| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Microbial Ecology Group, Institute of Plant Biology, University of Zürich, Zollikerstrasse 107, CH 8008, Zürich, Switzerland
Correspondence
Janine Kessi
janine.kessi{at}access.unizh.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; O'Toole & Raisbeck, 1998; Läuchli, 1993). Since many species of bacteria are able to transform the highly toxic selenium oxyanions selenate (
) and selenite (
) to the non-toxic elemental selenium (Se0), bioremediation strategies based on bacterially mediated reduction of Se oxyanions have been proposed. Dissimilatory reduction of selenite has been reported in many species of bacteria (Gerrard et al., 1974; Tomei et al., 1992, 1995; Losi & Frankenberger, 1997; Dungan et al., 2003), although the mechanism of reduction is poorly understood. As purple non-sulfur bacteria have been shown to be highly tolerant to transition metal oxyanions (Moore & Kaplan, 1992), it seemed interesting to investigate the mechanism of reduction of selenite in these organisms. Rhodospirillum rubrum and Rhodobacter capsulatus, which are phylogenetically distant within the alphaproteobacteria (Woese et al., 1984; Karr et al., 2003) and show diverse resistance toward transition metal oxyanions (Moore & Kaplan, 1992), were used as model organisms.
Various enzymic systems have been proposed to catalyse the reduction of selenite in bacteria. In Thauera selenatis, DeMoll-Decker & Macy (1993) have suggested that the reaction might be catalysed by a periplasmic dissimilatory nitrite reductase. They report that mutants lacking periplasmic nitrite reductase activity are unable to reduce either nitrite or selenite; however, the location of the mutation was not determined. Dissimilatory nitrate reduction and nitrous oxide reduction in purple non-sulfur bacteria are well documented (Ferguson et al., 1987). In contrast, dissimilatory nitrite reductase activity has been reported in only a few species, including Rhodopseudomonas palustris (Preuss & Klemme, 1983), some strains of Rhodobacter sphaeroides (Richardson et al., 1994), and some strains of Rb. capsulatus (Schwintner et al., 1998). In each one of these organisms, a copper-type nitrite reductase has been identified, suggesting that this enzyme may be generally present in the purple non-sulfur bacteria. Interestingly, Schwintner et al. (1998) have reported the presence of a nitrite reductase gene (nirK) in a species of Rb. sphaeroides which does not express nitrite reductase activity (Rb. sphaeroides 2.4.1). This suggests that the activity of this enzyme can be repressed under standard laboratory culture conditions. Confirming this hypothesis, a strain of Rb. capsulatus (BK5) only expresses nitrite reductase activity after repetitive subculture in the presence of 1 mM nitrate and a reduced carbon source (butyrate) (Richardson et al., 1994). In the present work, the influence of nitrite, separately and in combination with selenite, on cell growth and on the kinetics of selenite reduction were investigated. Considering that selenite has been reported to interfere with the metabolism of sulfite in Clostridium pasteurianum (Harrison et al., 1980), and that an inducible dissimilatory sulfite reductase from this organism efficiently reduces selenite (40 % of the sulfite reductase activity) (Harrison et al., 1984), it seemed interesting to determine whether sulfite and selenite metabolism also interfere in the purple non-sulfur bacteria or whether a sulfite reductase is functioning as a selenite reductase. Sulfite reductase activity has been measured in various species of photosynthetic bacteria (Peck et al., 1974). In the present work, the influence of sulfite, separately and in combination with selenite, on the cell growth and on the kinetics of selenite reduction was investigated.
Glutathione (GSH) has been proposed by Ganther (1966, 1968, 1971) to be involved in the reduction of selenite in mammalian tissues. Ganther (1971) has described the formation of selenodiglutathione (GS-Se-SG), which forms when selenous acid reacts with glutathione (H2SeO3+4GSH
GSSG+GS-Se-SG+3H2O), and has shown that GS-Se-SG is reduced to Se0 by the glutathione reductase. These reactions have been confirmed by Rabenstein & Tan (1988), who have conducted in vivo 77Se NMR spectroscopic analysis of Escherichia coli cultures amended with selenite. Once selenite reduction commences, a moderate signal is seen in the range expected for selenotrisulfide (550–700 p.p.m.). It weakens in intensity over a few hours with concomitant formation of a red precipitate in the culture medium that is characteristic of elemental selenium. Glutathione is present at millimolar levels in the cyanobacteria and the alpha-, beta- and gammaproteobacteria (Newton & Fahey, 1989). The high reactivity of glutathione with selenite (Ganther, 1971; Kessi & Hanselmann, 2004) makes glutathione an obvious candidate for participating in the dissimilatory reduction of selenite in organisms which are able to synthesize this metabolite. Supporting this hypothesis, it has been reported that the synthesis of the glutathione reductase is induced in cultures of Rb sphaeroides and E. coli amended with selenite (Bébien et al., 2001, 2002, respectively). As GS-Se-SG has also been shown to be a highly efficient oxidant of reduced bacterial thioredoxin (Björnstedt et al., 1992), thioredoxin and thioredoxin reductase may also be involved in the reduction of selenite in bacteria that are able to synthesize glutathione. Supporting this hypothesis, Bébien et al. (2002) have reported a large induction of the synthesis of thioredoxin and thioredoxin reductase in E. coli amended with selenite. According to their results, glutathione and glutathione reductase and/or thioredoxin and thioredoxin reductase may be involved in the dissimilatory reduction of selenite in organisms containing high levels of glutathione.
A comparison of the features of the abiotic (chemical) and the bacterial reduction of selenite with glutathione lends additional support to the involvement of glutathione in the biological reduction of selenite. Both the abiotic reduction of selenite with glutathione (Kessi & Hanselmann, 2004) and the reaction mediated by E. coli (Bébien et al., 2002) have been shown to produce superoxide anions. In addition, particles produced both chemically and biologically show the same size, spherical structure and characteristic orange-red colour (Kessi & Hanselmann, 2004).
In the present work, we used a physiological approach to investigate the mediation of glutathione in the dissimilatory reduction of selenite in Rs. rubrum. Buthionine sulfoximine (BSO, S-n-butyl homocysteine sulfoximine), which specifically inhibits the synthesis of 
-glutamylcysteine, an intermediary product of glutathione synthesis (Griffith, 1981; Griffith & Meister, 1979), was used to depress the intracellular glutathione level. With prokaryotic -glutamylcysteine synthetase, however, a reduced rate of inactivation and a decreased initial binding affinity of BSO were observed as compared with those of mammalian isoforms (Kelly et al., 2002). Therefore, relatively high BSO concentrations (0·5–3·0 mM) were needed to significantly decrease the concentration of glutathione in the bacterial cells.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Growth of cultures.
Rs. rubrum S1 (DSM 467) was grown under anoxic conditions in the light, as described previously (Kessi et al., 1999). Growth and selenite reduction kinetics were not modified when the cells were inoculated under anoxic conditions compared with inoculation in fresh culture medium and incubation in the dark before exposure to light. Rb. capsulatus B10 (ATCC 33303) was inoculated into fresh culture medium and left for 20 h at 30 °C in the dark in order to consume all oxygen before exposure to light. For analysis of headspace gases, cultures were grown in flasks with Minivert valves (Supelco).
To investigate the effects of nitrite and sulfite on growth, cultures were amended with selenite, nitrite or sulfite, or with a combination of selenite and nitrite or selenite and sulfite. The concentration of the amendments was selected so that neither cell growth nor selenite reduction was completely inhibited. Selenite and nitrite were added to the cultures for final concentrations of 0·5 mM. Sulfite was injected into the cultures for a final concentration of 0·3–0·4 mM. Individual stock solutions of nitrite, sulfite and selenite (100 mM each) were prepared in degassed, sterile culture medium under anoxic, aseptic conditions and added to the cultures so as to avoid air contamination. Such precautions ensured that culture growth was unaffected by contamination with oxygen and were required to preserve the chemical stability of added sulfite.
The in vivo effect of glutathione was investigated by chemically interfering with glutathione synthesis. BSO, a specific inhibitor of the -glutamylcysteine synthetase, was used to depress intracellular glutathione levels. Cultures were amended with BSO at concentrations from 0·5 to 3·0 mM, which efficiently decreased glutathione level but did not suppress either cell growth or selenite reduction. Stock solutions of BSO (250 mM) were prepared in degassed, sterile culture medium under anoxic, aseptic conditions and added to the cultures so as to avoid air contamination, and the culture was shaken immediately. BSO was injected inside the culture medium (not on its surface). This precaution avoided cell lysis after BSO injection.
Protein and selenite determination.
Determination of cell protein and selenite concentrations in the culture medium were performed as described previously (Kessi et al., 1999). All samples for selenite analysis were centrifuged for 10 min at 1000 g to improve separation of the organic and water phases following cyclohexane extraction of the selenium complex. Calibration curves were obtained with samples containing 0–200 nmol sodium selenite and exhibited linearity between absorption and selenite concentration (r2=0·999). The calibration curves were not modified by the addition of sulfite or nitrite at concentrations equal to the selenite amendments, nor by the addition of BSO at up to six times the selenite concentration. No modification of selenite concentration was measured after incubation of selenite (0·5 mM) with sulfite (0·3–0·4 mM) or nitrite (0·5 mM) in 50 mM Tris, pH 7·0, at room temperature for 24 h under anoxic conditions. Measurements were made in duplicate.
Nitrite analysis.
Nitrite was determined using the diazo dye method of Bratton et al. (1939). Fifty microlitres of cell-free culture medium containing 0–25 nmol 
were added to 950 µl of a 7 mg ml–1 solution of Spectroquant-NO2 in double-distilled H2O. The samples were incubated at 30 °C for 10 min and absorbance measured at 525 nm. All measurements were made in duplicate. Calibration curves were obtained with samples containing 0–25 nmol NaNO2 and exhibited linearity between absorption and nitrite concentration (r2=0·999). The calibration curves were not modified by addition of selenite at concentrations equal to the nitrite amendments.
Sulfite determination.
Sulfite concentrations were determined by oxidation with sulfite oxidase using potassium ferricyanide as electron acceptor. The reaction mixture contained the following: 520–720 µl double-distilled H2O, 15 µl 50 mM potassium ferricyanide, 50 µl 1 M Tris/HCl (pH 7·5), 10 µl 10 mM disodium EDTA and 200–400 µl of sample. All were mixed in a 1·5 ml plastic cuvette (1 cm path length) under anoxic conditions. All chemical solutions were degassed with an aspirator pump for about 40 min before use. Absorbance was measured at 420 nm, before and 20 min after the addition of 5 µl sulfite oxidase (about 10 U ml–1). All measurements were made in triplicate.
Sulfite concentrations were calculated as the difference between the two absorbance values using calibration curves; there was good linearity between sulfite concentration (0–200 nmol sulfite) and absorbance (r2=0·998–0·999). Sulfite concentrations were determined in both the presence and the absence of cells because of the observed affinity of sulfite for the cell surface during exponential growth. Ferricyanide absorption was determined in the presence and absence of sulfite oxidase; the difference in absorbance at 420 nm was corrected for bacterial iron reduction (Dobbin et al., 1996). It should be noted that the concentration of sulfite in the cultures was always lower than that calculated for the amendment (approx. 0·1–0·2 mM), suggesting that some sulfite oxidation occurred despite efforts to avoid exposure of solutions to oxygen. Sulfite oxidase exhibited a high specificity toward sulfite. No activity was expressed in the presence of selenite.
Determination of glutathione (GSH).
Cells were concentrated by centrifugation (4000 g for Rs. rubrum and 20 000 g for Rb. capsulatus) at 4 °C for 10 min and resuspended in 50 mM Tris (pH 7·0) to a concentration of 1·5–2·0 grams wet cells per five millilitres Concentrated cells kept on ice in glass tubes were disrupted using sonication. Three sonication steps of 60 s each at an intensity of 75 % in a Branson sonifier II (Model 250, Skan AG, Basel, Switzerland) were applied. Broken cells were centrifuged at 100 000 g at 4 °C for 1 h. Enzymic activity of the supernatant was stopped by treatment in a boiling water bath for 2 min. Denatured macromolecules were then centrifuged at 10 000 g at room temperature for 10 min; the supernatant solution was analysed for glutathione. The sum of reduced and oxidized glutathione was determined using glutathione reductase and Ellman's reagent, DTNB [5,5'-dithiobis(2-nitrobenzoic acid)], as described by Akerboom & Sies (1981). The reaction mixture contained the following: 410–880 µl 0·1 M phosphate buffer (pH 7·0), 50 µl of 4 mg ml–1 NADPH in 0·5 % NaHCO3, 20 µl of 1·5 mg ml–1 DTNB in 0·5 % NaHCO3, 30–500 µl culture supernatant, and 20 µl glutathione reductase (6 U ml–1). Samples containing all reagents, but no enzyme, were equilibrated at 25 °C for 30 min in the thermostatic compartment of the spectrophotometer; the reaction was started by addition of the enzyme. The rate of the reaction was recorded for 2 min at 412 nm in a spectrophotometer (Uvikon 860, Kontron, Zürich, Switzerland). In the samples obtained from stationary-phase cultures amended with BSO and containing low levels of glutathione, the same method was used except that the reaction mixture was incubated for 30 min at 25 °C in the dark and the final absorbance at 412 nm was measured. The sum of the reduced and oxidized glutathione concentrations (GSH+GSSG) was computed using calibration measurements with standards of the oxidized form of glutathione (GSSG). A good linearity of reaction rate versus GSSG concentration was obtained between 1·0 and 3·5 µM GSSG (r2=0·998) in the kinetic experiments and between 0·02 and 0·2 nM in the static measurements (r2=0·992). GSSG-reductase activity was not affected by the presence of BSO.
Identification of volatile selenium compounds.
Gas chromatographic separation of volatile sulfur and selenium compounds was achieved on a DANI 86 HT gas chromatograph (DANI, Instrumentazione Analitica, Monza, Italy) equipped with a sulfur-specific flame photometric detector (FPD). The temperature of both injector and detector was 250 °C. The column was a J&W GS-Q 30 mx0·53 mm internal diameter porous layer open tubular (PLOT) fused-silica column (Brechbuehler, Schlieren, Switzerland). The 35 °C initial temperature of the column was maintained for 3 min followed by a ramp of 17·5 °C min–1 to 225 °C. This final temperature was maintained for 7 min. Helium was used as the carrier gas at a flow rate of 4·2 ml min–1 (Fritz & Bachofen, 2000). A volume of 0·5–1·5 ml of the headspace of the cultures was injected onto the column. Peak identification was made by comparing the retention times of analytes to those of commercially available standards.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Growth of Rs. rubrum.
In Rs. rubrum, the effects of added ions were assessed by observing increases in lag time and/or decreases in growth rate and cell density attained at stationary phase. The results are presented in Fig. 1(a).
|
In the presence of selenite alone (0·5 mM), nitrite alone (0·5 mM) or sulfite alone (0·3–0·4 mM), the growth rate was not significantly modified compared to reference cultures free of any amendment. Also, the final cell density was preserved in the presence of sulfite, but decreased by 12 and 50 % in the presence of nitrite and selenite, respectively, compared to control cultures free of any amendment.
Amendment with a combination of nitrite and selenite increased the lag time to 57 h and decreased the growth rate by 40 % compared to cultures amended with selenite alone. The effects of a combination of nitrite and selenite were larger than the sum of the effects of each ion separately.
Interestingly, cultures of Rs. rubrum amended with a combination of selenite (0·5 mM) and sulfite (0·3–0·4 mM) reached a cell density corresponding to 70 % of the control cultures, which was 20 % more than that attained in cultures grown in the presence of selenite alone.
Reduction in Rs. rubrum.
Selenite (0·5 mM) was reduced at a very rapid rate (48 nmol ml–1 h–1), although the reaction was delayed to the end of exponential growth (Fig. 1b
).
Amendment with a combination of selenite (0·5 mM) and nitrite (0·5 mM) or selenite (0·5 mM) and sulfite (0·3–0·4 mM) did not modify the rate of the reaction, and, in each case, selenite reduction was delayed until the end of the exponential phase as in the cultures amended with selenite alone (compare Fig. 1a, b).
Interestingly, selenite reduction was strongly inhibited when sulfite was maintained at a concentration greater than 0·3 mM by supplementary additions of this ion a few hours before the start of reduction and during the first hours of reduction (Fig. 1b). Supplementary additions of nitrite did not affect the selenite reduction kinetics (not shown).
Nitrite and sulfite were efficiently reduced during exponential growth, with only 8 and 12 %, respectively, remaining in the cultures at the beginning of stationary phase (Fig. 1b). This may indicate that an active reduction system for both nitrite and sulfite was present in Rs. rubrum under the growth conditions used in these experiments. Nitrite consumption was decreased in cultures amended with selenite (0·5 mM) and nitrite (0·5 mM) together. Under these conditions, 25 % of the initial nitrite was still present in the cultures upon entry into stationary phase (not shown). In cultures containing selenite (0·5 mM) and sulfite (0·3–0·4 mM) together, only 3–4 % of the initial sulfite was detected in the cultures by the end of growth (not shown).
During exponential growth, none of the amended sulfite (0·3–0·4 mM) could be detected in the spent culture medium (data not shown). When the sulfite concentration was higher (0·9–1·0 mM), 25–30 % of the added sulfite was detectable in the spent culture medium. The extant sulfite concentration could only be accurately quantified in the presence of cells. An affinity of sulfite for cell surfaces was never observed in stationary phase cultures.
Selenite reduction in Rs. rubrum was considerably slower during stationary phase (20 nmol ml–1 h–1 maximal rate) than during exponential growth (48 nmol ml–1 h–1), and the presence of nitrite with selenite did not influence the selenite reduction rate (Fig. 4). In contrast, sulfite (0·3–0·4 mM) completely inhibited selenite reduction in stationary-phase experiments (Fig. 4). Interestingly, neither nitrite nor sulfite was transformed when added to the cultures at stationary phase (Fig. 5). Selenite reduction was always delayed by about 2 h after selenite injection in stationary-phase cultures of Rs. rubrum. The lag in reduction was increased by about 2 h when the added selenite was dissolved in fresh culture medium instead of in water (0·5 % of the culture volume, v/v) (not shown).
|
|
(a).
|
Amendment with selenite and nitrite severely decreased both growth rate and cell density achieved at stationary phase compared to cultures amended with selenite alone or nitrite alone. As in the cultures of Rs. rubrum, the effects of a combination of selenite and nitrite were more pronounced than the sum of the effects of each ion separately.
A combination of selenite (0·5 mM) and sulfite (0·3–0·4 mM) markedly decreased the growth rate for the first 95 h of incubation, but the cell density slowly increased again after that time. As in the case of selenite and nitrite, the effect of selenite and sulfite on the growth rate was more pronounced than that of each ion separately.
Reduction in Rb. capsulatus.
In contrast to Rs. rubrum, Rb. capsulatus transformed selenite immediately, though at a much lower rate (8·8 nmol ml–1 h–1) (Fig. 2b). It must be noted that the rate of selenite reduction at exponential phase in Rb. capsulatus depends on selenite concentration, and decreased markedly with increasing initial selenite concentration (not shown). Therefore, effects of nitrite or sulfite on the kinetics of selenite reduction can be assessed only by comparing cultures with an identical initial selenite concentration.
Selenite reduction was strongly inhibited in the presence of nitrite (0·5 mM), but sulfite did not have any effect on the reaction (Fig. 2b). Amendment with 0·5 mM nitrite alone resulted in a very slow disappearance of this ion during exponential phase: 60 % of the amended concentration was still present in the culture medium when the cultures entered stationary phase (Fig. 3), suggesting that a nitrite reductase is present in Rb. capsulatus, although of low activity, under the growth condition used. Moreover, nitrite concentrations varied little over 6–7 days in cultures amended with 0·5 mM nitrite and 0·5 mM selenite (Fig. 3). In Rb. capsulatus, as in Rs. rubrum, sulfite (0·3–0·4 mM) could not be detected in the spent culture medium (data not shown). When the sulfite concentration was higher (0·9–1·0 mM), 50 % of the added sulfite could be accounted for in Rb. capsulatus (see insert in Fig. 3). The extant sulfite concentration could be accurately quantified only in the presence of cells. As in Rs. rubrum, affinity of sulfite for the cell surface was never observed in stationary-phase cultures. During exponential growth, sulfite disappeared very rapidly from both the culture medium and cell surfaces of Rb. capsulatus (see insert in Fig. 3).
|
).
As observed also in the case of Rs. rubrum, the stationary-phase cultures of Rb. capsulatus were not able to transform either nitrite or sulfite (Fig. 5).
Influence of intracellular glutathione level on growth and selenite reduction
The influence of glutathione on growth and on the kinetics of selenite reduction was investigated principally in cultures of Rs. rubrum; only stationary-phase experiments were performed with Rb. capsulatus. BSO, which is a specific inhibitor of glutathione synthesis, was used to depress the intracellular glutathione level.
Growth.
Addition of selenite alone to Rs. rubrum up to a concentration of 0·55–0·60 mM in the culture medium only minimally affected lag time and growth rate, but the cell density attained at stationary phase was decreased by 50 % when compared to selenite-free cultures (Fig. 6a). Cultures amended with 1·0 or 3·0 mM BSO initially exhibited a slow growth rate compared to BSO-free controls; however, the growth rate equalled that of the controls about 22 h before entering stationary phase (Fig. 6a). The final cell density attained in cultures amended with BSO alone was decreased by only a few percent compared to controls (Fig. 6a). A combination of BSO and selenite resulted in a growth rate similar to that of cultures amended with BSO alone and to cell densities equal to those of cultures amended with selenite alone (Fig. 6a).
|
; Kessi et al., 1999). In the presence of BSO (1·0–3·0 mM), glutathione concentration in the cell cytoplasm and maximum selenite reduction rate were both significantly decreased during exponential and stationary phases compared to those of control cultures (Figs 6b and 7, Table 1). During exponential growth, however, no correlation could be found between the intracellular concentration of glutathione and the rate of selenite reduction. Cultures amended with 1·0 or 3·0 mM BSO exhibited the same selenite reduction rate (Fig. 6b, Table 1). In contrast, an increase of BSO concentration from 0·5 to 3·0 mM in stationary-phase cultures correlated with both a decrease of the intracellular glutathione level and a decrease of the selenite reduction rate. In cultures amended with 0·5 or 1·0 mM BSO, the lag time of the selenite reduction was increased, but the rate of the reaction was not decreased compared with cultures amended with selenite alone (Fig. 7, Table 1).
|
|
) did not allow a precise determination of the glutathione concentration of stationary-phase cells amended with BSO, it was possible to determine that the glutathione concentration in the cells grown in the presence of 1·0–3·0 mM BSO was approximately 0·01 mM (Table 1).
In stationary-phase experiments with Rb. capsulatus, the reduction of selenite was particularly fast. In the presence of BSO, not only was the initial rate of the reduction lower than that measured for the cultures amended with selenite only, but it decreased with time approximately 6–7 h after the beginning of the reaction. The kinetics of reduction varied little when the BSO concentration was increased from 0·5 to 3·0 mM (Fig. 8).
|
. Very similar results were obtained with Rb. capsulatus. Cultures of both organisms amended with selenite alone produced several methylated compounds, including methanethiol, dimethylsulfide, dimethylselenenylsulfide and dimethyldiselenide (Fig. 9a). In contrast, cultures amended with selenite and 3 mM BSO contained only a small amount of dimethylsulfide; all other methylated compounds were absent (Fig. 9b). Control experiments using freshly made anaerobic medium supplemented with 1·5–2·0 nM of dimethyldiselenide alone or dimethylselenide and 3 mM BSO demonstrated that volatile selenium compounds are not chemically transformed in the presence of this concentration of BSO (three experiments).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and a dissimilatory sulfite reductase has shown a high selenite reductase activity (Harrison et al., 1984), we investigated the possible involvement of these enzymes in the reduction of selenite in two species of purple non-sulfur bacteria, Rs. rubrum and Rb. capsulatus. We systematically tested the influence of nitrite and sulfite, separately and in combination with selenite, on cell growth and the kinetics of selenite reduction.
The results demonstrate that Rs. rubrum and Rb. capsulatus employ different strategies to cope with the various ions. Rs. rubrum separated reduction of selenite from that of either nitrite or sulfite, thus largely avoiding any negative impacts of these ions on selenite metabolism (Fig. 1b). In Rb. capsulatus, all ions were metabolized simultaneously, and interferences between or among pathways were more obvious. In both organisms, amendment with sulfite alone or nitrite alone did not modify or only weakly affected growth rate and final cell density, implying that the observed modifications of these parameters in the presence of selenite are not due to toxicity effects of nitrite or sulfite.
In Rs. rubrum, the successive transformation of nitrite and selenite on the one hand, and sulfite and selenite on the other, suggests that this organism is able to discriminate between each ion and so probably possesses very specific recognition systems. We therefore propose that selenite may be kept outside the cell until the end of exponential phase. This assumption is supported by the high reactivity of selenite with glutathione reported in Ganther (1971) and Kessi & Hanselmann (2004). Consistent with these results, selenite would rapidly react with glutathione or other cytoplasmic molecules containing -SH groups as soon as it entered the cell. This property of selenite excludes that it is transported into the cytoplasm before the reduction system is operating and implies that bacteria resistant to selenite are equipped with a tightly regulated selenite transport system. The existence of specific uptake systems for selenite has been reported by many authors and is reviewed by Heider & Boeck (1993). According to these considerations, it can be proposed that the strong inhibition of selenite reduction in Rs. rubrum in the presence of sulfite during stationary as well as exponential phase after supplementary additions of sulfite may be tentatively explained if one assumes inhibition of the selenite transport system by sulfite ions (see also below).
Both sulfite and nitrite markedly increased the lag time of Rs. rubrum, when selenite alone minimally affected this parameter (Fig. 1a). This suggests that sulfite metabolism and nitrite metabolism, in contrast to selenite metabolism, are not constitutively expressed. Consistent with this result, selenite was reduced, but neither sulfite nor nitrite was transformed during stationary phase in unadapted cultures of Rs. rubrum. This excludes common reduction pathways for selenite and nitrite on the one hand and selenite and sulfite on the other.
In Rb. capsulatus, in which all ions were metabolized simultaneously during exponential phase, growth was affected to a greater extent than in Rs. rubrum by the presence of a combination of selenite and nitrite or selenite and sulfite (Fig. 2a). Selenite and sulfite substantially decreased growth and cell density, compared to cultures amended with selenite alone, until selenite reduction was complete (90 h after the start of growth). This result shows that simultaneity of selenite and sulfite metabolism constituted a difficult challenge for the cells. However, the selenite reduction kinetics were not at all affected in the presence of sulfite during the growth phase. These two combined results reinforce the assumption that sulfite interferes with the selenite reduction pathway on the level of the transport system rather than on the level of the reductase.
As with Rs. rubrum (see above), the absence of reduction of both ions in stationary-phase experiments with Rb. capsulatus showed that the nitrite and sulfite reduction pathways, in contrast to the selenite reduction pathway, may not be constitutively expressed. Thus, exponential-phase experiments with Rb. capsulatus and stationary-phase experiments with both organisms offer evidence that the sulfite reductase may not, in fact, function as a dissimilatory selenite reductase, despite the relatively high selenite reductase activity determined by Harrison et al. (1984) for C. pasteurianum.
In the presence of selenite and nitrite, the growth and selenite reduction kinetics of Rb. capsulatus were significantly decreased compared with cultures amended with selenite only, indicating strong interferences between selenite and nitrite metabolism. However, in Rb. capsulatus, as in Rs. rubrum (see above), selenite was reduced during stationary phase in cultures that were not previously exposed to selenite (Fig. 4), indicating that selenite reductase is expressed constitutively. In contrast, nitrite was not transformed during stationary phase in non-adapted cultures of either Rb. capsulatus or Rs. rubrum (Fig. 5), indicaing that the dissimilatory nitrite reduction pathway may not be constitutively expressed, and therefore that neither the nitrite reductase nor the sulfite reductase functions as a selenite reductase in either of the two organisms considered.
This conclusion is at variance with that of DeMoll-Decker & Macy (1993), who propose that the periplasmic nitrite reductase of T. selenatis catalyses the reduction of selenite to elemental selenium. However, the large negative effects of nitrite on growth and on selenite reduction in cultures of Rb. capsulatus, along with the marked decrease of growth rate and final cell density attained in the presence of selenite and nitrite in Rs. rubrum (see Figs 1a and 2a), suggest that the pathways for nitrite and selenite reduction share common sequence(s). This is consistent with the suggestion of DeMoll-Decker & Macy (1993) that perhaps only one component of the nitrite respiratory system is involved in the reduction of selenite to Se0. The observed increase of an as-yet-unidentified cytochrome in the cultures of both Rs. rubrum and Rb. capsulatus exposed to selenite or nitrite (J. Kessi, unpublished results) suggests that the selenite and nitrite reduction pathways may share (an) electron transfer protein(s).
It must be noted that the dissimilatory nitrite reductases of various Thauera species, in contrast to the enzymes found in the purple non-sulfur bacteria (see Introduction), belong to the cd1-type nitrite reductases (Etchebehere & Tiedje, 2005). The variability of the nitrite reduction system among species of bacteria that are highly resistant to selenium oxyanions also supports the proposition that this enzymic system may not be involved in the bacterial reduction of selenite.
The non-constitutive expression of the sulfite reductase is not consistent with the fast disappearance of sulfite at the beginning of growth in Rb. capsulatus. One interpretation of this result is that the sulfite reduction system is rapidly activated during exponential growth.
In Rs. rubrum, a combination of sulfite and selenite increased the cell density attained at stationary phase compared with controls amended with selenite only (Fig. 1a). This result is consistent with those of Van Fleet-Stalder et al. (1997), who show that sulfur compounds positively influence selenite metabolism. These authors observe that low sulfate concentration limits biomass production and selenium volatilization and that high sulfate concentration enhances the release of organoselenium compounds in Rb. sphaeroides. The positive influence of sulfur compounds on selenite metabolism is also consistent with the close relationship between bacterially mediated selenite reduction and the abiotic reduction of selenite by glutathione (Kessi & Hanselmann, 2004), as well as with the observation that glutathione may be involved in the dissimilatory reduction of selenite in purple non-sulfur bacteria (see below).
Influence of the intracellular glutathione level on the reduction of selenite
A physiological approach was also used to investigate the involvement of glutathione in the reduction of selenite in Rs. rubrum. BSO, a specific inhibitor of glutathione synthesis, was used to depress the intracellular glutathione level, and to determine the effects of the intracellular concentration of glutathione on cell growth and on the kinetics of selenite reduction.
Effect on cell growth.
The lag time of Rs. rubrum amended with BSO was longer than that of the control cultures free of BSO, indicating that the cells developed a metabolic adaptation to this molecule. Cell integrity seemed to have been preserved, since the growth rate equalled that of cultures not amended with BSO 22 h before entering stationary phase, and the final cell density was decreased only minimally compared with BSO-free cultures.
Effect on selenite reduction.
A direct effect of BSO on the rate of selenite reduction must be characterized by an interdependence between the concentration of glutathione and the rate of the reaction when the concentration of glutathione is below the Km of the glutathione reductase for glutathione. In exponential-phase experiments, the rate of the reaction was decreased, despite the fact that the glutathione concentration was not decreased below the Km of the glutathione reductase (Fig. 6b, Table 1). This result indicates that the decrease of the reaction rate may not be based on the decrease of the glutathione concentration. In this case, the decrease of the rate of selenite reduction may be explained by a slow degradation of BSO during incubation of the cultures, liberating small quantities of sulfite into the culture medium. This hypothesis is supported by the strong inhibition of the reduction of selenite observed in the presence of sulfite in Rs. rubrum (see above). Induction of a metabolic adaptation to BSO during exponential growth is consistent with the increase of the lag time observed in cultures amended with BSO.
The influence of BSO on the selenite reduction kinetics via a decrease of the intracellular concentration of glutathione was more evident in stationary-phase cultures. In these experiments, a relationship was observed between the concentration of BSO, or the intracellular glutathione level, and the rate of selenite reduction. In stationary-phase cultures amended with 0·5–3·0 mM BSO, the intracellular glutathione concentration was significantly lower (0·01–0·05 mM, Table 1) than the Km (0·09 mM) of the glutathione for the glutathione reductase of Rs. rubrum (Libreros-Minotta et al., 1992). This may explain the significant decrease of the selenite reduction rate in the cultures amended with 1·5 and 3·0 mM BSO (Fig. 7, Table 1). This result strongly suggests the involvement of glutathione in the reduction of selenite in Rs. rubrum. The slight increase of the selenite reduction rate in the cultures amended with 0·5 and 1·0 mM BSO, accompanied by an increased lag time for the reaction, particularly in the cultures amended with 1·0 mM BSO, may be explained by the induction of the synthesis of glutathione and/or of the reductase(s) involved in the reduction of the selenodiglutathione (see below).
In both the growth and the stationary phases of the control cultures of Rs. rubrum free of BSO, the intracellular glutathione concentration (4·3 and 0·8 mM, respectively) was higher than the Km of glutathione for the glutathione reductase (0·09 mM). This explains why no correlation could be observed between the concentration of glutathione and the rate of selenite reduction in these cultures, and indicates that the concentration of glutathione is not a limiting factor for the reduction of selenite under conditions in which the intracellular glutathione level is not depressed.
The effect of BSO on the kinetics of selenite reduction in Rb. capsulatus was investigated at stationary phase only. The nearly equal decrease of the initial rate of selenite reduction in the presence of various concentrations of BSO (Fig. 8) could be interpreted as the result of the release of sulfite ions due to a slow degradation of BSO, as suggested for the exponential-phase experiments with Rb. rubrum. However, as the sulfite concentration is unlikely to be modified within a few hours, this interpretation would not explain the decrease of the reduction rate observed with time in the stationary-phase cultures of Rb. capsulatus. The decrease of the reduction rate with time can tentatively be interpreted as a decrease of the intracellular level of reduced glutathione during the reaction. Possibly the rate of glutathione reduction becomes a limiting factor in the selenite reduction process. However, this phenomenon cannot be interpreted without further investigations.
The involvement of glutathione in the selenite reduction pathway in purple non-sulfur bacteria is consistent with the high reactivity of selenite with glutathione (Ganther, 1971; Kessi & Hanselmann, 2004) and with many correlations observed between the abiotic reaction and the process mediated by E. coli and Rb. sphaeroides. i) Superoxide anions are produced in the abiotic reaction (Kessi & Hanselmann, 2004), and two types of superoxide dismutase are induced in E. coli amended with selenite (Bébien et al., 2002). ii) Selenodiglutathione is formed in both the abiotic reaction (Ganther, 1971; Kessi & Hanselmann, 2004) and the reaction mediated by E. coli (Rabenstein & Tan, 1988). iii) Selenodiglutathione is a good substrate for glutathione reductase (Ganther, 1971) and this enzyme is induced in both Rb. sphaeroides and E. coli amended with selenite (Bébien et al., 2001, 2002, respectively).
As selenodiglutathione is not only a good substrate for the glutathione reductase (Ganther, 1971) but also a highly efficient oxidant of reduced thioredoxin in E. coli (Björnstedt et al., 1992), thioredoxin may also be involved in the reduction, possibly depending on the organism and the redox state of the cells. Transposon mutagenesis has failed to identify the enzymic system(s) involved in the reduction of selenite in Rb. sphaeroides (Bébien et al., 2001). This may be explained by the ability of alphaproteobacteria to call upon two or more different enzymic systems to catalyse the reaction, and by the fact that double mutants in gor and trxB (glutathione reductase and thioredoxin reductase) or gshA and trxA (-glutamylcysteine sythetase and thioredoxin) grow poorly or are not viable in minimal media (Gleason & Holmgren, 1988).
The reduction of selenite with glutathione has been shown to produce selenodiglutathione together with superoxide anions (Kessi & Hanselmann, 2004). As glutathione reductase and thioredoxin reductase are involved not only in the reduction of the selenodiglutathione but also in the regulation of the expression of the oxidative stress system (Carmel-Harel & Storz, 2000), it is necessary to determine the part of the induction contributed by each function. Experiments are currently being undertaken in our laboratory to clarify this question.
Gas chromatographic analyses of culture headspace gases clearly indicate that inhibition of the synthesis of glutathione strongly decreases the formation of volatile selenite metabolites in both organisms studied (Fig. 9). It must be noted that the production of volatile sulfur compounds is also strongly inhibited. This suggests that the same enzymes mediate the production of both types of volatile metabolite. These observations are consistent with those of Harrison et al. (1980), who reported a large decrease in the evolution of hydrogen sulfide in cultures of C. pasteurianum and in cell-free extracts amended with selenite.
The assumed role of glutathione in the formation of methylated selenium compounds in purple non-sulfur bacteria is in agreement with the results reported by Van Fleet-Stalder et al. (1997), who observed a decrease in the synthesis of these compounds under sulfur-limited conditions.
It must be pointed out, however, that the involvement of glutathione in the selenite reduction pathway in bacteria is restricted to the alpha-, beta- and gammaproteobacteria and to the cyanobacteria, which are the only bacteria able to synthesize glutathione (Newton & Fahey, 1989). In all other bacterial species, which lack the glutathione-dependent reduction system, the pathway for the dissimilatory reduction of selenite inevitably differs from that of the alphaproteobacteria considered in this work. Bacteria deprived of glutathione contain high levels of other thiols, including coenzyme A, cysteine, lipoamide and pantetheine, with their corresponding thiol/disulfide oxidoreductases (Newton & Fahey, 1989). As selenite is highly reactive towards thiols (Painter, 1941; Ganther, 1971), such oxidoreductases are possibly involved in the reduction of selenite in these organisms.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bébien, M., Chauvin, J.-P., Adriano, J.-M., Grosse, S. & Verméglio, A. (2001). Effect of selenite on growth and protein synthesis in the phototrophic bacterium Rhodobacter sphaeroides. Appl Environ Microbiol 67, 4440–4447.
Bébien, M., Lagniel, G., Garin, J., Touati, D., Verméglio, A. & Labarre, J. (2002). Involvement of superoxide dismutases in the response of Escherichia coli to selenium oxides. J Bacteriol 184, 1556–1564.
Björnstedt, M., Kumar, S. & Holmgren, A. (1992). Selenodiglutathione is a highly efficient oxidant of reduced thioredoxin and a substrate for mammalian thioredoxin reductase. J Biol Chem 267, 8030–8034.
Bratton, A. C., Marshall, E. K., Babbitt, D. & Hendrickson, A. R. (1939). A new coupling component for sulfanilamide determination. J Biol Chem 128, 537–550.
Carmel-Harel, O. & Storz, G. (2000). Roles of the glutathione- and thioredoxin-dependent reduction systems in the Escherichia coli and Saccharomyces cerevisiae responses to oxidative stress. Annu Rev Microbiol 54, 439–461.[CrossRef][Medline]
DeMoll-Decker, H. & Macy, J. M. (1993). The periplasmic nitrite reductase of Thauera selenatis may catalyze the reduction of selenite to elemental selenium. Arch Microbiol 160, 241–247.
Dobbin, P. S., Warren, L. H., Cook, N. J., McEwan, A. G., Powell, A. K. & Richardson, D. J. (1996). Dissimilatory iron(III) reduction by Rhodobacter capsulatus. Microbiology 142, 765–774.
Dungan, R. S., Yates, S. R. & Frankenberger, W. T., Jr (2003). Transformations of selenate and selenite by Stenotrophomonas maltophilia isolated from a seleniferous agricultural drainage pond sediment. Environ Microbiol 5, 287–295.[CrossRef][Medline]
Etchebehere, C. & Tiedje, J. (2005). Presence of two different active nirS nitrite reductase genes in a denitrifying Thauera sp. from a high-nitrate-removal-rate reactor. Appl Environ Microbiol 71, 5642–5645.
Ferguson, S. J., Jackson, J. B. & McEwan, A. G. (1987). Anaerobic respiration in the Rhodospirillaceae: characterisation of pathways and evaluation of roles in redox balancing during photosynthesis. FEMS Microbiol Rev 46, 117–143.[CrossRef]
Fritz, M. & Bachofen, R. (2000). Volatile organic sulfur compounds in a meromictic alpine lake. Acta Hydrochim Hydrobiol 28, 185–192.[CrossRef]
Ganther, H. E. (1966). Enzymic synthesis of dimethylselenide from sodium selenite in mouse liver extracts. Biochemistry 5, 1089–1098.[CrossRef][Medline]
Ganther, H. E. (1968). Selenotrisulfides. Formation by the reaction of thiols with selenious acid. Biochemistry 7, 2898–2905.[CrossRef][Medline]
Ganther, H. E. (1971). Reduction of the selenotrisulfide derivative of glutathione to a persulfide analog by glutathione reductase. Biochemistry 10, 4089–4098.[CrossRef][Medline]
Gerrard, T. L., Telford, J. N. & Williams, H. H. (1974). Detection of selenium deposits in Escherichia coli by electron microscopy. J Bacteriol 119, 1057–1060.
Gleason, F. K. & Holmgren, A. (1988). Thioredoxin and related proteins in prokaryotes. FEMS Microbiol Rev 54, 271–298.[CrossRef]
Griffith, O. W. (1981). Depletion of glutathione by inhibition of biosynthesis. Methods Enzymol 77, 59–63.[Medline]
Griffith, O. W. & Meister, A. (1979). Potent and specific inhibition of glutathione synthesis by buthionine sulfoximine (S-n-butyl homocysteine sulfoximine). J Biol Chem 254, 7558–7560.
Harrison, G. I., Laishley, E. J. & Krouse, H. R. (1980). Stable isotope fractionation by Clostridium pasteurianum. 3. Effect of 
on the physiology and associated sulfur isotope fractionation during 
and 
reductions. Can J Microbiol 26, 952–958.[Medline]
Harrison, G., Curle, C. & Laishley, E. J. (1984). Purification and characterization of an inducible dissimilatory type sulfite reductase from Clostridium pasteurianum. Arch Microbiol 138, 72–78.[CrossRef][Medline]
Heider, J. & Boeck, A. (1993). Selenium metabolism in micro-organisms. Adv Microbial Physiol 35, 71–109.[Medline]
Karr, E. A., Sattley, W. M., Jung, D. O., Madigan, M. T. & Achenbach, L. A. (2003). Remarkable diversity of phototrophic purple bacteria in a permanently frozen antarctic lake. Appl Environ Microbiol 69, 4910–4914.
Kelly, B. S., Antholine, W. E. & Griffith, O. W. (2002). Escherichia coli -glutamylcysteine synthetase. Two active site metal ions affect substrate and inhibitor binding. J Biol Chem 277, 50–58.
Kessi, J. & Hanselmann, K. W. (2004). Similarities between the abiotic reduction of selenite with glutathione and the dissimilatory reaction mediated by Rhodospirillum rubrum and Escherichia coli. J Biol Chem 279, 50662–50669.
Kessi, J., Ramuz, M., Wehrli, E., Spycher, M. & Bachofen, R. (1999). Reduction of selenite and detoxification of elemental selenium by the phototrophic bacterium Rhodospirillum rubrum. Appl Environ Microbiol 65, 4734–4740.
Läuchli, A. (1993). Selenium in plants: uptake, functions, and environmental toxicity. Bot Acta 106, 455–468.
Libreros-Minotta, C. A., Pardo, J. P., Mendoza-Hernández, G. & Rendón, J. L. (1992). Purification and characterization of glutathione reductase from Rhodospirillum rubrum. Arch Biochem Biophys 298, 247–253.[CrossRef][Medline]
Losi, M. E. & Frankenberger, W. T., Jr (1997). Reduction of selenium oxyanions by Enterobacter cloacae SLD1a-1: isolation and growth of the bacterium and its expulsion of selenium particles. Appl Environ Microbiol 63, 3079–3084.[Abstract]
Moore, M. D. & Kaplan, S. (1992). Identification of intrinsic high-level resistance to rare-earth oxides and oxyanions in members of the class Proteobacteria: characterization of tellurite, selenite, and rhodium sesquioxide reduction in Rhodobacter sphaeroides. J Bacteriol 174, 1505–1514.
Newton, G. L. & Fahey, R. C. (1989). Glutathione in prokaryotes. In Glutathione: Metabolism and Physiological Functions, pp. 69–77. Edited by J. Viña. Boca Raton: CRC Press.
Ohlendorf, H. M. (1989). Bioaccumulation and effects of selenium in wildlife. In Selenium in Agriculture and the Environment, pp. 133–177. Edited by L. W. Jacobs. Madison, WI: American Society of Agronomy.
O'Toole, D. & Raisbeck, M. F. (1998). Magic numbers, elusive lesions: comparative pathology and toxicology of selenosis in waterfowl and mammalian species. In Environmental Chemistry of Selenium, pp. 355–395. Edited by W. T. Frankenberger, Jr & R. A. Engberg. New York: Marcel Dekker.
Painter, E. P. (1941). The chemistry and toxicity of selenium compounds with special reference to the selenium problem. Chem Rev 28, 179–213.[CrossRef]
Peck, H. D., Jr, Tedro, S. & Kamen, M. D. (1974). Sulfite reductase activity in extracts of various photosynthetic bacteria. Proc Natl Acad Sci U S A 71, 2404–2406.
Preuss, M. & Klemme, J.-H. (1983). Purification and characterization of a dissimilatory nitrite reductase from the phototrophic bacterium Rhodopseudomonas palustris. Z Naturforsch (C) 38, 933–938.
Rabenstein, D. L. & Tan, K.-S. (1988). 77Se NMR studies of bis(alkylthio)selenides of biological thiols. Magn Resonance Chem 26, 1079–1085.[CrossRef]
Richardson, D. J., Bell, L. C., Moir, J. W. B. & Ferguson, S. J. (1994). A denitrifying strain of Rhodobacter capsulatus. FEMS Microbiol Lett 120, 323–328.[CrossRef]
Schwintner, C., Sabaty, M., Berna, B., Cahors, S. & Richaud, P. (1998). Plasmid content and localization of the genes encoding the denitrification enzymes in two strains of Rhodobacter sphaeroides. FEMS Microbiol Lett 165, 313–321.[CrossRef][Medline]
Tomei, F. A., Barton, L. L., Lemanski, C. L. & Zocco, T. G. (1992). Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Can J Microbiol 38, 1328–1333.
Tomei, F. A., Barton, L. L., Lemanski, C. L., Zocco, T. G., Fink, N. H. & Sillerud, L. O. (1995). Transformation of selenate and selenite to elemental selenium by Desulfovibrio desulfuricans. J Ind Microbiol 14, 329–336.[CrossRef]
Van Fleet-Stalder, V., Gürleyük, H., Bachofen, R. & Chasteen, T. G. (1997). Effects of growth conditions on production of methyl selenides in cultures of Rhodobacter sphaeroides. J Ind Microbiol Biotechnol 19, 98–103.[CrossRef]
Woese, C. R., Stackebrandt, E., Weisburg, W. G. & 8 other authors (1984). The phylogeny of purple bacteria: the alpha subdivision. Syst Appl Microbiol 5, 315–326.[Medline]
Received 30 May 2005;
revised 11 October 2005;
accepted 16 November 2005.
This article has been cited by other articles:
|
|
 |
|
  L. Avoscan, M. Carriere, O. Proux, G. Sarret, J. Degrouard, J. Coves, and B. Gouget Enhanced Selenate Accumulation in Cupriavidus metallidurans CH34 Does Not Trigger a Detoxification Pathway Appl. Envir. Microbiol., April 1, 2009; 75(7): 2250 - 2252. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Antonioli, S. Lampis, I. Chesini, G. Vallini, S. Rinalducci, L. Zolla, and P. G. Righetti Stenotrophomonas maltophilia SeITE02, a New Bacterial Strain Suitable for Bioremediation of Selenite-Contaminated Environmental Matrices Appl. Envir. Microbiol., November 1, 2007; 73(21): 6854 - 6863. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
| J MED MICROBIOL | ALL SGM JOURNALS | |
, control cultures without amendment; 
, 0·5 mM nitrite; 
, 0·3–0·4 mM sulfite; 
, 0·5 mM selenite; 
, 0·5 mMselenite+0·5 mMnitrite; 
, 0·5 mM selenite+0·3–0·4 mM sulfite; 
, 1·0 mM BSO only; 
, 3·0 mM BSO only; 
, 1·0 mM BSO+0·5 mM selenite; 
, 3·0 mM BSO+0·5 mM selenite. In (b), 
