Chemosphere 81 (2010) 904–910

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Defining the surface adsorption and internalization of copper and cadmiumin a soil bacterium, Pseudomonas putida

Mindy W. Pabst a, Charles D. Miller b, Christian O. Dimkpa b, Anne J. Anderson c, Joan E. McLean d,*

a Utah Water Research Laboratory, Utah State University, 8200 Old Main Hill, Logan, UT 84322-8200, USAb Department of Biological Engineering, Utah State University, 4105 Old Main Hill, Logan, UT 8322-4105, USAc Department of Biology, Utah State University, 8200 Old Main Hill, Logan, UT 84322-8200, USAd Department of Biological Engineering, Utah Water Research Laboratory, Utah State University, 8200 Old Main Hill, Logan, UT 84322-8200, USA

a r t i c l e i n f o

Article history:Received 23 April 2010Received in revised form 28 July 2010Accepted 30 July 2010Available online 24 August 2010

Keywords:MetalsSorptionUptakeExtractionBacteriaPseudomonas

0045-6535/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2010.07.069

* Corresponding author. Tel.: +1 435 797 3199; faxE-mail address: joan.mclean@usu.edu (J.E. McLean

a b s t r a c t

The distribution of cadmium (Cd) and copper (Cu) ions onto and within two soil pseudomonads, Pseudo-monas putida strains KT2440 and Corvallis, was investigated using selective extraction procedures andmodeled using Langmuir isotherms. Cadmium and Cu associated differently with the surface, periplasmand cytoplasm of the two strains. Both of these pseudomonad cells allowed more Cu to pass into the peri-plasmic space and to the cytoplasm than Cd. The distribution of Cu among the cellular spaces was solu-tion concentration dependent, with limited amounts of Cu entering the cell at higher exposureconcentrations. The Langmuir isotherm with a single binding site fit well to the observed data for Cu cellassociation. Cadmium was mainly found on the surface of the cells. The capacity of surface exchange sitesfor Cd increased with solution concentration, possibly indicating a modification of surface functionalgroups with ion concentration. This surface sorption behavior of Cd was best described using a two-siteLangmuir model, whereas all other Cu and Cd associations were described using a one-site model.Although potentiometric titration identified differences in site densities for proton binding to the twostrains, these differences were not consistently displayed with Cu and Cd surface interactions.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Copper and cadmium are common toxic metal contaminants ofsoil and water being released into the environment by natural pro-cesses and as a result of human activities. Metals are classifiedaccording to the Hard Soft Acids Bases theory based on atomic size,electronegativity, and polarizability (Pearson, 1963). Cadmium is asoft acid and complexes primarily with sulfur containing com-pounds, whereas Cu2+ borders between hard and soft acids andcomplexes with a variety of inorganic ligands and organic func-tional groups. The association of Cu and Cd ions with differentfunctional groups results in the two metals differing in their envi-ronmental behavior and biological response.

The formation of complexes between metals and anionic func-tional groups on bacterial cell surfaces is viewed as a pre-requisitefor uptake of metals by the organism (Campbell et al., 2002). Oncesurface sorption occurs, the metal may be transported into theperiplasmic space of Gram-negative cells and be transported fur-ther into the cytoplasm. This process is a series of complexationreactions between the metal and biomolecules involved in surfacebinding, transportation, storage, and biofunctions. The interaction

ll rights reserved.

: +1 435 797 3663.).

of metals to solid surfaces is commonly modeled using linear ornon-linear isotherms, such as the Freundlich or Langmuir models,that describe the relationship between the activity of the free ionin solution at equilibrium (Ce) and the amount of metal associatedwith the solid surface (qads). Adsorption models have been used todescribe the removal of metals from solution by bacteria (Kulc-zycki et al., 2002; Pagnanelli et al., 2002; Loukidou et al., 2005).Metallic cation adsorption to bacterial surfaces contributes to theoverall retention and mobility of metals in soil and water systems(Huang et al., 2005; Claessens et al., 2006; Guine et al., 2006); anapplication of adsorption is the use of live or dead cells for removalof metals from wastewaters (e.g., Zouboilis et al., 2004; Aksu andDonmez, 2006). Many adsorption studies consider only the re-moval of the metal from solution, attributing this removal to sur-face reactions and neglecting further transport of the metal intothe cell.

Bacterial cells have mechanisms for uptake of Cu to ensureavailability of this essential trace nutrient for its role as a cofactorin many enzymes (Nies, 1999). Passage into the cytoplasm requiresboth outer and inner membrane transport systems. Copper is ex-tremely toxic to the cell above the metabolic need so its cytoplas-mic level is maintained by chaperone binding in the cytoplasm(Jordan et al., 2000), efflux systems in the cytoplasmic membrane(Lewinson et al., 2009), and binding proteins in the periplasm (Ba-

M.W. Pabst et al. / Chemosphere 81 (2010) 904–910 905

gai et al., 2008). In contrast, Cd is a nonessential element with nospecific uptake systems, gaining entry erroneously through othertransporters (Silver, 1998; Hao et al., 1999). However, bacterialcells do possess mechanisms for Cd efflux and binding to limit tox-icity (Pagès et al., 2007).

This study examines the accumulation of Cu and Cd on the sur-face and within the periplasm and cytoplasm of soil pseudomo-nads. Two isolates of Pseudomonas putida were selected forstudy: strain KT2440 was isolated from industrial sewage sludge(Nusslein et al., 1992) and strain Corvallis was isolated from fieldgrown-wheat (Anderson and Jasalavich, 1979). The genome ofKT2440 has been sequenced because of its potential to survive pol-lutant stress (Nelson et al., 2002). Both strains colonize plant roots(Buell and Anderson, 1993; Molina et al., 2005); however, isolateCorvallis has a modified cell wall lacking the O-antigen portionof its lipopolysaccharide (LPS) (Nusslein et al., 1992). Thus, wecompared Cu and Cd adsorption and uptake between the two pseu-domonads. Potentiometric titrations were used to characterize thetypes of charged functions on the bacterial surfaces that could actas the initial metal ion binding sites. The distribution of Cu and Cdon and within the cell was determined using chemical methods tosequentially access ionically bound ions at the bacterial surfaceand within the periplasm or cytoplasm. The surface binding, trans-portation, and internal location of metals are a series of metal com-plexation reactions with functional groups of biomolecules, eithermembrane bound or within the periplasm or cytoplasm. Compari-sons between the association of Cu and Cd with the two isolates ofP. putida were made using Langmuir isotherms.

2. Materials and methods

2.1. Bacterial stocks

P. putida isolate KT2440 (Nusslein et al., 1992) and isolate Cor-vallis (Anderson and Jasalavich, 1979) were grown to stationaryphase in liquid minimal medium (MM) containing per L: K2HPO4

10.5 g, KH2PO4 4.5 g, sodium citrate (2H2O) 0.5 g, (NH4)2SO4

1.0 g, sucrose 20 g and MgSO4 0.1 g. This medium was inoculatedwith stock cultures stored in 15% glycerol at �80 �C. Chemicalsused were Mallinckrodt Analyzed Reagents (AR), KY.

2.2. Potentiometric titration

Potentiometric titrations were performed to identify and quan-tify functional groups on the bacterial surfaces that may be respon-sible for metal adsorption. Stationary phase cells grown in MMwere resuspended into 100 mL of 0.01 M KNO3 to a density of108–109 cells mL�1, a bacterial suspension of 0.5 gdw L�1 (dw: dryweight). The suspensions were purged with N2 in a closed flaskfor 30 min to remove CO2, and N2 was bubbled into the flask con-tinuously throughout the titrations. Each suspension was titratedto the starting pH value of 3.6 using small aliquots of 1 M HCland back-titrated with 0.108 M NaOH to a pH of 10 (Borroket al., 2005a). Short-term cell exposure to pH values used for thistitration, as typically used in the literature (Borrok et al., 2005b),does not cause loss of cell viability (Borrok et al., 2004) or mem-brane damage (Guine et al., 2006). Titrations were performed intriplicate. Protofit GUI Version 2.1 was used to analyze the titrationdata (Turner and Fein, 2006). Protofit calculates the number of sitesavailable, the type of sites, the pKa (�log Ka, the acid dissociationconstant), and the site concentrations (log C) based on the surfacecomplexation model selected, the background electrolyte concen-tration, the titrant concentration, the solution volume, and theadsorbent mass. Titration data were fitted to the non-electrostaticsurface complexation model option within Protofit using two,

three, and four discrete sites. The four-site model provided the bestfit of the experimental data. This four-site model provided the bestfit of experimental data across a variety of bacteria (Borrok et al.,2005b). All cell mass is reported on a dry weight basis. The dryweight of the cells was determined by pelleting the cells via centri-fugation then drying the cells in a 100 �C oven for 24 h.

2.3. Metal solutions analysis

The metal solutions were prepared at concentrations from0.008 to 0.16 mM (0.5–10 mg L�1) of free Cu2+ ions using Cu(NO3)2

and from 0.004 to 0.1 mM (0.5–10 mg L�1) of Cd2+ using Cd(NO3)2

in a background of 0.01 M KNO3. Nitrate does not form complexeswith Cu or Cd. At the pH values of the prepared metal solutions, pH5, using the nitrate salts, Cu and Cd, as determined by geochemicalmodeling (GEOCHEM-PC), were present in solutions as the freemetal ions (Parker et al., 1995) with no precipitation. Solutionswere prepared fresh for each experiment. All solutions wereanalyzed by Atomic Absorption Spectrometry (AA, PerkinElmerAnalyst 800) or Inductively Coupled Plasma Mass Spectrometry(ICP-MS, Agilent 7500 series). The linear range of the AA for Cuwas from 0.002 to 0.08 mM (0.1–5 mg L�1) and for Cd, 0.0004–0.02 mM (0.04–2 mg L�1). The linear range used by the ICP-MSwas 7.8 � 10�5 to 0.02 mM for Cu and 4.5 � 10�5 to 0.009 mMfor Cd. Detection limit on a mass basis was 0.23 mmol kg�1 Cuand 0.13 mmol kg�1 Cd by ICP-MS.

2.4. Growth of bacteria

Thawed cells from frozen stocks (200 lL) were grown to earlystationary phase in 50 mL of MM with shaking at 220 rpm in a28 ± 2 �C incubator. Plate counts were performed to ensure a min-imum cell density of 108 cells mL�1. Aliquots of 25 mL mediumwith cells was placed into 30 mL plastic centrifuge tubes, centri-fuged, and washed in sterile deionized water (DI), centrifugedagain and resuspended in 15 mL of the metal solutions or 0.01 MKNO3 as a control. The samples were shaken at 220 rpm at25 ± 1 �C for 30 min, and then centrifuged to pellet the cells andobtain a supernatant. This exposure time was adequate to obtainsteady state conditions (Langley and Beveridge, 1999; Wu et al.,2009) yet minimize phase-dependent changes in cell activity. ThepH of the initial bacteria–metal solutions was 5.6. Chemical buffer-ing agents were not added to avoid complexation with Cu or Cd,but the pH was maintained throughout the study to within 0.2pH units without the need for pH adjustment. Experiments wereperformed in triplicate.

2.5. Analysis: solution phase

After centrifugation, the supernatant was removed from thecentrifuge tube with a pipette, filtered through a 0.2 lm nylon fil-ter, and preserved with nitric acid to pH 2, and the solution wasanalyzed for Cu or Cd using the AA or ICP-MS.

2.6. Analysis: cell compartmentalization

A series of extractions was used to determine the proportion forCu and Cd associated with surface exchange sites (qexch), the peri-plasmic space (qper), and the cytoplasm (qcyt). Metals associatedwith exchange sites on the cell surface were extracted by resus-pending the cell pellet in 15 mL of 5 mM Ca2+ as Ca(NO3)2 for15 min with gentle shaking at 80 rpm (Salt et al., 1997). The sus-pension was centrifuged at 10,000g and the cell pellet was treatedwith 15 mL of 20 mM EDTA (tetrasodium salt dehydrate) for 20 s(Kola and Wilkinson, 2005) to disrupt the outer membrane and re-lease periplasmic components (Vaara, 1992). The EDTA extraction

906 M.W. Pabst et al. / Chemosphere 81 (2010) 904–910

would include metals tightly bound to the outer membrane but notexchanged with the Ca in the first step and metals in the periplasm.After centrifugation, the remaining pellet was dissolved in 15 mL of50% concentrated nitric acid (trace metal grade, Fisher Science, PA)for 30 min. Each supernatant was filtered, acidified, and Cu and Cdanalyzed by AA or ICP-MS. The centrifuge tubes were weighed be-tween each extraction step to determine the mass of cells and li-quid associated with the cells.

The sum of Cu or Cd from the three extraction steps was desig-nated as the total amount of metal associated with the cell, qe. Themass balance, expressed as a percent, was between 90% and 105%when the initial solution concentration and the sum of extractedmetals were compared.

2.7. Quality control

Each test included: (1) 0.01 M KNO3 with no metals or cells, (2)0.01 M KNO3 with cells but no metals, and (3) 0.01 M KNO3 with nocells but with the test metal. These controls were included to pro-vide: (1) contamination levels for Cu or Cd through the procedures,(2) metal levels in unexposed cells, and (3) any losses of Cu and Cdnot due to cell adsorption/uptake throughout the experimentalprocedures. All blanks were below the detection limit and therewas no loss of Cu or Cd to laboratory glassware.

Fig. 1. Q �tot (protons exchange with adsorbent (mol H+ kg�1 dry weight/pH)) for P.putida strain Corvallis (d) and strain KT2440 (s). The model prediction is displayedas the solid line for Corvallis and the dash line for KT2440. Data points fromtriplicate titrations are displayed on the graph.

2.8. Langmuir model

The Langmuir model was used as a tool to compare the resultsfor the distribution of free metal ions to cells of P. putida KT2440and Corvallis isolates. The Langmuir model (Eq. (1)) describes thenon-linear relationship between the activity of the free ion in solu-tion at equilibrium (Ceq mmol L�1) and the amount of metal (q) onthe surface exchange sites (qexch), in the periplasmic space (qper), orin the cytoplasm (qcyt) in mmol kg�1.

q ¼ QobCeq

1þ bCeqð1Þ

where Qo is a sorption maximum and b is the magnitude of the ini-tial slope. The Langmuir model was developed to describe revers-ible adsorption of gases to glass surfaces but also has beenapplied to more complex systems, such as the sorption of metalsto pure minerals, soils, and bacteria. The model well describes theexponential relationship between Ceq and q observed for metalsorption to various surfaces. The model is used here for goodnessof fit and for statistical comparison of the behavior of the two iso-lates and not to interpret mechanistic details.

The best-fit model was obtained by directly fitting data to theLangmuir model by minimizing the residual sum of squares (Bol-ster and Hornberger, 2007). Also a two-surface Langmuir model(Eq. (2)) was used to describe the surface adsorption process forCd. This model assumes that the sorption occurs on two types ofsurfaces, each with different bonding energies, or on one surfacewith two distinct energies (Bolster and Hornberger, 2007).

q ¼ Q 1ob1Ceq

1þ b1Ceqþ Q2

ob2Ceq

1þ b2Ceqð2Þ

Joint confidence regions were calculated by creating a table ofcritical sum of squares for various Q and b values. The correspond-ing F value for the degrees of freedom indicates the 95% joint con-fidence regions. The joint confidence regions were used todetermine if the P. putida KT2440 and Corvallis isolates were sim-ilar in the association distribution of Cu and Cd for qexch, qper, andqcyt.

3. Results and discussion

3.1. Proton binding capacity of the bacterial cell surface

Acid–base exchange sites have been attributed to carboxyl andphosphodiesters (pH < 4.7), phosphomonoesters and sulfhydryl(neutral sites), and hydroxyl and amine (pH > 8), functional groupsthat may be found in the O-antigen chains, the core-LPS and the li-pid A layer of the lipopolysaccharide component of the outer leafletof the outer membrane of Gram-negative bacteria (Guine et al.,2006). The �log of the acid dissociation constants (pKa) and thesite densities (mmol kg�1

dw) describing these surface sites on thebacterial cells were determined through potentiometric titrationsand use of the Protofit model. A non-electrostatic model in Protofitwith four discrete sites provided the best fit to both the titrationcurves and Q �tot , the total proton buffering capacity function(Q �tot = dQtot/dpH, mol H+ kg�1

dw/pH) (Fig. 1). The four-site model(pKa1–pKa4) provided a reasonable estimate of the extent of pro-tonation of the cell wall across a variety of Gram-positive andGram-negative bacterial species and consortia (Borrok et al.,2005b). The two strains have similar types of surface functionalgroups since the pKas were the same (Table 1). The Corvallis strain,however, had statistically higher site densities for the first twosites compared with isolate KT2440, whereas the more basic bind-ing site had statistically the same density (Table 1). The densitiesdifferences may relate to the apparent lack of O-antigen chainson LPS for isolate Corvallis, making the negatively charged LPS-corestructures more available (Langley and Beveridge, 1999). The totalnumber of sites, however, was in agreement with the average ofthe total concentration of proton functional group sites for 36 bac-terial species (2560 ± 800 mmol kg�1

dw) (Borrok et al., 2005b).All Cu and Cd adsorption experiments were performed at a pH

of 5.6 ± 0.2 where the cell surface had a charge of 322 ±3.6 H+ mmolc kg�1

dw for the Corvallis isolate and 167 ± 3.4 H+

mmolc kg�1dw for the KT2440 isolate (Fig. 1). At this pH, carboxylic,

phosphoric, and sulfhydryl functional groups on the cell wouldbe available for exchange reactions with Cu or Cd ions. The pH re-mained within ±0.2 pH units throughout the interactions of themetal with the bacterial cells, without need for any pH adjustment.

3.2. Copper responses

Exposure of stationary phase KT2440 and Corvallis isolates to0.15 mM (10 mg Cu L�1) did not result in loss of cell culturability(data not shown). Other researchers have used similar concentra-

Table 1The pKa (�log of the acidity constant) and site density (mmol kg�1

dw) for the average and 95% confidence intervals for each site associated with the cells of the Corvallis and KT2440isolates as determined by potentiometric titrations and calculation by the ProtoFit GUI program (Turner and Fein, 2006).

pKa1 Site density(mmol kg�1)

pKa2 Site density(mmol kg�1)

pKa3 Site density(mmol kg�1)

pKa4 Site density(mmol kg�1)

Total site density(mmol kg�1)

Corvallis 4.8 ± 0.1 646 ± 56 6.4 ± 0.2 417 ± 22 8.6 ± 0.3 309 ± 81 10.4 ± 0.1 2138 ± 411 3510KT2004 5.1 ± 0.3 275 ± 146 6.4 ± 0.8 190 ± 80 7.8 ± 1.0 199 ± 75 10.7 ± 0.3 2884 ± 1898 3548

Fig. 2. Sorption isotherms of Cu associated with exchangeable sites (qexch), theperiplasmic space (qps), and the cytoplasm (qcyt) of P. putida Corvallis (A) and P.putida KT2440 (B). The points are averages of the experimental data (n = 3) with95% confidence interval error bars. The lines are the Langmuir model predictions.

M.W. Pabst et al. / Chemosphere 81 (2010) 904–910 907

tions of Cu with other pseudomonad strains without death of thecells or cell wall damage (Langley and Beveridge, 1999). With bothcell types, exposure to Cu resulted in cell pigmentation: a brightblue for the Corvallis cells and a green–blue for KT2440. The colorwas retained after Ca2+ treatments, suggesting that the outer mem-brane was intact. The intent of this procedure was to remove Cuassociated with surface ionic-exchangeable sites; however, pig-mentation was lost with EDTA extraction, a procedure that otherresearchers have used to quantify metal attachment to the cellmembrane (Kola and Wilkinson, 2005). Because of the prior extrac-tion of ionic-bound materials, this EDTA wash would haveremoved Cu that was more tightly bound to the outer membraneand Cu associated with the periplasmic space. For Gram-negativebacteria such as these pseudomonads, EDTA would have destabi-lized the LPS component of the outer membrane leaflet byremoving Ca2+ and Mg2+ (Vaara, 1992). The blue coloration of theCu-treated cells of the two pseudomonads could not be due to pre-cipitation since Cu does not form a solid phase at the pH of thisstudy (pH 5.6), as determined by geochemical modeling usingGEOCHEM. Rather, the observed blue color was caused by the com-plexation of Cu with specific periplasmic Cu-binding proteins: pro-teins that are thought to play a key role in copper homeostasis. Thecolor difference between cells of KT2440 and Corvallis suggestedthat the Cu-binding proteins may differ between the strains. Geno-mic sequencing of isolate KT2440 has revealed the presence ofgenes encoding periplasmic Cu-binding proteins (Canovas et al.,2003). Our transcriptome analysis of KT2440 cells showed, uponexposure to Cu2+, activation of the genes porD PP_0046, copA1PP_5380, copA1 PP_2205, and copB1 PP_5379 encoding Cu-bindingproteins in the outer membrane and periplasmic space (Milleret al., 2009).

The experimental values for the equilibrium solution concen-tration (Ceq) and the Cu associated with surface exchange sites,the periplasmic space, and the cytoplasm were described withthe Langmuir single site isotherm (Fig. 2 and Table 2). The Lang-muir isotherm provided a good fit of the exponential relationshipsfor compartmentalization within the periplasm and cytoplasmalthough they are not reversible surface adsorption reactions, anassumption of the model. The use of the model allows statisticalcomparison between fitted parameters for the two isolates withoutimplying mechanistic detail. The calculated Q values for all com-partments are within the range of the experimental values (Table2 and Fig. 2). The Langmuir parameters, Q and b, were not differentfor the two isolates within each of the compartments analyzed (Ta-ble 2). This finding is interesting because the titration study detect-ing proton-binding sites suggested that the Corvallis cells hadtwice the number of sites as the KT2440 cells at the pH of thisstudy. Assuming a 1:1 complex formation, the KT2440 surface siteswould be saturated with Cu at the sorption maximum whereasCorvallis would be at half saturation, implying that a portion ofthe surface sites were not available for Cu sorption with this iso-late. The Corvallis isolate, lacking the O-antigen, binds the sameamount of Cu to the surface of the cell as KT2440. This findingagreed with studies by Langley and Beveridge (1999) with Pseudo-monas aeruginosa and three isogenic LPS mutants, showing that thecharge on the O-antigen of the LPS did not affect the extent of Cusorption by these cells. They concluded that Cu was bound to sites

common to all strains, perhaps the phosphoryl groups in the core-lipid A region. At the estimated sorption maximum, Q, the majorityof the total Cu associated with the cell was on exchangeable sur-face sites (56–57%), 33–35% was associated with the periplasm,and 9% was in the cytoplasm for both cell types (Table 2).

The distribution among the three cell compartments was solu-tion concentration dependent (Fig. 2). This process is well illus-trated for KT2440 (Fig. 2B) where more Cu is associated with theperiplasmic space than on exchange sites at equilibrium solutionconcentrations below 0.02 mM. The concentration of Cu in thecytoplasm was constant (23 ± 5.8 mmol kg�1) at solution concen-trations above 0.008 mM, whereas the proportion of Cu on the sur-face and within the periplasm shifted with increasing solutionconcentration.

Although the calculated values of Q and b for the Langmuir iso-therm were not significantly different between strains, the mea-sured values showed that the strains differed slightly incompartmentalization with solution concentration. At the lowersolution concentrations, more Cu was surface associated for the

Table 2Langmuir parameters, Q and b, for Cu associated with the surface, periplasmic space and cytoplasm of P. putida strains Corvallis and KT2440.

Location Q (mmol kg�1) b (L kg�1) Sum of square error r2 Percent of total Cuassociated with eachcompartment of thecell based on Q values (%)

Corvallis (n = 42)Surface exchangeable 154 (a)a 209 (a) 1.61E + 04 0.928 57.4Periplasmic space 89.4 (b) 476 (b) 1.38E + 04 0.836 33.3Cytoplasm 25.1 (c) 547 (c) 1.71E + 03 0.750 9.4

KT2440 (n = 21)Surface exchangeable 184 (a) 65.0 (a) 1.54E + 03 0.981 56.2Periplasmic space 115 (b) 320 (b) 7.37E + 03 0.856 35.1Cytoplasm 28.7 (c) 175 (c) 4.54E + 02 0.835 8.7

a Q and b values followed by the same letter indicate that values were not significantly different for the two isolates within that cell compartment.

908 M.W. Pabst et al. / Chemosphere 81 (2010) 904–910

Corvallis cells, whereas more entered the periplasm for the KT2440cells (Fig. 2). The modified LPS structure of the Corvallis outer leaf-let may allow for more Cu to be held on the surface at these lowersolution concentrations.

Fig. 3. Sorption isotherms of Cd associated with exchangeable sites (qexch), theperiplasmic space (qps), and the cytoplasm (qcyt) of P. putida Corvallis (A) and P.putida KT2440 (B). The points are averages of the experimental data (n = 3) with95% confidence interval error bars. The lines are the Langmuir model predictions.The two-site Langmuir model was used to describe the surface exchange behaviorof Cd.

3.3. Cadmium response

Exposure of both isolates at stationary phase to0.09 mmol L�1 Cd (10 mg Cd L�1) did not affect culturability. Un-like Cu exposure, there was no change in pigmentation with theCd-exposed pseudomonad cells. Cadmium was distributed amongthe three cell compartments but was mainly held at the cell surface(Fig. 3), with 90% of the initial Cd associated with the surface of thecells. Unlike Cu (Fig. 2), an essential nutrient, there was no ob-served shift in the percent distribution of Cd into the periplasmor cytoplasm affected by exposure concentration for Cd. The max-imum concentration observed in the cytoplasm was less than1 mmol Cd kg�1. Cell protection from Cd was maintained by sur-face interactions with limited transport into the cell under our as-say conditions.

The Langmuir isotherm described the overall association of Cdwith the periplasm and cytoplasm (Fig. 3 and Table 3), but thisequation did not adequately describe Cd adsorption onto the cellsurface, underpredicting the amount of Cd on the surface at thelower exposure concentrations. A two-site Langmuir model (Eq.(2)) better described the sorption at lower concentrations, andthe model well described the increasing sorption of Cd onto the cellsurface with increasing exposure concentrations. This finding wassupported by previous studies with another Gram-negative bacte-rium Escherichia coli where the two-site Langmuir model wasneeded to describe sorption of Cd and Pb (Kulczycki et al., 2002).

The use of the two-site model to describe the sorption of Cdonto the cell surface implies two types of surface sites with differ-ent capacities and affinities. At low surface coverage, Cd is at-tracted to the surface of the cells. The second site of retention ofCd on the cell surface is predicted to have high capacity (Table3), exceeding the number of sites on the cell determined by poten-tiometer titration by 16 times for isolate KT2440. Cadmium expo-sure increased bacterial production of extracellular polymericsubstances (EPS) (Guibaud et al., 2005; Henriques et al., 2007),and Ueshima et al. (2008) concluded that binding of Cd+ to theEPS provided protection to the cell by restricting its contact withthe outer membrane layer. For other studies (data not shown)we have isolated EPS from the KT2440 strain, and thus the cells un-der test should have a preexisting EPS layer. Similarly, siderophoresecreted by the cell could bind the Cd outside of the cell. Cd-in-duced synthesis of siderophore has been documented for P. aeru-ginosa as well as for Alcaligenes and Streptomycete isolates (Giliset al., 1998; Dimkpa et al., 2008; Sinha and Mukherjee, 2008). In-

duced polyphosphate could provide another binding site throughgenerating insoluble Cd phosphate (Pagès et al., 2007). In addition,proton antiporter–metal efflux systems for Cd could generate amore alkaline environment around the bacteria, promoting theprecipitation of Cd (Gilis et al., 1998). There was, however, no evi-dence of precipitates forming, and the low pH maintained through-out this study (pH 5.6) would preclude such an event, asdetermined by geochemical modeling (GEOCHEM). The observedenhanced surface sorption with increasing solution concentrationmay therefore be explained as a cell protective strategy possiblydue to cell surface modification.

Table 3Langmuir parameters, Q and b, for Cd associated with the surface, periplasmic space and cytoplasm of P. putida strains Corvallis and KT2440.

Location Q (mmol kg�1) b (L kg�1) Sum of square error r2 Percent

Corvallis Langmuir parameters (n = 18)Surface exchangeable site 1 51.5 2141 aa 991 0.980 95.4Surface exchangeable site 2 378 8.76Periplasmic space 19.4 220 c 203 0.785 4.3Cytoplasm 1.12 600 e 0.51 0.863 0.25

KT2440 Langmuir parameters (n = 18)Surface exchangeable site 1 30.4 4053 b 1945 0.959 99.5Surface exchangeable site 2 2800 1.08Periplasmic space 12.9 88.9 d 82.0 0.739 0.45Cytoplasm 1.16 386 e 0.50 0.872 0.04

a Q and b values followed by the same letter indicate that values were not significantly different for the two isolates within that cell compartment.

M.W. Pabst et al. / Chemosphere 81 (2010) 904–910 909

The two isolates differed in the extent of partitioning of Cd ontothe cell surface and into the periplasm as predicted by Q and b (Ta-ble 3). The Langmuir parameters, however, described the same dis-tribution of Cd into the cytoplasm for the two isolates. Retention ofCd on the cell surface or storage in the periplasm protects the cellby limiting uptake into the cytoplasm. The two isolates both lim-ited the uptake into the cell where the predicted maximum was1.12–1.16 mmol kg�1 (Table 3). Predictions of Q maxima showedmore Cd would be held in the periplasm for isolate Corvallis,whereas strain KT2440 would hold more on the surface (Table3). Although the strategies could be different they would resultin the same extent of protection.

4. Summary

This research has illustrated that sorption studies should con-sider the association of metals with the bacterial surface and theuptake of metals into the bacteria. For Cu this distribution changedwith solution concentration as the exposed bacteria altered mech-anisms of interactions with Cu for protection. The cell regulated Cuuptake and distribution to ensure adequate supply of Cu into thecell for metabolic use. At higher Cu concentrations, the cell re-stricted Cu uptake with more of the Cu held on the cell surface,thus protecting the cell from damage. In contrast, the majority ofCd was held on the surface so application of isotherm models withthe assumption of only surface reactions may be appropriate.Again, however, the living cell could alter mechanisms of sorptionwith increasing cell protection from Cd through efflux and/orincreasing surface functional groups. The increasing surface sorp-tion capacity for Cd with increasing solution concentration maybe a defense mechanism involving, for example, EPS production,a response not observed with Cu at the concentrations tested.

We hypothesized that the distribution of Cu and Cd would differbetween the two isolates, since Corvallis lacks the O-antigen chainson LPS, making the negatively charged LPS-core structures moreavailable. Indeed differences were illustrated with the potentio-metric titration, but they were not apparent in responses to Cu.Further studies are needed to define why protons bind differen-tially than Cu2+ or Cd2+ to the pseudomonad cell surface. Addition-ally, the roles that the different potential sites play (carboxyl andphosphodiesters, phosphomonoesters, and hydroxyl and amine,and sulfhydryl functional groups) remain to be determined.

Acknowledgements

This work was supported US EPA Science to Achieve Results(STAR) Program Grant RD83090701. Additional support providedby the Utah Water Research Laboratory and the Utah AgriculturalExperiment Station, Utah State University.

References

Aksu, Z., Donmez, G., 2006. Binary biosorption of cadmium(II) and nickel(II) ontodried Chlorella vulgaris: co-ion effect on mono-component isotherm parameters.Process Biochem. 41, 860–868.

Anderson, A.J., Jasalavich, C., 1979. Agglutination of pseudomonad cells by plantproducts. Physiol. Plant Pathol. 15, 149–159.

Bagai, I., Rensing, C., Blackburn, N.J., McEvoy, M.M., 2008. Direct metal transferbetween periplasmic proteins identifies a bacterial copper chaperone.Biochemistry 47, 11408–11414.

Bolster, C.H., Hornberger, G.M., 2007. On the use of linearized Langmuir equations.Soil Sci. Soc. Am. J. 72, 1796–1806.

Borrok, D., Fein, J.B., Kulpa, C.F., 2004. Proton and Cd adsorption onto naturalbacterial consortia: testing universal adsorption behavior. Geochim.Cosmochim. Acta 68, 3231–3238.

Borrok, D., Borrok, M.J., Fein, J.B., Kiessling, L.L., 2005a. Link between chemotacticresponse to Ni2+ and its adsorption onto the Escherichia coli cell surface. Environ.Sci. Technol. 39, 5227–5233.

Borrok, D., Turner, B.F., Fein, J.B., 2005b. A universal surface complexationframework for modeling proton binding onto bacterial surfaces in geologicsettings. Am. J. Sci. 305, 826–853.

Buell, C.R., Anderson, A.J., 1993. Expression of the aggA locus of Pseudomonas putidain vitro and in planta as detected by the reported gene, xyIE. Mol. Plant–Microbe Interact. 6, 331–340.

Campbell, P.G.C., Errecalde, O., Fortin, C., Hiriart-Baer, V.P., Vigneault, B., 2002. Metalbioavailability to phytoplankton-applicability of the biotic ligand model. Comp.Biochem. Physiol., C 133, 189–206.

Canovas, D., Cases, I., de Lorenzo, V., 2003. Heavy metal tolerance and metalhomeostasis in Pseudomonas putida as revealed by complete genome analysis.Environ. Microbiol. 5, 1242–1256.

Claessens, J., van Lith, Y., Laverman, A.M., Van Cappellen, P., 2006. Acid–base activityof live bacteria: implications for quantifying cell wall charge. Geochim.Cosmochim. Acta 70, 267–276.

Dimkpa, C.O., Svatos, A., Dabrowska, P., Schmidt, A., Boland, W., Kothe, E., 2008.Involvement of siderophores in the reduction of metal-induced inhibition ofauxin synthesis in Streptomyces spp.. Chemosphere 74, 19–25.

Gilis, A., Corbisier, P., Baeyens, W., Taghavi, S., Mergeay, M., van der Lelie, D., 1998.Effect of the siderophore alcaligin E on the bioavailiability of Cd to Alcaligeneseutrophus CH34. J. Ind. Microbiol. Biotechnol. 20, 61–68.

Guibaud, G., Comte, S., Bordas, F., Dupuy, S., Baudu, M., 2005. Comparison of thecomplexation potential of extracellular polymeric substances (EPS), extractedfrom activated sludges and produced by pure bacteria strains for cadmium, leadand nickel. Chemosphere 59, 629–638.

Guine, V., Spadini, L., Sarret, G., Muris, M., Delolme, C., Gaudet, J.P., Martins, J.M.F.,2006. Zinc sorption to three gram-negative bacteria: combined titration,modeling, and EXAFS study. Environ. Sci. Technol. 40, 1806–1813.

Hao, Z., Reiske, H., Wilson, D., 1999. Characterization of cadmium uptake inLactobacillus plantarum and isolation of cadmium and manganese uptakemutants. Appl. Environ. Microbiol. 65, 4741–4745.

Henriques, I.D.S., Aga, D.A., Mendes, P., O’Connor, S.K., Love, N.G., 2007. Metabolicfootprinting: a new approach to identify physiological changes in complexmicrobial communities upon exposure to toxic chemicals. Environ. Sci. Technol.41, 3945–3951.

Huang, Q., Chen, W., Xu, L., 2005. Adsorption of copper and cadmium by Cu- and Cd-resistant bacteria and their composites with soil colloids and kaolinite.Geomicrobiol. J. 22, 227–236.

Jordan, I.K., Natale, D., Galperin, M.Y., 2000. Copper chaperones in bacteria:association with copper-transporting ATPases. Trends Biochem. Sci. 25, 480–481.

Kola, H., Wilkinson, K.J., 2005. Cadmium uptake by a green alga can be predicted byequilibrium modeling. Environ. Sci. Technol. 39, 3040–3047.

Kulczycki, E., Ferris, F.G., Fortin, D., 2002. Impact of cell wall structure on thebehavior of bacterial cells as sorbents of cadmium and lead. Geomicrobiol. J. 19,553–565.

Langley, S., Beveridge, T.J., 1999. Effect of O-side-chain-lipopolysaccharidechemistry on metal binding. Appl. Environ. Microbiol. 65, 489–498.

910 M.W. Pabst et al. / Chemosphere 81 (2010) 904–910

Lewinson, O., Allen, A.T., Rees, D.C., 2009. A P-type ATPase importer thatdiscriminates between essential and toxic transition metals. Proc. Natl. Acad.Sci. U. S. A. 106, 4677–4682.

Loukidou, M., Karapantsios, T., Zouboulis, A.I., Matis, K., 2005. Cadmium(II)biosorption by Aeromonas caviae: kinetic modeling. Sep. Sci. Technol. 40,1293–1311.

Miller, C.D., Pettee, B., Zhang, C., Pabst, M., McLean, J.E., Anderson, A.J., 2009. Copperand cadmium: responses in Pseudomonas putida KT2440. Lett. Appl. Microbiol.49, 775–783.

Molina, M.A., Godoy, P., Ramos-González, M.I., Muñoz, N., Ramos, J.L., Espinosa-Urgel, M., 2005. Role of iron and TonB system in colonization of corn seeds androots by Pseudomonas putida KT2440. Environ. Microbiol. 7, 443–449.

Nelson, K.E., Weinel, C., Paulsen, I.T., Dodson, R.J., Hilbert, H., Martins dos Santos,V.A., Fouts, D.E., Gill, S.R., Pop, M., Holmes, M., Brinka, C.L., Beanan, M., DeBo,R.T., Daugherty, S., Kolonay, J., Madupu, R., Nelson, W., White, O., Peterson, J.,Khouri, H., Hance, I., Lee, P.C., Holtzapple, E., Scanlan, D., Tran, K., Moazzez, A.,Utterback, T., Rizzo, M., Lee, K., Kosack, D., Moestl, D., Wedler, H., Lauber, J.,Stjepandic, D., Hoheisel, J., Straetz, M., Heim, S., Kiewitz, C., Eisen, J.A., Timmis,K.N., Düsterhöft, A., Tümmler, B., Fraser, C.M., 2002. Complete genome sequenceand comparative analysis of the metabolically versatile Pseudomonas putidaKT2440. Environ. Microbiol. 4, 799–808.

Nies, D.H., 1999. Microbial heavy-metal resistance. Appl. Microbial. Biotechnol. 51,730–750.

Nusslein, K., Maris, D., Timmis, K., Dwyer, D.F., 1992. Expression and transfer ofengineered catabolic pathways harbored by Pseudomonas spp. introduced intoactivated sludge microcosms. Appl. Environ. Microbiol. 58, 3380–3386.

Pagès, D., Sanchez, L., Conrod, S., Gidrol, X., Fekete, A., Schmitt-Kopplin, P., Heulin, T.,Achouak, W., 2007. Exploration of intraclonal adaptation mechanisms ofPseudomonas brassicacearum facing cadmium toxicity. Environ. Microbiol. 9,28020–22835.

Pagnanelli, F., Esposito, A., Toro, L., Veglio, F., 2002. Copper and cadmiumbiosorption onto Sphaerotilus natans: application and discrimination ofcommonly used adsorption models. Sep. Sci. Technol. 37, 677–699.

Parker, D.R., Norvell, W.A., Chaney, R.L., 1995. GEOCHEM-PC: A chemical speciationprogram for IBM and compatible personal computers. In: Loeppert, R.H.,Schwab, A.P., Goldberg, S. (Eds.), Chemical Equilibrium and Reaction Models.SSSA Spec. Publ. 42: ASA and SSSA, Madison, WI, pp. 253–269.

Pearson, R.G., 1963. Hard and soft acids and bases. J. Am. Chem. Soc. 85, 3533–3539.Salt, D.E., Pickering, I.J., Prince, R.C., Gleba, D., Dushenkov, S., Smith, R.D., Raskin, I.,

1997. Metal accumulation by aquacultured seedlings of Indian mustard.Environ. Sci. Technol. 31, 1636–1644.

Silver, S., 1998. Genes for all metals – a bacterial view of the periodic table: the 1996Thom Award Lecture. J. Ind. Microbiol. Biotechnol. 20, 1–12.

Sinha, S., Mukherjee, S.K., 2008. Cadmium-induced siderophore production by ahigh Cd-resistant bacterial strain relieved Cd toxicity in plants through rootcolonization. Curr. Microbiol. 56, 55–60.

Turner, B.F., Fein, J.B., 2006. Protofit: a program for determining surface protonationconstants from titration data. Comput. Geosci. 32, 1344–1356.

Ueshima, M., Ginn, B.R., Haack, E.A., Szymanowski, J.E.S., Fein, J.B., 2008. Cdadsorption onto Pseudomonas putida in the presence and absence ofextracellular polymeric substances. Geochim. Cosmochim. Acta 72, 5885–5895.

Vaara, M., 1992. Agents that increase the permeability of the outer membrane.Microbiol. Rev. 56, 395–411.

Wu, S.C., Peng, X.L., Cheung, K.C., Liu, S.L., Wong, M.H., 2009. Adsorption kinetics ofPb and Cd by two plant growth promoting rhizobacteria. Bioresour. Technol.100, 4559–4563.

Zouboilis, A.I., Loukidou, M., Matis, K., 2004. Biosorption of toxic metals fromaqueous solutions by bacteria strains isolated from metal-polluted soils.Process Biochem. 39, 909–916.