Bioaccessible Porosity in Soil Aggregates and Implications for

Nov 1, 2015 - Qingmei Liu , Qibin Li , Ning Wang , Dan Liu , Li Zan , Le Chang , Xuemei ... Holly A. Yu , Niamh Nic Daeid , Lorna A. Dawson , David A...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/est

Bioaccessible Porosity in Soil Aggregates and Implications for Biodegradation of High Molecular Weight Petroleum Compounds Ali Akbari and Subhasis Ghoshal* Department of Civil Engineering, McGill University, Montreal, Quebec H3A 0C3, Canada S Supporting Information *

ABSTRACT: We evaluated the role of soil aggregate pore size on biodegradation of essentially insoluble petroleum hydrocarbons that are biodegraded primarily at the oil−water interface. The size and spatial distribution of pores in aggregates sampled from biodegradation experiments of a clayey, aggregated, hydrocarbon-contaminated soil with relatively high bioremediation end point were characterized by image analyses of X-ray micro-CT scans and N2 adsorption. To determine the bioaccessible pore sizes, we performed separate experiments to assess the ability of hydrocarbon degrading bacteria isolated from the soil to pass through membranes with specific sized pores and to access hexadecane (model insoluble hydrocarbon). Hexadecane biodegradation occurred only when pores were 5 μm or larger, and did not occur when pores were 3 μm and smaller. In clayey aggregates, ∼ 25% of the aggregate volume was attributed to pores larger than 4 μm, which was comparable to that in aggregates from a sandy, hydrocarboncontaminated soil (∼23%) scanned for comparison. The ratio of volumes of inaccessible pores (4 μm) in the clayey aggregates was 0.32, whereas in the sandy aggregates it was approximately 10 times lower. The role of soil microstructure on attainable bioremediation end points could be qualitatively assessed in various soils by the aggregate characterization approach outlined herein.



INTRODUCTION The rates of biodegradation of organic compounds in soil and groundwater environments have been shown to be controlled by the rates of mass transfer from sorbed phase or nonaqueous phase liquid (NAPL) to the aqueous phase or by their intrinsic rates of biodegradation.1 However, there are exceptions to this dissolution−degradation framework. For example, direct uptake of oil-phase hydrocarbon compounds following contact between bacterial cells and the oil−water interface, is the major biodegradation mechanism of a fraction of poorly soluble petroleum hydrocarbon compounds2 (e.g., hexadecane, which has a solubility less than 1 ppb3). In addition, close proximity or attachment of hydrocarbon-degrading bacteria to NAPL may enhance the microbial uptake and dissolution rate of NAPL hydrocarbon solutes compared to abiotic systems by decreasing the diffusion distance and increasing the concentration gradient for dissolution.4 For example, Ortega-Calvo and Alexander reported higher biodegradation of naphthalene when bacteria were attached to a heptamethylnonane/naphthalene NAPL, compared to when attachment was inhibited by adding a nonionic surfactant.5 Interfacial uptake of poorly soluble NAPL components would be severely limited if these compounds are entrapped in small soil micropores which size excludes bacteria and thus prevents the direct contact between bacteria and NAPL, and this may contribute to a high bioremediation end point. © XXXX American Chemical Society

Petroleum NAPLs discharged into soils can migrate into micropores of soils by altering the wettability of soil mineral domains through a multistep process which starts with binding of the polar fractions of oil to soil mineral surfaces.6 Using confocal laser scanning microscopy Karimi-Lotfabad and Gray showed that NAPL were present in soil aggregate pores of various sizes including those smaller than 5 μm inside clay aggregates.7 In our recent study, 1 year of biotreatment of an aged, petroleum-contaminated, aggregated clayey soil from a subArctic site in Northwest Territories (NWT), Canada, resulted in biodegradation of nonvolatile petroleum hydrocarbons (>C16−C34) from an initial concentration of 1068.8 ± 138.0 to an end point of 525.8 ± 77.1 mg/kg-soil.8 The end point of nonvolatile hydrocarbon fraction was however significantly lower at 102.5 ± 20.5 mg/kg for an aged, petroleum− contaminated, sandy, sub-Arctic soil from Resolution Island (RI), Nunavut, Canada, after extended biotreatment.9 GC analyses of the residual hydrocarbons in the NWT soil clearly showed potentially biodegradable hydrocarbons (e.g., 30% of octadecane remained after 1 year), which suggests that the Received: July 26, 2015 Revised: October 27, 2015 Accepted: October 31, 2015

A

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology



Article

MATERIALS AND METHODS Micro-CT Scanning of Soil Aggregates. A high resolution X-ray micro-CT scanner (SkyScan 1172) was employed to image the interior architecture of several soil aggregates and to determine their pore size distributions and pore volumes. Intact, undisturbed macro- (dp > 2 mm) and meso-aggregates (0.6 < dp < 2 mm) were carefully sampled from biodegradation experiments (described in the Supporting Information), exposed to air until the aggregates moisture content dropped to ∼1%. Then, aggregates were placed with extreme care and fixed on pipet tips for scanning. Aggregates were air-dried to minimize changes in gray scale values of pixels representing soil moisture in different projections during the several hours needed for X-ray scanning a whole aggregate at high resolution. Rapid scans for about 20 min at low resolution (about 11 μm) of an NWT aggregate before and after exposing to air flow, when moisture content dropped to ∼1%, did not show significant alternation of soil aggregate microstructure. The pipet tips were mounted on the rotating base inside the micro-CT. The average intensity of X-ray transmission was adjusted at 50% by tuning the voltage and exposure time at each desired resolution. The resolution of acquired images was in the range of 1.04−3.28 μm, depending on the size of aggregate. To reduce scanning artifacts due to beam hardening, preferential absorption of low energy X-ray in the outer layer of object, a 0.5 mm aluminum filter was used. Images were acquired while sample was rotating over 360° with 0.28° step intervals. At each angular position three or four images were acquired and averaged. Image Analysis. The images obtained from scanning were reconstructed using NRECON software (SkyScan). However, to extract the information on intra-aggregate pores, the reconstructed images need to be thresholded to distinguish solid and void voxels and the boundary of aggregates need to be defined for the 500−1500 imaged cross sections of the aggregates. This cannot be reliably done for heterogeneous materials with irregular shapes such as soil aggregates with commercially available software, and therefore we developed our own code for image analysis as follows. The reconstructed two-dimensional (2D) images from micro-CT scanning were thresholded into two populations as soil (solids) and pores (air). Aggregates were air-dried to a moisture content of approximately 1%. Also, the volume of residual TPH after treatment was on the order of 0.1% of the total volume of the aggregates. Therefore, thresholding the images to only two phases, soil and air, did not cause a significant error in pore volume calculations. Thresholding is a critical step which is particularly complex in the case of clayey soil aggregates with constituents of various densities and nonuniformly distributed pore network. Moreover, fine soil particles could be similar in size or smaller than the attainable scanning resolution which leads to pixels with intermediate X-ray intensities between air and solid. Thus, simple thresholding based on a single global value cannot adequately characterize the aggregate structure. Among different thresholding methods that were investigated in our study (global thresholding, alternating mean thresholding and median filtering,16 two sigma smoothing and low pass filter,17 indicator kriging (IK)18), the IK method provided an image closest to real aggregate microstructure. In this study, the IK method with additional modifications to preserve all detectable micropores (1 pixel) of aggregates was used for image thresholding. The algorithm starts with the initial

residual hydrocarbon phase were not bioaccessible. We hypothesize that this biodegradation pattern is related to the pore structure of the aggregates that allowed hydrocarbon degraders to access a fraction of the pore volume, and yet prevented the accessibility to the remaining pore volume. An important first step toward evaluating this hypothesis is a detailed characterization of the pore size and spatial distributions in soil aggregates for determination of the bioaccessible and bioinaccessible pore volumes. It should be noted that in addition to pore size exclusion, irreversible sorption to soil background organic matter may contribute to limited bioaccessibility and bioremediation end point of hydrophobic compounds. Soil background organic matter contents were relatively low and comparable in NWT (2.3%) and RI (2.4%) soils. Different approaches have been employed for characterization of soil pore network microstructure, such as microscopic-imaging of sections of resin-impregnated soil aggregates,10 gas adsorption based methods, and X-ray computed tomography (CT) scanning.11 Physical sectioning of aggregates is time-consuming, allow only a limited number of sections to be imaged, and can lead to perturbation of the aggregate structure.12 Gas adsorption based methods for characterization of pore size distribution in soils do not provide information on the spatial distribution of pores. X-ray micro-CT has been successfully employed as a nondestructive method to study the three-dimensional pore structure of soil aggregates.13,14 Analyses of images acquired from CT-scanning provide information on spatial distribution, connectivity as well as diameter of pores inside the aggregate. Thus, X-ray CT-based methods can provide unique insights on the pore space in soil aggregates that can be accessed by bacteria where biodegradation through direct contact is possible. In this study, we characterized the pore network of aggregates of petroleum hydrocarbon-contaminated soils from NWT and RI subjected to long-term biodegradation in pilotscale bioreactors, by micro-CT scanning. In addition N2 adsorption analysis was used as a complementary pore size distribution characterization technique for pores smaller than 1 μm, which were too small to be characterized by micro-CT scanning at the resolutions practical for scanning of the entire aggregate. To obtain direct evidence of the minimum bioaccessible pore size for a dominant hydrocarbon degrader present in the contaminated soils used in our experiments, we assessed the biodegradation of hexadecane, a model poorly soluble petroleum hydrocarbon compound, in carefully designed bioreactors with fixed NAPL-water interface separated from the dominant hydrocarbon degrading bacteria in NWT and RI soils by membranes with specific pore diameters. It has been shown that microbial uptake of hexadecane occurs only at oil−water interfaces.15 The data from the CT and N2 adsorption analyses and the knowledge of pore sizes accessible by the dominant hydrocarbon degrader present in the soils was used to evaluate the potentially bioaccessible porosity (ratio of bioaccessible pore volume and total aggregate volume) and inaccessible pore volume fraction (ratio of bioinaccessible pore volume and total pore volume) of the aggregates. The potential contribution of these parameters in bioremediation end points of poorly soluble biodegradable hydrocarbon contaminants in soil aggregates is discussed. B

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology assignment of subsets of pixels to the background and the object using two low and high threshold values (T0,T1), and is followed by kriging and two filtering steps.18 The intermediate gray scale values are classified using indicator kriging and based on estimates of spatial covariance of the image. The details of algorithm and modifications introduced by authors are described in the Supporting Information. To avoid considering the air phase surrounding the aggregate in calculation of the aggregate porosity, several previous studies analyzed only the core of soil aggregates.17,19 A new procedure was developed in this study to determine the boundary of aggregate and analyze the whole aggregate structure. Considering just the core of aggregate may significantly bias the calculated porosity as the core of aggregates are usually denser than outer layers of aggregate. To determine the boundary of soil aggregates, we determined the alpha-shape of the aggregate body. The alpha-shape, as introduced by Edelsbrunner et al., is a formal shape of point sets based on Delaunay triangulation.20 The alpha value (1/maximum radius of triangles) was set at 0.02 to balance the degree of details in the boundary and a reasonable computing time. To minimize any artifacts arising from edge determination and particularly to avoid accounting the surrounding air as intra-aggregate porosity, we shrank the resulting alpha-shape to 95% of its original size, which was then applied as a mask to the IK binarized image. Further details of the shape analysis procedure are presented in the Supporting Information. The image analysis algorithm was programmed with MATLAB (Mathworks, Natick, MA). The analyses were performed using the Guillimin cluster of the Compute Canada High Performance Computing Facilities. Paraview21 (Kitware, Clifton Park, NY) was used for volume rendering and threedimensional (3D) representation of soil aggregate structure. Given the long computational time for image analysis, 6 representative aggregates chosen qualitatively based on reconstructed images from CT scanning among 14 total scanned aggregates were analyzed to extract pore network information. Information of scanned aggregates which were analyzed are presented in Table S2 (Supporting Information). Bioreactor Experiments to Determine Bioaccessible Pore Sizes. Accessibility of bacteria to a hexadecane layer through different pore sizes, and subsequent biodegradation of radiolabeled hexadecane (2.5 × 104 dpm/μL), as a model, essentially insoluble NAPL, was assessed in specially designed bioreactors. Bioreactor systems were custom-made by a glass blower at McGill University. The specific activity of hexadecane was adjusted by mixing radiolabeled hexadecane (hexadecane, n-[1−14C] specific activity, 0.225 uCi/ul without solvent; American Radiolabeled Chemicals Inc., St. Louis, MO ) and nonlabeled hexadecane (anhydrous, ≥ 99%, Sigma-Aldrich Canada Co., Canada). The schematic of the bioreactor is presented in Figure 1 which shows a vial fitted with membrane of specific pore size that separates the oil−water interface in the top chamber from the bacterial suspension in the lower chamber. In order to access to the hexadecane oil phase, bacteria need to pass through membrane pores. Nuclepore track-etched membranes (GE Healthcare Bio-Sciences, Canada) with pore sizes of ∼12, 5, 8, 3, and 0.4 μm were used. These membranes have channel-like pores with uniform diameters. The membranes were made of polycarbonate coated with polyvinylpyrrolidone (PVP) to improve the wettability. Membranes were slightly negatively charged with a zeta potential of −4.82 mV in Bushnell Haas (BH) mineral solution as measured by a streaming potential analyzer (EKA,

Figure 1. Schematic of the bioreactors with fixed oil−water interfacial areas fitted with membranes with different pore sizes. The aqueous phase was a BH mineral nutrient solution and bacteria was inoculated below the membrane. Hexadecane was used as a model NAPL phase.

Brookhaven Instruments Corporation, Holtsville, NY). The contact angle of membrane was measured as 50.1 ± 4.3 (as described in Supporting Information). Bioreactors were inoculated with bacterial pure culture of Dietzia maris (Dietzia maris; CA160; GQ870425 (99%)) which was isolated from the NWT soils, according to the method presented in the Supporting Information. More than 80% of identified alkB gene harboring bacterial community of NWT site soils were closely related to Dietzia maris and belonged to Corynebacterineae suborder.22 Corynebacterineae members were also detected as major alkB gene harboring hydrocarbon degrader in RI soil.23 The alkB gene is responsible for encoding the AlkB protein involved in initial activation of aliphatic hydrocarbon metabolism.24,25 The strain was nonmotile as determined by a motility assay described in Supporting Information. The zeta potential of the strain was −33.83 mV in BH mineral solution calculated based on the electrophoretic mobility measured by laser Doppler velocimetry (Zetasizer, Malvern). A MATH test (described in Supporting Information) showed that Dietzia maris cells were hydrophobic (74% attachment of bacteria to a hexadecane layer). The size (length and width) of bacterial cells as determined from environmental scanning microscopy (ESEM) and SEM images was 0.92 ± 0.13 and 0.70 ± 0.07 μm (n = 50). Details of the SEM imaging procedures are provided in Supporting Information. The biomineralization time profile of hexadecane was determined by monitoring the 14CO2 evolution over time. A headspace trap containing 2 mL of 2 M NaOH solution attached to the upper chamber of the bioreactor (Figure 1) was periodically sampled to determine the amount of 14CO2 produced. At the end of each set of experiments, 14C mass balance was performed by sampling the NAPL phase and aqueous phase and accounting for the 14CO2 produced. More than 90% of the initial radioactivity was recovered in all systems, indicating that there was minimal abiotic loss of hexadecane. Detailed information on characterization of bacteria and membrane as well as experimental set up are provided in the Supporting Information.



RESULTS AND DISCUSSION Characterization of Aggregate Microstructure. Pore diameters and associated pore volumes of NWT and RI soil aggregates sampled from bioreactors8,9 were determined with micro-CT scanning and N2 adsorption analyses. Figure 2A presents an example of a reconstructed image from CT scanning with a range of gray scale values presenting different constituents of soil particles with different densities. The higher C

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

The image analysis procedure employed also enabled us to precisely determine that 92−96% of total pore volume was connected as a large open pore in representative NWT aggregates, when analyzed in 3D. The connectivity of pore assigned pixels were determined based on the 26 connected neighborhood pixels of the cross-section and its two adjacent cross sections (slices). The heterogeneous distribution of pores in the aggregates is demonstrated in Figure 3, which shows 3D representations of aggregates from the clayey, fine-grained, NWT soil (Figure 3A,B) and from the coarse-grained, sandy RI soil (Figure 3C,D). The 3D images were obtained from volume rendering of all cross sections for each aggregate and highlights the differences in microstructure and pore sizes of sandy and clayey aggregates. To characterize the bioaccessible pore volume from CT images of each aggregate, we first determined the size distribution of pore areas in thresholded cross sections, such as in Figure 2B. The pore diameters were estimated based on a circular cross-sectional area of equal magnitude to the actual, irregular-shaped, pore cross-sectional area. These number distributions of equivalent pore diameter from 500−1500 cross sections (depending on the aggregate size) for each aggregate were then combined, and normalized to aggregate volume in order to facilitate comparison between aggregates with different volumes. The aggregate volume was calculated from processed CT images of the aggregate. The normalized size distributions for representative aggregates are shown in Figure 4A for the clayey NWT soil and the sandy RI soil. The dimensions and scanning resolutions for NWT and RI aggregates are presented in Table S2 (Supporting Information). There were a large number of small pores in all aggregates as determined by the number of pores of small diameters encountered in all cross sections (e.g., > 103 pore cross

Figure 2. (A) A reconstructed cross section from micro-CT scanning provided by NRECON software (SkyScan); scanning resolution, 2.96 μm. (B) The same image after thresholding with indicator kriging and follow-up processes to preserve the whole aggregate body. The white areas in image B represent the intra-aggregate pore areas.

the density, the atomic number, or both, the brighter is the pixel in the image. The images provide spatial distribution of pores which are heterogeneously distributed inside aggregates. The same cross section after image processing is shown in Figure 2B where intra-aggregate pores are distinguished from the surrounding void and soil particles. Determining the boundary of aggregate enabled us to perform image analysis for the whole aggregate body and thus eliminated bias related to selection of only the core of the aggregate. The bias could be substantial as demonstrated by the fact that the porosity of a 0.9 × 0.9 × 1.1 mm cube from the core of an aggregate from the NWT soil was calculated as 14%, compared to a porosity of 26% when the whole aggregate body was analyzed.

Figure 3. (A) Three-dimensional representation of a clayey soil aggregate (NWT1). (B) Cross sections of clayey aggregate (NWT1). (C) Threedimensional representation of a sandy soil aggregate (RI). (D) cross sections of sandy aggregate (RI). D

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 4. (A) The size distribution of pore cross sections calculated from all aggregate cross sections (500−1500 cross sections for different aggregates) normalized to total aggregate volume. (B) Cumulative volume of pores larger than specific pore sizes as percent of total aggregate volume. (B, inset) Blown-up plot of NWT1 for pores with 1−29 μm diameters; arrows indicate the range of bioaccessible porosity for two scenarios assuming 1 or 4 μm as minimum diameter that bacterium/bacteria can pass through.

The bioaccessible porosity values of the NWT fine-grained soil were comparable to the values calculated for the RI1 and RI2 aggregates (23.4 and 27.3%). The range of bioaccessible porosity values in the NWT and RI aggregates imply that at high residual saturation, for example, following an oil spill, there will be a high potential for biodegradation of hydrocarbons entrapped in the aggregate given that significant fraction of soil porosity will be bioaccessible to microorganism. The bioaccessible porosity could be calculated solely from CT scanning data. Given the limitation of CT-scanning for detecting pores smaller than its attainable resolution, N2 adsorption analysis (described in Supporting Information) was conducted to determine the volume of smaller pores in the range of 17 Å to 1 μm, the range of pore diameters relevant to gas adsorption analyses. A significantly high volume of pores (106 pore cross sections) at an assumed oil saturation of 10% could accommodate NAPL equivalent to approximately 1000 mg/kg soil. E

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 5. (A) Time profiles of mineralization of n-hexadecane to CO2 in bioreactors with fixed oil−water interfacial areas with membranes with different pore sizes. The error bars indicate the standard deviation of the mean of three replicate samples from three different bioreactors (n = 9), except in the case of 8 μm data where the results are from three replicates from one bioreactor (n = 3). (B) ESEM image of Dietzia maris on 3 μm membranes illustrating the aggregation of bacteria. (C) SEM image of Dietzia maris illustrating the aggregation of bacteria.

Effect of Pore Size on Hydrocarbon Biodegradation. The controlled pore experiments were designed to determine the range of pore diameters for which NAPL would be accessible to a hydrocarbon degrader, Dietzia maris, isolated from the clayey NWT soil. Figure 5A shows the percent of radiolabeled hexadecane mineralized to CO2 by bacteria during the experiments in systems with membranes of different pore diameters. Mineralization activity was taken as evidence that the NAPL was accessed by Dietzia maris cells by crossing the membrane through the pores. More than 10% of hexadecane was mineralized in 41 days in the systems without membrane and in bioreactors where the hexadecane and inoculated bacteria were separated with 5, 8, or 12 μm pore diameter membranes. In contrast, in the reactors fitted with 0.4 and 3 μm membranes, mineralization of hexadecane was not detected. In bioreactors without a membrane or with membranes with 5, 8, or 12 μm pore diameter, the attachment and growth of bacteria on fixed oil−water interface was clearly visible after about 2 weeks. Figure 5A shows the exponential pattern of mineralization of hexadecane in these systems, suggesting there was exponential bacterial growth following attachment to the NAPL-water interface and uptake of hexadecane. The lack of biodegradation in the systems with 0.4 μm pore sizes was expected, given the very low solubility of hexadecane on one hand and the larger size of bacteria on the other hand. However, although the 3 μm pore diameter is significantly larger than the diameter of a single cell of bacteria, the bacteria appear to have been unable to access the hexadecane−water interfacial area. These results are consistent with the findings by Strong et al.,27 who examined the in situ decomposition of plant organic matter in soils with different pore size distributions and found that the maximum biological activity in terms of carbon decomposition rate occurred in soils with large volume of pores in the range of 15−60 μm diameters, whereas, organic matter mostly remained nondegraded in soils with large volume of pores smaller than 4 μm. Several previous studies have reported unhindered bacterial motility in case of micro channels with widths as small as 1.117 or 2 μm.28,29 Mannik et al. studied swimming speed of Escherichia coli and Bacillus subtilis in micro channels etched in silicon wafers with widths between 0.3 and 5 μm under nutrient gradients that resulted in bacterial transport by chemotaxis.

They found that the average of E. coli cell velocity while passing through channels with width as small as 1.1 μm was comparable to the velocity observed in larger channels or nonconstricted chambers.26 They also reported that in the densely packed chambers where bacteria are pressed into channels, bacteria can penetrate pores as small as 0.4 μm, as a result of growth and deformation (flattening) inside the pores.26 In our experiments, given the nonmotile nature of Dietzia maris, chemotaxis-driven transport of bacteria is unlikely. The membranes were not in direct contact with hexadecane, and given very low solubility of hexadecane, there would be virtually no dissolved hexadecane in or diffusion toward the membrane, and initial adhesion or the growth of bacteria inside the pores was not possible. Contrary to our study, the above-mentioned studies were performed with motile bacterial strains. Furthermore, given the negative charge of bacteria and slightly negative charge of membranes the electrostatic attraction between bacteria and membrane pore walls and hindrance to transport through the pores is unlikely. Aggregates of coccus-shaped cells of Dietzia maris in solutions sampled from the lower chamber of the bioreactors was abundant in ESEM images acquired with minimal perturbation (without dehydration). The length and width of bacterial cell aggregates were 6.1 ± 1.6 and 4.8 ± 1.12 μm (n = 22) as determined from ESEM and SEM images such as Figure 5B and 5C. These figures show SEM images of bacterial suspension from the lower chamber of a bioreactor. Aggregation of different members of Rhodococcus genus, phylogenetically close to Dietzia genus has been reported, and aggregation was found to be correlated with hydrophobicity of bacterial cells.30 In natural soil environments, bacteria are predominantly present as aggregates, either freely floating or attached as biofilms on surfaces. Extracellular polymeric substances act as binding agents in such aggregates.31 The location of these patches of bacterial communities has been correlated to nutrient gradients, spatial locations of soil organic matter, and pores.12 In real soil environments, bacteria may also interact with soil minerals, mainly through electrostatic interactions.32 Bacteria may adhere at the surface of minerals, and this could cause changes in their viability or physiology.33 The bioreactor experiments were specifically designed to identify the exclusion pore size for bacteria rather F

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology



than serve as synthetic model aggregate. The membrane surface is different in composition from silica, clays, alumina, and mineral carbonates which are dominant minerals in soils. However, the membrane surface was negatively charged at pH ∼ 7 in the media solution of the bioreactors. Environmental Implications. Overall, our results suggest that in the case of fine-grained, clayey soils, such as the NWT soil, where a significant number of pores are smaller than or equal to the range of bacterial aggregate size and thus direct contact between entrapped oil in small micropores and bacteria would not be feasible, the bioaccessibility of a fraction of poorly soluble hydrocarbons will likely be limited. However, considering the significant volume of bioaccessible pores a significant biodegradation extent is potentially possible. In coarse-grained, less porous soils, where pores are predominantly bioaccessible, the bioremediation end point will be mainly governed by biodegradability of residual NAPL components. The soil aggregate characterization methods employed in this study could be useful tools for quantifying the bioaccessible porosity and bioaccessible/bioinaccessible pore volume fractions and for qualitative initial assessment of bioremediation effectiveness and end point for different soils. Characterization of the bioaccessible pore volumes can allow assessments of the size of the bacterial habitat inside soil aggregates and their potential for access to carbon sources, which can provide important insights on biodegradation activity related not only to bioremediation but also to greenhouse gas emissions from soil respiration and microbial enhanced oil recovery. With respect to bioremediation of petroleum hydrocarbon contaminated soils, the suite of tools for characterization of soil aggregates can be further improved by employing techniques that can determine the distribution of NAPL in soil aggregate pores. Given the low contrast between NAPL, water, and soil constituents due to the resolution limitations of micro-CT and the distributed mass of a relatively small amount of NAPL in the aggregates, the mapping of the NAPL in the aggregates may be challenging even by synchrotron CT scanning. However, next-generation scanning technologies at higher resolutions and lower energies might provide the opportunity to characterize even smaller pores and differentiate between materials/phases in soil aggregates. Furthermore, it is important to determine the limiting pore diameter for different, dominant hydrocarbon degrading bacteria in different contaminated soils. It is possible that some hydrocarbon-degrading bacteria will be able to access hydrocarbon in pores of size similar to that of single cells, whereas other bacteria may aggregate more extensively than Dietzia maris and, thus, only access pores at least several microns in diameter. This will significantly impact the bio(in)accessible pore volumes.



Article

AUTHOR INFORMATION

Corresponding Author

*Phone: (514) 398-6867. Fax: (514) 398-7361. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), and Imperial Oil, Ltd. Isolation of Dietzia maris strain from NWT petroleum hydrocarbon-contaminated soils was conducted by Dr. Wonjae Chang at McGill University. The authors thank the anonymous reviewers for their constructive and valuable comments.



REFERENCES

(1) Ghoshal, S.; Luthy, R. G. Bioavailability of hydrophobic organic compounds from nonaqueous-phase liquids: The biodegradation of naphthalene from coal tar. Environ. Toxicol. Chem. 1996, 15 (11), 1894−1900. (2) Leahy, J. G.; Colwell, R. R. Microbial degradation of hydrocarbons in the environment. Microbiol. Rev. 1990, 54 (3), 305−315. (3) Haynes, W. M. Aqueous Solubility and Henry’s Law Constants of Organic Compounds. In CRC Handbook of Chemistry and Physics. 96th ed. (Internet Version 2016) ed.; CRC Press/Taylor and Francis: Boca Raton, FL., 2016. (4) Seagren, E. A.; Rittmann, B. E.; Valocchi, A. J. Bioenhancement of NAPL pool dissolution: Experimental evaluation. J. Contam. Hydrol. 2002, 55 (1−2), 57−85. (5) Ortega-Calvo, J. J.; Alexander, M. Roles of bacterial attachment and spontaneous partitioning in the biodegradation of naphthalene initially present in nonaqueous-phase liquids. Appl. Environ. Microbiol. 1994, 60 (7), 2643−2646. (6) Huang, W.; Schlautman, M. A.; Weber, W. J. A distributed reactivity model for sorption by soils and sediments. 5. The influence of near-surface characteristics in mineral domains. Environ. Sci. Technol. 1996, 30 (10), 2993−3000. (7) Karimi-Lotfabad, S.; Gray, M. R. Characterization of contaminated soils using confocal laser scanning microscopy and cryogenicscanning electron microscopy. Environ. Sci. Technol. 2000, 34 (16), 3408−3414. (8) Akbari, A.; Ghoshal, S. Pilot-scale bioremediation of a petroleum hydrocarbon-contaminated clayey soil from a sub-Arctic site. J. Hazard. Mater. 2014, 280 (0), 595−602. (9) Chang, W.; Dyen, M.; Spagnuolo, L.; Simon, P.; Whyte, L.; Ghoshal, S. Biodegradation of semi- and non-volatile petroleum hydrocarbons in aged, contaminated soils from a sub-Arctic site: Laboratory pilot-scale experiments at site temperatures. Chemosphere 2010, 80 (3), 319−326. (10) Chun, H. C.; Giménez, D.; Yoon, S. W. Morphology, lacunarity and entropy of intra-aggregate pores: Aggregate size and soil management effects. Geoderma 2008, 146 (1−2), 83−93. (11) Chang, W.; Akbari, A.; Snelgrove, J.; Frigon, D.; Ghoshal, S. Biodegradation of petroleum hydrocarbons in contaminated clayey soils from a sub-arctic site: The role of aggregate size and microstructure. Chemosphere 2013, 91 (11), 1620−1626. (12) Nunan, N.; Wu, K.; Young, I. M.; Crawford, J. W.; Ritz, K. Spatial distribution of bacterial communities and their relationships with the micro-architecture of soil. FEMS Microbiol. Ecol. 2003, 44 (2), 203−215. (13) Werth, C. J.; Zhang, C.; Brusseau, M. L.; Oostrom, M.; Baumann, T. A review of non-invasive imaging methods and applications in contaminant hydrogeology research. J. Contam. Hydrol. 2010, 113 (1−4), 1−24.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b03618. Details of materials and methods of biodegradation experiments, N2 adsorption analysis, bioreactor experiments, isolation and characterization of Dietzia maris from petroleum hydrocarbon contaminated site soils, SEM and image analysis procedure for X-ray scans, and information on analyzed scanned aggregates. (PDF) G

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

(33) Cai, P.; Huang, Q.; Walker, S. L. Deposition and Survival of Escherichia coli O157:H7 on Clay Minerals in a Parallel Plate Flow System. Environ. Sci. Technol. 2013, 47 (4), 1896−1903.

(14) Dal Ferro, N.; Delmas, P.; Duwig, C.; Simonetti, G.; Morari, F. Coupling X-ray microtomography and mercury intrusion porosimetry to quantify aggregate structures of a cambisol under different fertilisation treatments. Soil Tillage Res. 2012, 119 (0), 13−21. (15) Holden, P. A.; LaMontagne, M. G.; Bruce, A. K.; Miller, W. G.; Lindow, S. E. Assessing the Role of Pseudomonas aeruginosa SurfaceActive Gene Expression in Hexadecane Biodegradation in Sand. Appl. Environ. Microbiol. 2002, 68 (5), 2509−2518. (16) Mardia, K. V.; Hainsworth, T. J. A spatial thresholding method for image segmentation. IEEE Trans. Pattern Anal. Machine Intell. 1988, 10 (6), 919−927. (17) Nunan, N.; Ritz, K.; Rivers, M.; Feeney, D. S.; Young, I. M. Investigating microbial micro-habitat structure using X-ray computed tomography. Geoderma 2006, 133 (3−4), 398−407. (18) Oh, W.; Lindquist, B. Image thresholding by indicator kriging. IEEE Trans. Pattern Anal. Machine Intell. 1999, 21 (7), 590−602. (19) Peth, S.; Horn, R.; Beckmann, F.; Donath, T.; Fischer, J.; Smucker, A. J. M. Three-dimensional quantification of intra-aggregate pore-space features using synchrotron-radiation-based microtomography. Soil Sci. Soc. Am. J. 2008, 72 (4), 897−907. (20) Edelsbrunner, H.; Kirkpatrick, D.; Seidel, R. On the shape of a set of points in the plane. IEEE Trans. Inf. Theory 1983, 29 (4), 551− 559. (21) Ahrens, J.; Geveci, B.; Law, C. 36 - ParaView: An End-User Tool for Large-Data Visualization. In Visualization Handbook; Hansen, C. D.; Johnson, C. R., Eds. Butterworth-Heinemann: Burlington, MA, 2005; pp 717−731. (22) Akbari, A.; Ghoshal, S. Effects of diurnal temperature variation on microbial community and petroleum hydrocarbon biodegradation in contaminated soils from a sub-Arctic site. Environ. Microbiol. 2015, http://dx.doi.org/10.1111/1462-2920.12846.10.1111/14622920.12846 (23) Chang, W.; Klemm, S.; Beaulieu, C.; Hawari, J.; Whyte, L.; Ghoshal, S. Petroleum Hydrocarbon Biodegradation under Seasonal Freeze−Thaw Soil Temperature Regimes in Contaminated Soils from a Sub-Arctic Site. Environ. Sci. Technol. 2011, 45 (3), 1061−1066. (24) Guibert, L.; Loviso, C.; Marcos, M.; Commendatore, M.; Dionisi, H.; Lozada, M. Alkane biodegradation genes from chronically polluted subantarctic coastal sediments and their shifts in response to oil exposure. Microb. Ecol. 2012, 64 (3), 605−616. (25) Hamamura, N.; Fukui, M.; Ward, D. M.; Inskeep, W. P. Assessing soil microbial populations responding to crude-oil amendment at different temperatures using phylogenetic, functional gene (alkB) and physiological analyses. Environ. Sci. Technol. 2008, 42 (20), 7580−7586. (26) Männik, J.; Driessen, R.; Galajda, P.; Keymer, J. E.; Dekker, C. Bacterial growth and motility in sub-micron constrictions. Proc. Natl. Acad. Sci. U. S. A. 2009, 106 (35), 14861−14866. (27) Strong, D. T.; Wever, H. D.; Merckx, R.; Recous, S. Spatial location of carbon decomposition in the soil pore system. Eur. J. Soil Sci. 2004, 55 (4), 739−750. (28) Biondi, S. A.; Quinn, J. A.; Goldfine, H. Random motility of swimming bacteria in restricted geometries. AIChE J. 1998, 44 (8), 1923−1929. (29) Binz, M.; Lee, A. P.; Edwards, C.; Nicolau, D. V. Motility of bacteria in microfluidic structures. Microelectron. Eng. 2010, 87 (5−8), 810−813. (30) Iwabuchi, N.; Sunairi, M.; Anzai, H.; Morisaki, H.; Nakajima, M. Relationships among colony morphotypes, cell-surface properties and bacterial adhesion to substrata in Rhodococcus. Colloids Surf., B 2003, 30 (1−2), 51−60. (31) Klebensberger, J.; Rui, O.; Fritz, E.; Schink, B.; Philipp, B. Cell aggregation of Pseudomonas aeruginosa strain PAO1 as an energydependent stress response during growth with sodium dodecyl sulfate. Arch. Microbiol. 2006, 185 (6), 417−427. (32) Rong, X.; Chen, W.; Huang, Q.; Cai, P.; Liang, W. Pseudomonas putida adhesion to goethite: Studied by equilibrium adsorption, SEM, FTIR and ITC. Colloids Surf., B 2010, 80 (1), 79− 85. H

DOI: 10.1021/acs.est.5b03618 Environ. Sci. Technol. XXXX, XXX, XXX−XXX