobic chemostat fed with benzoate, yeast extract and PCE. 16S rRNA gene sequences related to D. hafniense were detected in these PCE-enrichment cultures growing in the presence of H2 or acetate as electron donors. FISH with specific fluorescent oligonucleotides showed that D. hafniense cells represented 17% (H2/PCE) and 53% (acetate/PCE) of the respective enrichments. These cells were not detected in the original chemostat, probably because they were below the limit of detection and/or their ribosomal content was low and generated no detectable signal. Von Wintzingerode (1999) carried out a survey of the microbial consortium from a fluidized-bed reactor used to dechlorinate trichlorobenzene, and observed 16S rRNA gene sequences related to Dehalobacter/Desulfitobacterium spp. Specific microbial populations can also be enriched in bioreactors to treat metal-contaminated wastewater such as mentioned above with the sulfate-reducing fluidized-bed reactor developed to treat acidic metal-containing wastewater (Kaksonen, 2004).

Inoculation of upflow anaerobic sludge blanket (UASB) reactors has been achieved at the laboratory scale with D. hafniense strains DCB-2 and PCP-1 for the degradation of PCP, and with Desulfomonile tiedjei and Sulfurospirillum multivorans for the degradation of 3-chlorobenzoate and PCE, respectively. Christiansen & Ahring (1996b) inoculated strains DCB-2 in an UASB reactor containing sterilized granules that was subsequently fed with PCP and lactate. PCP was transformed into 3,4,5-TCP, with a transformation rate of up to 158 μmol Embedded Image. Colonization of the sterilized granules by strain DCB-2 was revealed by the use of specific antibodies, and showed that strain DCB-2 formed a net structure inside the granules. A PCP-degrading UASB reactor was also developed by inoculating D. hafniense strain PCP-1 in the reactor containing live granules (Tartakovsky, 1999). Proliferation of strain PCP-1 allowed a substantial increase of the volumetric PCP load from 19 to 300 (μmol Embedded Image) with a 99% PCP removal efficiency and a dechlorination efficiency equal to or greater than 91%. Following inoculation, the concentration of strain PCP-1 was found to increase from 106 to 1010 cells gVSS −1, as observed by competitive PCR. Gradual colonization of the granules by strain PCP-1 was observed by FISH (Lanthier, 2002). Small colonies of strain PCP-1 were observed after 3 weeks of operation, and after 9 weeks the outer biofilm layer was completely colonized by strain PCP-1 (Fig. 6). Contrary to the case with strain DCB-2, no PCP-1 microcolonies were observed in the inner part of the granules. It is possible that strain PCP-1 cells colonized the inner part of the granules as did strain DCB-2, but that not enough ribosomes were present to produce a detectable FISH signal. The deeper penetration and more uniform distribution of D. hafniense strain DCB-2 in the granules probably result from the fact that the granules were autoclaved. This means that there was no competition for space and substrate with other bacteria, which would promote a more uniform growth. This hypothesis is corroborated by results from Ahring (1992) with 3-chlorobenzoate-degrading UASB reactors inoculated with Desulfomonile tiedjei. They observed by immunofluorescence that the surface and the subsurface of the live granules were colonized by Desulfomonile tiedjei. The transformation rate of these reactors reached 54 μmol ggranule volatile solid −1 d−1. Furthermore, Horber (1998) developed two UASB reactors transforming PCE to DCE by inoculating S. multivorans either to autoclaved or live granules. Both reactors achieved an average transformation of PCE into DCE of 96% and 100% with a transformation rate of up to 1400 and 1900 μmolPCELsludge −1 d−1, respectively. Sulfurospirillum multivorans cells, as revealed by immunofluorescence, built a net-like structure at the surface of PCE-fed live granules, but also formed microcolonies principally in the centre of autoclaved granules.


Localization of strain PCP-1 within granules by FISH. Anaerobic granules were taken from a PCP-degrading upflow anaerobic sludge blanket reactor augmented with Desulfitobacterium hafniense PCP-1 after 3, 4, 5 and 9 weeks of operation (A-D). The granules were fixed and solidified with Paraplast, then cut in 7-μm slices. FISH was carried out with a strain PCP-1-specific, oligonucleotide fluorescent probe coupled to Cy3 (red). Cells were counterstained with DAPI (blue). The bar corresponds to 20 μm (adapted from Lanthier, 2002).

Lanthier (2005) developed a methanogenic fixed-film reactor acclimated with increasing PCP concentrations. One of the PCP-fed reactors degraded a PCP load of 1173 μmol Embedded Image after 225 days of operation, with approximately 60% degradation of 3-CP, the main chlorophenol residual intermediate. This reactor achieved a three-fold higher PCP removal rate than the best anaerobic PCP-degrading reactor reported (330–364 μmol Embedded Image) (Wu, 1993). FISH experiments showed that D. hafniense cells were present and scattered into the biofilm of the PCP-fed reactors and accounted for 19% of the population (Fig. 7). This proportion dropped to less than 1% in the control reactor not fed with PCP. It was hypothesized from the results obtained with the two types of reactors studied by Lanthier (2002, 2005) that colonization of the granules and the biofilm by D. hafniense probably protects the other microorganisms, especially the methanogens, from PCP toxicity. Desulfitobacterium hafniense may have fed on end-products generated by the fermentative bacteria usually found in the outer layer of the granules or by using H2 generated under methanogenic conditions.


Spatial arrangement of Desulfitobacterium hafniense in a PCP-fed anaerobic fixed film reactor. The biofilm sample was taken from the reactor after 146 days of operation. It was fixed, permeabilized and hybridized with fluorescent-labelled oligonucleotide probes specific for D. hafniense (purple). Total cells were counterstained with YOYO-1 (green) before being examined by confocal microscopy. The bar corresponds to 20 μm (adapted from Lanthier, 2005).

Concluding remarks

Desulfitobacterium spp. are very versatile microorganisms that can survive in a variety of environments. They have been isolated from different parts of the world. Several of them can sporulate and then can resist harsh environments. They can use a wide variety of electron acceptors, from nitrate to sulfite, metals, humic acids, and man-made or naturally occurring halogenated organic compounds. They also use H2 as electron donor below the threshold concentration that would allow sulfate reduction and methanogenesis. Furthermore, there is some evidence that syntrophic relationships occur between Desulfitobacterium spp. and sulfate-reducing bacteria, and this relationship is believed to occur in microbial consortiums in methanogenic PCP-degrading reactors. All these features make Desulfitobacterium spp. excellent candidates for the development of anaerobic bioremediation processes. In fact, desulfitobacteria have been detected in bioprocesses treating sites contaminated with halogenated organic compounds but also with high concentrations of toxic metals.

The characterization of Rdhases and their corresponding genes in Desulfitobacterium spp. has demonstrated the diversity of these genes in members of this genus. This diversity may reflect the differences in the specificity towards halogenated organic compounds among Desulfitobacterium strains. In this regard, the genome of strain DCB-2 revealed 3 Rdhases (CprA1, CprA5 and CrdA) and 4 putative ones (CprA2, CprA3, CprA4 and RdA2) that may be required to dehalogenate all the halogenated organic compounds that strain DCB-2 can dehalogenate. It will be important to express these putative Rdhases in heterologous systems to obtain larger amounts of these enzymes in order to perform 3D analysis, or to perform genetic modifications of the coding sequence to develop improved Rdhases (wider spectrum of dehalogenation, better dehalogenation efficiency). However, this is impaired by (1) the lack of knowledge about cofactors, especially cobalamine, associated with Rdhases, and (2) the absence of the said cofactors in bacterial hosts such as E. coli that are used for heterologous expression.


  • Editor: Jiri Damborsky


View Abstract