A

uoFiMaKuadt©

K

1

actiH{eLcHtt

0d

Aquatic Toxicology 84 (2007) 123–132

pH modulates transport rates of manganese and cadmium in the greenalga Chlamydomonas reinhardtii through non-competitive interactions:

Implications for an algal BLM

Laura Francois, Claude Fortin ∗, Peter G.C. CampbellINRS-Eau, Terre et Environnement, Universite du Quebec, 490 de la Couronne, Quebec, Que., Canada G1K 9A9

Received 15 November 2006; received in revised form 19 February 2007; accepted 27 February 2007

bstract

The influence of pH on short-term uptake of manganese and cadmium by the green alga Chlamydomonas reinhardtii was studied to betternderstand the nature of proton interactions with metal membrane transporters. Manganese and cadmium internalization fluxes (Jint) were measuredver a wide range of free metal ion concentrations from 1× 10−10 to 4× 10−4 M at several pH values (Mn: 5.0, 6.5 and 8.0; Cd: 5.0 and 6.5).or both metals, first-order biological internalization kinetics were observed but the maximum transport flux (Jmax) decreased when pH decreased,

n contradiction with the Biotic Ligand Model (BLM). This result suggested a non-competitive inhibition of metal uptake by the H+-ion. Aichaelis–Menten type inhibition model considering proton and calcium competition was tested. The metal biotic ligand stability constants

nd the stability constants for competitive binding of Ca2+ and H+ with the metal transporters were calculated: for manganese, KMn = 104.20 andCa = 103.71; for cadmium, KCd = 104.19 and KCa = 104.76; for both metal transport systems, KH was not a significant parameter. Furthermore, metal

ptake was not significantly influenced by the pH of the antecedent growth medium, suggesting that increases in metal fluxes as the pH is raisedre caused by conformational changes of the surface transport proteins rather than by the synthesis of additional transport sites. Our resultsemonstrate that the BLM in its present state does not properly describe the true influence of pH on manganese and cadmium uptake by algae andhat a non-competitive inhibition component must be integrated.

2007 Elsevier B.V. All rights reserved.

ompe

tuatim

tu“ta

eywords: Manganese; Cadmium; pH; Biotic Ligand Model; Phytoplankton; C

. Introduction

According to the vast majority of studies on metal uptake byquatic biota, the free ion activity, {Mz+}, reflects the metal’shemical activity, which in turn determines the importance ofhe metal’s reactions with the biologically active surface bind-ng sites (Free Ion Activity Model, FIAM; Campbell, 1995).owever, when the pH is not constant, the simple knowledge ofMz+} is insufficient to predict the biological response to metalxposure (Campbell, 1995; Campbell et al., 2002). The Bioticigand Model, or BLM (Di Toro et al., 2001), was designed toorrect for this shortcoming of the FIAM. The BLM considers

+/Mz+ competition, explicitly introducing the reactions of pro-

ons with the biotic ligand. Protons are expected to compete withhe metal for binding sites X−, such that when {H+} increases,

∗ Corresponding author. Tel.: +1 418 654 3770; fax: +1 418 654 2600.E-mail address: claude.fortin@ete.inrs.ca (C. Fortin).

1datbun

166-445X/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.aquatox.2007.02.019

tition; Calcium

he number of available X− sites decreases, thus reducing metalptake and toxicity (Campbell and Stokes, 1985; Campbell etl., 2002; Wilkinson and Buffle, 2004). In the development ofhe BLM to date it has been assumed, rather uncritically, that thenfluence of protons is limited to this “competitive inhibition”

echanism.In addition to H+/Mz+ competition at the biological surface,

here are other ways in which pH variations can influence metalptake. For example, protons could inhibit metal uptake in anon-competitive” manner, by causing changes in the conforma-ion of the proteins involved in trans-membrane metal transportnd modifying the metal internalization rate (Harrison et al.,986). According to Parent and Campbell (1994), and as pre-icted by the FIAM, aluminium bioavailability to the unicellularlga Chlorella pyrenoidosa at a given pH was a function of

he free aluminium ion concentration [Al3+] (for pH valuesetween 4.3 and 6.0). Furthermore, for a fixed value of [Al3+], Alptake and toxicity increased with pH. Nevertheless, the authorsoted that the FIAM did not explain the data correctly: despite

1 Toxic

cgoKthsstpceet(ieewiMrfl

�d

tbtiaatT(hegmcstlp

iebp2nuettp

tncboiaoetco

2

2

pds(BcmaitFsntbSrsTanp7MbB

abb[

2D(18 M� cm) and dried in a laminar flow hood. Material destined

24 L. Francois et al. / Aquatic

onsidering Al3+/H+ competition for binding at the algal surface,rowth inhibition showed an unexplained residual dependencyn pH (enhanced protection as the pH was lowered). Likewise,ola and Wilkinson (2005) observed an important decrease in

he cadmium internalization flux (Jint) in Chlamydomonas rein-ardtii when pH was decreased from 7.0 to 5.2. Because metalpeciation remained unchanged over this pH range, the authorsuggested that this decrease was due to direct proton competi-ion for membrane transport sites. Nonetheless, the BLM did notredict the uptake very well and they suggested that allosterichanges of the transport proteins induced variations in the appar-nt binding constant. Both Heijerick et al. (2002) and Wildet al. (2006) observed a Zn EC50 (expressed as free Zn2+)hat did not increase linearly with {H+} in two green algaePseudokirchneriella subcapitata and Chlorella sp.). A sharpncrease in toxicity was noted at pH > 6, which could not bexplained by simple proton competition. In a recent study, Fortint al. (2007) confirmed that uranyl ion uptake in C. reinhardtiias inhibited by protons. Their results showed that competitive

nhibition of the uptake was important (KMMUO2 pH 5 � KMM

UO2 pH 7;ichaelis–Menten half-saturation constant in �M, which cor-

esponds to the free UO22+ ion concentration when uranyl

ux is equal to half of the maximum uptake flux JMMUO2

in

mol m−2 min−1), but non-competitive inhibition was also evi-ent (JMM

UO2 pH 5 < JMMUO2 pH 7).

In all the above cases, deviations from the BLM have beenentatively explained by invoking physiological changes inducedy pH, such as conformational changes in metal transport pro-eins leading to changes in internalization rates, or increasesn the number of metal transporters with pH. Chowdhurynd Blust (2001) studied 85Sr uptake in the common carpnd focused on the nature of the surface binding of stron-ium and protons, as the pH was varied between 5.0 and 8.5.hey observed that H+ inhibited Sr2+ uptake, modifying JMM

Srmaximum Sr2+ uptake flux in �mol kg−1 h−1), but not thealf-saturation constant KMM

Sr . To describe the overall protonffect, a model of partial non-competitive inhibition was sug-ested and the data obtained followed the Michaelis–Mentenodel. This model implies that the H+-ion interacts with the

ell surface in a different region from where the metal bindingites are located (so that Sr2+ binding is not directly dis-urbed); however metal transport decreases because H+ bindingeads to changes in the functional characteristics of the trans-orter.

Because there is no theoretical or experimental basis allow-ng one to neglect the direct biological effects of pH changes,mpirical models (based on regression analyses) have ofteneen used successfully to adjust the BLM over a range ofH values (Heijerick et al., 2003; De Schamphelaere et al.,003). However, these models lack a mechanistic underpin-ing and there is thus a need to better understand why metalptake changes when pH varies, and to integrate the influ-

nce of pH into a BLM-type model. This reasoning led uso ask the following question: are protons simply competi-ors for metal-specific binding sites (with no influence on thehysiology) or do they also influence trans-membrane metal

fub

ology 84 (2007) 123–132

ransport flux capacity (e.g., the intrinsic rate of metal inter-alization and/or the number of transport sites)? We tried tolarify the influence of protons on manganese and cadmiumioavailability in a green alga by observing the influence of pHn metal internalization fluxes. Short-term measures of max-mum internalization fluxes over a pH range were performednd a mechanistic metal internalization flux model was devel-ped to help discriminate between the two different inhibitoryffects of H+: the competitive effect (binding of protons reducinghe number of available transport sites) versus a physiologi-al non-competitive effect (on transport site affinity, kineticsr number).

. Materials and methods

.1. Test organism and cultures

The unicellular green alga C. reinhardtii was chosen for theresent study for its ability to grow over a wide pH range (5–8,ata not shown). Pure cultures were obtained from the Univer-ity of Toronto Culture Collection of Algae and Cyanobacteriastrain UTCC 11, Department of Ecology and Evolutionaryiology, University of Toronto, Toronto, Ont., Canada). Asyn-hronous batch cultures were grown axenically in 100 mL ofodified high salt medium MHSM-1 (Macfie et al., 1994; Fortin

nd Campbell, 2000), see Table 1. Cultures were maintainedn 250 mL Erlenmeyer glass flasks under constant rotary agi-ation (100 rpm), illumination (130 �E m−2 s−1, Cool Whiteluorescent tubes) and temperature (20.0± 0.2 ◦C). The ionictrength of the medium was 6 meq L−1. For regular mainte-ance, approximately 2 mL of culture were transferred weeklyo a fresh, sterile medium at pH 7 with 10−2 M of HEPESuffer (2-[N-morpholino]ethanesulfonic acid], purchased fromigma–Aldrich, St. Louis, MO, USA); this operation wasepeated up to 10 weeks, at which point a new culture wastarted from the original culture that had been kept on agar.his was a supplementary precaution to ensure maintenance ofxenic conditions. To identify the exponential growth phase, cellumbers in batch cultures were determined using an electronicarticle counter (1:10 dilution with ISOTON II isotonic solution,0 �m aperture, model MultisizerTM 3 Counter®, Beckman,iami, FL, USA). Cultures were periodically checked for

acterial contamination by plating on nutrient agar (Difco-acto).

Buffers known to be non-complexing for the selected met-ls (Wehr et al., 1986) were chosen: for media with pHetween 7 and 8, HEPES was used, and for solutions with pHetween 5.0 and 6.5, MES (N-[2-hydroxyethyl]piperazine-N′-2-ethanesulfonic acid]; 10−2 M; Sigma–Aldrich) was used.

All polycarbonate containers and glassware were soaked for4 h in 10% (w/v) HNO3 (ACS grade, EMD Chemical Inc.,armstadt, Germany), rinsed seven times with ultra pure water

or algal cultures was also autoclaved at 121 ◦C for 15 min. Saltssed for cultures and experiments were of analytical grade oretter.

L. Francois et al. / Aquatic Toxicology 84 (2007) 123–132 125

Table 1Molar concentrations of modified high salt media used for algal growth and uptake experiments (carbonates are assumed to be at equilibrium with the atmosphere)

Ions MHSM-1 (growth) MHSM-C (growthbefore experiment)

MHSM-R1 (rinsingbefore exposure)

MHSM-E (exposure) MHSM-R2 (rinsingafter exposure)

NH4 9.37× 10−4 9.37× 10−4 9.37× 10−4 9.37× 10−4 9.37× 10−4

Cl 5.98× 10−6 3.88× 10−6 – − −K 4.22× 10−3 4.22× 10−3 4.22× 10−3 4.22× 10−3 4.22× 10−3

PO4 1.37× 10−4 1.37× 10−4 – − −NO3 5.07× 10−3 5.07× 10−3 5.07× 10−3 5.07× 10−3 5.07× 10−3

SO4 8.12× 10−5 8.12× 10−5 8.12× 10−5 8.12× 10−5 8.12× 10−5

Mg 8.12× 10−5 8.12× 10−5 8.12× 10−5 8.12× 10−5 8.12× 10−5

Ca 6.80× 10−5 6.80× 10−5 6.80× 10−5 6.80× 10−5 6.80× 10−5

Na 1.02× 10−4 1.02× 10−4 1.02× 10−4 1.02× 10−4 1.02× 10−4

BO3 3.01× 10−6 3.01× 10−6 – – –Mn 2.10× 10−6 1.00× 10−9 – Variable −Cd – – – Variable −EDTA 8.06× 10−7 8.06× 10−7 – – 1.00× 10−4

Fe 5.92× 10−7 5.92× 10−7 – – –MoO4 3.00× 10−8 3.00× 10−8 – – –Zn 2.43× 10−8 2.43× 10−8 – – –Co 1.09× 10−8 1.09× 10−8 – – –C −11 −11

B

buffe

2

tstogabtpmcarm1

cWorStrr

2

2

1A1

a(lCcp(CdApabCtToaitd(pMejbato

u 7.04× 10 7.04× 10 –uffera 10−2 10−2 0

a HEPES was used to buffer pH between 7.0 and 8.0 while MES was used to

.2. Selection of metals

The main selection criterion for the metals to be tested washeir speciation in aqueous solution. The use of a metal with veryimple speciation, minimally affected by pH variations, facili-ates data interpretation (e.g., absence of significant covariancef hydroxo- and carbonato-metal complexes with pH). Man-anese, an essential metal, and cadmium, generally considereds a non-essential toxic metal, were good candidates: they areoth environmentally relevant and the aquo-ion, M(H2O)6

2+, ishe dominant metal form (98–99%) in simple media betweenH 5 and 8 for manganese and between pH 5 and 7 for cad-ium. At pH 8, formation of cadmium carbonate complexes

an be significant and over our wide metal concentration rangen over saturation in CdCO3(s) might have occurred; for thiseason we limited the upper pH range to 7 for cadmium. Bothetals are available as convenient radioisotopes, 54Mn and

09Cd (2 mCi �mol−1, for manganese, and 28 mCi �mol−1, foradmium; 0.1 M HCl solutions purchased from Perkin-Elmer,oodbridge, Ont., Canada), which facilitated measurements

f metal uptake. Metal stock solutions were diluted with non-adioactive metal (1000 �g mL−1, 2–5% HNO3, ICP Standard,CP Science, Que., Canada) to adjust metal concentrations in

he exposure media, when needed, so as to ensure sufficientadioactivity while minimizing needless consumption of theadioisotope.

.3. Metal uptake experiments

.3.1. Influence of pH

Widely used uptake methods were employed (Knauer et al.,

997; Fortin and Campbell, 2000; Kola and Wilkinson, 2005).n aliquot of the main algal culture was transferred into a00 mL of fresh medium MHSM-C (Table 1) in order to have

tmce

– –10−2 0

r pH between 5.0 and 6.5.

n initial algal density of 40,000 cells mL−1. Because MHSM-1Table 1) is very rich in manganese, its total concentration wasowered from 2× 10−6 M (MHSM-1) to 1× 10−9 M (MHSM-) to provide more environmentally realistic background Mnoncentrations in the test culture cells. To be consistent in ourrotocol, all experiments were carried out in the same mannereven when measuring Cd uptake, algae were grown in MHSM-); preliminary algal growth experiments showed no significantifference when algae were grown in MHSM-1 or MHSM–C.fter 48 h, which corresponded to the end of the exponentialhase of growth (algal density∼800,000 cells mL−1), sufficientlgae were harvested on a filter membrane (2.0 �m, polycar-onate membrane, IsoporeTM; Millipore, Mississauga, Ont.,anada) using a vacuum pressure ≤10 cm Hg, and rinsed five

imes with 10 mL of a simplified medium (MHSM-R1 medium,able 1). The rinsing step was designed to minimize carry-overf metals (micronutrients) as well as algal exudates and lig-nds from the growth medium. Next, algae were re-suspendedn 5 mL of the simplified medium MHSM-R1 and size distribu-ion, average surface area, and cell abundance were immediatelyetermined using the particle counter. Cells were then distributedfinal cell density of 15,000 cells mL−1) in 250 mL Erlenmeyerolycarbonate flasks containing 100 mL of the exposure mediumHSM-E. This exposure medium was prepared 24 h prior to the

xperiment to ensure metal equilibrium in solution, and filteredust before inoculation through a 0.2 �m polycarbonate mem-rane to remove any particulates. Each exposure condition (pHnd concentration) was tested in triplicate and repeated at leastwice on different batch cultures. We chose an exposure timef 45 min as a compromise between our desire, on one hand,

o minimize the impact of cell exudates, which could affect

etal speciation in solution, and also to minimize changes inell abundance or volume that could modify exposure param-ters, and, on the other hand, to maximize gamma emissions

1 Toxic

foctEmcsgw((udtBNogitsp

pawwTa1

mlmtRoct

m1iAttpsatimmeooO

elto

(pmocwa

2

u2t8fmtcoceaOiisn

2

a(cbcfia1H

2

ita

2

26 L. Francois et al. / Aquatic

rom our harvested cells. At the end of each exposure, 2 mLf each replicate were used to determine cell density and spe-ific surface area whereas the rest of the aliquot was filteredhrough two superimposed 2.0 �m porosity filter membranes.xperiments were conducted with low cell densities, to mini-ize metal depletion through uptake and adsorption by the algal

ells. A 2-mL aliquot of the filtrate was used to determine if aignificant loss of dissolved metal had occurred; if the loss wasreater than 10%, the data were excluded. The harvested cellsere then rinsed five times with 10 mL of MHSM-R2 medium

Table 1) containing 10−4 M of ethylenediamine tetra-acetate,Na4EDTA; Sigma–Aldrich). This rinsing step is commonlysed to remove metals adsorbed to the algal surface, allowing aistinction between absorbed metal (retained in the algal cells onhe top filter) and adsorbed metal (collected in the filtrate)—seeates et al. (1982), Mehta et al. (2002), Hassler et al. (2004).ote that the absorbed metal increased linearly as a functionf time (<60 min), and the regression line extended to the ori-in (data not shown), indicating that the discrimination betweenntracellular metal and adsorbed metal using EDTA was effec-ive. The metal radioactivity retained by the bottom filter wasubtracted from the radioactivity of the top filter to correct forassive metal adsorption to the membrane.

Each 2 mL-aliquot collected during the experiment waslaced in a 4 mL glass vial. The filters were carefully foldednd inserted into identical vials and 2 mL of ultra-pure waterere added to provide similar counting matrices. Every sampleas analysed by a gamma counter (Wallac 1480 WizardTM 3,urku, Finland). Counting windows of 600–820 keV (main peakt 700 keV) for 54Mn and 16–36 keV (main peak at 22 keV) for09Cd were used; the counting time was set to 2000 s and/or to aaximum collection of 100,000 counts, ensuring a sufficiently

ow uncertainty (<0.3%). Counts per minute were converted intoetal molar concentrations using detector efficiency, radioac-

ive decay and exposure solution radioisotope specific activity.esults were normalized for the total algal surface area presentn the filter. In a typical uptake experiment, the C. reinhardtiiells had a total surface area of approximately 70 �m2 and aotal volume of approximately 55 �m3.

The operationally defined intracellular metal fraction waseasured at increasing free metal ion concentrations (range:× 10−10 to 4× 10−4 M); a relatively broad range was used

n order to approach saturation of trans-membrane transport.lthough such concentrations are higher than those encoun-

ered in the environment, conditions approaching saturation ofhe transporters are essential for the model to converge and torovide good estimates of uptake parameters. Between four andix independent experiments were performed at each pH testednd for both metals. These repetitions were necessary to coverhe full range of concentrations and to test for reproducibil-ty among different cell batch cultures. The pH of the growth

edium (MHSM-C) was the same as the pH of the exposureedium (MHSM-E) which was checked minutes prior to the

xperiment and readjusted when required (to within a rangef ±0.02) using NaOH (0.5 M, Sigma Ultra, Sigma–Aldrich)r HNO3 (0.5 M, trace metal grade, Fisher Scientific, Ottawa,nt., Canada). At the beginning of the study, adsorption in each

oo

ology 84 (2007) 123–132

xperiment was measured on only one of the three replicates; inater experiments, it was measured on all three replicates. Suchriplicate measurements of adsorption were obtained for at leastne experiment for each set of conditions.

The influence of the pH of the antecedent growth mediumMHSM-C) on the saturation of trans-membrane metal trans-orters was also tested twice for each set of conditions. Foranganese, cells were grown at pH 8 and exposed at pH 6.5, and

thers were grown at pH 6.5 and exposed at pH 8. Similarly foradmium, cells were grown at pH 5 and exposed at 6.5 and othersere grown at 6.5 and exposed at 5. For all these experiments,

dsorption was also measured, for all three replicates.

.3.2. Influence of CaCalcium is known to be a good inhibitor of divalent metal

ptake in algae (Hauck et al., 2002; Slaveykova and Wilkinson,002; Mosulen et al., 2003; Kola and Wilkinson, 2005). Weested for this effect by increasing Ca2+ concentrations at pH

for manganese (three independent experiments) and pH 6.5or cadmium (two independent experiments). The total nominaletal concentration was kept constant at 0.5 �M, a concen-

ration well below the transport saturation level. The Ca2+

oncentration was varied from 0 to 1× 10−3 M by the additionf the appropriate volumes of stock solutions prepared fromalcium nitrate salt (Sigma–Aldrich). The ionic strength of thexposure solutions was maintained at a constant 6 meq L−1 bydjusting the concentration of KNO3 (Caledon, Georgetown,nt., Canada) in the medium. We thus assumed that potassium

ons in the low mM concentration range would have a negligiblempact on Cd and Mn uptake. Indeed, Heijerick et al. (2002)howed that K+ ions (concentration range: 0.08–1.12 mM) hado effect on zinc toxicity to green algae.

.3.3. Influence of CrBecause the decay product of 54Mn is 54Cr, and chromium is

potentially toxic metal, we tested the influence of chromiumIII) on manganese–membrane interactions. Chromium con-entrations encountered in our experiment were estimated toe between 2.5× 10−12 and 6.6× 10−11 M. The effect ofhromium on manganese uptake was tested at pH 6.5, with axed free manganese concentration of (3.4± 0.4)× 10−10 Mnd eight nominal concentrations of chromium between.5× 10−12 and 1× 10−8 M (Cr(NO3)3, 10,000 �g mL−1, 4%NO3, ICP Standard, SCP Science).

.4. Speciation modelling

Metal speciation in all media was calculated with the chem-cal equilibrium software MINEQL+ (v. 4.5) with an updatedhermodynamic database prepared from a recent review of avail-ble thermodynamic data (Schecher, 2001; Martell et al., 2004).

.5. Statistical analysis

Unless indicated otherwise, results are given as mean valuesf three replicates ±1S.E. All regressions and the optimizationf model parameters and their standard errors to the observed

L. Francois et al. / Aquatic Toxicology 84 (2007) 123–132 127

Table 2Estimated model parameters (standard error intervals of the equilibrium constants, log K, are presented between square brackets)

Mn JMMMn (�mol m−2 min−1) KMM

Mn (�M) log KCa log KMn Jmax (�mol m−2 min−1)

pH 8.0 11.2 ± 1.2 77 ± 31 11.2 ± 1.2pH 6.5 8.5 ± 1.0 253 ± 66 3.70 [3.56–3.81] 4.20 [4.11–4.27] 5.8 ± 0.2pH 5.0 2.6 ± 0.3 85 ± 25 2.6 ± 1.0

Cd JMMCd KMM

Cd log KCa log KCd Jmax

pp

upIow

3

3

3

c

Ftflp(

sopewsdit

H 6.5 57.2 ± 4.9 312 ± 51H 5.0 24.6 ± 2.7 156 ± 37

ptake fluxes were determined using the Graph Pad Prism 4 com-uter programme (Graph Pad software, San Diego, CA, USA).n order to compare internalization fluxes between different setsf growth and exposure conditions, Monte Carlo simulationsere performed, (see Section 4.2).

. Results

.1. Metal internalization fluxes

.1.1. Influence of pHMetal internalization fluxes as a function of the free metal ion

oncentration (Fig. 1) showed, at all pHs, a saturation plateau

ig. 1. Manganese (a) and cadmium (b) internalization fluxes as a function ofhe free metal ion concentration, at pH 8 (�), pH 6.5 (�), pH 5 (©). Predicteduxes were made using Eq. (12), for pH 8 (solid line), pH 6.5 (dashed line) andH 5 (dotted line). Standard errors are given when larger than the symbol sizen = 3).

oa

wda6a8

3

soflw0ia

gJcic

3

tcnsnw5

57.2 ± 4.94.75 [4.63–4.85] 4.19 [4.18–4.21] 34.9 ± 1.2

tarting when the free ion concentration reached∼0.1 mM. Thebserved behaviour is typical of a Michaelis–Menten saturationrocess (estimated parameters are presented in Table 2). How-ver, the maximum Mn internalization fluxes, JMM

Mn , increasedith pH, a trend inconsistent with the BLM. Moreover, half-

aturation constants KMMm are expected to increase when pH

ecreases but we obtained incoherent results. Indeed, KMMMn

ncreased from pH 8.0 to 6.5 but then decreased at pH 5.0 backo the value obtained at pH 8. In a manner similar to the resultsbtained for Mn, JMM

Cd and KMMCd also decreased between pH 6.5

nd 5.0.Metal internalization fluxes were not sensitive to the pH at

hich the cells had been grown (Fig. 2). Metal uptake wasetermined as a function of free metal–ion concentration forlgae cultivated at a low pH and exposed at a high pH (Mn: pH.5→ pH 8.0, Fig. 2a; Cd: pH 5.0→ pH 6.5, Fig. 2c) and forlgae cultivated at a high pH and exposed at a low pH (Mn: pH.0→ pH 6.5, Fig. 2b; Cd: pH 6.5→ pH 5.0, Fig. 2d).

.1.2. Influence of Ca and CrMetal internalization fluxes as a function of [Ca2+] showed a

ignificant inhibitive effect from calcium (Fig. 4). In the presencef a 10,000-fold excess of calcium (10−3 M), Mn internalizationuxes decreased 2-fold (from 0.080 to 0.037 �mol m−2 min−1)hereas cadmium uptake decreased more than 40-fold (from.43 to 0.01 �mol m−2 min−1). The decrease started from nom-nal concentrations of calcium greater than 1× 10−5 M for Mnnd greater than 1× 10−6 M for Cd.

Addition of an excess of chromium (III) relative to man-anese at pH 6.5, revealed no significant influence on Mn uptake:int = (1.4± 0.1)× 10−4 �mol m−2 min−1 when the chromiumoncentration was varied from 1.5× 10−12 to 1× 10−8 M (dom-nant metal form, Cr(OH)2

+, ∼99%), and [Mn2+] was heldonstant (3.4× 10−10 M).

.2. Metal adsorption

In contrast to the internalization results, metal adsorptiono the cell surface was linear over the entire range of free iononcentrations and over the entire pH range tested, with determi-ation coefficients comprised between 0.87 and 0.99 (Fig. 3). No

ignificant pH effect was observed for Mn (slopes were not sig-ificantly different, P = 0.26) but cadmium adsorption increasedith pH: at pH 6.5 the slope was 28.9± 1.8, whereas at pH.0 it was 11.7± 0.4 (slopes significantly different, P < 0.001).

128 L. Francois et al. / Aquatic Toxicology 84 (2007) 123–132

Fig. 2. Manganese and cadmium internalization fluxes as a function of the freemetal concentration for cells exposed to: (a) Mn at pH 8 (�,�); (b) Mn at pH 6.5(�, �); (c) Cd at pH 6.5 (�, �); and (d) Cd at pH 5 (�, ©). Monte Carlo simula-tions (solid lines) and the 95% confidence intervals (dashed and dotted lines) arealso plotted. Standard errors are given when larger than the symbol size (n = 3).

Fig. 3. Manganese (a) and cadmium (b) adsorption as a function of the freemetal concentration, at pH 8 (�), pH 6.5 (�), pH 5 (©); (a): for algae cultivatedat pH 6.5 and exposed at pH 8 (�), for algae cultivated at pH 8 and exposedat pH 6.5 (�); (b): for algae cultivated at pH 5 and exposed at pH 6.5 (�),for algae cultivated at pH 6.5 and exposed at pH 5 (�). Linear regressions((t

Chasacn

4

4

tsp

J

iw

Mn: 0.96 < r2 < 0.99; Cd: 0.87 < r2 < 0.97) are plotted: pH 8 (solid line), pH 6.5dashed line) and pH 5 (dotted line). Standard errors are given when larger thanhe symbol size (n = 3).

ulturing the algae at a different pH from the pH of exposuread little influence on Mn adsorption when the cells were grownt pH 8 and exposed at pH 6.5 (P = 0.11) but the slopes weretatistically different from controls when the cells were grownt pH 6.5 and exposed at pH 8 (P = 0.04). For Cd, adsorptionurves for algae cultivated and exposed at different pHs wereot significantly different (P > 0.4).

. Discussion

.1. Modelling metal internalization fluxes

For C. reinhardtii, both manganese and cadmium internaliza-ion fluxes could be modelled by a typical Michaelis–Mentenaturation process (Eq. (1)); estimated model parameters areresented in Table 2.

int = JMMm [M2+]

KMM + [M2+](1)

m

In this simple form, the Michaelis–Menten equation does notntegrate competition reactions. If we look at metal interactionsith transport sites from a thermodynamic equilibrium point of

Toxic

vb

{J

wMicdc

tc

b

{

t

{

J

wiaJcccwuaktci

mo

ucd

J

woKruTe1c

cwork of Kola and Wilkinson (2005) that showed no significantinhibition of cadmium uptake by C. reinhardtii. Also, Kehres etal. (2000) showed that magnesium did not inhibit Mn uptake inSalmonella typhimurium and Escherichia coli. However, Hauck

L. Francois et al. / Aquatic

iew, the reaction of the free divalent metal ion, M2+, with theiotic ligand, X−, can be represented as follows:

X− +M2+ KM←→ X−Mkint−→ X− + (M2+)Cell (2)

X−M} = KM[M2+]{ X−} (3)

int = kintKM[M2+]{ X−} (4)

here KM is the equilibrium constant for the reaction of metalwith the biotic ligand X−, kint is the rate of metal internal-

zation, and Jint is the metal internalization flux. To simplify,harges on complexes and water molecules in the metal coor-ination sphere are not shown; braces {} correspond to surfaceoncentrations and square brackets [] to solution concentrations.

If we consider all possible surface complexation reactions, aransport site can bind a proton H+ (Eq. (5)), another competingation Cn+ (Eq. (6)), or the free metal M2+ (Eq. (7)):

X− + H+ KH←→ X− H KH = { X− H}{ X−}[H+]

(5)

X− + Cn+ KC←→ X− C KC = { X− C}{ X−}[Cn+]

(6)

X− +M2+ KM←→ X−M KM = { X−M}{ X−}[M2+]

(7)

The total concentration of sites is given by the following massalance equation:

X}T = { X−} + { X− H} + { X− C} + { X−M} (8)

To express this concentration as a function of the concentra-ion of free sites, we combined Eqs. (5), (6), (7) and (8):

X}T = { X−}(1+KH[H+]+KC[Cn+]+KM[M2+]) (9)

Eqs. (4) and (9) combined give:

int = JmaxKM[M2+]

(1+KH[H+]+KC[Cn+]+KM[M2+])(10)

here Jmax = kint { X}T. According to the BLM, protons willnhibit metal uptake competitively: as the pH decreases, [H+]nd the product KH[H+] increase, leading to a decrease inint. The maximum internalization rate Jmax is assumed to beonstant. However, protons could also inhibit metal uptake non-ompetitively, by affecting either KM or Jmax. A rise in pHausing an increase in the biotic ligand affinity for a metal (KM)ill result in a steeper slope of Jint as a function of [M2+] but willltimately reach the same upper limit (Jmax). A rise in pH causingn increase in Jmax could be explained either by an increase inint (conformational changes of metal transport proteins leadingo an increase in the intrinsic metal internalization rate, as typi-ally observed for a partial non-competitive inhibitor) or by an

ncrease in { X}T (changes in the number of transporter sites).

Since Eq. (10) contains many adjustable parameters, we esti-ated several parameters under simple conditions where only

ne parameter was varied. Our initial step was to examine metal

Ftf(t

ology 84 (2007) 123–132 129

ptake as a function of [Ca2+] at constant pH and metal con-entrations using non-linear regression and Eq. (11), in order toetermine the calcium binding constant KCa:

int = A

1+KCa[Ca2+](11)

here A is an adjustable parameter regrouping all constant termsf Eq. (10). Note here that we assumed that both KM[M2+] andH[H+] are �1. The first assumption is reasonable given the

elatively low [M2+] (0.5 �M) that we used with respect to sat-ration concentrations, and the second will be discussed below.his analysis yielded values for KCa: 103.70 M−1 with a standardrror interval of [103.56–103.81] for manganese binding sites; and04.75 M−1 with a standard error interval of [104.63–104.85] foradmium binding sites (Fig. 4, Table 2).

We assumed that Ca was the only significant competingation (Cn+, Eq. (10)). Magnesium was not tested based on the

ig. 4. Manganese (a) and cadmium (b) internalization fluxes as a function ofhe free calcium concentration, at pH 8 (�) for manganese (a) and at pH 6.5 (�)or cadmium (b). Predicted fluxes were calculated using Eq. (11), for manganesesolid line) and cadmium (dashed line). Standard errors are given when largerhan the symbol size (n = 3). Insets show the same data on a log 10 scale.

1 Toxic

eecOtt

tM(g(ta

J

KaIoKsrGematatp[

mgTt

4

thDctutbtma

K

tvosipuhefUmpotit

4

wdHvncpbiirpaa

cmbii{pasCtftipot

30 L. Francois et al. / Aquatic

t al. (2002) reported that MgCl2 reduced manganese uptake inpiphytic lichens, but to a smaller extent than CaCl2. Ideally, allations should be tested to incorporate all potential inhibitors.mitting a significant inhibitor would result in an underestima-

ion of KM, provided that the inhibitor concentration is constanthroughout.

These KCa values were used to estimate the model parame-ers for manganese and cadmium, using a competitive inhibition

ichaelis–Menten model. Maximum internalization flux valueswith Jmax = JMM

m in Eq. (1)) were obtained at pH 8.0 for man-anese (11± 1 �mol m−2 min−1) and at pH 6.5 for cadmium57± 5 �mol m−2 min−1). At these pH values, we consideredhat KH [H+]� 1 and Eq. (10) was simplified to estimate KMnnd KCd:

int = JmaxKM[M2+]

(1+KCa[Ca2+]+KM[M2+])(12)

Through non-linear regression analysis, we obtainedMn = 104.20 with a standard error interval of [104.11to 104.27]nd KCd = 104.19 with a standard error interval of [104.18–104.21].ntroducing KCa and KM in Eq. (10) along with the Jint databtained at pH 5 should have enabled us to estimate a value forH. However, the overlap in the data at concentrations below

aturation levels at all tested pH values prevented us from nar-owing down a precise value for the binding of protons (KH).iven that all measurements were done in triplicate, several

xperiments (using different batch cultures) were used to modeletal uptake, thus incorporating the inherent biological vari-

bility in our graphs. The resultant scatter no doubt contributeso the inconsistent trend in KMM

m as a function of pH (i.e. KMMm

re expected to increase when pH decreases). Although pro-on competition most probably occurs, it remains an insensitivearameter in our Cd and Mn uptake model, indicating that KHH+]� 1 even at the lowest pH tested (5.0) such that KH� 105.

These estimated parameters were introduced into our finalodel (Eq. (12)) and Jmax values at pH 6.5 and 5.0 for man-

anese, and at pH 5.0 for cadmium were calculated (Fig. 1,able 2). From the observed changes in Jmax with pH (Fig. 1),

he influence of protons is obviously of a non-competitive nature.

.2. Influence of growth medium pH

The influence of acclimation (or absence of acclimation) ofhe cells to the experimental medium pH on metal uptake ratesas been evoked before to explain discrepancies. For example,e Schamphelaere and Janssen (2006) compared their study of

opper toxicity to P. subcapitata and Chlorella vulgaris withhe work of Macfie et al. (1994) on metal (Cd, Co, Cu and Ni)ptake in C. reinhardtii; they suggested that the difference inhe magnitude of the pH effect (lower in C. reinhardtii) coulde partially explained by the fact that the cells were acclimatedo the experimental pH for 4 days before being exposed to the

etal whereas De Schamphelaere and Janssen (2006) did notcclimate their algae.

Means and standard errors of each variable (Jmax, KCa andMn or KCd; see modelling in Section 4.1) were used to calculate

iTrc

ology 84 (2007) 123–132

he theoretical internalization flux and its 95% confidence inter-al using Monte Carlo simulations. The simulations were basedn the experimental results for algae grown and exposed at theame pH. A total of 10,000 simulations were performed, assum-ng a normally distributed (Gaussian) random error and errorropagation laws were used. The simulations results alloweds to determine if a variation in the pH of the growth mediumad a significant effect on metal uptake and helped comparexperimental results obtained for algae grown at a pH dif-erent from the one used for the metal exposure (Fig. 2).ptake of Cd or Mn was not affected by the pH of the growthedium; all results obtained with cells grown at a different

H than the exposure medium pH were consistent with thosebtained with cells grown and exposed at the same pH. Notehat, in Fig. 2, the line for the lower limit of the confidencenterval is interrupted because of the log 10 scale represen-ation.

.3. Mechanism of proton inhibition of Mn and Cd uptake

As expected, the main observation in our experimentsas that, when pH decreased, metal uptake also decreased,emonstrating the inhibition of metal uptake by protons.owever, Michaelis–Menten modelling (Table 2) gave JMM

malues that increased with pH, which is a clear indicator of aon-competitive inhibition phenomenon. Considering the non-ompetitive inhibition suggested by the behaviour of Jmax, alausible explanation for the uptake mechanism is that protonsind to the algal membrane at sites that differ from those involvedn surface metal complexation. The consequence of this bindings a change in conformation of the metal binding sites, whichesults in lower uptake rates of the metal. This mechanism hasreviously been suggested (Harrison et al., 1986; Chowdhurynd Blust, 2001; Kola and Wilkinson, 2005; Fortin et al., 2007)nd our present data set clearly supports it.

A change of Jmax can be interpreted as a change of kint (ahange in the intrinsic rate of transmembrane uptake of theetal) or as a change in the total number of sites capable of

inding and internalizing the metal, { X}T. Non-competitivenhibition could result from an irreversible binding of annhibitor, disabling the transporter (a decrease in “active”

X}T). As the pH is lowered, a growing proportion of trans-ort sites could be deactivated. However, our experiments wherelgae were cultivated at a pH different from the pH of expo-ure showed no sign of irreversible binding. Indeed, Mn andd uptake were not affected by the pH of the growth medium;

he exposure medium pH was the key parameter. From a dif-erent perspective, it could be argued that cells produce moreransporters at higher pH values (an increase in { X}T). Thiss however very unlikely considering that cells grown at a lowH but exposed at a high pH gave results consistent with thosebtained with algae grown and exposed at a high pH. The syn-hesis of new transporters is not an instantaneous process and is

mprobable in such a short time scale (exposure time = 45 min).herefore, the observed change in Jmax most probably reflects a

eversible conformational change in the uptake sites, inducing ahange in kint. The hypothesis that protons are simply competi-

Toxic

tb

4

sswK1ec

pt(Kp(uwilsmt(fveHuaoac

4

o1opldt

ptpraseb

4

VbHpotvataigbewimmpcr

5

asbH41citoKiBaqctrpcc

A

L. Francois et al. / Aquatic

ors for manganese and cadmium uptake in C. reinhardtii muste rejected.

.4. Biotic ligand affinity

Very few studies have proposed conditional stability con-tants for Mn binding to transport sites in algae. Moreover, thepecies tested were marine algae, which makes comparisonsith our results hazardous. Nevertheless, the values reported forMn by Slaveykova and Wilkinson (2005) ranged from 107 to03.1 (some species exhibited a biphasic transport profile). Ourstimated value for KMn (104.20) lies within this range albeitloser to the lower limit.

Values for KCd found for C. pyrenoidosa (105.5 M−1 atH 7.0) and for C. vulgaris (104.4 M−1 at pH 6.8) are closeo our results (KCd = 104.19)—see Slaveykova and Wilkinson2005) and references therein. In the case of C. reinhardtii,ola and Wilkinson (2005) reported a much lower Jmax atH 7.0 (1.3× 10−11 mol cm−2 min−1) than we have observedJmax = 5.72× 10−9 mol cm−2 min−1) and they observed a sat-ration at much lower concentrations (around [Cd2+] = 10−6 M,hereas our saturation plateau started at ∼10−4 M), resulting

n a conditional constant cKCd = 1/KMMm = 106 M−1 (calcu-

ated from Michaelis–Menten modelling which assumes noignificant inhibition from other cations) almost two orders ofagnitude higher than ours (KCd = 104.19). On the other hand,

he binding affinity of Ca for the Cd transport sites was closeKCa = 104.5) to ours (KCa = 104.75). They also estimated a valueor proton binding to Cd transport sites (KH = 105.2), but thisalue did not successfully explain the data and they consid-red a second proton binding site on the same transporter (2+ + X−↔ X–H2; KH2 = 1011.3). The large differences inptake parameters might reflect the composition of the growthnd exposure media used by Kola and Wilkinson (2005) (10−2 Mf morpholino propane sulfonic acid, ionic strength of 10−2 Mdjusted with NaNO3, and most notably the absence of Ca), orome from differences in the origin of the algal strains used.

.5. Adsorption

No significant pH effect on manganese adsorption wasbserved, thus confirming previous observations (Schenck et al.,988 and references therein). Also, no significant saturation wasbserved over our very wide concentration range. Saturation canrovide an accurate estimate of the number of non-specific (noteading to internalization) adsorption sites. We thus could notetermine an adsorption equilibrium constant for Mn binding tohe algal surface.

Unlike manganese, cadmium adsorption was influenced byH variations in the same manner as cadmium uptake; whenhe pH decreased, cadmium adsorption decreased as well. TheH at which the algae had been cultured did not influence theesults either, indicating that no major changes in the surface Cd

dsorption sites occur within this pH range (5.0–6.5). Again, noignificant saturation was observed and no accurate adsorptionquilibrium constant for Cd binding to the algal surface coulde computed.

sseR

ology 84 (2007) 123–132 131

.6. Implications for an algal BLM

The BLM is based on numerous hypotheses (Campbell, 1995;an Leeuwen, 1999; Di Toro et al., 2001; Campbell et al., 2002)ut two, in particular, were tested in this work: (i) cations like+ or Ca2+ can compete with the metal Mz+ in surface com-lexation reactions; and (ii) the nature of the biological surfacer its ionic transporters remain unchanged during the exposureo the metal. Even though we observed metal uptake over aery wide concentration range (reaching concentrations wellbove normal environmental values), the knowledge of satura-ion trans-membrane fluxes allowed the development of a moreccurate model for the mechanistic interpretation of processesnvolved in manganese and cadmium uptake. Our results sug-est that when pH varies, the changes in metal uptake cannote explained by the BLM theory. Should this non-competitiveffect of pH be applicable to all metal–algae interactions, thisould warrant a re-evaluation of the BLM limitations concern-

ng the influence of protons. Otherwise, the understanding ofetal uptake based on the model will be more empirical thanechanistic (and calculated constants more conditional) when

H varies; or BLM applications will be limited to media with nohanging pH conditions, which does not reflect relevant envi-onmental situations.

. Conclusion

The influence of pH on short-term uptake of manganesend cadmium by C. reinhardtii was studied to better under-tand the effect of protons on metal trans-membrane fluxes. Foroth metals, first-order biological internalization was observed.owever, maximum transport fluxes (Jmax) decreased more than-fold for manganese between pH 8.0 and 5.0, and decreased.6-fold for cadmium between pH 6.5 and 5.0, suggesting a non-ompetitive inhibition by H+-ions. A Michaelis–Menten typenhibition model was tested: the metal stability constants forransport sites and the stability constants for competitive bindingf Ca2+ and H+ were calculated. Surprisingly, for both metals,H was not a significant parameter, indicating that competitive

nhibition of protons (the type of inhibition considered in theLM) is negligible. We also observed that growing the algae atpH different from the exposure pH did not influence subse-

uent metal uptake, leading us to conclude that protons induceonformational changes in the membrane bound transport pro-ein that result in pH-sensitive internalization rates (kint). Ouresults demonstrate that the BLM in its present state does notroperly describe the true influence of pH on manganese andadmium uptake by algae and that a non-competitive inhibitionomponent must be integrated.

cknowledgments

Financial support to C. Fortin provided through the new

cientist initiative of the Fonds Quebecois de la rechercheur la nature et les technologies (FQRNT) and the discov-ry grant program of the Natural Sciences and Engineeringesearch Council of Canada (NSERC) is acknowledged.

1 Toxic

Pp

R

B

C

C

C

C

D

D

D

F

F

H

H

H

H

H

K

K

K

M

M

M

M

P

S

S

S

S

V

W

W

32 L. Francois et al. / Aquatic

.G.C. Campbell is supported by the Canada Research Chairrogram.

eferences

ates, S.S., Tessier, A., Campbell, P.G.C., Buffle, J., 1982. Zinc adsorptionand transport by Chlamydomonas variabilis and Scenedesmus subspica-tus (Chlorophycae) grown in semicontinuous culture. J. Phycol. 18, 521–529.

ampbell, P.G.C., 1995. Interactions between trace metals and aquatic organ-isms: a critique of the free ion activity model. In: Tessier, A., Turner, D.R.(Eds.), Metal Speciation and Bioavailability in Aquatic Systems. IUPAC,John Wiley and Sons, Chichester, UK, pp. 45–102.

ampbell, P.G.C., Stokes, P.M., 1985. Acidification and toxicity of metals toaquatic biota. Can. J. Fish. Aquat. Sci. 42, 2034–2049.

ampbell, P.G.C., Errecalde, O., Fortin, C., Hiriart-Baer, V.P., Vigneault, B.,2002. Metal bioavaibility to phytoplankton—applicability of the biotic lig-and model. Comp. Biochem. Physiol. Part C 133, 189–206.

howdhury, M.J., Blust, R., 2001. A mechanistic model for the uptake of water-borne strontium in the common carp (Cyprinus carpio L.). Environ. Sci.Technol. 35, 669–675.

e Schamphelaere, K.A.C., Janssen, C.R., 2006. Bioavailability models forpredicting copper toxicity to fresh green microalgae as a function of waterchemistry. Environ. Sci. Technol. 40, 4514–4522.

e Schamphelaere, K.A.C., Vasconcelos, F.M., Heijerick, D.G., Tack, F.M.G.,Delbeke, K., Allen, H.E., Janssen, C.R., 2003. Development and fieldvalidation of a predictive copper toxicity model for the green algaPseudokirchneriella subcapitata. Environ. Toxicol. Chem. 22, 2454–2465.

i Toro, D., Allen, H.E., Bergman, H.L., Meyer, J.S., Paquin, P.R., Santore,R.C., 2001. Biotic ligand model of the acute toxicity of metals. Technicalbasis. Environ. Toxicol. Chem. 20, 2383–2396.

ortin, C., Campbell, P.G.C., 2000. Silver uptake by the green alga Chlamy-domonas reinhardtii in relation to chemical speciation: influence of chloride.Environ. Toxicol. Chem. 19, 2769–2778.

ortin, C., Denison, F.H., Garnier-Laplace, J., 2007. Metal–phytoplankton inter-actions: modeling the effect of competing ions (H+, Ca2+ and Mg2+) onuranium uptake. Environ. Toxicol. Chem. 26, 242–248.

arrison, G.I., Campbell, P.G.C., Tessier, A., 1986. Effects of pH changes onzinc uptake by Chlamydomonas variabilis grown in batch culture. Can. J.Fish. Aquat. Sci. 43, 687–693.

assler, C.S., Slaveykova, V.I., Wilkinson, K.J., 2004. Discriminating betweenintra- and extracellular metals using chemical extractions. Limnol.Oceanogr. Methods 2, 237–247.

auck, M., Mulack, C., Paul, A., 2002. Manganese uptake in the epiphyticlichens Hypogymnia physodes and Lecanora conizaeoides. Environ. Exp.

Bot. 48, 107–117.

eijerick, D.G., De Schamphelaere, K.A.C., Janssen, C.R., 2002. Biotic ligandmodel development predicting Zn toxicity to the alga Pseudokirchneriellasubcapitata: possibilities and limitations. Comp. Biochem. Physiol. C Tox-icol. Pharmacol. 133, 207–218.

W

ology 84 (2007) 123–132

eijerick, D.G., Janssen, C.R., De Coen, W.M., 2003. The combined effectsof hardness, pH, and dissolved organic carbon on the chronic toxicity ofZn to D. magna: development of a surface response model. Arch. Environ.Contam. Toxicol. 44, 210–217.

ehres, D.G., Zaharik, M.L., Finlay, B.B., Maguire, M.E., 2000. The NRAMPproteins of Salmonella typhimurium and Escherichia coli are selectivemanganese transporters involved in the response to reactive oxygen. Mol.Microbiol. 36, 1085–1100.

nauer, K., Behra, R., Sigg, L., 1997. Adsorption and uptake of copper by thegreen alga Scenedesmus subspicatus (Chlorophyta). J. Phycol. 33, 596–601.

ola, H., Wilkinson, K.J., 2005. Cadmium uptake by a green alga can be pre-dicted by equilibrium modelling. Environ. Sci. Technol. 39, 3040–3047.

acfie, S.M., Tarmohamed, Y., Welbourn, P.M., 1994. Effects of cadmium,cobalt, copper, and nickel on growth of the green alga Chlamydomonasreinhardtii: the influences of the cell wall and pH. Arch. Environ. Contam.Toxicol. 27, 454–458.

artell, A.E., Smith, R.M., Motekaitis, R.J., 2004. Critical Stability Constantsof Metal Complexes Database, Version 8. 0. U.S. Department of Commerce,Gaithersburg, MD, USA.

ehta, S.K., Singh, A., Gaur, J.P., 2002. Kinetics of adsorption and uptake ofCu2+ by Chlorella vulgaris: influence of pH, temperature, culture age, andcations. J. Environ. Sci. Health Part a-Tox./Hazard. Subst. Environ. Eng. 37,399–414.

osulen, S., Domınguez, M.J., Vigara, J., Vılchez, C., Guiraum, A., Vega, J.M.,2003. Metal toxicity in Chlamydomonas reinhardtii. Effect on sulfate andnitrate assimilation. Biomol. Eng. 20, 199–203.

arent, L., Campbell, P.G.C., 1994. Aluminium bioavaibility to the green algaChlorella pyrenoidosa in acidified synthetic soft water. Environ. Toxicol.Chem. 13, 587–598.

checher, W.D., 2001. MINEQL+: Version 4.5, Environmental Research Soft-ware. Hallowell, ME, USA.

chenck, R.C., Tessier, A., Campbell, P.G.C., 1988. The effect of pH on ironand manganese uptake by a green alga. Limnol. Oceanogr. 33, 538–550.

laveykova, V.I., Wilkinson, K.J., 2002. Physicochemical aspects of lead bioac-cumulation by Chlorella vulgaris. Environ. Sci. Technol. 36, 969–975.

laveykova, V.I., Wilkinson, K.J., 2005. Predicting the bioavailability of metalsand metal complexes: critical review of the Biotic Ligand Model. Environ.Chem. 2, 9–24.

an Leeuwen, H.P., 1999. Metal speciation dynamics and bioavailability: inertand labile complexes. Environ. Sci. Technol. 33, 3743–3748.

ehr, J.D., Brown, L.M., Vanderelst, I.E., 1986. Hydrogen ion buffering of cul-ture media for algae from moderately acidic, oligotrophic waters. J. Phycol.22, 88–94.

ilde, K.L., Stauber, J.L., Markich, S.J., Franklin, N.M., Brown, P.L., 2006.The effect of pH on the uptake and toxicity of copper and zinc in a trop-ical freshwater alga (Chlorella sp.). Arch. Environ. Contam. Toxicol. 51,174–185.

ilkinson, K.J., Buffle, J., 2004. Critical evaluation of physicochemical param-eters and process for modelling the biological uptake of trace metals inenvironmental (aquatic) systems. In: van Leeuwen, P., Koster, W. (Eds.),Physicochemical Kinetics and Transport at Biointerfaces. John Wiley andSons, Chichester, UK, pp. 445–533.