Environ. Sci. Technol. 2003, 37, 701-706
Cd Bioaccumulation by a Freshwater Bacterium, Rhodospirillum rubrum A . S M I E J A N , † K . J . W I L K I N S O N , * ,‡ A N D C. ROSSIER† Laboratory of Bacteriology and Microbial Ecology, Department of Botany and Plant Biology, University of Geneva (Bastions), 3 Place de l’Universite´, 1211 Geneva 4, Switzerland, and Analytical and Biophysical Environmental Chemistry (CABE), University of Geneva (Sciences II), 30 Quai Ernest Ansermet, 1211 Geneva 4, Switzerland
Cd bioaccumulation by Rhodospirillum rubrum, a Gramnegative freshwater bacterium, was studied in a synthetic medium. The free ion (Cd2+) was the best predictor of the Cd internalization fluxes. Representation of the shortterm uptake fluxes as a function of [Cd2+] in the medium demonstrated a linear relationship, as would be expected for a rate-limiting, first-order internalization with a single transporter. Nonetheless, several different accumulation profiles were observed, depending on the Cd concentration. Cd uptake was regulated differently for concentrations above and below 10-6 M (or was regulated only above [Cd2+] ) 10-6 M). Short-and long-term studies revealed that regulation was rapidly initiated for the highest Cd concentrations examined, effectively decreasing both adsorbed and internalized Cd. Anodic stripping voltammetry demonstrated that a Cd complexing ligand was produced within minutes upon exposure to 5 × 10-6 M Cd2+ and that an extracellular sequestration of Cd was one mechanism regulating Cd uptake. Competition studies with other cations revealed a competitive inhibition of Cd uptake by Zn and an uptake enhancement in the presence of Mn and Cu.
of the mechanistic studies have been performed at high, environmentally unrealistic, concentrations of metal with little attention paid to the chemical speciation of the metal in solution. One aim of this study was therefore to examine metal bioaccumulation as a function of the free metal ion concentration in solution ([Mz+]) and to verify if the free ion was the best predictor of internalization fluxes and biological effects (8) at both low and high [Cd2+]. In its simplest form, there are two important and testable hypotheses of the FIAM: (i) a linear relationship between uptake and [Mz+] (internalization rates should be directly proportional to the activity of the free metal ion in solution) and (ii) at a given concentration of Mz+ (bulk concentration of Mz+ below that for which the membrane transporter sites are saturated), internalization should be linear with time. In this case, neither the physical transport nor the chemical reaction of the metal at the biological interface is rate-limiting, and internalization fluxes should be proportional to [Mz+], irrespective of the actual chemical species that reacts with the biologically sensitive site. In this study, we compared Cd accumulation at [Cd2+] typical of contaminated aquatic environments with that observed at the higher concentrations generally employed for the mechanistic studies. Because bacterial resistance mechanisms are well-documented at high metal concentrations (9, 10), a goal of this study was to determine whether these mechanisms are similar at (lower) more environmentally relevant conditions and if the presence of bacterial resistance will modify our capacity to predict uptake fluxes. A Gram-negative bacteria, Rhodospirillum rubrum, was examined because its cell wall, more complex than that of the Gram-positive cells, has an outer membrane that controls the passage of some substances in and out of the cell. At first view, this membrane might provide an additional barrier for the biological internalization of metals that might make it difficult to predict uptake fluxes. Finally, the influence of several other metal ions was examined to verify if the Cd uptake system differentiates between structurally similar ions and provide another rigorous test of the simple uptake models.
Materials and Methods Introduction In aquatic systems, chemical speciation is known to be an important factor that influences biological availability. Many trace metal uptake models including the free ion activity model (FIAM) and the biotic ligand model (BLM) assume that biological internalization is rate-limiting and first-order and that toxicity can be related to body burdens or uptake fluxes. In this case, the transport rate across the membrane (uptake flux) can be assumed to be directly proportional to the free metal ion concentration in solution (basis of the FIAM) or to the concentration of surface transporter-bound metal (basis of the BLM). While this assumption has often been verified for the uptake of toxic trace metals by phytoplankton (e.g., refs 1-3) and fish (e.g., ref 4), there are relatively few quantitative examinations of trace metal uptake rates to aquatic bacteria (e.g., refs 5 and 6) despite a relatively large number of papers that have examined the molecular mechanisms of the uptake process (e.g., ref 7). Indeed, many * Corresponding author phone: (4122)702 6053; fax: (4122)702 6069; e-mail:
[email protected]. † University of Geneva (Bastions). ‡ University of Geneva (Sciences II). 10.1021/es025901h CCC: $25.00 Published on Web 01/07/2003
2003 American Chemical Society
Organism and Growth Medium. Rhodospirillum rubrum S1 (DSM No. 467), is a Gram-negative bacterium that belongs to the purple nonsulfur bacteria group (anoxygenic phototrophic bacteria). R. rubrum was cultivated anaerobically in Sistrom medium (11, 12) at pH 7 and 30 °C under constant light conditions on a rotary shaker at 100 rpm. At the end of the exponential growth phase (ca. 6 × 108 cells mL-1 after 3.5 days), the bacteria were harvested by a 15-min centrifugation at 1200g. The pellet was washed (1×) in 50 mL of a 0.05 M KNO3 solution, then recentrifuged, and resuspended (2×) in an experimental medium (see below). Experimental Medium. The experimental medium was purposely chosen to be inert to the bacteria and simple enough to allow accurate speciation calculations. Cells from the bacterial concentrate (ca. 1010 cells mL-1) were inoculated into a defined volume of 0.02 M HEPES (hydroxy piperazine ethane sulfonic acid) at pH 7.0 without a carbon source. The final cell concentration for the bioaccumulation experiments was ca. 5 × 107 cells mL-1. During the short-term experiments, pH was sufficiently buffered to ensure a constant value. Cd was added as the nitrate salt. Cadmium speciation was determined using MINEQL+ 3.01a with updated equilibrium constants. Under the experimental conditions employed here, VOL. 37, NO. 4, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
701
TABLE 1. Cd2+ Concentrations Determined prior to and following Diluton with Sistrom Mediuma
a
Cd2+ in HEPES medium (M)
Cd2+ following dilution with Sistrom (M)
10-3 10-4 10-5 10-6 10-7 10-8
1.5 × 10-6 1.0 × 10-8 3.5 × 10-10 2.0 × 10-11 1.9 × 10-12 1.9 × 10-13
Ionic strength (100 mM) and pH (7) were maintained constant.
the HEPES should not complex the cadmium (13) so that [Cd2+] was nearly equal to the total Cd added to the solutions. Growth Inhibition Test. To investigate the effects of [Cd2+] on cell growth, 107 cells mL-1 were incubated for 24 h at 30 °C in sterile polypropylene tubes containing 0.02 M HEPES and variable [Cd] at pH 7. Effects on cell growth were determined indirectly by cell density measurements following a 10-fold dilution with the Sistrom growth medium and further incubation. Growth was evaluated by comparing cell densities obtained 50 h after transfer to the Sistrom medium. The NTA and EDTA present in the Sistrom medium effectively ended the Cd exposure by greatly reducing the residual Cd2+ concentrations (Table 1). Bioaccumulation Experiments. After resuspension of R. rubrum in the experimental medium containing Cd, samples were filtered through a 0.22-µm hydrophilic polymer membrane (PVDF Durapore membrane, Millipore) at defined times. The filtrate was acidified to pH 2 with ultrapure nitric acid to allow for measurements of total dissolved Cd. Intracellular (internalized) and adsorbed Cd were determined using an EDTA washing procedure described below (3, 14). Cd concentrations (total dissolved, adsorbed, and internalized Cd) were measured either by atomic absorption (PhilipsPye-Unicam SP9) or by ICP-MS (Hewlett-Packard 4500) for the lower concentrations (e10-7 M Cd). In the short-term experiments, cells were harvested after a defined contact time (0.5, 1.5, 4.5, 9, and 13.5 min) with the experimental medium. At each time, 10 mL of an EDTA stock solution at pH 7 was added to 150 mL of the sample in a polypropylene flask where it was shaken for 1 min at 150 rpm. The concentration of the EDTA stock solution was 0.02 M for experiments performed at 10-7 and 10-8 M Cd or 0.2 M for experiments with 10-4, 10-5, and 10-6 M Cd. Suspensions were then filtered with a polysulfone filtration device using a manual vacuum pump. Filters with bacteria were rinsed with a further 5 mL of diluted (10×) EDTA solution, digested in 25 mL of a 1 M HNO3 solution that was heated for 10 min at 100 °C, and left overnight at room temperature. The non-EDTA-extractable fraction obtained by measuring the Cd in the bacterial digests corresponded to a cell “internalized” Cd fraction. In the short-term experiments, adsorbed Cd was not determined. Data were rejected if the mass balance of the EDTA wash plus the internalized Cd exceeded a 10% variation from the initial dissolved concentration. The Cd internalization flux (Jint) was determined from the slope of internalized Cd as a function of time for initial data points ( 18 mΩ‚cm; TOC < 2 µg of C L-1), and all experiments were performed under laminar flow conditions.
Results and Discussion Growth Inhibition. The Sistrom medium usually employed for the growth of R. rubrum was not appropriate to rigorously study Cd bioavailability because it contains too many Cd chelating agents. Although it is possible to perform thermodynamic calculations to determine Cd speciation, such
FIGURE 1. Growth inhibition: optical density (660 nm) of the bacterial suspensions following a 24-h incubatoin in Cd-containing HEPES medium followed by a 50-h recovery in Sistrom growth medium. Error bars represent standard deviations, N ) 3. calculations are estimates at best in such complicated media. Indeed, preliminary experiments in the Sistrom medium (data not shown) revealed no significant differences among growth curves in the presence of 10-8-10-3 M added Cd with respect to the control. It was clear from these experiments that the complexing agents present in the medium drastically reduced the free [Cd2+] and suppressed toxic effects. Therefore, to evaluate the acute toxicity of the Cd solutions, bacteria were first exposed to a Cd-containing, noncomplexing medium for 24 h. Subsequent dilutions into the Sistrom medium effectively reduced [Cd2+] by ca. 1000× (Table 1). No growth inhibition was observed over 50 h for initial [Cd2+] ranging from 10-8 M to 5 × 10-6 M (Figure 1). On the other hand, for g10-5 M Cd2+, a lag in the resumption of exponential growth was observed. After 50 h of “recovery” in the Sistrom medium, cell densities were 43% of control values for 10-5 M Cd2+. For 10-4 and 10-3 M Cd2+, significant lasting reductions in viable cell numbers were noted (observations performed up to 116 h). Complementary experiments (not shown) confirmed that 50 h were not enough for the cells to adapt to these Cd concentrations. In subsequent experiments, cell numbers were maintained constant among the different experimental conditions by examining uptake over short-term manipulations. Cd Uptake. Short term (0-14 min) internalization fluxes were determined for [Cd2+] from 10-8 to 10-4 M (Figure 2). A simple linear uptake was not observed at all [Cd2+]. At low Cd concentrations (10-7 and 10-8 M), Cd accumulation was indeed linear as is assumed in the simple, thermodynamically based uptake models such as the FIAM and BLM. On the other hand, for the highest Cd concentrations examined (10-6-10-4 M), the net internalization fluxes (slopes in the lines in Figure 2) decreased with time. This result suggested that, for the higher Cd concentrations, regulatory mechanisms played an important role even for the short-term uptake kinetics examined here. Indeed, these data suggested that regulation, where present, was initiated in less than 5 min. Nonetheless, the representation of the slopes of the initial linear portion of the short-term uptake curves as a function of [Cd2+] showed a linear relationship as would be expected for a rate limiting, first-order internalization (Figure 3). Furthermore, the observation of a single first-order uptake curve implied that a single transporter was involved since the interaction of Cd with a second transporter with another affinity would be reflected by a break in the slope of Figure 3 (e.g., ref 18). The previous experiments suggested that R. rubrum showed some resistance to high Cd concentrations (g10-6 M). To get more insight into the possible regulation mechanisms, uptake was therefore examined for a 2-h exposure to 5 × 10-6 M Cd2+ (Figure 4). Initial cadmium adsorption
FIGURE 2. Short-term (14 min) uptake of Cd. Points represent internalized (non-EDTA-extractable) Cd as funtion of exposure time. Error bars represent standard deviations, N ) 4.
FIGURE 3. Cd internalization fluxes determined in the initial 1.5 min of an uptake experiment (cf. Figure 2) as a function of [Cd2+] in the external medium. Error bars represent standard deviations that are given when larger than the symbol size, N ) 4.
FIGURE 4. Representative Cd adsorption to the bacterial surface (EDTA-extractable Cd, b) and Cd internalization (non-EDTAextractable Cd, 0) as a function of exposure time following an exposure to 5 × 10-6 M Cd2+. on the bacterial surface was significant and rapid, on the order of 1 min, with an adsorption maximum of 1.2 × 10-10 mol cm-2 (45 µmol g-1 dry weight) in the first few minutes VOL. 37, NO. 4, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
703
of the experiment. Adsorbed Cd subsequently decreased to extremely low levels (
