Aquatic Toxicology 75 (2005) 127–135

Contrasting effects of chloride on the toxicity of silver totwo green algae,Pseudokirchneriella subcapitataand

Chlamydomonas reinhardtii

Dae-Young Lee1, Claude Fortin, Peter G.C. Campbell∗

INRS-Eau, Terre et Environnement, Universit´e du Quebec, Institut National de la Recherche Scientifique,490 de la Couronne, Quebec City, Que., Canada G1K 9A9

Received 17 January 2005; received in revised form 20 June 2005; accepted 29 June 2005

Abstract

Short-term silver toxicity was determined for two freshwater algae,Pseudokirchneriella subcapitataandChlamydomonasreinhardtii, in the presence and absence of chloride. Silver speciation in the exposure media was controlled and algal growth wasmeasured over 6 h. ForP. subcapitata, an alga with low Ag uptake fluxes, silver toxicity could be predicted on the basis of thefree Ag+ concentration, in the presence or absence of significant complexation by chloride ions, as predicted by the biotic ligandmodel (BLM). ForC. reinhardtii, an alga with high Ag uptake fluxes, silver toxicity was better predicted by the concentration ofall labile dissolved Ag species than by free silver, a result that is consistent with diffusion through the unstirred layer surroundingthe cell surface being the rate-limiting step in silver uptake. For both species, growth inhibition could be predicted on the basiso acellulara©

K

1

s

f

C

hetes

,enncethe

rom

0

f the Ag intracellular quota in the presence or absence of chloride, indicating that silver toxicity is a direct result of intrccumulation rather than cell surface interactions.2005 Elsevier B.V. All rights reserved.

eywords:Biotic ligand model; Intracellular metal; Growth inhibition; Phytoplankton; Silver; Speciation

. Introduction

Freshwater animals exhibit a wide range of silverensitivities with LC50 values ranging from ng/L (pM)

∗ Corresponding author. Tel.: +1 418 654 2538;ax: +1 418 654 2600.E-mail address:peter.campbell@ete.inrs.ca (P.G.C. Campbell).

1 Present address: Wastewater Technology Centre, Environmentanada, 867 Lakeshore Road, Burlington, Ont., Canada L7R 4A6.

to mg/L (�M), and the profound influence of tspeciation of silver on its toxicity to aquatic vertebraand invertebrates is well recognized (Wood et al.2002). Silver toxicity to algae, however, has belargely unexplored, despite the ecological importaof phytoplankton and their recognized role inbiogeochemical cycling of metals (Sigg et al., 1995).A literature search yielded only five EC50 values forthe effects of silver on freshwater algae, ranging f24 nM (Pseudokirchneriella subcapitata, formerly

166-445X/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.aquatox.2005.06.011

128 D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135

known as Selenastrum capricornutum; (U.S.EPA,1987)) to 190 nM (Scenedesmus acutiformis; (Stokes,1981)). However, this compilation is of limited valuesince the EC50 values were all reported as nominaltotal silver concentrations, with no consideration ofsilver speciation or of silver loss from the exposuremedia during the toxicity tests.

In contrast to the situation with silver, the literatureis replete with references to the toxicity of other (diva-lent) metals to algae. Indeed, many of the pioneeringexperiments that led to the development of the free-ion model (FIM) (Campbell, 1995; Morel and Hering,1993), and its more recent incarnation the biotic ligandmodel or BLM (Paquin et al., 2002), were performedwith unicellular algae. If the BLM approach appliesto silver–algae interactions, then the uptake and toxic-ity of silver should vary as a function of the free Ag+

ion concentration (or activity) in the exposure medium.For a fixed dissolved silver concentration, complex-ation with chloride should lead to a decrease in theuptake and toxicity of the metal, whereas for a fixedfree Ag+ concentration, an increase in the concentra-tion of AgCln(1−n)+ complexes should not affect silveruptake or toxicity. In our earlier work on the uptake ofsilver by freshwater algae in the presence of chlorideas a complexing ligand, we demonstrated that this wasnot always the case. When Ag uptake was investigatedfor algae that internalized silver slowly (e.g.,P. sub-capitataandChlorella pyrenoidosa), BLM predictionsfor media containing chloride proved reasonably accu-r ree[ -t id ceb ies.U ciescd egu-l thertt ccuro h aC takeo ep

lgalu ity,i M

as a descriptive model for silver toxicity to algae inthe presence or absence of chloride. We assumed thatsilver toxicity would be related to intracellular silveraccumulation and therefore, that it would be defined bythe factors that affect silver bioavailability, as observedin short-term uptake studies. Specifically, on the basisof our previous Ag bioavailability studies (Fortin andCampbell, 2000; Lee et al., 2004), we expected thatqualitatively the toxicity of silver (expressed in termsof total dissolved [Ag]) would be lower in the presenceof chloride than in its absence, due to chloride com-plexation and the consequent decrease in the free Ag+

concentration. For algae with slow Ag uptake rates, weanticipated that Ag toxicity would be determined bythe free Ag+ concentration, as predicted by the BLMbut that for algae with fast Ag uptake rates, the toxicresponse would be greater than that predicted on thebasis of [Ag+], due to the contribution of labile Agspecies to the diffusive flux of silver through the dif-fusion layer surrounding the cell surface. Finally, ifthe toxicity of silver involves its binding to intracellu-lar ligands, we might expect silver toxicity to vary as afunction of the intracellular Ag cell quota for both ‘fast’and ‘slow’ algae, regardless of the presence or absenceof chloride in the growth media. To test these variouspredictions, we carried out short-term growth assayswith two freshwater algae,C. reinhardtii (a ‘fast’ Aguptake species) andP. subcapitata(a ‘slow’ Ag uptakespecies).

2

2

eena eC da)atcm up-p toa fil-ts resw en-m tant

ate (i.e., Ag bioavailability was determined by the fAg+]) (Lee et al., 2004). However, if silver internalizaion is very rapid (e.g.,Chlamydomonas reinhardti),iffusion from the bulk solution to the algal surfaecomes rate-limiting, and the BLM no longer applnder these latter circumstances, all labile Ag speontribute to silver uptake and since AgCln complexesissociate readily, the overall uptake process is r

ated by the total dissolved silver concentration rahan by [Ag+] (Fortin and Campbell, 2000). For allhree algal species, silver uptake appeared to only via facilitated cation transport (perhaps througu(I) transporter); there was no evidence for the upf neutral (AgCl0) or anionic (AgCl2−) species in thresence of chloride.

In the present study, we set out to determine if aptake of silver could be used to predict its toxic

.e., we sought to determine the validity of the BL

. Materials and methods

.1. Test organisms and culture preparation

The experiments were performed with two grlgae,C. reinhardtii (University of Toronto Culturollection, UTCC 11; UTCC, Toronto, Ont., CanandP. subcapitata(formerly known asS. capricornu-um, UTCC 37) grown in a high salt medium (Table 1,olumn 2) modified fromMacfie et al. (1994). Theedium was sterilized by autoclaving and then slemented with a filter-sterilized trace metal mixvoid metal precipitation; the trace metal mix wasered through a polycarbonate membrane (0.2�m poreize, Poretics, Minnetonka, MN, USA). Batch cultuere grown axenically in 250 mL polycarbonate Erleyer flasks (100 mL culture volume) under cons

D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135 129

Table 1Composition of the algal culture medium (modified fromMacfie et al. (1994)), the silver-exposure medium, and the simplified culture-rinsingmediuma

Components Culture medium (mol/L) Ag-exposure medium (mol/L) Simplified culture-rinsingmedium (with Ag) (mol/L)

Ag – Variable (Table 2) 1.00× 10−7

Cl 5.98× 10−6 Variable (Table 2) 5.00× 10−6

K 4.22× 10−3 4.00× 10−3 4.00× 10−3

PO4 1.37× 10−4 – –NO3 5.07× 10−3 5.07→ 1.87× 10−3b 5.07× 10−3

Na 1.02× 10−4 5.00× 10−6 5.00× 10−6

a For ingredients common to all media, and the minor components for the culture media, please refer toFortin and Campbell (2000).b The ionic strength in each medium was maintained at a constant level (6 meq. L−1) by replacing chloride with nitrate as the chloride

concentration was decreased.

temperature at 20◦C and an illumination of about100�E m−2 s−1 (cool white fluorescent lamps) withrotary agitation at 50 rpm. To identify the exponentialgrowth phase, the batch cultures were monitored dailyby removing a small aliquot (1 mL) and determiningthe cell densities using an electronic particle counter(model MultisizerTM 3, Coulter Electronics, Toronto,Ont., Canada). Cultures were checked regularly foraxenicity by plating onto nutrient agar (Difco-Bactoagar, Liverpool, New South Wales, Australia).

2.2. Reagents and glassware

All exposure solutions were initially prepared fromultrapure water (18 M� cm; Milli-Q3RO/Milli-Q2 sys-tem, Bedford, MA, USA), filtered to eliminate any fineparticulates to which silver could adsorb (0.2�m poly-carbonate filters), and then adjusted to pH 7.0± 0.1 (nopH buffers).

To minimize metal contamination and adsorption,reagents were prepared in containers made fromTeflon®, polycarbonate, polypropylene, or polyethy-lene. Containers were soaked for at least 24 h in 10%HNO3, thoroughly rinsed seven times with ultrapurewater, and dried prior to use in a class 100 laminarflow hood under a positive pressure of filtered air.Reagents used in this study were of analytical grade orbetter. Silver radioactivity (110mAg; 136 mCi mmol−1;Risø National Laboratory, Rosklide, Denmark) wasm 14W i-t ,C in-d ck

solutions (0.2% HNO3) of non-radioactive and radioac-tive silver were kept at pH <2 and 4◦C in the dark.

2.3. Experimental procedures

The experiments designed to determine Ag toxicityto C. reinhardtii andP. subcapitatawere performedover a short time period (6 h) under two experimentalregimes (see details below in Section2.3.1). Algalcultures in mid-exponential growth phase werecollected by gentle filtration (vacuum pressure of<10 cm Hg) onto a 2.0�m pore-size polycarbonatemembrane filter (Poretics, Minnetonka, MN, USA).The algal cells were then resuspended in 10 mL of theculture medium and their size distribution, averagesurface area and concentration were determined usingthe electronic particle counter. Typical surface areaswere 85–95�m2 for C. reinhardtii and 65–75�m2

for P. subcapitatain the exponential growth phase.The resuspended cells were inoculated into a seriesof batch cultures containing various amounts ofradioactive and non-radioactive Ag (see Section2.3.1and Table 2) to give an initial culture density of∼50,000 cells mL−1 for C. reinhardtii and∼120,000cells mL−1 for P. subcapitata. The silver exposuretests were carried out under conditions similar to thoseused for growing the stock culture (20◦C and a lightregime of 100�E m−2 s−1). A short growth period(6 h) was selected to limit silver depletion in the testm thee icale dM icd es/

easured on a liquid scintillation counter (14inSpectralTM, Wallac, Turku, Finland) after add

ion of 5 mL of Eco-Lume scintillation cocktail (ICNosta Mesa, CA, USA) using the full spectral wow (1–1024 keV) with blank subtraction. Acidic sto

edia over the test period. Silver speciation inxposure solutions was computed with a chemquilibrium program, MINEQL+ (Schecher ancAvoy, 2001), using an updated thermodynamatabase (http://www.inrs-ete.uquebec.ca/activit

130 D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135

Table 2Concentrations of silver and chloride in silver-exposure media

Treatments Species Elements Control Batch 1 Batch 2 Batch 3 Batch 4

Test 1 (absence of Cl) C. reinhardtii Nominal [Ag]tot (nM) 0 10 20 30 50[Cl] (�M) 5 5 5 5 5Nominal [Ag+] (nM) 0 10 20 30 50Final [Ag]tot (nM)a 0 1.3 3.3 11.1 24.5

P. subcapitata Nominal [Ag]tot (nM) 0 10 30 50 80[Cl] (�M) 5 5 5 5 5Nominal [Ag+] (nM) 0 10 30 50 80Final [Ag]tot (nM)a 0 0.2 3.3 15.5 36.0

Test 2 (presence of Cl) C. reinhardtii Nominal [Ag]tot (nM) 0 25 41 60 80[Cl] (�M) 5 804 1600 2400 3200Nominal [Ag+] (nM) 0 10 10 10 10Final [Ag]tot (nM)a 0 6.7 15.6 32.9 45.4

P. subcapitata Nominal [Ag]tot (nM) 0 25 41 60 80[Cl] (�M) 5 804 1600 2400 3200Nominal [Ag+] (nM) 0 10 10 10 10Final [Ag]tot (nM)a 0 3.1 10.8 17.5 38.0

a Aqueous silver concentration at the end of 6 h exposure period.

groupes/biogeo/personal.htm). The thermodynamicdata were derived from the default database suppliedwith the MINEQL+ software (v. 4.5), but wereupdated by comparison with the NIST Critical Stabil-ity Constants of Metal Complexes Standard ReferenceDatabase (Martell and Smith, 2001).

2.3.1. Tests for silver toxicityThe experiments were designed to determine the

toxic response ofC. reinhardtii andP. subcapitatainbatch cultures during a 6 h Ag exposure period. Thegrowth rates of the cultures were determined as a func-tion of the added silver concentration in the presenceor absence of AgCln(1−n)+ chloro complexes.

Under the first experimental regime, growth assayswere conducted over a wide range of the nominal totalsilver concentrations [Ag]T, and at a low fixed chlo-ride concentration ([Cl] = 5�M). Since this low levelof chloride does not affect silver speciation (i.e., sil-ver exists almost entirely as free Ag+), these tests wereregarded as being conducted in the “absence of chlo-ride” (Test 1,Table 2). In the second regime, silvertoxicity was monitored over a wide range of the [Ag]Tvalues, where [Ag+] was fixed at 10 nM throughoutthe entire series by using appropriate combinations of[Ag]T and [Cl]. These tests were designated as the“presence of chloride” condition (Test 2,Table 2). Theionic strength in each medium was maintained at a con-

stant level of 6 meq. L−1 by replacing chloride withnitrate as the chloride concentration was decreased.The toxicity experiments were run in duplicate withtwo parallel incubations for any given set of experi-mental conditions.

At the end of Ag exposure, sub-samples (1 mL) weretaken in duplicate, i.e., from each replicate run, and thecell density was determined with the electronic particlecounter as described above. Growth rates were calcu-lated according to the following equation:

µ = (ln Nt − ln Nt0)

(t − t0)

whereµ is the growth rate,t− t0 is the Ag exposureperiod (6 h in the present study), andNt andNt0 arethe cell densities at timet and at the beginning of test,respectively. The relative growth rate was determinedas the ratio ofµAg (growth rate of the culturesexposed to Ag) toµ0 (growth rate of the controlculture). The growth data were plotted against averageAg exposure concentrations and fitted using SigmaPlot 2002 software and its built-in non-linear “fourparameter logistic curve” regression tool. The Excelmacro REGTOX© (based on the Marquardt algorithm;(http://eric.vindimian.9online.fr/endownload.html))was used to calculate EC50s. The confidence intervalson the parameters were estimated by a bootstrapnon-parametric simulation. On one occasion (Table 2:

D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135 131

Test 1, Batch 4, forC. reinhardtii), we observeda decrease in cell density due to cell lysis; for thepurpose of fitting the algorithm, growth was thusconsidered to be null.

2.3.2. Determination of silver accumulationTo estimate intracellular Ag accumulation by the

two algal species and its relationship with toxicity,we collected sub-samples (50 mL) at the end of theAg exposures from each replicate flask. Cells werecollected on two superimposed 2.0-�m polycarbonatemembrane filters and Ag accumulation was assessedby measuring the110mAg radioactivity of the filters. Todetermine the true intracellular silver fraction,110mAgadsorbed to the algal cell surfaces was removed byrinsing the filters four times with 10 mL of simplifiedculture medium containing 100 nM non-radioactive Ag(Table 1, column 4) (Fortin and Campbell, 2000). Theradioactivity of the lower filter was subtracted from themeasurement of the upper filter to correct for passiveadsorption of110mAg radioactivity onto the filters. Thefiltrates were also sampled for110mAg radioactivity todetermine the Ag concentration remaining in the aque-ous phase.

3. Results and discussion

3.1. Exposure conditions

lgalcp int sil-v spitet sese en-t lvere nei aveu s thea al[ usp s thei cu-l t,t bye om-

Fig. 1. Silver toxicity to algae in Test 1 (expressed as relative growthrates) as a function of the mean dissolved silver concentration inthe chloride-free exposure media. Symbols represent silver toxicityto C. reinhardtii (�) andP. subcapitata(�) in the absence of Cl.Experiments were run in duplicate and both data points are shownfor each experimental condition. The curves represent non-linearregressions through the data points for each alga (r2 = 0.93 forP.subcapitataand 0.99 forC. reinhardtii).

inal silver concentration but did not check for silverlosses.

3.2. Silver toxicity under two exposure regimes

3.2.1. Test 1: silver toxicity in the absence ofchloride

Ag toxicity expressed as a reduction in relativegrowth rate increased progressively with the mean [Ag]for both species (Fig. 1). The relative growth ratedecreased gradually from 100 (control, no silver) to13% of the control value at the highest mean [Ag] of58 nM forP. subcapitataand to−11% of the controllevel (i.e., a population decrease) at the highest mean[Ag] of 37 nM for C. reinhardtii. From the growthcurves, the EC50(Ag) value (i.e., the mean Ag concen-tration required to inhibit growth by 50%) was calcu-lated to be 26 nM (95% CI 21–31 nm) forP. subcapitataand 12 nM (95% CI 11–13 nM) forC. reinhardtii. Asindicated by its lower EC50 value,C. reinhardtiiwasmore susceptible to Ag toxicity thanP. subcapitata.These EC50(Ag) values are among the lowest reportedin the literature for freshwater algae (Table 3). Note thatin the present experiments in the absence of chloride,

At the end of the 6 h incubation periods, the aells were collected on polycarbonate filters (2�morosity). Measurements of total “dissolved” Ag

he filtrates showed that very significant losses ofer had occurred during the growth assays, dehe choice of a relatively short growth period; losxceeded 90% in the media with low silver concrations and declined to about 50% in the high sixposures (Table 2). To take into account this decli

n ambient silver during the exposure period, we hsed the mean silver concentration, calculated arithmetic mean of the nominal [Ag] and the fin

Ag] (i.e., the concentration of Ag in the aqueohase at the end of the 6 h exposure period), a

nput value for the equilibrium model used to calate the free Ag+ concentration. Although imperfechis approach is an improvement over that usedarlier workers, who systematically reported the n

132 D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135

Table 3Toxicity of silver toChlamydomonas reinhardtiiandPseudokirchneriella subcapitata—comparison with literature values for other freshwateralgae

Algal species Response [Ag] (nM) Toxicity parameter Reference

Chlorella vulgaris Growth (14 d) 470 Algistatic Hutchinson and Stokes (1975)90 Detectable inhibition

Scenedesmus acuminatus Growth (8 d) 70 EC50 Stokes (1981)470 Algistatic Hutchinson (1973)

Scenedesmus acutiformis Growth (8 d) 190 EC50 Stokes (1981)Haematococcus capensis Growth (14 d) 930 EC80 Hutchinson (1973)Microcystis aeruginosa Growth (8 d) 655 Incipient inhibition Bringmann and Kuhn (1978)

Nostoc muscorum Survival (21 d) 26 LC50 Rai and Raizada (1985)Photosynthesis 26 90% decrease

Scenedesmus quadricauda Growth (7 d) 88 Incipient inhibition Bringmann and Kuhn (1980)Selenastrum capricornutum Growth (14–21 d) 59 EC50 Turbak et al. (1986)Selenastrum capricornutum Growth 24 EC50 US EPA 1978 (cited inU.S.EPA, 1987)Pseudokirchneriella subcapitataa Growth (6 h) 26 EC50 This studyChlamydomonas reinhardtii Growth (6 h) 12 EC50 This study

a Alga formerly known asSelenastrum capricornutum.

>98% of the silver was present as the free Ag+ ion, andthus EC50(Ag+) ∼ EC50(Ag).

3.2.2. Test 2: silver toxicity in the presence ofchloride

For both species, silver toxicity was attenuated inthe presence of chloride; for a given dissolved silverconcentration, the inhibition of algal growth was lessin the presence of chloride than in its absence (com-pareFigs. 1 and 2for each alga). For example, for[Ag]d ∼20 nM, the relative growth ofC. reinhardtiiwas ∼20% in the absence of chloride, or∼70% inthe presence of chloride; for similar values of [Ag]d,the relative growth ofP. subcapitatawas∼60% in theabsence of chloride and∼85% in its presence. Thisresult agrees qualitatively with our prediction that thetoxicity of silver (expressed in terms of total dissolved[Ag]) would be lower in the presence of chloride than inits absence due to chloride complexation and the conse-quent decrease in the free Ag+ concentration. However,on closer examination it is clear that the two algae dif-fer in their responses to the presence of AgCln

(1−n)+complexes.

The second test was designed to test the BLM pre-diction that silver toxicity should be independent ofthe presence or absence of chloro complexes in theexposure medium. The composition of each chloride-containing medium was adjusted so that the initial free

Fig. 2. Silver toxicity in Test 2 (expressed as relative growth rates)as a function of the mean dissolved silver concentration in the expo-sure medium. The media initially contained a constant free Ag+

concentration (10 nM) and increasing concentrations of both silverand chloride (seeTable 2). Symbols represent silver toxicity toC.reinhardtii (�) andP. subcapitata(�) in the presence of chloride.Experiments were run in duplicate and both data points are shownfor each experimental condition. The dotted line represents the rel-ative growth rate predicted for a nominal free Ag+ concentration of10 nM, based onFig. 1.

D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135 133

[Ag+] in the external medium att= 0 was constant(10 nM; seeTable 2), a value predicted fromFig. 1to inhibit algal growth by∼20% (see dotted horizontalline in Fig. 2). Note that in the absence of silver, i.e.,in the control cultures, increasing the chloride concen-tration from 5�M to 5 mM did not affect algal growthrates (results not shown). If the BLM applied in thepresent case, algal growth should have remained con-stant at about 80% of the control value. ForP. subcapi-tata, the alga with the slower Ag uptake rate (Lee et al.,2004), the reduction in algal growth did indeed remainreasonably constant at the predicted level regardlessof the total Ag concentrations (Fig. 2, open squares);relative algal growth was virtually unaffected as [Ag]o

dincreased from 25 to 80 nM. The slope of the linearregression between relative growth and the mean dis-solved Ag concentration was not significantly differentfrom zero (P> 0.05). The small deviation of our datafrom the ideal BLM prediction (i.e., the growth rateappeared to decrease slightly with the increase of sil-ver concentration) may reflect greater silver depletionfrom the system at low ambient silver concentration.

Under the identical Ag exposure regime,C. rein-hardtii behaved differently fromP. subcapitata. Rel-ative growth did not remain constant but decreasedprogressively as the concentration of total silver (andAgCln(1−n)+ complexes) increased (Fig. 2, filled cir-cles). The slope of the regression between relative

growth and the mean dissolved Ag concentration wasvery significantly different from zero (P< 0.0001). Thisresponse agrees well with our predictions, which werebased on the earlier observation that silver uptake bythis alga also increased in media with constant [Ag+]but increasing concentrations of AgCln

(1−n)+ com-plexes (Fortin and Campbell, 2000). For algal cellswith an exceptionally high Ag internalization rate suchas C. reinhardtii, Ag bioavailability can be limitedby transport through the unstirred diffusion layer sur-rounding the cell. The overall uptake rate is influencedby the concentration gradient of labile silver complexes(Ag+ + AgCln(1−n)+) between the bulk solution and thealgal surface (Pinheiro et al., 2004). The high Ag inter-nalization rates ofC. reinhardtiiresult in deviations ofAg uptake from the BLM prediction and in the failureof the BLM to predict subsequent cellular Ag toxicity.

The differing results for the two algal species nicelyillustrate the dangers of relying on the free-metal ionconcentration to predict metal toxicity under conditionswhere metal uptake is controlled not by the rate of inter-nalization but rather by the rate of metal supply fromthe exposure medium.Hudson and Morel (1993)wereamong the first to distinguish clearly between the ther-modynamic and kinetic models of metal uptake andbioavailability, but this distinction seems to have beenlargely forgotten in the current collective enthusiasmfor the biotic ligand model.

F tion of e( () of ch e datap

ig. 3. Silver toxicity to algae (relative growth rates) as a func�) of chloride. (B)P. subcapitatain the absence (�) or presence�oints for each alga (r2 = 0.86 for both curves).

the silver cell quota. (A)C. reinhardtii in the absence (©) or presencloride. The curves represent non-linear regressions through th

134 D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135

3.3. Silver toxicity and cell quota

When Ag toxicity was expressed on the basis of Agcell quotas instead of the mean total Ag concentrationsin the exposure media, the response curves (relativegrowth rate versus intracellular silver) in the absenceor presence of chloride overlapped (Fig. 3A and B).This result suggests that once the silver had entered thealgal cells its effects were then similar, regardless ofthe external chloride concentration.

Since the relative growth ofC. reinhardtiideclinedin both Tests 1 and 2, we were able to calculate EC50values expressed as the intracellular Ag concentra-tion required to inhibit growth by 50%: 74 amol cell−1

(95% CI 63–80 amol cell−1) in the absence of Cl(Fig. 3A, open circles); 75 amol cell−1 (95% CI68–85 amol cell−1) in the presence of Cl (Fig. 3A,closed circles). These values are much closer to eachother than are the EC50 values expressed in terms ofmean dissolved Ag: 56 nM (95% CI 50–64 nM) in thepresence of chloride (Fig. 2, closed circles); 12 nM(95% CI 11–13 nM) in the absence of chloride (Fig. 1,closed circles), indicating that intracellular silver is amuch better predictor of toxicity than is the dissolvedsilver concentration. ForP. subcapitata, an EC50 cellquota could only be calculated in the absence of chlo-ride (120 amol cell−1; 95% CI 110–132 amol cell−1),since in Test 2 the growth rate did not decline signif-icantly as the dissolved Ag concentration increased;this EC50 quota is higher than that forC. reinhardtii( hel

4

t-t nedb es-e ium,ar rms Agca s-t thet ellu-l s a

good predictor of growth inhibition in the presence orabsence of chloride, indicating that silver toxicity is adirect result of intracellular accumulation rather thancell surface interactions.

Acknowledgements

This work was supported by grants from KodakCanada, the Natural Sciences and EngineeringResearch Council of Canada, the Quebec Fonds deformation de chercheurs et l’aidea la recherche andthe Canadian Network of Toxicology Centers. P.G.C.Campbell is supported by the Canada Research ChairProgram.

References

Bringmann, G., Kuhn, R., 1978. Limiting values for the noxiouseffects of water pollutant material to blue algae (Micro-cystis aeruginosa) and green algae (Scenedesmus quadri-cauda) in cell multiplication inhibition tests. Vom Wasser 50,45–60.

Bringmann, G., Kuhn, R., 1980. Comparison of the toxicity thresh-olds of water pollutants to bacteria algae and protozoa in the cellmultiplication inhibition. Water Res. 14, 231–241.

Campbell, P.G.C., 1995. Interactions between trace metals and organ-isms: a critique of the free-ion activity model. In: Tessier, A.,Turner, D. (Eds.), Metal Speciation and Bioavailability in AquaticSystems. IUPAC Series on Analytical and Physical Chemistry of

pp.

F alga,n:78.

H t bytion

H eavyater

H algal573.

L ride-,

M s ofalgad

74 amol cell−1), a result that is consistent with tower sensitivity ofP. subcapitatato silver.

. Conclusions

For P. subcapitata, an alga exhibiting slow shorerm Ag uptake rates, silver toxicity was determiy the free Ag+ concentration, regardless of the prnce or absence of chloride in the exposure meds predicted by the BLM. In contrast, growth ofC.einhardtii, an alga characterized by fast short-teilver uptake rates, varied as a function of labileomplexes in solution (Ag+ + AgCln(1−n)+), reflectingdiffusional limitation on Ag bioavailability. This di

inction between the two algae disappeared whenoxic response was expressed in terms of intracar Ag accumulation, i.e., the silver cell quota wa

Environmental Systems. J. Wiley & Sons, Chichester, UK,45–102.

ortin, C., Campbell, P.G.C., 2000. Silver uptake by the greenChlamydomonas reinhardtii, in relation to chemical speciatioinfluence of chloride. Environ. Toxicol. Chem. 19, 2769–27

udson, R.J.M., Morel, F.M.M., 1993. Trace metal transpormarine microorganisms: implications of metal coordinakinetics. Deep-Sea Res. 40, 129–150.

utchinson, T.C., 1973. Comparative studies of the toxicity of hmetals to phytoplankton and their synergistic interactions. WPollut. Res. J. Can. 8, 68–90.

utchinson, T.C., Stokes, P.M., 1975. Heavy metal toxicity andbioassays. Water quality parameters. No. ASTM STP No.Am. Soc. Test. Mater. 320–343.

ee, D.Y., Fortin, C., Campbell, P.G.C., 2004. Influence of chloon silver uptake by two green algae,Pseudokirchneriella subcapitataandChlorella pyrenoidosa. Environ. Toxicol. Chem. 231012–1018.

acfie, S.M., Tarmohamed, Y., Welbourn, P.M., 1994. Effectcadmium, cobalt, copper, and nickel on growth of the greenChlamydomonas reinhardtii: the influences of the cell wall anpH. Arch. Environ. Contam. Toxicol. 27, 454–458.

D.-Y. Lee et al. / Aquatic Toxicology 75 (2005) 127–135 135

Martell, A.E., Smith, R.M., 2001. NIST Critically Selected Stabil-ity Constants of Metal Complexes v. 6.0. National Institute ofStandards and Technology, Gaithersburg, MD, USA.

Morel, F.M.M., Hering, J.G., 1993. Principles and Applications ofAquatic Chemistry. J. Wiley & Sons Ltd., New York, NY.

Paquin, P.R., Gorsuch, J.W., Apte, S., Batley, G.E., Bowles, K.C.,Campbell, P.G.C., Delos, C.G., Di Toro, D.M., Dwyer, R.L.,Galvez, F., Gensemer, R.W., Goss, G.G., Hogstrand, C., Janssen,C.R., McGeer, J.C., Naddy, R.B., Playle, R.C., Santore, R.C.,Schneider, U., Stubblefield, W.A., Wood, C.M., Wu, K.B., 2002.The biotic ligand model: a historical overview. Comp. Biochem.Physiol. 133, 3–35.

Pinheiro, J.P., Galceran, J., Van Leeuwen, H.P., 2004. Metal speci-ation dynamics and bioavailability: bulk depletion effects. Envi-ron. Sci. Technol. 38, 2397–2405.

Rai, L.C., Raizada, M., 1985. Effect of nickel and silver ions on sur-vival, growth carbon fixation and nitrogenase activity inNostocmuscorum: regulation of toxicity by EDTA and calcium. J. Gen.Appl. Microbiol. 31, 329–337.

Schecher, W.D., McAvoy, D., 2001. MINEQL+: a Chemical Equi-librium Modeling System. (4.5). Hallowell, ME, USA, Environ-mental Research Software.

Sigg, L., Kuhn, A., Xue, H.B., Kiefer, E., Kistler, D., 1995. Cyclesof trace-elements (copper and zinc) in a eutrophic lake—role ofspeciation and sedimentation. Aquatic Chemistry. Advances inChemistry Series, No. 244. American Chemical Society, Wash-ington, DC, pp. 177–194.

Stokes, P.M., 1981. Multiple metal tolerance in copper tolerant greenalgae. J. Plant Nutr. 3, 667–678.

Turbak, S.C., Olson, S.B., Mcfeters, G.A., 1986. Comparison of algalassay systems for detecting waterborne herbicides and metals.Water Res. 20, 91–96.

U.S.EPA 1987. Ambient Aquatic Life Water Quality Criteria for Sil-ver. U.S. Environmental Protection Agency, Office of Researchand Development, Environmental Research Laboratories No.440/5-87-011.

Wood, C.M., La Point, T.W., Armstrong, D.E., Birge, W.J.,Brauner, C.J., Brix, K.V., Call, D.J., Crecelius, E.A., Davies,P.H., Gorsuch, J.W., Hogstrand, C., Mahony, J.D., McGeer,J.C., O’Connor, T.P., 2002. Biological effects of silver. In:Andren, A.W., Bober, T.W. (Eds.), Silver in the Environment:Transport, Fate and Effects. Society of Environmental Tox-icology and Chemistry, SETAC Press, Pensacola, FL, USA,pp. 27–63.