Zinc isotope fractionation during surface adsorption and intracellular incorporation by bacteria

Chemical Geology 366 (2014) 42–51

Contents lists available at ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Zinc isotope fractionation during surface adsorption and intracellular incorporation by bacteria Fotios-Christos A. Kafantaris 1, David M. Borrok ⁎,2 Department of Geological Sciences, University of Texas at El Paso, El Paso, TX 79968, USA

a r t i c l e

i n f o

Article history: Received 14 June 2013 Received in revised form 26 October 2013 Accepted 15 December 2013 Available online 25 December 2013 Editor: Carla M. Koretsky Keywords: Zn Zinc Isotopes Adsorption Uptake Bacteria

a b s t r a c t Zinc (Zn) isotopes are fractionated during biogeochemical processing by microorganisms. Uncertainties remain, however, regarding the roles of cell surface adsorption and speciation of aqueous Zn on the extents of isotopic fractionation. In this study, we conducted bacterial surface adsorption and intracellular incorporation experiments using Zn and representative Gram-positive (Bacillus subtilis) and Gram-negative (Pseudomonas mendocina, Escherichia coli) bacterial species, as well as a natural bacterial consortium derived from soil. Under conditions of high Zn:bacteria ratio, surface complexes preferentially incorporated the heavier isotopes of Zn, resulting in an average Δ66Znadsorbed-solution of + 0.46‰ (αadsorbed-solution ≈ 1.00046). Adsorption experiments conducted under conditions of low Zn:bacteria ratio appear to have been complicated by the presence of dissolved organic exudates that competed with surface functional group sites for Zn. We were able to empirically model this process to show that very small amounts of Zn-organic complexes with fractionation factors in the range of α = 1.002 to 1.003 could account for the observed δ66Zn of the experimental solutions. For the intracellular incorporation experiments, the presence of 0.2 and 2 mg/L of Zn (as Zn-citrate) resulted in a Δ66Znincorporated-solution ranging from −0.2‰ to +0.5‰, depending upon the bacterial species and the growth phase. The addition of 0.2 and 2 mg/L Zn2+ to the growth medium appeared to create a metal stress response (or at least a change in metal processing) in P. mendocina that resulted in a positive Δ66Znincorporated-solution of up to + 2.04‰. Our study suggests that Zn isotopes have the potential to be used to elucidate metal-binding pathways associated with microorganisms in natural systems, but that the interpretation of these effects is likely complicated by factors such as competing surface interactions and differences in bacterial species and metal speciation. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Zinc (Zn) is a critical element for biological functioning (Olhaberry et al., 1983; Shankar and Prasad, 1998; Hambidge, 2000; Andreini et al., 2006a,b; Maret, 2009), and the stable isotopes of Zn are substantially fractionated during biological processing (e.g., John et al., 2007). For these reasons, a number of investigations have examined Zn isotopic fractionation during plant uptake (Weiss et al., 2004; Viers et al., 2007; Moynier et al., 2009; Arnold et al., 2010; Caldelas et al., 2011; Jouvin et al., 2012; Tang et al., 2012) and interaction with yeast (Zhu et al., 2002), diatoms (Gélabert et al., 2006; John et al., 2007) and bacteria (Wanty et al., 2013). Moreover, Peel et al. (2009) and Andersen et al. (2011) have shown that Zn isotopes of particulate matter and diatom frustules, respectively, can serve as proxies for the biochemical cycling of Zn.

⁎ Corresponding author. Tel.: +1 337 482 2888. E-mail address: [email protected] (D.M. Borrok). 1 Current address: Department of Earth Sciences, Indiana University–Purdue University Indianapolis, Indianapolis, IN 46202, USA. 2 Current address: School of Geosciences, University of Louisiana at Lafayette, Lafayette, LA 70504, USA. 0009-2541/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chemgeo.2013.12.007

Despite the increasing interest in understanding how Zn isotopes fractionate during biological cycling, there are still substantial uncertainties. In plant systems, for example, variable concentrations of Zn, as well as differences in biochemical pathways among different plant species, can lead to substantially different Zn isotopic signatures among plants (Caldelas et al., 2011; Jouvin et al., 2012; Tang et al., 2012). In microorganism systems, the relative importance of surface complexation versus intracellular incorporation is uncertain. For example, Gélabert et al. (2006) showed that during sorption and uptake, freshwater and seawater diatoms were preferentially associated with heavier Zn isotopes (Δ66Zncell-solution of +0.35‰ and +0.27‰, respectively), while in a separate study John et al. (2007) found that seawater diatoms incorporated the lighter Zn isotopes (Δ66Zncell-solution ranged from −0.20‰ to −0.80‰, depending on the Zn homeostasis pathway), but adsorbed the heavier Zn isotopes. In addition, there is still uncertainty regarding the magnitude of Zn isotopic fractionation among different bacterial species and under differing conditions of Zn stress. In order to try and address some of these uncertainties, we investigated the isotopic fractionation of Zn during surface adsorption and intracellular incorporation by representative Gram-positive (Bacillus subtilis) and Gram-negative (Pseudomonas mendocina and Escherichia coli) bacterial species and a natural bacterial consortium isolated from

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

soil. The laboratory bacteria strains were chosen because of their different surface chemical properties. Moreover, Zn homeostasis mechanisms for each of these bacterial species have been described previously (e.g., Beard et al., 2000; Outten and O'Halloran, 2001; Cánovas et al., 2003; Moore and Helmann, 2005). We conducted separate batch adsorption experiments and bacterial growth experiments under a variety of chemical conditions. Samples of the associated fluids and/or bacterial cells (collected as a function of pH, time, or growth phase) were prepared and analyzed for their concentrations and isotopic compositions of Zn. Here we present the results of these experiments and discuss their implications for understanding Zn isotope variations in natural systems. 2. Methods All volumetric flasks, sample bottles, test tubes, and pipette tips used for experiments were acid-washed in sub-boiling 10% HCl and rinsed 3 times with ultra-pure (18.2 MΩ) water. Chemical reagents and growth media were prepared using ultra-pure water. Ultra-pure acids and bases were used for experiments and sample digestion and preparation. 2.1. Bacteria collection, growth, and harvesting B. subtilis (Gram-positive), P. mendocina, and E. coli (both Gramnegative) were chosen as representative laboratory strains of bacteria and were acquired through the Fein lab at the University of Notre Dame. The cellular surface of Gram-positive bacteria consists principally of a thick and rigid peptidoglycan layer (25 to 30 nm) bound together with secondary polymers such as teichoic acids (Konhauser, 2006). The surface of Gram-negative bacteria consists of a flexible outer membrane of lipopolysaccharides covering a thin (3 nm) peptidoglycan layer (Beveridge and Koval, 1981; Konhauser, 2006). B. subtilis was grown by inoculating (via sterile loop) an autoclaved medium of 30 g of Trypticase Soy Broth (TSB) plus 10 g of yeast extract dissolved in 2 L of ultra-pure water (Kenney et al., 2012). P. mendocina and E. coli were grown in the same manner but in a basal medium modified slightly from that described by Hersman et al. (2001). The nutrient medium included the addition of the following solutes per 1 L of ultra-pure water: 0.5 g K2HPO4, 1.0 g NH4Cl, 0.2 g MgSO4·7H2O, 0.2 g CaCl2·2H2O, 8.33 g succinate disodium salt, 30 mM Fe-EDTA, 4.77 g HEPES buffer, 1 g of glucose (for E. coli only), and 0.125 mL of trace elements solution (5 mg MnSO4·H2O, 6.5 mg CoSO4·7H2O, 3.3 mg ZnSO4 and 2.4 mg MoO3 per 100 mL of water; Hersman et al., 2001). Zn is an essential trace nutrient and the bacteria cannot grow effectively without it. It was eliminated from the growth experiments but replaced with another form of Zn (Zn-citrate or Zn+2). The elimination of the ZnSO4 was important so we could have an isotopically homogeneous starting point. For all experiments, bacteria were grown aerobically at 25 °C in an incubator-shaker. All inoculations and sample handling were conducted under sterile conditions in a Class 100 clean bench. A consortium of natural bacterial species was obtained from nearsurface soils in the El Paso, Texas, region. Five hundred grams of soil were sampled using a sterile plastic bag and nitrile gloves. Thirty grams of the soil sample were used to inoculate 2000 mL of sterilized basal medium. After several days of growth, the suspended soil was allowed to settle and 1 mL of the top part of the solution was used to inoculate a second sterilized flask of basal medium. This transfer “diluted out” the soil used for the initial inoculation (e.g., Borrok et al., 2004). Even though the natural consortium was used as an analog for natural bacterial species, we cannot exclude the possibility that other microorganisms like fungi may have been present. Furthermore, it is likely that the microbial populations shifted slightly from batch to batch based on small changes in soil conditions and inoculation. Bacteria used for adsorption experiments were harvested from the media after 48 h through centrifugation. Previous work has shown that the laboratory strains are near the stationary growth phase after 48 h (e.g., Guiné et al., 2006; Wei et al., 2011). It was not possible to

43

determine the growth phase for the natural consortium because multiple bacterial species were present. After the bacteria pellet was removed from the growth medium, it was re-suspended in a 15 mL test tube in a wash solution of 0.02 M MgSO4 adjusted to a pH of 1.5. The bacterial suspension was centrifuged (7 min at 3800 g) and the supernatant removed and replaced with fresh wash solution. This washing process was repeated 5 times to remove any surface-bound Zn acquired from the growth medium (e.g., Navarrete et al., 2011; Kenney et al., 2012). During the final wash, the bacterial pellet was centrifuged at 5000 g for 1 h and the supernatant removed. The moist mass of the bacteria pellet was recorded and the pellet was used for experiments. The effectiveness of the washing procedure was evaluated by measuring the concentrations of Zn in the supernatant solutions using an ICP-OES.

2.2. Variable pH surface adsorption experiments Batch adsorption experiments were designed to isolate Zn complexation reactions associated with bacterial surfaces. Bacterial surfaces contain a variety of organic-acid functional group sites (phosphoryl, carboxyl, hydroxyl, and some sulfhydryl, etc., depending on the species and Gram-type) that deprotonate as a function of pH (e.g., Beveridge, 1989; Kelly et al., 2002; McClure et al., 2003; Guiné et al., 2006; Mishra et al., 2010). The deprotonated sites form surface complexes with Zn and other metals, impacting the transport and fate of metals in natural systems (e.g., Yee and Fein, 2001; Ngwenya et al., 2003; Ginn and Fein, 2008). These surface interactions are also a first step in intracellular uptake of metal (e.g., Borrok et al., 2005a). Batch adsorption experiments were conducted as a function of pH in experimental solutions with constant Zn/bacteria ratios using live cells of B. subtilis or the natural consortium and using live and dead cells of P. mendocina. Experimental component concentrations are summarized in Table 1. Although the live cells were not thought to be actively metabolizing in the nutrient-free experimental solutions, the dead cell experiment was done to further test this possibility. In this case, P. mendocina cells were suspended in the experimental stock solution and treated for 12 h with UV radiation prior to the addition of Zn. We tested the effectiveness of the UV treatment by trying to culture the affected cells and no growth was observed. All batch adsorption experiments were performed in a stock solution with a 0.01 M NaClO4 electrolyte. A measured volume of stock solution was amended with Zn (from an in-house Zn wire standard) and a pre-weighed bacterial pellet was suspended in the mixture. A control experiment where no bacterial pellet was added to the stock solution was also conducted to confirm that Zn did not interact with the experimental apparatus. The control experiment was subjected to the same procedures, including filtration and sample processing, that were part of the other experiments. The final Zn and bacteria concentrations for each experiment are provided in Table 1. Thirty milliliters of the homogenized experimental stock solution were added to each of eight 50 mL reaction vessels. The pH of each vessel was individually adjusted between pH 2.5 and ~ 7 using 0.2 mL aliquots of dilute HNO3 or dilute NaOH (this small volume addition had a negligible impact on Zn concentrations). The pH of the experimental solutions was measured using a Thermo Scientific Orion Star Log™ pH Meter that was calibrated prior to each experiment. The reaction vessels were placed on a shaker table for 2 h after which time the final (equilibrium) pH of each vessel was recorded. Previous work has shown that a period of 2 h is enough time for bacteria–metal surface complexes to reach apparent equilibrium (e.g., Fowle and Fein, 2000). The reaction vessels were then centrifuged and the supernatant filtered through a pre-cleaned (with 2% nitric acid) 0.45 μm nylon filter. The supernatant was preserved for later analyses by adding 0.2 mL of concentrated HNO3. The solutions were later analyzed for their Zn concentrations using an ICP-OES and were further prepared for isotopic analysis (see below).

44

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

Table 1 Experimental conditions. Exp.

Type

Cell conditions

Metal

Bacterial species

Zn (mg/L)

Bacteria conc. (g/L)

A1 A2 A3 A4 A5 A6 I1 I2 I3 I4 I5 I6 I7

Adsorption Adsorption Adsorption Adsorption Adsorption Adsorption Incorporation Incorporation Incorporation Incorporation Incorporation Incorporation Incorporation

Alive — not metabolizing Alive — not metabolizing Alive — not metabolizing Alive — not metabolizing Alive — not metabolizing Dead Alive — metabolizing Alive — metabolizing Alive — metabolizing Alive — metabolizing Alive — metabolizing Alive — metabolizing Alive — metabolizing

Zn2+ Zn2+ Zn2+ Zn2+ Zn2+ Zn2+ Zn-citrate Zn-citrate Zn-citrate Zn-citrate Zn-citrate Zn2+ Zn2+

P. mendocina P. mendocina B. subtilis B. subtilis Natural consortium P. mendocina P. mendocina P. mendocina P. mendocina Natural consortium E. coli P. mendocina P. mendocina

2.6 20.0 2.1 19.6 8.8 3.1 0.2 2.0 20.0 2.0 2.0 0.2 2.0

8.08 1.28 6.29 1.07 9.96 2.68 NA NA NA NA NA NA NA

NA = not applicable.

2.3. Intracellular incorporation experiments

2.4. Analysis of Zn concentrations

Intracellular incorporation experiments were conducted to examine how the cellular uptake and internal processing of Zn fractionates Zn isotopes. We attempted to isolate these reactions from the surface reactions by carefully washing the surfaces of the cells prior to their digestion. In order to test for possible differences in Zn isotopic fractionation attributable to aqueous Zn speciation (and its subsequent homoeostasis pathways) two sets of experiments were completed, one using Zn-citrate and the other using Zn+2. Zn-citrate was selected because it serves as a proxy for Zn bound to organic acids and natural organic matter. For example, citric acid — among other carboxylates — is a principal organic anion exuded from root plants in soil (Roelofs et al., 2001). In addition to Znorganic complexes, Zn+2 is also a primary form of Zn in many natural waters. Cellular incorporation experiments were conducted with differing concentrations of Zn (0.2 mg/L, 2 mg/L, and 20 mg/L) and with different bacterial species (P. mendocina, E. coli, and natural consortium). The various component combinations used for the intracellular incorporation experiments are summarized in Table 1. Erlenmeyer flasks of sterilized basal growth media were inoculated with the appropriate bacterial species or consortium using the procedures described above for the adsorption experiments. The flasks were gently agitated at room temperature using an orbital shaker for the duration of the experiments. Immediately prior to sampling, the flasks were agitated more vigorously to ensure homogenization of the bacterial cells in the medium solution. One milliliter samples were collected daily from the experimental solutions using a sterile pipette and analyzed for their optical densities at a wavelength of 600 nm using a Hach™ UV–vis spectrophotometer. The UV–vis measurements, a proxy for bacterial cell counts, were used to construct growth curves for each of the experiments. Samples for isotopic analysis were collected from each experiment during different phases of growth roughly corresponding to the log, stationary, and death phases. The sampling process was identical to that described for the adsorption experiments. However, in this case the bacteria pellets were washed and processed. After harvesting, the pellets were washed 5 times using the same electrolyte solution used for cleaning bacteria for the adsorption experiments. The wash solutions were collected to evaluate the effectiveness of the washes in removing surface bound Zn. After the final wash, the bacteria pellets were digested in 3 to 4 mL of concentrated HNO3. The digests were twice evaporated to dryness and re-dissolved in 1 mL of concentrated HNO3 in order to oxidize dissolved organics. The final solutions were diluted with pure water to 10 mL. Sample splits were analyzed using ICP-OES, while the remaining samples were preserved for isotopic preparation and analysis.

Concentrations of Zn were determined for the supernatant solutions collected from the adsorption experiments and the supernatant and bacterial digests collected from the intracellular incorporation experiments. These solutions were analyzed using a Perkin-Elmer™ Optima 5300DV ICP-OES instrument with Zn standard solutions diluted from a Zn CertiSPEX™ atomic adsorption standard. Using this method, the reporting limit for Zn was less than 8 ppb. Uncertainty, based on replicate analyses, was less than ±5%. 2.5. Isotopic preparation and analysis Supernatant solutions and bacterial digests were prepared for isotope analysis following the method of Borrok et al. (2007). Briefly, 0.1 to 5 mL of the sample solutions were evaporated to dryness (to achieve masses in the range from about 150 to 1000 ng of Zn) and dissolved in 10 N HCl for loading onto an anion-exchange column of pre-cleaned 100–200 mesh AG MP-1 resin (Bio-Rad®). Matrix elements were eluted from the column while retaining Zn by passing through approximately 7 mL of HCl with a normality N1. Zn was then eluted from the column by passing through pure water. The separated Zn was evaporated and dissolved in a volume of 2% HNO3 to achieve a concentration of ~150 μg/L. These Zn solutions were analyzed using an ICP-OES, which confirmed 100% recovery of Zn within the analytical uncertainty of the analyses. All column separations were performed in a Hepa-filtered laminar flow clean bench in a class 100 clean room. Samples were analyzed for their Zn isotopic compositions using a Nu Instruments™ Multi-collector ICP-Mass Spectrometer (MC-ICP-MS) housed in the Center for Earth and Environmental Isotope Research at the University of Texas at El Paso. The analytical procedures used for Zn analysis have been described previously (e.g., Borrok et al., 2007; Thapalia et al., 2010; Aranda et al., 2012) and are only briefly summarized here. The samples were introduced to the instrument using a desolvating nebulizer system (Nu DSN 100). The voltages for masses 64, 66, 67, and 68 were simultaneously measured. Here we report δ66Zn, in standard delta notation which is the difference in the 66Zn/64Zn ratio of the sample relative to that of standard multiplied by 1000 to achieve units of “per mil” as shown in Eq. (1). 2


 3 66 Zn=64 Zn sample 6 7  −15  1000: δ Zn ¼ 4 66 Zn=64 Zn STD 66

ð1Þ

ave

In this case, the bracketing standards we used (also passed through the column chemistry) were the same as the Zn used for the

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

The necessary experimental designs (described above) do not always allow for a complete mass and isotopic mass balance to be performed. For example, in the Zn adsorption experiments, we can only monitor the concentrations and isotopic changes of Zn in the solution phase. This is because it is not possible to fully eliminate the co-mingling of adsorbed Zn on the bacteria pellets with free Zn in the aqueous solution that surrounds the pellets. Attempts to isolate the adsorbed Zn through centrifugation and washing may influence the isotopic composition of this reservoir. However, in an effort to add confidence to our assumption of mass balance of the system calculated from the solution reservoir alone, we performed a control experiment with 10 mg/L of Zn that did not contain bacteria. ICP-OES measurements of the solution confirmed that Zn did not adsorb to the apparatus or filter during processing. The measured δ 66 Zn of the control sample was − 0.041‰ ± 0.06‰. Hence, the assumption of mass balance between Zn in solution and Zn associated with the bacterial pellet is supported. It is similarly challenging to demonstrate mass balance in the Zn incorporation experiments, as the amount of Zn added to the medium solution is much greater than that taken up by the bacteria. Based on the mass of Zn measured in the digested bacterial pellets, we calculated that less than 0.5% of the total Zn in solution is transferred to the inside of the bacteria. Hence, the δ66Zn of Zn in solution is not expected to appreciably change in these experiments. Nevertheless, we did measure the δ66Zn of the solution phase for the incorporation experiments with Zn-citrate to test the assumption that changes in the solution phase were small. The δ66Zn values for the 11 solutions tested were quite uniform with a 2σ variation of 0.15‰. Hence our assumption that isotopic fractionations in solution are small and not measurable was confirmed. 3. Results and discussion 3.1. Variable pH adsorption experiments Adsorption experiments conducted as a function of pH can be broken into 2 groups; (1) those conducted at “high” Zn:bacteria component ratios, and (2) those conducted at a “low” Zn:bacteria component ratios (see Table 1). The high Zn:bacteria ratio experiments were designed such that there was much more Zn available (particularly at lower pH values) than there were bacterial surface functional group sites. On the other hand, at low Zn:bacteria ratios there was an abundance of

a

100

80

Zn adsorbed (%)

2.6. Mass balance constraints

functional group sites relative to the amount of Zn in solution. This generalized comparison is based on the molar abundance of Zn relative to that of functional group sites estimated by Borrok et al. (2005b), assuming the formation of monodentate binding complexes. Fig. 1a and Table 2 present the results of the adsorption experiments conducted at high Zn:bacteria ratios. B. subtilis and P. mendocina bacterial species display similar patterns of increasing adsorption as a function of increasing pH. However, only a modest amount of adsorption occurs at the lower pH values. For example, at pH 6 only ~ 10% and ~25% of the total Zn is adsorbed to B. subtilis and P. mendocina, respectively. Most of the adsorption (i.e., the adsorption edge) in this system occurs above pH 6. This pattern likely reflects the limited amount of deprotonated functional group sites available at low pH relative to the amount of Zn in the system. The available binding sites are likely saturated with Zn. Further deprotonation of functional group sites at neutral pH provides more available sites for adsorption resulting in the observed adsorption edge. Geochemical modeling (using Visual Minteq™), however, does indicate the possibility of the precipitation of hydrozincite, Zn5(CO3)2(OH)6, for a limited number of samples (three) that were subjected to pH N 7.1. The possible amounts of precipitation are difficult to estimate, as the concentration of Zn in solution available for precipitation changes with the extent of surface adsorption. If precipitation did occur in these samples, it would appear in Fig. 1a as if there were an increase in the extent of adsorption. Fig. 1b and Table 2 present the results of adsorption experiments conducted at low Zn:bacteria ratios. Despite some small differences in

60

40

20

0 2

3

4

5

6

7

8

pH

b 100 80

Zn adsorbed (%)

experiments (a Zn wire standard for the adsorption experiments and some incorporation experiments and Zn-citrate for the remaining incorporation experiments; Table 1). Using this convention, a δ66Zn of zero means no fractionation. Mass bias was corrected using the standardsample-standard bracketing technique (Borrok et al., 2007; Thapalia et al., 2010; Aranda et al., 2012). The addition of an external normalization element (i.e., a Cu dopant) did not provide better results, as the mass bias was too stable to develop a meaningful relationship between the dopant and sample. Sample measurements were generally replicated 2 to 3 times with an external precision better than ± 0.1‰. However, selected samples were analyzed more (as much as 11 times over the course of 2 months) and the average precision of these measurements was ± 0.058‰. Procedural replicates also fell within this range of uncertainty. Sixteen procedural blanks for column separations were completed and they contained on average 3.3 ± 4.1 ng of Zn. The blank contributions were too small to impact the measured ratios. Finally, the mass dependency of the Zn isotope measurements was checked by plotting δ66Zn vs. δ67Zn/1.5 and δ68Zn/2 (not shown). The best-fit line had an R2 value of 0.999 and a slope of 1.005, confirming the proper mass dependency.

45

60

40 P. mendocina - dead Natural consortium B. subtilis P. mendocina

20

0

2

3

4

5

6

7

8

pH Fig. 1. Percentage of Zn adsorbed to bacterial surfaces plotted as a function of pH under (a) conditions of high Zn:bacteria ratio and (b) conditions of low Zn:bacteria ratio (see Table 1 for component concentrations). Error bars associated with Zn concentration data fall within the size of the symbols.

46

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

Table 2 Results from adsorption experiments. Exp.

pH

Fraction adsorbed

δ66Zn averagea,b

2σ

nc

(solution) A1

A2

A3

A4

A5

A6

a b c d

4.73 5.15 4.03 3.07 2.55 6.51 6.38 6.15 5.79 6.57 7.10 7.29 7.47 6.28 5.09 4.60 3.57 2.87 6.71 3.69 2.93 4.73 6.38 5.43 6.98 5.11 5.45 5.59 6.74 6.7 6.98 7.51 3.6 6.12 5.12 2.92 6.82 6.57 6.2 3.61 4.34 5.81 5.36 4.73 5.22 3.99 3.15 6.7 7.28

0.75 0.83 0.59 0.27 0.14 0.88 0.91 0.88 0.22 0.39 0.39 0.74 0.80 0.28 0.24 0.22 0.17 0.13 0.95 0.40 0.23 0.80 0.91 0.89 0.94 0.87 0.09 0.10 0.20 0.34 0.65 0.86 0.01 0.08 0.07 0.01 0.82 0.80 0.76 0.26 0.43 0.69 0.57 0.51 0.80 0.60 0.36 0.88 0.89

0.443 1.509 0.511 0.140 0.250 2.139 0.893 1.374 −0.089 −0.177 −0.355 −0.335 −0.371 −0.199 −0.043 −0.233 0.045 −0.076 0.694 0.407 0.232 0.244 −0.015 0.128 0.004 0.192 −0.036 −0.116 −0.074 −0.140 −0.247 −0.288 −0.157 −0.108 −0.055 0.007 −0.126 −0.202 −0.171 0.018 0.004 −0.097 −0.005 0.020 0.965 0.037 −0.002 2.683 0.389

0.006 0.033 0.006 0.014 0.101 0.045 0.104 0.100 0.107 0.059 0.089 0.064 0.091 0.071 0.072 0.019 0.035 0.053 0.034 0.021 0.040 0.039 0.078 0.061 0.062 0.061 0.060 0.039 0.053 0.090 0.034 0.053 0.024 0.066 0.092 0.015 0.017 0.014 0.046 0.019 0.057 0.039 0.022 0.036 0.073 0.091 0.104 0.055 0.048

2 2 2 2 2 2 2 2 3 11 3 3 3 3 2 2 2 2 2 2 2 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 2 4

Empirical modeld f Znfree

f Znorg

0.15 0.07 0.30 0.60 0.76 0.02 0.00 0.02 0.76 0.59 0.59 0.24 0.18 0.70 0.74 0.76 0.81 0.85 0.02 0.57 0.74 0.17 0.06 0.08 0.03 0.10 0.91 0.90 0.80 0.66 0.35 0.14 0.99 0.92 0.93 0.99 – – – – – – – – – – – – –

0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.016 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 0.005 – – – – – – – – – – – – –

Isotopic ratio of the experimental solution. δ66Zn of the bracketing standard was zero per mil. Number of analytical replicates. The relative proportions of Zn associated with free Zn+2 and dissolved Zn-organic complexes that were estimated for the empirical model. See text for details.

component concentrations (Table 1), B. subtilis, P. mendocina (both live and dead cells), and the natural consortium show roughly similar extents of adsorption as a function of pH. In this case, however, the abundance of available functional group sites — due to the high bacterial concentrations — results in a different adsorption pattern than that observed for the high Zn:bacteria experiments (Fig. 1a). In this case, the abundance of available sites greatly increases the adsorption of Zn at low pH. For example, by pH 5, at least 50% of the total Zn was adsorbed in each of the experimental systems. The adsorption of Zn reached a maximum of about 95% at pH ~ 6 for the B. subtilis and P. mendocina systems, whereas peak adsorption for the natural consortium was at pH ~ 6.8. Precipitation of hydrozincite (or other Zn minerals) was not expected in these experiments, as the pH for all but one sample did not exceed 7.0, and the concentrations of Zn used in these experiments were sufficiently low to limit this process.

3.1.1. Zn isotopes in high Zn:bacteria ratio adsorption experiments Fig. 2a presents the results of Zn isotopic measurements from the high Zn:bacteria ratio adsorption experiments plotted as a function of the fraction of Zn adsorbed. For comparison we have also included the results from previous Zn adsorption experiments with B. subtilis that were conducted under nearly identical conditions by Borrok et al. (2008). Despite some variability, particularly when very low amounts of Zn were adsorbed, the δ66Zn of the experimental solutions decreased as a function of the fraction of Zn adsorbed (Fig. 2a). This suggests that the heavier isotopes of Zn were preferentially adsorbed. It is unlikely that the precipitation of hydrozincite influenced the isotopic data. Previous work has demonstrated that the precipitation of hydrozincite from solution does not measurably fractionate Zn isotopes (Wanty et al., 2013). The magnitude of Zn isotopic fractionation was roughly similar for B. subtilis and P. mendocina. Hence, in an effort to develop a fractionation

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

a 0.5

0

δ66Zn

-0.5 Equilibrium Model

-1 Rayleigh Model

-1.5

-2

0

0.2

0.4

0.6

0.8

1

Fraction adsorbed

b

3 P. mendocina - dead Natural consortium B. subtilis P. mendocina P. mendocina - model B. subtilis - model

2.5

δ66Zn

2 1.5 1 0.5

47

functional group sites are similar to those described for Zn-acetate, -phosphate and -citrate complexes (Borg and Liu, 2010; Fujii and Albarède, 2012), it would suggest that these bonds are much stronger 2 than dissolved Zn(H2O)+ 6 , assuming that bond length and bond strength correlate. For example, the shortest Zn–O distance for aqueous Zn coordinated with six water molecules is 2.062 Å, while the Zn–O for Zn-acetate complexes is 1.953 Å (Borg and Liu, 2010). Acetate is thought to be a reasonable proxy for carboxylic Zn-organic site complexes (e.g., Yee and Fein, 2003). Ab initio theoretical calculations by Fujii and Albarède (2012) have also predicted the preferential incorporation of 66Zn over 64Zn in Zn-phosphate and -citrate complexes in plants. Hence from a theoretical standpoint, our observations of isotopically heavier Zn associated with bacterial surface complexes seem reasonable. In addition to theoretical calculations, our findings for Grampositive and Gram-negative bacterial species are similar to experimental work by Gélabert et al. (2006) that measured separation factors Δ66Znsolid-solution of + 0.43‰ and + 0.27‰ for of the sorption of Zn onto two separate diatom species. Wanty et al. (2013) demonstrated that hydrozincite precipitation alone does not fractionate Zn isotopes. However, when hydrozincite precipitated on the surfaces of a natural bacterial consortia, it incorporated the heavier Zn isotopes (Δ66Znbacteria-solution = + 0.35‰), likely inheriting the effect from surface adsorption. Our results are also similar in magnitude to the fractionation associated with the complexation of dissolved Zn with the high-affinity sites (HAS) of purified humic acid Δ66ZnHAS-solution of +0.40‰, Jouvin et al., 2009). Finally, the magnitude of Zn isotopic fractionation measured here is comparable to that for adsorption onto some mineral substrates such as ferryhydrite (Δ66Znsolid-solution of +0.52‰; Balistrieri et al., 2008).

0 -0.5 0

0.2

0.4

0.6

0.8

1

Fraction adsorbed Fig. 2. δ66Zn of experimental solutions plotted as a function of the fraction of Zn adsorbed under (a) conditions of high Zn:bacteria ratio and (b) conditions of low Zn:bacteria ratio. The additional gray squares in panel a are data from Borrok et al. (2008) collected for B. subtilis under identical experimental conditions. The gray solid line and curve in panel a represent the best-fit equilibrium and corresponding Rayleigh model fits for all the isotope data, respectively (αadsorbed-solution = 1.00046; Δ66Znadsorbed-solution = +0.46‰ for the equilibrium model). Red and blue solid curves in both panels represent the empirical model fits of the isotopic data in the presence of dissolved Zn-organic complexes (see Section 3.1 for details). 2σ uncertainties for δ66Zn are plotted, but largely fall within the size of the symbols.

factor that can be useful in natural systems where multiple bacterial species are present, we attempted to fit the entire data set using both equilibrium and Rayleigh fractionation models (Fig. 2a). Both models were based on best-fit linear regressions of the data forced to go through the origin. The R2 value of the fit for the equilibrium model (eliminating the 3 largest outliers) is 0.77. Based on the best-fit model we estimate an overall equilibrium separation factor (Δ66Znadsorbed-solution) of +0.46‰ (an αadsorbed-solution ≈ 1.00046). The best-fit for the Rayleigh model (eliminating the 3 largest outliers) was achieved with a separation factor of + 0.32‰ (αadsorbed-solution ≈ 1.00032; Fig. 2a). Although the data collected under the highest amounts of adsorption are limited, it appears that the equilibrium fractionation model provides a better fit for the available data. This result is consistent with previous studies that have examined Zn isotopic fractionation during adsorption onto mineral substrates (e.g., Pokrovsky et al., 2005; Balistrieri et al., 2008). The fractionation process is most likely controlled by the relative strength (i.e., coordinating environment and corresponding bond stiffness) of the bacterial surface complexes formed with Zn relative to complexes (e.g., Urey, 1947; that of dissolved Zn forming Zn(H2O)+2 6 Schauble, 2004). If bonding relationships for bacterial-surface organic

3.1.2. Zn isotopes in low Zn:bacteria ratio adsorption experiments Fig. 2b presents the results of Zn isotopic measurements from the low Zn:bacteria adsorption experiments. In this case, the δ66Zn of the experimental solutions is characterized by positive values for both P. mendocina (live and dead cells) and B. subtilis at the highest fractions of Zn adsorbed. For example, with both live and dead cells of P. mendocina, the δ66Zn started slightly positive (from 0 to 70% adsorption), but then increased to N+2.0‰ when the Zn fraction adsorbed was N0.9. This means that the lighter Zn isotopes were preferentially associated with the bacteria in these experiments. This is wholly contrary to the behavior observed for the high Zn:bacteria adsorption experiments (Fig. 2a). The δ66Zn of the solution for the natural consortium experiment was within error of zero when the fraction of adsorption was less than about 0.7, but was slightly negative (average −0.22‰) when more Zn was adsorbed (Fig. 2b). This result is more consistent with the finding of the high Zn:bacteria ratio experiments, but the magnitude of the fractionation is smaller by comparison. Several explanations are possible for the differences in the Zn isotopic behavior observed in the high versus low Zn:bacteria ratio adsorption experiments. As discussed above, precipitation reactions do not appear to have influenced the Zn isotopic data. Other possible explanations include, (1) differences in cell surface bonding chemistry, (2) the release of intracellular Zn accumulated during the growth process, or (3) the presence of dissolved Zn-organic complexes. The first possibility can generally be ruled out because it is unlikely that strongly-bound Zn surface complexes, regardless of their chemistries, would preferentially incorporate the lighter isotopes of Zn. It is also difficult to conceive of a functional group moiety that would greatly influence the Zn isotopic signature at higher pH values in the low Zn:bacteria experiments, while not influencing Zn isotopes at all over the same pH range in the high Zn:bacteria experiments. A second possibility is that small amounts of Zn internally incorporated in the bacteria during growth were released (or somehow became isotopically equilibrated) with the Zn solutions used for the adsorption experiments. Although the surfaces of the bacteria were washed to

48

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

remove surficial Zn prior to experimentation, the bacterial cells inevitably incorporated small amounts of Zn during their growth. If some of this internal Zn were released during the adsorption experiments, then one might well expect the largest isotopic fractionations would be observed in those experiments where more bacteria were present. However, for this to be an effective mechanism to explain the observed isotopic differences, there would have to be enough intracellular Zn (relative to the Zn added to the solution) to impact the overall Zn isotopic composition. We found that P. mendocina and B. subtilis incorporated only a trace amount of intracellular Zn during their growth (1.5 and 0.4 μg/moist gram of bacteria). Mass balance calculations indicate that there is simply not enough internal Zn in the bacteria to measurably influence the isotopic composition of the experimental solutions. Even if the δ66Zn of the internal Zn was + 10‰, it could only have increased the bulk isotopic composition by a maximum of + 0.014‰ and + 0.005‰ for the low Zn:bacteria adsorption experiments with P. mendocina and B. subtilis, respectively. The most likely explanation is that small amounts of dissolved organic exudates produced by the bacteria strongly complexed a fraction of the total Zn, forming two separate dissolved Zn reservoirs in solution; (1) the free Zn+2 (Znfree) fraction and (2) a dissolved organically complexed Zn (Znorg) fraction. Assuming that the dissolved Zn-organic complexes (with a high δ66Zn) have a higher affinity for Zn than does the bacterial surface, the measured δ66Zn of the solution that we measure would then represent the combination of isotopic signatures of both Znfree and Znorg. Hence, we would expect the δ66Zn of the experimental solutions to increase as a function of increasing pH, because the remaining Znfree would become adsorbed to the bacterial surfaces and the isotopically heavier dissolved organic Zn would be the only dissolved form of Zn left behind at higher pH. This scenario generally fits our observations, so we tested it further using an empirical modeling approach. The goal of this modeling was to determine if it was feasible for the δ66Zn of the experimental solutions to have developed in a way that approximates the observations if dissolved Zn complexes were present. We estimated the amount of the dissolved Zn-organic complex that could be present based on the difference between the maximum amount of Zn adsorbed and the total amount of Zn added to the low Zn:bacteria experimental systems. In these systems there were enough surface functional group sites present to adsorb 100% of the Zn, so we speculate that variations from 100% adsorption could have been attributable to the presence of dissolved Zn-organic complexes. Hence, dissolved Zn-organic complexes comprised as much as 10‰ and 3% of the total Zn in the low Zn:bacteria experiments with P. mendocina and B. subtilis, respectively. These values were used as a starting point for the empirical model. The model was constructed by calculating a Zn isotopic mass balance at all of the experimental pH points, using the following information and assumptions: (1) the known fraction of Zn adsorbed at every point, (2) the known separation factor for Zn adsorption onto bacterial surfaces (Δ66Znbacteria-solution = + 0.46‰) that was determined in the high Zn:bacteria ratio experiments, (3) the measured δ66Zn of the bulk solution (free Zn+2 + dissolved organic Zn complexes) and (4) the assumed percentage of strong dissolved Zn-organic complexes in the system (10% and 3% for P. mendocina and B. subtilis, respectively). We let the separation factor for the dissolved Zn-organic complex (i.e., Δ66ZnZn_org-free) vary in order to best-fit the observed data. We additionally assumed that the concentration of Zn-organic complexes was constant throughout an individual experiment and that the δ66Zn of the dissolved Zn-organic complexes was constant. We used the same approach to model both the low and high Zn:bacteria ratio experiments with live cells of P. mendocina and B. subtilis. However, for the high Zn:bacteria ratio experiments the amounts of dissolved Zn-organic complexes were scaled by the percentage differences in the amounts of bacteria in the experimental solutions (note that concentrations of bacteria were much lower for the high Zn-bacteria ratio experiments). The fractions of Znads, Znfree, and Znorg for each point in the model are presented in Table 2.

In Fig. 2 we have plotted the model results that provided the best fits to the data. The best fits involved Δ66ZnZn_org-free separation factors of + 2.0‰ and + 3.0‰ for P. mendocina and B. subtilis, respectively. Although the models are not perfect, they provide reasonable firstapproximation fits of all the systems, suggesting that the presence of isotopically heavy dissolved Zn-organic complexes is a viable explanation for the differences in the low and high Zn:bacteria ratio experiments. Better fits to the data could be achieved by varying more parameters and/or disregarding the assumption of a constant abundance of Zn-organic complexes. In the low Zn:bacteria ratio experiments (Fig. 2b) the larger amounts of bacteria and lower amounts of total Zn amplified the isotopic contribution of the dissolved Zn-organic complexes. In the high Zn:bacteria ratio experiments (Fig. 2a) there was simply too much Znfree, relative to the amount of dissolved Zn-organic complexes (i.e., very little dissolved organic exudate was produced relative to the amount of total Zn) to appreciably influence the overall isotopic composition of the solution phase (Table 2). Assuming the production of dissolved organic molecules is the most likely explanation for our results; it also appears that the adsorption experiment with dead cells of P. mendocina must have also been strongly influenced by dissolved organic ligands (Fig. 2b). Previous work by Ngwenya (2007) has shown that under neutral pH conditions, bacterial cell exudates formed from the lysis of dead cells lead to the formation of dissolved Zn-organic complexes. A similar process may have impacted the production of dissolved organic molecules in our experiments. 3.2. Intracellular incorporation experiments Zinc is an essential trace nutrient for bacteria and other organisms contributing to the formation of enzymes and serving as a cofactor for regulatory proteins (Outten and O'Halloran, 2001). Normal operation of the cell occurs over a narrow concentration range of cytoplasmic Zn2 + (Outten and O'Halloran, 2001). In order for the cell to sustain these healthy levels of Zn, it utilizes a series of metal-transport proteins to control both the influx and efflux of Zn within the cell (Finney and O'Halloran, 2003; Nies, 2007). Our intracellular incorporation experiments evaluated Zn isotope fractionations related to the following: (1) “Metal stress” — whether changes in Zn homeostasis caused by metal stress in a single bacterial species (P. mendocina) resulted in changes in δ66Zn, (2) “Different species” — whether differences in the Zn homeostasis machinery among different bacterial species (P. mendocina, E. coli, and natural consortium) resulted in differences in δ66Zn when grown with the same concentrations of Zn; and (3) “Zn speciation” — whether changes in the speciation and possible bioavailability of Zn lead to differences in δ66Zn in a single bacterial species (P. mendocina) when compared at the same total concentration of Zn. For the latter experiments, Zn-citrate and free Zn+2 were used. The Zn in the form of Zn-citrate was not predicted to appreciably change speciation or precipitate under the experimental conditions. However, geochemical modeling using Visual Minteq™ indicated that the free Zn+2 added to the growth medium was re-distributed primarily among disspecies and that the solutions solved ZnHPO4, Nfree Zn+2, NZnNH+2 3 were oversaturated with respect to Zn3(PO4)2∙4H2O and hydrozincite. The pHs of the experimental systems were ~ 8.0 at day 3, but were only monitored sporadically. 3.2.1. Isolation of intracellular Zn For the intracellular incorporation experiments, we attempted to remove surface-bound Zn from the bacteria by washing them 5 times in a pH 1.5, 0.02 M MgSO4 electrolyte solution prior to digesting them (Navarrete et al., 2011). We evaluated the effectiveness of the washing procedures by analyzing the concentrations of Zn in the wash fluids after each of the five washing cycles. To ensure any potential influence from adsorbed Zn was limited, we rejected any samples that had more than 5% “surface contamination.” Mass balance constraints suggest that under conditions of less than 5% adsorbed Zn the influence of the

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

a

Table 3 Incorporation experiments.

UV absorption (600 nm)

1.6

Exp.

Day

Zn incorporated (μg/g)

δ66Zn averagea,b

2σ

n

1.2

I1

0.8

I2

0.4

I3

3 5 7 3 5 7 3 8 11 5 7 11 4 6 10 3 5 7 3 5 7

4.45 9.20 3.84 9.90 9.58 16.40 115.0 176.9 509.1 17.77 14.95 29.47 14.55 12.82 16.55 0.17 0.44 0.74 0.78 0.70 0.65

−0.048 0.510 0.028 −0.092 −0.247 −0.235 1.027 1.504 1.051 0.059 0.027 −0.041 0.078 0.226 0.032 1.402 2.042 1.696 1.754 0.772 0.461

0.101 0.080 0.184 0.080 0.037 0.106 0.076 0.104 0.098 0.102 0.096 0.051 0.085 0.066 0.027 0.031 0.008 0.101 0.081 0.211 0.006

2 2 2 2 2 3 2 2 3 2 5 2 2 3 2 2 2 2 2 3 2

0 0

2

4

6

8

10

12

I4

Time (days) I5

b Zn incorporated log (mg/g)

10-1

I6

I7

10-2

a

10-3

10-4

b

3

7

11

Time (days)

c 2.5

δ66Zn of bacteria

49

2 1.5 1 0.5 0

-0.5

3

7

11

Time (days) P. mendocina - 0.2 mg/L Zn (Zn-citrate) P. mendocina - 2 mg/L Zn (Zn-citrate) P. mendocina - 20 mg/L Zn (Zn-citrate) Natural consortium - 2 mg/L Zn (Zn-citrate) E. coli - 2 mg/L Zn (Zn-citrate) P. mendocina - 0.2 mg/L Zn2+ P. mendocina - 2 mg/L Zn2+

Fig. 3. Cell density in solution (a), Zn concentration (b), and δ66Zn results (c) for digested bacteria cells plotted as a function of time for intracellular incorporation experiments. Absorbance at 600 nm was used as a proxy for cell density. Zn concentrations are expressed (on a log scale) in mg of Zn per moist gram of digested bacteria pellet. 2σ error bars are plotted for the isotope data.

adsorbed phase on the δ66Zn of the bacteria would be less than 0.1‰. For most samples the concentrations of Zn decreased from the 1st to the 5th wash cycle and only small amounts of Zn — much less than 5% — remained (Table S-1). However, we were forced to reject all the samples from a 20 mg/L Zn-citrate experiment with P. mendocina. 3.2.2. Bacterial growth Fig. 3a illustrates the growth behavior for the bacteria grown under the tested conditions (Tables 1 and 3). The results demonstrate that when compared at different concentrations of Zn (in the form of Zncitrate), the growth of P. mendocina was substantially impacted at the

Isotopic ratio of the digested cells. δ66Zn of the bracketing standard was zero per mil.

highest Zn concentrations. A concentration of 20 mg/L of Zn (calculated as the concentration of Zn, not Zn-citrate) resulted in a longer lag time prior to exponential growth and this phase itself was not as rapid as those in the lower concentration experiments (Fig. 3a). It is thought that this lag is a toxicity effect related to the time it takes for the bacteria to adjust to these conditions (Fisher and Jones, 1981; Cabrero et al., 1998). When compared at the same concentrations of Zn-citrate (2 mg/L), E. coli and the natural consortium exhibited similarly shaped growth curves, while P. mendocina experienced more rapid growth and was in the death phase before the other species reached their maxima (Fig. 3a). Finally, a comparison of P. mendocina grown in Zn-citrate vs. Zn2 + (at 0.2 and 2 mg/L total Zn) shows very similar timing of growth phases (Fig. 3a). Duplicate growth experiments (data not shown) have shown that the exact magnitudes of absorbance at 600 nm change slightly during repeat experiments (probably due to small changes in the amount of inoculum used), but that the shapes and timing of the curves are consistent. 3.2.3. Intracellular Zn concentrations Fig. 3b presents the concentrations of Zn (in mg of Zn per moist gram of bacteria) in the digested bacteria sampled at 3 different stages during the intracellular incorporation experiments. The data for the 20 mg/L experiment are not presented because adsorbed Zn may have influenced these results (see Section 3.2.1). The extent of Zn incorporation in P. mendocina, E. coli, and the natural consortium exposed to the same external concentrations of Zn (2 mg/L) was roughly similar, varying over a narrow range from 0.01 to 0.03 mg/g, and did not change much as a function of growth phase (Fig. 3a and b). In comparison, the results from the experiment with P. mendocina at 0.2 mg/L of Zn (as Zn-citrate) show measurably less incorporation of Zn, suggesting that Zn incorporation is influenced by the concentration of Zn in the growth medium. When compared at the same concentrations of total Zn (2 and 0.2 mg/L), less Zn was incorporated into P. mendocina cells in the experiments with free Zn+2 as compared with those using Zn-citrate (Fig. 3a). This suggests that the presence of Zn-citrate leads to more bioaccumulation of Zn than does free Zn+ 2 in the tested bacteria. 3.2.4. Zinc isotopic fractionation during intracellular incorporation The δ66Zn values measured for digested cells sampled at different periods of growth are presented in Fig. 3c. The data for the 20 mg/L

50

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

experiment are not presented because adsorbed Zn may have influenced these results (see Section 3.2.1). The results from the comparison of different levels of metal stress (0.2 mg/L and 2 mg/L of Zn as Zn-citrate) with P. mendocina indicate that there was little difference in Zn isotope fractionation despite a measurable difference in the amount of Zn incorporated (Fig. 3b). P. mendocina cells subjected to 2 mg/L of Zn (as Zn-citrate) show slightly negative δ66Zn with a Δ66Znincorporated-solution = − 0.09‰, − 0.25‰, and − 0.24‰ for samples collected on days 3, 5, and 7, respectively (Fig. 3c). The 0.2 mg/L Zn (as Zn-citrate) experiment showed less isotope fractionation with a Δ66Znincorporated-solution = −0.04‰ and +0.03‰ for samples collected on days 3 and 7, respectively (Fig. 3c). The sample collected on day 5 had a much higher δ66Zn of +0.51‰. This appears to be anomalous; however, we also found that substantially more Zn was incorporated in this sample (Fig. 3b). The δ66Zn values of digested cells of different bacterial species (E. coli, P. mendocina, and the natural consortium) subjected to the same concentrations of Zn (2 mg/L as Zn-citrate) were small, but measurable. For example, the δ66Zn in both the E. coli and the natural consortium experiments tended to be slightly positive (up to +0.23‰ for the day 3 E. coli sample), while the δ66Zn for all the P. mendocina samples was slightly negative. These small changes in Zn isotopic fractionation may be a reflection of differences in the chemistries of Zn-transporter proteins, their amounts, and/or the extents of their activity. For example, Leedjärv et al. (2008) indicates that certain genes of each family, although common among most bacteria, can still show different functionality in different bacterial species. The experiments with P. mendocina comparing the effect of Zn speciation (Zn-citrate vs. free Zn+2 at 0.2 mg/L and 2 mg/L of total Zn) are also presented in Fig. 3c. At both concentrations of Zn, the addition of free Zn+2 (as compared to Zn-citrate) resulted in substantial preferential uptake of the heavier Zn isotopes in the solid phase. For example, in the 0.2 mg/L free Zn+2 experiment the δ66Zn was +1.40‰, +2.04‰, and +1.70‰ for samples collected at days 3, 5, and 7. The explanation for the large difference in Zn isotope fractionation between free Zn+2 versus Zn-citrate in P. mendocina is not clear. At least part of the observed effects could be attributable to changes in aqueous speciation and the precipitation of Zn. A precipitate that preferentially incorporates the heavier Zn isotopes would be consistent with our data. Fujii and Albarède (2012) have demonstrated that aqueous Zn-phosphate complexes, the precursors of Zn-phosphate precipitates, are isotopically heavier than coexisting free Zn+2 by up to 1.0‰. On the other hand, the effects on Zn isotope fractionation could be attributable to differences in the cellular processing of the different aqueous Zn-species. Because free metals are more toxic to bacteria than most metal-organic complexes like Zn-citrate, it is possible that a metal stress response was triggered in the experiments with free Zn+ 2. The principal Zn transporters controlling cell homeostasis in most bacteria include influx (ZiP, Zur) and efflux (ATPase, CDF) proteins (Rossbach et al., 2000; Hantke, 2001; Eide, 2006; Nies, 2007; Leedjärv et al., 2008; Waldron and Robinson, 2009). For Pseudomonas species, the P-type ATPase (ZntA) Zn transporter is common as part of cellular efflux mechanisms (Nies, 2007; Leedjärv et al., 2008), as is CzcCBA2, which is a chemiosmotic transporter shown to contribute to Zn resistance in P. putida (Leedjärv et al., 2008). In addition, siderophore formation in Pseudomonas species may be stimulated by elevated Zn concentrations. This has been shown previously for Pseudomonas fluorescens (Rossbach et al., 2000). This sort of additional gene expression (and the production of more or different Zn-binding proteins) acting in response to increased levels of free (toxic) Zn may be what is reflected in our isotopic results. Our experiment with 2 mg/L free Zn+2 shows a distinctive decrease in δ66Zn from +1.75 to +0.77 and +0.46 as the cells continue to grow (and die) from day 3 to days 5 and 7, respectively. The gradual decrease in δ66Zn could indicate that the cells are being adversely impacted by toxic concentrations of Zn, which may have caused a relative decrease

in the activity of various Zn-binding proteins over time. Changes in the δ66Zn attributable to Zn precipitation are less likely, since the extent of precipitation should be relatively stable over the duration of the experiment.

4. Conclusions In this study, we performed laboratory experiments using a variety of bacterial species and experimental conditions to understand Zn isotopic fractionation behavior for bacterial-surface adsorption and intracellular incorporation. Our key findings include the following: 1. Bacterial surface adsorption produces an equilibrium isotopic fractionation that preferentially incorporates the heavier Zn isotopes into the surface complexes. We were able to estimate an average equilibrium separation factor (inclusive of the tested Gram-positive and Gram-negative bacterial species) of Δ66Znadsorbed-solution = +0.46‰ (αadsorbed-solution ≈ 1.00046). 2. Bacterial surface adsorption experiments conducted under conditions of low Zn:bacteria ratio appear to have been complicated by the presence of dissolved organic exudates that competed with surface functional group sites for free Zn+2. We were able to empirically model this process to show that very small amounts of Zn-organic complexes with fractionation factors in the range of α = 1.002 to 1.003 could account for the observed δ66Zn in the experimental solutions. The unforeseen influence of dissolved organic molecules serves as an important lesson for these and other experiments that attempt to experimentally determine isotopic fractionations in aqueous systems with bacteria (or other biological systems). More importantly, this observation hints at the possibility that small amounts of dissolved organic molecules that strongly bind metals may play a big role in determining the isotopic variations of Zn in natural water systems. 3. When bacteria (P. mendocina, E. coli, and our natural consortium) were grown in solutions with Zn-citrate where Zn concentrations (0.2 and 2 mg/L) might be typical of some natural waters or (more likely) sediment systems, only modest amounts of Zn isotopic fractionation were observed. The small differences in δ66Zn among the bacterial species likely reflect subtle changes in their Zn homeostasis machinery and its functionality. If this holds true, it may suggest that careful measurements of δ66Zn can be paired with other molecular and spectroscopic techniques to better understand the biological cycling and processing of Zn at the cellular scale. These species-dependent variations also suggest the need for caution when interpreting the cause for Zn isotopic variations in natural systems and in the geologic record. 4. The replacement of Zn-citrate for an equal amount of free Zn+2 in the growth medium had the impact of greatly increasing the δ66Zn of Zn incorporated within the solid phase. Although the experiments with free Zn+2 were likely impacted by precipitation reactions, isotopic changes also appeared to be attributable to differences in the intracellular processing of free Zn+ 2. Previous work by John et al. (2007) showed that different levels of free Zn+ 2 trigger different homeostasis pathways in diatoms. Intracellular free Zn+2 may have been incorporated into strong metal binding proteins released by P. mendocina cells in order to detoxify their internal environments. Although more work would be needed to isolate key reactions and to test this hypothesis, our observations suggest that Zn isotopic fractionations in bacteria could be used to elucidate metal-binding pathways related to metal toxicity. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.chemgeo.2013.12.007.

F.-C.A. Kafantaris, D.M. Borrok / Chemical Geology 366 (2014) 42–51

Acknowledgments This publication was made possible through funding by the National Science Foundation (NSF) grant 0745345, the Center for Earth and Environmental Isotope Research (NSF MRI grant 0820986) and a GSA student research grant for Kafantaris. Comments from Corey Archer, Sune Nielsen, and 2 anonymous reviewers were influential in helping to shape this manuscript. References Andersen, M.B., Vance, D., Archer, C., Anderson, R.F., Ellwood, M.J., Allen, C.S., 2011. The Zn abundance and isotopic composition of diatom frustules, a proxy for Zn availability in ocean surface seawater. Earth Planet. Sci. Lett. 301, 137–145. Andreini, C., Banci, L., Bertini, I., Rosato, A., 2006a. Counting the zinc-proteins encoded in the human genome. J. Proteome Res. 5, 196–201. Andreini, C., Banci, L., Bertini, I., Rosato, A., 2006b. Zinc through the three domains of life. J. Proteome Res. 5, 3173–3178. Aranda, S., Borrok, D.M., Wanty, R.B., Balistrieri, L.S., 2012. Zinc isotope investigation of surface and pore waters in a mountain watershed impacted by acid rock drainage. Sci. Total Environ. 420, 202–213. Arnold, T.I.M., Kirk, G.J.D., Wissuwa, M., Frei, M., Zhao, F.-J., Mason, T.F.D., Weiss, D.J., 2010. Evidence for the mechanisms of zinc uptake by rice using isotope fractionation. Plant Cell Environ. 33, 370–381. Balistrieri, L.S., Borrok, D.M., Wanty, R.B., Ridley, W.I., 2008. Fractionation of Cu and Zn isotopes during adsorption onto amorphous Fe (III) oxyhydroxide: experimental mixing of acid rock drainage and ambient river water. Geochim. Cosmochim. Acta 72, 311–328. Beard, S.J., Hashim, R., Wu, G., Binet, M.R.B., Hughes, M.N., Poole, R.K., 2000. Evidence for the transport of zinc(II) ions via the Pit inorganic phosphate transport system in Escherichia coli. FEMS Microbiol. Lett. 184, 231–235. Beveridge, T.J., 1989. Role of cellular design in bacterial metal accumulation and mineralization. Rev. Microbiol. 43, 147–171. Beveridge, T.J., Koval, S.F., 1981. Binding of metals to cell envelopes of Escherichia coli K-12. Appl. Environ. Microbiol. 42, 325–335. Borg, S.J., Liu, W., 2010. An XAS study of zinc speciation in aqueous acetate solutions at 25–200 °C. Nucl. Instrum. Meth. Phys. Res. A 619, 276–279. Borrok, D., Fein, J.B., Kulpa, C.F., 2004. Proton and Cd adsorption onto natural bacterial consortia: testing universal adsorption behavior. Geochim. Cosmochim. Acta 68, 3231–3238. Borrok, D., Borrok, M.J., Fein, J.B., Kiessling, L.L., 2005a. Link between chemotactic response to Ni2+ and its adsorption onto the Escherichia coli cell surface. Environ. Sci. Technol. 39, 5227–5233. Borrok, D., Turner, B.F., Fein, J.B., 2005b. A universal surface complexation framework for modeling proton binding onto bacterial surfaces in geologic settings. Am. J. Sci. 305, 826–853. Borrok, D.M., Wanty, R., Ridley, W., Wolf, R., Lamothe, P., Adams, M., 2007. Separation of copper, iron, and zinc from complex aqueous solutions for isotopic measurement. Chem. Geol. 242, 400–414. Borrok, D.M., Ridley, W.I., Wanty, R.B., 2008. Isotopic fractionation of Cu and Zn during adsorption onto bacterial surfaces. Geochim. Cosmochim. Acta 72, A99. Cabrero, A., Fernandez, S., Mirada, F., Garcia, J., 1998. Effects of copper and zinc on the activated sludge bacteria growth kinetics. Water Res. 32, 1355–1362. Caldelas, C., Dong, S., Araus, J.L., Weiss, D.J., 2011. Zinc isotopic fractionation in Phragmites australis in response to toxic levels of zinc. J. Exp. Bot. 62, 2169–2178. Cánovas, D., Cases, I., De Lorenzo, V., 2003. Heavy metal tolerance and metal homeostasis in Pseudomonas putida as revealed by complete genome analysis. Environ. Microbiol. 5, 1242–1256. Eide, D.J., 2006. Zinc transporters and the cellular trafficking of zinc. Biochim. Biophys. Acta (BBA)—Molecular Cell Res. 1763, 711–722. Finney, L.A., O'Halloran, T.V., 2003. Transition metal speciation in the cell: insights from the chemistry of metal ion receptors. Sci. Signal. 300, 931. Fisher, N.S., Jones, G.J., 1981. Heavy metals and marine phytoplankton: correlation of toxicity and sulfhydryl-binding. J. Phycol. 17, 108–111. Fowle, D.A., Fein, J.B., 2000. Experimental measurements of the reversibility of metal– bacteria adsorption reactions. Chem. Geol. 168, 27–36. Fujii, T., Albarède, F., 2012. Ab initio calculation of the Zn isotope effect in phosphates, citrates, and malates and applications to plants and soil. PLoS ONE 7, e30726. Gélabert, A., Pokrovsky, O., Viers, J., Schott, J., Boudou, A., Feurtet-Mazel, A., 2006. Interaction between zinc and freshwater and marine diatom species: surface complexation and Zn isotope fractionation. Geochim. Cosmochim. Acta 70, 839–857. Ginn, B.R., Fein, J.B., 2008. The effect of species diversity on metal adsorption onto bacteria. Geochim. Cosmochim. Acta 72, 3939–3948. Guiné, V., Spadini, L., Sarret, G., Muris, M., Delolme, C., Gaudet, J.P., Martins, J., 2006. Zinc sorption to three Gram-negative bacteria: combined titration, modeling, and EXAFS study. Environ. Sci. Technol. 40, 1806–1813. Hambidge, M., 2000. Zinc and health: current status and future directions. J. Nutr. 130, 1344S–1349S. Hantke, K., 2001. Bacterial zinc transporters and regulators. Biometals 14, 239–249. Hersman, L., Forsythe, J., Ticknor, L., Maurice, P., 2001. Growth of Pseudomonas mendocina on Fe(III) (hydr)oxides. Appl. Environ. Microbiol. 67, 4448–4453. John, S.G., Geis, R.W., Saito, M.A., Boyle, E.A., 2007. Zinc isotope fractionation during highaffinity and low-affinity zinc transport by the marine diatom Thalassiosira oceanica. Limnol. Oceanogr. 52 (6), 2710–2714.

51

Jouvin, D., Louvat, P., Juillot, F., Maréchal, C.N., Benedetti, M.F., 2009. Zinc isotopic fractionation: why organic matters. Environ. Sci. Technol. 43, 5747–5754. Jouvin, D., Weiss, D.J., Mason, T.F.M., Bravin, M.N., Louvat, P., Zhao, F., Ferec, F., Hinsinger, P., Benedetti, M.F., 2012. Stable isotopes of Cu and Zn in higher plants: evidence for Cu reduction at the root surface and two conceptual models for isotopic fractionation processes. Environ. Sci. Technol. 46, 2652–2660. Kelly, S., Kemner, K., Fein, J., Fowle, D., Boyanov, M., Bunker, B., Yee, N., 2002. X-ray absorption fine structure determination of pH-dependent U-bacterial cell wall interactions. Geochim. Cosmochim. Acta 66, 3855–3871. Kenney, J.P.L., Song, Z., Bunker, B.A., Fein, J.B., 2012. An experimental study of Au removal from solution by non-metabolizing bacterial cells and their exudates. Geochim. Cosmochim. Acta 87, 51–60. Konhauser, K.O., 2006. Introduction to Geomicrobiology.Wiley-Blackwell Press, Oxford. Leedjärv, A., Ivask, A., Virta, M., 2008. Interplay of different transporters in the mediation of divalent heavy metal resistance in Pseudomonas putida KT2440. J. Bacteriol. 190, 2680–2689. Maret, W., 2009. Molecular aspects of human cellular zinc homeostasis: redox control of zinc potentials and zinc signals. Biometals 22, 149–157. McClure, C.P., Rusche, K.M., Peariso, K., Jackman, J.E., Fierke, C.A., Penner-Hahn, J.E., 2003. EXAFS studies of the zinc sites of UDP-(3-O-acyl)-N-acetylglucosamine deacetylase (LpxC). J. Inorg. Biochem. 94, 78–85. Mishra, B., Boyanov, M., Bunker, B.A., Kelly, S.D., Kemner, K.M., Fein, J.B., 2010. High- and low-affinity binding sites for Cd on the bacterial cell walls of Bacillus subtilis and Shewanella oneidensis. Geochim. Cosmochim. Acta 74, 4219–4233. Moore, C.M., Helmann, J.D., 2005. Metal ion homeostasis in Bacillus subtilis. Curr. Opin. Microbiol. 8, 188–195. Moynier, F., Pichat, S., Pons, M.L., Fike, D., Balter, V., Albarède, F., 2009. Isotopic fractionation and transport mechanisms of Zn in plants. Chem. Geol. 267, 125–130. Navarrete, J.U., Borrok, D.M., Viveros, M., Ellzey, J.T., 2011. Copper isotope fractionation during surface adsorption and intracellular incorporation by bacteria. Geochim. Cosmochim. Acta 75, 784–799. Ngwenya, B.T., 2007. Enhanced adsorption of zinc is associated with aging and lysis of bacterial cells in batch incubations. Chemosphere 67 (10), 1982–1992. Ngwenya, B.T., Sutherland, I.W., Kennedy, L., 2003. Comparison of the acid–base behaviour and metal adsorption characteristics of a Gram-negative bacterium with other strains. Appl. Geochem. 18, 527–538. Nies, D.H., 2007. How cells control zinc homeostasis. Science 317, 1695. Olhaberry, J., Leary, W., Reyes, A., Lockett, C., 1983. Biochemistry of zinc. S. Afr. Med. J. 64, 894–895. Outten, C.E., O'Halloran, T.V., 2001. Femtomolar sensitivity of metalloregulatory proteins controlling zinc homeostasis. Science 292, 2488–2492. Peel, K., Weiss, D., Sigg, L., 2009. Zinc isotope composition of settling particles as a proxy for biogeochemical processes in lakes: Insights from the eutrophic Lake Greifen, Switzerland. Limnol. Oceanogr. 54 (5), 1699–1708. Pokrovsky, O., Viers, J., Freydier, R., 2005. Zinc stable isotope fractionation during its adsorption on oxides and hydroxides. J. Colloid Interface Sci. 291, 192–200. Roelofs, R., Rengel, Z., Cawthray, G., Dixon, K., Lambers, H., 2001. Exudation of carboxylates in Australian Proteaceae: chemical composition. Plant Cell Environ. 24, 891–904. Rossbach, S., Wilson, T.L., Kukuk, M.L., Carty, H.A., 2000. Elevated zinc induces siderophore biosynthesis genes and a zntA‐like gene in Pseudomonas fluorescens. FEMS Microbiol. Lett. 191, 61–70. Schauble, E.A., 2004. Applying stable isotope fractionation theory to new systems. Rev. Mineral. Geochem. 55, 65–111. Shankar, A.H., Prasad, A.S., 1998. Zinc and immune function: the biological basis of altered resistance to infection. Am. J. Clin. Nutr. 68, 447S–463S. Tang, Y.T., Cloquet, C., Sterckeman, T., Echevarria, G., Carignan, J., Qiu, R.L., Morel, J.L., 2012. Fractionation of stable Zn isotopes in the field-grown Zn hyperaccumulator Noccaea caerulescens and the Zn-tolerant plant Silene vulgaris. Environ. Sci. Technol. 46, 9972–9979. Thapalia, A., Borrok, D.M., Van Metre, P.C., Musgrove, M.L., Landa, E.R., 2010. Zn and Cu isotopes as tracers of anthropogenic contamination in a sediment core from an urban lake. Environ. Sci. Technol. 44, 1544–1550. Urey, H.C., 1947. The thermodynamic properties of isotopic substances. J. Chem. Soc. (Resumed) 562–581. Viers, J., Oliva, P., Nonell, A., Gélabert, A., Sonke, J.E., Freydier, R., Gainville, R., Dupré, B., 2007. Evidence of Zn isotopic fractionation in a soil–plant system of a pristine tropical watershed (Nsimi, Cameroon). Chem. Geol. 239, 124–137. Waldron, K.J., Robinson, N.J., 2009. How do bacterial cells ensure that metalloproteins get the correct metal? Nat. Rev. Microbiol. 7, 25–35. Wanty, R.B., Podda, F., De Giudici, G., Cidu, R., Lattanzi, P., 2013. Zinc isotope and transition-element dynamics accompanying hydrozincite biomineralization in the Rio Naracauli, Sardinia, Italy. Chem. Geol. 337–338, 1–10. Wei, X., Fang, L., Cai, P., Huang, Q., Chen, H., Liang, W., Rong, X., 2011. Influence of extracellular polymeric substances (EPS) on Cd adsorption by bacteria. Environ. Pollut. 159, 1369–1374. Weiss, D., Mason, T.F.D., Zhao, F., Kirk, G., Coles, B., Horstwood, M., 2004. Isotopic discrimination of zinc in higher plants. New Phytol. 165, 703–710. Yee, N., Fein, J.B., 2001. Cd adsorption onto bacterial surfaces: a universal adsorption edge? Geochim. Cosmochim. Acta 65, 2037–2042. Yee, N., Fein, J.B., 2003. Quantifying metal adsorption onto bacteria mixtures: a test and application of the surface complexation model. Geomicrobiol J. 20, 43–60. Zhu, X., Guo, Y., Williams, R., O'Nions, R., Matthews, A., Belshaw, N., Canters, G., De Waal, E., Weser, U., Burgess, B., 2002. Mass fractionation processes of transition metal isotopes. Earth Planet. Sci. Lett. 200, 47–62.