URANIUM ACCUMULATION AND TOXICITY IN THE GREEN ALGACHLAMYDOMONAS REINHARDTII IS MODULATED BY pH

MICHEL LAVOIE,y SÉBASTIEN SABATIER,z JACQUELINE GARNIER-LAPLACE,z and CLAUDE FORTIN*yyInstitut National de la Recherche Scientifique, Centre Eau Terre Environnement (INRS-ETE), Qu�ebec, Canada

zInstitut de Radioprotection et Sûret�e Nucl�eaire, Service de Recherche et d’Expertise sur les Risques Environnementaux, Centre de Cadarache, BP3 13115,Saint Paul lez Durance, France

(Submitted 11 October 2013; Returned for Revision 20 January 2014; Accepted 24 February 2014)

Abstract: The effects of pH on metal uptake and toxicity in aquatic organisms are currently poorly understood and remain an evolvingtopic in studies about the biotic ligand model (BLM). In the present study, the authors investigated how pHmay influence long-term (4 d)uranium (U) accumulation and chronic toxicity in batch cultures of the freshwater green alga Chlamydomonas reinhardtii. The toxicityexpressed as a function of the free uranyl ion wasmuch greater at pH 7 (effective concentration, 50% [EC50]¼ 1.8� 10�9MUO2

2þ) thanat pH 5 (EC50¼ 1.2� 10�7MUO2

2þ). The net accumulation rate of U in algal cells wasmuch higher at pH 7 than at pH 5 for the same free[UO2

2þ], but the cells exposed at pH 5 were also more sensitive to intracellular U than the cells at pH 7 with EC50s of 4.0� 10�15 and7.1� 10�13mol of internalized U cell�1, respectively. The higher cellular sensitivity to U at pH 5 than at pH 7 could be explained partly bythe increase in cytosolic U binding to algal soluble proteins or enzymes at pH 5 as observed by subcellular fractionation. To predict Uaccumulation and toxicity in algae accurately, the important modulating effects of pH on U accumulation and U cellular sensitivity shouldbe considered in the BLM. Environ Toxicol Chem 2014;33:1372–1379. # 2014 SETAC

Keywords: Uranium Toxicity Subcellular distribution Accumulation pH

INTRODUCTION

It is now well recognized that metal uptake and toxicity inaquatic organisms are strongly linked to metal speciation in theaqueous environment. A large body of evidence indicates thatmetal accumulation and its resulting toxicity in a range of aquaticorganisms is often proportional to the free ion concentrations insolution at constant pH and hardness; this has been conceptual-ized in the so-called free ion activity model (FIAM) [1].Incorporating the competition between cations (protons, calci-um, and magnesium) for binding sites at the biological surfaceand the resulting decrease in metal uptake and toxicity led to aderivative of the FIAM, the biotic ligand model (BLM) [2–5].However, some of the current premises behind the BLMmay notbe respected [6]. For example, from a physical perspective, themodel ignores dynamic reactions occurring at the surface of thecells by assuming a thermodynamic equilibrium between themetal and the surface [7]. Furthermore, from a biologicalperspective, the BLM overlooks noncompetitive interactions ofessential/nonessential metal and protons with the algal cells(e.g., synthesis of metal transport systems) and the subsequentrepercussions on toxic metal accumulation [8,9]. Such inter-actions usually appear within a few hours and may stronglyaffect chronic toxicity [10,11]. Reports have also shown,however, that it may sometimes occur rapidly within a short-term exposure (< 30min) [10,12].

Until recently, most of the research on the applicability of theBLM has focused on common divalent metals (Zn, Cd, Cu, Pb),which are characterized by a relatively simple metal speciation(free ion is the dominating species at pH< 7) in inorganic culture

medium [2,3]. Relatively fewer studies have been carried outwith metals having more complex aqueous speciation, such asuranium (U; see Goulet et al. [13] for a recent review). Perhapsdue to concerns regarding the use of depleted U in armedconflicts [14], most of the literature on U uptake and toxicity inaquatic organisms has been published in the last 15 yr or so.Among these publications, only a few have determined Utoxicity in algae over long-term exposures [15–17]. As aresult, U toxicity as well as the modulating effects of majorcations and pH on U toxicity remain poorly understood inalgae [9,15,18,19]. Such quantitative data are essential to derivepertinent water quality guidelines for U [20]. A betterunderstanding of the mechanistic basis of U uptake and toxicityis also needed to refine the BLM.

Background aqueous total U concentrations in freshwatersystems are generally below 4 nM U, but concentrations up to6mM U have been observed in Canadian mining areas [13].Uranium is a class A metal present as various hydroxo-,carbonato-, and poorly soluble phosphato-complexes in aqueoussolution at neutral pH. Its speciation is also highly sensitiveto pH, that is, speciation undergoes major changes as pHdecreases to 5, where the free uranyl ion (UO2

2þ) predom-inates [13]. We thus decided to investigate the effect of pH on Uuptake and toxicity in a well-known freshwater alga, Chlamy-domonas reinhardtii, cultivated in chemically well-definedculture media. First, we measured the toxicity of U at 2 differentpHs (7 and 5) in long-term batch algal cultures. Second, tounderstand U toxicity better, the amount of intracellular andadsorbed U was measured after 4 d of growth at pH 7 and pH 5.Third, the subcellular U distribution within soluble (cytosol) andinsoluble (membranes, cell walls, flagella, organelles) algalpools was then studied to look at the possible relationshipbetween U sensitivity and U cellular distribution. Finally, weattempted to model long-term U accumulation based on theshort-term U transport kinetics measured in the same algal

All Supplemental Data may be found in the online version of this article.* Address correspondence to claude.fortin@ete.inrs.ca.Published online 5 March 2014 in Wiley Online Library

(wileyonlinelibrary.com).DOI: 10.1002/etc.2565

Environmental Toxicology and Chemistry, Vol. 33, No. 6, pp. 1372–1379, 2014# 2014 SETAC

Printed in the USA

1372

species (data from Fortin et al. [9]). Comparing the modeled andthe experimentally measured U content values allowed us toverify the extent of algal acclimation to U and pH during thelong-term exposure.

MATERIALS AND METHODS

Test organisms and culture conditions

The unicellular green alga C. reinhardtii P.A. Dang wasobtained from the Culture Collection of Algae and Protozoa(culture 11/32B). Asynchronous batch cultures were grownaxenically (as determined visually by culture plating) in 100mLof MHSM-F10 (modified high salt medium—low phosphorusand low trace metal concentrations) culture medium with no pHbuffers (Table 1). Stock algae cultures were maintained in 250-mL Erlenmeyer glass flasks under constant rotary agitation(60 rpm), illumination (60–80mmolm�2 s�1), and temperature(20� 1 8C). A 2-mL subsample of algal culture was transferredto a sterile culture medium every week using a sterile serologicalpipette with an initial pH of 7.0. Culture media (without tracemetal) and material (Erlenmeyers and filtration apparatus)destined for algal cultures were autoclaved at 121 8C for 15min.Culture media were then supplemented with a filter-sterilizedtrace metal mix (0.2mm polycarbonate membrane, 47mm). Allplasticware and glassware were soaked for at least 24 h in 10%(w/v) HNO3, rinsed 7 times with ultrapure water (18 MV cm;Nanopure grade), and then dried under a class 100 laminar flowhood, where all manipulations that required precautions againstpossible contamination by airborne particulates or organismswere performed.

Optimizing the culture medium composition

The MHSM-F10 medium was designed specifically to limitthe formation of poorly soluble U complexes and simplify the U

speciation. The total phosphorus, potassium, chloride, micro-nutrients (Cu, Mn, Zn, Co, Fe, B, Mo), and ethylenediaminete-traacetic acid (EDTA) concentrations were lowered by 10-foldwith respect to theMHSM-1 culture medium [21]. The pH bufferwas also omitted in the MHSM-F10 medium to avoid possiblecomplexation of U by the buffers and potential effects onmembrane permeability [22]. The growth rates of C. reinhardtiiremained the same when cultured in the MHSM-1 or theMHSM-F10 culture media (data not shown).

Preparing the exposure media

Natural U from an acidic stock solution of UO2(NO3)2 wasused to spike the majority of experimental media, albeit the 233Uradioisotope (83 kBq mmol�1 as uranyl nitrate; CERCA-LEAFramatome) was used for 1 experiment. The 233U solution wasprepared in containers made of fluorinated ethylene propylenecopolymer (Teflon, Nalge Nunc) to minimize losses byadsorption to the container’s walls. The exposure mediaconsisting ofMHSM-F10mediumwith Uwere left to equilibrate1 d after adding U and adjusting the pH. Uranium speciation inthe exposure solutions was calculated with the chemicalspeciation model MINEQLþ (Ver. 4.6; [23]) with updatedequilibrium constants [24], using total metal concentrations asinput data. All calculations were constrained to a fixed input pH(7 or 5) at equilibrium with the atmosphere (pCO2¼ 10�3.5 atm).

Long-term exposure of batch cultures to uranium

Exponentially growing cells from the stock cultures weregently harvested on 2mm polycarbonate filters (Poretics) using avacuum pressure of <10 cm Hg. Harvested cells were rinsed 5times withMHSM-F10medium and then inoculated at 5000 cellsmL�1 in 3, 250-mL polycarbonate Erlenmeyers, each of whichcontained 125mL of exposure media (MHSM-F10þU).Cellular densities were determined with an electronic particlecounter (Coulter Z2 particle counter, 100-mm aperture; Beck-man) after appropriate dilution in Isoton II electrolyte. Theexposure media encompassed a range of total U concentrations(total [U]) between 0.1mM to 20mM and 0.1mM to 0.6mMat pH 7 and pH 5, respectively (including control cultures withoutadded U). During the U exposure, the pH of the media wereverified and, if needed, adjusted daily. The ionic strength waskept constant in all exposure media. Total dissolved natural U or233U concentrations were monitored by inductively coupledplasma–atomic emission spectroscopy (ICP–AES; Perkin Opti-ma 4300 DV) or liquid scintillation counting (Quantulus 1220,Wallac) after 6 h, 24 h, 48 h, 72 h, and 96 h of exposure. Algaewere harvested from the batch cultures (100–115mL) onto 2-mmpolycarbonate filters at the end of the exposure (96 h). The filtratewas used to measure total dissolved natural U or 233Uconcentrations by ICP–AES or liquid scintillation, respectively,after 6 h, 24 h, 48 h, 72 h, and 96 h of exposure. Cell densitieswere also monitored at each time point to calculate the maximumspecific growth rate (m) using Equation 1.

m ¼ lnNt48 � lnNt24

t48 � t24ð1Þ

where t24 and t48 are 24 h and 48 h, respectively, and Nt24 andNt48 are the cellular density at time 24 h and 48 h, respectively.

The harvested cells were then resuspended in 10mL of5� 10�5M EDTA (dissolved in simplified MHSM-R media,described in Table 1) to remove the operationally definedadsorbed U on the surface of algal cells [9,18]. The suspensionswere centrifuged at 10 000 g for 10min at 15 8C to separate the

Table 1. Total molar concentrations of constituents of the MHSM-F10culture medium and MHSM-R adapted from the MHSM-1 mediuma

CompoundsMHSM-F10 MHSM-R

M M

NH4 9.37� 10�4 9.37� 10�4

Cl 5.98� 10�7–

K 4.22� 10�4–

PO4 1.37� 10�5–

CO3 atm atmNO3 5.07� 10�3 5.07� 10�3

SO4 8.12� 10�5 8.12� 10�5

Mg 8.12� 10�5 8.12� 10�5

Ca 6.80� 10�5 6.80� 10�5

Na 5.00–7.00� 10�4 5.00–7.00� 10�4

BO3 3.01� 10�7–

Mn 2.09� 10�7–

EDTA 8.06� 10�8 5� 10�5

Fe 5.92� 10�8–

MoO4 1.87� 10�9–

Zn 1.22� 10�9–

Co 1.09� 10�9–

Cu 3.50� 10�12–

aThe carbonate concentration is assumed to be at equilibrium with theatmosphere. The pH of the medium was adjusted daily to 5.0 or 7.0 withNaOH (the total Na concentrations thus varied slightly as outlined in thetable) and no pH buffer was used. Ionic strengths of culture and rinse mediawere kept between 3.8 and 4.0meqL�1 [21].MHSM-F10¼modified high salt medium—low phosphorus and low tracemetals concentrations; MHSM-R¼modified high salt medium—rinse;atm¼ atmosphere.

Acidification increases cell sensitivity to uranium Environ Toxicol Chem 33, 2014 1373

cells (pellet) from the cell surface adsorbed metal (supernatant).The pellet was kept for intracellular distribution analyses(described in Intracellular distribution), whereas the natural Uor 233U content in the supernatant was analyzed using ICP-AESand liquid scintillation counting spectrometry, respectively.Prior to scintillation counting, the supernatants containing 233Uwere evaporated and the U redissolved with 1mL 1% v/v HNO3

followed by adding 19mL of Instagel scintillation cocktail(Packard Instrument).

Intracellular distribution

The algal pellets (obtained after separating the adsorbed Ufractions; method described in Long-term exposure of batchcultures to uranium) were resuspended in 10mL of simplifiedculture media containing 5� 10�5M EDTA (MHSM-R) toprevent U redistribution from the cytosol to insoluble cellulardebris. The distribution of U among the operationally definedsoluble (cytosol) and insoluble (cell membranes, cell walls,flagella, organelles) fractions was then investigated by sonica-tion for 4min on ice.

The disrupted cells were then centrifuged at 100 000 g for30min at 15 8C to separate the insoluble fraction (pellet) fromthe cytosol (supernatant). The 233U content of the supernatantwas measured by liquid scintillation as outlined in Long-termexposure of batch cultures to uranium. The pellet containing233U was mineralized for more than 24 h with concentratedHNO3 (65%w/v), evaporated on a sand bed, redissolved with1mL 1% v/v HNO3, and finally analyzed by liquid scintillationwith 19mL of scintillation cocktail. For the experiments usingnatural U, the cytosol fraction was acidified with HNO3 (finalconcentration of 2% v/v), whereas the pellet was mineralized inconcentrated HNO3, evaporated, and diluted in Milli-Q water toreach a final HNO3 concentration of 2% v/v. Mass balancecalculations comparing accumulated (insoluble and solublefractions) and adsorbed U to depletion of total ([Utot]) inexposure media agreed within 27%, indicating that adsorptivelosses on the Erlenmeyers, centrifugation tubes, pipette tips, andsonicator probes were less than 27%. Indeed, adsorptive lossesof U on the walls of the Erlenmeyer flasks (containing various Uconcentrations but no algae) over a 30-h period were low(<15%, data not shown).

Statistics

Significant differences among the U subcellular distributionswere evaluated with the nonparametric statistical test ofKruskall–Wallis using Systat software (Ver. 10.0) and resultswere considered to be significant if p< 0.05. Unless otherwiseindicated, errors are given as� 1 standard deviation.

Uranium toxicity modeling

Effective U concentrations inhibiting the algal growth rate by50% in batch cultures (EC50) were calculated by the nonlinearregression using a 4-parameter curve fitting approach (Equa-tion 2):

mt ¼ mmin þmmax � mmin

1þ ð½U�=EC50Þhill slope ð2Þ

where mt is the growth rate for a given U exposure concentration(normalized with respect to the growth rate of control cultures);mmax and mmin represent the modeled maximum and minimumrelative growth rates; [U] is the medium total or free Uconcentration; hill slope is the fitted parameter to the curve; and

EC50 is the effective concentration where 50% growth ratereduction is observed.

Steady-state uranium accumulation modeling

Intracellular U contents measured after 4 d of growth(approximating steady-state U intracellular concentrations)were modeled as a function of total or free U concentrationswith the following procedures and equations. First, we computedthe U uptake rates (Vint) as a function of total or free Uconcentrations at pH 7 or pH 5. These calculations were achievedby using the kinetic parameters of U transport systems inC. reinhardtii derived by Fortin et al. [9] for short-term (30min)exposures of algal cells to U at pH 7 or pH 5. The uptake rates(Vint) as a function of [Utot] were modeled with Equation 3incorporating the Michaelis–Menten parameters Km (half-saturation constant) and Vmax (maximal uptake rate achieved atU saturation of the transport systems) describingU uptake at pH 7(Km¼ 1.1� 0.2mM Utot; Vmax¼ [2.87� 0.28]� 10�17mol Ucell�1 min�1) or pH 5 (Km¼ 0.51� 0.07mM Utot; Vmax¼[6.93� 0.35]� 10�18mol U cell�1 min�1).

V int ¼ Vmax½Utot�Km þ ½Utot� ð3Þ

The U uptake rate (Vint) as a function of UO22þ was modeled

with Equation 4, which is a derivative of the Michaelis–Mentenequation incorporating cation (Hþ, Ca2þ, and Mg2þ) competi-tion with UO2

2þ for uptake.

Vint ¼ Vmax½UO2þ2 �KU

1þ ½UO2þ2 �KU

� �þ ½Hþ�KHU

� �þ ½Ca2þ�KCaU

� �þ ½Mg2þ�KMgU

� �

ð4Þ

where Vmax is the maximal UO22þ uptake flux for which

the U transport system achieved saturation at pH 7 (�1.8� 10�15mol U cell�1 min�1, as estimated by a first localminimum between modeled and observed values) or pH 5[6.93� 0.35]� 10�18mol U cell�1 min�1 [9]; the series KU,KH

U, KCaU , and KMg

U are the affinity constants of UO22þ, Hþ, Ca2þ,

and Mg2þ, respectively, for the U transport system (KU¼ 107.7

M�1; KHU ¼ 106.1 M�1; KCa

U ¼KMgU ¼ 104.2 M�1) [9]; [Hþ] is the

concentration of protons; and [Ca2þ] and [Mg2þ] are theconcentrations of the free Ca and Mg ion, respectively.

The second step was to compute the steady-state Uintracellular accumulation (likely achieved in a long-termexperiment), by taking into account the cellular division rate.At both exposure levels of pH, the steady-state intracellular Uconcentration ([Ucell], i.e., U content per cell) would thus be afunction of the steady-state U uptake rates (Vint computed withEquation 3 or Equation 4) divided by the specific growth rate (mcomputed with Equation 1) (Equation 5) [25]:

Ucell½ � ¼ V int

mð5Þ

RESULTS AND DISCUSSION

Dissolved uranium toxicity on cellular growth rate

Despite a decrease in the measured total [U] over time (seeSupplemental Data, Figures S1 and Figure S2), we chose toexpress U toxicity on C. reinhardtii as a function of the initial

1374 Environ Toxicol Chem 33, 2014 M. Lavoie et al.

total [U] or free [UO22þ]. This depletion in total [U] observed

at pH 5 and pH 7 was due to an important U accumulation by C.reinhardtii; the decrease in total [U] was mainly accounted forby U accumulation and adsorption by the algae (mass balanceagreed within� 27%). The apparent toxicity of free uranyl ionson the growth rate of C. reinhardtii cultivated in batch culturesdramatically decreased by approximately 60-fold as the pHdecreased from 7 (EC50¼ [1.8� 0.3]� 10�9M initial UO2

2þ)to 5 (EC50¼ [1.23� 0.02]� 10�7M initial UO2

2þ). Aninverse trend was observed when normalizing the toxicitywith the total initial [U] (the apparent EC50 decreased from[1.5� 0.5]� 10�5M to [3.35� 0.06]� 10�7M total initial [U]as the pH decreased from 7 to 5), because the proportion of freeions was much smaller at pH 7 than at pH 5 (Figure 1). ThisEC50 at pH 5, expressed as total initial U, is in the same order ofmagnitude as the EC50s obtained by Pradines et al. [26](endpoint, 72 h cell yield) and by Herlory et al. [27] (endpoint,5 h nonphotochemical quenching) using the same algal speciesat pH 5, but in a slightly different culture media. The EC50smeasured by Charles et al. [19] for the tropical alga Chlorellasp. growing at pH 7 varied between 2.4� 10�7 and3.1� 10�7M total U for hardness of 8 and 40mg CaCO3

L�1, respectively, that is, a range of hardness encompassing thehardness of the MHSM-F10 medium, which is approximately17mg CaCO3 L�1. We estimated that these EC50 values

correspond to 1.8� 10�10 and 2.1� 10�10M UO22þ based on

their medium composition. These EC50s are 1 order ofmagnitude smaller than that determined for C. reinhardtii at pH7 in the present study. In another study, Trenfield et al. [17]found an EC50 of 1.3� 10�8 UO2

2þ for the unicellularphotosynthetic protist Euglena gracilis at pH 6, which fallsbetween our own values at pH 5 and pH 7.

Note, however, that thermodynamic equilibrium calculationssuggest that at U concentrations greater than approximately3� 10�6M total U at pH 7, the present study’s exposure mediawould be oversaturated with respect to (UO2)3(PO4)2(s) andcould thus result in underestimating the EC50 at pH 7 (i.e., EC50at pH 7� [1.5� 0.5]� 10�5M total initial U). Indeed, ifphosphorus precipitation occurs, cellular growth may beinhibited by the decrease in available phosphorus at high Uconcentrations. Figure 1A shows that cell growth first becomesinhibited at 2� 10�5M, a concentration where the U to P ratio isapproximately 3:2, identical to the ratio of the oversaturatedmineral phase. At this point, not enough phosphorus would beavailable to sustain algal growth. As such, our observed EC50may not be due to U toxicity but more likely to the quasi-absenceof phosphorus for growth.

Our results nevertheless suggest that C. reinhardtii is moretolerant to the free uranyl ion at more acidic pH (compareFigures 1B and 1D). Within a narrower range of pH (6.5–5.7),

0.0 5.0e-10 1.0e-9 1.5e-9 2.0e-9

0.0

0.2

0.4

0.6

0.8

1.0

1.2 B

1e-7 1e-6 1e-5 1e-4

Rel

ativ

e gr

owth

rate

(µ/µ

0)

0.0

0.2

0.4

0.6

0.8

1.0

1.2 A

Initial [UO22+] (M)

0.0 5.0e-8 1.0e-7 1.5e-7 2.0e-7 2.5e-7

0.0

0.2

0.4

0.6

0.8

1.0

1.2 D

Total initial [U] (M)

0 2e-7 4e-7 6e-7

Rel

ativ

e gr

owth

rate

(µ/µ

0)

0.0

0.2

0.4

0.6

0.8

1.0

1.2 C

Figure 1. Uranium (U) toxicity on algal growth rate (m) measured in batch cultures relative to control cultures (m0; without added U). Toxicity is expressed asmeasured initial total (A, C) or free (B, D) U concentrations at pH 7 (A, B) or pH 5 (C, D). The curves represent nonlinear regressions (Equation 3) through the datapoints (r2> 0.85). Error bars on the y-axis are the standard deviations in growth rate of 3 replicate algal cultures. Error bars on the x-axis are the standard deviationsin measured total initial [U] concentrations (A, C) or calculated initial [UO2

2þ] (B, D).

Acidification increases cell sensitivity to uranium Environ Toxicol Chem 33, 2014 1375

Franklin et al. [15] also reported in another freshwater alga,Chlorella sp., a decreasing toxicity of U (based on totalmeasured U) as the pH decreased, with EC50s from1.8� 10�7M U to 3.3� 10�7M U, respectively. Whenexpressed as a function of UO2

2þ, the EC50 at pH 6.5(2.2� 10�8MUO2

2þ) is 5 times higher than that observed at pH5.7 (4.3� 10�9M), a trend that contradicts what we observed inthe present study. The causes of the apparent effect of pH on Utoxicity revealed in C. reinhardtii will be explored further in thenext sections.

pH affects algal sensitivity to cellular uranium

Surface boundmetal has been suggested in the past as a proxyfor a metal’s reactivity and thus for metal toxicity [28]. When theEC50 was expressed as the amount of adsorbed U per cell, weobserved that more U was bound to the cell surface at pH 7([8.2� 0.4]� 10�15mol of adsorbed U cell�1) than at pH 5([7.8� 1.9]� 10�16mol of adsorbed U cell�1) (Figure 2). Thegap between the 2 values could be explained by the changes inspeciation between the 2 experimental pHs and by protoncompetition, with the latter being perhaps more important thanthe former.

The internalized U per cell required to inhibit the cellulargrowth rate of C. reinhardtii by 50% varied as a function of pH(Figure 2) indicating that the cellular tolerance to intracellular Uis affected by the exposure pH. Surprisingly, intracellular Uappeared to be more toxic at pH 5 than at pH 7. Indeed, theinternalized U for which the cellular growth rate decreased by50% was reduced by more than 2 orders of magnitude as the pHdecreased from 7 (EC50¼ [7.1� 1.8]� 10�13mol of internal-ized U cell�1) to 5 (EC50¼ [4.0� 1.1]� 10�15mol ofinternalized U cell�1). The mechanisms whereby pH affectsthe sensitivity of cellular-bound U could be related to differentcells’ sensitivity to U as the external pH changes or to differencesin U subcellular distributions. The latter hypothesis is exploredfurther in the next sections.

Intracellular distribution—limitations

Subcellular fractionation protocols using mechanical ho-mogenization techniques and differential centrifugations includemany caveats, such as the breakage or aggregation of organelles,the leakage of soluble constituents from organelles, and overlapbetween centrifugal fractions [29–31]. Results must thus beinterpreted with caution, especially among studies in whichdifferent subcellular fractionation protocols and differentexperimental conditions have been used. For example, in thepresent study, large organelles (e.g., vacuoles, chloroplasts)might have been broken during the sonication step and theorganelles’ internal content may thus have leaked from the“insoluble” to the “soluble” fraction.

Intracellular distribution and uranium toxicity

A high proportion of the U accumulated by the algal cellswas associated with the insoluble fraction (>95% and >85%at pH 7 and pH 5, respectively, Figure 3). Such results are inaccordance with our previous study [18] in which we showedthat cellular U was located mainly in insoluble algal fragments(� 75% of the internalized U) of the alga C. reinhardtii after ashort (30min) exposure time to 2� 10�7M total U at pH 5. Thecytosolic U fraction was consistently higher at pH 5 than at pH 7(p< 0.01, Kruskall–Wallis; note that the U concentrations in thecytosolic fraction for the algae exposed to total initial [U] lowerthan 9� 10�7M at pH 7 were below the detection limit of theICP–AES; i.e., the percentage cytosolic U never exceeded 3% to

Intracellular or adsorbed U (mol/cell)

1e-17 1e-16 1e-15 1e-14 1e-13

Rel

ativ

e gr

owth

rate

(u/u

0 )

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Intracellular or adsorbed U (mol/cell)

1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12 1e-11

Rel

ativ

e gr

owth

rate

(µ/µ

0)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Intracellular U Adsorbed U

A

B

Figure 2. Uranium (U) toxicity on algal growth rate (m) measured in batchcultures relative to control cultures (m0; without added U) at pH 7 (A) or pH 5(B). Toxicity is expressed as a function of the amount of intracellular oradsorbed U (mol cell�1). The curves represent nonlinear regressions(Equation 3) through the data points (r2> 0.84). Error bars on the y-axis arethe standard deviations in relative growth rate of 3 replicate algal cultures.Error bars on the x-axis are the standard deviations in intracellular oradsorbed U.

Total initial [U] (M)

5-e16-e17-e1

Cyt

osol

ic u

rani

um (%

)

0

2

4

6

8

10

12

14

16

pH 5pH 7

Figure 3. Percentage of intracellular uranium (U) in the cytosol as a functionof total measured initial [U] for the cells exposed at pH 7 and pH 5. Error barson the y-axis are the standard deviations in cytosolic U percentage of 3replicate algal cultures. Error bars on the x-axis are the standard deviations intotal measured initial [U].

1376 Environ Toxicol Chem 33, 2014 M. Lavoie et al.

9%, depending on the quantity of biomass available). It thusfollows that the increased sensitivity to intracellular U at pH 5could be explained partly by the increase in cytosolic U bindingto algal soluble proteins or enzymes at pH 5. However, therelatively modest quantitative difference between U partitioningas a function of pH, albeit a more than 2 orders of magnitudechange in intracellular U sensitivity, also suggests that theinteracting influence of pH and accumulated-U affects otherunknownmetabolic processes leading toU toxicity. For example,our crude subcellular separation of U within the algae did notallow us to determine whether the increased tolerance tocellular U at pH 7 could be caused by a more efficientdetoxification of U in insoluble granules at this pH. Indeed, hardcations such as uranyl ions could be expected to bind tointracellular granular concretions such as polyphosphates in thecellular milieu [32]. Such changes in the U detoxificationefficiency as the pH varies could be another possibility explainingthemodulating effect of pHon the cellular sensitivity toU. To ourknowledge, the present study provides the first evidence thatexternal pH may influence internal handling of metals in algae.Therefore, additional studies are warranted.

Modeling long-term uranium uptake

The important gap between the modeled and the measuredlong-term U accumulation at both pHs, when plotted against thetotal initial [U] and free [UO2

2þ] (Figure 4), can be explained at

least in part by the depletion in total dissolved [U] over the long-term exposure, which would lead to a lower measured thanmodeled long-term U accumulation. For the lower and nontoxicinitial U concentrations at pH 7 and pH 5, the modeled steady-state intracellular U was overestimated compared to the long-term U accumulation measured after 4 d of exposure (Figure 4).However, the difference between the modeled and the measuredvalues narrowed for higher nontoxic (at pH 7) and toxic initial[U]. The better agreement between modeled and measuredvalues at pH 7 for high but nontoxic initial [U] compared to lowinitial [U] suggests that the algae regulated the net Uaccumulation within the long-term U exposure because therelative depletion of the dissolved total [U] was not smaller athigh relative to low U exposure concentrations (SupplementalData, Figure S1). Hence, the total U depletion would not favor ahigher long-term U accumulation at higher [U]. We hypothesizethat the cells either upregulate U transport in response tohigher U concentration or, alternatively, down regulate Uaccumulation in response to lower U concentration.

When we compared U accumulation at a low, nontoxic totalinitial [U] at both pHs (i.e., 10�7M), U accumulation was similareven though the initial [UO2

2þ] was much greater (474 times)at pH 5 (6.5� 10�8M UO2

2þ) than at pH 7 (1.37� 10�10MUO2

2þ) (Figure 4). This indicates that the U net accumulationrate is much greater at pH 7 than at pH 5 for the same initial[UO2

2þ] and explains in part the lower toxicity of the free ion at

Total initial [U] (M)

10-7 10-6

Intra

cellu

lar U

(mol

/cel

l)

10-17

10-16

10-15

10-14

10-13

10-7 10-6 10-5

10-16

10-15

10-14

10-13

10-12

10-11

MeasuredModeled

A

B

[UO22+] (M)

10-7 10-6

10-10 10-9

C

D

Figure 4. Long-term (4 d) uranium (U) accumulation as a function of the total initial [U] (A, B) and free [UO22þ] (C, D) at pH 7 (A, C) and pH 5 (B, D). The

modeled solid curves represent the steady-state intracellular U content predicted by using Michaëlis–Menten kinetic parameters derived from short-termexperiments in Fortin et al. [9] and Equations 3 and 5 (A, B) and Equations 4 and 5 (C, D).

Acidification increases cell sensitivity to uranium Environ Toxicol Chem 33, 2014 1377

more acidic pH as shown in Figure 1. Simple competitionbetween protons and uranyl ions may partly explain the lower Uaccumulation (as a function of [UO2

2þ]) at pH 5 than at pH 7.However, Fortin et al. [9] have shown in short-term U exposureexperiments that protons noncompetitively inhibit U uptake inC. reinhardtii; with Jmax (maximum uptake flux; defined as theproduct of the metal internalization rate constant kint and of thenumber of transport sites) values being more than 2 orders ofmagnitude larger at pH 7 compared with those at pH 5. Such aresponse of algal physiology to exposure pH may explain themuch higher U accumulation at pH 7 than at pH 5 whenconsidering only the free ion concentrations.

The precise mechanisms behind this regulation of Uaccumulation and toxicity thus remain unknown but reveal ahigher level of complexity than the BLM usually assumes.Within our experiments, thermodynamic calculations usingMINEQL show that the free Mn, Co, and Zn concentrationsdecrease by less than 10% as the pH increases from 5 to 7. Thisdoes not support an upregulation (rise in Jmax or Vmax) ofessential metal transport system (by which UO2

2þ could beinadvertently taken up) in response to decreasing availability ofessential metals. Moreover, the labile Cu and Fe concentrationstend to increase as the pH increases, which would not intuitivelyinduce an up-regulation of Cu and Fe metal transport (and ofUO2

2þ) as the pH increases.Note that the relative standard error (estimated from error

propagation calculations) of the mean modeled steady-stateintracellular U concentration was always below 30% and wasthus not sensitive to the uncertainty of all input parameters (K,Vmax, and m). Hence, our conclusions are unaffected by theuncertainty of the modeled long-term U uptake.

CONCLUSION

Our results show that pH affects U toxicity in C. reinhardtiinot only by modulating net U accumulation, but also bychanging algal sensitivity to internalized U. We found someevidence that such a pH effect on cellular sensitivity to U toxicitymight be explained in part by the subcellular distribution in thesensitive cytosol fractions. Modeling the long-termintracellular U concentrations suggests an active physiologicalregulation of U uptake as a function of the U exposureconcentration within the 4-d exposure. The present study bringsinsights to U chronic toxicity in C. reinhardtii at varyingexposure pH and provides useful data for developing a BLMfor U and for deriving adequate water quality guidelines.

SUPPLEMENTAL DATA

Figures S1 and S2. (25 KB PDF).

Acknowledgement—We acknowledge the technical assistance provided byM. Morello, V. Camilleri, and G. Grasset. The editorial assistance providedby B. Barst was appreciated. The present study was part of theRadioprotection de l’environnement à l’homme research program fundedby the Institute for Radioprotection and Nuclear Safety. M. Lavoie held ascholarship from the Fonds de recherche du Qu�ebec, Nature ettechnologies. C. Fortin is supported by the Canada Research Chair program.

REFERENCES

1. Campbell PGC. 1995. Interactions between trace metals and aquaticorganisms: A critique of the free-ion activity model. In Tessier A, TurnerDR, eds,Metal Speciation and Bioavailability in Aquatic Systems. JohnWiley & Sons, New York, NY, USA, pp 45–102.

2. Paquin PR, Gorsuch JW, Apte S, Batley GE, Bowles KC, CampbellPGC, Delos CG, Di Toro DM, Dwyer RL, Galvez F, Gensemer RW,Goss GG, Hogstrand C, Janssen CR,McGeer JC, Naddy RB, Playle RC,Santore RC, Schneider U, Stubblefield WA, Wood CM, Wu KB. 2002.The biotic ligand model: A historical overview. Comp Biochem PhysiolC 133:3–35.

3. Campbell PGC, Err�ecalde O, Fortin C, Hiriart-Baer VP, Vigneault B.2002. Metal bioavailability to phytoplankton: Applicability of the bioticligand model. Comp Biochem Physiol C 133:189–206.

4. Di Toro DM, Allen HE, Bergman HL, Meyer JS, Paquin PR, SantoreRC. 2001. Biotic ligand model of the acute toxicity of metals. 1.Technical basis. Environ Toxicol Chem 20:2383–2396.

5. Slaveykova VI, Wilkinson KJ. 2005. Predicting the bioavailability ofmetals and metal complexes: Critical review of the biotic ligand model.Environ Chem 2:9–24.

6. Campbell PGC, Fortin C. 2013. Biotic ligand model. In F�erard J-F,Blaise C, eds, Encyclopedia of Aquatic Ecotoxicology. Springer,Netherlands, Dordrecht, pp 237–246.

7. Duval JF. 2013. Dynamics of metal uptake by charged biointerphases:Bioavailability and bulk depletion. Physical Chemistry ChemicalPhysics 15:7873–7888.

8. François L, Fortin C, Campbell PGC. 2007. pHmodulates transport ratesof manganese and cadmium in the green alga Chlamydomonasreinhardtii through non-competitive interactions: Implications for analgal BLM. Aquat Toxicol 84:123–132.

9. Fortin C, Denison FH, Garnier-Laplace J. 2007. Metal-phytoplanktoninteractions: Modeling the effect of competing ions (Hþ, Ca2þ, andMg2þ) on uranium uptake. Environ Toxicol Chem 26:242–248.

10. Sunda WG, Huntsman SA. 1986. Relationships among growth rate,cellular manganese concentrations and manganese transport kinetics inestuarine and oceanic species of the diatom Thalassiosira. J Phycol22:259–270.

11. Sunda WG, Huntsman SA. 1992. Feedback interactions between zincand phytoplankton in seawater. Limnol Oceanogr 37:25–40.

12. Chen Z, Zhu L, Wilkinson KJ. 2010. Validation of the biotic ligandmodel in metal mixtures: Bioaccumulation of lead and copper. EnvironSci Technol 44:3580–3586.

13. Goulet RR, Fortin C, SpryDJ. 2011. Uranium. InWoodCM, Farrell AP,Brauner CJ, eds, Fish Physiology: Homeostasis and Toxicology of Non-EssentialMetals, Vol. 31B-Fish Physiology. Academic, NewYork, NY,USA, pp 391–428.

14. Bleise A, Danesi PR, BurkartW. 2003. Properties, use and health effectsof depleted uranium (DU): A general overview. J Environ Radioact64:93–112.

15. Franklin NM, Stauber JL, Markich SJ, Lim RP. 2000. pH-dependenttoxicity of copper and uranium to a tropical freshwater alga (Chlorellasp.). Aquat Toxicol 48:275–289.

16. Trenfield MA, Ng JC, Noller BN, Markich SJ, van Dam RA. 2011.Dissolved organic carbon reduces uranium bioavailability and toxicity.2. Uranium[VI] speciation and toxicity to three tropical freshwaterorganisms. Environ Sci Technol 45:3082–3089.

17. Trenfield MA, Ng JC, Noller B, Markich SJ, van Dam RA. 2012.Dissolved organic carbon reduces uranium toxicity to the unicellulareukaryote Euglena gracilis. Ecotoxicology 21:1013–1023.

18. Fortin C, Dutel L, Garnier-Laplace J. 2004. Uranium complexation anduptake by a green alga in relation to chemical speciation: The importanceof the free uranyl ion. Environ Toxicol Chem 23:974–981.

19. Charles AL, Markich SJ, Stauber JL, De Filippis LF. 2002. The effect ofwater hardness on the toxicity of uranium to a tropical freshwater algaChlorella sp. Aquat Toxicol 60:61–73.

20. Sheppard SC, Sheppard MI, Gallerand MO, Sanipelli B. 2005.Derivation of ecotoxicity thresholds for uranium. J Environ Radioactiv79:55–83.

21. Lavoie M, Le Faucheur S, Fortin C, Campbell PGC. 2009. Cadmiumdetoxification strategies in two phytoplankton species: Metal binding bynewly synthesized thiolated peptides and metal sequestration ingranules. Aquat Toxicol 92:65–75.

22. Twiss MR, Err�ecalde O, Fortin C, Campbell PGC, Jumarie C, DenizeauF, Berkelaar E, Hale B, Van Rees K. 2001. Coupling the use of computerchemical speciation models and culture techniques in laboratoryinvestigations of trace metal toxicity. Chemical Speciation andBioavailability 13:9–24.

23. Schecher WD, McAvoy DC. 1992. MINEQL þ : A softwareenvironment for chemical equilibrium modeling. Comput EnvironUrban Syst 16:65–76.

24. Denison FH, Garnier-Laplace J. 2005. The effects of database parameteruncertainty on uranium(VI) equilibrium calculations. Geochim Cosmo-chim Acta 69:2183–2191.

1378 Environ Toxicol Chem 33, 2014 M. Lavoie et al.

25. Sunda WG, Huntsman SA. 1998. Processes regulating cellular metalaccumulation and physiological effects: Phytoplankton as modelsystems. Sci Total Environ 219:165–181.

26. Pradines C, Wiktor V, Camilleri V, Gilbin R. 2005. Development ofbiochemical methods to estimate the subcellular impact of uraniumexposure on Chlamydomonas reinhardtii. Radioprotection 40:S163–S168.

27. Herlory O, Bonzom J-M, Gilbin R. 2013. Sensitivity evaluation of thegreen alga Chlamydomonas reinhardtii to uranium by pulseamplitude modulated (PAM) fluorometry. Aquat Toxicol 140–141:288–294.

28. De Schamphelaere KAC, Stauber JL,WildeKL,Markich SJ, Brown PL,Franklin NM, Creighton NM, Janssen CR. 2005. Toward a biotic ligandmodel for freshwater green algae: Surface-bound and internal copper are

better predictors of toxicity than free Cu2þ-ion activity when pH isvaried. Environ Sci Technol 39:2067–2072.

29. DeDuveC. 1975. Exploring cells with a centrifuge. Science 189:186–194.30. Wallace WG, Lee B-G, Luoma SN. 2003. Subcellular compartmentali-

zation of Cd and Zn in two bivalves. I. Significance of metal-sensitivefractions (MSF) and biological detoxified metal (BDM).Mar Ecol ProgSer 249:183–197.

31. Lavoie M, Bernier J, Fortin C, Campbell PGC. 2009. Cell homogeniza-tion and subcellular fractionation in two phytoplanktonic algae:Implications for the assessment of metal subcellular distributions.Limnol Oceanogr-Meth 7:277–286.

32. Marigómez I, Soto M, Cajaraville MP, Angulo E, Giamberini L. 2002.Cellular and subcellular distribution of metals in molluscs.Microsc ReshTechniq 56:358–392.

Acidification increases cell sensitivity to uranium Environ Toxicol Chem 33, 2014 1379