PAH Bioremediation with Rhodococcus rhodochrous ATCC 21198: Impact of Cell Immobilization and Surfactant Use on PAH Treatment and Post-Remediation Toxicity

Abstract

Polycyclic aromatic hydrocarbons (PAHs) are prevalent environmental contaminants that are harmful to ecological and human health. Bioremediation is a promising technique for remediating PAHs in the environment, however bioremediation often results in the accumulation of toxic PAH metabolites. The objectives of this research were to demonstrate the cometabolic treatment of a mixture of PAHs by a pure bacterial culture, Rhodococcus rhodochrous ATCC 21198, and investigate PAH metabolites and toxicity. Additionally, the surfactant Tween ^® 80 and cell immobilization techniques were used to enhance bioremediation. Total PAH removal ranged from 70–95% for fluorene, 44-89% for phenanthrene, 86-97% for anthracene, and 6.5-78% for pyrene. Maximum removal was achieved with immobilized cells in the presence of Tween ^® 80. Investigation of PAH metabolites produced by 21198 revealed a complex mixture of hydroxylated compounds, quinones, and ring-fission products. Toxicity appeared to increase after bioremediation, manifesting as mortality and developmental effects in embryonic zebrafish. 21198’s ability to rapidly transform PAHs of a variety of molecular structures and sizes suggests that 21198 can be a valuable microorganism for catalyzing PAH remediation, however implementing further treatment processes to address toxic PAH metabolites should be pursued to help lower post-remediation toxicity in future studies.

1.0. Introduction

Polycyclic aromatic hydrocarbons (PAHs) are a class of ubiquitous environmental contaminants that arise from both natural and anthropogenic sources and exhibit mutagenic, teratogenic, and carcinogenic properties (Gitipour et al., 2018; Mojiri et al., 2019). The broad range of physicochemical properties of PAHs reflects their range of molecular structures and sizes. A variety of characteristics can be used to organize PAHs into categories, including: molecular weight (low molecular weight (LMW) for three fused benzene rings or fewer, high molecular weight (HMW) for more than three fused benzene rings), structure (linear, angular, cluster), and aromaticity (alternant and non-alternant) (Patel et al., 2020; Sakshi and Haritash, 2020).

Remediation of PAHs can be challenging due to their hydrophobicity and environmental stability (Sakshi et al., 2019; Zhou et al., 2023). An attractive option for the remediation of PAH contaminated material is bioremediation, as it is less expensive and less invasive than physical and chemical removal strategies (Kuppusamy et al., 2017; Sakshi and Haritash, 2020; Thacharodi et al., 2023). However, bioremediation has unique challenges associated with its implementation, including (1) limitations of PAH metabolizing bacteria, (2) limited PAH bioavailability, and (3) the formation and accumulation of toxic PAH transformation products (Abdel-Shafy and Mansour, 2016; Khan et al., 2015; Kumar et al., 2021).

Microorganisms can transform PAHs through metabolism, where the contaminants are used as carbon and energy sources, or cometabolism, where the contaminants are transformed fortuitously without any carbon or energy gain (Hazen, 2010). Although many microorganisms have been isolated that metabolize PAHs (Ghosal et al., 2016; Imam et al., 2022; Lu et al., 2011), the application of these microorganisms is limited, as they are dependent on PAHs for their survival and are unable to degrade the contaminants to very low concentrations (Hazen, 2010; Khan et al., 2015). Microorganisms that cometabolize contaminants are not limited in this way and are therefore better equipped to degrade PAHs to remediation goal concentrations. However, complete mineralization is difficult to achieve with cometabolism, thus the potential to accumulate toxic PAH transformation products should be considered. Cometabolism can occur without the addition of exogenous substrates by relying on a microorganism’s energy reserves to support activity (Alvarez-Cohen and McCarty, 1991; Frascari et al., 2015; Henrysson and McCarty, 1993). However, contaminant transformation capacity is more limited for resting cells compared to growing cells. Therefore, sustained cometabolism requires the addition of a substrate that provide carbon and energy for the bacteria (Criddle, 1993; Kim et al., 2020).

Whether a process is metabolic or cometabolic, one of the biggest challenges in PAH bioremediation is their bioavailability (Duan et al., 2015). In the environment PAHs sorb strongly to soil and sediment due to their hydrophobicity. Sorbed PAHs are not accessible to microorganisms, therefore treatment is often limited by the desorption rate of PAHs from the solid phase to the aqueous phase, as concentrations are only reduced by microorganisms in the aqueous phase (Lamichhane et al., 2016; Mohan et al., 2006). A common method used to overcome the bioavailability barrier is the use of surfactants to solubilize sorbed PAHs through a process known as soil washing (Cheng et al., 2018; Mulligan et al., 2001). Surfactants aid in the solubilization of PAHs by creating micelles that are suspended in the aqueous phase but have hydrophobic centers capable of accumulating PAHs (Mulligan et al., 2001; Peng et al., 2011).

Surfactants are typically categorized as cationic, anionic, or nonionic. Nonionic surfactants are most commonly used for soil washing, as cationic surfactants are prone to adsorption to soil and anionic surfactants prone to precipitation out of the aqueous phase (Bai et al., 2019; Zhang and Zhu, 2010). The nonionic surfactant Tween ^® 80 (T80) has been studied extensively for soil washing applications, particularly in combination with bioremediation. It has been shown to have minimal toxicity and disruption to its areas of use, including minimal to no impacts to fungal communities, plants, and soil microbiome (Cheng et al., 2017). However, all surfactants, including T80, are toxic to microorganisms at high enough concentrations, as they can disrupt cellular membranes and cause cell lysis (Gharibzadeh et al., 2016). Because of the inherent toxicity of surfactants, protecting cells from surfactants is necessary for extending the lifetime of bacteria used for bioremediation.

One approach used to protect cells from surfactant toxicity is immobilizing cells in hydrogels made from biocompatible polymers such as alginate, polyvinyl alcohol, or chitosan (Mehrotra et al., 2021). The benefits of using immobilized cells in bioremediation applications has been discussed extensively in a variety of reviews, and include advantages such as increasing the activity and viability of biomass, providing opportunities for cell recovery and reuse, and protecting cells from harsh environmental conditions (Bayat et al., 2015; Mehrotra et al., 2021; Partovinia and Rasekh, 2018; Patel et al., 2020). Of studies that have investigated PAH bioremediation using surfactants and immobilized cells, improved PAH treatment with immobilized cells compared to suspended cells has consistently been demonstrated and attributed to increased resistance to surfactant toxicity, improved mass transfer performance, bacterial growth retention in the hydrogel matrix, and sorption of PAHs to hydrogels (Chen et al., 2021; Wen et al., 2021; Xu et al., 2019). However, these studies have focused exclusively on PAH metabolizing bacteria, and have not investigated PAH transformation products or toxicity impacts of these remediation strategies. Several studies have reported increases in toxicity after chemical or biological treatment (Andersson et al., 2009; Chibwe et al., 2015; Schrlau et al., 2017; Wang et al., 2021), due to the formation of transformation products that are more bioavailable and more toxic that their parent compounds. Despite these findings, it is still very uncommon for transformation products or toxicity to be monitored in PAH remediation studies (Titaley et al. 2020).

Rhodococcus rhodochrous ATCC 21198 (21198) was the gram-positive pure bacterial culture used in the present bioremediation study. The genera Rhodococcus is very popular in remediation applications because of its many useful qualities such as its tolerance of harsh environments, metabolic diversity, and expression of enzymes that can transform a wide range of relevant environmental contaminants (Krivoruchko et al., 2019; Nazari et al., 2022). Previous studies with 21198 have demonstrated 21198’s ability to cometabolize a variety of chlorinated solvents and 1,4-dioxane as resting and growing cells (Bealessio et al., 2023; Rasmussen et al., 2020; Rolston et al., 2019). Most recently, it was discovered that 21198 also has the ability to transform monoaromatic hydrocarbons via a combination of metabolism and cometabolism (Huizenga and Semprini, 2023a), motivating further investigation of 21198’s ability to remediate aromatic hydrocarbon contaminants in the present study.

The objectives of this research were to demonstrate the cometabolic treatment of a mixture of four PAHs, each representing a different structural class: fluorene (non-alternant), phenanthrene (bent), anthracene (linear), and pyrene (cluster), by 21198. The cometabolic treatment was studied in batch incubations that included: suspended cells; suspended cell with T80; immobilized cells; and immobilized cells with T80. The impact of T80 and cell immobilization on PAH treatment was evaluated not only with PAH removal, but also with formation of PAH transformation products and post-remediation toxicity, which are often excluded aspects of remediation studies. This study is the first to investigate 21198’s ability to remediate PAHs, and to the authors knowledge, the first study of cometabolic treatment of PAHs with bacteria immobilized in hydrogels.

2.0. Materials and Methods

2.1. Chemicals

A full list of the PAHs and isotope labelled PAHs included in this study is provided in Tables S1 and S2. Compressed gases were purchased from Gas Innovations (La Porte, TX): Isobutane (99.99%) and Airgas Inc. (Radnor, PA): Oxygen (100%). 1-Butanol (99.4%) was purchased from Acros Organics (MA). Polymers were purchased from Sigma Aldrich (St. Louis, MO): polyvinyl alcohol (>99%) with a molecular weight distribution of 85,000-124,000 and Cape Crystal Brands (Summit, NJ): sodium alginate of food grade purity. Crosslinkers were purchased from Honeywell (Charlotte, NC): Boric acid (>99.8%) and Millipore Sigma (Burlington, MA): Calcium chloride dihydrate (>95%). Tween ^® 80 (T80) was purchased from Sigma Aldrich (St. Louis, MO). All solvents used were of analytical grade purity and purchased from various vendors.

2.2. Cell Culturing

The pure culture Rhodococcus rhodochrous ATCC 21198 (21198) was obtained from Dr. Michael Hyman at North Carolina State University. 21198 was maintained as a pure culture on minimal agar plates in a sterile airtight container with 1.5% v/v isobutane supplied as the sole carbon and energy source. Batch growth reactors were prepared as described in (Huizenga and Semprini, 2023a). Briefly, an inoculum scraped from a minimal plate was added to a 500 mL Wheaton glass media bottle containing 300 mL of sterile mineral salt media (MSM) (Kottegoda et al. 2015) amended with 10% v/v isobutane in excess based on the oxygen content of a sealed bottle.

Growth reactors were incubated in the dark at 30 °C on a 150 RPM shaker table. After 5 days of incubation, cells in late exponential growth phase were concentrated via centrifugation as described in Murnane et al. (2021), and biomass content of the concentrate was measured using a total suspended solids (TSS) analysis (Baird et al. 2017). Concentrated cell solutions were stored at 4 °C for no longer than two days before use in experiments.

2.3. Cell Immobilization

21198 cells were immobilized in sodium alginate (NaAlg)/polyvinyl alcohol (PVA) beads using a modified version of the technique described in Harris et al. (2024). Briefly, isobutane grown 21198 cells were added to a 2% w/v NaAlg/3% w/v PVA mixture at a concentration of 0.5 mg cells/mL. This mixture was transferred to 60 mL sterile syringes with 16-gauge blunt tip needles and dropped into a 500 mL crosslinking solution consisting of boric acid (3% w/v) and calcium chloride (1.5% w/v) with a syringe pump to create spherical hydrogel beads. Beads were crosslinked for 1 hour before being rinsed with autoclaved water.

To increase the biomass concentration and activity in the beads, rinsed beads were transferred to growth reactors prepared as described in section 2.2. The beads were present in the reactors (~20 g per reactor) rather than a suspended cell inoculum, and bottles were not placed on the shaker table to help maintain bead structure. This process was repeated without the addition of cells to produce abiotic control beads. One set NaAlg and PVA polymer solutions were prepared and split to produce separate abiotic and active bead batches to maximize consistency between bead batches.

Isobutane consumption was monitored during incubation with gas chromatography-flame ionization detection (GC-FID) method using a Series 6890 Hewlett Packard gas chromatograph equipped with a capillary column (Agilent DB-624 UI 30m x 0.53mm) and a flame ionization detector operating under conditions described in Huizenga and Semprini (2023a). Beads were removed from growth reactors following the five-day incubation period, rinsed with autoclaved deionized (DI) water, and stored in an autoclaved 50 mM phosphate buffer (pH = 7) for approximately 24 hours until use in batch experiments.

2.3.1. T80 Sensitivity and Transformation Tests

The sensitivity of suspended and immobilized 21198 was assessed in batch tests with a range of T80 concentrations. Batch bottles were prepared in 27 mL vials with butyl septa crimp tops. Stock solutions with a range T80 concentrations (0 mg/L to 5000 mg/L) in MSM were prepared, and 10 mL of solution was added to the 27 mL vials. Each concentration was prepared in triplicate. 21198 was added to bottles as suspended cells (0.5 mg cells/bottle) or immobilized cells (~2 mg cells/bottle). Note that beads used in these experiments were not incubated.

After sealing the vials, 0.2 mL of isobutane was added to the headspace with a syringe and subsequently monitored with GC-FID. A second spike of isobutane was added to bottles after the initial spike was depleted to ensure cell activity was maintained. Equilibrium was maintained at 30 °C on a 150 RPM shaker table for the duration of the experiments. Zero-order biomass normalized rates of isobutane utilization were determined and compared to evaluate the impact of T80 on cell activity.

Because T80 is a biodegradable surfactant (Nazari et al., 2022), a set of batch tests were conducted to determine the ability of 21198 to degrade T80. These tests were carried out in 125 mL Wheaton glass media bottles (nominal volume 155 mL) with butyl septa caps. The aqueous phase of batch bottles was 100 mL of 10 times diluted MSM with 1 g/L T80, to which 30 mg of isobutane grown 21198 was added. Triplicate bottles were prepared with or without the presence of acetylene, a known monooxygenase inhibitor for 21198 (Rolston et al., 2019). Batch bottles were placed on a 150 RPM shaker table in a 30 °C room during the two-day incubation period. During this time, 2 mL samples were periodically removed from the bottles with a syringe through the cap septa, centrifuged at 5000 RPM for 2 minutes, and were analyzed for T80 using the fluorescent spectroscopy method described in Mousset et al. (2013).

2.3.2. Oxygen Uptake Tests

To compare the activity of suspended cells and immobilized cells, oxygen uptake experiments were conducted with cells before and after immobilization and bead incubation, with the same cell concentrate immobilized in beads tested as suspended cells. Batch tests were conducted in 27 mL vials with butyl septa crimp tops. Each bottle contained 10 mL MSM with 40 mg/L 1-butanol, and approximately 1 mg of suspended cells or 1 g of beads. Active bottles were prepared in triplicates, while abiotic controls were prepared in duplicates. Headspace oxygen was monitored after addition of cells using a Hewlet Packard 5890 Series Gas Chromatograph equipped with thermal conductivity detector (GC-TCD) and capillary column (Supelco 60/80 Caroboxen 1000) as described in Murnane et al. (2021). Helium flowed at 30 mL per minute as the carrier gas with the oven temperature held constant at 40 °C. Equilibrium was maintained at 30 °C on a 150 RPM shaker table for the duration of the experiments. The zero-order biomass normalized oxygen utilization rate $({\mathrm{k}}_{\mathrm{O}2})$ for suspended cells was compared to that of the immobilized cells to estimate the immobilized biomass in beads using the equation 1:

$$\text{Immobilized biomass}\left(\frac{\text{mg}\text{biomass}}{\text{g beads}}\right)=\frac{{\text{immobilized cells k}}_{{\mathrm{O}}_2}\left(\frac{\text{mg}{\mathrm{O}}_2}{\text{hour}\ast \text{g beads}}\right)}{{\text{suspended cells k}}_{{\mathrm{O}}_2}\left(\frac{\text{mg}{\mathrm{O}}_2}{\text{hour}\ast \text{mg biomass}}\right)}$$ [1]

2.4. Batch Bottle Preparation

PAH bioremediation experiments were carried out in 125 mL Wheaton glass media bottles (nominal volume 155 mL) with butyl septa caps. Experiments were conducted in four sets; set 1 with suspended cells, set 2 with suspended cells and T80, set 3 with immobilized cells, and set 4 with immobilized cells and T80. The aqueous phase of batch bottles was 100 mL of 10 times diluted MSM, with T80 added to the solution for batch sets 2 and 4 to achieve an aqueous concentration of 1 g/L. PAHs were added to batch bottles from concentrated dimethyl sulfoxide (DMSO) stock solutions to reach approximate initial concentrations of 1 mg/L of fluorene and phenanthrene, 0.5 mg/L anthracene, and 0.1 mg/L pyrene (Table S7). Initial concentrations reflect the PAH’s aqueous solubility limits. 1-Butanol, which is utilized by 21198 as a growth substrate, was added as a neat liquid at an initial concentration of 40 mg/L. 1-Butanol was previously demonstrated to be the superior growth substrate to support aromatic hydrocarbon degradation by 21198 compared to its primary substrate (isobutane) due to 1-butanol’s rapid uptake and lack of competition for enzymes degrading aromatic hydrocarbons (Huizenga and Semprini, 2023a). The oxygen demand associated with the 1-butanol utilization did not deplete the available oxygen in sealed bottles.

Batch bottles were placed on a 150 RPM shaker table in a 30 °C room before cell addition to allow dissolution of the PAHs. 21198 was added to active bottles as either 30 mg of suspended cells or 28 g of beads (representing ~30 mg cells) at time zero, after which initial samples were taken. Batch bottles were kept in the dark at 30 °C on a 100 RPM shaker table for the duration of the experiment. Active bottles were prepared in triplicates, while autoclaved cell controls (prepared as active bottles with autoclaved cells or abiotic beads), and abiotic controls, (prepared as active bottles with no cells) were prepared in duplicates.

Batch bottles were opened in a laminar flow hood every 24 hours to refresh oxygen, and an additional dose of 1-butanol was added every three days to active bottles. Batch sets 3 and 4 had a reduced headspace volume due to the bead volume, and therefore received an additional 5 mL of pure oxygen at the start of the experiment and every subsequent 24 hours to match the headspace oxygen content in batch sets 1 and 2. Beads were harvested from batch bottles after the final timepoint samples were taken and tested for oxygen uptake as described in section 2.3.1 to compare bead activity before and after PAH exposure.

2.5. Chemical Analyses

2.5.1. Parent PAH Quantification

Throughout the course of the batch experiments, 0.7 mL liquid samples were taken from the batch bottles using a sterile glass syringe through the cap septa. These samples were transferred to microcentrifuge tubes and spiked with an isotope-labeled PAH surrogate mixture (Table S2) to monitor PAH losses that occurred during sample processing and storage. Samples were extracted using the following liquid-liquid extraction (LLE) method. Equal volumes of ethyl acetate were added to microcentrifuge tubes containing the 0.7 mL liquid sample, vortexed for five minutes, and then centrifuged at 8000 RPM for 2 minutes to separate the ethyl acetate and aqueous phase. The ethyl acetate phase was removed and stored in an amber autosampler vial with a polyethylene terephthalate (PET) lined screw cap. This process was repeated three times for each sample for a final extract volume of ~2 mL. Extracts were then dried under a nitrogen stream to approximately 225 μL before and spiked with 75 μL of acenaphthene-d10 in ethyl acetate as an internal standard for a final sample volume of 300 μL. Samples were then analyzed with GC-MS for parent PAHs using a previously reported method (Huizenga and Semprini, 2023b).

2.5.2. Qualitative PAH and PAH Metabolite Monitoring

Concurrently with GC-MS samples, 2 mL liquid samples were taken from active bottles for fluorescent spectroscopy analysis over the duration of the experiments. Samples were transferred to a four-sided quartz cuvette (Starna Cells) and scanned using a Varian Cary Eclipse fluorescent spectrometer. Samples were scanned in 3D mode through excitation and emission wavelength ranges of 200 to 300 nm and 300 to 450 nm, respectively. The scan speed was 600 nm/min, and excitation and emission slit widths were both 5 nm. Excitation-emission matrices (EEMs) were processed further with parallel factor analysis (PARAFAC) using the drEEM MATLAB toolbox (Murphy et al., 2013) as described in (Huizenga and Semprini, 2023b).

Separate PARAFAC models were generated with EEMs from active batch bottles for each experiment set (n = 21 or 24) as well as with the combined EEMs from all experiments (n = 90). The most appropriate number of components for each model was determined based on the trends of three metrics: percent explained, core consistency, and sum of squared error. All final models were validated with split half analysis (Bro, 1997; Murphy et al., 2013).

2.5.3. PAH Metabolite Analysis

At the end of each batch experiment, the remaining liquid volume in the replicate batch bottles were combined to form three large volume samples: active (A), autoclaved control (C), and abiotic control (X) that ranged from 190 to 235 mL. Samples were stored at −20 ° C until they were extracted using a previously described solid-phase extraction (SPE) method (Schrlau et al., 2017) using Bond Elut Plexa (500 mg, 6 mL) cartridges (Agilent Technologies, New Castle, DE). One method amendment was made for samples from batch sets 3 and 4, as cartridge clogging from free polymer released from the beads was a barrier for successful extraction. Boric acid and calcium chloride were added to samples after thawing to re-crosslink free polymer in solution. After five minutes of mixing on a magnetic stir plate, precipitate was removed from the sample before extraction.

Extracts were split 80/20 for toxicity and chemical analysis. Ultra-high performance liquid chromatography (UPLC – Sciex ExionLC AD) coupled to high-resolution mass spectrometry (HRMS – Sciex ZenoTOF 7600) were used for the identification of PAH transformation products. Separation was achieved with an ACE Excel column (1.7 μm; 2.1x100 mm, Hichrom Limited, UK) and a gradient of nanopure water with 0.1% formic acid and acetonitrile with 0.1% formic acid. Samples (2 μL injection) were analyzed in both positive and negative electrospray ionization (ESI) modes. Full details of the UPLC-HRMS parameters are provided in Tables S3 and S4. Raw data was processed using Sciex OS software version 3.0.

2.6. Embryonic Zebrafish Assay

The extract fractions dedicated to toxicity testing, described in the previous section, were blown to dryness under a nitrogen stream and reconstituted in DMSO. Toxicity was evaluated using the embryonic zebrafish assay described in Truong et al. (2016) at the Sinnhuber Aquatic Research Laboratory (SARL) at Oregon State University (Corvallis, OR, USA). Zebrafish were maintained according to Institutional Animal Care and Use committee protocols.

To summarize, dechorionated tropical 5D wild-type zebrafish (Danio rerio) embryos were statically exposed to extracts at 6 hours post fertilization (hpf) for 120 hours. A 12-point concentration curve was run in duplicate plates for each extract with 7 animals exposed per concentration per plate. Mortality (MO 24) and spontaneous movement (SM 24) were assessed at 24 hpf, while 10 developmental endpoints were assessed at 120 hpf. These endpoints consisted of mortality (MORT), deformation of axis (AXIS), brain (BRN), eye/snout/jaw (CRAN), lower trunk (LTRK), circulation/somite/swim bladder (MUSC), notochord (NC), pigmentation (PIG), heart or yolk sac malformation or edema (EDEM), and unresponsive to touch (TR). Changes in toxicity before and after treatment were assessed by comparing embryo responses of extracts from control batch bottles to extracts from active batch bottles. Extraction blanks and vehicle controls (DMSO only) were included for each extract set. DMSO media concentrations did not exceed the safe threshold of 1% DMSO in EM (Hoyberghs et al., 2021).

2.7. Statistical Analysis

Statistical significance of PAH transformation was determined using one-way ANOVA analysis and Fisher’s least squared difference test assuming equal variance among comparison groups with statistical significance reported for p ≤ 0.05. Statistical analysis was carried out in MATLAB. Statistical significance of toxicity responses in embryonic zebrafish assays was determined according to previously published procedures (Truong et al., 2014)

3.0. Results and Discussion

3.1. T80 Sensitivity and Transformation Tests

Results from the T80 sensitivity tests, shown in Figure 1, indicate that the tolerance of T80 for suspended and immobilized cells is approximately 1 g/L.

Fig. 1.

Isobutane consumption for suspended (A) and immobilized (B) 21198 with various concentrations of T80 (mg/L). Error bars represent standard deviation among triplicates

No difference was observed in isobutane consumption for T80 concentrations of 1 g/L and below compared to controls for both spikes of isobutane over the course of their incubation times, indicating that these concentrations did not cause significant cell distress. This is in good agreement with previously reported T80 tolerances for bacterial strains in the Rhodococcus genera, which typically range from 1 g/L to 2 g/L (Chen et al., 2012; Hu et al., 2020; Lee et al., 2006). Conversely, it was observed that T80 concentrations of 5 g/L and 2.5 g/L significantly reduced isobutane consumption after approximately 8 hours of incubation, and completely halted isobutane consumption after 24 hours of incubation. Isobutane consumption and calculated rates are reported in Table S5. The 1 g/L T80 concentration selected for batch experiments is well above its critical micelle concentration of approximately 15 mg/L (Mahmood and Al-Koofee, 2013; Pinto and Moore, 2000), therefore T80 exists predominantly as micelles at this concentration.

It should be noted that the biomass normalized isobutane consumption rates were significantly different between suspended and immobilized cells for these experiments. Zero-order isobutane consumption rates for suspended and immobilized cells without T80 were 0.15 ± 0.013 and 0.003 ± 0.0007 mg isobutane/mg biomass/hour, respectively. This order of magnitude difference in rates is due to the use of beads directly after they were prepared, rather than being used after an incubation period as was done with beads used in PAH batch experiments. Immobilized cell inactivity has been observed previously and may be due to acid stress responses such as changes to the cell membrane, enzyme expression, or metabolic pathways (Guan and Liu, 2020). This inactivity may also be due to cell death within the beads, as the pH of the crosslinking solution is ~4, well below the optimal pH of 6.5-7 for 21198 (Shields-Menard et al., 2017).

Rapid removal of T80 from the aqueous was observed in T80 transformation tests with suspended cells of 21198. T80 concentrations reached below the method quantification limit (MQL) of 56 mg/L after ~2 days, as shown in Figure 2.

Fig. 2.

Transformation of T80 over time by suspended 21198 cells in solution with (x’s) or without (circles) the presence of a monooxygenase inhibitor (acetylene). Method quantification limit (MQL) is shown in red. Error bars represent standard deviation among triplicates.

Although T80 has been documented consistently as a biodegradable surfactant, the mechanism of biodegradation is not fully known (Li et al., 2017; Nazar et al., 2022; Xu et al., 2019). It is hypothesized that this is a cometabolic process involving an esterase, as has been reported with other bacterial strains (Wang et al., 2011). This hypothesis for 21198 is supported by similar T80 removal in active cells and monooxygenase-inhibited cells and inability to culture 21198 with T80 as a carbon source. Thus, enhancement of PAH solubility from T80 micelles is expected to decline with extended incubation with 21198. Nevertheless, T80 exists as micelles in the system for at least 50 hours, or approximately half of the incubation time. However, T80 transformation was likely not competitive to PAH transformation, as the transformation of these substrates is catalyzed by different enzymes.

3.2. Oxygen Uptake Tests

A comparison of the rate of oxygen uptake (k_O2) for suspended cells compared to immobilized cells, as shown in Figure S1, resulted in an estimated biomass concentration in the beads of 1.1 ± 0.4 mg cells/g bead. Considering the initial immobilized biomass loading of 0.5 mg cells/g bead, it is estimated that bead biomass content was approximately doubled after incubation. No statistically significant difference was observed in zero-order oxygen utilization rate for used beads in either diluted MSM (set 3) or T80 solution (set 4), indicating that PAH exposure and transformation did not significantly impact cell respiration or viability, nor did the incubation with T80. Therefore, significant changes to bead integrity and 21198 activity were not observed after use in the batch experiments.

3.3. Qualitative PAH and PAH Metabolite Monitoring

Treatment progress was assessed in quasi-real time with the fluorescent spectroscopy measurements, with further interpretation was made possible with the PARAFAC modeling results. Preliminary PARAFAC models for individual experiments revealed a high degree of similarity between the components generated to model the data sets. Therefore, it was appropriate to combine data sets and generate one PARAFAC model to allow for comparison of component scores between all experiments. This was expected due to the similarity in experimental set up between the four batch set ups.

Only scans from the first timepoint samples of experiment set three were found to be outliers due to excessively high leverages and were therefore removed from the data set. Fitting metrics for preliminary models with 3 to 7 components are shown in Table S6. The final model using all data sets required four PARAFAC components to appropriately model the data set, with 79% core consistency, 96.5% explained, and 5.7E6 sum of squares errors (SSE). Components were identified by comparing component Stokes shift and spectra to standard scans and literature values, as shown in Figure S2.

Component one (C1) was identified as a mixture of PAH metabolites and cellular material, component two (C2) was identified to be a mixture of anthracene and phenanthrene, component three (C3) was identified to be fluorene, and component four (C4) was identified to be cellular material. Pyrene was not represented in a PARAFAC component likely due to its low initial concentration and obscurant fluorescent compounds present in the bulk sample. The temporal trends of each component are shown in Figure 3.

Fig. 3.

Temporal trends of the four PARAFAC components (C1, C2, C3, and C4) for each experimental set: set 1/suspended cells (A), set 2/suspended cells+T80 (B), set 3/immobilized cells (C), set 4/immobilized cells+T80 (D). Error bars represent standard deviation among triplicates

Trends in component maximum fluorescent intensity (F_max) with time supported their suggested composition. The decrease in F_max of C3 indicates fluorene was transformed in all the treatments. Different rates and degrees of treatment at indicated for anthracene and phenanthrene, based on the C2 results, with the C4 treatment of immobilized cells with T80 being most effective. C1 increased with time for all experiments, which is likely due to the formation of fluorescent PAH metabolites. The presence of C1 in initial timepoint samples may be due to the cellular material present in set 1 and set 2 samples, as the cellular material does fluoresce in the same region as C1. Lower initial amounts of C1 in sets 3 and 4 were expected, as minimal cellular material was present in the aqueous phase due to cell immobilization in the hydrogel. The emergence of C1 with time also informs the nature of these metabolites. Because C1 concentrations increase with time, these are hypothesized to be dead-end products, rather than intermediates in the PAH degradation process.

The location of their spectra in this range of excitation and emission wavelengths suggests that they contain at least two fused benzene rings, as compounds with a single benzene ring do not fluoresce in the region C1 appears (Schwarz and Wasik, 1976). However, this region does match reported spectra for PAH-quinones reported in the literature (Peng et al., 2018). C2 and C3 both decreased with time, representing the transformation of PAHs that occurred over the 5-day incubation. A less apparent decrease in C2 was observed in set 1 and set 3, possible due to spectral interferences from the cellular material or production of PAH metabolites that are not represented by C1. Finally, C4 remained relatively constant throughout the experiments, which is to be expected, as it arises from sample turbidity. Furthermore, its minor presence in set 3 and set 4 support its correlation to suspended cellular material in the aqueous samples, as this was very minor in experiments that used immobilized cells.

3.4. Parent PAH Quantification Results

While EEM-PARAFAC allowed for rapid, qualitative information regarding the PAH concentration changes with time, GC-MS results fully elucidate the quantitative changes in PAH concentrations throughout the experiments. The reduction in PAH concentration for active batch bottles in each experiment set are shown in Figure 4. The equivalent data for controls in each experiment are shown in Figure S3.

Fig. 4.

Normalized concentrations of PAHs over time for experiment set 1/suspended cells (A), set 2/suspended cells+T80 (B), set 3/immobilized cells (C), set 4/immobilized cells+T80 (D). Note that concentrations were normalized to their initial concentrations, reported in Table S6. Error bars represent standard deviations among triplicates.

A combination of abiotic and biotic losses attributed to the removal of PAHs shown in Figure 2. The two significant sources of loss considered in this experimental set up were abiotic losses via adsorption of PAHs to glass and hydrogel beads, and biotic losses of PAHs via cometabolism by 21198. Adsorption of PAHs to the hydrogels is suggested by the trends in PAH concentrations for batch set 3 (Figure 2C), as there was a rapid decrease in PAH concentration in the first 12 hours of incubation, followed by slower decreases for the remaining ~100 hours. This same observation can be seen in the controls illustrated in Figure S3. Therefore, the hydrogel beads may be an additional sink for PAHs via adsorption. However, adsorption of the PAHs to beads in set 4 was not observed in active or control bottles, presumably due to the increased solubility of PAHs with the presence of T80.

3.4.1. Comparison of Quantitative and Qualitative PAH Measurements

A high degree of similarity was observed between GC-MS fluorene results and C3 PARAFAC results as shown in Figure S4, indicating that the PARAFAC model was able to isolate trends in fluorene concentrations accurately throughout the experiments. The same comparison for anthracene and phenanthrene GC-MS results and C2 PARAFAC results were not as straightforward as fluorene, as they were not modeled as separate components. This may be due to the rates of anthracene and phenanthrene degradation being more similar in sets 2 and 4 compared to sets 1 and 3. The more similar the trends are for anthracene and phenanthrene, the better the PARAFAC model can use one component to model them together. Phenanthrene and anthracene share very similar spectra, thus it is expected that they be grouped together, however as their concentrations more greatly differ as they are degraded at different rates, the less their individual trends will resemble the combined trend modeled by C2. It is common for PARAFAC components to be the combined spectra of multiple compounds, particularly as mixture complexity increases (Hung et al., 2022; Seopela et al., 2021). Consequently, EEM-PARAFAC analysis provides limited quantitative data but rapid qualitative data, therefore it is best used in combination with a quantitative targeted analysis such as GC-MS.

3.4.2. Total and Biotic PAH Losses

While the transformation of the PAHs by 21198 alone is a novel finding, the impact of immobilization and presence of T80 were of interest in this study as well. To further compare the different treatments in the experiment sets, the total losses and biotic losses (losses in active bottles minus losses in autoclaved cell control bottles) were calculated and are presented in Figure 5. The impact of T80 in suspended and immobilized cells was investigated by comparing set 1 to 2 and 3 to 4, respectively, while the impact of immobilization in batch tests without and with T80 was investigated by comparing set 1 to 3 and 2 to 4, respectively.

Fig. 5.

Total losses (solid + stripe filled filled) and abiotic losses (stripe filled) for each experiment set organized by PAH. Error bars represent one standard deviation. Values are provided in Table S10 while p-values are reported in Table S8.

The presence of T80 did not appear to impact total fluorene, phenanthrene and anthracene losses, except for suspended cells transforming fluorene, which decreased from 90.9% to 70.5% from set 1 to set 2. Total pyrene losses increased in the presence of T80 for suspended and immobilized cells (6.5% to 22% and 52.7% to 78.3%, respectively).

In comparing PAH total losses for suspended and immobilized cells, total phenanthrene losses increased from 43.5-49.5% to 82.7-89.3% with immobilized cells. Similarly, total pyrene losses increased from 6.5-22% to 52.7-78.3% with immobilized cells. Higher fluorene losses were also observed with immobilized cells compared to suspended cells in the presence of T80 (70.5% to 91.5%). This suggests that immobilized cells transformed more of the total PAH mixture than the suspended cells. This could be due to multiple factors, including greater cell biomass, higher cometabolic activity, or more interactions between PAHs and cells facilitated by the hydrogel matrix, as discussed in similar studies that have made the same observations (Wang 2019, Wen 2021, Xu 2019, Chen 2021).

When accounting for abiotic losses in controls for each set, trends differ slightly from trends of total PAH losses as shown in the solid filled bars in Figure 5. For immobilized cells, the difference between total and biotic PAH losses is largest for batch set 3, in which abiotic losses due to sorption were the greatest. Thus, biotic losses (biotransformation) for all PAHs in batch set 3 were significantly lower than those for batch set 4, which also used beads but had T80 present. Similar PAH losses with control beads were reported in studies summarized in Table 1 and may be due to a combination of sorption to beads, sorption to glassware, volatilization, or reactions with sunlight. However, it cannot be determined with the existing measurements how much each contributed to the total losses, as it is possible that PAHs which initially sorbed to the hydrogel matrix were then desorbed and transformed by 21198. Therefore, these observations are for apparent biotic PAH losses (biotransformation), as actual biotic losses are underestimated by assuming all abiotic losses were irreversible (e.g., sorbed PAHs could not desorb).

Table 1.

Comparison of PAH removal and control bead PAH loss for studies of PAH biodegradation with immobilized bacteria. Note that Fla is fluoranthene, while Flu is fluorene.

Bead formula	Appx. Biomass Loading	Surfactant	Time incubated (days)	PAHs studied	Bead PAH removal (%)	Control Bead PAH Removal (%)	Ref.
3% Ca-Alg	1/40 (w/v)	N/A	42	Phe Fla	63.16 56.94	17.79 13.40	Wang et al., 2019
10% PVA 0.8% Na-Alg 0.7% ALNPs	1/5 (v/v)	2.5 g/L T80	30	Phe Fla Pyr	92.21 87.97 88.81	34.76 30.53 26.21	Wen et al. 2021
10% PVA 0.8% Na-Alg 0.7% ALNPs	1/10 (v/v)	2.5 g/L T80	30	Phe Fla Pyr	89.14 81.25 78.33	37.96 33.44 29.21	Xu et al. 2019
3% PVA 1% Na-Alg	1/4 (v/v)	4 g/L Triton X	4	Phe Fla Pyr	42.06 66.84 68.70	−7.31 29.59 22.92	Chen et al. 2021 *
3% PVA 2% Na-Alg	1/1000 (w/w)	1 g/L T80	5	Phe Flu Pyr Ant	89.26 91.41 78.19 97.11	15.53 18.97 9.76 28.92	Present Study

^*16 priority PAHs were degraded simultaneously, the 3 PAHs shown were reported to be most significantly degraded

The lack of methods to quantify live immobilized biomass is a significant limitation of this study, thus it is possible that more cells were present in the beads than was discernable from the oxygen uptake tests used to compare suspended and immobilized cells (Figure S1). This also makes it difficult to compare PAH removal efficiencies and rates with similar studies, as other studies have the same limitation. Nevertheless, biotic and abiotic PAHs observed in this study are similar to those reported with other microorganisms and bead formulations.

Several studies have immobilized PAH-metabolizing bacteria at lower biomass loadings in beads than the present study for use in PAH treatment studies, results from which are summarized in Table 1. Comparable PAH removal to lower biomass beads incubated for an extended time was achieved by the beads in this study, despite the processes in the present study being exclusively cometabolic. Furthermore, the PAH transformation in the present study is the work of one microorganism, rather than a bacterial consortium as in Wen et al. (2019) and Wang et al. (2019).

3.4.3. PAH Transformation Rates

The impact of immobilization and T80 presence was also investigated in terms of PAH transformation rates, as both rate and extent of treatment are important parameters of remediation technologies. PAH transformation followed pseudo-first order kinetics well, as typical for cometabolic processes (Geng et al., 2022). Pseudo-first order rates derived from both total and apparent biotic losses are shown in Figure 6. It should be noted that rates derived from total losses significantly differed from the rates derived from apparent biotic losses for set 3, where significant PAH losses were observed in abiotic bead controls. This was presumably due to adsorption of PAHs to the bead polymer matrix, however as noted in section 3.4.1, the estimates for biotic PAH losses are likely underestimates.

Fig. 6.

Pseudo-first order transformation rate of PAHs in active bottles derived from total PAH losses (A) and PAH losses normalized to abiotic losses (B). Error bars represent compounded standard deviation for active bars (n=3). Values are reported in Table S10, p-values are reported in Table S8, Fit Coefficients are reported in Table S9

Transformation rates for the four PAHs all ranged within one order of magnitude, indicating that PAH structure did not significantly impact degradation rate, as has been reported previously (Wammer and Peters, 2005; Wang et al., 2019; Wen et al., 2021). The presence of T80 decreased transformation rates of fluorene and anthracene by a factor of ~1.8 and ~1.3 for suspended and immobilized cells, respectively. Contrarily, the presence of T80 increased the phenanthrene transformation rate by a factor of ~1.7 for immobilized cells. It is possible that the decrease in transformation rate for the LMW PAHs when T80 is present may be due to the increase in bioavailability and transformation of pyrene. The decrease in LMW PAH transformation rates may also be due to cell stress responses from the surfactant (Cheng et al., 2018, 2017). Pyrene transformation rates were only derivable when surfactant was present, suggesting that surfactant was required to support pyrene transformation. The rate of pyrene degradation increased by a factor of ~2.2 from set 2 to set 4. It was expected that the presence of surfactant would impact pyrene degradation the most, as it is the least soluble and therefore least accessible to the bacteria. Thus, its bioavailability is most improved by the presence of surfactant.

Independent of T80 presence, transformation rates of fluorene and phenanthrene were faster for immobilized cells compared to suspended cells. The rate of anthracene transformation was significantly faster for immobilized cells only in the presence of T80. The same factors discussed previously that may have contributed to the increase in total PAH losses with immobilized cells are likely to contribute to this observation as well.

3.5. PAH Metabolite Investigation Results

The investigation of PAH metabolites produced from the bioremediation batch set 1 revealed a complex metabolite mixture consisting of several different compound classes. 21 unique compounds were identified with level 2 confidence based on Schymanski et al. (2015), meaning that these compounds matched the MS/MS2 and retention time of a reference standard. A table of these compounds is provided in Table S11. However, because not all isomers of each compound identified were analyzed, it is not confirmed that the specific isomers reported in Table S11 are the exact isomer(s) present in the mixture. The 21 unique compounds were categorized into the following five categories: hydroxylated, ring fission, quinone, and hydroxy/quinone PAH transformation products. Multiple positions on each parent PAH are susceptible to enzymatic transformation, thus transformation pathways are presented for a generic PAH structure in Figure 7.

Fig. 7.

Proposed generic pathway for PAH transformation by 21198. Compounds in brackets represent undetected intermediates, while others represent compound classes identified in treated samples with high confidence: I – hydroxylated, II – dihydroxylated, III – ring fission, IV – quinone, and V – hydroxy-/dihydroxy-quinone.

The compound classes identified in remediated samples indicate that cometabolism was primarily initiated by a monooxygenase. This was an expected finding, as 21198 can express many monooxygenases that have broad substrate specificity, particularly the short-chain alkane monooxygenase (SCAM) highly expressed when grown on isobutane (Bealessio et al., 2023; Murnane et al., 2021). Transformation of PAHs without the presence of isobutane indicates that significant expression of SCAM, or other enzymes involved in PAH transformation, remained in isobutane grown cells even in the absence of isobutane in the remediation batch tests. This agrees with observations in previous studies with 21198 cometabolizing chlorinated solvents, 1,4-dioxane, and aromatic hydrocarbons (Bealessio et al., 2023; Huizenga and Semprini, 2023a).

Hydroxylated PAHs are often reported in PAH biodegradation studies with Rhodococcus strains (Finkelstein et al., 2003; Leneva et al., 2009; Luo et al., 2016, Ma et al. 2023).

Di-hydroxylated products with alcohol groups in the non-ortho position is likely the results of two separate monooxygenase attacks followed by abiotic hydrolysis of the epoxide, while products with alcohol groups in the ortho position are likely the results of enzymatic hydrolysis of the epoxide intermediate, producing a trans- stereoisomer intermediate (Xiao and Li, 2022). Di-hydroxylated products in the ortho position are often attributed to a naphthalene dioxygenase, however this would produce a cis- stereoisomer intermediate (Sakshi and Haritash, 2020). Because a naphthalene dioxygenase has not been annotated in 21198’s genome (Shields-Menard et al., 2014), these intermediates were not considered for this system.

Another class of compounds found in 21198’s PAH metabolite mixtures was ring fission products. Diphenic acid and benzocoumarin, among others, were identified in metabolite mixtures as ring fission products of phenanthrene and anthracene, suggesting the activity of intradiol-ring cleaving dioxygenase, which is present in 21198’s genome. Ring-fission products have been reported for other Rhodococcus strains as well (Finkelstein et al., 2003; Jin et al., 2015; Leneva et al., 2009; Liu et al., 2021; Ma et al., 2023). Diphenic acid has been proposed as an intermediate in phenanthrene metabolism, and a precursor to pyrocatechol for the TCA cycle (Lu et al., 2024). Single ring products were not identified in 21198’s samples, however the method used to identify PAH transformation products was not optimized for compounds of this size.

Significant biotic losses of pyrene were observed in sets 2 and 4, thus pyrene can be degraded by 21198, however, solubility limitations may have prevented significant transformation in sets 1 and 3. Thus, no pyrene metabolites were detected with high confidence in samples from batch set 1. Pyrene degradation by Rhodococcus strains has been reported as a significant member of a microbial consortia (Elyamine et al., 2021), however it is less commonly reported to be degraded by a single pure culture. Liu et al. (2021) reported degradation of pyrene by a Rhodococcus ruber strain, which degraded pyrene primarily with a catechol-1,2-dioxygenase, which has been annotated in 21198’s genome (Shields-Menard et al., 2014). However, many other Rhodococcus strains have been reported to degrade pyrene via a naphthalene dioxygenase, which 21198 does not possess (Jin et al., 2015; Sakshi et al., 2021; Subashchandrabose et al., 2019; Walter et al., 1991). Further investigation of enzyme expression and targeted analytical techniques would help elucidate the PAH transformation pathways and metabolites formed by 21198.

3.6. Embryonic Zebrafish Assay

Although parent PAH concentrations were significantly reduced in all remediation studies, post-remediation toxicity increased for all batch sets, as shown in Figure 8. Several endpoints (SM24, BRN, MUSC, NC, PIG, and TR) were not observed in exposed embryos, and are therefore not shown in Figure 6. Lowest effect levels (LELs) are reported as the amount of treated material (mL) embryos are exposed to in the microwell media (μL).

Fig. 8.

Estimated lowest effect level (LEL) based on exposure to the volume of the sample represented in the microwells. Extracts are grouped by sample type; active (A), autoclaved control (C), abiotic control (X), media extraction blanks (M), T80 control (T), and not remediated (NR). Only observed endpoints are reported in figure: mortality at 24 hpf (MO24) mortality at 120 hpf (MORT), deformation of axis (AXIS), eye/snout/jaw (CRAN), lower trunk (LTRK), and heart or yolk sac malformation and edema (EDEM). Additional columns report any effect except mortality, and any effect.

The initial PAH mixture was not toxic to zebrafish, as observed in unremediated controls (NR) from the present study. This is in agreement with individual PAH toxicity assessments conducted previously, where only pyrene was observed to be toxic to embryonic zebrafish with a LEL (50 μM) that exceeds pyrene concentrations in the present study (Geier et al., 2018). After 94 to 115 hours of treatment of the PAH mixture, mortality and morphology impacts were observed in zebrafish embryos, indicating that mutagenic metabolites accumulated during remediation. Edema and cranial and axis deformation were observed exclusively in treated samples, suggesting that PAH metabolites are responsible for eliciting these responses. Mortality assessed at 120 hpf was also significant in three of the four remediated extracts. Extract 3A was the least concentrated and only elicited toxic responses at the highest concentration tests, therefore it is likely that the tested concentration was not high enough to cause mortality.

Mortality was observed in all abiotic controls that contained T80 (2X,4X, 2T, 4T), thus T80 was likely extracted by the SPE protocol and caused a toxic response in the zebrafish. This agrees with previous work that reported mortality of zebrafish embryos after exposure to polysorbate 20 and 80 (Ginzburg et al., 2018). Interestingly, this was not observed in 2C which did contain T80. This may be due to sorption of the T80 to the cell material that lowered their extraction efficiency. Similar results from active extracts with T80 present (2A and 4A) may indicate that SPE was impacted by the presence of cells, however because T80 was degraded during these experiments (Fig. 2), T80 concentrations in active bottles were likely significantly lower than that in control bottles at the end of the experiments. Controls from experiment 1 (1C and 1X) were contaminated with cells and thus contained a small amount of PAH metabolites, supported by GC-MS data shown in Figure S3, therefore toxicity observed in zebrafish embryos exposed to these extracts do not accurately represent the toxicity of untreated samples. No toxicity was observed in media extraction blanks or solvent blanks.

Although the full content of the treated extracts is not known, 15 of the 21 compounds that have been identified were assessed previously for toxicity using embryonic zebrafish assays, as reported in Table S11 (Geier et al., 2018; Knecht et al., 2013). From these comparisons, we see good agreement between the morphology endpoints observed in our study (MORT, CRAN, AXIS, EDEM) to endpoints observed with transformation product compound classes that are present in remediated samples. Several of these compound classes, including anthraquinones, phenanthrene quinones, and hydroxy-fluorenones have been found to be acutely toxic, with concentrations as low as 1.5 μM eliciting toxic responses in zebrafish embryos (Knecht et al., 2013). Therefore, it is feasible that the minor PAH transformation observed in contaminated controls from batch set 1 resulted in toxic responses. Nevertheless, the consistent trends in toxicity before and after treatment with 21198 indicates that formation of transformation products is an important factor in assessing overall remediation success and warrants further efforts to resolve the complex mixture of PAH transformation products produced by 21198.

4.0. Conclusions

The present study demonstrated the ability of 21198 to transform a mixture of PAHs in a variety of conditions that combined T80 and cell immobilization. PAH transformation was monitored qualitatively with fluorescence spectroscopy, and quantitatively with GC-MS, with good agreement between the methods. Total PAH removal ranged from 70-95% for fluorene, 44-89% for phenanthrene, 86-97% for anthracene, and 6.5-78% for pyrene. The rate and extent of PAH removal was maximized with immobilized cells in the presence of T80, despite T80’s rapid degradation by 21198. Preliminary identification of PAH metabolites produced by 21198 revealed a complex mixture of primarily oxygenated PAH metabolites. These metabolites were found to be more toxic than their parent compounds, based on mortality and developmental effects observed in zebrafish embryos exposed to remediated material. The importance of including metrics other than PAH removal, such as formation of PAH metabolites and post-remediation toxicity, is highlighted by this study. This first demonstration of 21198’s ability to rapidly transform PAHs of various molecular structures and sizes suggests that 21198 can be a valuable bacterium in catalyzing the transformation of PAHs. However, more work is needed to elucidate 21198’s PAH transformation pathways, including enzymes involved and intermediates produced. Future work should capitalize on the enhanced activity of 21198 immobilized in hydrogel beads to explore reuse of hydrogel beads, and continuous flow systems. Implementing further downstream processes to treat the PAH transformation products produced during initial treatment with 21198 should be pursued to help lower post-treatment toxicity in future studies.

Environmental Implication.

Polycyclic aromatic hydrocarbons (PAHs) continue to be ubiquitous environmental contaminants that are known to be hazardous to ecological and human health. Because of their toxicity and abundance, remediation of PAHs from contaminated materials is of great interest. This work explores bioremediation of PAHs via cometabolism with a bacterial pure culture, in both suspended and immobilized form, that was found to be very effective at treating a mixture of four PAHs. However, this work also demonstrated an increase in toxicity of remediated material, cautioning that exclusively monitoring PAH concentrations does not fully capture the effects of a remediation strategy.