an x-ray absorption fine structure study of au adsorbed onto the …fein/publications/song...

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 86 (2012) 103–117

An X-ray Absorption Fine Structure study of Au adsorbedonto the non-metabolizing cells of two soil bacterial species

Zhen Song a, Janice P.L. Kenney b, Jeremy B. Fein b, Bruce A. Bunker a,⇑

a Dept. of Physics, University of Notre Dame, Notre Dame, IN 46556, USAb Dept. of Civil Engineering & Geological Sciences, University of Notre Dame, Notre Dame, IN 46556, USA

Received 11 May 2011; accepted in revised form 28 February 2012; available online 14 March 2012

Abstract

Gram-positive and Gram-negative bacterial cells can remove Au from Au(III)–chloride solutions, and the extent ofremoval is strongly pH dependent. In order to determine the removal mechanisms, X-ray Absorption Fine Structure (XAFS)spectroscopy experiments were conducted on non-metabolizing biomass of Bacillus subtilis and Pseudomonas putida with fixedAu(III) concentrations over a range of bacterial concentrations and pH values. X-ray Absorption Near Edge Structure(XANES) and Extended X-ray Absorption Fine Structure (EXAFS) data on both bacterial species indicate that more than90% of the Au atoms on the bacterial cell walls were reduced to Au(I). In contrast to what has been observed for Au(III)interaction with metabolizing bacterial cells, no Au(0) or Au–Au nearest neighbors were observed in our experimental sys-tems. All of the removed Au was present as adsorbed bacterial surface complexes. For both species, the XAFS data suggestthat although Au–chloride–hydroxide aqueous complexes dominate the speciation of Au in solution, Au on the bacterial cellwall is characterized predominantly by binding of Au atoms to sulfhydryl functional groups and amine and/or carboxyl func-tional groups, and the relative importance of the sulfhydryl groups increases with increasing pH and with decreasing Au load-ing. The XAFS data for both microorganism species suggest that adsorption is the first step in the formation of Aunanoparticles by bacteria, and the results enhance our ability to account for the behavior of Au in bacteria-bearing geologicsystems.� 2012 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Microorganisms play an important role in the biogeo-chemical cycling of Au, and biomineralization of Au hasbeen suggested as a mechanism responsible for the forma-tion of secondary Au deposits in soil environments (e.g.,Watterson, 1992; Bischoff, 1994, 1997; Falconer et al.,2006; Reith et al., 2007). Studies of Au(I) and Au(III) bio-accumulation by a wide range of types of microorganisms(bacteria, fungi and yeasts) indicate that the extent of Auremoval from solution decreases with increasing pH, thatbacteria show a greater affinity for Au than do the othertypes of organisms studied, and that Au uptake can be re-

0016-7037/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.gca.2012.02.030

⇑ Corresponding author.E-mail address: [email protected] (B.A. Bunker).

versed by exposure of the biomass to a 0.1 M thioureasolution (Nakajima, 2003; Tsuruta, 2004). A number of lab-oratory studies have observed Au(0) nanoparticle forma-tion from Au(I) and Au(III) solutions in the presence ofmetabolizing bacteria (Beveridge and Murray, 1976; Kash-efi et al., 2001; Karthikeyan and Beveridge, 2002; Lengkeand Southam, 2005, 2006; Reith et al., 2005; Lengkeet al., 2006a,b; Reith et al., 2006; Lengke et al., 2007). How-ever, the mechanisms of Au reduction and precipitation,and the potential role of Au adsorption onto the cells in thisprocess, are not well understood.

Previous XAFS studies have investigated the valencestate and speciation of Au in aqueous solutions (Fargeset al., 1993; Pokrovski et al., 2009a,b) and on mineral andmicrobial surfaces (Berrodier et al., 2004; Lengke et al.,2007). Farges et al. (1993) studied the speciation of Au(III)in aqueous chloride solutions, and demonstrated that

104 Z. Song et al. / Geochimica et Cosmochimica Acta 86 (2012) 103–117

Au(III) is present as 4-coordinated, square–planar com-plexes. At low pH, AuCl�4 is the dominant complex, andwith increasing pH, chloride ligands are replaced byhydroxide ligands. Pokrovski et al. (2009a,b) studied Aucomplexation and speciation in chloride- and in sulfur-bearing hydrothermal fluids. Their XANES and EXAFSresults suggest that in Au–chloride systems, AuðIIIÞCl�4 isthe dominant Au species at low temperature (<150 �C),while at higher temperatures, Au(III) reduces to Au(I), inthe form of a linear AuðIÞCl�2 complex, and to Au(0). Fur-thermore, in sulfur-bearing hydrothermal fluids underacidic and neutral-to-basic conditions, the atomic environ-ment of dissolved Au involves two sulfur atoms in a lineargeometry around Au(I) at similar atomic distances, with anaverage distance of 2.29 ± 0.01 A. Berrodier et al., 2004investigated Au(III) adsorbed from chloride solutions ontoferrihydrite, goethite, and boehmite mineral surfaces,reporting that Au(III) is adsorbed onto these surfaces dom-inantly as inner-sphere, square–planar complexes. The sur-face speciation of Au on these surfaces varies with pH, withAu(III)O4 surface complexes at pH >6 and Au(III)(O,Cl)4

at pH <6, and with the number of Cl ligands increasingwith decreasing pH. In this study, Au(I) and Au(0) werenot observed. Lengke et al. (2007) found precipitation ofAu(0) nanoparticles from Au(III)–chloride solution bymetabolizing cyanobacteria, and indicated that organic sul-fur released from cyanobacteria bound with the reducedAu(I) to form an intermediate Au(I)-sulfide species beforeAu(0) nanoparticles formed. These previous studies indi-cate that at ambient temperature, Au(III) adsorbed ontothe mineral surfaces studied does not change valence state,but that interaction of metabolizing bacteria with aqueousAu(III) can lead to Au reduction to Au(I) and eventuallyto Au(0) by the biomass. Although it is clear that bacteriacan cause Au reduction, the role of adsorption in this pro-cess and the molecular binding environments of Au on bac-terial cell walls have not been determined.

In this study, we measured the valence state and specia-tion of the Au that was removed from Au(III)–chloridesolutions by non-metabolizing common soil bacteria: theGram-positive species Bacillus subtilis, and the Gram-nega-tive species Pseudomonas putida. We conducted the XAFSexperiments as a function of pH at two different Au load-ings. Our objectives were to determine the functionalgroups involved in the initial binding of Au to the bacterialcell walls, and to determine whether the binding mechanismchanges as a function of Au loading and pH.

2. MATERIALS AND METHODS

2.1. Bacteria growth and sorption experiments

B. subtilis and P. putida bacterial cells were incubatedaerobically in 3 mL of trypticase soy broth (TSB) growthmedium with the addition of 0.5% yeast extract at 32 �Cfor 24 h, and then transferred to 2 L of the same growthmedium and incubated for another 24 h. The bacteria wereharvested by centrifugation at 8100g for 30 min, then trans-ferred to test tubes, and washed five times in 0.1 M NaClO4

electrolyte. After the final wash, the bacterial pellet was

transferred to a pre-weighed test tube and resuspended withsmall amount of clean 0.1 M NaClO4 electrolyte. The sus-pension was centrifuged at 8100g for two 30-min intervals,decanting the bacteria-free supernatant after each interval,and the wet mass of the bacterial pellet was determined.

All Au standards and samples were made from diluting1000 ppm Au ICP-OES standard solution (1000 ± 3 ppmAu(III) in 10% HCl) used for inductively coupled plasma-optical emission spectroscopy (ICP-OES) (Assurancegrade; Spex Certiprep). Suspensions of non-metabolizingB. subtilis and P. putida cells in 0.1 M NaClO4 or 0.1 MNaNO3 (see the Supplementary material) were exposed to5 ppm Au(III)–chloride solutions for 2 h. The pH valuesof the suspensions were adjusted using small quantities of0.1–1.0 M HCl or NaOH and measured using a ThermoOrion model 420A bench-top pH meter. The pH of eachexperimental solution was monitored every 30 min, and ad-justed if required. The final pH was measured after 2 h oftotal reaction time and reported in Table 1. The chlorideconcentration is approximately 500 ppm in the suspensions.Sixteen biomass samples were prepared as described in Ta-ble 1. The pH values were chosen to represent different Auremoval behavior regimes, based on the results from Nak-ajima (2003) and Tsuruta (2004).

After the bacterial suspensions were mixed for 2 h, theywere centrifuged at 8100g for 10 min. The supernatant wasfiltered through a 0.45 lm disposable nylon membrane, andthe concentration of Au remaining in solution was analyzedusing ICP-OES. The concentration of adsorbed Au was cal-culated by the difference in concentration of Au in the start-ing solution and the final measured concentration of Au inthe supernatant. Biomass-free controls were conducted anddemonstrated that no significant Au adsorption onto thepolypropylene reaction vessels occurred.

2.2. XAFS spectroscopy

The wet biomass pellets from the batch sorption exper-iments were collected and maintained in an ice-cooled ves-sel during transport to the beamline. All X-ray absorptionspectroscopy measurements were made within 48 h of sam-ple preparation. Fluorescence Au LIII-edge (11,919 eV)XAFS measurements were made on the biomass pelletswhich were loaded into slotted Teflon holders and coveredwith Kapton film. The XAFS measurements were per-formed at the MRCAT sector 10-ID beamline (Segreet al., 2000) at the Advanced Photon Source at ArgonneNational Laboratory.

The energy of the incident X-ray beam was scanned byusing a Si (111) reflection plane of a cryogenically-cooleddouble-crystal monochromator. The undulator was tunedto its second harmonic, and tapered to an X-ray energyspread of approximately 3.5 keV to reduce the variationin the incident intensity to less than 15% over the wholescanned energy range. Higher harmonics were rejectedusing a Rh-coated mirror.

The incident and transmitted X-ray fluxes were mea-sured by ion chambers and the Au LIIIa,b fluorescencewas monitored by either a 5-grid Stern–Heald–Lytle detec-tor or a 4-element Vortex (ME-4) detector. The incident ion

Table 1List of Au biomass samples for XAFS analysis.

Sample IDa BacterialConcentration(wet mass) g/L

Bacterial species pH

AuB1 1.0 B. subtilis 3.0AuB2 1.0 B. subtilis 4.3AuB3 1.0 B. subtilis 5.2AuB4 1.0 B. subtilis 5.9AuB5 7.0 B. subtilis 3.4AuB6 7.0 B. subtilis 4.6AuB7 7.0 B. subtilis 5.5AuB8 7.0 B. subtilis 6.6AuP1 1.0 P. putida 2.8AuP2 1.0 P. putida 4.0AuP3 1.0 P. putida 5.4AuP4 1.0 P. putida 7.0AuP5 1.0 P. putida 7.4AuP6 7.0 P. putida 3.2AuP7 7.0 P. putida 4.9AuP8 7.0 P. putida 5.9

a Initial Au concentration in solution was 5 ppm for all sixteenbiomass samples.

Z. Song et al. / Geochimica et Cosmochimica Acta 86 (2012) 103–117 105

chamber was filled with 33% nitrogen gas and 67% heliumgas. The transmitted and reference ion chambers were filledwith 100% nitrogen gas. The fluorescence detector in theStern–Heald geometry (Stern and Heald, 1983) was filledwith argon gas. The signal/background ratio was high en-ough that no X-ray filter was necessary. The incident X-ray beam profile was 0.6 mm2. Linearity tests (Kemneret al., 1994) indicated less than 0.05% nonlinearity for a50% decrease in incident X-ray intensity. The XAFS datawere aligned using simultaneously collected Au foil data.Au(III)–hydroxide–chloride (2 mM) solutions with pH3.0, 5.0, 7.0, 9.0 and 11.0, a series of powder standard com-pounds detailed in Section 3.2.1, and a 5 lm thick Au foilwere measured under the same beamline conditions. Thestandard compounds were acquired from Alfa Aesar Com-pany, Ward Hill, MA and Sigma–Aldrich Inc., St. Louis,MO, and the purity of the compounds was no less than99.9%.

By monitoring the XANES of consecutive scans on thesame spot on the sample, we found that photo-reductionof the room-temperature biomass samples occurred withintwenty seconds during exposure to the X-ray beam, leadingto reduction of Au(III) to Au(0) during the measurement.To minimize this radiation damage, the samples werequick-frozen with liquid nitrogen (�196 �C). Fast-scanXANES spectra of all the samples were measured at bothroom temperature and liquid nitrogen temperature. Bycomparing XANES spectra taken at room temperatureand liquid nitrogen temperature for each sample, we foundthat freezing the samples did not change the valence state orlocal atomic environment, and that photo-reduction did notappear until after 5 min of X-ray exposure in the frozensamples. Therefore, all XAFS data used in analysis of bio-mass samples were taken at liquid nitrogen temperature. Tokeep the measurement conditions consistent with the Au-biomass samples, we also used the data from the low-tem-perature solution samples in the analysis.

Quick scans (continuous scanning of the monochroma-tor with data sampled every 1 eV in the entire scanningrange) were used with an integration time of 0.05 s per datapoint. Multiple 0.6 mm2 spots on the samples were mea-sured. Five XANES scans (scan range from 100 eV belowto 100 eV above the Au LIII edge) were performed at thefirst spot of each sample. Three EXAFS scans (scan rangefrom 150 eV below to 790 eV above the Au LIII edge) wereperformed at the remaining spots (the number of spots var-ies between 12 and 25 depending on the specific sample toobtain required data quality) of each sample.

2.3. XAFS spectra analysis

XANES spectra provide information on the valencestate of Au in our samples. The spectrum of Au(III) has asharp pre-edge feature and the edge positions (determinedby the position of the maximum of dl/dE) for the Au(I)and Au(III) standards differ by approximately 4 eV (Leng-ke et al., 2006b), making it straightforward to distinguishthese two cases. Additionally, Au(0) has two characteristicpeaks on XANES spectra at approximately 11,947 and11,970 eV. To conduct a semi-quantitative analysis, linearcombination fits of Au aqueous standards, powder stan-dards and Au metal XANES spectra were performed onthe XANES data from the Au-biomass samples.

EXAFS data were analyzed using codes from the UW-XAFS package (Stern et al., 1995). Photo-reduction wasobserved for the third or fourth scan on the same spot ofeach sample. By comparing XANES spectra, those scanswithout radiation damage were selected, averaged, normal-ized and background subtracted using ATHENA (Raveland Newville, 2005), which is a graphical interface to IFEF-FIT (Newville, 2001). Background subtraction used theAUTOBK method with Rbkg, the maximum frequency ofthe background, set to 1.1 A. The Fourier transform rangein k space was 2.0–10.2 A�1, and symmetrical Hanning win-dow function was used with dk = 1.0 A�1 (Newville et al.,1993). Theoretical amplitudes were calculated using Feff6(Rehr et al., 1992) and experimental data were fit usingIFEFFIT. Four of the calculated single-scattering paths,(Au–N, Au–O, Au–S, and Au–C) were used to fit the spec-tra from the biomass samples, and in order to exclude thepossibility of some minor Au–phosphoryl binding, or Aunanoparticle formation, additional Au–P and Au–Au scat-tering paths were allowed into the fitting model in selectedfits. Significant reduction (>50%) in “reduced chi square”

(v2m) and R factor values were used as the standard for the

better data-model fitting. The v2m is defined as the normal

statistical v2 divided by the number of degrees of freedomm in the fit, where m is given by the number of “independentpoints” (determined by the data range used in the analysis)in the data minus the number of parameters allowed float inthe fit (Bunker, 2010). The R factor is the sum of the differ-ences of the model and data value squared divided by thesum of the data value squared. The fitting range was setto 1.15–2.9 A for all the Au biomass sample data sets and1.15–4.5 A for all the standards. In order to break correla-tions between fitting parameters and to reduce the possibil-ity of obtaining wrong fitting results from a single k-


weighting value, simultaneous fitting of data sets with mul-tiple k-weighting (k1,k2,k3) was performed (Kelly et al.,2002).

3. RESULTS AND DISCUSSION

3.1. Au sorption data

As can be seen in Fig. 1, the extent of Au adsorptionvaries strongly as a function of pH. In all sample sets stud-ied except for samples with 7.0 g/L P. putida cells, the ex-tent of adsorption was nearly 100% at pH �3.0, starteddecreasing at pH �4.5, and decreased dramatically betweenpH �5.5 and 7.5 with increasing pH. These observationsare consistent with anion-like adsorption behavior. Thebacterial cell wall becomes more negatively charged withincreasing pH, and it is likely that increasing electrostaticrepulsion between the negatively charged Au(III)–hydrox-ide–chloride aqueous complexes and the bacteria is respon-sible for the decrease in adsorption that we observed withincreasing pH. As for the samples with 7.0 g/L P. putida

cells, the extent of adsorption was approximately 75% atpH �3.0, decreased at pH �5.0 and increased to nearly80% at pH �6.0. This unusual adsorption behavior maybe due to elevated concentrations of bacterial exudatesfrom the higher P. putida cell concentration in these exper-iments. However, we did not measure dissolved organic car-bon concentrations in these samples and thereforeadditional measurements would be required in order to re-solve this issue.

3.2. Analysis of XANES spectra

3.2.1. XANES for Au solution standards, powder compounds

and Au foil

As a basis for interpretation of our Au-biomass data,XANES spectra were collected at room temperature forthe following standards: Au(III)–chloride (HAuCl4�3H2O),Au(III)–acetate (Au(O2CCH3)3), Au(I)–sulfide (Au2S),Au(I)–thiosulfate (Na3Au(S2O3)2�2H2O), Au(I) thiomalate

Fig. 1. The percentage of Au removed from solution after 2 h ofexposure to 1.0 and 7.0 g/L wet mass of B. subtilis and P. putida

using a total of 5 ppm Au in 0.1 M NaClO4.

(NaO2CCH2CH(SAu)CO2Na�xH2O), Au(I)–chloro(tri-phenylphosphine) (C18H15AuClP) and 5 lm thick Au foil(Fig. 2a). The three stable oxidation states of Au: Au(0),Au(I), and Au(III) (Patai and Rappoport, 1999) are in-cluded in these compounds. A sharp XANES feature char-acteristic of Au(III) compounds was observed and theabsorption edge of Au(III) compounds shifted to about4 eV lower than Au(I) compounds and Au foil. This sharppre-edge feature and the shifted absorption edge of Au(III)are known to be caused by a dipole-allowed atomic 2p to 5dtransition. This transition is allowed for the atomic d8 stateof Au(III) but is forbidden for the d10 state of Au(I) andAu(0) (Pantelouris et al., 1995). Additionally, the post-edgepeak at approximately 11,947 eV on Au foil XANES spec-tra is a clear signature of Au(0). We use these XANES fea-tures to determine the oxidation states of Au in the biomasssamples. XANES spectra of Au(I)–sulfide, Au(I)–thiosul-fate and Au(I)–thiomalate are nearly identical, suggestinga similar Au–S local atomic environment in these com-pounds. The feature located at �11,930 eV represents mul-tiple scattering from the S–Au–S linear structure (Elder andEidsness, 1987; Bau, 1998; Pokrovski et al., 2009b), and thisfeature can be used as a “fingerprint” of this structure.

A series of 2 mM Au(III)–chloride solutions of pH 3.0,5.0, 7.0, 9.0 and 11.0 was studied. The Au LIII edge XANESspectra of the solution standards Fig. 2b) show that Au spe-ciation in solution changes with pH, and the valence stateof Au remains as Au(III) in all samples studied. For solu-tions at pH 3.0 and 5.0, the XANES spectra look nearlyidentical; the multiple-scattering feature at 11,934 eV is inagreement with Au being present as the square planarAuCl�4 anion (Farges et al., 1993), indicating that the dom-inant Au species in these two solutions is AuCl�4 . For solu-tions at pH 7.0 and 9.0, the multiple-scattering featurerepresenting AuCl�4 decreases slightly in magnitude andthe multiple-scattering feature at 11,945 eV which is relatedto the Au–OH environment (Berrodier et al., 2004) in-creases. These changes indicate mixed Au speciation ofAu–Cl and Au–OH complexes under these pH conditions.For the solution at pH 11.0, the multiple-scattering featurerepresenting AuCl�4 disappears and the multiple-scatteringfeature at 11,945 eV indicates that the dominant Au speciesin the solution is AuðOHÞ�4 . These results are in agreementwith the study by Farges et al. (1993) of Au speciation inchloride solutions with higher Au concentrations.

3.2.2. XANES for Au-biomass samples

The Au-L3 XANES spectra of the Au biomass samples(Fig. 2c and d) confirm the change in Au-speciation andAu valence state as Au is removed from solution. The ab-sence of the characteristic Au(0) peaks at approximately11,947 eV on all Au-biomass XANES spectra indicates thatAu(0) represents less than a few percent of the total Au inthese samples, and that virtually all of the Au atoms arepresent as adsorbed bacterial surface complexes.

A comparison between the XANES spectra of the stan-dard Au compounds and those of the Au biomass samples(Fig. 2e) suggests significant Au–S binding in all of the bio-mass samples, and that the Au–S signal is especially strongin samples in Group B. The multiple-scattering feature

Fig. 2. XANES spectra of (a) standard Au powder compounds and Au foil, (b) 2 mM Au(III)–chloride solutions at pH 3.0, 5.0, 6.0, 7.0, 9.0and 11.0, (c) B. subtilis biomass samples, (d) P. putida biomass samples, and (e) comparison between compounds and biomass samples.


located at �11,930 eV in all Au biomass samples likely rep-resents the S–Au–S linear structure that we observed in theAu(I)-sulfide, Au(I)-thiosulfate and Au(I) thiomalate com-pounds. No significant multiple-scattering feature is seenat 11,934 eV (representing the square planar AuCl�4 com-plex) or at 11,945 eV (representing the AuðOHÞ�4 complex),

suggesting that there is negligible Au–Cl or Au–OH bindingin the biomass samples.

The sharp pre-edge features of the Au-biomass XANESspectra divide the Au-biomass samples into two groups.For some samples (here denoted Group A, and includingsamples AuB1, AuB2, AuB3, AuB4, AuP1, AuP2, and


AuP3), the small pre-edge feature suggests a small amountof Au(III) atoms in the sample, while the edge position(approximately 11,923 eV) implies that the majority of Auatoms are present as Au(I). Linear combination fits withXANES spectra of Au(I) thiomalate (to represent Au(I)),Au(III)–chloride solution at pH 11.0 (to represent Au(III)),and 5 lm Au foil (to represent Au(0)) in the energy range11,900–11,960 eV indicate that, in all the Group A Au-bio-mass samples, more than 90% of the Au atoms are presentas Au(I), and that the remaining Au atoms are present onthe biomass as Au(III). For the remaining samples (GroupB, including samples AuB5, AuB6, AuB7, AuB8, AuP4,AuP5, AuP6, AuP7, and AuP8), linear combination fitswith XANES spectra of the same Au standards in the en-ergy range 11,900–11,960 eV indicate that virtually all theAu atoms are present as Au(I).

3.3. Analysis of EXAFS spectra for Au solution and Au(I)

thiomalate powder standards

The fitting results for the standards serve as a founda-tion for interpreting the EXAFS spectra of the Au-biomasssamples in this study. The fitting results for the Au(III)–chloride solutions and the Au(I) thiomalate powder areshown in Table 2. The EXAFS spectra of the Au(III)–chlo-ride solutions at pH 3.0 and 5.0 were fit with four Cl atomsin a square planar geometry about a central Au atom. TheAu–Cl single-scattering path (path length = 2.27 A) andmultiple-scattering Au–Cl–Cl–Au and Au–Cl–Au–Cl paths(path length = 4.56 A) were included in the fitting process.The Au–Cl bond length was found to be 2.27(±0.01) A; theamplitude reduction factor S2

0 was 0.85(±0.05); and the De-bye–Waller factor r2 (variation in average radial distance)was 0.002(±0.001) A2. The EXAFS spectrum of theAu(III)–chloride solution at pH 11.0 was fit with four Oatoms in square planar geometry about Au. The single-scat-tering Au–O path (path length = 1.97 A) and multiple-scat-tering Au–O–O–Au and Au–O–Au–O paths (pathlengths = 3.94 A) were included in the fitting process. TheAu–O bond length was found to be 1.96(±0.01) A; S2

0 wascalculated to be 0.76(±0.05); and the Debye–Waller factorr2 was 0.002(±0.002) A2. The EXAFS spectra of theAu(III)–chloride solutions at pH 7.0 and 9.0 were fit witha linear combination of Au–O and Au–Cl single-scatteringand multiple-scattering paths. Constraints on S2

0 and r2

were based on the results of previous fittings of the Au–Cl and Au–O bonds. These results are consistent with thosereported by Farges et al. (1993).

The EXAFS data of the Au(I) thiomalate standard wasfit with two S atoms in a linear geometry with respect toAu. A single-scattering Au–S (path length = 2.30 A) andmultiple-scattering Au–S–S–Au and Au–S–Au–S paths(path lengths = 4.60 A) were included in the fitting process.The Au–S bond length was found to be 2.30(±0.01) A; S2

0

was found to be 0.78(±0.07); and the Debye–Waller factorr2 was found to be 0.002(±0.001) A2. These results are con-sistent with previously reported values (Elder and Eidsness,1987; Bau, 1998; Pokrovski et al., 2009b).

Comparing the EXAFS oscillations as a function ofelectron wavenumber (Fig. 3a) and the Fourier transformed

data (Fig. 4a) for Au bound to O, Cl and S in the first shell,a phase difference between Au bound to O and Au bound toCl and S can be clearly seen. However, because Cl and Sdiffer in atomic number by just one, the backscatteringamplitudes from these two atoms are similar, making it dif-ficult to distinguish signals of Au–Cl from signals of Au–Sbinding in an unknown sample. Fortunately, the multiple-scattering XANES features (Durham et al., 1982) locatedat 11,930 and 11,934 for linear Au(I)–S binding and squareplanar Au(III)–Cl binding (Fig. 2e) can be used to distin-guish between Au–Cl and Au–S binding in a sample. Inaddition, the Au(I)–Cl complex is not stable at room tem-perature and will disproportionate to a mixture ofAu(III)–chloride and Au metal (Gammons et al., 1997; Pa-tai and Rappoport, 1999; Lengke et al., 2006a,b). There-fore, Au(I)-Cl can be excluded from models of ourexperimental system.

3.4. Au binding to biomass: EXAFS qualitative analysis

3.4.1. B. subtilis biomass samples (AuB)

The magnitude and real part of the Fourier transformed(FT) data of the Au – B. subtilis biomass samples (AuB) fortwo different Au/bacteria ratios and four different pH val-ues are shown in Fig. 4c and d. As can be seen by compar-ing the magnitude and real part of the FT data of thebiomass samples and the solution standards at the samepH values (Fig. 5), the amplitude and position of the firstpeak differ significantly. The amplitude of the first peak inFT data of the solution standards (Fig. 4a) decreases andthe peak position shifts to smaller radial distance valuesas pH increases. In contrast, the amplitude of the first peakin the FT data of the biomass samples increases and thepeak position shifts to larger radial distance values as pHincreases, indicating significant differences in the local coor-dination environment of Au between the biomass samplesand the Au chloride solutions. For samples with 1.0 g/Lbacteria, it can be clearly seen that the FT first shell ampli-tude increases and it shifts towards higher distances as pHincreases. These changes with pH can be attributed to a dif-ferent percentage of Au atoms bound to different functionalgroups under different pH conditions on bacterial cell walls.

Previous studies have demonstrated that metal sorptiononto non-metabolizing bacteria is dominated by carboxyl,phosphoryl and sulfhydryl functional groups on the cells(Hennig et al., 2001; Kelly et al., 2002; Boyanov et al.,2003; Toner et al., 2005; Guine et al., 2006; Mishra et al.,2009). Considering that 12% (by dry weight) of a typicalbacterial cell consists of nitrogen, an important element inproteins (Madigan et al., 2002), nitrogen-containing func-tional groups such as amine groups may also play a roleadsorbing Au atoms in our experiment systems. Just as itis difficult to distinguish between Au–Cl and Au–S bondsusing EXAFS, it is also difficult to distinguish Au–O fromAu–N bonds in an unknown sample. Therefore, Au–car-boxyl, –phosphoryl, and –amine bonds are likely to exhibitsimilar first shell amplitudes and peak positions in the FTdata. Thus, the changes of the amplitude of the first peakin the FT data and the peak position as pH increases seenin the EXAFS data could indicate an increase of sulfhydryl

Table 2Fit results for Au(III)–hydroxide–chloride (2 mM) solutions and Au(I) thiomalate powder.

Standard Path R (A) N S20 r2 (A2) E0 (eV) v2

m R factor

Auaqa pH3 Au–Cl 2.27 ± 0.01 4b 0.85 ± 0.05 0.002 ± 0.001 7.2 ± 0.6 43 0.007Auaqa pH5 Au–Cl 2.27 ± 0.01 4b 0.84 ± 0.05 0.002 ± 0.001 7.2 ± 0.6 43 0.007Auaqa pH7 Au–O 1.96 ± 0.01 1.44 ± 0.22c 0.76d 0.002d 6.5d 38 0.018

Au–Cl 2.27 ± 0.01 2.56 ± 0.13c 0.85d 0.002d 7.2d

Auaqa pH9 Au–O 1.96 ± 0.01 2.14 ± 0.10c 0.76d 0.002d 6.5d 38 0.018Au–Cl 2.27 ± 0.01 1.86 ± 0.17c 0.85d 0.002d 7.2d

Auaqa pH11 Au–O 1.96 ± 0.01 4b 0.76 ± 0.05 0.002 ± 0.002 6.5 ± 0.6 36 0.002Au(I) thiomalate Au–S 2.30 ± 0.01 2e 0.78 ± 0.07 0.002 ± 0.001 6.8 ± 1.2 39 0.002

a Represents Au(III)–hydroxide–chloride (2 mM) solutions.b Coordination number of Au–Cl and Au–O was fixed to 4 in the fitting.c The sum of coordination numbers of Au–O and Au–Cl was fixed to 4 in the fitting.d S2

0, r2, and E0 were fixed to best fit values obtained from pH 3, 5 and 11 solution standards in the fitting.e Coordination number of Au–S was fixed to 2 in the fitting.


binding (Au–S), which agrees with the XANES results.Moreover, since the chloride concentration in the solutionused in the batch adsorption experiments is high, and EX-AFS spectroscopy alone is unable to distinguish betweenAu–Cl and Au–S bonds, we cannot completely rule outthe possibility of a small amount of Au–Cl binding on thebacteria.

FT data for samples with 7.0 g/L bacteria (lower Auloading) show similar changes with respect to pH, but areless pronounced than we observed for the samples with1.0 g/L bacteria. Comparing samples with different Auloading at the same pH indicates that the first peak ampli-tude of the FT data of samples with lower Au loading in-creases and shifts to larger distance compared to sampleswith higher Au loading. This result suggests that a higherpercentage of Au atoms are bound to sulfhydryl groups un-der conditions of lower Au loading on bacterial cells rela-tive to the higher loading conditions.

3.4.2. P. putida biomass samples (AuP)

The magnitude and real part of the FT data of P. putida

biomass samples (AuP) at two different Au/bacteria ratiosand at four different pH values are shown in Fig. 4e andf. Comparisons of the FT data of the two bacterial speciesshow similar peaks and similar trends of the changes inpeak amplitude and position, suggesting similar atomicbinding environments between these the two bacteria.Additionally, at the same pH and Au loading, the magni-tude of the first peak of the FT data for P. putida samplesAuP1, AuP2, and AuP3 (1.0 g/L bacteria under low pHconditions) increases and the position shifts to higher radialdistance values than those of the corresponding B. subtilis

(AuB) samples, indicating a higher percentage of Au atomsadsorbed to sulfhydryl groups in the AuP samples relativeto the AuB samples under the same conditions. The EX-AFS spectra of AuP samples 4, 5, (1.0 g/L bacteria underhigher pH conditions) 6, 7, and 8 (7.0 g/L bacteria) arenearly identical to each other as well as to the spectrumof the Au(I) thiomalate standard, suggesting significantsulfhydryl binding in these samples. There are three reso-nance signals that can be clearly seen in the magnitude of

the FT data of P. putida biomass samples (AuP1–AuP5)(see the Supplementary material) with a larger k range(2.0–14.0 (A�1)). These three resonance peaks are verylikely to represent the mixed N/O and S first coordinationshells and second C shell.

The qualitative analysis of the biomass samples indicatesthat, for all AuB samples and for AuP samples 1, 2 and 3,Au atoms are bound to a mixture of amine/carboxyl/phos-phoryl and sulfhydryl functional groups on the bacterialcells, and the relative number of Au-bound sulfhydrylgroups increases with increasing pH. For the rest of theAuP samples, sulfhydryl groups appear to be the dominantbinding site.

3.5. Au binding to biomass: EXAFS quantitative analysis

3.5.1. B. subtilis biomass samples (AuB)

Based on the qualitative analysis discussed above, theEXAFS data from the eight AuB samples were first fit inde-pendently using a linear combination of Au–amine andAu–sulfhydryl binding. Three single-scattering paths wereused: Au–N, Au–S and Au–C. Each data set was fit simul-taneously with k weights equal to 1, 2, and 3 in Fourier-transformed R-space to decrease the correlation betweenthe fitting parameters. Modeling the data from the eightAuB samples with these two functional groups yields goodfit results. Additionally, to check for the possibility of phos-phoryl binding in the samples, a second shell Au–P path(path length allowed to vary between 2.6 and 3.3 A) wasadded and the coordination number was allowed to vary.The result of the fit indicates negligible Au–phosphorylbinding, and forcing the Au–P path into the fitting makesthe fit results worse, increasing both the v2

m , and R-factorsignificantly. Similar results were obtained when we addedan Au–Au path into the fitting. The calculated percentageof Au–Au bonds in the sample was close to zero, indicatingthat there are no XAFS-observable Au nanoparticles in thesample, a result which is consistent with the XANES datafor these samples. Nearly identical fit results can be gener-ated by fitting these data with a linear combination of Au–carboxyl and Au–sulfhydryl binding. This is not surprising;

Fig. 3. k3 weighed v(k) data for (a) Au solution standards andAu(I) thiomalate powder compound, (b) B. subtilis biomasssamples, and (c) P. putida biomass samples.


as discussed earlier, N and O have similar backscatteringsignatures so the contribution of these two first neighboratoms cannot be distinguished from each other.

By fitting the data from different samples independently,our results indicate that Au atoms in the B. subtilis biomass

are bound to a mixture of amine and/or carboxyl and sulf-hydryl functional groups. The fitting results also provideconstraints on the coordination numbers, bond distances,and Debye–Waller factors of the Au bonds in the samples.However, the correlation between the fitting parameters,such as coordination numbers and Debye–Waller factorsor energy shift and bond distances in each individual fit re-sult in large uncertainties for the value of these variablesobtained from the fit. For the final fits of the data, all datafrom these eight samples were fit simultaneously at the k-weights of 1, 2, and 3. Based on the fact that the local atom-ic environments of Au atoms are similar in these eight sam-ples, the energy shift, Debye–Waller factor, and radialdistance were constrained to be the same for the same pathin all samples in the fitting process. In other words, thestructures of Au bound to the various functional groupsdo not change for these samples, only their relative num-bers do. This has the effect of reducing the correlation be-tween fitting parameters, leading to smaller uncertaintiesin these parameters (Kelly et al., 2002).

The final fitting results for the AuB samples are listed inTable 3, and the magnitude and real part of the FT dataand fits are shown in Fig. 6a and b, respectively. The rela-tive numbers of Au bound to either N/O or S as a functionof pH for the AuB samples are shown in Fig. 8a. The bonddistances suggest an inner-sphere binding mechanism. Forsamples AuB1, AuB2, AuB3 and AuB4, the number of Nand/or O atoms in the first shell decreases with increasingpH and the number of S atoms in the first shell follows areverse trend. The sum of coordination numbers of Nand/or O and S for all the samples is close to two, indicat-ing bidentatebinding of the Au atoms. For samples AuB5,AuB6, AuB7, and AuB8, sulfhydryl groups become thedominant binding site and the percentage of Au–amine/car-boxyl binding becomes much smaller than in the sampleswith the higher Au:biomass ratio (AuB1–AuB4). However,the same trend in coordination numbers with respect to pHis observed. As was the case for the samples with the higherAu:biomass ratio, our results are consistent with bidentatebinding of Au in these samples, with the sum of the firstshell coordination numbers close to two.

In addition to the Au–N, Au–S and Au–C single scatter-ing paths initially included in the fitting, a fourth singlescattering path (Au–C2, corresponding to a second-shellcarbon in the sulfhydryl group) should in principle be in-cluded. Attempts to include this indicated that large disor-der in the path distance reduces the amplitude enough tomake it unobservable. Guine et al. (2006) and Mishraet al. (2009, 2010) studied Zn and Cd adsorption to bacteriaand observed Zn and Cd bound to sulfhydryl sites on bac-terial surfaces in some samples. No second shell carbonatom in the sulfhydryl group was reported in these studieseither, presumably for the same reason.

3.5.2. P. putida biomass samples (AuP)

The AuP data are analyzed following a similar approachto that applied to the AuB data. The data from the eightAuP samples were first modeled with a linear combinationof Au binding to amine groups and sulfhydryl groups, andwere at first fit individually. An Au–P path was then

Fig. 4. Magnitude of the Fourier transformed (FT) data (FT k range 2.0–10.2 (A�1)) for (a) Au solution standards and Au(I) thiomalatepowder compound, (c) B. subtilis biomass samples, and (e) P. putida biomass samples; real part of the FT data plotted in r range 1.2–2.9 A for(b) Au solution standards and Au(I) thiomalate powder compound, (d) B. subtilis biomass samples, and (f) P. putida biomass samples.


allowed in the fitting to check the possibility of phosphorylbinding; and an Au–Au path was added to check the possi-bility of Au nanoparticle formation. For the same reasonsas discussed in the previous section, neither XAFS-observa-ble Au–phosphoryl binding nor Au nanoparticle formationwere observed in any of these samples. The fit results

showed similar Au binding environments in samplesAuP1, AuP2, and AuP3 (1.0 g/L bacteria under low pHconditions) to those in AuB samples. Additionally, the datafrom AuP samples 4, 5, (1.0 g/L bacteria under high pHconditions) 6, 7, and 8 (7.0 g/L bacteria) yielded improvedfits with pure sulfhydryl binding relative to models without

Fig. 5. Magnitude (a) and real part plotted in r range 1.2–2.9 A (b)of the Fourier transformed (FT) data (FT k range 2.0–10.2 (A�1))for comparison between chosen Au solution standards and biomasssamples. Arrows in the figure show the shift of the first EXAFSpeak position as a function of pH for selected Au solutionstandards and biomass samples. Note that the solution standardsand biomass samples shift in opposite directions.


sulfhydryl binding considered. In order to reduce the corre-lations between fitting parameters, all eight samples werethen fit simultaneously at the k-weights of 1, 2 and 3. Thecalculated coordination numbers for N in samples AuP4,AuP5, AuP6, AuP7 and AuP8 are approximately 0, andthe calculated coordination numbers for S are close to 2,which is consistent with the individual fit results and indi-cates pure di-sulfhydryl binding. Finally, samples AuP1,AuP2, and AuP3 were then simultaneously fit with Au–N,Au–S and Au–C paths considered, and samples AuP4,AuP5, AuP6, AuP7, and AuP8 simultaneously fit with asingle Au–S path to generate the best fit values.

The final fitting results for the AuP samples are listed inTable 4, and the magnitude and real part of the FT dataand fits are shown in Fig. 7a and b, respectively. The rela-tive numbers of Au bound to either N/O or S as a functionof pH for the AuP samples are shown in Fig. 8b. The bonddistances are consistent with those calculated for the AuB

samples. For samples AuP1, AuP2, and AuP3, the numberof N atoms in the first shell decreases with increasing pHwhereas the number of S atoms in the first shell follows areverse trend. The sum of coordination numbers of N andS for all AuP samples is close to 2, confirming bidentatebinding.

3.6. Discussion

The results from our XAFS experiments indicate thatthe coordination environment of Au within the biomasssamples varies as a function of both solution pH and the ex-tent of Au loading on the bacterial cells. The experimentsbegan with all of the Au in the systems present as aqueousAu(III)–hydroxide–chloride complexes. Exposure to eitherbacterial species resulted in pH-dependent removal of Aufrom solution, and virtually all of the Au atoms that be-came associated with the biomass were reduced to Au(I)by these non-metabolizing cells. In contrast to what hasbeen observed in some studies of Au(III) interaction withmetabolizing bacterial cells (e.g., Karthikeyan and Bever-idge, 2002; Lengke and Southam, 2005, 2006; Reith andMcPhail, 2006; Lengke et al., 2006a,b, 2007; Jian et al.,2009; Reith et al., 2009), complete reduction to Au(0) andnanoparticle formation did not occur in our experiments.No Au(0) (as evidenced by the XANES data) or Au–Aunearest neighbors (from the EXAFS data) was observedin our experimental systems. All the Au in the biomass sam-ples was present as adsorbed species, and the process ofadsorption in all samples appears to include the breakdownof aqueous Au(III)–hydroxide–chloride complexes and for-mation of chloride-free Au complexes with bacterial cellwall functional groups.

The EXAFS data indicate that, to a large extent, similarAu binding mechanisms occur between the two bacterialspecies studied here, despite their differing cell wall struc-tures. Amine/carboxyl groups and sulfhydryl groups arethe main binding sites for the Au in all of the biomasssamples studied. For the B. subtilis samples with lowerAu loading (5 ppm Au, 7.0 g/L bacteria concentration;AuB5–AuB8), sulfhydryl groups are dominant. For the P.

putida samples with lower Au loading (5 ppm Au, 7.0 g/Lbacteria concentration; AuP6–AuP8), di-sulfhydryl bindingoccurs exclusively. For the B. subtilis samples with higherAu loading (5 ppm Au, 1.0 g/L bacteria concentration;AuB1–AuB4) and the P. putida samples with higher Auloading under pH <7 (5 ppm Au, 1.0 g/L bacteria concen-tration; AuP1–AuP3), amine/carboxyl groups are as impor-tant as the sulfhydryl groups in binding the Au. For P.

putida samples with higher Au loading at pH >7 (5 ppmAu, 1.0 g/L bacteria concentration; AuP4 and AuP5), di-sulfhydryl binding was observed as well. Additionally, forall the AuB and AuP samples in which Au is bound to morethan one type of functional groups, as pH increases (andhence as Au loading decreases), the relative importance ofsulfhydryl groups increases.

There have been few reports of metal binding to aminegroups on bacterial cell walls. Amine groups are positivelycharged under a broad range of pH values; the pKa valuefor amine functional groups on many molecules is near 11

Table 3Fit results for eight Au-B. subtilis biomass samples.

Sample ID AuB1 AuB2 AuB3 AuB4 AuB5 AuB6 AuB7 AuB8

Au–N N 1.3 ± 0.3 1.3 ± 0.3 1.2 ± 0.3 1.0 ± 0.3 0.5 ± 0.2 0.5 ± 0.2 0.4 ± 0.1 0.3 ± 0.1R (A) 1.97 ± 0.01a

r2 (A2) 0.002 ± 0.001a

E0 (eV) �1.8 ± 1.1a

Au–S N 0.7 ± 0.2 0.8 ± 0.2 1.0 ± 0.3 1.3 ± 0.3 1.6 ± 0.4 1.7 ± 0.3 1.8 ± 0.4 1.9 ± 0.3R (A) 2.29 ± 0.01a

r2 (A2) 0.002 ± 0.002a

E0 (eV) 4.3 ± 1.8a

Au–C N 1.8 ± 0.8 1.7 ± 0.9 1.6 ± 0.8 1.4 ± 0.7 0.7 ± 0.5 0.7 ± 0.5 0.7 ± 0.5 0.5 ± 0.4R (A) 2.98 ± 0.02a

r2 (A2) 0.002 ± 0.002a

E0 (eV) �1.5 ± 0.7a

v2m 58a

R factor 0.005a

a Best fit values and goodness of the fit were obtained from a simultaneous fit of eight samples. R, r2, and E0 of each path were constrainedto be the same for the eight samples.

Fig. 6. Data and fits of (a) the magnitude of the FT of B. subtilis

biomass data, (b) the real part of the FT of B. subtilis biomass dataplotted over the fitting range of 1.15–2.9 A.

Fig. 7. Data and fits of (a) the magnitude of the FT of P. putida

biomass data, (b) the real part of the FT of P. putida biomass dataplotted over the fitting range of 1.15–2.9 A.


Table 4Fit results for eight Au-P. putida biomass samples.

Sample ID AuP1 AuP2 AuP3 AuP4 AuP5 AuP6 AuP7 AuP8

Au–N N 1.1 ± 0.3 1.1 ± 0.3 1.0 ± 0.3 –b

R (A) 1.97 ± 0.01a –b

r2 (A2) 0.002 ± 0.001a –b

E0 (eV) 1.2 ± 0.7a –b

Au–S N 0.8 ± 0.3 0.9 ± 0.3 1.2 ± 0.3 2.0 ± 0.1 1.9 ± 0.1 2.1 ± 0.1 2.0 ± 0.1 2.0 ± 0.2R (A) 2.29 ± 0.01a 2.30 ± 0.01c

r2 (A2) 0.002 ± 0.001a 0.002 ± 0.001c

E0 (eV) 6.9 ± 2.6a 6.5 ± 0.5c

Au–C N 1.3 ± 0.8 1.3 ± 0.7 1.1 ± 0.5 –b

R (A) 2.98 ± 0.02a –b

r2 (A2) 0.002 ± 0.002a –b

E0 (eV) 0.2 ± 0.8a –b

v2m 39a 32c

R factor 0.001a 0.009c

a Best fit values and goodness of the fit were obtained from a simultaneous fit of three samples AuP1, AuP2, and AuP3. R, r2, and E0 of eachpath were constrained to be the same for the three samples.

b Au–N and Au–C paths were not included in fitting samples AuP4–AuP8.c Best fit values and goodness of the fit were obtained from a simultaneous fit of three samples AuP4–AuP8. R, r2, and E0 of each path were

constrained to be the same for the five samples.


(Sorrell, 2005). Most aqueous metal ions also carry positivecharges; therefore, aqueous metal cations are not likely toattach to amine groups. However, for metals such as theoxidized forms of aqueous Au, As, Se, and Cr that are pres-ent as negatively charged complexes in solution, bindingonto amine groups on bacterial cell walls may becomeimportant, especially at low pH where electrostatic repul-sive forces between the aqueous anions and the negativelycharged cell wall weaken.

In our experimental system, sulfhydryl groups becomeincreasingly important in binding Au in samples with lowerAu loading, a result that is consistent with the bacterialadsorption behavior of Zn and Cd (e.g. Guine et al.,2006; Mishra et al., 2009, 2010). Also, note that at the sameAu/bacteria ratio and under similar pH conditions, moreAu atoms are bound to sulfhydryl groups in P. putida sam-ples than in B. subtilis samples. This may indicate a higherdensity of sulfhydryl groups within P. putida cell walls thanon B. subtilis cells. In most natural geological systems, Auand many other metals are present in low concentrations,and bacterial cell wall sulfhydryl groups may control thebacterial adsorption behavior of these metals.

XAFS results alone cannot differentiate between Au thatis adsorbed onto bacterial cell walls and Au that may bepresent inside the cells. However, our experiments involvedonly non-metabolizing cells that were thoroughly washedwith 0.1 M NaClO4 prior to use; no external electron do-nors or sources of carbon were included in the experiments;and the exposure time was limited to only 2 h. In addition,experiments conducted under similar conditions (Nakaj-ima, 2003; Tsuruta, 2004) demonstrated that Au removalfrom solution can be readily reversed by exposure of thebiomass to a 0.1 M thiourea solution. Therefore, the Au re-moval from solution that we observed in our experimentswas most likely due to adsorption reactions, and the speci-

ation of Au that we determined from the XAFS data mostlikely reflect the cell wall binding environments for eachbacterial species studied.

In our experiments, Au was present initially as aqueousAu(III)–hydroxide–chloride complexes, which were con-verted during the experiments to Au(I) that was bound tothe bacterial cells. The XAS data provide unequivocal evi-dence for the Au reduction and binding mechanisms, butthe data do not provide constraints on the relative timingof the reduction and adsorption reactions that occurred.It is possible that reduction occurs on the cell wall afteradsorption of Au(III). We observed no Au–Cl signal inthe Au on the biomass, so dechlorination of the aqueousAu–hydroxide–chloride complexes must occur prior toadsorption. Another possibility is that the reduction occursin solution prior to adsorption of the Au, with electronssupplied to the aqueous Au(III)–hydroxide–chloride com-plexes by bacterial exudates, forming an intermediate(and unidentified) aqueous Au(I) species which eventuallyadsorbs onto the bacteria. Speciation studies of the aqueousphase during reduction and measurements of Au reductionpotential in the presence of isolated bacterial exudates areneeded to distinguish between these two mechanisms.

This study provides constraints on what are likely theinitial stages of Au biomineralization and the formationof Au(0) nanoparticles by bacteria. Previous studies havedemonstrated that metal and sulfate reducing bacteria canreduce and precipitate aqueous Au(III) to form Au(0)nanoparticles at the solution–cell wall interface (e.g., Kart-hikeyan and Beveridge, 2002; Lengke and Southam, 2005,2006; Reith and McPhail, 2006; Lengke et al., 2006a,b,2007). In addition, bacteria have been reported to metabol-ically reduce Au(III) complexes to Au(0) nanoparticleslikely with the formation of intermediate Au(I)–S com-plexes via fast passive adsorption mechanism (Jian et al.,

Fig. 8. Au bound to the different first shell atoms as a function ofpH in (a) B. subtilis biomass samples and (b) P. putida biomasssamples according to the proposed EXAFS model fitting.


2009; Reith et al., 2009). Our results not only provide directmolecular-level evidence of this mechanism, but also sug-gest that the first step in this process is the adsorption ofAu to sulfhydryl and amine and/or carboxyl binding siteson bacterial cell walls, and that these sites may be the loca-tions where Au reduction occurs. Moreover, this researchimplies that sulfhydryl binding and subsequent reductionof Au on bacterial cell walls may play an important rolein affecting the speciation and distribution of Au in near-surface geologic systems.

4. CONCLUSIONS

The Au LIII XAFS measurements of aqueous Au stan-dards, standard Au compounds and biomass samples ex-posed to Au(III) solutions provide information about thevalence state of Au in the biomass samples as well as theaverage local atomic environment of Au atoms adsorbedto the cell walls. Both XANES and EXAFS data of Au bio-mass samples indicate that more than 90% of the adsorbedAu, which started as Au(III) in solution, was reduced toAu(I) atoms by non-metabolizing B. subtilis and P. putida

bacterial cells, possibly due to oxidation of electron-trans-port chain enzymes located within the bacterial cell walls.

In contrast to what has been observed for Au(III) interac-tion with actively metabolizing metal-reducing bacterialcells, no Au(0) or Au–Au nearest neighbors were observedin our experimental systems.

The EXAFS data suggest that dechlorination accompa-nies bacterial adsorption of Au from Au(III)–hydroxide–chloride solutions under all conditions studied here. TheEXAFS data also suggest that Au on the B. subtilis andP. putida cell walls is present almost exclusively as Au(I)bound to a mixture of amine/carboxyl sites as well as tosulfhydryl functional groups. The relative importance ofthe sulfhydryl groups increases with increasing pH and withdecreasing Au loading on the cell walls. The coordinationnumbers for all of the biomass samples studied here areconsistent with bidentate binding of Au on the bacterial cellwalls, and the bond distances indicate inner-sphere bindingof the Au. Collectively, the XAFS data presented here pro-vide a framework for understanding the initial stages of for-mation of Au(0) nanoparticles by bacteria from Au(III)solutions, and the results enhance our ability to accountfor the behavior of Au in bacteria-bearing geologic systems.

ACKNOWLEDGEMENTS

This study was supported by the funding provided by the Na-tional Science Foundation through an Environmental MolecularScience Institute (EMSI) Grant (EAR-0221966) to the Universityof Notre Dame. MRCAT is supported by the US Department ofEnergy under contract DE-FG02-94-ER-45525 and the memberinstitutions. Use of the Advanced Photon Source was supportedby the US Department of Energy under contract W-31-109-Eng-38. Z.S. was supported by the EMSI grant and by a Bayer Predoc-toral Fellowship provided through the Center for EnvironmentalScience and Technology (CEST) at Notre Dame. We thank Jenni-fer Szymanowski for her help in preparing samples, and MRCATstaff for their beamline support. We also thank Bhoopesh Mishraand Maxim Boyanov for their suggestions in XAFS experimentand data analysis. ICP-OES measurements were supported byCEST. Two anonymous journal reviews of an earlier version of thismanuscript were very helpful in improving the presentation of theresearch.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.gca.2012.02.030.

REFERENCES

Bau R. (1998) Crystal structure of the antiarthritic drug Authiomalate (Myochrysine): a double-helical geometry in thesolid state. J. Am. Chem. Soc. 120, 9380–9381.

Berrodier I., Farges F., Benedetti M., Winterer M., Brown, Jr., G.E. and Deveughele M. (2004) Adsorption mechanisms oftrivalent Au on iron- and aluminum-(oxy) hydroxides. Part 1:X-ray absorption and Raman scattering spectroscopic studiesof Au(III) adsorbed on ferrihydrite, goethite, and boehmite.Geochim. Cosmochim. Acta. 68, 3019–3042.

Beveridge T. J. and Murray R. G. E. (1976) Uptake and retentionof metals by cell walls of Bacillus subtilis. J. Bacteriol. 127,

1502.


Bischoff G. C. O. (1994) Au-adsorbing bacteria as colonisers onalluvial placer Au. N. Jb. Geol. Palaont. Abh. 194, 187–209.

Bischoff G. C. O. (1997) The biological origin of bacterioform Aufrom Australia. N. Jb. Geol. Palaont. Abh. H6, 329–338.

Boyanov M. I., Kelly S. D., Kemner K. M., Bunker B. A., Fein J.B. and Fowle D. A. (2003) Adsorption of cadmium to Bacillus

subtilis bacterial cell walls: a pH-dependent X-ray absorptionfine structure spectroscopy study. Geochim. Cosmochim. Acta

67, 3299–3311.

Bunker G. B. (2010) Introduction to XAFS: A Practical Guide to

X-ray Absorption Fine Structure Spectroscopy, first ed. Cam-bridge University Press.

Durham P. J., Pendry J. B. and Hodges C. H. (1982) Calculation ofX-ray absorption near-edge structure, XANES. Comp. Phys.

Commun. 25, 193–206.

Elder R. C. and Eidsness M. K. (1987) Synchrotron X-ray studiesof metal-based drugs and metabolites. Chem. Rev. 87, 1027–

1046.

Falconer D. M., Craw D., Youngson J. H. and Faure K. (2006) Auand sulphide minerals in Tertiary quartz pebble conglomerateAu placers, Southland, New Zealand. Ore Geol. Rev. 28,

525–545.

Farges F., Sharps J. A. and Brown, Jr., G. E. (1993) Localenvironment around Au(III) in aqueous chloride solutions: anEXAFS spectroscopy study. Geochim. Cosmochim. Acta 57,

1243–1252.

Gammons C. H., Yu Y. M. and Williams-Jones A. E. (1997) Thedisproportionation of gold(I) chloride complexes at 25 to 200degrees C. Geochim. Cosmochim. Acta 61, 1971–1983.

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

Hennig C., Panak P. J., Reich T., Rossberg A., Raff J., Selenska-Pobell S., Matz W., Bucher J. J., Bernhard G. and Nitsche H.(2001) EXAFS investigation of uranium(VI) complexes formedat Bacillus cereus and Bacillus sphaericus surfaces. Radiochim.

Acta 89, 625–631.

Jian X., Wasinger E. C., Lockard J. V., ChenL X. and He C. (2009)Highly sensitive and selective Au(I) recognition by a metallo-regulator in Ralstonia metallidurans. J. Am. Chem. Soc. 131(31),

10869–10871.

Karthikeyan S. and Beveridge T. J. (2002) Pseudomonas aeruginosa

biofilms react with and precipitate toxic soluble Au. Environ.

Microbiol. 4, 667–675.

Kashefi K., Tor J. M., Nevin K. P. and Lovely D. R. (2001)Reductive precipitation of Au by dissimilatory Fe(III)-reducingbacteria and Archaea. Appl. Environ. Microbiol. 67, 3275–

3279.

Kelly S. D., Kemner K. M., Fein J. B., Fowle D. A., Boyanov M.I., Bunker B. A. and Yee N. (2002) X-ray absorption finestructure determination of pH-dependent U-bacterial cell wallinteractions. Geochim. Cosmochim. Acta 66, 3855–3871.

Kemner K. M., Kropf A. J. and Bunker B. A. (1994) A lowtemperature total electron yield detector for X-ray absorptionfine structure spectra. Rev. Sci. Instrum. 65, 3667–3669.

Lengke M. F. and Southam G. (2005) The effect of thiosulfate-oxidizing bacteria on the stability of the Au–thiosulfatecomplex. Geochim. Cosmochim. Acta 69, 3759.

Lengke M. and Southam G. (2006) Bioaccumulation of Au bysulfate-reducing bacteria cultured in the presence of Au(I)–thiosulfate complex. Geochim. Cosmochim. Acta 70, 3646–3661.

Lengke M. F., Fleet M. E. and Southam G. (2006a) Morphology ofAu nanoparticles synthesized by filamentous Cyanobacteriafrom Au(I)–thiosulfate and Au(III)–chloride complexes. Lang-

muir 22, 2780–2787.

Lengke M. F., Ravel B., Fleet M. E., Wanger G., Gordon R. A.and Southam G. (2006b) Mechanisms of Au bioaccumulationby filamentous Cyanobacteria from Au(III)–chloride complex.Environ. Sci. Technol. 40, 6304–6309.

Lengke M. F., Ravel B., Fleet M. E., Wanger G., Gordon R. A.and Southam G. (2007) Precipitation of Au by the reaction ofaqueous Au(III) chloride with cyanobacteria at 25–80 �C –studied by X-ray absorption spectroscopy. Can. J. Chem. 85,

651–659.

Madigan M. T., Martinko J. M. and Parker J. (2002) Brock

Biology of Microorganisms, 10 ed. Prentice Hall Inc..Mishra B., Boyanov M. I., Bunker B. A., Kelly S. D., Kemner K.

M., Nerenberg R., Read-Daily B. L. and Fein J. B. (2009) AnX-ray absorption spectroscopy study of Cd binding ontobacterial consortia. Geochim. Cosmochim. Acta 73, 4311–4325.

Mishra B., Boyanov M., Bunker B. A., Kelly S. D., Kemner K. M.and Fein J. B. (2010) High- and low-affinity binding sites for Cdon the bacterial cell walls of Bacillus subtilis and Shewanella

oneidensis. Geochim. Cosmochim. Acta 74, 4219–4233.

Nakajima A. (2003) Accumulation of Au by microorganisms.World J. Microbiol. Biotech. 19, 369–374.

Newville M. (2001) IFEFFIT: interactive XAFS analysis andFEFF fitting. J. Synchrotron Radiat. 8, 322–324.

Newville M., Livins P., Yacoby Y., Rehr J. J. and Stern E. A.(1993) Near-edge X-ray-absorption fine structure of Pb – acomparison of theory and experiment. Phys. Rev. B 47, 14126–

14131.

Pantelouris A., Kueper G., Hormes J., Feldmann C. and Jansen M.(1995) Anionic Au in Cs3AuO and Rb3AuO established byX-ray absorption spectroscopy. J. Am. Chem. Soc. 117, 11749–

11753.

Patai S. and Rappoport Z. (1999) The chemistry of organic

derivatives of Au and silver, first ed. Wiley.Pokrovski G. S., Tagirov B. R., Schott J., Bazarkina E. F.,

Hazemann J. L. and Proux O. (2009a) An in situ X-rayabsorption spectroscopy study of Au–chloride complexing inhydrothermal fluids. Chem. Geol. 259, 17–29.

Pokrovski G. S., Tagirov B. R., Schott J., Hazemann J. L. andProux O. (2009b) A new view on Au speciation in sulfur-bearing hydrothermal fluids from in situ X-ray absorptionspectroscopy and quantum-chemical modeling. Geochim. Cos-

mochim. Acta 73, 5406–5427.

Ravel B. and Newville M. (2005) Athena, artemis, hephaestus: dataanalysis for X-ray absorption spectroscopy using IFEFFIT. J.

Synchrotron Radiat. 12, 537–541.

Rehr J. J., Zabinsky S. I. and Albers R. C. (1992) High-ordermultiple scattering calculations of X-ray-absorption fine struc-ture. Phys. Rev. Lett. 69, 3397–3400.

Reith F. and McPhail D. C. (2006) Effect of resident microbiota onthe solubilization of Au in soil from the Tomakin Park AuMine, New South Wales, Australia. Geochim. Cosmochim. Acta

70, 1421–1438.

Reith F., McPhail D. C. and Christy A. G. (2005) Bacillus cereus,Au and associated elements in soil and other regolith samplesfrom Tomakin Park Au Mine in southeastern New SouthWales, Australia. J. Geochem. Explor. 85, 81–98.

Reith F., Lengke M., Falconer D., Craw D. and Southam G.(2007) The geomicrobiology of Au. ISME J. 1, 567–584.

Reith F., Etschmann B., Grosse C., Moors H., Benotmane M. A.,Monsieurs P., Grass G., Doonan C. and Vogt S., et al. (2009)Mechanisms of Au biomineralization in the bacterium Cupri-

avidus metallidurans. Proc. Natl. Acad. Sci. USA 106, 17757–

17762.

Reith F., Rogers S. L., McPhail D. C. and Webb D. (2006)Biomineralization of Gold: Biofilms on Bacterioform Gold.Science 313, 233–236.


Segre C. U., Leyarovska N. E., Chapman L. D., Lavender W. M.,Plag P. W., King A. S., Kropf A. J., Bunker B. A., Kemner K.M., Dutta P., Duran R. S. and Kaduk J. (2000) The MRCAT

Insertion Device Beamline at the Advanced Photon Source.Synchrotron Radiat. Instrum: Eleven US Conference CP521,pp. 419–422.

Sorrell T. N. (2005) Organic Chemistry, second ed. UniversityScience Books.

Stern E. A. and Heald S. M. (1983) Basic principles andapplications of EXAFS. Handb. Synchrotron Radiat. 10, 995–

1014.

Stern E. A., Newville M., Ravel B., Yacoby Y. and Haskel D.(1995) The uwxafs analysis package – philosophy and details.Physica B209, 117–120.

Toner B., Manceau A., Marcus M. A., Millet D. B. and Sposito G.(2005) Zinc sorption by a bacterial biofilm. Environ. Sci.

Technol. 39, 8288–8294.

Tsuruta T. (2004) Biosorption and recycling of Au using variousmicroorganisms. J. Gen. Appl. Microbiol. 50, 221–228.

Watterson J. R. (1992) Preliminary evidence for the involvement ofbudding bacteria in the origin of Alaskan placer Au. Geology

20, 315–318.

Associate editor: Peggy A. O’Day

an x-ray absorption fine structure study of au adsorbed onto the …fein/publications/song...

Documents