Environ. Sci. Technol. 2003, 37, 300-307
Sorption versus Biomineralization of Pb(II) within Burkholderia cepacia Biofilms A L E X I S S . T E M P L E T O N , * ,† T H O M A S P . T R A I N O R , †,‡ ALFRED M. SPORMANN,§ MATHEW NEWVILLE,‡ STEVEN R. SUTTON,‡ ALICE DOHNALKOVA,| YURI GORBY,| AND G O R D O N E . B R O W N , J R . †,⊥ Surface & Aqueous Geochemistry Group, Department of Geological & Environmental Sciences, Stanford University, Stanford, California 94035-2115, Consortium for Advanced Radiation Sources, The University of Chicago, Building 434-A, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, Illinois 60439, Department of Civil and Environmental Engineering, Stanford University, Stanford, California 94305, Environmental Molecular Science Laboratory, Pacific Northwest National Laboratory, 908 Battelle Boulevard, Richland, Washington 99352, and Stanford Synchrotron Radiation Laboratory, SLAC, Stanford University, Stanford, California 94309-0210
X-ray spectroscopy measurements have been combined with macroscopic uptake data and transmission electron microscopy (TEM) results to show that Pb(II) uptake by Burkholderia cepacia is due to simultaneous sorption and biomineralization processes. X-ray microprobe mapping of B. cepacia biofilms formed on R-Al2O3 surfaces shows that Pb(II) is distributed heterogeneously throughout the biofilms because of the formation of Pb “hot spots”. EXAFS data and TEM observations show that the enhanced Pb accumulation is due to the formation of nanoscale crystals of pyromorphite (Pb5(PO4)3(OH)) adjacent to the outermembrane of a fraction of the total population of B. cepacia cells. In contrast, B. cepacia cell suspensions or biofilms that were heat-killed or pretreated with X-rays do not form pyromorphite, which suggests that metabolic activity is required. Precipitation of pyromorphite occurs over several orders of magnitude in [H+] and [Pb] and accounts for approximately 90% of the total Pb uptake below pH 4.5 but only 45-60% at near-neutral pH because of the formation of additional Pb(II) adsorption complexes. Structural fits of Pb LIII EXAFS data collected for heat-treated cells at near-neutral pH suggest that Pb(II) forms inner-sphere adsorption complexes with carboxyl functional groups in the biofilms. * Corresponding author present address: Scripps Institution of Oceanography, Marine Biology Research Division, 9500 Gilman Dr., La Jolla, CA 92093-0202; e-mail:
[email protected]; phone: (858)822-1426; fax: (858)534-7373. † Department of Geological & Environmental Sciences, Stanford University. ‡ Argonne National Laboratory. § Department of Civil and Environmental Engineering, Stanford University. | Pacific Northwest National Laboratory. ⊥ SLAC, Stanford University. 300
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003
Introduction Pb is one of the most common toxic heavy metals found in contaminated soils and mine-tailing environments. The mobility and bioavailability of Pb is significantly mitigated by sorption and precipitation reactions, particularly at nearneutral to alkaline pH. Sorption of Pb to reactive, amphoteric functional groups on mineral surfaces such as aluminum and iron (hydr)oxide phases and clays has been wellcharacterized via spectroscopic techniques to determine the mode of Pb binding (e.g., refs 1-4). The results of these studies indicate that the types of Pb(II) surface complexes formed are strongly dependent upon the composition and structure of the reactive surface. In many soil and aquatic systems, sorption and precipitation reactions at biological surfaces, particularly in biofilm environments where bacteria attach to mineral surfaces and produce a matrix of highly hydrated exopolysaccharides, may also be important for scavenging toxic metals such as Pb (e.g., ref 5). However, the speciation of Pb complexed to bacterial surfaces has not been directly probed using spectroscopic methods. Strong interest in sorption to biological surfaces (i.e., biosorption) as a method of Pb ion immobilization has resulted in numerous macroscopic uptake studies of Pb by plants, macro- and micro-algae, fungi, and Gram-positive and Gram-negative bacteria (e.g., refs 6-10). Metal binding to bacterial surfaces occurs over a large pH range because of the low isoelectric point of most bacterial surfaces. Surface functional groups such as carboxyl, sulfuryl, and phosphoryl moieties show a strong affinity for Pb. Surface-complexation modeling of Pb sorption on Gram-positive bacterial surfaces, in particular Bacillus species, has invoked Pb surface complexes with carboxyl functional groups at low to neutral pH and phosphoryl groups at higher pH (11). Similar behavior is expected on Gram-negative bacterial surfaces, given the similar site densities of carboxyl and phosphoryl groups determined from potentiometric titration studies (12). Direct determination of the mode of Pb binding to bacterial surfaces at the molecular level would significantly complement the surface complexation modeling of Pbbacterial interactions. However, only a few X-ray spectroscopic studies have addressed the mode of Pb uptake on biological surfaces, such as fungal cell walls, to differentiate between the role of carboxyl and phosphoryl functional groups in Pb uptake (13). In addition, there is little quantitative information on the relative importance of adsorption versus precipitation reactions for metal sequestration when microorganisms are challenged by elevated concentrations of toxic metals such as Pb. X-ray spectroscopic methods, coupled with electron microscopy, are particularly well-suited for differentiating among the large variety of metal-ion species that may be associated with bacterial surfaces. We have chosen to examine the mode of Pb uptake, in particular the relationship between sorption and precipitation of Pb, in washed cell suspensions and monolayer biofilms of a common Gram-negative soil bacterium, Burkholderia cepacia. B. cepacia will form biofilms on numerous surfaces (e.g., glass slides and metal (hydr)oxide surfaces, such as quartz, corundum, hematite, and goethite), and these biofilms can form a significant sink for Pb at micromolar to millimolar Pb concentrations (14, 15). In the present study, the distribution of Pb within B. cepacia biofilms has been probed on the micron scale using X-ray microscopy at the Pb LIII edge to determine whether Pb accumulation is spatially heterogeneous or homogeneous across the biofilms. Extended X-ray absorption fine-structure (EXAFS) spectroscopy 10.1021/es025972g CCC: $25.00
2003 American Chemical Society Published on Web 12/10/2002
FIGURE 1. Pb LIII X-ray fluorescence maps, optical images collected at 50×, and grazing-incidence (GI-)EXAFS data for B. cepacia biofilms on r-Al2O3 incubated with 10-4.3 M Pb(NO3)2, pH 6. False color scale from blue (300 cps Pb FY) to white (3000 cps Pb FY). Sample A2 was directly removed from growth media and rinsed with 0.01M NaNO3 prior to incubation with Pb. Sample A1 was preexposed to X-rays after the rinse and then incubated with Pb (control). X-ray spot size was 5 µm × 5 µm. has been used to investigate the dominant mode of metal binding and to identify whether the Pb sequestered within the biofilms is in a sorbed or biomineralized form. These spectroscopic data are coupled with TEM observations to detect any Pb-bearing precipitates.
Materials and Methods Biofilm/Single-Crystal Experiments (Series A). Burkholderia cepacia biofilms were grown on R-Al2O3 (1-102) single-crystal substrates for 1 week in an annular reactor as described in Templeton et al. (14), using a minimal defined medium (200 µM CaCl2, 150 µM MgSO4, 90 µM (NH4)2SO4, 150 µM KNO3, 10 µM NaHCO3, 25 µM KH2PO4, and 1 mM sodium acetate as a carbon and energy source) at pH 6. After biofilm growth, the B. cepacia/R-Al2O3 crystal was withdrawn from the reactor, rinsed with a 0.01 M NaNO3 solution, and either immersed directly into a 10-4.2 M Pb(NO3)2 solution, pH 6 (experiment A2), in a CO2-free glovebox or exposed to an X-ray beam for 5 min to stop any metabolic activity prior to incubation with the 10-4.2 M Pb(NO3)2 solution (experiment A1). To examine the distribution of Pb within the B. cepacia biofilm, X-ray fluorescent mapping was performed on GeoSoilEnviroCARS beamline 13-IDC at the Advanced Photon Source, Argonne National Laboratory. A 250 µm × 250 µm collimated beam was tuned to 13.5 keV using a doublecrystal Si(111) monochromator. The X-rays were subsequently focused down to ∼5 µm × 5 µm using a KirkpatrickBaez mirror assembly (22). The B. cepacia/R-Al2O3 crystal was mounted on a motorized stage and aligned at 45° to the
incident beam and at 45° to a 13-element Ge solid-state detector. A 50× microscope objective was placed adjacent to the sample to image the bacterial distribution on the crystal surface. The Pb LR fluorescence signal was mapped across a 100 µm × 100 µm region on each crystal surface at the edge of a contiguous region of the biofilm. A nitrogen-filled ionization chamber was used to normalize the beam intensity for each pixel in the map. The sample stage was scanned in step sizes of 5 µm with a collection time of 10 s per step. The penetration depth of the analysis was through the whole biofilm (∼1 µm). For the final maps, the Pb LR fluorescence peak was integrated and background subtracted to derive net counts and plotted using a false-color scale to show variations in fluorescence yield (Figure 1). Surface concentrations of Pb within the biofilm were derived by comparing the Pb fluorescence counts to a 0.55-µm glass thin film standard (NBS SRM 1833) containing 15.19 µg of Pb/cm2 (9.9 wt %). Parallel grazing-incidence EXAFS experiments with the B. cepacia/R-Al2O3 (1-102) biofilms reacted with 10-4.2 M Pb(NO3)2 solution (pH 6) (i.e., duplicates of experiments A1 and A2) were conducted at beamline 6-2 at the Stanford Synchrotron Radiation Laboratory (SSRL), using the SSRL grazing-incidence apparatus. The incident X-rays were focused using a Pt-coated mirror upstream of a Si(111) monochromator and collimated to 5 mm horizontal by 0.05 mm vertical. Pb LIII EXAFS data were collected at an incidence angle of 150° using a 13-element Ge array detector aligned perpendicular to the incident beam. VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
301
Pb Uptake by Cell Suspensions (Series B and C). B. cepacia (ATCC 17616) was grown aerobically in several flasks containing 0.5 L of the defined minimal medium listed above. The flasks were constantly shaken at 150 rpm at 30°C. At late-exponential growth phase (∼4 days), the cells (and exopolysaccharides) were centrifuged at 5000 rpm and washed twice by resuspending the cells in 50 mL of 0.01 M NaNO3 (pH 6) to remove excess media. Prior to and after the washing procedure, 1 mL of the cell suspensions was incubated with a Live/Dead stain (Molecular Probes, Inc.) and inspected using an E600 Nikon epifluorescent microscope. Approximately >95% of the bacterial cells were green (live) in the assays. Subsequently, the cells were split into aliquots of approximately 150 mg cells (wet weight) and resuspended in 50 mL of variable concentration Pb(NO3)2 solutions at an initial pH of 4.5. The pH was adjusted by additions of 0.01 M NaOH or HCl at the beginning of the equilibrations and adjusted, if necessary, every 6 h, over a 24-h incubation period. During the equilibration with Pb, the cell suspensions were purged with N2 gas to prevent the saturation and precipitation of lead carbonate species. To probe the pH dependence of Pb uptake by B. cepacia, experiments were conducted with a fixed initial [Pb] of 10-4.4 M Pb(NO3)2 in 0.01 M NaNO3, and pH was varied between 2.75 and 7.5 (series B; fixed [Pb], varied pH). To examine changes in the mode of Pb(II) binding to B. cepacia over a large range in [Pb], B. cepacia was incubated with Pb(NO3)2 solutions in 0.01 M NaNO3 with initial [Pb] ranging between 10-6.3 and 10-3.3 M at pH 6 (series C; varied [Pb], fixed pH). Cell viability for samples across the pH and [Pb] range was confirmed using the Live/Dead stain. After equilibration, the suspensions were centrifuged at 15 000 rpm, and the supernatant was decanted and acidified to measure the residual [Pb] by ICP analysis. Pb LIII EXAFS measurements for the Pb-reacted B. cepacia samples from the series B and C experiments were performed at SSRL on wiggler beamline 4-3. The incident X-ray beam was monochromatized using a Si(111) double-crystal monochromator with 2 mm × 20 mm slits prior to the first nitrogenfilled ionization chamber (I0). Each 2 mm × 20 mm Teflon sample holder was placed in the beam path between the first and the second ionization chambers (I0 and I1) at 45° to a xenon-filled Stearn-Heald-type Lytle detector (16) with a 6× As filter to reduce elastic scatter and background fluorescence. An elemental Pb calibration foil was placed between the second and the third ionization chambers, and the first inflection of the absorption edge was taken at 13055 eV (log I1/I2). Scans through the near-edge region were collected in 1-eV steps and through the EXAFS region in 0.5Å-1 steps. Typically 12-16 successive scans were averaged together, and the data were background subtracted and normalized to a k3-weighted spline. E0 for the EXAFS was taken at 13070 eV. The EXAFS spectra were fit using a linear combination fitting procedure using the DATFIT module of EXAFSPAK program, with a library of more than 30 empirical model spectra, which include lead carbonates, phosphates, sulfates, Pb bound to humic substances, and a variety of organic model compounds (see refs 17 and 18). Data were fit over the k-range of 1-9 Å-1, and the fits were optimized by minimizing the residual. The fitting routine used fractions of only two model spectra to describe all of the EXAFS spectra collected in this study: (i) Pb sorbed to heat-treated B. cepacia (spectra C056) and (ii) pyromorphite (Pb5(PO4)3(OH)). The inclusion of additional model spectra did not significantly improve any of the fits. The fraction of the total Pb in each sample that is associated with B. cepacia as a sorption complex versus pyromorphite is given directly by the fraction of each model spectrum in the fit. Several samples from the series B and C experiments were run in duplicate to test the reproducibility 302
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003
of the EXAFS spectra. EXAFS spectra for comparable samples were typically fit with proportions of pyromorphite and adsorbed Pb that varied less than 5%. Structural fitting of the model EXAFS spectra collected for Pb sorbed onto heat-treated B. cepacia (spectra C056) was performed using the OPT module of the EXAFSPAK program (19) to determine the local coordination environment of Pb in the sample. Scattering contributions to the first and second shells in the Fourier transforms (FTs) of the EXAFS spectra were isolated by back-transforming windowed regions of the FT. The filtered data were fit to determine the identity of the back-scatterers contributing to each shell and to obtain starting values for the final fits of the raw data within the k-range of ∼2.5-10 Å-1. Phase and amplitude functions for Pb-O, Pb-S, Pb-Pb, Pb-C, and Pb-P were derived from model compounds using Feff 7.0 (20). Variables in the fitting procedure included energy shift (∆E0), coordination number (N), radial distance (R), a Debye-Wallertype disorder factor (σ2) for each shell, and a scale factor (So) applied to all shells in a fit. In the fitting procedure, N and R values were derived by nonlinear least-squares fitting, while σ2 and So were fixed to 0.01 Å2 and 0.8, respectively. E0 was derived from the best fit for the first shell Pb-O. R values are accurate to approximately (0.03 Å for first shell distances and (0.05 Å for second neighbors. N values are typically accurate within ∼30% (21), although they are strongly correlated with So. Transmission Electron Microscopy. TEM images were collected for a B. cepacia sample incubated with 10-4.2 M [Pb]i at pH 6 (same conditions as the series C experiments). After a 24-h incubation with Pb, the cells were fixed in 2.5% glutaraldehyde, embedded in 2% Noble agar, subjected to an ethanol dehydration series, infiltrated using hard-grade LR White, and cured at 60°C for 24 h. The 50-nm-thick sections of the polymerized blocks were mounted on copper mesh grids with Formvar support film sputtered with carbon. The sections were observed directly without any post-staining to visualize only electron-dense deposits (i.e., Pb-bearing phases). After the images were collected, the grids were immersed in a 1% uranyl acetate solution for 5 min to visualize cell morphology. Transmission electron micrographs were obtained using a 200-kV JEOL 2010 high-resolution analytical electron microscope coupled with Oxford Link ISIS electron dispersive spectoscopy (EDS) system at the Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory.
Results Pb Distributions in the Biofilm/Single-Crystal Experiments. Pb LR fluorescence maps were generated by rastering a 5 µm × 5 µm beam at 13.5 keV (above the Pb LIII edge) across at least a 100 µm × 100 µm region of each B. cepacia/R-Al2O3 sample. In Figure 1, samples A1 (X-ray-treated) and A2 (untreated) show an enhanced Pb signal in the dark biofilm regions as compared to “clean”, bare R-Al2O3 crystal regions because of significant sorption of Pb by the bacterial cells and exopolysaccharide matrix. The biofilm regions of sample A1 show only small variations (less than 20%) in Pb counts measured per pixel, indicating that Pb is relatively homogeneously distributed and that no discrete accumulation of Pb phases occurs in the X-ray-inactivated biofilm. Comparison of the average fluorescence yield obtained from bulk biofilm regions with the thin-film standard provides a surface coverage estimate of approximately 4.8 µmol of Pb/m2 crystal surface area. This is in reasonable agreement with Pb uptake predictions of ∼6 µmol/m2 at 10-4.2 M Pb made in an independent study (14). In the Pb LR fluorescence map for sample A2, the Pb counts vary by more than an order of magnitude between the brightest 5 µm × 5 µm pixels versus regions of lowest Pb
FIGURE 2. Transmission electron micrographs of B. cepacia incubated with 10-4.3 M Pb(NO3)2, pH 6. (A) No additional electron-dense stains were added; inset: selected-area diffraction pattern of the crystalline Pb precipitate. (B) TEM grid was immersed in uranyl acetate to visualize cell structures. counts. Two of the “hottest” regions appear to be concentrated within the isolated microcolonies on the crystal surface. A third hot spot also occurs within the contiguous biofilm. These observations suggest that when the bacterial cells are removed from growth media and transferred to a Pb(NO3)2 solution, a strongly heterogeneous distribution of Pb within the biofilm occurs as well as enhanced Pb uptake. Grazing-incidence Pb LIII EXAFS spectra collected for two duplicate samples of A1 and A2 are also shown in Figure 1. The EXAFS spectra for the X-ray-treated sample (A1) shows a relatively simple oscillatory pattern whereas the untreated sample (A2), containing Pb hot spots, exhibits a strong beat pattern and shoulder in the first oscillation at k ∼4 Å-1 because of the formation of an additional Pb phase. Comparison to EXAFS spectra collected for heat-treated B. cepacia and untreated cell suspensions, as presented and discussed below, indicates that the additional Pb phase in the active biofilms is a lead phosphate precipitate. Pb Uptake Data and Pb EXAFS Spectra for CellSuspension Experiments. Uptake as a Function of pH; Series B. Washed cell suspensions of B. cepacia incubated with variable pH, 10-4.4 M Pb(NO3)2 solutions show a steep adsorption edge at ∼pH 4.4 (series B; Figure 3). The initial assumption was that EXAFS data would highlight changes in the dominant functional groups involved in Pb binding at low versus high pH. Instead, the EXAFS data show that a variable level of adsorption and precipitation is occurring across the pH range. EXAFS spectra collected at low pH values (i.e., pH 4) directly match the spectrum of the lead phosphate mineral pyromorphite (Pb5(PO4)3(OH)), as shown in the inset to Figure 3. This suggests that a large fraction of the Pb uptake at low pH can be attributed to the formation of a precipitated form of Pb that at least has local structural order around Pb similar to that of pyromorphite. At higher pH, the EXAFS data only partially resemble pyromorphite, particularly the strong shoulder at k ∼4 Å-1 (see the series C EXAFS spectra collected at pH 6 in Figure 4). Quantitative fits of the series B EXAFS spectra (see Materials and Methods) show that Pb reacted with B. cepacia at low pH results in the formation of pyromorphite (90% of the total uptake), whereas at higher pH a mixture of sorbed Pb and biomineralized Pb occurs (58% pyromorphite, 42% sorbed at pH 6; sample ii in Figure 4). Uptake as a Function of [Pb]; Series C. Washed cell suspensions of B. cepacia incubated with a variable [Pb] at
FIGURE 3. Series B: Pb(II) uptake as a function of pH for washedcell suspensions incubated with 10-4.4 M Pb(NO3)2. Inset: Comparison of the EXAFS spectra for the Pb associated with B. cepacia at ∼pH 4 and EXAFS spectrum for the model compound pyromorphite, Pb5(PO4)3(OH). pH 6 show steep uptake behavior at low [Pb] (
