The microbial reduction of metals has attracted recent interest as these transformations can play crucial roles in the cycling of both inorganic and organic species in a range of environments and, if harnessed, may offer the basis for a wide range of innovative biotechnological processes. Under certain conditions, however, microbial metal reduction can also mobilise toxic metals with potentially calamitous effects on human health. This review focuses on recent research on the reduction of a wide range of metals including Fe(III), Mn(IV) and other more toxic metals such as Cr(VI), Hg(II), Co(III), Pd(II), Au(III), Ag(I), Mo(VI) and V(V). The reduction of metalloids including As(V) and Se(VI) and radionuclides including U(VI), Np(V) and Tc(VII) is also reviewed. Rapid advances over the last decade have resulted in a detailed understanding of some of these transformations at a molecular level. Where known, the mechanisms of metal reduction are discussed, alongside the environmental impact of such transformations and possible biotechnological applications that could utilise these activities.
Although it has been known for over a century that micro-organisms have the potential to reduce metals, more recent observations showing that a diversity of specialist Bacteria and Archaea can use such activities to conserve energy for growth under anaerobic conditions (for example ‘dissimilatory’ processes described in [1–4]) have opened up new and fascinating areas of research with potentially exciting practical applications. Micro-organisms have also evolved metal resistance processes that often incorporate changes in the oxidation state of toxic metals. Several such resistance mechanisms, which do not support anaerobic growth, have been studied in detail using the tools of molecular biology. Three important examples have been reviewed recently in this journal: resistance to Hg(II) , As(V)  and Ag(II) . The molecular basis of respiratory metal reduction processes has not, however, been studied in such fine detail, although rapid advances are expected in this area with the imminent availability of complete genome sequences for key metal-reducing bacteria, in combination with genomic and proteomic tools. This research is being driven forward both by the need to understand the fundamental basis of biogeochemical cycles of several key elements, and also by the possibility of harnessing such activities for a range of biotechnological applications. These include the bioremediation of metal-contaminated land and water , the oxidation of xenobiotics under anaerobic conditions , metal recovery in combination with the formation of novel biocatalysts , and even the generation of electricity from sediments . The aims of this review are to give an overview of the range of metals (and metalloids) reduced by micro-organisms, the mechanisms involved and the environmental impact of such transformations. Where appropriate, possible applications for these processes will be discussed. Given the availability of several excellent recent reviews on microbial metal reduction resistance determinants, this review will focus mainly on dissimilatory metal reduction processes.
Reduction of Fe(III) and Mn(IV)
A wide range of Archaea and Bacteria are able to conserve energy though the reduction of Fe(III) (ferric iron) to Fe(II) (ferrous iron). Many of these organisms are also able to grow through the reduction of Mn(VI) to Mn(II). The environmental relevance of Fe(III) and Mn(VI) reduction has been well documented [12, 13]. Indeed, geochemical and microbiological evidence suggests that the reduction of Fe(III) may have been an early form of respiration on Earth , and is a candidate for the basis of life on other planets . On modern Earth, Fe(III) can be the dominant electron acceptor for microbial respiration in many subsurface environments . As such, Fe(III)-reducing communities can be responsible for the majority of the organic matter oxidised in such environments . Recent studies have shown that several important xenobiotics that contaminate aquifers can also be degraded under anaerobic conditions by Fe(III)- and Mn(IV)-reducing micro-organisms [9, 18, 19]. In addition to playing an important role in the degradation of organic material, Fe(III)- and Mn(IV)-reducing micro-organisms can influence the mineralogy of sediments through the reductive dissolution of insoluble Fe(III) and Mn(IV) oxides. These processes can result in the release of potentially toxic levels of Fe(II) and Mn(II), and also trace metals that were bound by the Fe(III) or Mn(IV) minerals. Depending on the chemistry of the water, a range of reduced minerals can also be formed including magnetite (Fe3O4), siderite (FeCO3) and rhodochrosite (MnCO3), resulting in a change in structure of the sediments. Finally, Fe(III)- and Mn(IV)-reducing micro-organisms can also impact on the fate of other high-valency contaminant metals through direct enzymatic reduction, and also via indirect reduction catalysed by biogenic Fe(II). The bioreduction of U(VI), Cr(VI) and Tc(VII) by Fe(III)-reducing micro-organisms will be discussed in detail in later sections, and can result in immobilisation of these potentially toxic and mobile metals in sediments .
Diversity of Fe(III)-reducing organisms
Early studies on the reduction of Fe(III) and Mn(IV) (e.g. ) focused on organisms that can grow predominantly via fermentation of sugars such as glucose, with metals utilised as minor electron acceptors, typically <5% of the reducing equivalents used for metal reduction . It has not been demonstrated conclusively that metal reduction via this form of metabolism increases cell yields. More recent studies have demonstrated that Bacteria including a Vibrio sp.  and Wolinella succinogenes and Archaea such as Archaeoglobus fulgidus, Pyrococcus furiosus and Pyrodictium abyssi can couple the oxidation of hydrogen and short-chain fatty acids to the reduction of Fe(III), although it has not been shown that these reactions support growth. Sulfate-reducing bacteria also reduce Fe(III)  and other metals (see later sections), although again growth has not been demonstrated using these electron acceptors. A recent paper suggesting the contrary  has been disputed on the basis of inconsistencies in the cell yields when grown on a range of metals .
The first organisms that were unequivocally shown to conserve energy for growth through the reduction of Fe(III) or Mn(IV) were Shewanella oneidensis (formerly Alteromonas putrefaciens and then Shewanella putrefaciens) and Geobacter metallireducens (formerly strain GS-15) [4, 25, 26]. Since these pioneering studies in the late 1980s numerous organisms have been isolated that can grow using Fe(III) and Mn(IV) as electron acceptors. A detailed overview of more than 30 of these strains is given in a recent review . Most organisms that are known to grow through the reduction of Fe(III) or Mn(IV) are relatives of G. metallireducens and fall within the family Geobacteraceae, in the δ subdivision of the class Proteobacteria. This group comprises the genera Geobacter, Desulfuromonas, Desulfuromusa and Pelobacter [24, 27]). These organisms, with the exception of the Pelobacter species, are able to completely oxidise a wide range of organic compounds including acetate when respiring using Fe(III) or Mn(IV); Pelobacter species are more restricted in the range of electron donors utilised (e.g. ethanol, lactate, formate and hydrogen). Some members of the family Geobacteraceae are also able to use aromatic compounds including toluene, phenol and benzoate as electron donors for metal reduction . This is in contrast to S. oneidensis and close relatives in the γ subdivision of the Proteobacteria (a range of Shewanella, Ferrimonas and Aeromonas species) that are generally able to use only a restricted range of small organic acids and hydrogen as electron donors for Fe(III) and Mn(IV) reduction. Other Fe(III)-respiring bacteria that have been characterised recently include Geothrix fermentans, Geovibrio ferrireducens and the related thermophile Deferribacter thermophilus, Ferribacter limneticum, which is unusual in that it can reduce Fe(III) but not Mn(IV), and Sulfurospirillum barnesii (see Sections 4.1 and 4.2). Acidophilic bacteria that are able to grow through the reduction of Fe(III) include Thiobacillus ferrooxidans which uses sulfur as an electron donor for metal reduction . Several hyperthermophilic Archaea and Bacteria have also been shown to grow using Fe(III) as an electron acceptor including Pyrobaculum islandicum, Pyrobaculum aerophilum and Thermotoga maritima. The environmental distribution of Fe(III)-reducing prokaryotes remains poorly understood, but most studies show that members of the family Geobacteraceae are the key components of Fe(III)-reducing communities in subsurface environments (for example [34–38]).
Mechanisms of Fe(III) and Mn(IV) reduction
The mechanisms of Fe(III) reduction, and to a lesser degree Mn(IV) reduction, have been studied in most detail in S. oneidensis and Geobacter sulfurreducens. Indeed research on these organisms has been given added impetus through the availability of the genome sequences (available at http://www.tigr.org), and suitable genetic systems for the generation of deletion mutants for both of these organisms [39, 40]. Although the terminal reductase has yet to be identified unequivocally in either organism, the involvement of c-type cytochromes is implicated in electron transport to Fe(III) and Mn(IV) by several studies [41–46]. In some examples, activities have also been localised to the outer membrane or surface of the cell, consistent with a role in direct transfer of electrons to Fe(III) and Mn(IV) oxides that are highly insoluble at circumneutral pH [41, 47–50]; see Figs. 1A and 2. In addition to the proposed direct transfer of electrons to Fe(III) and Mn(IV) minerals, soluble ‘electron shuttles’ are also able to transfer electrons between metal-reducing prokaryotes and the mineral surface (Fig. 1B). This mechanism alleviates the requirement for direct contact between the micro-organism and mineral. For example, humics and other extracellular quinones are utilised as electron acceptors by Fe(III)-reducing bacteria , and the reduced hydroquinone moieties are able to abiotically transfer electrons to Fe(III) minerals. The oxidised humic is then available for reduction by the micro-organism, leading to further rounds of electron shuttling to the insoluble mineral . Very low concentrations of an electron shuttle, e.g. 100 nM of the humic analogue anthraquinone-2,6-disulfonate, can rapidly accelerate the reduction of Fe(III) oxides , and possibly other insoluble metal oxides such as Mn(IV). The environmental significance of such processes, however, remains to be confirmed. The secretion of soluble electron shuttles by actively respiring Fe(III) and Mn(IV) reducers has also been proposed for both S. oneidensis and G. sulfurreducens, and remains hotly debated for the Geobacter species. Early studies suggested release of a small soluble c-type cytochrome by G. sulfurreducens, but more recent studies have suggested that this protein is not an effective electron shuttle . Studies have also suggested that a small quinone-containing extracellular electron shuttle is released by S. oneidensis, and may also promote electron transfer to Fe(III) and Mn(IV) minerals . Finally, an important new discovery was made recently when it was shown that G. metallireducens synthesised pili and flagella when grown on insoluble Fe(III) or Mn(IV) minerals, but not soluble forms of the metals . These results suggest that Geobacter species sense when soluble electron acceptors are depleted and synthesise the appropriate appendages that allow movement to Fe(III) and Mn(IV) minerals and subsequent attachment.
Reduction of other transition metals
Reduction of V(V)
It has been suggested that the microbial reduction of V(V) could be responsible for the precipitation of the element in anaerobic environments and could also be used to remediate vanadium-contaminated ore-processing waste streams . Despite an early study that showed V(V) reduction by Micrococcus lactilyticus, Desulfovibrio desulfuricans and Clostridium pasteurianum, followed by the observation that the ability to reduce V(V) was widespread amongst soil bacteria and fungi , few recent studies have addressed the bioreduction of this metal. To date, most work has focused on two pseudomonads: P. vanadiumreductans and P. isachenkovii, isolated from a waste stream from a ferrovanadium factory and seawater respectively . Anaerobic cells were able to utilise a wide range of electron donors including hydrogen, sugars and amino acids. V(V) was reduced to blue-coloured V(IV), and possibly further reduced to V(III); the latter indicated by the formation of a black precipitate and by its reaction with the reagent tairon .
Reduction of Cr(VI)
The widespread use of chromium in the metals industries and subsequent contamination problems have led to a lot of interest in this metal. Although trace quantities are required for some metabolic activities, e.g. glucose and lipid metabolism, chromium is considered toxic, and is designated a priority pollutant in many countries. Two oxidation states dominate: Cr(VI) is the most toxic and mobile form commonly encountered, with Cr(III) less soluble and less toxic. Indeed, Cr(III) is considered 1000 times less mutagenic than Cr(VI) . Current treatment involves reduction of Cr(VI) to Cr(III) using chemical reductants at low pH, followed by adjustment to near-neutral pH and subsequent precipitation of Cr(III). Recent studies, however, have shown that micro-organisms can also reduce Cr(VI) efficiently at circumneutral pH, and could be used to treat Cr(VI)-contaminated water.
A wide range of facultative anaerobes are able to reduce Cr(VI) to Cr(III) including Escherichia coli, pseudomonads, S. oneidensis and Aeromonas species (see  for a more exhaustive list and key references). Anaerobic conditions are generally required to induce maximum activity against Cr(VI), but some enzyme systems operate under aerobic conditions, for examples the soluble NAD(P)H-dependent reductases of Pseudomonas ambigua G-1,  and Pseudomonas putida. The former study is also worthy of mention as it showed reduction of Cr(VI) via a Cr(V) intermediate, also noted recently in preparations of Cr(VI) reduced by the membrane fraction of anaerobically grown S. putrefaciens (now S. oneidensis) . Thus, Cr(VI) bioreduction seems to be initiated by a one-electron transfer from the reductase. Obligate anaerobes are also able to enzymatically reduce Cr(VI), and anaerobic growth coupled to Cr(VI) reduction has been reported for a sulfate-reducing bacterium . The reduction of Cr(VI) by sulfate-reducing bacteria is particularly well studied (e.g. ) and has been shown to be catalysed by cytochrome c3. Other studies have also implicated the involvement of cytochromes in Cr(VI) reduction by bacteria: cytochrome c in Enterobacter cloacae and cytochromes b and d in E. coli. Most studies have focused on planktonic cells, but more recent studies have shown that biofilms of sulfate-reducing bacteria also reduce and precipitate Cr(VI). Cr(VI) reduction was thought to be enzymatic in this study; reduction by sulfide was discounted because sulfate reduction was inhibited dramatically in the presence of chromate . Immobilised cells of Bacillus sp.  and Pseudomonas fluorescens LB 300  have also been used to treat Cr(VI)-contaminated water.
Environmental factors that affect Cr(VI) reduction were reviewed recently and include competing electron acceptors, pH, temperature, redox potential and the presence of other metals . A recent study has also demonstrated that the presence of complexing agents can promote Cr(VI) reduction, possibly through protection of the metal reductase by chelation of Cr(III) or intermediates formed . The type of electron donor supplied can also have an effect on the rate and extent of Cr(VI) reduction. Optimal electron donors, in keeping with other dissimilatory metal reduction processes described in this review, are low-molecular-mass carbohydrates, amino acids and fatty acids. Degradation of a range of aromatics including phenol, p-cresol and benzene by P. putida DMP-1 has also been coupled to the reduction of Cr(VI) by E. coli 33456 in co-culture . Similar results have also been reported for a mixed culture of phenol-degrading micro-organisms and the Cr(VI)-reducing E. coli strain . Finally, indirect mechanisms that also promote Cr(VI) reduction in contaminated sediments are catalysed by biogenic sulfide [74, 75] and Fe(II) . Experiments using contaminated sediments from Norman, Oklahoma have, however, suggested that indirect mechanisms may not always be the critical control on Cr solubility, with direct enzymatic Cr(VI) reduction by a consortium of methanogens implicated .
Reduction of Mo(VI)
Although microbial Mo(VI) reduction could play a role in the Mo cycle, for example leading to the concentration of insoluble Mo in anaerobic marine sediments and reduction spots in rocks , comparatively few studies have addressed this process. Early work suggested that Pseudomonas guillermondii and a Micrococcus species could reduce Mo(VI) to molybdenum blue . More recently, similar activities have been identified in cultures of a molybdenum-resistant Enterobacter species . The organism was grown under anaerobic conditions in glucose-containing medium supplemented with 200 mM Mo(VI). Reduction of Mo(VI) was accompanied by a change in colour, as Mo(V) formed and complexed with phosphate in the medium to form methylene blue . The use of metabolic inhibitors suggested that the electrons for Mo(VI) reduction were derived from the glycolytic pathway, and the terminal reductase was downstream from cytochrome b.
The ability to reduce Mo(VI) has also been identified in pre-grown cells of the sulfate-reducing bacterium D. desulfuricans. Here lactate or hydrogen was used as an electron donor for enzymatic metal reduction. In parallel experiments conducted using cultures supplied with sulfate, Mo(VI) was also reduced by biogenic sulfide via an indirect mechanism . Cells of D. desulfuricans immobilised in a bioreactor have also been used to remove Mo(VI) from solution at high efficiency . The organism was unable to grow using Mo(VI) as an electron acceptor , and it is unlikely that actively growing cultures of sulfate-reducing bacteria would play a direct role in reducing high concentrations of Mo(VI) in the environment given the toxicity of molybdate to these organisms . Finally, the thermophilic acidophiles Sulfolobus brierleyi and Sulfolobus acidocaldarius are able to couple S0 oxidation to Mo(VI) reduction . Similar activities have also been noted in T. ferrooxidans.
Reduction of Co(III)
The reduction of Co(III) has received recent attention because radioactive cobalt-60 can be a problematic contaminant at some US Department of Energy sites where radioactive waste has been stored. Co(III) is especially mobile when complexed with EDTA, and recent studies have focused on the ability of Fe(III)-reducing bacteria to retard the mobility of the metal through reduction to Co(II) [84, 85]. The Co(II) formed does not associate strongly with EDTA (it is over 25 orders of magnitude less thermodynamically stable than Co(III)EDTA), and absorbs to soils, offering potential for in situ immobilisation of the metal in contaminated soils. However, Gorby et al.  have demonstrated that reduced Co(II)EDTA can also transfer electrons abiotically to solid Mn(IV) oxides, effectively acting as an electron shuttle between the bacterial cell and the metal oxide. Mn(IV) minerals could, therefore, play a role in maintaining concentrations of mobile Co(III)EDTA in the subsurface. The precise mechanisms of dissimilatory Co(III) reduction remain to be investigated.
Reduction of Pd(II)
The reduction of soluble Pd(II) to insoluble Pd(0) has also attracted much interest as this enzymatic process may be used to recover Pd from industrial catalysts  and may also be used to synthesise nano-scale bioinorganic catalysts of considerable commercial potential . Interest in this area is driven by the widespread use of platinum group metals (PGMs) including Pd in automotive catalytic converters required to reduce gaseous emissions, and problems associated with their recycling. With approximately 5 g of PGM per catalyst the consumption of PGMs together was 2.5×106 oz in 1994, with only 0.4×106 oz recovered . The lifetime of a catalyst is only approx. 50.000 miles although many fail sooner, and future shortages and higher prices may be predicted. Chemical and electrochemical treatments are made difficult by complex solution chemistry.
Initial experiments aimed at reducing and recovering Pd were based on the use of D. desulfuricans because it is active against a wide range of metals including Fe(III), Mn(IV), U(VI), Cr(VI), Tc(VII) via hydrogenase or cytochrome c3. Cells were able to reduce 0.5 mM Pd(II) (as Pd(NH3)4Cl) with a range of electron donors including pyruvate, formate or H2. Although the enzyme responsible for Pd(II) reduction has not been identified, the involvement of a periplasmic hydrogenase is implicated by the use of hydrogen as electron donor and inhibition by treatment with 0.5 mM Cu2+. Transmission electron microscope studies in combination with energy dispersive X-ray microanalysis (EDS) confirmed precipitation in the periplasm, with X-ray diffraction studies confirming reduction to Pd(0) (see Fig. 3).
More recent studies have focused on the recovery of Pd(II) at a range of pH values and also from acid (aqua regia) leachates from spent automotive catalysts . Inhibition by chloride ions was reported and may necessitate optimisation of leaching from catalysts to minimise the formation of PdCl42− prior to bioreduction and recovery. Delivery of reducing power to an immobilised biocatalyst has also been studied in a novel electrobioreactor . A biofilm of D. desulfuricans was immobilised on a Pd–Ag membrane that transported atomic hydrogen to the cells, minimising loss of gaseous hydrogen. Pd(0) reduction and recovery was fed by H2 on demand and the authors proposed that this system was both clean and economic, with no generation of secondary wastes. Interestingly, Pd(0) recovered in the electrobioreactor proved a better catalyst that its chemical counterpart, as determined by hydrogen liberation from hypophosphite , and reduction of several target organics (Macaskie, personal communication).
Reduction of Au(III) and Ag(I)
Early studies suggested that c-type cytochromes of the Fe(III)-reducing bacterium G. metallireducens were able to transfer electrons to soluble Au(III) . In subsequent studies a wide range of dissimilatory reducing Fe(III)-reducing Bacteria and Archaea, including the hyperthermophilic Archaea P. islandicum and P. furiosus, the hyperthermophilic bacterium T. maritima, and the mesophilic bacteria Shewanella alga and G. ferrireducens were shown to reduce Au(III) (as gold chloride) to insoluble Au(0) . Organic electron donors were not utilised, and the site of precipitation of Au(0) varied between organisms. Most strains tested precipitated Au on the surface of the cell, while G. ferrireducens precipitated Au(0) in the periplasm. The ability to reduce Au(III) seems to be species-specific, and closely related organisms with similar activities against a range of metals have differing activities against Au(III) . For example, unlike P. islandicum, a close relative, P. aerophilum, is unable to reduce Au(III). The obligate requirement for hydrogen as an electron donor would suggest that a hydrogenase is involved in Au(III) reduction, and given the direct reduction of other metals by hydrogenase  it is tempting to hypothesise that hydrogenases may play a direct role in Au(III) reduction. This is yet to be tested. The precise involvement of Fe(III)-reducing bacteria in the deposition of Au ores also warrants attention, as it has been argued that these organisms are present in high- and moderate-temperature sedimentary environments where Au deposits have been recovered . The microbial reduction of Ag(I) has also been studied, but in little detail. Early reports noted that the reduction of Ag(I) may account for resistance to silver in some micro-organisms , but more recent studies have uncovered alternative strategies for resistance to Ag(I) in organisms isolated from hospital burns wards where silver may be used as a biocide . Several studies, including that of Fu et al. , have shown biosorption of Ag(I) to the surface of cells (in this case a Lactobacillus sp.), followed by reduction to Ag(0). Here the mechanism for Ag(I) reduction remains unknown.
Reduction of Hg(II)
A well-studied metal resistance system is encoded by genes of the mer or mercury resistance operon. Here, Hg(II) is transported into the cell via the MerT transporter protein, and detoxified by reduction to relatively non-toxic volatile elemental mercury by an intracellular mercuric reductase (MerA). This system is described in detail in a chapter by Ann Summers in this volume , and will be discussed only very briefly in this review. The biotechnological potential of this process has also been described recently, focusing on the use of mercury-resistant bacteria and the proteins that they encode . Applications include the bioremediation of Hg-contaminated water and the development of Hg(II)-detecting biosensors . Finally, in addition to the MerA-mediated mechanism of Hg reduction other enzymes are also able to reduce Hg(II). A novel Fe2+-dependent mechanism for mercury reduction has been characterised in the membrane fraction of T. ferrooxidans, which may involve cytochrome c oxidase , and c-type cytochromes of G. metallireducens also reduce Hg(II) . In no cases, however, has the reduction of Hg(II) been shown to support microbial growth.
Reduction of metalloids
Reduction of As(V)
Arsenic exists as the arsenate anion (As(V)) in oxic environments, and binds very strongly to sediments. Under anaerobic conditions, As(V) is reduced to the more mobile and toxic As(III). The mobilisation of arsenic in sediments poses a potential threat to the lives of millions worldwide, for example those in the Ganges–Meghna–Brahmaputra delta plain who rely on contaminated wells for their drinking water . In addition, As(V) can be the dominant electron acceptor for carbon oxidation in some environments such as Mono lake, California . As with several other metal reduction processes discussed, As(V) reduction can be catalysed by biotic or abiotic mechanisms. Early evidence for a microbial role in As(V) reduction in sediments has been reviewed extensively by Oremland and Stolz . These include inhibition of As(V) reduction by treatment of sediments with heat and a wide range of antimicrobials. In addition, As(V) reduction was inhibited by air but enhanced by an atmosphere of hydrogen suggesting the involvement of anaerobic metal-reducing bacteria.
Indeed several organisms capable of growing through dissimilatory reduction of As(V) have now been isolated. Chrysiogenes arsenatis is a strict anaerobe that was isolated from wastewater from a gold mine . This organism contains a periplasmic arsenate reductase consisting of two subunits (masses 87 and 29 kDa) which contain Mo, Fe, S and Zn cofactors . Sulfurospirillum arsenophilum (previously strain MIT-13) is a Gram-negative vibroid microaerobic sulfur-reducing bacterium isolated from arsenic-contaminated watershed sediments in eastern Massachusetts. It is a very close relative of S. barnesii which can also reduce As(V), and has a broad activity against metals including Fe(III) and Se(VI) (see Sections 2.1 and 4.2). A recent review has mentioned the preliminary purification and characterisation of an arsenate reductase from this organism, which constituted a trimeric complex of mass 120 kDa, consisting of subunits of 65 kDa, 31 kDa and 22 kDa . A Gram-positive sulfate-reducing bacterium Desulfitomaculum auripigmentum has also been described that reduces As(V) followed by sulfate, resulting in the formation of orpiment (As2S3) . Reduction of As(V) to As(III) by the ArcC reductase also forms the basis of a well-studied microbial arsenic resistance mechanism, preceding efflux of As(III) from the cell. This mechanism has been reviewed in detail in this journal recently  and will not be discussed further. It should be noted, however, that ArcC-mediated As(V) reduction does not support microbial growth, and there is currently little evidence linking this mechanism of As(V) reduction to the biogeochemical cycling of arsenic. Finally, on this note, it is worth mentioning that a new organism has recently been isolated that is able to use As(III) as an electron donor for aerobic growth  closing the biological As cycle.
Reduction of Se(VI) and Se(IV) and other group VIB elements
Reduction of toxic Se(VI) and Se(IV) to relatively unreactive Se(0) results in its removal from water. Several studies have demonstrated that these transformations can be catalysed by microbes (for an overview see ). For example, the ability to reduce Se(VI) is widespread in sediments, with biological reduction unequivocally demonstrated in 10 out of 11 sediment types . Also, Se(VI) is not reduced chemically under physiological conditions of pH and temperature, and Se(VI) reduction is inhibited by autoclaving of sediments. Se(VI) reduction is also inhibited by chromate and tungstate, but not by molybdate, an inhibitor of sulfate reducers [101, 102]. These results suggest the involvement of a molybdenum-containing enzyme, recently backed up by the study of Schroder et al.  who identified a Mo-containing membrane-bound Se(VI) reductase in Thauera selanatis as described below.
Organisms that are known to reduce Se(VI) enzymatically include W. succinogenes, D. desulfuricans, Pseudomonas stutzeri, E. cloacae and E. coli. In these examples, Se(VI) reduction does not support growth, and seems to be incidental to the physiology of the organism. In at least one organism (E. coli), the involvement of broad-specificity nitrate reductases is implicated by biochemical studies . In addition to these rather non-specific reactions, specialist organisms are known to conserve energy through Se(VI) reduction including T. selanatis and S. barnesii (originally Geospirillum barnesii strain SES-3 ) and two other Bacilli (Bacillus arsenicoselanatis and B. selenitireducens), both isolated from Mono lake, California . Of these four model organisms, the mechanism of Se(VI) reduction is best understood in T. selanatis. A periplasmic complex of approximately 180 kDa (with subunits of masses 96, 40 and 23 kDa) has been characterised, and shown to contain Mo, Fe and acid-labile sulfur. Specificity for Se(VI) is high, with a Km of 16 μM. The enzyme was unable to reduce nitrate, nitrite, chlorate, chlorite or sulfate. Earlier studies had suggested that the enzyme that catalysed further reduction of Se(IV) (selenite) to elemental selenium in the organism was a component of the nitrate-respiratory system, possibly a periplasmic nitrite reductase . It was also noted in this study that T. selanatis was unable to grow through the reduction of selenite. Biochemical studies are not so advanced in S. barnesii, although the enzyme activity contrasts with that characterised in T. selanatis as it is localised in the membrane fraction, and may have a wider substrate specificity . Whole cells of S. barnesii can use a very wide range of electron acceptors (including Se(VI) in addition to As(V), nitrate, fumarate and thiosulfate), and despite the proposed wide specificity of the Se(VI) reductase, the authors of this paper suggest that a suite of specific terminal reductases are required for growth on the different electron acceptors. This hypothesis is based on the observed differential cytochrome content and cytochrome activity when grown on alternative electron acceptors; at least three different b-type and two different c-type cytochromes have been detected .
Tellurite (TeO32−) reduction has also been studied in several organisms, mainly in the context of resistance to this toxic oxyanion. Indeed, the antibacterial properties of Te(IV) have been known for more than 70 years; in the pre-antibiotic era Te(IV) was used to treat a range of bacterial infections and Te(IV) remains an ingredient of several selective media (e.g. for verocytotoxigenic E. coli O157). Basal levels of resistance to toxic Te(IV) have been attributed to the activity of a membrane-bound nitrate reductase in E. coli. An additional Te(IV) reductase was detected in the soluble fraction of anaerobically grown cells. Growth using Te(IV) as an electron acceptor was also reported in an engineered strain overexpressing nitrate reductase, but was not thought to be physiologically relevant in wild-type cells of E. coli. Rhodobacter sphaeroides has also been reported to reduce Te(IV) (as well as oxyanions of Se), with a requirement for a functional photosynthetic electron transfer chain under photosynthetic conditions, or cytochromes bc1 and c2 when grown aerobically . Again, metal reduction was discussed in the context of resistance to the metalloids. Finally, plasmids are known to encode several distinct resistance determinants for Te(IV), and again Te(IV) reduction is implicated as the resistance mechanism as elemental Te is laid down in Te-resistant bacteria . Other mechanisms of resistance may be important, however, involving cysteine-metabolising enzymes and methyl transferases . Finally, Te(IV) (and Se(VI)/Se(IV)) reduction and precipitation by sulfate-reducing bacteria have also been reported, in the order Te(IV)>S(VI)>Se(IV), which is in contrast to that predicted by the redox potentials alone . To date there have been no reports of microbial growth coupled to the reduction of Te(IV) by non-genetically engineered micro-organisms.
One last member of group VIB that warrants brief discussion is polonium. 210Po is the terminal member of the 238U decay series, produced by the decay of 210Pb via 210Bi. 210Po has attracted recent attention, accounting for all α-emitting activity in some samples of groundwater from the Central Florida Phosphate District , and has also been identified as a troublesome isotope contaminating and fouling drilling equipment during oil exploration in areas rich in uranium ores . Little is known about the geochemistry of Po, but there is evidence that factors affecting the sulfur cycle may affect Po availability. Indeed, Cherrier et al.  demonstrated the microbial uptake of Po into the cellular pool, indicating the potential for assimilative Po reduction and a biochemical role analogous to that of S. Given that a range of anaerobic bacteria are also able to reduce Se and Te oxyanions, it is likely that dissimilatory reduction of Po is possible.
Reduction of actinides and fission products
Although there are natural sources of radioactivity, the release of anthropogenic radionuclides into the environment is significant and a subject of intense public concern. Large quantities of radionuclides were released as a consequence of nuclear weapons testing in the 1950s and 1960s, while the controlled discharge of process effluents produced by industrial activities allied to the generation of nuclear power continues to introduce radioactivity in the environment today. Indeed, research programmes aimed at remediating large areas of land contaminated by radionuclides in the USA, at so-called superfund sites, have resulted in significant advances in the understanding of the mechanisms of metal and radionuclide reduction in the subsurface (for examples see http://www.lbl.gov/NABIR/). Indeed, because many radionuclides of concern are both redox-active and less soluble when reduced, bioreduction offers much promise for controlling the solubility and mobility of target radionuclides in contaminated sediments, e.g. the reduction of U(VI) (the uranyl ion; UO22+) to U(IV) (uraninite; UO2) [3, 118] or the reduction of the fission product Tc(VII) (the pertechnetate ion; TcO4−) to Tc(IV) (TcO2) .
Reduction of U(VI)
The first demonstration of dissimilatory U(VI) reduction was by Lovley and coworkers  who reported that the Fe(III)-reducing bacteria G. metallireducens (previously designated strain GS-15) and S. oneidensis (formerly A. putrefaciens and then S. putrefaciens) can conserve energy for anaerobic growth via the reduction of U(VI). It should be noted, however, that the ability to reduce U(VI) enzymatically is not restricted to Fe(III)-reducing bacteria. Other organisms including a Clostridium sp.  and the sulfate-reducing bacteria D. desulfuricans and Desulfovibrio vulgaris also reduce uranium, but are unable to conserve energy for growth via this transformation. To date, D. vulgaris remains the only organism in which the enzyme system responsible for U(VI) reduction has been characterised in detail. Purified tetrahaem cytochrome c3 was shown to function as a U(VI) reductase in vitro, in combination with hydrogenase, its physiological electron donor . In vivo studies using a cytochrome c3 mutant of the close relative D. desulfuricans strain G20 confirmed a role for cytochrome c3 in hydrogen-dependent U(VI) reduction, but suggested additional pathways from organic electron donors to U(VI), that bypassed the cytochrome .
More recent studies have identified a homologous cytochrome (PpcA), a trihaem periplasmic cytochrome c7 of the Fe(III)-reducing bacterium G. sulfurreducens that may also play a role in U(VI) reduction . The protein was able to reduce U(VI) in vitro, while a ppcA deletion mutant supplied with acetate as an electron donor had lower activity against U(VI) . Additional (if indirect) evidence linking the activity of this periplasmic protein with U(VI) reduction in vitro included the precipitation of the reduced product U(IV) in the periplasm (see Fig. 4), and the lack of impact of protease treatment of whole cells on the ability to reduce U(VI) . This final result is particularly important, as it implies that U(VI) and Fe(III) are reduced by different mechanisms in G. sulfurreducens. U(VI) would seem to be reduced in the periplasm, while the reduction of insoluble Fe(III) oxides was inhibited dramatically by protease treatment, presumably due to removal of surface-bound cytochromes required for reduction of the extracellular electron acceptor. The mechanism of U(VI) reduction by a S. putrefaciens strain has also been investigated . A novel screening method was used to identify mutants that were unable to reduce U(VI). Evidence was presented to suggest that the mechanism of U(VI) reduction was distinct from those of Fe(III) and Mn(IV) reduction, but may share components of the nitrite-reducing pathway .
Reduction of Np(V) and Pu(IV)
Although 238U remains the priority pollutant in most medium- and low-level radioactive wastes, other actinides including 230Th, 237Np, 241Pu and 241Am can also be present [116, 124]. Th(IV) and Am(III) are stable across most Eh values encountered in radionuclide-contaminated waters (Fig. 5) but the potentials for Pu(V)/Pu(IV) and Np(V)/Np(IV), in common with that of U(VI)/U(IV), are more electropositive than the standard redox potential of ferrihydrite/Fe2+ (approximately 0 V ). Thus, Fe(III)-reducing bacteria have the metabolic potential to reduce these radionuclides enzymatically, or via Fe(II) produced from the reduction of Fe(III) oxides. This is significant because the tetravalent actinides are amenable to bioremediation due to their high ligand-complexing abilities , and are also immobilised in sediments containing active biomass . Thus, although it is possible for Fe(III)-reducing bacteria to reduce and precipitate actinides in one step, e.g. the reduction of soluble U(VI) to insoluble U(IV) (see above), some transformations do not result in direct formation of an insoluble mineral phase but in the formation of a cation more amenable to bioprecipitation. This is illustrated when considering highly soluble Np(V) (NpO2+), which was reduced to soluble Np(IV) by S. putrefaciens, with the Np(IV) removed as an insoluble phosphate biomineral by a phosphate-liberating Citrobacter sp. . Also, some studies have suggested that the reduction of Pu(IV) to Pu(III) can be achieved by Fe(III)-reducing bacteria, although the Pu(III) was reported to reoxidise spontaneously . Although this may lead to solubilisation of sediment-bound Pu(IV), it will yield a trivalent actinide that is also amenable to bioremediation using a range of microbially produced ligands . The biochemical basis of these transformations remains uncharacterised.
Reduction of Tc(VII)
The fission product technetium is another long-lived radionuclide that is present in nuclear waste and has attracted considerable recent interest. This is due to a combination of its mobility as the soluble pertechnetate ion (Tc(VII); TcO4−), bioavailability as an analogue of sulfate and a long half-life (2.13×105 years) . Like Np(V), Tc(VII) has weak ligand-complexing capabilities and is difficult to remove from solution using conventional ‘chemical’ approaches. Several reduced forms of the radionuclide are insoluble, however, and metal-reducing micro-organisms can reduce Tc(VII) and precipitate the radionuclide as a low-valency oxide.
Although microbial metabolism was known to decrease the solubility of Tc from earlier studies [129, 130], Lloyd and Macaskie were the first to unequivocally demonstrate direct enzymatic reduction of Tc(VII) by micro-organisms . In this study, a novel phosphorimager technique was used to confirm reduction of the radionuclide by S. putrefaciens and G. metallireducens, with similar activities subsequently detected in laboratory cultures of R. sphaeroides, Paracoccus denitrificans, some pseudomonads , E. coli and a range of sulfate-reducing bacteria [64, 132, 133]. Other workers have used this technique to show that T. ferrooxidans and Thiobacillus thiooxidans and the hyperthermophile P. islandicum are also able to reduce Tc(VII). It should be stressed that Tc(VII) reduction has not been shown to support growth in any of these studies, and seems to be a fortuitous biochemical side reaction in the organisms studied to date. Finally, X-ray absorption spectroscopy studies have recently identified insoluble Tc(IV) as the final oxidation state produced when Tc(VII) is reduced enzymatically by G. sulfurreducens, E. coli (Lloyd and Sole, unpublished) and S. putrefaciens. Recent studies have also shown that Tc(VII) can be reduced via indirect microbial processes via, for example, biogenic sulfide , Fe(II)  or U(IV) . Tc(VII) reduction and precipitation by biogenic Fe(II) is particularly efficient, and may offer a potentially useful mechanism for the remediation of Tc-contaminated sediments containing active concentrations of Fe(III)-reducing bacteria .
The biochemical basis of Tc(VII) reduction has been best studied in E. coli. Initial studies demonstrated that anaerobic, but not aerobic, cultures of E. coli reduced Tc(VII) with the reduced radionuclide precipitated within the cell . Results obtained from studies conducted with wild-type cells and 34 defined mutants defective in the synthesis of regulatory or electron transfer proteins were used to construct a model for Tc(VII) reduction by E. coli (Fig. 6). The central tenet of this model is that the hydrogenase 3 component of formate hydrogenlyase (FHL) catalyses the transfer of electrons from dihydrogen to Tc(VII). According to this model, the formate dehydrogenase component (FdhH) is required only if formate, or a precursor, is supplied as an electron donor for Tc(VII) reduction in place of hydrogen. This model has been validated by the observations that a mutant unable to synthesise hydrogenase 3 was unable to reduce Tc(VII) when either hydrogen or formate was supplied as an electron donor .
The identification of hydrogenase 3 of FHL as the Tc(VII) reductase of E. coli opened up the way for a programme to screen for organisms with naturally enhanced activities against Tc(VII). Several organisms documented to have naturally high activities of FHL or uptake hydrogenase were tested, resulting in the identification of several strains of sulfate-reducing bacteria that were able to couple the oxidation of formate or hydrogen to Tc(VII) reduction . Rates of reduction in some strains were approximately 64-fold greater than those recorded in anaerobic cultures of E. coli. D. desulfuricans and related strains  were also able to utilise formate as an efficient electron donor for Tc(VII) reduction. This is consistent with the existence of a rudimentary FHL complex (consisting of a formate dehydrogenase coupled to a hydrogenase via a cytochrome) located in the periplasm of these strains . Accordingly, the site of reduced Tc precipitation was identified as the periplasm in D. desulfuricans, and more recent studies have confirmed a role for a periplasmic Ni–Fe hydrogenase in Tc(VII) reduction by a relative in the δ subclass of the Proteobacteria, the sulfate-reducing bacterium Desulfovibrio fructosovorans. Subsequent studies on the development of a bioprocess to treat Tc(VII)-contaminated water have focused on the use of immobilised cells of sulfate-reducing bacteria such as D. desulfuricans which are robust and capable of treating low concentrations of Tc(VII) against a high background of contaminating nitrate ions, which is often noted in nuclear waste [132, 137].
Although the full environmental relevance of metal reduction processes has only recently become apparent, rapid advances in the understanding of these important biotransformations have been made. However, we still have much to learn about the precise mechanisms involved, and the full impact of such reactions on a range of biogeochemical cycles. Given the availability of genomic sequences for key metal-reducing micro-organisms, new post-genomic and proteomic approaches and the possibility of combining these tools with advanced techniques from other branches of science and technology (e.g. isotopic, spectroscopic and computational tools), rapid advances in these areas are predicted.
The author thanks the UK Natural Environment Research Council (NERC) and the Natural and Accelerated Bioremediation Research (NABIR) programme of the US Department of Energy for financial support through Grants NER/A/S/2001/00960 and DE-FG02-02ER63422 respectively.