Uptake of Hydrophobic Metal Complexes by Three Freshwater Algae

Apr 6, 2009 - Rio Tinto Alcan - Centre de Recherche et Développement. Arvida, 1955 boulevard Mellon, C.P. 1250, Jonquie`re, Québec,. Canada G7S 4K8...
0 downloads 0 Views 249KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 3308–3314

Uptake of Hydrophobic Metal Complexes by Three Freshwater Algae: Unexpected Influence of pH A M I E L B O U L L E M A N T , †,‡ MICHEL LAVOIE,† CLAUDE FORTIN,† AND P E T E R G . C . C A M P B E L L * ,† INRS-Eau Terre et Environnement, Universite´ du Que´bec, 490 de la Couronne, Que´bec, Que´bec, Canada G1K 9A9, and Rio Tinto Alcan - Centre de Recherche et De´veloppement Arvida, 1955 boulevard Mellon, C.P. 1250, Jonquie`re, Que´bec, Canada G7S 4K8

Received October 6, 2008. Revised manuscript received March 2, 2009. Accepted March 11, 2009.

Cadmium forms neutral, lipophilic Cd(L)20 complexes with diethyldithiocarbamate (DDC) and with ethylxanthate (XANT). Uptake of these complexes by three unicellular freshwater green algae (Chlamydomonas reinhardtii, Chlorella fusca, and Pseudokirchneriella subcapitata) was determined at two pH values (7.0 and 5.5) and compared to uptake of the free, uncomplexed Cd2+ ion. Uptake of the lipophilic complexes over time, characterized by high initial uptake rates but tending toward a plateau after about 30 min, could be modeled successfully as the result of the following processes: firstorder uptake from solution, partitioning of the accumulated Cd into two internal pools (labile and nonlabile), and first-order loss of Cd from the labile pool. At pH 7.0 initial uptake rates for both Cd(L)20 complexes were much higher than for Cd2+ alone (e.g., up to ∼90 times higher for comparable dissolved Cd concentrations of ∼0.4 nM). However, the initial uptake rates for the lipophilic complexes dropped dramatically when the pH was lowered from 7.0 to 5.5 (2- to 60-fold decrease, depending on the algal species and the nature of the neutral complex). Loss rates for the accumulated complexes also decreased at the lower pH. The lipophilicity of the neutral complexes, as estimated from their octanol-water distribution coefficient (Dow), was not affected by the decrease in pH from 7.0 to 5.5. We thus conclude that the acidification of the external medium, i.e., the interaction of protons with the algal membrane, strongly affects algal membrane permeability.

Introduction Interactions of metals with aquatic organisms are normally well described within the constraints of the Free Ion Activity Model (FIAM) (1) or its more recent incarnation, the Biotic Ligand Model (BLM) (2). In its present form, the BLM takes into account the influence of three water quality parameters on metal bioavailability: pH ([H+]), hardness ([Ca2+], [Mg2+]), and dissolved organic matter ([DOM]). In the case of pH, the model considers H+-ion competition for metal-binding sites at the biological surface (the so-called “biotic ligand”), and * Corresponding author phone: 1-418-654-2538; fax: 1-418-6542600; e-mail: [email protected]. † INRS-Eau Terre et Environnement, Universite´ du Que´bec. ‡ Rio Tinto Alcan - Centre de Recherche et De´veloppement Arvida. 3308

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009

for metal-binding sites in solution (dissolved ligands, including natural DOM). For biological targets ranging from unicellular algae to fish, this modeling approach has proved remarkably successful in describing the uptake and (acute) toxicity of divalent cationic metals such as cadmium (3-5). The BLM assumes that intact metal complexes cannot cross biological membranes. This assumption is based upon metals forming anionic, hydrophilic complexes with ligands such as EDTA and NTA (e.g., Cd(EDTA)2- and Cd(NTA)-). This idea that complexed metal will be biologically unavailable breaks down when the complex is electrically neutral and lipophilic. Such cases constitute one of the wellrecognized classes of exceptions to the BLM, where intact complexes enter the living cell by passive diffusion across the plasma membrane (i.e., bypassing the normal facilitated transport route). Early examples of this phenomenon were reported by Poldoski (6) and by Florence and co-workers (7-9), demonstrating exceptions to the FIAM/BLM for Cd and Cu complexes with diethyldithiocarbamate, 8-hydroxyquinoline, and xanthates. Recent laboratory studies (10, 11) have confirmed this early work, and recent field observations suggest that such complexes are indeed present in some natural waters (12). Without exception, these prior experiments with neutral lipophilic metal complexes were carried out at circumneutral pH. Given that the complexes are presumed to enter the cell by passive diffusion across the phospholipid membrane, it was not expected that this uptake route would be sensitive to pH changes (unlike the normal facilitated transport uptake route, which is known to be affected by H+-ion competition). However, in a recent study (13) we demonstrated that uptake of the Cd(DDC)20 complex by a unicellular alga, Pseudokirchneriella subcapitata, was lower at pH 6.0 and 5.5 than at pH 7.0. The present work was undertaken to determine the generality of this unexpected pH effect. We initially chose one alga, C. reinhardtii, and studied the short-term uptake and loss at pH 7.0 and 5.5 of two neutral complexes, Cd(DDC)20 and Cd(XANT)20, with differing lipophilicities. We then examined the uptake and loss of the same neutral complex, Cd(DDC)20, by two other unicellular algae (Chlorella fusca var. vacuolata, Pseudokirchneriella subcapitata) at two pH values (7.0 and 5.5). Uptake kinetics were different for the two complexes, and for a given complex they also varied from one algal species to another, but in all cases uptake and loss were lower at pH 5.5 than at neutral pH.

Experimental Procedures Reagents and Glassware. All glass, Teflon, and polycarbonate containers were soaked for at least 24 h in 10% HNO3, thoroughly rinsed seven times with ultrapure water (resistivity >17 Mohms · cm; Milli-Q3RO/Milli-Q2 system, Bedford, MA) and dried under a laminar flow hood prior to use. All stock solutions were prepared with ultrapure water. Reagents used for culture media and experiments were of analytical grade or better. Sodium diethyldithiocarbamate trihydrate (DDC: ACS quality, purity >99%) was purchased from Sigma Aldrich (St. Louis, MO) and potassium ethylxanthate (XANT, purity >98%) came from Alfa Aesar (Ward Hill, MA). Stock solutions of these ligands were prepared in methanol (ACS/HPLC grade; Fisher Scientific, Fairlawn, NJ). Noncomplexing buffers (10 mM) were used to buffer the exposure solutions and culture media at pH 7.0 (N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid], HEPES, purity >99.5%; Sigma Aldrich), and pH 5.5 (2-[N-morpholino]ethanesulfonic acid, MES, purity 10.1021/es802832u CCC: $40.75

 2009 American Chemical Society

Published on Web 04/06/2009

>99.5%; Sigma Aldrich). Radioactive cadmium (109Cd; 1 and 2.2 × 106 Bq · mmol-1, respectively, for experiments at pH 7.0 and 5.5) was purchased from Amersham Bioscience (Buckinghamshire, UK) and supplied as a 0.1 M HCl solution. Organisms and Culture Conditions. The metal uptake experiments were carried out with three species of unicellular freshwater algae: Pseudokirchneriella subcapitata (Korshikov) Hindak (formerly known as Selenastrum capricornutum); Chlamydomonas reinhardtii; and Chlorella fusca var. vacuolata (formerly known as Chlorella pyrenoı¨dosa). The algae were obtained from the University of Toronto Culture Collection (UTCC, Department of Ecology and Evolutionary Biology, Toronto, ON, Canada: strains UTCC37, UTCC11 and UTCC89, respectively). The original strain and stock cultures were maintained axenically in modified High Salt Medium (MHSM) (14). The composition of the solutions used to grow the algae, to rinse the algal cells, and to expose them to Cd can be found in Table S1. The basic medium was sterilized by autoclaving, and then supplemented with a filter-sterilized trace metal mix (0.2 µm polycarbonate membranes; 47 mm; Poretics, Osmonics, Livermore, CA). Batch cultures were maintained in 250 mL polycarbonate Erlenmeyer flasks under constant illumination at 130 µE · m-2 · s-1 (Cool White Fluorescent Tubes), with rotary agitation at 100 rpm and a temperature of 20 °C. For regular maintenance, a ∼2 mL sample of the culture was transferred to a fresh, sterile medium (pH 7.0) every week. Uptake Experiments. To minimize the influence of the algal cells on the chemistry of the exposure medium (i.e., depletion of metal complexes in solution), and to estimate the Cd internalization constant (ki), we chose to measure Cd accumulation in short-term uptake experiments (1000-fold excess of organic ligands (DDC, 1 µM; XANT, 100 µM) to ensure that virtually all of the Cd was present as the Cd(L)20 complex (see Supporting Information). The uptake experiments were performed in a simplified algal growth medium (i.e., the major anions and cations of the MHSM medium, minus the trace elements (see Table S1)). The design for the time-course experiments involved two exposure variables: pH and the organic ligand L. To prepare the exposure media a fresh organic ligand stock solution in methanol was prepared (100 mL of methanol containing about 50 mg of organic ligands, accurately weighed) a few hours prior to use. The appropriate volumes of ligand solution and diluted 109Cd solution were added to two 1-L Teflon bottles (control and assay) containing the simplified growth medium and the appropriate buffer. These solutions were allowed to equilibrate for 24 h at the ambient laboratory temperature (∼21 °C), filtered (0.2-µm), and then dispensed into nine individual 250-mL flasks (three time points × triplicate flasks) and inoculated with the algal cells. The algae used for the inoculum were grown at pH 7.0 as described above, and exponentially growing cells were harvested after 48 h (C. reinhardtii) or 72 h (P. subcapitata and C. fusca) by centrifugation (20,000g, 10 min). For experiments at pH 5.5, the algal cells were acclimated for a minimum of 8 h at the new pH (controlled by bubbling the suspension intermittently with pure CO2 (g)), and then centrifuged. A pH 5.5 was chosen to probe the uptake mechanism, and because it is an environmentally relevant pH (particularly in waters on the Canadian Precambrian Shield); pH 5.5 media are also readily tolerated by freshwater chlorophytes. To remove any algal exudates present in the initial algal pellet, the collected cells were washed three times with

simplified MHSM (10 mL), with centrifuging for 5 min (20,000g) after each wash. The final pellet was resuspended in simplified MHSM (10 mL; >106 cells · mL-1). The concentration of algal cells was determined rapidly with an electronic particle counter (Multisizer 3, 70 µm aperture; Coulter Electronics, Toronto, ON, Canada), and the appropriate aliquot of the algal suspension was added to each experimental medium (100 mL, in 250-mL polycarbonate flasks) to attain an initial cell density of ∼15,000 cells · mL-1. Uptake of the Cd(L)20 complexes was monitored by collecting the algal cells after different exposure times in the 3 to 40 min range; normally three exposure times were selected for each run, and for each time triplicate flasks were sacrificed, this kinetic experimental design being repeated three times (i.e., 3 × 3 × 3 ) 27 flasks used for each exposure condition). For each assay, a 2-mL subsample of the exposure medium was collected at the beginning of the exposure period for total Cd concentration determinations and after addition of the algal inoculum a second 2-mL subsample was removed for cell concentration and surface area measurements using the electronic particle counter. For each time point, a 96-mL subsample of the algal suspension was gently filtered onto two superimposed 0.4-µm polycarbonate filters and the harvested cells were rinsed three times with 10 mL of MHSMR, a simplified rinse medium. Filtration blanks of exposure media without algae were systematically determined (2 × 96 mL, one control medium without 109Cd and one assay medium with the radioisotope). To correct for passive retention of radioactivity by the polycarbonate filters, the activity of the lower filter was subtracted from that measured on the upper filter; uptake values were then normalized for the total algal surface area. Exposure duration was measured from the addition of the algal cells to the exposure media until the moment when the 96-mL subsample had completely passed through the filter membrane. Filter radioactivity was measured with a gamma counter (LBK Wallac Compugamma model 1282, Turku, Finland), in vials containing 2 mL of water to ensure constant sample geometry. A counting window of 16-32 keV was used (main peak at 22 keV) and the counting time was set to 2,000 s or to a maximum collection of 100,000 counts. Counts per minute (cpm) were converted into Cd molar concentrations, taking into account the detector efficiency, radioactive decay, and the 109Cd specific activity. 109Cd activities in the filtrates from the algal collection step (final time point) were measured to ensure that the exposure conditions hadn’t changed appreciably during the experiment (e.g., due to algal uptake of Cd); losses of Cd from solution were always less than 5%. Depuration Experiments. To determine the rate at which the Cd(L)20 complexes were eliminated from the algae, we first exposed the algal cells to the radio-labeled lipophilic complexes for 30 min, i.e., until the steady-state condition was approached. The algal cells were then collected by filtration (single 0.2 µm polycarbonate filter), rinsed rapidly on the filter with simplified MHSM (3 × 10 mL), and resuspended in simplified MHSM containing neither Cd nor the organic ligand. Algal cells were collected (and cell numbers measured) after t ) 0, 3, 12, and 30 min and their residual radioactivity was determined as described above for the uptake experiments. Determination of Octanol-Water Distribution Coefficients. Octanol-water distribution coefficients of the Cd(L)20 complexes were determined by a shake-flask method adapted from Paschke et al. (15). Details can be found in the Supporting Information. Ligand and metal concentrations were increased for determinations of the Dow values, to facilitate analyses in the octanol and aqueous phases: [Cd] ) 38 nM, labeled with 109Cd, specific activity 1.0 × 107 Bq · µmol-1; [DDC] ) 0.1 mM at pH 7.0, and 10 mM at pH 5.5; [XANT] ) 10 mM; pH 7.0 and 5.5. VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3309

Note that in this simplified approach, it is assumed that fL is constant, independent of the value of [Cd]A.

Results

FIGURE 1. General scheme for the bioaccumulation of Cd in the presence of ligands forming lipophilic Cd(L)20 complexes. Cadmium Speciation in the Exposure Media. Cadmium speciation in the exposure solutions was computed with the MINEQL+ chemical equilibrium program (15, 16). The thermodynamic data from the default database supplied with the MINEQL+ software (v. 4.5) were updated by comparison with the NIST Critical Stability Constants of Metal Complexes Standard Reference Database (17) and the appropriate complexation and acidity constants were added for DDC and XANT as ligands. The derivation of these constants is presented in the Supporting Information. Speciation calculations with these data confirmed that g96% of the Cd is present as the Cd(L)20 complex in the presence of DDC (1.0 µM) or XANT (100 µM), regardless of the solution pH. The remainder of the Cd was present almost entirely as the CdL+ complex, with the calculated contribution of the free Cd2+ ion being less than 0.1% (Table S2). Subcellular Partitioning of Cd. To determine the subcellular location of Cd in algal cells that had accumulated lipophilic Cd(L)20 complexes, a suspension of C. reinhardtii cells (150,000 cells · mL-1) was exposed to 109Cd(DDC)20 (0.38 nM) for 30 min. The suspension was centrifuged (10 min at 20,000g) and the pellet was rinsed (3 × 50 mL of MHSM-R solution, 1 × 50 mL of 0.2 mM EDTA, 2 × 50 mL of MHSM-R) and resuspended in a small volume (2.5 mL) of MHSM-R. This concentrated suspension of cells (∼ 30 × 106 cells · mL-1) was then subjected to sonication to rupture the cells, followed by differential centrifugation to obtain five subcellular fractions (cell debris, including cells wall fragments; granules; organelles; heat-stable peptides; heat-denaturable proteins), the 109Cd activity of which was determined by gamma counting. Details of the fractionation procedure are provided in the Supporting Information. Theoretical Considerations. The passive uptake of Cd(L)20 complexes can be described by a two-compartment model (Figure 1) derived from the Spacie and Hamelink model for the bioaccumulation of hydrophobic compounds (18). The model (eq 1) takes into account the internalization of the Cd in the complexed form (ki in L · m-2 · min-1) and the labile forms of Cd that can be eliminated from the unicellular alga (ke in min-1): d[Cd]A ) ki[Cd(L)02] - ke·fL·([Cd]A) dt

(1)

where [Cd]A is the total Cd concentration in the alga, normalized for the algal surface area (nmol · m-2), [Cd(L)20] is the concentration of the lipophilic complex in the exposure solution (nmol · L-1), and fL is the fraction of labile Cd concentration within the alga that is available for loss to the external medium. The uptake constant ki was determined from the initial uptake rates (i.e., when [Cd]A was very low and the second term in eq 1 was negligible), whereas the elimination constant ke and the labile fraction fL were estimated from the depuration experiments. The time-course of [Cd]A accumulation can be represented by eq 2: [Cd]A ) [Cd(L)02]·

3310

9

ki ·(1 - e-ke·fL·t) ke·fL

(2)

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009

Uptake of Cd(DDC)20 and Cd(XANT)20 by C. reinhardtii. We initially tested one of the algal species, C. reinhardtii, and followed the uptake at pH 7.0 and 5.5 of two neutral complexes, Cd(DDC)20 and Cd(XANT)20, with differing lipophilicities. Uptake curves for both complexes typically showed a rapid initial uptake phase followed by a trend toward a plateau (Figure 2). Note that the uptake curves all pass through the origin, since we are measuring the rate of uptake of the radio-labeled complexes, and [109Cd]A ) 0 at t ) 0. The initial uptake point, determined after 3 min, was used to estimate ki, the uptake rate constant. When algal cells were collected after 30 min and resuspended in a clean medium, some but not all of the accumulated Cd was eliminated (Figure 3). A first-order loss rate, ke, was determined for the elimination of the “labile” Cd fraction from the algal cells. Values for ki, ke, and fL are compiled in Table 1. For comparison purposes, the initial uptake rate constants for free Cd2+ in the absence of organic ligands were also determined at the two pH values. For equivalent dissolved Cd concentrations (0.38 nM), uptake rates were much higher for the Cd(L)20 complexes than for free Cd2+ (e.g., ∼90 times higher at pH 7.0 for both DDC and XANT; Table 1), a result that is consistent with the literature (6, 10, 11). This enhancement was still present at pH 5.5, but was less marked (23× for DDC, 42× for XANT). The decrease in pH from 7.0 to 5.5 had a negligible effect on uptake of Cd2+ by C. reinhardtii, but reduced uptake of the lipophilic complexes (6.5-fold for Cd(DDC)20; 3-fold for Cd(XANT)20). The unexpected effect of pH on uptake of neutral lipophilic complexes is thus confirmed. Loss rates for the labile intracellular Cd fraction were also affected by pH, in a manner similar to the uptake rates but to a lesser degree. For both lipophilic complexes, ke values were lower at pH 5.5 than at pH 7.0, this effect being slightly less for Cd(DDC)20 than for Cd(XANT)20: ratio ke(pH 7.0)/ ke(pH 5.5) ) 1.5 and 1.9, respectively. Uptake of Cd(DDC)20 by the Three Algal Species. In a second series of experiments we used a single lipophilic complex, Cd(DDC)20, as the probe and determined the uptake and elimination kinetics for the other two test algae, P. subcapitata and C. fusca. The uptake and elimination curves were similar in form for all three algae (Figures S2 and S3), but important interspecies differences were evident (Table 1). At pH 7.0 the initial uptake rate constants, ki, decreased in the order C. fusca g C. reinhardtii > P. subcapitata, whereas at pH 5 the order was slightly different (C. reinhardtii > C. fusca > P. subcapitata), but in all cases the initial uptake rates were much greater than those for free Cd2+ (Table 1, right column). With respect to the pH effect initially observed with C. reinhardtii, the suppression of uptake of the Cd(DDC)20 complex on decreasing the pH from 7.0 to 5.5 was even more important for the other two algae. The ratios ki(pH 7.0)/ ki(pH 5.5) decreased in the order P. subcapitata (60) > C. fusca (56) > C. reinhardtii (6.5). The rates of elimination of the labile intracellular Cd fraction were also lower at pH 5.5 than at pH 7.0, but this effect was much less marked than the effect on the uptake rates; the ratios ke(pH 7.0)/ke(pH 5.5) decreased in the order C. reinhardtii (1.5) > P. subcapitata (1.4) > C. fusca (1.0). Octanol-Water Distribution Coefficients. As expected, both Cd(L)20 complexes partitioned into water-saturated octanol (Table 2), demonstrating their lipophilicity. At pH 7.0 and for Cd:L ratios similar to those used for the algal experiments, the Dow values determined for the Cd(DDC)20 complex were higher than those measured for Cd(XANT)20.

FIGURE 2. Time course of the uptake of Cd(L)20 complexes at pH 7.0 and pH 5.5 by C. reinhardtii. The symbols represent the experimental points for three independent time-course experiments ((SD). The solid curve is modeled with eq 2 and the dashed lines represent 95% confidence intervals for the modeled uptake.

Discussion

FIGURE 3. Time course of the elimination of Cd from C. reinhardtii algal cells that were preloaded by exposure to Cd(DDC)20 and resuspended in fresh metal-free medium. The symbols represent the experimental points for three independent time-course experiments ((SD).

Determination of the Dow value for Cd(DDC)20 at pH 5.5 was, however, complicated by the fact that much of the free DDCwas extracted into the octanol phase. To maintain the concentration of DDC- remaining in the aqueous phase at a level sufficient to ensure that all the Cd was complexed as Cd(DDC)20, we increased the total DDC concentration from 1 to 100 µM. Under these conditions the Dow values determined for the Cd(DDC)20 complex remained higher than those measured for Cd(XANT)20.

General Considerations. Uptake of lipophilic organic solutes by living cells involves diffusion of the solute from the bulk solution to the cell surface, its partitioning between the aqueous solvent and the cell membrane, followed by its diffusion into the cytoplasm (Figure 1) (19). In principle, the uptake of lipophilic M(L)n0 complexes should follow the same route (20). The results of the subcellular partitioning experiment, run with Cd(DDC)20 and C. reinhardtii, support this conceptual model, in that after 30 min exposure, more than 80% of the radio-labeled Cd was found to be associated with the cytosol and organelle fractions, not with the cell wall (Table S6). Uptake measurements are reasonably straightforward, given that Cd speciation can be calculated and thus [M(L)n0] is well-constrained and values of the initial uptake constant, normalized for the total surface area of algae (ki; L · m-2 · min-1), can be readily extracted. In the present case, we used these constants to calculate the metal internalization flux (mol Cd · cm-2 · s-1) and compared this flux to the maximum possible diffusive flux of the Cd(L)20 complexes from the bulk solution to the algal surface, Jdif (Table S3). At pH 7.0 the measured uptake fluxes approached but were always less than the calculated diffusive flux across the phycosphere. In their study of the uptake of the lipophilic Cu(oxine)20 complex by five marine phytoplankton species at pH 8, Croot et al. (11) also noted that the initial uptake of the complex occurred close to the diffusion limit. At pH 5.5, however, the uptake fluxes measured in our experiments were well below the maximum possible flux from the bulk solution (see Discussion below). VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3311

TABLE 1. Kinetic Parameters for the Uptake and Elimination of Two Lipophilic Complexes: (A) Cd(DDC)20 and (B) Cd(XANT)20, by the Test Algae, at pH 7.0 and 5.5 (means ± SE, N = 3) (Initial Uptake Rate Constants for Cd2+ are Presented for Comparison) (A) Cd(DDC)20

5.5

7.0 5.5

-2

-1

alga

ki (10 · L · m · min )

C. reinhardtii P. subcapitata C. fusca C. reinhardtii P. subcapitata C. fusca

14 ( 2 8.7 ( 1.5 15 ( 1.4 2.1 ( 0.7 0.15 ( 0.02 0.27 ( 0.03

C. reinhardtii C. reinhardtii

12 ( 2 3.9 ( 0.8

pH 7.0

-1

Cd2+ -1

fL

ki (10 · L · m-2 · min-1)

0.34 ( 0.06 0.52 ( 0.05 0.30 ( 0.02 0.91 ( 0.02 0.16 ( 0.02 0.62 ( 0.09

0.15 ( 0.01 1.7 ( 0.1 1.8 ( 0.4 0.09 ( 0.01 0.007 ( 0.001

ke (min ) 0.33 ( 0.06 0.15 ( 0.01 0.21 ( 0.01 0.22 ( 0.002 0.11 ( 0.02 0.22 ( 0.03 (B) Cd(XANT)20 0.69 ( 0.02 0.36 ( 0.01

-1

0.07 ( 0.01

0.37 ( 0.02 0.34 ( 0.03

TABLE 2. Octanol-Water Distribution Coefficient (Dow) at pH 5.5 and 7.0 for Inorganic Cd and for Two Neutral Cd(L)20 Complexes (>99.9%)

pH 7.0 pH 5.5

Cd2+

Cd(DDC)20

Cd(XANT)20

5 × 10-4 (6 × 10-5 5 × 10-3 (1 × 10-3

270 ( 28 481 ( 73 a

60 ( 4 56 ( 1

a Value obtained with 10 mM DDC at pH 5.5 instead of 100 µM because of an unexpected extraction of DDC by the octanol (means ( SE, N ) 6).

The aforementioned analogy between lipophilic metal complexes and hydrophobic organic compounds (HOCs) breaks down once the metal complex enters the cell. Whereas HOCs remain intact within the cell and tend to partition to internal lipid phases (21), M(L)n0 complexes are subject to dissociation or ligand-exchange reactions in the intracellular environment; a priori, we cannot be sure what proportion of the M(L)n0 complex remains intact within the cell. If the M(L)n0 complex dissociates within the cell and M binds to various intracellular binding sites (e.g., glutathione, phytochelatins, proteins), the metal will tend to remain within the cell when the cells are removed from the exposure medium and placed in a metal-free solution (cf. 22, 23). In the present case, the elimination kinetics (Figures 3 and S3) bear this out; only a portion of the accumulated Cd is lost when the algal cells are resuspended in a fresh medium. In our modeling, we have assumed that the proportion of labile Cd is constant, i.e., that it does not change as a function of Cd accumulation. This is likely an oversimplification; as Cd builds up in the algal cells, one might expect the proportion of labile Cd to increase. However, when we tested the effect of allowing the proportion of labile Cd to increase as a function of the total algal Cd concentration, up to the value of fL that we measured in the algal cells collected after 30 min exposure, the effects on the modeled accumulation curves were modest and the agreement between observed and predicted accumulation was not noticeably improved. Despite this uncertainty regarding intracellular Cd speciation, the Cd accumulation curves calculated with eq 2, using the experimentally determined values ki, ke, and fL, match the trend in the observed values quite well (compare the experimental points with the solid curves in Figures 2 and S2); the only real exception is for the system Cd(XANT)20 at pH 5.5 for C. reinhardtii, where the predicted uptake curve exceeds the observed accumulation (suggesting that ki may be overestimated or ke underestimated in this case). Note that the net accumulation curves plotted in Figures 2 and S2 are not the result of simply fitting a curve through the experimental points. Rather, they correspond to the predicted accumulation based on independent measurements of ki, ke, and fL. We will use these three parameters when comparing the different algal species and when considering the effects 3312

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009

of pH changes on the uptake of the lipophilic Cd complexes. We initially consider uptake at pH 7.0, since this is closer to the pH employed by earlier workers in their studies on marine algae, and then consider the results obtained at pH 5.5. Comparison of Different Algal Species (pH 7.0). Differences in overall uptake among the three algal species can be ascribed to two factors acting simultaneously: (i) their membrane permeability, which would be expected to affect both the uptake and depuration rates; and (ii) the labile intracellular fraction (fL), which will affect the depuration rate (see eq 1). In the following section we first consider uptake of the lipophilic metal complexes, and then discuss the loss of metal from pre-exposed cells. Uptake. Previous workers (10, 11) have reported their uptake results in two ways, either in terms of a membrane permeability coefficient (Pmint) or as biological removal rate constants (kbio). In the first case, the permeability coefficient Pmint (cm · s-1) has been used to describe the flux of the lipophilic complex across the algal membrane: int Pm )

Klw·Dm Jint ) ∆z A·(∆[M(L)02])

(3)

where Klw is the partition coefficient between the aqueous solution and the algal membrane; Dm is the diffusion coefficient within the membrane (cm2 · s-1); ∆z is the membrane thickness (cm); A is the algal surface area (cm2 · cell-1); ∆[M(L)20] is the initial concentration gradient of the complex between the outside and the inside of the cell (mol · cm-3), with [M(L)20] ) 0 mol · cm-3 at t ) 0; and Jint is the metal internalization flux (mol · cell-1 · s-1). Permeabilities calculated in this manner for the Cd(DDC)20 complex at pH 7.0 did not vary greatly among the three test algae (from 14 to 26 × 10-4 cm · s-1) and the value for Cd(XANT)20 with C. reinhardtii (19 × 10-4 cm · s-1) was within this range. Values reported for marine phytoplankton are, with one exception, less than 8 × 10-4 cm · s-1 (Table S4), suggesting that the membranes of marine species may be somewhat less permeable toward lipophilic metal complexes than those of freshwater chlorophytes. Differences in the lipid composition of the membranes would be expected to affect Pmint values. Alternatively, the higher values in the present study may

simply reflect the shorter initial time step in our experiments (3 min compared to 10 min (11) or 30 min (10)). Note that the Pmint values for Cd(DDC)20 and Cd(XANT)20, as determined with C. reinhardtii, are virtually identical (Table S4), despite the differences in their octanol-water distribution coefficients (Table 2). For lipophilic organic molecules, Pmint is normally correlated positively with the Kow value of the permeant (19), reflecting the fact that Pmint incorporates Klw (eq 3), but this does not appear to be the case here, suggesting that differences in the Dm term must offset the greater lipophilicity of the Cd(DDC)20 complex. In an aqueous medium, the diffusion coefficients of the two complexes would not be expected to differ greatly, but within the membrane, a nonaqueous medium, they may be affected by interactions between the M(L)02 complexes and the constituent membrane phospholipids. In their study of the uptake of the M(DDC)20 complexes of Cd, Cu, and Pb by the marine diatom Thalassiosira weissflogii, Phinney and Bruland (10) reported a similar discrepancy between (calculated) Kow values and measured permeabilities. Uptake of lipophilic metal complexes by unicellular microorganisms has also been quantified in terms of a biological removal rate (10, 11), measured during the initial uptake phase and described as d[M(L)02] ) -kbio·[cells]·[M(L)02] dt

(4)

where kbio is the second-order uptake rate constant (L · cell-1 · h-1) for the population of microorganisms, and [cells] and [M(L)20] are the concentration of cells and metal complex, respectively, in the exposure medium. Expressed in this manner, kbio reflects both the permeability of the algal cells and their surface area. In our experiments we purposefully used low cell densities so as to minimize the change in [M(L)20] during the course of the experiment. Since we could not determine kbio on the basis of a measured ∆ [M(L)20] in solution, we instead calculated the removal of M(L)20 from solution indirectly, based on the measured accumulation in the algal cells and the total number of cells (kbio ) Jobs (Table S3) × algal surface area (cm2 · cell-1) ÷ [M(L)20]). Values of kbio calculated in this manner for experiments run at pH 7.0 proved more variable than Pm, ranging from 2.9 × 10-9 L · cell-1 · h-1 (P. subcapitata) to 9.9 × 10-9 L · cell-1 · h-1 (C. fusca) and reflecting the differences in algal specific surface areas (68, 94, and 107 µm2 at pH 7.0 for P. subcapitata, C. reinhardtii, and C. fusca, respectively). Literature values of kbio for marine phytoplankton (Table S4) cover a wider range (0.01 to 570 × 10-9 L · cell-1 · h-1), again largely driven by differences in surface area (11). Loss. When algal cells were exposed to Cd(DDC)20 for 30 min and then resuspended in fresh (metal-free) medium at pH 7.0, Cd loss rates (ke) varied from 0.15 min-1 (P. subcapitata) to 0.33 min-1 (C. reinhardtii). We could find no other measurements of loss rates for lipophilic metal complexes in the literature. To facilitate comparison with the uptake measurements, we used eq 3 to calculate Pmefflux values for the efflux of the labile fraction of the metal complex (Table S5). For Cd(DDC)20 the permeability coefficients for efflux were similar for all three algae (1 to 1.3 × 10-4 cm · s-1). For C. reinhardtii, the loss rate of the more hydrophobic complex, Cd(DDC)20, was lower than that for Cd(XANT)20 (Table 1, ke: 0.33 vs 0.69 min-1; Table S5, Pmefflux 1.3 × 10-4 cm · s-1 vs 4.7 × 10-4 cm · s-1), a trend that is consistent with the behavior of hydrophobic organic molecules (24). However, it should be noted that for both lipophilic complexes, the permeability coefficients for efflux (1 to 4.7 × 10-4 cm · s-1) were up to an order of magnitude lower than the Pmint values measured in the uptake experiments (14 to 26 × 10-4 cm · s-1). This discrepancy raises the possibility that the Cd lost from

the algal cells was not the original CdL20 complex. Lee et al. (25) did observe short-term loss of Cd from the labile intracellular pool of Cd in a marine diatom (Thalassiosira weissflogii) and suggested that the Cd was lost as a phytochelatin-Cd complex. However, their first-order rate constants for Cd loss (ke ) 0.0018 to 0.0026 min-1) were more than 100× lower than those reported here. In addition to ke and Pmefflux values, the elimination experiments also yield information about the intracellular fate of the Cd(L)20 complexes. Based on the values for fL (Table 1), it is clear that a substantial portion of the Cd transported into the cells dissociates from the DDC or XANT ligands and binds to other intracellular ligands (cf. (10)). For cells exposed to Cd(DDC)20 at pH 7.0 for 30 min, the nonlabile fraction, Cdnl, varied from 0.48 (P. subcapitata) to 0.70 (C. fusca). These differences among algal species presumably reflect variations in the concentration and nature of the competing ligands. pH Effect. Based on the results obtained with two neutral complexes, Cd(DDC)20 and Cd(XANT)20, and three different algal species, acidification from pH 7.0 to 5.5 consistently leads to lower overall uptake of the lipophilic metal complexes. The decrease in pH affects both metal uptake and loss: the uptake constant, ki, and the related permeability constant, Pm, are lower at pH 5.5 than at pH 7.0, as is the efflux constant, ke. This pH effect confirms our earlier observations (13), but otherwise is without precedent for contaminants entering the cell by a passive diffusion process since all prior studies were carried out with marine algae at constant pH. The origins of the pH effect are not obvious. In eq 3, used to describe the passive diffusion of lipophilic molecules across biological membranes, parameters A or Klw could in principle be pH-dependent. However, for algal cultures grown at pH 7.0 or 5.5, the cell surface areas, A, did not vary by more than 10% (94.1 ( 1.6 µm2 · cell-1 at pH 7; 86.7 ( 2.6 µm2 · cell-1 at pH 5.5). Similarly, pH had only a weak influence on the measured Dow values for the two complexes (Table 2), suggesting that Klw should also be relatively insensitive to the pH change from 7.0 to 5.5. This reasoning leads us to consider pH-induced changes in membrane properties, e.g., membrane fluidity. At the lower pH a greater proportion of the phospholipid head groups will be protonated, and we speculate that the reduced charge may allow tighter packing of the phospholipids, leading to lower membrane fluidity and a lower value for Dm, the diffusion coefficient of the complex within the membrane. Pyrene-labeled lipids might be used to probe for potential effects of pH on membrane fluidity (26). Note, too, that some algal species seem to adapt to pH changes by modifying the properties and even the composition of their cell membranes (e.g., (27)). Implications for the Biotic Ligand Model (BLM). In our control media, i.e., in the absence of DDC or XANT, algal uptake of Cd was strongly pH- and species-dependent (Table 1, right column). Effects of acidification on metal cation uptake are normally explained in terms of pH-induced changes in metal speciation in solution, together with increased competition between Mz+ and the H+-ion for binding to membrane-bound metal transporters (1, 5). In the control media Cd was present almost entirely as the free Cd2+ ion (∼99%), regardless of the pH, and thus within the construct of the BLM the suppression of Cd uptake at the acidic pH would normally be explained solely in terms of increased competition between Cd2+ and H+ at the algal surface. However, if this competition were the only factor involved, one might expect the effect to be comparable among species, but this is not the case: ratios of kipH 7.0/kipH 5.5 for uptake of Cd2+ (0.38 nM) are 263, 24, and 2 for P. subcapitata, C. fusca, and C. reinhardtii, respectively. The same order of pH-sensitivity is noted for uptake of Cd(DDC)20, for which the kipH 7.0/kipH 5.5 ratios also decline in the sequence P. VOL. 43, NO. 9, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3313

subcapitata (60) > C. fusca (56) > C. reinhardtii (6.5). On the basis of these results, and considering also the recent demonstration of noncompetitive inhibition of metal uptake by the H+-ion (28), we conclude that the effects of acidification on the uptake (and toxicity) of metal cations involve more than just simple H+/Mz+ competition for membrane binding sites. Our results with the lipophilic metal complexes indicate that the interaction of protons with the algal surface affects membrane properties, and that these changes influence not only the permeability of the membrane lipophilic substrates, but also the functioning of the metal cation transporters embedded in the membrane.

Acknowledgments This work was supported by grants from the Canadian Network of Toxicology Centres (Environment Canada) and from the Natural Sciences and Engineering Research Council of Canada (NSERC). Helpful discussions with Laura Sigg (EAWAG) and Se´bastien Sauve´ (Universite´ de Montre´al) are gratefully acknowledged, as is the technical assistance of Lise Rancourt (INRS). P.G.C.C. is supported by the Canada Research Chair Program.

Supporting Information Available Algal growth medium and solutions used in the manipulation of the algal cells; thermodynamic stability constants for Cd(DDC)n and Cd(XANT)n complexes; calculated diffusive fluxes for Cd(L)20 complexes in the algal boundary layer and across the cell membrane; permeability coefficients and biological removal rates for Cd(L)20 complexes; comparison of the uptake and depuration kinetics for Cd(DDC)20 with all three test algae; experimental details for the determination of the subcellular distribution of Cd in algal cells (Chlamydomonas reinhardtii) pre-exposed to Cd(DDC)20 and for the determination of the octanol-water distribution coefficients for the lipophilic Cd(L)20 complexes. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Campbell, P. G. C. Interactions between trace metals and organisms: a critique of the free-ion activity model. In Metal Speciation and Bioavailability in Aquatic Systems, Tessier, A., Turner, D., Eds.; J. Wiley & Sons: Chichester, UK, 1995; Chapter 2, pp 45-102. (2) Gorsuch, J. W.; Janssen, C. R.; Lee, C. M.; Reiley, M. C. Special Issue: The Biotic Ligand Model for metals - current research, future directions, regulatory implications. Comp. Biochem. Physiol. 2002, C133, 1–343. (3) Campbell, P. G. C.; Erre´calde, O.; Fortin, C.; Hiriart-Baer, V. P.; Vigneault, B. Metal bioavailability to phytoplankton - applicability of the Biotic Ligand Model. Comp. Biochem. Physiol. 2002, C133, 185–202. (4) Niyogi, S.; Wood, C. M. Biotic Ligand Model, a flexible tool for developing site-specific water quality guidelines for metals. Environ. Sci. Technol. 2004, 38, 6177–6192. (5) Slaveykova, V. I.; Wilkinson, K. J. Predicting the bioavailability of metals and metal complexes: Critical review of the Biotic Ligand Model. Environ. Chem. 2005, 2, 9–24. (6) Poldoski, J. E. Cadmium bio-accumulation assays - their relationship to various ionic equilibria in Lake Superior water. Environ. Sci. Technol. 1979, 13, 701–706. (7) Florence, T. M.; Lumsden, B. G.; Fardy, J. J. Evaluation of physicochemical techniques for the determination of the fraction of dissolved copper toxic to the marine diatom Nitzchia closterium. Anal. Chim. Acta 1983, 151, 281–295.

3314

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 9, 2009

(8) Ahsanullah, M.; Florence, T. M. Toxicity of copper to the marine amphipod Allorchestes compressa in the presence of water- and lipid-soluble ligands. Mar. Biol. 1984, 84, 41–45. (9) Florence, T. M.; Stauber, J. L. Toxicity of copper complexes to the marine diatom Nitzschia closterium. Aquat. Toxicol. 1986, 8, 11–26. (10) Phinney, J. T.; Bruland, K. W. Uptake of lipophilic organic Cu, Cd and Pb complexes in the coastal diatom Thalassiosira weissflogii. Environ. Sci. Technol. 1994, 28, 1781–1790. (11) Croot, P. L.; Karlson, B.; Elteren, J. T.; Kroon, J. J. Uptake of 64 Cu-oxine by marine phytoplankton. Environ. Sci. Technol. 1999, 33, 3615–3621. (12) Turner, A.; Williamson, I. Octanol-water partitioning of chemical constituents in river water and treated sewage effluent. Water Res. 2005, 39, 4325–4334. (13) Boullemant, A.; Vigneault, B.; Fortin, C.; Campbell, P. G. C. Uptake of neutral metal complexes by a green alga: Influence of pH and humic substances. Aust. J. Chem. 2004, 57, 931–936. (14) Macfie, S. M.; Tarmohamed, Y.; Welbourn, P. M. Effects of cadmium, cobalt, copper, and nickel on growth of the green alga Chlamydomonas reinhardtii: The influences of the cell wall and pH. Arch. Environ. Contam. Toxicol. 1994, 27, 454–458. (15) Paschke, A.; Neitzel, P. L.; Walther, W.; Schu ¨u ¨rmann, G. Octanol/ water partition coefficient of selected herbicides: Determination using shake-flask method and reversed-phase high-performance liquid chromatography. J. Chem. Eng. Data 2004, 49, 1639– 1642. (16) Schecher, W. D.; McAvoy, D. MINEQL+: A Chemical Equilibrium Modeling System. [4.5]; Environmental Research Software: Hallowell, ME, 2001. (17) Martell, A. E.; Smith, R. M. NIST Critically Selected Stability Constants of Metal Complexes v. 8.0; NIST Standard Reference Data Base 46; National Institute of Standards and Technology: Gaithersburg, MD, 2004. (18) Spacie, A.; Hamelink, J. L. Alternative models for describing the bioconcentration of organics in fish. Environ. Toxicol. Chem. 1982, 1, 309–320. (19) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; John Wiley and Sons Ltd.: New York, 1993; pp 1-550. (20) Kaiser, S. M.; Escher, B. I. The evaluation of liposome-water partitioning of 8-hydroxyquinolines and their copper complexes. Environ. Sci. Technol. 2006, 40, 1784–1791. (21) Skoglund, R. S.; Stange, K.; Swackhamer, D. L. A kinetics model for predicting the accumulation of PCBs in phytoplankton. Environ. Sci. Technol. 1996, 30, 2113–2120. (22) Hassler, C. S.; Slaveykova, V. I.; Wilkinson, K. J. Discrimination between intra- and extracellular metals using chemical extractions. Limnol. Oceanogr. Methods 2004, 2, 237–247. (23) Erre´calde, O.; Campbell, P. G. C. Cadmium and zinc bioavailability to Selenastrum capricornutum (Chlorophyceae): Accidental metal uptake and toxicity in the presence of citrate. J. Phycol. 2000, 36, 473–483. (24) Kooijman, S. A. L. M.; Jager, T.; Kooi, B. W. The relationship between elimination rates and partition coefficients. Chemosphere 2004, 57, 745–753. (25) Lee, J. G.; Ahner, B. A.; Morel, F. M. M. Export of cadmium and phytochelatin by the marine diatom Thalassiosira weissflogii. Environ. Sci. Technol. 1996, 30, 1814–1821. (26) Somerharju, P. Pyrene-labeled lipids as tools in membrane biophysics and cell biology. Chem. Phys. Lipids 2002, 116, 57– 74. (27) Tatsuzawa, H.; Takizawa, E.; Wada, M.; Yamamoto, Y. Fatty acid and lipid composition of the acidophilic green alga Chlamydomonas sp. J. Phycol. 1996, 32, 598–601. (28) Franc¸ois, L.; Fortin, C.; Campbell, P. G. C. pH modulates transport rates of manganese and cadmium in the green alga Chlamydomonas reinhardtii through non-competitive interactions: Implications for an algal BLM. Aquat. Toxicol. 2007, 84, 123–132.

ES802832U