INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Approximately 1,500 former manufactured gas plant (MGP) sites in the United States are estimated to exhibit contaminated soil and groundwater due to coal and oil gasification and liquefaction operations (9). Similarly, at least 700 identified sites in the United States are contaminated with creosote as a result of improper handling and disposal of materials during wood-preserving activities (18). Both of these classes of sites are considered to pose a significant potential health risk for humans and wildlife, since the wastes generated in these processes (primarily coal tars and related substances) contain numerous toxic, carcinogenic, and/or mutagenic compounds.

The most notable class of hazardous compounds found in both coal tar and its derivatives (e.g., creosote) is the polycyclic aromatic hydrocarbons, or PAHs (19, 20), which consist of two or more benzene rings fused into a single aromatic structure. Mammalian liver enzymes (cytochromes P-450 and epoxide hydrolase) oxidize certain PAHs to fjord- and bay-region diol-epoxides (2, 11, 17, 29, 30); these moieties form covalent adducts with DNA (17, 28). Therefore, many PAHs are genotoxic and/or carcinogenic (1, 2, 8, 14, 23) and promote similar effects of other compounds (6). Thus, a total of 16 PAHs have been included on the U.S. Environmental Protection Agency (EPA)'s priority pollutant list (13).

Bioremediation has long been proposed as a treatment technology for the decontamination of PAH-contaminated soils. Numerous bacteria are known to catabolize various two-, three-, and four-ring PAHs as sole sources of carbon and energy (for a review, see reference 5), thus making them good candidate species for site remediation applications. The efficacy of bioremediation approaches, particularly when applied in situ, depends on overcoming any potential nutrient limitations within the soil system to be remediated. In the case of hydrocarbon-contaminated soils, the limiting nutrient is most frequently either phosphorus or nitrogen or, in some cases, both of these. This can be ameliorated via the subsurface injection of soluble nutrients; however, the resultant very high concentrations of nutrients in the immediate vicinity of such injection wells has been observed to lead to excessive localized microbial growth, with concomitant "biofouling" of the wells (4).

The use of gaseous nutrients (N and P compounds with sufficiently high vapor pressures to allow their conversion to a gas under environmental conditions) has been demonstrated in situ as a means of better distributing nutrients throughout the system in support of soil bioremediation. Triethylphosphate (TEP) and tributylphosphate (TBP), although mildly toxic and corrosive irritants, are nonetheless the safest phosphorus compounds which can readily be gasified (in comparison with, for example, phosgene or the carcinogenic trimethylphosphate). They have thus been utilized as phosphorus sources (4, 21) in a patented process (15, 16). Similarly, gaseous nitrous oxide has been used to supply nitrogen (4, 21). The delivery of gaseous nutrients has been shown to enhance the in situ remediation of chlorinated solvents and volatile organic compounds (4, 21), as well as C₄-C₁₀ alkanes and monoaromatic hydrocarbons (e.g., benzene, toluene, ethylbenzene, and xylene) (24). It has not, however, been documented as a means of enhancing the remediation of PAH-contaminated soils.

This paper presents the results of liquid-culture studies of the abilities of organic phosphates (TEP and TBP) and N₂O to support degradation of PAH by bacteria present in MGP and other petroleum-contaminated site soils. Liquid-culture conditions, although clearly not representative of field conditions, were chosen in order to evaluate microbial performance under conditions of optimal bioavailability. In general, we have found that while removal of some PAHs in some soils does appear to be significantly stimulated through the use of alternative sources of N and P, this effect is not universal. There appears to be considerable site-specific variability based on differences in microbiology, soil chemistry, and/or soil structure, implying that soil treatability evaluations will have to be conducted on a case-by-case basis. Our results will serve as a starting point for studies on the use of gaseous N and P sources to support PAH bioremediation in soil column microcosms which are more representative of site conditions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Soils. Soil samples were obtained from sites with a history of industrial activities leading to PAH contamination. The MGP soil is a loamy sand (86% sand, 5% clay, 9% silt) from a New Jersey site, whereas the oil field soil is a crude oil-contaminated sandy loam (63% sand, 3% clay, 34% silt) obtained from the vicinity of a wellhead in a drilling field in southern Illinois. PAH concentrations for the two soils are given in Table 1. The oil field soil, although very high (ca. 16% [data not shown]) in total petroleum hydrocarbons, actually contained relatively modest levels of PAH, as can be seen in Table 1. Each of these soils was air dried (to ca. 3% moisture) and homogenized immediately prior to use.

Culture conditions. Homogenized soil samples (500 mg) were mixed with 50 ml of sterile media (0.1 ml of Wolfe's vitamins [2-mg · liter¹ biotin, 2-mg · liter¹ folic acid, 10-mg · liter¹ pyridoxine HCl, 5-mg · liter¹ thiamine HCl, 5-mg · liter¹ riboflavin, 5-mg · liter¹ nicotinic acid, 5-mg · liter¹ pantothenic acid, 0.1-mg · liter¹ cyanocobalamine, 5-mg · liter¹ p-aminobenzoic acid, 5-mg · liter¹ thioctic acid], 0.1 ml of trace minerals [100-mg · liter¹ ZnSO₄, 300-mg · liter¹ H₃BO₃, 300-mg · liter¹ CoCl, 10-mg · liter¹ CuCl], and 0.8 ml of N- and P-free Winogradsky medium [pH 7.2] [62.5-g · liter¹ MgSO₄ · 7H₂O, 31.25-g · liter¹ NaCl, 1.25-g · liter¹ FeSO₄, 1.25-g · liter¹ MnSO₄] per 100 ml of sterile deionized water) in 125-ml serum bottles. In order to assess the degree of N and P limitation on PAH degradation inherent in each soil, [¹⁴C] phenanthrene mineralization was measured in cultures of each of the three soils which received no supplemental N or P, N only (as NH₄Cl), P only (as KH₂PO₄), or both N and P. Six combinations were then investigated for N and P supplementation: NH₄Cl plus KH₂PO₄; N₂O plus KH₂PO₄; NH₄Cl plus TEP; NH₄Cl plus TBP; N₂O plus TEP; and N₂O plus TBP. Within each condition, duplicate cultures were employed. In all cases, addition of N and P sources was normalized on a molar basis to provide 9.2 mM N and 3.7 mM P. When N₂O was used, it was added by injection to sealed bottles. In order to ensure the presence of at least some bioavailable PAH, all cultures were also supplemented with 50 µl of a PAH-containing extract from a second MGP soil (approx. 12,000 ppm of total PAH; Table 1) dissolved in n,n-dimethylformamide. One set of cultures (duplicates of each condition) was further supplemented with ¹⁴C-phenanthrene for mineralization determinations (see below), while one received no radiolabel and was used to simultaneously measure the extent of disappearance of multiple PAHs. Both sets of cultures were incubated at room temperature (approximately 25°C) with shaking at 170 rpm. Poisoned controls (10 mg of HgCl₂ per culture) were also conducted in duplicate.

Mineralization of ¹⁴C-PAH. CO₂ traps were made by wrapping stainless steel wire around the necks of 12-by-32-mm borosilicate glass autosampler vials and pushing the wire through 20-mm Teflon silicone-lined septa. These assemblies were placed in the serum bottles, which were then crimped with aluminum seals. Syringes were used to inject 1 ml of 0.5 M NaOH into each CO₂ trap. Periodically, the CO₂-containing NaOH solution was withdrawn from the traps, mixed with 5 ml of Ultima Gold high-flashpoint liquid scintillation cocktail (Packard, Meriden, Conn.) and counted in a liquid scintillation counter (Packard model 2200CA Tri-Carb). Fresh NaOH was then added to the CO₂ traps. Cultures containing ¹⁴C-phenanthrene typically received ca. 80,000 to 100,000 dpm of PAH in 20 µl of methanol.

Extraction and high-pressure liquid chromatography (HPLC) analysis. Soil samples were centrifuged (10 min, 5,000 × g) in stainless-steel containers to separate solid and aqueous phases. Soil solids were mixed with anhydrous sodium sulfate (1:1) and ground with a mortar and pestle to form a fine powder. Sonication was performed according to EPA method 3550A (27) using 1:1 hexane-acetone (30 ml) as the solvent and was repeated three times. The extracts were combined and vacuum filtered before evaporation. The aqueous phases of various cultures were extracted threefold with methylene chloride as per EPA method 3510B (27). These extracts were then dried by passage through anhydrous sodium sulfate. Both solid and aqueous extracts were evaporated to dryness under a stream of N₂ in a Turbovap evaporator (Zymark, Hopkinton, Mass.) and exchanged into acetonitrile (ACN) (1 ml). Ten microliters of this solution was analyzed by reverse-phase HPLC (EPA method 8310 [27]) using a Supelcosil LC-PAH column (15 cm by 4.6 mm) and a Waters HPLC system coupled to a diode-array detector (Waters model 996). The following gradient was used, with a flow rate of 1.5 ml · min¹ throughout: 0 min, 60% H₂O-40% ACN; 25 min, 100% ACN (hold for 2 min); 33 min, 60% H₂O-40% ACN. Identities of individual PAHs were verified by comparing the retention times and the absorbance spectra and quantified by comparison with five-point standard curves (all r² values were >0.988).

Chemicals. [9-¹⁴C]phenanthrene (reported purity, 98%) was purchased from Sigma (St. Louis, Mo.). TEP (99% pure) and TBP (98%) were purchased from Aldrich (Milwaukee, Wis.); N₂O (ultra-high purity) was from Matheson Gas Products (Joliet, Ill.). NH₄Cl and KH₂PO₄ were purchased from Mallinckrodt Chemicals (Paris, Ky.), and NaOH and HPLC-grade solvents were purchased from Fisher Chemicals (Fairway, N.J.). Authentic PAH standards for use in HPLC analysis were obtained from Ultra Scientific (Kingstown, R.I.).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Nutrient limitations were assessed for each site soil by determining the degree of phenanthrene mineralization which occurred in the absence of any supplemental N or P; this was then compared to that which was supported by either nutrient singly or when the two were combined. Data for these trials are shown in Fig. 1. Both soils were strongly nutrient limited. The MGP soil showed an especially strong N limitation and a P limitation which was also significant, while the oil field soil was greatly limited by both N and P; in this case, neither nutrient alone was capable of enhancing ¹⁴CO₂ release at all relative to results with unsupplemented conditions.

View larger version (38K): [in this window] [in a new window]   FIG. 1.   Extent of [14C]phenanthrene mineralization (during a total incubation time of 5 weeks) in each soil in the absence of N or P supplementation or when NH4Cl and KH2PO4 were added either singly or in combination.

Cultures of the microbial communities from the two soils were examined for their ability to mineralize phenanthrene under various conditions of N and P supplementation. We also determined, using reverse-phase HPLC, disappearance of a range of three- to six-ring PAHs from cultures under the same set of nutrient supplementation conditions.

In the case of the MGP soil, there was little effect of varying nutrient compositions on mineralization of spiked [¹⁴C] phenanthrene, as can be seen in Fig. 2. All nutrient regimes examined sustained between 50 and 70% conversion to CO₂, with the only significant difference between the different treatments being a slightly longer lag time prior to mineralization in the three conditions in which N₂O served as the nitrogen source. Cultures receiving HgCl₂ (10 mg) were still capable of mineralizing 7.3% of the input phenanthrene.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Mineralization of phenanthrene and pyrene by the microbial community associated with the MGP soil under various conditions of N and P supplementation.

Under conventional nutrient additions (NH₄ and PO₄), the microbial community present in the MGP soil displayed significant removal of all three-ring compounds examined, as well as some elimination of several four- and five-ring PAHs (Table 2). A separate experiment, in which mineralization of ¹⁴C-pyrene was measured (data not shown), indicated that no mineralization of this PAH occurred; thus, it seems possible that the loss of pyrene seen in this soil may be due to cometabolic effects (7), which may also account for the loss of other four- and five-ring compounds. No loss of six-ring PAHs was observed in this soil (data not shown). The data in Table 2 clearly show that substitution of alternative N and P sources does not, in most cases, enhance microbial PAH degradation in this soil. The combination of NH₄ and TBP results in enhanced removal of phenanthrene and anthracene (relative to results with NH₄ and PO₄); however, losses of fluorene and fluoranthene are no greater under these conditions, and four- and five-ring compounds are unaffected. Most other nutrient regimes supported degradation of only the most labile compounds (fluorene, phenanthrene, and anthracene).

In the crude oil-contaminated wellhead soil, as in the MGP soil, mineralization of phenanthrene was greatest in the NH₄Cl-plus-KH₂PO₄ cultures (Fig. 3), again exceeding 80%. In this soil, use of any of the alternative nutrients substantially reduced the mineralization of phenanthrene. Pairwise comparisons of the different conditions showed clear trends, since all three of the NH₄-containing cultures outperformed their N₂O-containing counterparts. Furthermore, the trend in which PO₄ outperformed TEP which, in turn, outperformed TBP was true for both the NH₄ and N₂O sets of cultures. Killed controls evolved 1.7% of the input ¹⁴C as CO₂.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Mineralization of phenanthrene and pyrene by the microbial community associated with the crude oil-contaminated oil field soil under various conditions of N and P supplementation.

The extent of PAH removal from the oil field soil was considerably higher than that from the MGP site soil. Table 3 shows that nearly all compounds examined were significantly removed from this soil (which, as described above, is strongly N and P limited for mineralization of phenanthrene) under the NH₄-plus-PO₄ supplementation; this includes five-ring PAHs. Again, inasmuch as no significant mineralization of pyrene was seen in this case (data not shown), it seems possible that some of the losses of high-molecular-weight (HMW) PAHs are cometabolic in nature. As with the MGP soil, inclusion of an alternative source of either N or P significantly impeded PAH removal for many compounds, although essentially complete removal of some (phenanthrene, anthracene, and fluorene) did still occur under all nutrient regimes. In the case of this soil, it appears that N₂O is more capable of serving as a nitrogen source than either TBP or TEP is as a phosphorus source, since the N₂O-PO₄ combination supports a broader and more extensive removal of HMW species than do any conditions including either of the alkyphosphates.

Among the soils examined in this work, it is clear that the bacteria present in the oil field soil were more adept at removing PAHs, including the HMW species. This is interesting because although the soil in question has very high levels of total petroleum hydrocarbons (approximately 16.2%), the levels of PAHs are actually quite low. As can be seen in Table 1, the initial levels of total PAH in this soil were no higher than approximately 20 ppm. The superior ability of the microbial community in this soil to degrade PAH may therefore be the result of a long period of constant or recurring low-level exposure to these compounds.

It has been stated by other authors that the P contained in organophosphorus molecules, such as TEP, is not available to all microbes (4). In fact, the selective pressure exerted through the addition of TEP was looked upon as an advantage in the case of cometabolic remediation of TCE- and PCE-contaminated groundwaters by type II methanotrophs, which are capable of utilizing TEP as a phosphorus source. Coinjection of TEP and methane into a solvent-contaminated aquifer thus caused up to 1,000-fold increases in the activities of these bacteria, while the total bacterial biomass remained essentially unchanged (4, 21). Literature on the taxonomic distribution and extent of TEP and TBP utilization is limited. A mixed culture of Pseudomonas strains was capable of releasing PO₄ from TEP; this was examined for the removal (via precipitation of HUO₂PO₄) of UO₂²⁺ from uranium-contaminated mine wastewaters (26). Similarly, other Pseudomonas strains (22), as well as strains of Hyphomicrobium (10) and Acinetobacter (25), are known to be able to utilize TBP, TEP, and/or trimethylphosphate as sole sources of P.

Our results, however, imply that the ability to utilize TEP or TBP and N₂O, at least under slurry conditions, is not a universal attribute among PAH-degrading bacteria. The ability of alternative nutrient combinations to actually enhance PAH removal relative to results with NH₄-plus-PO₄-supplemented cultures seems to have been restricted to the MGP soil and was, even in this case, seen only with lower-molecular-weight compounds.

It is possible that some of our results were due to the effects of the solubility of TEP and TBP on their availability to PAH-degrading bacteria. As a consequence of its longer aliphatic groups and resultant greater hydrophobicity, TBP is more resistant to solubilization than TBP; for example, we observed that the former tended to form persistent globules in the culture media, whereas the latter did not. It is, however, difficult to attribute these observations to the effects of solubility alone, since the same adaptations (e.g., lipid-rich outer cell layers, production of biosurfactants) which confer the capability to take up HMW PAHs upon bacteria such as Mycobacterium (3, 12) and Sphingomonas (3) would be expected to have the same effect on TBP. It is therefore possible that the increased removal of some PAHs which was occasionally seen with TBP (Table 2), as well as the high mineralization of phenanthrene supported by both alkylphosphates in the MGP soil (Fig. 2), may reflect the participation of some of these bacteria. These species may actually be favored by the use of more hydrophobic nutrient sources, both because of increased compatibility with their uptake systems and a more favorable distribution within the microcosm. Hydrophobic nutrient sources, such as TBP, might be expected to partition onto soil organic matter, which might be beneficial to organisms such as those listed above, many of which tend to be adherent in nature (3). We have isolated several Sphingomonas strains from these two soils (3a); further examination of these isolates' relative abilities to utilize these alternative nutrient sources may help to address these questions.

The diffusivity of TEP and TBP is approximately 5 orders of magnitude higher in the gaseous phase than when these compounds are dissolved in water (4). Thus distribution would be expected to improve somewhat in vadose soil (4, 21), and the issue of the availability of P to soil surface adherents should be less of a factor than it may have been in these preliminary experiments. The results of the experiments described here indicate that it may not always be feasible to support remediation of PAH-contaminated soils with gaseous sources of N and P and that a great deal of site-specific variability can be expected. It is clear, however, that a true conclusion regarding the applicability of gaseous N and P supplements to in situ soil remediation will require soil column experiments which will better approximate the environmental conditions and behavior of bacteria in the vadose zone. It will also require a better understanding of which members of microbial communities involved in PAH degradation thrive and function under the different regimes and whether or not conditions can be devised to better optimize the performance of the entire community. Experiments in these areas are ongoing in this laboratory.