Biological Recovery of Platinum Complexes from Diluted Aqueous Streams by Axenic Cultures

The widespread use of platinum in high-tech and catalytic applications has led to the production of diverse Pt loaded wastewaters. Effective recovery strategies are needed for the treatment of low concentrated waste streams to prevent pollution and to stimulate recovery of this precious resource. The biological recovery of five common environmental Pt-complexes was studied under acidic conditions; the chloro-complexes PtCl42- and PtCl62-, the amine-complex Pt(NH3)4Cl2 and the pharmaceutical complexes cisplatin and carboplatin. Five bacterial species were screened on their platinum recovery potential; the Gram-negative species Shewanella oneidensis MR-1, Cupriavidus metallidurans CH34, Geobacter metallireducens, and Pseudomonas stutzeri, and the Gram-positive species Bacillus toyonensis. Overall, PtCl42- and PtCl62- were completely recovered by all bacterial species while only S. oneidensis and C. metallidurans were able to recover cisplatin quantitatively (99%), all in the presence of H2 as electron donor at pH 2. Carboplatin was only partly recovered (max. 25% at pH 7), whereas no recovery was observed in the case of the Pt-tetraamine complex. Transmission electron microscopy (TEM) revealed the presence of both intra- and extracellular platinum particles. Flow cytometry based microbial viability assessment demonstrated the decrease in number of intact bacterial cells during platinum reduction and indicated C. metallidurans to be the most resistant species. This study showed the effective and complete biological recovery of three common Pt-complexes, and estimated the fate and transport of the Pt-complexes in wastewater treatment plants and the natural environment.


Introduction
The growing importance and use of platinum in clean and high-tech products in the last 30 years have induced the production of Pt loaded waste streams and the accumulation of platinum in the environment [1,2]. For example, deterioration of automotive catalysts a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 leads to the emission of Pt particles into the environment, part of which gets drained by stormwater into sewers [3]. Platinum is also the crucial building block of chemotherapeutic drugs such as cisplatin and carboplatin, and the excreted human metabolites contaminate both hospital and municipal wastewaters [1]. Finally, liquid waste streams (often diluted) containing platinum are also produced from the application of industrial catalysts, the manufacturing of jewelry and electronics, and both primary mining and precious metal recovery activities [2,3].
The resulting residual platinum appears in different complexes in wastewater, with inorganic or organic ligands, such as cisplatin (cis-PtCl 2 [NH 3 ] 2 ), carboplatin (cis-(Pt [NH 3 ] 2 [1,1-cyclobutanedicarboxylato])), and their metabolites, chloro-complexes Pt(II) Cl 4 2and Pt(IV)Cl 6 2-or amine-complexes such as Pt(NH 3 ) 4 Cl 2 , resulting from leaching or metal refinery processes [4][5][6]. The metal's fate in a wastewater treatment plant or in the receiving environment depends largely on the metal's speciation and the matrix composition of the waste stream [2]. An effective removal of the precious metal is advised to both lower the pollutant load in the environment, based on the pollution prevention principle, and since the behavior and impact of species such as cisplatin is mainly unknown in the environment [7]. Moreover, platinum's high market value (av. 34.7 $ g -1 in 2015 [8]) and criticality stimulate the effective recovery and valorization of critical resources [9,10].
Case by case, it should be questioned if the targeted Pt-complex could be removed and recovered from the waste stream, and whether this recovery could be interesting from an economical point of view [2]. Biotechnologies based on living biomass can serve as low-cost and green treatment techniques to recover platinum at low concentrations [11,12]. The effective removal of platinum by different axenic cultures has been demonstrated before; PtCl 4 2and PtCl 6 2-were sorbed by Shewanella putrefaciens [12], PtCl 6 2-was reduced by Shewanella algae [13] and an undefined Pt-complex was reduced by Cupriavidus metallidurans [14]. However, the metal speciation can hamper an effective metal removal [2]. Complex waste streams such as highly acidic saline streams originating from metal refinery processes can be considered too challenging for conventional biological wastewater treatment plants (WWTP). They require specialized mixed cultures adapted to the prevalent conditions [11,15]. The aim of this study was to further elaborate the biological recovery of different synthetic platinum complexes, representative for diluted Pt containing wastewaters of interest. It is important to explore the fate of these common Pt-complexes once they have entered a wastewater treatment plant or the environment. Therefore, this study investigates the relationship between Pt-speciation and the observed recovery by axenic cultures and the effect of the different Pt-complexes on the cell viability. The studied Pt-complexes include; chloro-complexes PtCl 4 2and PtCl 6 2-, present in e.g. run-off waters or industrial process streams, cisplatin, carboplatin, and a Pt-amine complex Pt(NH 3 ) 4 Cl 2 , being found in metal refinery streams. A series of axenic bacterial species (Shewanella oneidensis MR-1, Cupriavidus metallidurans CH34, Geobacter metallireducens, Bacillus toyonensis, and Pseudomonas stutzeri), commonly present in wastewater plants treating metal contaminated effluents, was examined. In this research, it was evaluated which bacteria recovered platinum, under which conditions (with and without electron donor) and how this was affected by the metal speciation. The morphology and metal speciation of the precipitated Pt particles were investigated, as well as the viability of the axenic cultures during Pt recovery.
The results of this study can be used as a prediction of the fate and transport of Pt salts in wastewater treatment plants.

Bacterial cultures and growth conditions
Shewanella oneidensis MR-1 was obtained from the BCCM/LMG Bacteria Collection (Gent, Belgium; LMG 19005) and Cupriavidus metallidurans CH34 was obtained from SCK•CEN (Mol, Belgium). Both species were grown aerobically in Lysogeny broth (LB-Lennox) medium overnight at 28˚C. Geobacter metallireducens was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ; ATCC 53774) and was cultivated anaerobically in DSMZ medium 579 at 28˚C for 7 days. Bacillus toyonensis and Pseudomonas stutzeri were isolated from a wastewater treatment plant which processes metal streams, and were grown aerobically in Lysogeny broth (LB-Lennox) medium for 48 h at 28˚C.

Platinum recovery experiments
The five different bacterial species were used to study the recovery of five different Pt-complexes; K 2 PtCl 4 (Sigma-Aldrich, USA), K 2 PtCl 6 (Sigma-Aldrich, USA), Pt(NH 3 ) 4 .2HCO 3 (Alfa Aesar, Germany), cisplatin (Alfa Aesar, Germany), and carboplatin (Alfa Aesar, Germany). Cells from the cultures S. oneidensis, C. metallidurans, B. toyonensis, and P. stutzeri were harvested by centrifugation (7000 g, 7 min) and were then washed twice with 25 mL phosphate buffer (8.5 g L -1 Na 2 HPO 4 .7H 2 O and 3 g L -1 KH 2 PO 4 ). The washed cells were suspended in the phosphate buffer to a final optical density of 1 (OD 610nm ) and added to 120 mL glass serum bottles (50 mL cell suspensions). The glass bottles were flushed by 20 repeated cycles of N 2 overpressure and vacuum underpressure. In the case H 2 was used as electron donor, the headspace was replaced with 100% H 2 -gas. In the case of formate and acetate, 50 mM formate or 12.5 mM acetate was dosed to the biomass suspension. Subsequently, platinum was dosed to a final concentration of 100 mg L -1 Pt (2.5 g L -1 Pt 1 M HCl stock). The glass bottles were incubated and continuously mixed at 100 rpm and 28˚C during the experiment. The protocol was slightly modified for experiments based on G. metallireducens. The cells were centrifuged at 8000 g for 7 min, washed and suspended in 30 mM NaHCO 3 to a final optical density of 0.31. All steps were executed in an anaerobic closet (37˚C, 80% N 2 /20% CO 2 ).
The washed biomass suspensions were all characterized by pH 7.0-7.1 prior to Pt dosage. The pH was adjusted to pH 1.8-2.2 at the start of the experiment. Samples were analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES) and Pt recovery efficiencies were calculated after 48 h.

Inductively coupled plasma optical emission spectrometry (ICP-OES)
Experimental details on ICP-OES analysis were previously described by Maes et al. [11]. The platinum concentrations were determined with a Spectro Arcos ICP-OES (Spectro Analytical Instruments GmbH, Kleve, Germany). applied as electron donor, except for the aerobic S. oneidensis sample with Pt(IV)Cl 6 2-(with formate). Experimental details on X-ray absorption spectroscopy were previously described by Maes et al. [11]. All μXAS spectroscopy measurements were performed using the microprobe beamline X27A at the National Synchrotron Light Source (NSLS), Upton, NY. Aliquots of fully hydrated Pt-biomass samples were transferred to an air-tight polypropylene bag to prevent drying. Samples were secured to an x, y, z motorized stage 45 degrees to the incident beam and 13-element HGe Canberra fluorescence detector. The beam spot-size on the sample was maintained at ca. 15 μm.
By means of X-ray Absorption Near Edge Structure (XANES) and Extended X-ray Absorption Fine Structure (EXAFS) spectroscopies, the oxidation state and first-shell coordination environment of the Pt phase associated with the biomass was examined. All spectra were collected at room temperature. X-ray fluorescence was measured from 150 eV below to 800 eV above the Pt L3-edge. Absolute x-ray energy calibration was based on the first inflection point of standard Pt (11 919 eV) metal foil, which was collected in transmission mode as an internal calibration during each scan. Normalization, calibration and averaging of the XAS spectra and ab initio fitting of the EXAFS region of the spectra were performed using Athena and Artemis software [17].

Transmission electron microscopy (TEM)
After the recovery experiments were finished, the bacterial suspensions were washed twice with the according washing buffer; 30 mM NaHCO 3 for G. metallireducens, phosphate buffer for all other species. Samples were stored overnight and the supernatants were removed. The TEM analysis was then performed as previously described by Maes et al. [11], using a Zeiss TEM900 transmission electron microscope (Carl Zeiss, Oberkochen, Germany).

Cell viability by flow cytometry analysis
Partial viability of the strains was assessed during recovery experiments by dual stain flow cytometry as described elsewhere [18]. Briefly, bacterial cells were stained with a mixed SYBR 1 Green I (SG, 10 000x concentrate, Invitrogen) and propidium iodide (PI, 20 μM, Invitrogen) staining solution which resolves membrane damaged from intact bacterial cells. At the start of the recovery experiments the suspension was acidified to pH 2 and the studied Pt-complex and H 2 -gas were added. Samples taken during batch experiments were diluted immediately 100 times in sterile, 0.22 μm filtered phosphate buffer and stored at +4˚C until further analysis. Prior to analysis, samples were if necessary further diluted to approximately 10 6 cells mL -1 and stained with 10 μL mL -1 of the staining solution (final concentration of 1x SG and 4 μM PI). The stained samples were incubated for 20 minutes in the dark at 37˚C and immediately analyzed on a BD FACSVerse (BD Biosciences, Erembodgem, Belgium) equipped with a 20 mW 488 nm blue laser, 40 mW 405 nm violet laser and a 640 nm red laser. Green and red fluorescence intensities corresponding to respectively the SG and PI emission wavelengths were collected through a 527 ± 32 nm band pass and 700 ± 54 nm band pass filter. All samples were collected and analyzed in triplicate within 24 hours of sampling. Cell counts were extracted from manually drawn gates on the green vs. red fluorescence intensity plots as described elsewhere [19]. The limit of detection was 33.3 x 10 3 cells mL -1 .

Recovery of platinum complexes by axenic cultures
The recovery of five different platinum complexes was tested using five axenic bacterial cultures. The selection of Pt-complexes consisted of two Pt-chloro complexes (Pt(II)Cl . The biological platinum recovery was studied using three Gram-negative (G -) species (Shewanella oneidensis MR-1, Cupriavidus metallidurans CH34 and Geobacter metallireducens), each known for their ability to reduce metals, as shown for palladium [14,20,21], and extended with the Gram-positive (G + ) species Bacillus toyonensis and the Gspecies Pseudomonas stutzeri, both isolated from a wastewater treatment plant which processes metal streams.

Influence of Pt-speciation on microbial recovery
The recovery of the five Pt-complexes was tested at pH 2 by using each of the cultures, in the presence of H 2 , formate or acetate as electron donor to study dissimilatory metal reduction, and without electron donor to evaluate sorption. The Pt-chloro complex Pt(II)Cl 4 2was recovered completely with H 2 ( Table 1) and a black microbial Pt suspension was formed (indicative for Pt reduction) [22]. The addition of formate or acetate as electron donor was less effective; 56-79% and 6-19% PtCl 4 2was recovered with formate and acetate, respectively (S1 Table). In experiments studying sorption, 6-25% of the dosed Pt was recovered on the biomass, however no color change was observed. The kinetics of the microbial reduction of Pt(IV)Cl 6 2were slower compared to Pt(II)Cl 4 2-; whereas almost complete recovery was observed in the case of H 2 and 49-87% recovery with formate, no substantial Pt recovery was obtained using acetate (Table 1, S1 and S2 Tables). The sorption control showed a limited removal of 2-8%. Based on these Pt recovery efficiencies, hydrogen gas was preferred as sole electron donor for further experiments.
Cisplatin and carboplatin showed a very different behavior. Although both species sorbed on the biomass to a very little amount, only cisplatin was reduced successfully by the bacteria. Finally, the Pt-tetraamine complex was neither sorbed on any of the studied microbial species, nor reduced by these same species. , but at a slower rate (mainly between 24-48 h). Table 1. An overview of the platinum recovery efficiencies (%) at pH 2 is given; the Pt recovery was investigated with and without (sorption control) the addition of H 2 -gas. The platinum recovery was studied using five different bacterial species and five Pt-complexes (n = 1). All recoveries were measured after 48 h, except for: * 68 h, ** 107 h and *** 168 h, and **** 320 h. The chemical reduction was studied for all Pt-species using H 2 -gas. In the case of cisplatin, S. oneidensis and C. metallidurans could fully recover the complex, while B. toyonensis and P. stutzeri recovered at maximum 10%. In contrast, all bacterial species showed limited recovery of carboplatin.

Pt-species
To explore the recovery potential of the Pt-complexes under circumneutral conditions as a proxy for their fate in wastewater treatment plants or the environment, their recovery was also examined at pH 7 in the absence of an electron donor (S1 Table). In general, lower recovery efficiencies were noted under sorptive conditions compared to the dissimilatory reduction, showing the need for an electron donor to obtain full recovery. By using S. oneidensis and C. metallidurans, the platinum recovery at neutral pH was observed to be very similar to the sorptive removal at acidic pH. For B. toyonensis and P. stutzeri, a better sorption was noted under neutral conditions for mainly cisplatin, carboplatin, and PtCl 4 2on the long term (28-40% PtCl 4 2was recovered after 117-144 h) (S1 Table).

Platinum speciation analysis
A characterization of the bacteria-metal interaction using X-ray absorption spectroscopy, was executed on the species Shewanella oneidensis MR-1, Cupriavidus metallidurans CH34 and Geobacter metallireducens removing the platinum chloro-complexes Pt(II)Cl 4 2and Pt(IV) Cl 6 2- . Anaerobically grown S. oneidensis was included in this speciation analysis for comparison but was not further investigated in this study.
The X-ray energy of peak fluorescence along with the overall shape of the Pt K-edge XANES spectra provides measures of the average oxidation state and bonding environment of Pt associated with the biomass. The peak fluorescence of all the spectra in Fig 1 has shifted to lower energies relative to the Pt(II) aqueous spectrum indicating that the Pt associated with the biomass underwent reduction during the recovery process. The spectral patterns between peak fluorescence and 11 600 eV for Pt(II) recovered by aerobic C. metallidurans and anaerobic S. oneidensis are consistent with metallic Pt(0) particles [11], while the spectral patterns of anaerobic G. metallireducens and aerobic S. oneidensis are similar to the Pt(II) aqueous spectrum suggesting only partial reduction of the available Pt. The EXAFS spectra and Fourier Transforms (FTs) shown in the Supporting Information provide further evidence for these interpretations of the XANES spectra (see S1 Fig).
The XANES spectra also show how the level of Pt recovery (amount of Pt incorporated into the biomass) is correlated with each organism's ability to reduce Pt(II) to metallic Pt(0) nanoparticles. Prior to normalization the edge jump in the XANES spectra reflects the number of Pt atoms fluorescing within the x-ray beam, and therefore, the height of the edge jump provides a measure of the relative amount of Pt per unit of biomass assuming equivalent biomass densities and sample thickness for all biomass samples. These assumptions are reasonable given that all underwent the same biomass separation method and were loaded in identical sample holders. The edge jump values (values shown in square brackets in Fig 1) increase as the Pt XANES spectra more closely resemble the XANES spectrum of metallic Pt(0) nanoparticles [11]. Without independent measures of Pt:biomass, the XANES edge jump only provides a relative measure of Pt recovery.
All bacterial strains are less efficient at reducing and recovering aqueous Pt(IV) ions under anaerobic conditions according to the spectral edge jumps reported in Fig 2 (see values in square brackets). C. metallidurans was below detection and therefore its spectrum is not shown. Anaerobic G. metallireducens and S. oneidensis reduced approximately half as much Pt relative to when Pt(II) is the starting aqueous species. It needs to be noted that the mismatch between the amount of Pt that was reduced to Pt(0) (XANES spectra) and the total amount of Pt that was recovered from solution (ICP data) corresponds with unreduced or partly reduced Pt. The energy of peak fluorescence in the XANES spectrum is a less reliable diagnostic given that the peak fluorescence of Pt(IV) aqueous standard occurs at a similar energy to reduced Pt (0) in biomass. However, the EXAFS spectra indicate that only the anaerobically grown S. oneidensis is capable of reducing significant amounts of Pt(IV) to Pt(0). The FT's of these EXAFS spectra do suggest that a minor fraction of the Pt(IV) could have been reduced by anaerobic G. metallireducens and aerobic S. oneidensis (see S2 Fig).

Platinum particle morphology
The characteristics of the Pt precipitates in and on the microbial biomass were studied using transmission electron microscopy (TEM). All bacterial suspensions that showed visible signs of Pt reduction were analyzed (i.e. black discoloration). The morphology of the platinum particles differed considerably according to the studied bacterial species and the recovered Pt-complex (Fig 3 and S3 Fig). , as could be concluded from the particle size distributions (S4 and S5 Figs). For example, 24% (with formate) to 80% (with H 2 ) of the total particle surface area was allocated to particles larger than 100 nm for the reduction of Pt(IV)Cl 6 2by G. metallireducens, compared to 15-23% for the reduction of Pt(II)Cl 4 2by the same bacterial species. Precipitated cisplatin also formed mainly larger particles (min. 71% of the particles were larger than 100 nm). The Pt-speciation influenced the location of the formed precipitates as well, as observed for B. toyonensis. Whereas PtCl 4 2precipitated as dispersed, small intra-and extracellular particles, large Pt clusters completely filled the cells in the case of PtCl 6 2- (Fig 3). Overall, both intra-and extracellular particles were observed for all three Pt-complexes, depending on the bacterial species. The increased importance of large particles in the particle surface area also corresponds to the increase in edge jump value for Pt(II)Cl 4 2-(see Fig 1 and S4 Fig). S. oneidensis precipitated particles mainly on the cell wall and in the periplasmic space. Depending on the conditions (Pt-complex, electron donor), C. metallidurans tended to precipitate platinum into larger clusters, which can be observed in the case of cisplatin where very large clusters were located near the bacterial cells. Numerous small particles can be observed in the presence of G. metallireducens, which can cluster together to bigger particles such as in case of PtCl 6 2-. A similar clustering was observed after the reduction of PtCl 6 2by B. toyonensis; cells were completely filled with precipitated platinum. P. stutzeri reduced the Pt-chloro complexes into larger particles as well.
Next to the applied bacterial species and recovered Pt-complex, also the choice of electron donor influenced the precipitation of platinum. In general, larger particles were formed in the presence of H 2 -gas, compared to the presence of formate. Small uniformly dispersed particles were only observed in case of formate induced reduction (with S. oneidensis and G. metallireducens) (S3 Fig), which is in contrast with the palladium (Pd) study from De Windt et al. [20], that observed more small Pd particles in case H 2 was used compared to formate.
Platinum precipitates formed in a previous study by Shewanella algae with Pt(IV) and lactate showed similarities with our results [13]. Pt particles of about 5 nm were observed in the periplasmic space, which is similar to the case of formate induced PtCl 6 2reduction by S. oneidensis, which resulted in particles of 4.8 nm (mean size), mainly formed in the periplasmic space and on the cell wall of the cells. In general, more intracellular Pt particles were formed in the case of the platinum(IV) chloro-complex, as was previously observed by Maes et al. when using halophilic mixed cultures [11].

Membrane integrity of axenic cultures during platinum recovery
The recovery of Pt-complexes and the prevalent conditions of the target wastewater will influence the viability of the present microorganisms and could affect their metal recovery potential. Furthermore, bacteria might behave differently towards the studied Pt-complexes according to the different intrinsic and chemical characteristics. Therefore, the cell viability of the axenic cultures was investigated during Pt-complex recovery experiments through flow cytometry based membrane integrity staining [23]. This staining makes it possible to distinguish cells with an intact cellular membrane, referred to as intact cells, from damaged cells. The influence of the Pt-speciation upon recovery was investigated using S. oneidensis, as the model organism (Fig 4).
The viability of the Shewanella culture decreased from approximately 10 9 intact cells mL -1 to below the detection limit (33.3 x 10 3 cells mL -1 ) within 2 hours after dosing of PtCl 4

2-
. For PtCl 6 2and cisplatin, the number of intact cells decreased to the detection limit within 4 and 6 hours, respectively. The addition of the Pt-tetraamine complex and carboplatin lowered the amount of intact cells until approximately 10 5 cells mL -1 . These complexes were not reduced during the experiment, suggesting a damaging effect of the Pt reduction and the formation of intra-and extracellular precipitates. Furthermore, carboplatin is less toxic than cisplatin in chemotherapy treatments since it is more stable (due to a more stable leaving group), which might explain the more limited interaction in this study as well [24]. Although this study showed a slower decrease in intact cells in the case of PtCl 4 2compared to PtCl 6 2-(based on the S. oneidensis culture), PtCl 6 2was previously found to be more toxic to C. metallidurans than PtCl 4 2-; minimal inhibitory concentrations of 39 mg L -1 and 3.4 mg L -1 were determined for respectively the Pt(II) and Pt(IV)-chloro complex [25]. In this study, the cell viability was however mainly linked to the precipitation of Pt particles instead of the intrinsic toxicity of the metal salts.
The PtCl 4 2complex was selected as a toxic and industrially relevant Pt-complex to study the vulnerability of the different bacterial organisms towards Pt-complex toxicity (Fig 5). Among the screened cultures, differences could be observed. The decrease in intact cells was almost identical for B. toyonensis and P. stutzeri; the detection limit was reached within 1 hour, indicating a severe damaging effect during Pt recovery. C. metallidurans appeared to be the most resistant species; 1.2 x 10 6 intact cells mL -1 were still measured after 2 hours and 1.3 x 10 5 intact cells mL -1 after 6 hours. This might be explained by the presence of metal resistance gene clusters in this species, enabling cell detoxification, as was demonstrated before for precious metals gold and silver [26,27]. Still, the amount of intact cells of all bacterial suspensions Overall, the affected viability of the cultures will have been caused by the combined effect of (1) the acidic pH, (2) the exposure to the Pt-complexes, as an increased cell membrane permeability by Pt(IV) ions was observed before [22] and (3) the precipitation of Pt particles, previously shown for an Acinetobacter species [28]. The substantial effect of the acidic pH on the viability is shown for the S. oneidensis suspension containing cisplatin; 3.7 x 10 8 intact cells mL -1 were still detected after 6 h when working at pH 5, while only 3.3 x 10 4 intact cells mL -1 were measured at pH 2 (i.e. limit of detection). Next to metal toxicity, the pH of the metal containing wastewaters will be another important challenge to obtain an effective biological metal recovery.

Interaction mechanisms between bacteria and Pt-species
The bacterial interaction in these biorecovery processes is believed to consist of two concomitant steps; (1) the initial sorption of Pt to the cell, which is followed by (2) the microbial reduction of the sorbed Pt-molecules [29,30]. Differences observed during the first sorptive step might be partly related to the different cell wall structure of G + and Gspecies. In general, the recovery of deprotonated complexes is favored by a low pH as functional groups on the bacterial surface become protonated at low pH [31]. Furthermore, the speciation and valence of Pt-complexes is highly dependent on the pH and chloride concentration. Chloro, hydroxyl or hydrated complexes can be formed depending on the conditions, characterized by a complicated chemistry [32][33][34][35]. For example, the Pt-chloro complexes PtCl 4 2and PtCl 6 2-, which are important species under acidic and saline conditions, are characterized by a negative charge at low pH [36]. This enables electrostatic interactions with positively charged binding sites such as amine groups and presumably results in effective recovery [12,36]. The observed slower recovery and precipitation of Pt(IV) can be explained by the sequential transformation of Pt(IV) through Pt(II) to Pt(0) [37]. Additionally, PtCl 6 2is expected to be more difficult to reduce, based on the slightly lower standard reduction potential of this Pt-complex: E 0 (Pt(IV)Cl 6 2-, PtCl 4 2-) = + 0.726 V vs. SHE compared to E 0 (PtCl 4 2-) = + 0.758 V vs. SHE [38]. The first reduction step of Pt(IV) to Pt(II) will have slowed down and limited the platinum recovery performance. Riddin et al. [37] proposed a dislocated two-step reduction of Pt(IV) by sulphate-reducing bacteria, in which Pt(IV) was reduced to Pt(II) by a cytoplasmic hydrogenase and Pt(II) further to Pt(0) by a periplasmic hydrogenase.
Furthermore, none of the bacterial species was able to recover the Pt-tetraamine complex. Since any sorption was lacking, the neutral complex Pt(NH 3 ) 4 Cl 2 will probably have formed, preventing any interaction with the bacterial surface and making a biological treatment ineffective. The uncharged cisplatin partitions (partially) into the hydrated complexes cis-[PtCl(NH 3 ) 2 (H 2 O)] + and cis-[Pt(NH 3 ) 2 (H 2 O) 2 ] 2+ when low chloride concentrations are present [33]. The formation of these positively charged Pt-complexes might explain the limited sorption by protonated functional groups at low pH. Still, complete recovery was possible at low pH in the presence of H 2 -gas. The limited sorption of, for example, a residual amount of the uncharged mother compound can thus be sufficient to induce the full reduction of this Pt-complex, as long as a capable microbial species and an electron donor are present. The second chemotherapy complex carboplatin has been observed to mainly remain stable in wastewater and was characterized in our study by a limited recovery under all tested conditions [39]. The formation of Pt precipitates was only observed for cisplatin and not for carboplatin. The sorptive recovery of both chemotherapy complexes was studied before by using activated sludge, revealing the least sorption for carboplatin (70% vs. 96% for cisplatin) [32].
Different mechanisms might be responsible for the recovery potential of the studied bacterial species. The recovery of platinum was investigated before using the related marine species Shewanella algae, which was able to recover 90% PtCl 6 2within 1 hour using lactate as electron donor (C 0 = 200 mg L -1 Pt at pH 7) [13]. To our knowledge, the recovery of platinum by the anaerobic species Geobacter metallireducens has not been studied yet, but the reduction of palladium was demonstrated recently by the related Geobacter sulfurreducens [21,40]. The dissimilatory metal reducing bacteria Shewanella and Geobacter were found to use different hydrogenases and cytochromes, being present in the outer membrane, periplasm or cytoplasm, to transfer electrons to reduce the metals [40,41]. The heavy metalresistant and metallophilic species Cupriavidus metallidurans is well-studied for the reduction and precipitation of palladium and gold, initiated by the expression of different metal resistance genes [26,42,43]. Gauthier et al. [14] demonstrated the recovery of platinum and palladium by Cupriavidus metallidurans and Cupriavidus necator species in the presence of hydrogen gas; 70-74% Pt and 96-100% Pd were recovered from a mixed metal acidic leachate (pH 1.4; 24 h). The Pseudomonas stutzeri species has been shown to reduce selenate and selenite and to produce silver nanoparticles [44,45]. It is hypothesized that siderophores, produced by the Pseudomonas species, are involved in a detoxification strategy of the species, by extracellularly complexing and reducing various metals [46]. The only studied G + species, Bacillus toyonensis, has not been utilized yet in metal recovery studies, although Bacillus species have been shown to reduce palladium [47]. The recovery potential of these axenic cultures should be further explored under real stream conditions.