Reduction of Selenite and Detoxification of Elemental Selenium by the ...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1999, p. 4734–4740 0099-2240/99/$04.00⫹0 Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Vol. 65, No. 11

Reduction of Selenite and Detoxification of Elemental Selenium by the Phototrophic Bacterium Rhodospirillum rubrum J. KESSI,1* M. RAMUZ,2 E. WEHRLI,3 M. SPYCHER,4

AND

R. BACHOFEN1

1

Institute of Plant Biology, University of Zurich, CH-8008 Zurich, Institute of Physiology, University of Zurich, CH-8057 Zurich,2 Laboratory for Electron Microscopy, ETH-Zentrum, CH-8092 Zurich,3 and Laboratory for Electron Microscopy, University Hospital, CH-8091 Zurich,4 Switzerland Received 1 March 1999/Accepted 15 July 1999

The effect of selenite on growth kinetics, the ability of cultures to reduce selenite, and the mechanism of detoxification of selenium were investigated by using Rhodospirillum rubrum. Anoxic photosynthetic cultures were able to completely reduce as much as 1.5 mM selenite, whereas in aerobic cultures a 0.5 mM selenite concentration was only reduced to about 0.375 mM. The presence of selenite in the culture medium strongly affected cell division. In the presence of a selenite concentration of 1.5 mM cultures reached final cell densities that were only about 15% of the control final cell density. The cell density remained nearly constant during the stationary phase for all of the selenite concentrations tested, showing that the cells were not severely damaged by the presence of selenite or elemental selenium. Particles containing elemental selenium were observed in the cytoplasm, which led to an increase in the buoyant density of the cells. Interestingly, the change in the buoyant density was reversed after selenite reduction was complete; the buoyant density of the cells returned to the buoyant density of the control cells. This demonstrated that R. rubrum expels elemental selenium across the plasma membrane and the cell wall. Accordingly, electron-dense particles were more numerous in the cells during the reduction phase than after the reduction phase. shown to be particularly resistant to a variety of metal and transition metal oxyanions, including selenium. This resistance is attributed to the capacity of the organisms to reduce Se oxyanions to their elemental ground state (11), which is poorly soluble and thus less toxic than the initial oxyanions. Multiple detoxification processes may occur during selenite reduction by microorganisms since elemental selenium has been described as being deposited in the cytoplasm (13, 15, 16), in the periplasmic space (5), and outside the cell (6, 9, 19). According to Tomei et al. (16), the particles containing elemental selenium found outside cells are released by cell lysis, while Losi and Frankenberger (9) suggested that the reduction reaction occurs close to the membrane, possibly as a result of a membrane-associated reductase(s), and that the precipitate is rapidly expelled by a membrane efflux pump. On the other hand, elemental selenium deposited inside or outside cells has been described as being in spherical or spherical to ovalshaped structures (9, 16), fibrillar and granular structures (13), or amorphous aggregates (5, 19). Interestingly, in Escherichia coli elemental selenium deposition has been observed both in the periplasmic space (5) and in the cytoplasm (13). In this paper we describe the ability of Rhodospirillum rubrum, a purple nonsulfur bacterium, to reduce selenite to its elemental state and the mechanism of detoxification of elemental selenium in this organism.

Selenium is a normally occurring trace element. It is essential for humans and animals but is very toxic at higher concentrations. While in some regions of the world part of the daily food intake is artificially enriched with selenium for health reasons, other regions (e.g., some parts of the San Joaquin Valley in central California) are polluted with selenium (8). The greatest abundance of selenium is in igneous rocks, but high concentrations are also present in some sedimentary rocks and fossil fuels (12). The following three main forms of elemental selenium, Se0, have been described: a red amorphous form, a black amorphous form, and a grey hexagonal form. The red and black amorphous allotropes are the forms that are most likely to occur in soils. Red amorphous Se0 originates when Se0 precipitates in aqueous solution (4). At temperatures greater than 30°C, red amorphous Se0 gradually reverts to the black amorphous form (3). This form is then slowly transformed into the much more stable grey hexagonal allotrope or is reoxidized, depending on the redox conditions and the pH of the soil. Oxidation can occur through inorganic reactions or by the action of microorganisms (4). In aerated soils and aquatic environments, selenium occurs predominantly in the form of selenite and selenate oxyanions (SeO3⫺2 and SeO4⫺2), which are freely available to living organisms. Human activities, such as coal mining and fuel refining, as well as industrial uses of selenium (e.g., in photocopy machines, electronics, glass manufacturing, chemicals, and pigments), affect the biological availability of selenium. Chemical detoxification of metal- and metalloid-polluted sites has proven to be very expensive and often results in secondary effects in the environment. Consequently, more sustainable biological solutions need to be found. Phototrophic microorganisms belonging to the group containing the purple bacteria have been

MATERIALS AND METHODS All chemicals were the purest grade available. Folin-Ciocalteu reagent and selenite were purchased from Merck (Darmstadt, Germany), and diaminonaphthalene was purchased from Aldrich (Buchs, Switzerland). Bradford reagent was obtained from Bio-Rad (Glattbrugg, Switzerland), and thioglycolic acid was obtained from Fluka (Buchs, Switzerland). Growth of bacteria. The growth medium used for phototrophic bacteria was prepared as described by Sistrom (14), except that casein hydrolysate was omitted and 0.2 mM NaHCO3 was added, which resulted in better reproducibility of the growth kinetics. Both aerobic and anaerobic cultures were grown in this medium, which contained succinate as a carbon source. Aerobic cultures were grown in 250-ml Erlenmeyer flasks containing 100 ml of medium at 30°C in the dark on a rotatory shaker at 160 rpm. The medium used for anaerobic cultures

* Corresponding author. Mailing address: Institute of Plant Biology, University of Zurich, Zollikerstrasse 107, CH-8008 Zurich, Switzerland. Phone: 01 634 82 11. Fax: 01 634 82 04. E-mail: Janine.Kessi @access.unizh.ch. 4734

VOL. 65, 1999

REDUCTION OF SELENITE AND DETOXIFICATION OF SELENIUM

4735

FIG. 2. Time course of anoxic phototrophic growth of R. rubrum in the presence of different selenite concentrations. Symbols: F, control cells (no selenite); E, 0.5 mM; ■, 1.0 mM; 䊐, 1.5 mM; }, 2.0 mM. Selenite was added at zero time. Each curve shows means based on the results of two or three experiments.

FIG. 1. Time course of growth and selenite reduction by R. rubrum. (A) Oxic conditions in the dark. (B) Anoxic conditions in the light. Symbols: F, control cells (no selenite); ■, cells grown in the presence of 0.5 mM selenite; E, selenite concentration. Selenite was added at zero time. Each curve shows means based on the results of two or three experiments.

was evacuated with an aspirator pump for about 1 h, and 150-ml flasks with rubber septa were filled with 125 ml of the medium. The gas phase was exchanged with N2 by 10 cycles consisting of 1 atm of vacuum followed by the addition of N2 to a pressure of 1.5 atm. R. rubrum S1 (⫽ DSM 467) was grown in agar stabs and was used as the inoculum for 125-ml flasks. Incubation was at 30°C in the presence of incandescent light (35 W/m2) with gentle stirring. When cultures reached the end of the exponential growth phase, the cells were used to inoculate a new culture flask. The volume of culture used for inoculation was calculated so that the starting cell concentration corresponded to an absorbance at 650 nm of 0.01 with a 2-mm path length. Three transfers were done before the cultures were used for experiments. Spectrophotometric measurements of cultures. Absorbance at 650 nm was measured with model Uvikon 860 spectrophotometer equipped with a second sample position close to the photomultiplier (Kontron, Zurich, Switzerland) by using 2-mm-path-length cuvettes and undiluted samples. Protein content determination. The protein content of cells was determined by using a modification of the method of Lowry et al. (10). An amount of cells corresponding to 200 ␮l of a cell suspension with an absorbance at 650 nm of 0.1 (path length, 2 mm) was centrifuged for 15 min at 15,000 ⫻ g and resuspended in 125 ␮l of 0.1 N NaOH. At this point the sample could be frozen and stored. Then 875 ␮l of a solution containing 0.025% copper sulfate, 0.050% sodium tartrate, and 2.5% sodium carbonate was added, and the sample was incubated at room temperature for 10 min. After addition of 250 ␮l of Folin-Ciocalteu solution (diluted 1/6 with H2O) and incubation for an additional 3 h, the absorbance at 750 nm of the copper-protein complex was determined by using a blank containing all of the reagents except protein. Selenium-containing samples were centrifuged at 15,000 ⫻ g for 10 min before measurements were obtained. All measurements were done in triplicate. Bovine serum albumin was used as the standard. The protein contents of the fractions of the density gradients were determined by using the method of Bradford (1) and bovine serum albumin as the standard. All measurements were done in duplicate.

Electron microscopy. Cells or vesicles were fixed in 2.5% glutaraldehyde for 60 min (samples were diluted with 5% aqueous glutaraldehyde), washed with running water, and embedded in low-melting-point agarose. Agar blocks (approximately 1 by 1 by 1 mm) were fixed in 1% OsO4 in running water for 60 min, dehydrated with ethanol and acetone, and embedded in Epon-Araldit. Sections cut from the Epon-Araldit preparation were contrasted with uranyl acetate and lead citrate as described by Hess (7). For energy-dispersive X-ray (EDX) analysis, whole cells were applied to carbon-coated transmission electron microscopy grids, dried at room temperature, and coated with 5 nm of carbon before measurements were obtained. The EDX analysis was performed with a Philips model CM12 electron microscope equipped with an EDAX-DX4 microanalysis system (Philips, Eindhoven, The Netherlands). Selenite content determination. Selenite contents were determined spectrophotometrically by using a modification of the method of Watkinson (18). First, 10 ml of 0.1 M HCl, 0.5 ml of 0.1 M EDTA, 0.5 ml of 0.1 M NaF, and 0.5 ml of 0.1 M disodium oxalate were mixed in a 25- to 30-ml glass tube. A 50- to 250-␮l sample containing 100 to 200 nmol of selenite was added, and then 2.5 ml of 0.1% 2,3-diaminonaphthalene in 0.1 M HCl was added. After the contents were mixed, the tubes were incubated at 40°C for 40 min and then cooled to room temperature. The selenium–2,3-diaminonaphthalene complex was extracted with

FIG. 3. Time course of selenite reduction by R. rubrum during anoxic phototrophic growth in the presence of different selenite concentrations. Symbols: E, 0.5 mM; ■, 1.0 mM; }, 2.0 mM. Selenite was added at zero time. Each curve shows means based on the results of two experiments.

4736

KESSI ET AL.

FIG. 4. Time course of the specific extracellular turbidity (absorbance at 650 nm normalized to the protein content of the cells) of R. rubrum cultures during anoxic photosynthetic growth in the presence of different selenite concentrations. Symbols: F, control cells (no selenite); E, 0.5 mM; 䊐, 1.0 mM; }, 2.0 mM. Selenite was added at zero time. Abs. 650 nm, absorbance at 650 nm. 6 ml of cyclohexane by shaking the tubes vigorously for about 1 min. The absorbance at 377 nm of the organic phase was determined by using a 1-cmpath-length cuvette. When necessary, phase separation was accelerated by centrifuging the tubes for 10 min at 3,000 ⫻ g. All manipulations were done in the

APPL. ENVIRON. MICROBIOL. dark. Calibration curves were obtained by using 0, 50, 100, 150, and 200 nmol of selenite. The data showed that there was a perfect linear relationship between selenite concentration and absorption (correlation factor, 0.998 to 0.999). All measurements were done in duplicate. Sucrose gradient centrifugation. Polyallomer tubes (17 ml; Beckman Instruments, Zurich, Switzerland) were filled with four layers of sucrose as follows: 3.5 ml of 2.5 M sucrose, 5.0 ml of 2.0 M sucrose, 3.0 ml of 1.5 M sucrose, and 3.0 ml of 0.1 M sucrose. A 2-ml culture sample was overlaid, and centrifugation was performed at 20°C and 60,000 ⫻ g for 2 h by using a type SW 28 rotor. Samples from cultures containing 1.5 and 2.0 mM selenite were concentrated 2⫻ before centrifugation. Fractions (0.5 ml) were collected from the tubes manually from the top by using a precision pipette. Isolation of selenium-containing particles formed by the bacteria. One and one-half liters of cells was centrifuged at 3,000 ⫻ g and 4°C for 10 min 1 to 2 days after the reduction phase was complete. The pellet was discarded, and the cell-free medium was centrifuged at 100,000 ⫻ g and 4°C for 40 min. The supernatant was discarded, and the pellet with the selenium-containing particles was resuspended in about 20 ml of 50 mM Tris-HCl (pH 7.5). The suspension was washed twice in the same buffer by repeating the two centrifugation steps. Preparation of selenium-containing particles in a cell-free spent medium. A 250-␮l portion of 0.1 M selenite was added to 50 ml of cell-free medium obtained from a 0.5 mM selenite culture in which reduction was complete. A 1.5-ml portion of 80% thioglycolic acid in H2O was added, and the solution was thoroughly mixed and left at room temperature. Orange-red selenium particles formed slowly in about 1 to 1.5 h. The preparation was centrifuged at 100,000 ⫻ g and 4°C for 40 min, the particles were resuspended in 2 ml of 50 mM Tris-HCl (pH 7.5), and then the preparation was centrifuged at 15,000 ⫻ g and 4°C for 20 min.

RESULTS When R. rubrum was grown under oxic conditions, very small differences in growth kinetics were observed when we compared the control cultures and cultures containing 0.5 mM

FIG. 5. Thin sections of R. rubrum cells grown in the presence of 0.5 mM selenite. (A) Growth under oxic conditions in the dark. (B) Growth under anoxic conditions in the light. Elemental selenium was localized in the electron-dense particles which were present in the cytoplasm of both aerobically and anaerobically grown cells. (see Fig. 7A).

VOL. 65, 1999

REDUCTION OF SELENITE AND DETOXIFICATION OF SELENIUM

4737

FIG. 6. Thin sections of selenium-containing particles. (A) Section obtained after bacterial reduction in a culture amended with 0.5 mM selenite. (B) Section obtained after reduction of selenite with thioglycolate in spent medium.

selenite, showing that growth of the bacteria was only minimally influenced by the presence of selenite ions (Fig. 1A). The selenite concentration remained constant during the exponential growth phase; then it decreased to about 0.375 mM in the transition phase and slowly increased again during the stationary phase (Fig. 1A). During anoxic phototrophic growth, the 0.5 mM selenitecontaining cultures had lower cell concentrations than the control cultures beginning slightly before the mid-exponential phase, and the cell concentrations were about 40% of the control culture under stationary conditions (Fig. 1B). On the other hand, 0.5 mM selenite was completely reduced during the transition from the exponential phase to the stationary phase (Fig. 1B). Under both oxic and anoxic conditions, selenite reduction started only at the end of the exponential phase, independent of the time when selenite was added to the culture. There was no difference in lag time between the control and selenitecontaining cultures. In the absence of selenite, the cell densities of anaerobic cultures were clearly greater than the cell densities of aerobic cultures. This was due to the significant energy contribution of photosynthesis under anaerobic growth conditions, while aerobically growing cells obtained energy only from the carbon source supplied. Under both oxic and anoxic conditions, decreases in selenite concentration in the medium paralleled the appearance of an orange-red color due to the formation of the orange-red allotropic form of elemental selenium (see above). Growth kinetics data obtained at different selenite concentrations showed that increasing the selenite concentration drastically reduced the maximum attainable cell concentration (Fig. 2). The cell concentrations of cultures containing 1.5 mM selenite reached only about 15% of the cell concentration of the control in the stationary phase. Increasing in the selenite concentration to 2.0 mM had only a slight additional effect on growth. The cell protein concentration decreased slightly at the

beginning of the stationary phase in cultures amended with 0.5 and 1.0 mM selenite. In the other cultures, the cell protein concentration remained more or less constant during the stationary phase. The decrease in the selenite concentration during growth is shown in Fig. 3. This decrease was fast and was complete at selenite concentrations of 0.5 and 1.0 mM. In both cases total transformation of selenite occurred in the same period of time, whereas the cell densities of the cultures amended with 1.0 mM selenite were about 30% lower than the cell densities of the cultures amended with 0.5 mM selenite. Thus, the reduction rate was proportional to the initial selenite concentration but independent of the cell density of a culture. In cultures amended with 2 mM selenite only the beginning of the reduction was fast. The rate drastically decreased after a few hours, showing that the cultures were no longer able to maintain the reaction. Both the growth kinetics and the extent of selenite reduction were highly reproducible when cultures were started from the same bacterial colony, but both of these parameters varied when cultures were started from different colonies. In some cultures amended with 1.5 mM selenite reduction was complete with the same slope as the slope in cultures amended with 0.5 and 1.0 mM selenite, while in other cultures the reduction kinetics were similar to the reduction kinetics of cultures containing 2.0 mM selenite (data not shown). On the other hand, the reduction kinetics slope, as shown in Fig. 3, was nearly the same for every experiment, but the start of the reaction, which always coincided with the cultures entering the stationary phase (Fig. 2 and 3), could be delayed for a few hours. The specific extracellular turbidity (turbidity not due to cell material) was represented by the absorbance at 650 nm normalized to the cell protein concentration. Figure 4 shows that this turbidity increased with time in cultures amended with selenite, starting at the beginning of the reduction phase, and also increased with selenite concentration. This indicates that

4738

KESSI ET AL.

APPL. ENVIRON. MICROBIOL.

particulate material resulting from the reduction of selenite was released into the culture medium. Electron micrographs of oxic and anoxic cells grown with selenite are shown in Fig. 5A and B, respectively. In both cases, high-electron-density particles having a regular geometrical shape were present in the cytoplasm. In anoxic cultures such particles were also found in the culture medium. They could be sedimented by centrifugation at 100,000 ⫻ g without changing their regular geometrical shape (Fig. 6A) or their orange-red color. The particles contained about 20 mg of protein/mmol of selenium. The orange-red amorphous allotrope produced by the bacterial cells was stable for months in aged cultures, as was the chemically reduced selenium in a cell-free spent medium obtained from a stationary-phase selenite-containing culture. In contrast, selenite reduced by thioglycolate in fresh medium was converted to a black form within a few days (data not shown). An electron micrograph of selenium-containing particles formed by reduction of selenite with thioglycolate in a stationary-phase medium is shown in Fig. 6B. The shape of the particles was similar to the shape of the particles produced by the bacterial cells, but the particles were 15 to 20 times larger. When preparations of cells and selenium-containing particles obtained from the medium of a culture amended with 0.5 mM selenite were analyzed by using EDX analysis, the electrondense particles produced specific selenium absorption peaks at 1.37 keV (peak SeL␣), 11.22 keV (peak SeK␣), and 12.49 keV (peak SeK␤) (Fig. 7). The effect of the selenium-containing particles on the buoyant density of the cells was obvious in the distribution pattern obtained after gradient centrifugation. A discontinuous gradient resulted in sharp bands when centrifugation proceeded to isopycnic equilibrium. Cells grown with selenite obtained at about the middle of the reduction phase migrated to the top of the 2.5 M sucrose layer, whereas control cells migrated to the 2.0 M sucrose layer (Fig. 8). More interestingly, the buoyant density of the cells decreased again after reduction of selenite was complete. Indeed, large portions of the cells exposed to 0.5, 1.0, and 1.5 mM selenite and a smaller portion of the cells grown with 2.0 mM selenite exhibited a density equal to the density at the boundary of the 2.0 M sucrose layer (Fig. 8 and Table 1). This clearly demonstrated that there was only a transitory increase in the buoyant density of cells in the presence of selenite. This was confirmed by electron micrographs of cells grown with 1.0 mM selenite and sampled during and after the reduction phase. Clearly, more electron-dense particles were present in the cytoplasm during reduction than after reduction was complete (Fig. 9). Soluble proteins were present in the upper fractions of the density gradients. The concentration of these proteins increased with the age and the selenite concentration of the cultures (Fig. 8). DISCUSSION The ability of R. rubrum to reduce selenite is clearly greater under anoxic phototrophic growth conditions, under which selenite concentrations up to 1.5 mM are completely reduced, than under oxic growth conditions, under which a selenite concentration of 0.5 mM is only reduced to about 0.375 mM (Fig. 1A and 3). The slight increase in selenite concentration during the stationary phase under oxic conditions suggests that elemental selenium is slowly reoxidized; no reoxidation is evident under anoxic conditions. Selenite reduction is closely related to the growth kinetics of cultures and occurs only when the cells reach the transition

FIG. 7. EDX analysis of electron-dense particles formed by R. rubrum in photosynthetically grown cultures amended with 0.5 mM selenite. (A) Particles in the cell cytoplasm. (B) Particles in the culture medium. Energy levels (in kiloelectron volts) are indicated on the x axis. The emission lines for selenium are at 1.37 keV (peak SeL␣), 11.22 keV (peak SeK␣), and 12.49 keV (peak SeK␤).

between the exponential and stationary phases. This finding is consistent with the results of Van Fleet-Stalder et al. (17), who reported that maximum production of volatile selenium compounds occurs during the late stationary phase. This suggests that both reduction reactions are controlled by stationaryphase regulatory molecules. Increasing the selenite concentration in the culture medium leads to a drastically lower cell concentration in the stationary phase (Fig. 2), suggesting that selenite markedly affects the cell division process and is a strong stress factor for R. rubrum. The great stability of the orange-red allotropic form of Se0 produced by the bacteria or precipitated in a cell-free medium obtained from a stationary-phase culture implies that Se0 is tightly bound to some compound produced by the cells and is protected from transformation into the black form. A protein content of about 20 mg of protein/mmol of selenium is always found in suspensions of selenium-containing particles isolated from culture media after growth in the presence of 0.5 mM selenite, suggesting that a selenium-protein complex is present. The slight decreases in cell protein concentration observed during growth of cultures containing 0.5 and 1.0 mM selenite (Fig. 2) may be due to the excretion of such a complex. This hypothesis is supported by the fact that stable orange-red selenium-containing particles are formed in cell-free spent medium (Fig. 6B). A selenium-protein complex may also be partially responsible for the protein present in the top fractions of

VOL. 65, 1999

REDUCTION OF SELENITE AND DETOXIFICATION OF SELENIUM

4739

FIG. 8. Protein concentration profiles after sucrose density gradient centrifugation of R. rubrum cells grown under anoxic conditions in the presence of different concentrations of selenite. (A) Control cells (no selenite). (B) Cells grown in the presence of 0.5 mM selenite. (C) Cells grown in the presence of 1.0 mM selenite. (D) Cells grown in the presence of 1.5 mM selenite. (E) Cells grown in the presence of 2.0 mM selenite. Cultures were analyzed during the reduction phase (F) and after completion of the reduction phase (■). Preparations were centrifuged for 2 h at 60,000 ⫻ g by using a sucrose step gradient (see text).

the sucrose gradient. Other proteins present in these fractions probably originate from cells lysed during centrifugation. Electron micrographs showing intact cells after selenite reduction (Fig. 9B) and growth kinetics showing that the cell protein concentration in the stationary phase is rather constant (Fig. 2) suggest that cells are not severely damaged in the presence of selenite. On the other hand, large amounts of selenium-containing particles are present in the culture medium after selenite reduction (Fig. 4, 6A, and 7B), indicating that R. rubrum is able to efficiently transport elemental selenium out of the cell. This hypothesis is supported by the results of ultracentrifugation experiments showing that the buoyant density of cells increases in the presence of selenite during the

TABLE 1. Change in the buoyant density of cells grown in the presence of different concentrations of selenitea Selenite concn in culture medium at zero time (mM)

Cells sedimenting on top of the 2.0 M sucrose layer during the reduction phase (% of total) (n ⫽ 2)

Cells sedimenting on top of the 2.0 M sucrose layer after the reduction phase (% of total) (n ⫽ 2)

0.0 0.5 1.0 1.5 2.0

100.0 88.2 ⫾ 1.5 4.2 ⫾ 1.2 7.2 ⫾ 0.8 3.3 ⫾ 1.8

100.0 100.0 ⫾ 1.5 87.4 ⫾ 1.2 82.4 ⫾ 1.7 15.2 ⫾ 2.0

a Cultures were analyzed during reduction and 1 to 2 days after reduction was complete.

reduction phase and then reverts to the buoyant density of control cells after the reaction is complete (Fig. 8). This hypothesis, however, differs from that of Tomei et al. (16). These authors also observed that selenium-containing particles formed in the cytoplasma of Desulfovibrio desulfuricans growing in the presence of selenite and that red elemental selenium accumulated in the media in the stationary phase. Nevertheless, they concluded that release of elemental selenium into the culture medium occurs by cell lysis. Losi and Frankenberger (9) observed more or less spherical protrusions on the surfaces of Enterobacter cloacae cells grown in the presence of selenite, as well as selenium-containing particles in the culture medium, but no intracellular Se was present. These authors suggested that selenite reduction occurs via a membrane-associated reductase(s), followed by rapid expulsion of the Se particles. This mechanism is not consistent with our results without reservation. Transport of selenium through the membrane by a classic transport mechanism, such as a membrane channel, would imply that the selenium-containing particles present in the cytoplasm have to become disaggregated to form small particles the size of a molecular complex. These small particles would be transported out of the cell, and the large particles observed in the culture medium (Fig. 6A and 7B) would then be formed by extracellular aggregation. Such a mechanism would require an extremely large amount of energy. Thus, we suggest that a vesicular mechanism of excretion occurs in R. rubrum. Vesicular excretion in bacteria is still controversial. However, Zusheng et

4740

KESSI ET AL.

APPL. ENVIRON. MICROBIOL.

FIG. 9. Electron micrographs of R. rubrum cells grown phototrophically in the presence of 1 mM selenite. (A) Cells during reduction. A total of 73% of the cells contained electron-dense particles in their cytoplasm (n ⫽ 65). (B) Two days after reduction was complete. Electron-dense particles were found in only 4.2% of the cells (n ⫽ 70).

al. (20) described naturally produced membrane vesicles isolated from 15 strains of gram-negative bacteria. Also, export and intercellular transfer of DNA via membrane blebs have been observed in Neisseria gonorrhoeae (2). It will be a challenge to elucidate whether elemental selenium is expelled in the form of small atomic aggregates or whether vesicular expulsion occurs in R. rubrum. ACKNOWLEDGMENTS Janine Kessi thanks T. G. Chasteen of Sam Houston University, Houston, Tex., for very fruitful discussions, as well as for reading the manuscript and judicious comments. We thank H. P. Gautschi of the Laboratory for Electron Microscopy, University Hospital, Zurich, Switzerland, for preparing figures from the EDX analysis spectra. REFERENCES 1. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–252. 2. Dorward, D. E., C. F. Garon, and R. C. Judd. 1989. Export and intercellular transfer of DNA via membrane blebs of Neisseria gonorrhoeae. J. Bacteriol. 171:2499–2505. 3. Gattow, G., and G. Heinrich. 1964. Thermochemistry of selenium. II. Conversions of crystalline selenium modifications. III. Conversion of amorphous selenium modifications. Z. Anorg. Allg. Chem. 331:256–288. 4. Geering, H. R., E. E. Cary, L. H. P. Jones, and W. H. Allaway. 1968. Solubility and redox criteria for the possible forms of selenium in soils. Soil Sci. Soc. Am. Proc. 32:35–40. 5. Gerrard, T. L., J. N. Telford, and H. H. Williams. 1974. Detection of selenium deposits in Escherichia coli by electron microscopy. J. Bacteriol. 119: 1057–1060. 6. Harrison, G. I., E. J. Laishely, and H. R. Krouse. 1980. Stable isotope fractionation by Clostridium pasteurianum. 3. Effect of SeO3⫺2 on the physiology and associated sulfur isotope fractionation during SO3⫺2 and SO4⫺2 reductions. Can. J. Microbiol. 26:952–958. 7. Hess, W. M. 1966. Fixation and staining of fungus hyphae and host plant root

tissues for electron microscopy. Stain Technol. 41:27–35. 8. La ¨uchli, A. 1993. Selenium in plants: uptake, function and environmental toxicity. Bot. Acta 106:455–468. 9. Losi, M. E., and W. T. Frankenberger, 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. 10. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265–275. 11. Moore, M. D., and S. Kaplan. 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. 12. Ohlendorf, H. M. 1989. Bioaccumulation and effects of selenium in wildlife, p. 133–177. In L. W. Jacobs (ed.), Selenium in agriculture and the environment. American Society of Agronomy, Madison, Wis. 13. Silverberg, B. A., P. T. S. Wong, and Y. K. Chau. 1976. Localization of selenium in bacterial cells using TEM and energy dispersive X-ray analysis. Arch. Microbiol. 107:1–6. 14. Sistrom, W. R. 1960. A requirement for sodium in the growth of Rhodopseudomonas sphaeroides. J. Gen. Microbiol. 22:778–785. 15. Tomei, F. A., L. L. Barton, C. L. Lemanski, and T. G. Zocco. 1992. Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Can. J. Microbiol. 38:1328–1333. 16. Tomei, F. A., L. L. Barton, C. L. Lemanski, T. G. Zocco, N. H. Fink, and L. O. Sillerud. 1995. Transformation of selenate and selenite to elemental selenium by Desulfovibrio desulfuricans. J. Ind. Microbiol. 14:329–336. 17. Van Fleet-Stalder, V., H. Gu ¨rleyu ¨k, R. Bachofen, and T. G. Chasteen. 1997. Effects of the variation of growth conditions on the production of methyl selenides in cultures of Rhodobacter sphaeroides 2. 4. 1. amended with selenium oxyanions. J. Ind. Microbiol. Biotechnol. 19:98–103. 18. Watkinson, J. H. 1966. Fluorimetric determination of selenium in biological material with 2,3-diaminonaphtalene. Anal. Chem. 38:92–97. 19. Yamada, A., M. Miyashita, K. Inoue, and T. Matsunaga. 1997. Extracellular reduction of selenite by a novel marine photosynthetic bacterium. Appl. Microbiol. Biotechnol. 48:367–372. 20. Zusheng, L., A. J. Clarke, and T. J. Beveridge. 1998. Gram-negative bacteria produce membrane vesicles which are capable of killing other bacteria. J. Bacteriol. 180:5478–5483.