CSIRO PUBLISHINGResearch Paper

H. Kola et al., Environ. Chem. 2004, 1, 172–179. doi:10.1071/EN04061 www.publish.csiro.au/journals/env

Cadmium Adsorption by Chlamydomonas reinhardtiiand its Interaction with the Cell Wall Proteins

Heliana Kola,A,B Luis M. Laglera,A Nalini Parthasarathy,A and Kevin J. WilkinsonA,C

A Analytical and Biophysical Environmental Chemistry (CABE), University of Geneva,1211 Geneva 4, Switzerland.

B Environmental and Analytical Chemistry, Battelle Geneva Research Center,1227 Carouge-Geneva, Switzerland.

C Corresponding author. Email: kevin.wilkinson@cabe.unige.ch

Environmental Context. In natural waters, trace metals levels are largely controlled by microbiology;organisms take up, metabolize, store, and detoxify the metals. However, aquatic organisms may regulatetheir own uptake via dynamic processes that result in a system that is far from equilibrium. By examiningthe model title alga with a battery of techniques, a more realistic assessment of metal uptake and metalregulatory processes could be gained.

Abstract. Cadmium adsorption by a wild type strain of Chlamydomonas reinhardtii and a cell wall-less mutantwas quantified as a function of Cd speciation in a well-defined aqueous medium. For both strains, Cd adsorption tothe cell surface was not predicted by a single-site (Langmuirian) model. Indeed, no saturation of the cell wall wasobserved, even for Cd concentrations in excess of 5 × 10−3 M.A continual production of Cd binding sites appeared tobe responsible for the observed increase of Cd adsorption with time. SDS-page separations and measurements of theprotein content of algal supernatants demonstrated that organic matter was released by the algae, both in the presenceand absence of Cd. Both the nature (e.g. polysaccharides, proteins) and the quantity of exudate production wasinfluenced by the physicochemistry of the external medium. Measurements using the permeation liquid membrane(PLM) and anodic stripping voltammetry (ASV) demonstrated that dissolved cadmium was rapidly complexed bythe organic exudates produced by the algae.

Keywords. algae — bioavailability — Cd — proteins — speciation

Manuscript received: 28 July 2004.Final version: 1 November 2004.

Introduction

In natural waters, biological systems play an important rolein regulating trace metal concentrations and speciation (e.g.lakes[1] and oceans[2]) by decreasing concentrations throughadsorption and internalization, and by modifying chemicalspeciation through release of extracellular organic matter.In order to better understand the mechanisms by whichmetals are sequestered, metabolized, stored, and detoxifiedby aquatic organisms, model unicellular systems have beenextensively developed. C. reinhardtii is a green alga that isoften used to study the deleterious effects of toxic trace met-als through measurements of photosynthesis,[3,4] growth,[5–7]nutrient uptake inhibition,[8] ultrastructural changes,[9] etc.Furthermore, the recent availability of the complete genomefor C. reinhardtii[10] has given an added impetus to the useof this organism for metal bioavailability studies, for fun-damental investigations of trace metal effects, and for thedevelopment of eukaryotic biosensors.[11–15]

Nearly all descriptions of metal uptake and adsorption areequilibrium-based models that exclude biological parame-ters such as the production of organic exudates, metal efflux,or the synthesis of uptake sites.[16,17] For example, metaladsorption to the surface of microorganisms is most oftendescribed by a single-site model (e.g. Langmuir isotherm)with[18,19] or without[20–22] a charge correction. Nonetheless,at least two important conditions are required for a Lang-muir isotherm to apply: (a) equilibrium is attained betweenthe adsorbed and dissolved metal and (b) all surface sitesare equally active. In reality, for chemically heterogeneousand multiply charged surfaces, the tendency to form surfacecomplexes decreases with an increased metal loading of thesurface due to both chemical and Coulombic effects.[23,24]The equilibrium assumption implies that the concentrationof the free aquo metal ion in solution will be central to pre-dicting metal adsorption. It is thus important to quantify therole and nature of the metal-binding ligands that are produced

© CSIRO 2004 172 1448-2517/04/030172

Cadmium Adsorption by Chlamydomonas reinhardtii

and excreted by aquatic organisms. For example, Xue andSigg[25] have demonstrated that the binding of Cu(ii) to exu-dates produced by C. reinhardtii had a more significant effecton Cu speciation than did the adsorption to algal surfaces. Inthat case, the complexing capacity of the algal filtrates var-ied with cell number, cell storage, the duration of the Cutreatment, and the culture conditions. In another example, aten-fold increase in exudate production was observed duringthe stationary as compared to the logarithmic growth phaseof the green alga Fritschiella tuberose lying.[26] The natureof the exudates depends strongly on the biological speciesand the exposure conditions. For Scenedesmus quadricauda,Chlorella kessleri, and Raphidocelis subcapitata, exposure toCd increased the protein-to-carbohydrate ratio of the exudateswhereas exposure to Cl had an opposite effect.[27] Finally,in the case of C. reinhardtii, algae are known to secretehydrophilic organic signal molecules that affect quorum sens-ing regulation in naturally encountered bacteria,[28] even inthe absence of an apparent external stress.

Given that the organism can modify its own metal uptake,the objective of the present study was to precisely describeCd adsorption to C. reinhardtii by taking into account themost important biological, physicochemical, and environ-mental parameters. Several techniques were combined inorder to quantitatively determine free Cd or Cd bound eitherto extracellular ligands in the bulk solution or to the cellsurface.

Experimental

Cell Cultures and Isolation

Two strains of C. reinhardtii were selected for the present study: thewild-type (WT) 2137 obtained from M. Goldschmidt-Clermont (Uni-versity of Geneva) and a (cell wall-less) mutant (CW-2) obtained fromS. Waffenschmidt (University of Cologne) that continually releaseslayers of its cell wall[29] into solution. C. reinhardtii is a goodmodel species due to its well known culturing techniques in simpleinorganic culture media, its fast growth,[30] and a complete knowl-edge of its genetic makeup. In this study, algae were grown in amodified tris(hydroxymethyl)aminomethane/acetate/phosphate (TAP)medium,[31] in which the concentrations of components were reducedfour-fold in order to avoid a cellular aggregation that can occur in thehigh ionic strength of the TAP growth medium. Algal cultures wereplaced in an incubator (Multitron, Infors) at 20◦C under a 12-h dark/12-h light regime (50 µmol photons m−2 s−1) using a rotary agitation of100 rpm. Cells attained a stationary growth phase of 3 × 106 cells mL−1

after two days of incubation. In these experiments, cells were collectedin their late exponential growth phase (∼2 × 106 cells mL−1) using a 7-min centrifugation at 2700g. They were subsequently washed twice in asolution containing 10−2 M morpholinopropanesulphonic acid (MOPS;Fluka) at pH 7. The ionic strength of the wash solution was adjusted to10−2 M by adding NaNO3.

Experimental Conditions

Cd adsorption experiments were performed in the same defined mediumused for washing the cells with the addition of known quantities ofcadmium and/or ligands. For [Cd2+] <5 × 10−7 M, a total cadmiumconcentration of 10−6 M was buffered using nitrilotriacetic acid (NTA),citric acid, or diglycolic acid by maintaining the ligand-to-cadmiumratio greater than 5. Cadmium speciation calculations were performedusing the VISMINTEQ+ (ver. 2.0a) and MINEQL+ (ver. 3.01b) chemi-cal equilibrium programs. Stability constants were updated using theNIST stability constant database[32] and ionic strength correctionswere performed using the Davies equation. In the absence of added

ligand, [Cd2+] represented ≥98% of the total Cd in solution. Allaqueous solutions were prepared with MilliQ water (R > 18.2 M� cm;TOC < 2 µg L−1). pH adjustments were made with ultrapure HNO3(Baker Instra-Analyzed Reagents).

Cd Adsorption

Adsorbed Cd was distinguished from cellular (internalized) Cd usingan EDTA extraction[33] that followed a 10 min exposure to different[Cd2+] levels. In order to reduce Cd depletion in the bulk solution, theexperiments were conducted using relatively low cell densities (7 × 105–106 cells mL−1 in 50 mL, corresponding to ca. 1–2 cm2 mL−1). Algalsurface areas and average diameters were determined with a CoulterMultisizer II particle counter with a 50 µm aperture. Centrifugation(3 min, 2600g) was employed to separate the algae from the experi-mental solutions. A small volume (5 mL) of supernatant was filtered(1.2 µm) and analyzed for dissolved Cd following acidification (0.4%HNO3). The algal pellet was washed two or three times in MOPS solu-tion (30 mL, 10−2 M, pH 7.0), then resuspended for 10 min in an EDTAsolution (20 mL 10−3 M EDTA, in 10−2 M MOPS, pH 7.0). Followinga final centrifugation, the supernatant was acidified (1%) and used forthe determination of surface-bound (adsorbed) Cd.[33,34] Under theseconditions, Cd efflux was negligible such that the EDTA extractible Cdcould be attributed uniquely to the adsorbed fraction.[34] Adsorbed Cdwas also analyzed as a function of exposure time. In that case, dupli-cate 10 mL aliquots were sampled and centrifuged (2600g, 3 min) 5, 10,15, 20, and 25 min after the addition of the algae to the experimentalmedium. As previously, algal pellets were washed and centrifuged twicebefore the extraction of the adsorbed Cd using the 10 min extraction inEDTA/MOPS (10−3 M/10−2 M).

Biological and analytical variability were evaluated for each experi-mental condition by processing two runs and by repeating the experimentat a later date using a second culture of C. reinhardtii. For all experi-ments, the mass balance (sum of the dissolved, adsorbed, and internal-ized Cd fractions) was relatively constant over time. Furthermore, lossesof Cd from the experimental medium did not exceed 10% for the WTstrain and 15% for the CW-2 strains (data not shown). Adsorbed anddissolved Cd concentrations were analyzed by flame atomic absorptionspectrometry (AAS) or inductively coupled plasma mass spectrometry(ICP-MS).

Quantification and Characterization of Algal Exudates

Bradford Assay

In order to examine whether Cd influenced the release of cellularproteins, two 10 mL aliquots of concentrated algal cultures were addedto 70 mL of experimental media (10−2 M MOPS) with and without10−7 M Cd. Two 6-mL samples (containing 10–15 cm2 mL−1 algae)were removed and centrifuged (2600g, 3 min) following Cd accumu-lation times of 10 or 60 min. The same procedure was employed foralgae washed with 10−2 M N-hydroxyethylpiperazine-N′-2 ethanesul-fonic acid (HEPES) or 10−2 M KH2PO4/K2HPO4. Protein contentwas measured in 100 µL of algal filtrate (1.2 µm) using the Brad-ford microassay[35] (detection limit 10 mg L−1) in which absorbance(595 nm) was measured after a 20 min reaction with 1 mL of Bradfordreagent (Sigma). Standards containing 0–80 µg bovine serum albuminin a final volume of 100 µL (adjusted with Milli-Q water) were preparedin duplicate from a 1 g L−1 albumin stock solution (Sigma).

SDS-page

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis SDS-page was used to qualitatively analyze extracellular proteins. 6 mL ofalgal supernatant were concentrated to 60 µL using a centricon filtrationdevice (10 kDa cutoff).The concentrate was transferred into microtubes,boiled for 5 min with 30 µL of 10% β-mercaptoethanol, and analyzedon a 7.5% gel.[36] Gels were stained by shaking them for 5 h in an aque-ous solution of Coomassie brilliant blue R250 stain containing 50%methanol and 10% acetic acid. Following staining, the gels were thor-oughly washed with Milli-Q water. Purified proteins of the W6 layer

173

H. Kola et al.

of intact walls isolated from C. reinhardtii bald-2 cells[37] (i.e. mutantswithout flagella) were obtained from S. Waffenschmidt.

Total Organic Carbon and Nitrogen Analyses

TOC and N content were determined in algal supernatants iso-lated as above except that the pH buffer was replaced by a carbon-freebuffer (10−2 M KH2PO4/K2HPO4, pH 7.0). Following resuspensionof the cells in the experimental media, the supernatants were isolated(2600g, 3 min) then acidified with HCl (final concentration 0.1% v/v)and degassed with ultrapure air (CO2-free) before measurements onShimadzu analyzers (Model TOC-5000 or TOC-VCPH/CPN).

Chemical Speciation Measurements

Preparation of Solutions

Algae (1.5–2.5 cm2 mL−1) were exposed for 10–60 min to an exper-imental medium containing or in the absence of 10−7 M Cd thenseparated by centrifugation. The supernatant was divided into twofractions for measurements of [Cd2+] by the PLM technique or ofelectrochemically labile Cd by ASV. For exposures without Cd, speci-ation measurements were performed on algal supernatants spiked with50 nM Cd.

PLM Measurements

The diffusion cell used for PLM measurements has been describedpreviously.[38] A polypropylene membrane (Celgard 2500, Hoechst;porosity 45%, pore size 0.04 µm, thickness 25 µm) was impregnatedwith a 1/1 (v/v) toluene/phenyl hexane (Fluka) mixture containing1,10-didecyl [18]crown-6 ether (10−1 M, Kryptofix22DD) and lauricacid (10−1 M, Fluka). Following a rinse with Milli-Q water to removeexcess solvent, a 7.2 cm2 portion of membrane was placed betweentwo Plexiglas half cells. The source solution containing the Cd (80 mL,pH 7.0) and the strip solution (80 mL) containing 5 × 10−4 M trans-1,2 diaminocyclohexane tetraacetic acid (CDTA, Fluka; pH 6.4) wereplaced in two separated compartments of the cell. Solutions were stirredcontinuously at 480 rpm. Aliquots of the solutions were sampled peri-odically for analysis by ICP-MS. In the experimental setup that wasemployed here, metal transport across PLM was diffusion-controlledso that the measured fluxes across the PLM membrane were directlyproportional to the concentration of [Cd2+] in the source solution.[39]

ASV Measurements

A flow-through electrochemical cell, in which both the workingand auxiliary electrodes were in contact with the inner channel ofthe cell, was used for the ASV measurements. The working electrodewas a microfabricated square array of eight Ir-based Hg-plated micro-electrodes. Mercury was deposited on the Ir surface by maintaininga potential of −400 mV while circulating a deoxygenated solutionof Hg acetate (5 mM) in perchloric acid (50 mM). Electrode radii(4.6 ± 0.1 µm) were determined from the total charge accumulatedduring Hg deposition. The auxiliary electrode was an Ir microsurfaceplaced on the same board as the working electrode. The reference elec-trode was a homemade miniaturized Ag|AgCl|KCl (3 M) electrode witha NaNO3 (0.1 M) bridge. A more detailed description of the elec-trochemical system and its operating principles has been publishedpreviously.[40,41] Electrochemical measurements were performed usinga deposition potential of −1350 mV and deposition times from 2 to10 min depending on the Cd concentration (a linear dependence of peakheights on deposition times was observed). During the stripping step,the potential was scanned from −1120 to +100 mV, using a square-wave modulation (frequency 150 Hz, scan rate 2 mV, pulse amplitude50 mV). Well-defined peaks were found around −600 mV and no con-tamination from Zn, Pb, or Cu was detected in any sample. All sampleswere degassed with N2 before being pumped through the cell. A mea-sured detection limit of 0.8 nM was well below the lowest concentrationfound in the samples (2.8 nM). Coefficients of variation on replicatemeasurements (n = 3) were always lower than 5%.

0

1

2

3

Time (min)0 5 10 15 20 25

[Cd]

diss

(�

10�

7 M

)[C

d]ad

s (�

10�

11 m

olcm

�2)

1

2

3

4

(a)

(b)

Fig. 1. (a) Adsorbed (i.e. EDTA-extractable) Cd and (b) dissolved Cdas a function of time for the wild type (•) and cell wall-less strains (◦) ofC. reinhardtii. Initial [Cd] was 3–3.5 × 10−7 M, pH 7.0, 10−2 M MOPS,I = 10−2 M.

Results and Discussion

Cd Adsorption by C. reinhardtii WT and CW-2 Strains

Metal adsorption to algal cells is generally a rapid processwith equilibrium being attained in the first few minutes ofexposure.[33,42] Contrary to this expectation, adsorbed Cdfor both the wild type and cell wall-less strains of C. rein-hardtii did not attain constant values, but instead increasedwith time for all studied Cd concentrations (e.g. Fig. 1a). Arapid depletion of dissolved [Cd] that depended on the ini-tial [Cd] and cell densities was generally observed in thefirst few minutes of the accumulation experiments (Fig. 1b).The observed initial decrease in dissolved Cd was attributedin large part to Cd adsorption, while 5–10% could beaccounted for by Cd internalization. The further increase inEDTA-extractable Cd with time for the WT strain (Fig. 1a),1.1 × 10−12 mol cm−2 min−1, was nearly equal to the Cdinternalization flux, while for the CW-2 strain, this rate wasfive times smaller. For the wild type strain, the rate of Cdadsorption, obtained from the slope of the temporal increaseof adsorbed Cd, was plotted against [Cd2+] in the bulk media.A slope of 0.63 (R2 = 0.94) best described the logarithmicrepresentation of the data (Fig. 2).

The previous data strongly suggested that equilibrium wasnot attained between the cell surface and the bulk solution.It is possible to estimate times to reach chemical equi-librium based upon the reasonable assumption that waterloss from the inner hydration sphere of the cation is therate-limiting step (Eigen mechanism).[43] Based upon thismechanism, equilibrium between 2 × 10−9 mol cm−2 of Cdbinding sites (maximal observed value, see Fig. 3b) and

174

Cadmium Adsorption by Chlamydomonas reinhardtii

2 × 10−7 M of dissolved Cd should be achieved in less than1 min. Similarly, mass transport considerations[44] suggesta rapid equilibration, occurring in less than 2 min underthese conditions. These calculations contrast with the resultsobserved in Fig. 1a where equilibrium was not achieved evenafter 25 min. Although the formation rate of the surface com-plex could be controlled the dissociation rate of a competingH+, Ca2+, Mg2+, or a transition metal complex rather thanthe Cd water-loss kinetics,[45] this is unlikely given that, inthe case of a limiting adsorption rate, plots of adsorbed Cd asa function of [Cd2+] should show progressively decreasingslopes for longer exposure times as equilibrium is increas-ingly approached. This does not appear to be the case since,for an exposure of the WT strain to [Cd2+] ranging from 10−9

to 10−5 M, slopes of the (log) adsorbed Cd as a function of(log) [Cd2+] were relatively constant between 0.56 and 0.59

log[Cd2�] (M)

�9 �8 �7 �6 �5 �4 �3

log

rate

of C

d ad

sorp

tion

(mol

cm�

2 m

in�

1 )

�14

�13

�12

�11

�10

Fig. 2. Logarithmic representation of the rate of Cd adsorption by thewild type C. reinhardtii as a function of [Cd2+] in the absence (•) andpresence of the ligands: citric acid (�), diglycolic acid (♦), and NTA(�).The line represents a linear regression of the log–log representation:slope 0.63, R2 = 0.94. Error bars correspond to standard deviations whenlarger than the symbol size (n = 2–4).

log[Cd2�] (M)

�9 �8 �7 �6 �5

log[

Cd]

ads

(mol

cm�

2 )

�13

�12

�11

�10

�9(a)

log[Cd2�] (M)

�10 �9 �8 �7 �6 �5 �4 �3

(b)

Fig. 3. (a) Cd adsorption by the wild type C. reinhardtii as a function of [Cd2+] for a 5 (◦), 20 (�),and 60 (�) min equilibration times. The unbroken (5-min exposure, slope 0.56, R2 = 0.98), dotted (20-minexposure, slope 0.59, R2 = 0.98), and dashed (60-min exposure, slope 0.59, R2 = 0.96) lines represent thelinear regressions of (log) adsorbed Cd as a function of (log) [Cd2+] for the wild type strain. Error barsrepresent standard deviations when larger than the symbol size (n = 2). (b) Logarithmic representation ofadsorbed Cd by the wild type (filled symbols) and cell wall-less (open symbols) strains of C. reinhardtii asa function of [Cd2+] in the absence (circles) and presence of citric acid (squares) or NTA (triangles). Cellswere exposed to Cd for 10 min. The dashed line represents the linear regression (slope 0.5, R2 = 0.98) of(log) adsorbed Cd as a function of (log) [Cd2+] for the wild type strain (slope 0.55, R2 = 0.99 for the range of[Cd2+] used in Fig. 3a). Error bars represent standard deviations when larger than the symbol size (n = 2–6).

as the exposure time increased from 5 to 60 min (Fig. 3a).Furthermore, for a constant (10 min) exposure of both theWT and CW-2 strains, adsorbed Cd increased with increas-ing [Cd2+] in the bulk media from 10−10 to 10−3 M withno clear sign of a saturation plateau (Fig. 3b), even though aslight reduction in the slope may have been observed for theCW-2 strain for [Cd2+] > 10−5 M. For similar cell-surfaceareas, the WT strain adsorbed significantly more Cd (gen-erally ten-fold) than the CW-2 strain. It was impossible toperform experiments at higher Cd concentrations due to apotential precipitation of Cd above 10−3 M and the sensitiv-ity of the CW-2 strain to Cd; indeed, cell numbers decreaseddrastically for exposures to greater than 10−4 M Cd2+.

Adsorption cannot be accounted for by a Langmuir (orsimilar one-site) model. When plotting metal adsorbed tothe microorganisms against bulk solution concentrations,the observation of less than unity slopes is often attributedto the presence of chemically heterogeneous complexingsites or to a multiply charged surface (or both explanationssimultaneously).[18,19,46] In this case, neither classical normodified adsorption isotherms could reasonably be used todescribe the data due to the observation of a constant slope inFig. 3a and the absence of a saturation plateau in Fig. 3b. Onthe other hand, the mass transport and Eigen mechanism cal-culations suggested that adsorptive reaction kinetics were notresponsible for the observed absence of equilibrium betweenthe cell surface and the bulk solution.

Another plausible explanation for the observed increase ofCd adsorption with time would correspond to a production ofthe Cd binding sites. In this respect, the cell wall of wild typeC. reinhardtii is a seven-layered structure, with three con-spicuous layers W2, W4, and W6 (separated by spacers W3and W5) composed mainly of fibres and glycoproteins.[47]Goodenough and Heuser[37] have demonstrated that C. rein-hardtii cells can synthesize a primitive outer layer of their cell

175

H. Kola et al.

walls in less than 1 h following the addition of autolysin, a cellwall disrupting agent.[47] Indeed, if rapid cell wall renewal isresponsible for data showing an increasing adsorption of Cdwith time, then the rate of Cd adsorption for the WT strainshould be significantly higher than that of the CW-2 strain.Asdemonstrated above (Fig. 1a), for similar cell surface areas,the rate of Cd adsorption to the WT cells was five-fold higherthan that observed for the CW-2 strain.

Exudate Release

In the exposure media containing 10−7 M Cd in 10−2 MKH2PO4/K2HPO4, algal filtrates showed increasing con-centrations of total organic carbon and relatively constantnitrogen concentrations over time (Fig. 4a). The high C/Nratio and the apparent stability of the N content in the exu-dates suggested that the exudates in this buffer were primarilypolysaccharides. On the other hand, protein contents of thealgal filtrates increased as a function of time when the phos-phate buffer was replaced with 10−2 M MOPS (Fig. 4).Similar trends were observed for proteins, TOC, and N inthe filtrates of exposure media with or without added Cd,suggesting that exudate production was not in response to aCd stress but rather due to a constitutive process that occurredunder the studied conditions. The nature of the pH buffer hadan important influence on both the nature and quantity of exu-date production (Fig. 4b), with a higher protein productionobserved in the Good buffers (MOPS, HEPES) as opposed

Time (min)0 10 20 30 40 50 60

Pro

tein

con

tent

(�g

cm�

2 )

0.2

0.4

0.6

0.0

Pro

tein

s, c

arbo

n an

dni

trog

en c

onte

nt (

�g

cm�

2 )

0.0

0.1

0.2

0.3

0.4

0.5 (a)

(b)

Fig. 4. (a) Total protein content (triangles), total organic carbon (cir-cles), and total nitrogen (squares) released into the bulk media in thepresence (open symbols) or absence (filled symbols) of Cd. (b) Totalprotein content released into the bulk media in the presence of differ-ent buffer systems without Cd: 10−2 M KH2PO4/K2HPO4 (•), 10−2 MHEPES (◦), and 10−2 M MOPS (�). In all cases, the pH was adjustedto 7.0 at an ionic strength of 10−2 M.

to the phosphate buffer. SDS page gels of the extracellu-lar proteins of the wild type strain demonstrated that theywere present in both media, both in the absence and pres-ence of Cd. Furthermore, the majority of released proteinshad molar masses between 50–60 kDa. For longer contacttimes, the GP1 and GP2 glycoproteins, two of the main con-stituents of the outer layer of the cell wall of C. reinhardtii,[47]were also observed (Fig. 5). It should nonetheless be notedthat the centrifugation of the algal suspensions may havecaused limited cell rupture that may also have resulted inthe release of some intracellular compounds into the fil-trate. Moreover, the use of high-density algal suspensions(10 cm2 mL−1) in the exudate characterization experiments,in contrast to the bioaccumulation experiments, would surelyhave accentuated the observed effects. Nonetheless, based onFigs 4 and 5, the cells were clearly able to release exudates intothe experimental medium under the experimental conditionsused in the bioaccumulation experiments. The Cd complex-ing capacities of the exudates were examined in the followingSection.

Cd(II) Speciation Measurements: PLM and ASV

Measurements of [Cd2+] were performed using the PLMtechnique while mobile, labile Cd complexes were deter-mined using ASV. In order to increase the sensitivity of thespeciation measurements, exudate production was optimized

MW W6

GP1

GP2

GP3

97 kDa

66 kDa

45 kDa

30 kDa

�1 �2 �3 �4�1 �2 �3 �4

Fig. 5. SDS page of concentrated algal supernatant in the presence(+) and absence (−) of 10−7 M Cd after 10 (1), 20 (2), 30 (3), and60 min (4) (MW: molecular weight; W6: purified cell wall proteins ofthe W6 layer).

176

Cadmium Adsorption by Chlamydomonas reinhardtii

[Cd]

labi

le o

r [C

d2�

] (%

of d

isso

lved

Cd)

0

20

40

60

80

100

[Cd]

labi

le

(% o

f dis

solv

ed C

d)

0

20

40

60

80

100(c)

(d)

[Cd]

diss

, [C

d]la

bile

, or

[Cd2

�]

(nM

)

0

20

40

60

80

100

Time (min) Time (min)

0 10 20 30 40 50 60 0 10 20 30 40 50 60

[Cd]

labi

le o

r [C

d2�

](%

of d

isso

lved

Cd)

0

20

40

60

80

100

(a)

(b)

Fig. 6. (a) Dissolved (�), ASV-labile (●), and free (�) [Cd] as a function of time. (b) ASV-labile (●) andfree (�) [Cd] normalized by the total dissolved Cd in the bulk media (algal density was 2.5 cm2 mL−1).(c) Labile Cd at algal densites of 1.5 (�), 2.0 (�), and 2.5 (●) cm2 mL−1. The initial dissolved [Cd] was10−7 M for all experiments. (d) Labile (●�) and free (��) Cd as a function of contact time of the alga withthe bulk media in the absence of Cd after adding a 50 nM Cd spike (algal density was 1.5 (●�) or 2.5 (��)cm2 mL−1).

by increasing exposure times and cell densities (ca. 2×) inthe experimental media. Under these conditions, for algaeexposed to 10−7 M Cd, free (PLM), electrochemically labile(ASV), and total dissolved (ICP-MS, 0.22 µm filtration) Cddecreased with time. Nonetheless, both the electrochemicallylabile and free Cd decreased at a greater rate than did the totaldissolved Cd (Fig. 6a). As expected, the measured concentra-tion of labile Cd was greater than the concentration of Cd2+measured by the PLM (Fig. 6a,b) due to the dependence of thevoltammetric signal on the concentration of both the free ionand some of its complexes.[41] The ASV/PLM ratio rangedfrom near unity values at the beginning of the experimentsto values between 4 and 8 at the end of the accumulationtime. The rapid decrease in both the PLM and ASV signalscoupled to the absence of a shift in the potential of the reox-idation peak in the presence of exudates strongly suggestedeither that the ligands had a high affinity for Cd (i.e. kineti-cally inert complexes) or that the complexes were large andthus relatively immobile (in the time frame of the voltammet-ric experiment). Based upon an estimated stability constantof 103.3–105.3 M−1 for the interaction of Cd with the cellwall (assuming equilibrium and Cd saturation at 2 × 10−9

mol cm−2, Fig. 3) and the results from the protein gels, thesecond explanation would appear more likely.

The effect of algal cell density was also important (Fig. 6c)with a more pronounced decrease in the ASV-labile Cd whenthe algal density was increased from 1.5 to 2.5 cm2 mL−1.Indeed, following a 20-min exposure, 40–60% of the dis-solved Cd was bound in ASV-labile complexes. A similardecrease was observed when Cd was added to the algal super-natant obtained for cells that were not exposed to Cd (Fig. 6d).Such results support our previous observation that protein

production was a constitutive process rather than a responseto Cd stress.

Implications for Bioaccumulation Studies

It is possible to speculate that the nature of the exudatesand their observation for nearly all conditions suggestedthat the production of exudates by algae and their releaseinto the bulk media may be important in natural watersin reducing free metal concentrations to low levels.[48] Inthis study, theoretical models describing metal adsorption tohomogenous or heterogeneous algal surfaces with or withoutcharge correction[18,19] were unable to relate Cd adsorptionby C. reinhardtii to the speciation of Cd in the bulk media.The observation of an increase in adsorbed Cd with timesuggested that Cd binding sites were being continuously pro-duced. An important implication of this result is that themetal-uptake sites will never be at simple chemical equilib-rium with respect to the bulk solution (a requirement of boththe biotic ligand and free ion activity models). Cd adsorptionto C. reinhardtii is thus a dynamic rather than an equilib-rium process involving several adsorption steps. Becausethe CW-2 strain is thought to produce the same amount ofcell wall material as the wild type strain,[29] the observeddifferences in Cd adsorption for the two strains are likelydue to a greater release of the cell wall components for theCW-2 mutant. Indeed, the rate of Cd adsorption for CW-2strain was five times lower than that observed for the WTstrain (Fig. 1a). The observation of a less than unity slopein the (log) adsorbed Cd – (log) [Cd2+] plot was attributedto cell surface production in addition to the more usual het-erogeneity and charge effects.[46,49] It is important to notethat a rapid rate of Cd internalization further distinguishes

177

H. Kola et al.

C. reinhardtii from several other algal species typicallyused in bioaccumulation and toxicity studies. Nonetheless,whereas the rapid uptake coupled with a continuous adsorp-tion was likely responsible for the observed depletion ofdissolved Cd in the bulk solution, [Cd2+] was also signif-icantly reduced by the complexation of Cd by the exudates.While the production of metal-binding ligands has been pre-viously observed in response to a Cd stress,[50] the continualrelease of the cell wall components into the medium undera variety of physicochemical conditions was unexpected,especially for such short exposure times (Fig. 6).

Finally, although several authors[51] advocate the use ofN-substituted aminosulfonic acids buffers such as HEPESfor metal bioaccumulation and toxicity experiments due totheir low complexation capacity,[52] exudate production wasclearly enhanced in their presence.[53] On the other hand,while macronutriment buffers such as the phosphate buffermay have less effect on the cell wall, they are generallynot recommended due to their potential metabolic roles ortheir ability to precipitate or complex polyvalent cations andthus reduce biouptake.[54] Based upon these results, ‘ideal’buffers likely do not exist for metal bioaccumulation or toxi-city experiments. Inevitable effects of the buffers will need tobe taken into account and controlled, where possible, throughthe reduction of cellular densities, exposure times, and by theselection of suitable buffers. The use of chemostats may alsobe a useful means to minimize changes in both metal concen-trations (typically verified) and speciation (rarely verified).

Conclusions

Cd adsorption to living cells, in this case, C. reinhardtii,is a complex, dynamic process that cannot be describedby simple equilibrium models. Slow Cd adsorption dynam-ics was attributed to the production of new binding sites,especially for low Cd concentrations. Due to the continualrelease of cell wall proteins and the production of bindingsites on the cell wall, equilibrium was never achieved underthe studied conditions. The cell wall was highly complex-ing: wild type C. reinhardtii bound ten-fold more Cd thanthe cell wall-less strain. Under the studied conditions, theproduction and release of cell wall components by algaein media with or without Cd demonstrated that exudaterelease was not in response to Cd stress but rather a consti-tutive process. The ubiquitous release of the exudates greatlyaffected Cd speciation in the exposure media by increasingCd complexation.

Acknowledgements

The authors acknowledge the fifth European frameworkBIOSPEC project (EVK1-CT-2001–00005) and the SwissNational Science Foundation (200020–101788) for providingfunding for this work. The authors thank S. Waffenschmidt(University of Cologne) and M. Goldschmidt-Clermont (Uni-versity of Geneva) for providing the CW-2 strains, purifiedproteins of the W6 layer of the cell wall, wild type C. rein-hardtii (2137), as well as other important information on the

algal strains. The assistance of M. Martin for the ICP-MS andC. Gehin-Delval for the nitrogen measurements is also appre-ciated. Critical reviews of a previous draft of the manuscriptby I. Worms and D. Simon and discussion with H. P. vanLeeuwen were extremely helpful.

References

[1] L. Sigg, in Aquatic Surface Chemistry (Ed. W. Stumm) 1987,p. 319 (Wiley-Interscience: New York, NY).

[2] F. M. M. Morel, N. M. Price, Science 2003, 300, 944.doi:10.1126/SCIENCE.1083545

[3] K. Nagel, J. Voigt, Microbiol. Res. 1995, 150, 105.[4] M. N. Prasad, K. Drej, A. Skawinska, K. Strzalka, K. Stralka,

Bull. Environ. Contam. Toxicol. 1998, 60, 306. doi:10.1007/S001289900626

[5] B. Lustigman, L. H. Lee, C. Weiss-Magasic, Bull. Environ.Contam. Toxicol. 1995, 55, 65. doi:10.1007/BF00212390

[6] D. Ben-Bassat, G. Shelef, N. Gruner, H. I. Shuval, Nature 1972,240, 43.

[7] C. Weiss-Magasic, B. Lustigman, L. H. Lee, Bull. Environ.Contam. Toxicol. 1997, 59, 828. doi:10.1007/S001289900556

[8] M. Devriese, V. Tsakaloudi, I. Garbayo, R. Leon, C. Vilchez,J. Vigara, Plant Physiol. Biochem. 2001, 39, 443. doi:10.1016/S0981-9428(01)01257-8

[9] U. Irmer, I. Wachholz, H. Schaefer, D. W. Lorch, Environ.Experim. Bot. 1986, 26, 97. doi:10.1016/0098-8472(86)90002-X

[10] www.biology.duke.edu/genome_chlamy[11] P. Rubinelli, S. Siripornadulsil, G.-R. Fei, R. T. Sayre, Planta

2002, 215, 1. doi:10.1007/S00425-001-0711-3[12] J. L. Moseley, T. Allinger, S. Herzog, P. Hoerth, E. Wehinger,

S. Merchant, M. Hippler, EMBO J. 2002, 21, 6709. doi:10.1093/EMBOJ/CDF666

[13] J. L. Moseley, M. D. Page, N. P. Alder, M. Eriksson, J. Quinn,F. Soto, S. M. Theg, M. Hippler, S. Merchant, Plant Cell 2002,14, 673. doi:10.1105/TPC.010420

[14] T. Sasaki, N. Kurano, S. Miyachi, Plant Cell Physiol. 1998,39, 131.

[15] S. Siripornadulsil, S. Traina, D. P. S. Verma, R. T. Sayre, PlantCell 2002, 14, 2837. doi:10.1105/TPC.004853

[16] N. Mirimanoff, K. J. Wilkinson, Environ. Sci. Technol. 2000,34, 616. doi:10.1021/ES990744G

[17] C. S. Hassler, K. J. Wilkinson, Environ. Toxicol. Chem. 2003,22, 620.

[18] A. C. C. Plette, M. F. Benedetti, W. H. van Riemsdijk, Environ.Sci. Technol. 1996, 30, 1902. doi:10.1021/ES950568L

[19] A. C. C. Plette, M. M. Nederlof, E. J. M. Temminghoff, W. H. VanRiemsdijk, Environ. Toxicol. Chem. 1999, 18, 1882.

[20] R. H. Crist, K. Oberholser, D. Schwartz, J. Marzoff, D. Ryder,D. R. Crist, Environ. Sci. Technol. 1988, 22, 755.

[21] A. Tessier, J. Buffle, P. G. C. Campbell, in Chemical andBiological Regulation of Biological Systems (Eds J. Buffle,R. R. Devitre) 1994 (Lewis: New York, NY).

[22] E. Kiefer, L. Sigg, P. Schosseler, Environ. Sci. Technol. 1997,31, 759. doi:10.1021/ES960415D

[23] J. Buffle, in Complexation Reaction in Aquatic Systems: AnAnalytical Approach 1988, pp. 304 (Ellis Horwood: Chichester).

[24] H. B. Xue, W. Stumm, L. Sigg, Water Res. 1988, 22, 917.doi:10.1016/0043-1354(88)90029-2

[25] H. Xue, L. Sigg, Water Res. 1990, 24, 1129. doi:10.1016/0043-1354(90)90176-7

[26] M. Melkonian, Z. Pflanzenphysiol. 1979, 94, 125.[27] B. Marsalek, R. Rojickova, Z. Naturforsch. C: J. Biosci. 1996,

51, 646.[28] M. Teplitski, H. Chen, S. Rajamani, M. Gao, M. Merighi,

R. T. Sayre, J. B. Robinson, B. G. Rolfe, W. D. Bauer, PlantPhysiol. 2004, 134, 137. doi:10.1104/PP.103.029918

[29] R. D. Davis, A. Plaskitt, Genet. Res. Camb. 1971, 17, 33.

178

Cadmium Adsorption by Chlamydomonas reinhardtii

[30] G. E. Walsh, Environ. Toxicol. Chem. 1988, 7, 979.[31] E. H. Harris, in Culture and Storage Methods 1989, p. 25

(Academic Press: San Diego, CA).[32] A. E. Martel, R. M. Smith, in Critically Selected Stability

Constants of Metal Complexes, Data Base ver. 5.0 1998 (NIST:Gaithersburg, MD).

[33] C. S. Hassler, V. I. Slaveykova, K. J. Wilkinson, Limnol.Oceanogr. Methods 2004, 2, 237.

[34] H. Kola, K. J. Wilkinson, Environ. Sci. Technol. 2004, submitted.[35] N. J. Kruger, in Protein Protocols Handbook (Ed. J. M. Walker)

1996, p. 15 (Humana Press: Totowa, NJ).[36] J. M. Walker, in The Protein Protocols Handbook (Ed.

J. M. Walker) 1996, pp. 55 (Humana Press: Totowa, NJ).[37] U. W. Goodenough, J. E. Heuser, J. Cell Biol. 1985, 101, 1550.

doi:10.1083/JCB.101.4.1550[38] N. Parthasarathy, M. Pelletier, J. Buffle, J. Chromatogr. A 2004,

1025, 33. doi:10.1016/J.CHROMA.2003.10.083[39] P. R. Danesi, Sep. Sci. Technol. 1984, 19, 857.[40] M. L. Tercier, J. Buffle, Anal. Chem. 1996, 68, 3670. doi:10.1021/

AC960265P[41] J. Pei, M.-L. Tercier-Waeber, J. Buffle, Anal. Chem. 2000, 72,

161. doi:10.1021/AC990628W[42] S. S. Bates, A. Tessier, P. G. C. Campbell, J. Buffle, J. Phycol.

1982, 18, 521.[43] R. J. M. Hudson, F. M. M. Morel, Deep-Sea Res. 1993, 40,

129. doi:10.1016/0967-0637(93)90057-A

[44] H. P. Van Leeuwen, in Environmental Particles (Eds J. Buffle,H. P. Van Leeuwen) 1992, pp. 554 (Lewis: Boca Raton, FL).

[45] B. Raspor, H. W. Nürnberg, P. Valenta, M. Branica, J. Electroanal.Chem. 1980, 115, 293. doi:10.1016/0368-1874(80)80415-1

[46] J. P. Pinheiro, A. M. Mota, H. P. van Leeuwen, Colloids Surf. A:Physicochem. Eng. Aspects 1999, 151, 181. doi:10.1016/S0927-7757(98)00701-8

[47] U. W. Goodenough, B. Gebhart, R. P. Mecham, J. E. Heuser,J. Cell Biol. 1986, 103, 405. doi:10.1083/JCB.103.2.405

[48] R. M. Town, M. Filella, Limnol. Oceanogr. 2000, 45, 1341.[49] K. J. Wilkinson, J. Buffle, in Physicochemical Kinetics and

Transport at Biointerfaces (Eds H. P. van Leeuwen, W. Köster)2004, pp. 447 (John Wiley: Chichester).

[50] A. Smiejan, K. J. Wilkinson, C. Rossier, Environ. Sci. Technol.2003, 37, 701. doi:10.1021/ES025901H

[51] M. R. Twiss, J.-C. Auclair, M. N. Charlton, Can. J. Fish. Aquat.Sci. 2000, 57, 870. doi:10.1139/CJFAS-57-4-870

[52] H. E. Mash, Y.-P. Chin, L. Sigg, R. Hari, H. Xue, Anal. Chem.2003, 75, 671. doi:10.1021/AC0261101

[53] M. T. S. D. Vasconcelos, M. F. C. Leal, Environ. Toxicol. Chem.2002, 21, 404.

[54] D. Khummongkol, G. S. Canterford, C. Fryer, Biotechnol.Bioeng. 1982, 24, 2643.

179