1 University of Tulsa, Tulsa, OK, ² University of Oklahoma, Norman, OK, ³ Oklahoma State University, Stillwater, OK. Current affiliation: ^* University of Oklahoma, Norman, OK, ^** University of Toronto, Toronto, Ontario, Canada, ^*** Geo-Microbial Technologies, Ochelata, OK

Introduction

By definition, the bioavailability, or availability of chemicals in soil to ecological receptors, can only be determined by measuring the uptake and/or metabolism of chemicals by organisms. In soils, bioavailability is usually described in terms of chemical uptake by soil-dwelling macroscopic organisms or the ability of microbes to metabolize chemicals. Only a small fraction of total organic chemicals in soil is actually bioavailable, and this will vary with soil composition. Total chemical levels in a soil are usually determined following vigorous extraction with organic solvents or supercritical fluid, often grossly overestimating the bioavailable fraction of chemicals. The estimation of the bioavailable fraction of contaminants in soils is essential for the development of soil quality guidelines and to focus soil remediation efforts.

Soil bioassays provide an indirect estimate of bioavailability and may be confounded by other soil physical/chemical characteristics (e.g., particle size distribution). However, if the bioavailable fraction of chemical could be related to toxicological responses then it would be possible to pattern and develop chemical methods to mimic bioavailability.

One method of estimating bioavailability is measuring residues of contaminants in soil organisms. Another approach is to model bioavailability using an organism surrogate, or biomimetic model, such as a passive sampling device (PSD). PSDs mimic the way organisms bind organic molecules by providing lipophilic matrices which accumulate lipophilic organic contaminants from the external medium (Huckins et al. 1990). PSDs have a number of advantages over live organisms including ease of deployment, low production and maintenance costs, transportability, and applications to a wide variety of soils. Once lipophilic organics accumulate in PSDs and they are analytically identified and quantified, it is necessary to interpret these residues by calibration to residue levels and biological responses in soil organisms, and to microbial activity. Chemical residues and kinetics in earthworms (ie. bioavailability) related to biological responses such as growth and survival are termed critical body residues (CBRs) (Fitzgerald et al. 1996; Lanno and McCarty 1997) and can be compared to residues and kinetics of chemicals in PSDs in order to make bioavailability comparisons. Bioavailability, as estimated by PSDs, would also be useful in assessing the bioremediation potential of soil and as a guide for corrective measures. The rate of bioremediation of crude oil typically decreases rapidly after an initial period of several weeks of intense activity, even though microbes with the requisite enzymatic capability may be present, and compounds may still be chemically extracted from the soil. It has been assumed that these compounds are no longer bioavailable, although this assumption has been difficult to test due to lack of methods to assess bioavailability. PSDs may provide the necessary methods.

The two most promising PSD technologies in assessing the bioavailability of chemicals in soil are semi-permeable membrane devices (SPMDs) and solid phase microextraction (SPME) fibers (Supelco Inc., Bellefonte, PA). SPMDs are membranes composed of low density polyethylene (LDPE) layflat tubing filled with a known weight of neutral lipid (triolein) and have been shown to effectively mimic the function of bipolar lipid membranes in uptake of lipophilic contaminants (Huckins et al. 1990). SPMDs have been used effectively to assess toxicant bioavailability in water and sediment systems (Huckins et al. 1990, 1993), but their application to soil systems remains to be examined in detail. Solid phase microextraction (SPME) fibers represent an even simpler PSD technology that enables the sampling of volatile and non-volatile hydrophobic organics with the added benefit of no extraction and concentration procedures prior to GC analysis (Verbruggen 1999; Parkerton 1998).

The objective of this research is to examine the potential of various biomimetic surrogates in estimating the bioavailability of petroleum hydrocarbons from soil and validate this using biological responses. Bioavailability, as estimated by PSDs, will be compared to microbial activity and the number of microorganisms capable of degrading specific chemical residues accumulated in PSDs in order to more accurately determine the factors limiting biodegradation. The importance of this approach lies in its immediate applications to assessing bioremediation potential and in ecological risk assessment by providing a quantitative measure of the fraction of total chemical that is bioavailable from soil. The ultimate objective of this research is to be able to use PSDs as a rapid, inexpensive, screening tool for estimating the bioavailability of nonpolar contaminants to determine if toxicity testing or site cleanup is actually necessary. If a biomimetic surrogate for estimating the bioavailability of chemicals in soils is calibrated to biological responses, this would result in reduced reliance on toxicity testing for many chemicals.

Materials and Methods

Site selection

Field sites selected for this study were from the Nature Conservancy Tallgrass Prairie Preserve, Pawhuska, OK. Although this area is a nature preserve, there are still many operating oil wells. A recent pipeline break (January 1999) occurred in an area which is composed of two lobes (North and South; Fig. 1). Preliminary TPH measurements on the recent spill site (unpublished data from K. Sublette, U. of Tulsa) suggest that the North lobe is more heavily contaminated (mean approx. 31,000 mg/kg) than the South lobe (mean approx. 4,400 mg/kg), providing a gradient of recent contamination at this site. Two different types of reference sites were also used in the tests. SPMDs were deployed at three reference sites upslope of the north lobe of the oil spill in order to establish baseline tall grass prairie conditions and the variability of these baseline conditions. A tilled, untreated area that serves as a control for the tilling of the oil-contaminated areas served as a soil treatment control. Prairie seed hay was added to the tilled prairie and to the North Lobe. Half of each lobe of contamination was treated with nutrients (100 C:1 N: 0.3 P:0.3K, N as ammonium nitrate, P as superphosphate, K as potassium oxide) to examine the effect of amendments on the bioavailability and degradation of hydrocarbons.

Application of PSDs

Four SPMDs were deployed in each of the prairie and tilled reference areas and in the two sections of the north lobe of the spill. Three SPMDs were placed in each half of the south lobe of contamination, for a total of 22 SPMDs. The SPMDs were placed at a depth of approximately 15 cm and covered with the soil that was removed in the preparation of the hole. Orange flagging tape was attached to each SPMD and the are was marked with four orange flags for ease of location. The SPMDs were deployed under field conditions for three weeks and then sent to Environmental Sampling Technologies (St. Joseph, MO) for dialysis and GPC cleanup. Hydrocarbon content of the dialysate will be analyzed by GC-MS.

Bioavailable petroleum hydrocarbon (BPH), using a solid-phase microextraction fibre, will be determined according to the method of Parkerton and Stone (1998). One g of soil was suspended in 250 mL flask filled with reagent grade water, placed on a stir plate, and sealed with a stopper adapted to accommodate the SPME/holder assembly. Minimal head space is present in the flask, the SPME plunger is depressed and the SPME deployed into the solution and allowed to equilibrate for 16-24 hours as the solution is stirred. The SPME is then removed, rinsed with water to remove any particles, and is deployed directly to a GC injection port for chromatographic quantification.

The TPH level of the soils were analyzed by standard EPA Methods 418.1 (IR) and 8015-B (GC) by Soil Analytical Services, Inc. (SASI), College Station, Texas. Soil samples were shipped by overnight delivery to SASI in completely filled glass jars with Teflonä -lined lids. On two occasions samples were also evaluated in-house (University of Tulsa, Dept. Chemistry, supervised by Dr. William Potter) with PetroflagÔ (Dexil Corp.), which uses a spectrophotometer to measure the turbidity produced by hydrocarbon-surfactant micelles after suspending the soil in a proprietary solution.

Soil samples from the treatment areas and from two uncontaminated, undisturbed areas in May, 1999 for soil chemistry analysis (Janzen, 1993). Ion chromatography (IC) was used to measure the level of chloride, sulfate, nitrate, and phosphate in water extracts from the samples, ICP (inductively coupled plasma spectrometry) was performed on the same samples to determine the values for calcium, magnesium, and sodium (University of Tulsa, Dept. Chemistry, supervised by Dr. William Potter).

Soil sampling for microbial enumeration

Soil was collected for microbial enumeration with a sterile polyethylene centrifuge tube (50 mL size, approximately 5 cm diameter by 18 cm long) pushed into the soil at a depth of 1 cm to 8 cm in depth. Four samples were taken per sampling location, all within a one-foot radius of a SPMD, and collected into a sterile WhirlpacÔ bag for transport back to the laboratory. Soil was stored in the WhirlpacÔ bag at 4^o C for no more than 3 days before being composited by thorough mixing in a sterile glass beaker and a subsample withdrawn for microbial enumeration and for estimation of soil moisture. Soil moisture was determined by gravimetric measurement of two 10 g samples (wet weight) after oven drying. All bacterial counts were expressed as per gram of soil (dry weight).

Aerobic heterotrophic bacteria and hydrocarbon-degraders: A soil suspension of composited soil was made for a series of ten-fold dilutions by removing 5 g of soil from each WhirlpacÔ bag into a sterile 50 mL tube containing 25 mL of sterile isotonic saline (0.85% NaCl, pH 7.0) containing 100 ug/mL cycloheximide (cxy, Sigma Chemical Co., St. Louis, MO.) as an anti-fungal agent, mixed thoroughly by vortexing, then diluted and spread as described below for aerobic heterotrophic bacteria and hydrocarbon-degraders. Dilutions of the soil suspension were spread-plated (three replicates per dilution) on the following types of media: PCA (Plate Count Agar-Difco) with cxy for total aerobic heterotrophs; mineral salts (MS, Duncan et al., 1999) agar with trace metals and cxy containing naphthalene (NAP) added as crystals on the bottom of the plate lid for naphthalene-degraders. All plates were incubated at room temperature; colonies on PCA plates were counted after 48 hrs and again at one week, while colonies on NAP plates were counted after one week and again at two or three weeks, depending on their rate of growth. Individual colonies on NAP plates were transferred with a sterile toothpick from their original plates to two different sets of plates, one containing the same medium as the original plate, the other consisting of MS agar without any added hydrocarbon, in order to confirm the ability of these isolates to utilize hydrocarbons by their abundant growth, and their lack of growth on the latter.

Preliminary results show that SPMDs deployed in the north lobe of the spill site accumulated aromatic hydrocarbons (Fig. 1). Analysis by GC-FID showed the presence of hydrocarbons in the C18-24 range. Further analysis by HPLC showed the presence of a number of PAHs. Levels of two PAHs present in the SPMDs, phenanthrene (PHE) and benzo(a)pyrene (BaP), were quantified by comparison to known standards. The two SPMDs deployed in the north lobe contained mean total PHE and BaP levels of 7.4 and 5.0 g, respectively.

As seen in previous studies of crude-oil contaminated sites in the Tallgrass Prairie (Duncan et al., 1999), the numbers of hydrocarbon-degrading bacteria (here represented by those aerobic heterotrophs capable of growth on naphthalene) were elevated in comparison to samples taken from the tilled, uncontaminated prairie and a nearby aged spill site (SO, Figure 2).

In general, TPH levels decreased over the course of the season, as expected (Figure 3), with the exception of the Oct. 9 PetroflagÔ samples. Soil samples collected shortly before tilling and fertilizer application took place confirm that there was no contamination with brine (Table 1).

SPMDs deployed in the north lobe of the spill site accumulated PAHs from the soil. This results are significant in demonstrating that PSDs, such as SPMDs, can accumulate lipophilic contaminants from soil in under field conditions. Further chemical analysis is necessary to

determine if other hydrocarbons (e.g. aliphatic) were also absorbed by the SPMDs. Under laboratory conditions, SPMDs have been observed to take up greater levels of hydrocarbons, so the conditions that are required for enhanced uptake of hydrophobic contaminants for soil (e.g., moisture content) need to be determined. Work in progress also involves the determination of bioavailable hydrocarbons from contaminated soils using solid-phase microextraction (SPME) techniques and also the tests examining the bioaccumulation and toxicity of hydrocarbons from these soils by earthworms.

The microbial data confirm the SPMD results that hydrocarbons are bioavailable at this time. Interestingly, a sample taken in October suggests that the number of naphthalene-degraders in the tilled site and the aged spill site may be elevated in comparison to the number in undisturbed prairie. Work to be accomplished includes evaluating differences in bacterial species composition at the various sites, using traditional methods of characterization as well as molecular genetic techniques, in order to assess with greater sensitivity and precision the microbial response to hydrocarbons, and to distinguish such a response from that resulting from disturbance (e.g., tilling).

Heterogeneity of the distribution of oil prevents strict comparisons between soils sampled at different times using alternative methods of measuring TPH levels, nevertheless, the discrepancy between the Oct. 9 levels measured by PetroflagÔ , and those expected from continuing the downward trend seen with GC and IR, make us hesitant to recommend the exclusive use of PetroflagÔ for TPH measurements at this time in spite of its low cost and rapidity.

Acknowledgements

Soil chemistry and PetroflagÔ TPH analyses were performed by the following students under the supervision of Dr. William Potter, Dept. Chemistry, University of Tulsa, Tulsa, OK: A. Burn, N. Potter, J. Nguyen, A. Vinh, A. Taylor, N. Carter, Y. Tekleab, J. Middlebrook, C. Waggoner, and T. Sublette.

Literature cited

Table 1: Soil chemistry

* values in ppm

Note: samples were collected before the area was tilled and fertilizer applied. The site designations refer to manipulations that were performed after the soil samples were collected for chemical analysis.

N-t: North Lobe, tilled

N-t, f: North Lobe, tilled and fertilized

S-t: South Lobe, tilled

S-t, f: South Lobe, tilled and fertilized

Prairie 1, Prairie 2: adjacent, uncontaminated control areas

B

C

Figure 1: Polycyclic aromatic hydrocarbons (PAH) detected in

Semi-permeable Membrane Devices (SPMDs) deployed for two weeks in the north lobe of the Tall Grass Prairie spill site during early July, 1999. Chromatograms A and B are replicate SPMDs, while chromatogram C is a certified reference soil containing PAHs. SPMDs deployed in prairie reference areas contained no detectable hydrocarbons.

Figure 2: Numbers of culturable bacteria in the spill sites versus uncontaminated sites. Upper figure: Numbers of aerobic heterotrophic bacteria, from PCA. Lower figure: Numbers of naphthalene-degrading bacteria, from MS + naphthalene. Also indicated are the dates when the spill occurred and tilling and nutrient amendment were performed. "N" refers to samples taken from the North lobe, "S", samples from the South Lobe. "SO" is a nearby site contaminated by a crude oil spill in 1991.