Environ. Sci. Technol. 2004, 38, 4842-4848
Solid-Phase Microextraction To Predict Bioavailability and Accumulation of Organic Micropollutants in Terrestrial Organisms after Exposure to a Field-Contaminated Soil L E O N V A N D E R W A L , * ,† TJALLING JAGER,‡ ROEL H. L. J. FLEUREN,§ ARJAN BARENDREGT,† THEO L. SINNIGE,† CORNELIS A. M. VAN GESTEL,# AND JOOP L. M. HERMENS† Institute for Risk Assessment Sciences, Utrecht University, P.O. Box 80176, 3508 TD Utrecht, The Netherlands, Department of Theoretical Biology, Faculty of Earth and Life Sciences, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands, National Institute of Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands, and Department of Animal Ecology, Institute of Ecological Science, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands
The risk posed by soil contaminants strongly depends on their bioavailability. In this study, a partition-based sampling method was applied as a tool to estimate bioavailability in soil. The accumulation of organic micropollutants was measured in two earthworm species (Eisenia andrei and Aporrectodea caliginosa) and in 30µm poly(dimethylsiloxane) (PDMS)-coated solid-phase micro extraction (SPME) fibers after exposure to two fieldcontaminated soils. Within 10 days, steady state in earthworms was reached, and within 20 days in the SPME fibers. Steadystate concentrations in both earthworm species were linearly related to concentrations in fibers over a 10 000fold range of concentrations. Measured concentrations in earthworms were compared to levels calculated via equilibrium partitioning theory and total concentrations of contaminants in soil. In addition, freely dissolved concentrations of contaminants in pore water, derived from SPME measurements, were used to calculate concentrations in earthworms. Measured concentrations in earthworms were close to estimated concentrations from the SPME fiber measurements. Freely dissolved concentrations of contaminants in pore water, derived from SPME measurements, were used to calculate bioconcentration factors (BCF) in earthworms. A plot of log BCFs against the octanolwater partition coefficient (log Kow) was linear up to a * Corresponding author phone: +46-90-7869324; fax: +46-90128133; e-mail:
[email protected]. † Utrecht University. ‡ Department of Theoretical Biology, Faculty of Earth and Life Sciences, Vrije Universiteit. § National Institute of Public Health and the Environment. # Department of Animal Ecology, Institute of Ecological Science, Vrije Universiteit. 4842
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 18, 2004
log Kow of 8. These results show that measuring concentrations of hydrophobic chemicals in a PDMS-coated fiber represents a simple tool to estimate internal concentrations of chemicals in biota exposed to soil.
Introduction Contaminated soils are still a major issue of concern in many countries. Several factors, such as soil characteristics (pH, organic matter content, etc.), chemical properties of the contaminants (Kow) and environmental conditions (precipitation, temperature) affect the exposure and potential hazard to biota. The most complicating and uncertain factor in sitespecific risk assessment is lack of information about the actual bioavailability of soil contaminants. Bioavailability is an important factor in risk assessment and depends on the characteristics of soil and contaminants as well as the organism and its trophic position and habitat (1). Bioavailability may also change over time due to aging effects (2, 3). Various chemical methods have been proposed and applied to measure the bioavailable fraction of organic contaminants in soil and sediment, including mild extraction techniques using a variety of (combinations of) solvents (46). Supercritical fluid (7) extractions and extractions with Tenax TA (8-12) have been proposed and applied to quantify desorption rates or the rapidly desorbing fraction from soil and sediment as a measure of bioavailability. Passive samplers, such as C18 disks and semipermeable membrane devices (13-15), have also been used. Negligible-depletion solid-phase micro extraction (ndSPME) was introduced some years ago as a partition-based extraction or sampling technique to measure freely dissolved concentrations in several matrixes (16-18) and to mimic accumulation in biota (19-21). The basic assumption of the approach is that, because only a very small amount is extracted from a solution, the extraction does not influence the existing equilibrium between the bound and free form of a chemical, and the SPME fiber only measures the freely dissolved concentration. This approach has been extended to measure free concentrations in sediment pore water (22) by equilibrating the SPME fiber in a sediment-water system. Because of the small amount of the hydrophobic phase introduced, the extraction does not affect the concentration in the sediment and freely dissolved pore water concentrations (CPWD) can be calculated from:
CPWD ) CSPME/KSPME-W
(1)
In this equation, CSPME is the concentration in the SPME fiber and KSPME-W is the SPME-water partition coefficient. This same approach has been applied to predict bioconcentration in a sediment-dwelling organism (23). The advantage of these negligible-depletion extractions is that the extraction itself does not affect the soil system and in that sense it closely resembles the exposure of soil and sediment living organisms. A focus on freely dissolved concentrations assumes that the uptake route proceeds via the aqueous phase. One can argue about different routes of exposure in soil (i.e., by food, aqueous, or gaseous phase), but uptake into earthworms is well described by pore-water concentrations, irrespective of the dominant route of exposure (24). The objective of this study was to test the SPME method as a tool to measure bioavailability in a fieldcontaminated soil. In a previous study, the feasibility of the approach was shown in a laboratory-contaminated soil (25). 10.1021/es035318g CCC: $27.50
2004 American Chemical Society Published on Web 08/12/2004
TABLE 1. Characteristics (mean ( SD; N ) 2) of the Different Soils Used in This Study clay pH (KCl) % org C a
site 1
site 2
site 3
7.71 ((0.028) 4.37 ((0.14)
nd 7.70 ((0.014) 6.57 ((0.15)
24.3% 7.51 ((0.014) 8.53 ((0.17)
nda
nd ) not determined.
In the present study, accumulation experiments with two species of earthworms were performed in three soils originating from a contaminated site in Rotterdam, The Netherlands. Steady-state concentrations were measured in the earthworms and SPME fibers coated with poly(dimethylsiloxane) (PDMS) were exposed to the same soils until steady state was reached. For one soil, an exposure profile (concentration in fiber against time) was made to select the appropriate equilibration time. The suitability of SPME analysis was tested by comparing the body residues in earthworms to a classical equilibrium partitioning model and pore water concentrations estimated from total soil concentrations and Kow, as well as pore water concentrations derived by SPME. Although the focus of the study is on developing and testing a technique to measure bioavailability and bioaccumulation in soil organisms, the same data set was also used to calculate BCFs based on SPME measured pore water concentrations and to study the relationship between these BCFs and Kow.
Materials and Methods Exposure Conditions. Soil Samples. Soil was obtained from a polder called “de Esch” in Rotterdam (The Netherlands), amended with dredge sludge material in the 1970s. The sludge contained various contaminants, such as polycyclic aromatic hydrocarbons (PAHs), telodrin, dieldrin, and chlorobenzenes, which can now be found in the soil. Three locations from this site were sampled and this soil was used in laboratory experiments. The soil characteristics are shown in Table 1. Exposure of SPME Fibers. Poly(dimethylsiloxane) (PDMS, 30 µm) coated fibers were obtained from Supelco (Bellafonte, CA). Fibers with a length of 10 cm (n ) 3) were exposed to 5 g (d.w.) of Esch soil at maximum water holding capacity (WHC) in a 30-mL glass vial with glass stopper. Five mL of sodium azide solution (10 mM) was added to prevent bacterial growth. The amount of hydrophobic phase on each fiber was 1.3 µL. Fibers were exposed to site 3 soil for 1, 2 (n ) 2), 3 (n ) 2), 4, 5, 6, 7, 8, 15, 16, 22, 23, 29, 30, 42, and 43 days to obtain uptake curves and equilibration times for all compounds. Based on this study, a 20-day exposure time was chosen in the tests with the other two soils. The soils were exposed for 20 days in triplicate with three fibers in each replicate. Shaking during exposure was performed on a “rock and roller” shaker (Snijders, The Netherlands) with a rolling speed of 10 rpm and a rocking frequency of 5 cycles/ min at 35° at a temperature of 20 °C. Steady-state concentrations in the fibers were estimated using a one-compartment model:
Cf(t) ) Cf∞(1 - e-kt)
(2)
The concentration on the fiber at steady state (Cf∞) and the rate constant k are derived from fitting the measured concentrations Cf(t) at several exposure times using Graphpad Prism 3.0 (San Diego, CA). Exposure of Earthworms. Two species of earthworms were selected that differ in their behavior: Eisenia andrei (Ea), an epigeic compost worm inhabiting compost and manure
heaps, and Aporrectodea caliginosa (Ac) a soil-dwelling endogeic species. These worms were cultured at the National Institute for Public Health and the Environment (RIVM, Bilthoven, The Netherlands). Ten earthworms per time point were exposed to 500 g of moist contaminated soil, at 50% of the WHC (WHC50), in glass jars for 1, 3, 5, 7, 10, 14, 21, and 28 days. After exposure, they were put on wet filter paper for 24 h, at 20 °C for E. andrei and 10 °C for A. caliginosa, to allow them to empty their gut, and frozen at -20 °C until extraction. Chemical analysis was performed in three or four samples, with 2 earthworms per sample, resulting in at least triplicate concentration measurements. Detailed information about the exposure of earthworms is given by Jager et al. (26). Chemical Analysis. Chemicals. Test compounds used for identification and quantification were obtained from different suppliers. Hexachlorobenzene (HCB) (Pestanal 98%) and telodrin (99.9%) were purchased from Riedel-de Hae¨n, (Seelze, Germany). 2,2′,3,5′-Tetrachlorobiphenyl (PCB #44), 2,2′,3,5′,6-pentachlorobiphenyl (PCB #95), 2,3,3′,4′,6-pentachlorobiphenyl (PCB #110), 2,3,3′,5,6-pentachlorobiphenyl (PCB #112), 2,2′,3,4,4′,5′-hexachlorobiphenyl (PCB #138), 2,2′,3,4′,5′,6-hexachlorobiphenyl (PCB #149), 2,2′,4,4′,5,5′hexachlorobiphenyl (PCB #153), 2,2′,4,4′,6,6′-hexachlorobiphenyl (PCB #155), 2,2′,3,3′,4,5,6′-heptachlorobiphenyl (PCB #174), 2,2′,3,3′,5,6,6′-heptachlorobiphenyl (PCB #179), and 2,2′,3,4,4′,5,5′-heptachlorobiphenyl (PCB #180) (PCB numbering according to IUPAC) were obtained from Dr. Ehrenstorfer (GmbH, Augsburg, Germany) as solutions in isooctane. PCB #138 (99.6%) was also obtained from Promochem (Wesel, Germany). Extraction and Chemical Analysis of Soil. The soil sample was mixed with sodium sulfate, allowed to dry overnight, and was Soxhlet extracted for 12 h with hexane/acetone (3:1, v/v, 70 °C). The extract was concentrated on a rotary evaporator. The clean up was performed on an aluminum oxide column with isooctane as eluent. Next, the extract was concentrated under nitrogen to 2 mL and eluted over a silica gel column (2% water) with 11 mL isooctane (first fraction) and 10 mL 20% diethyl ether in isooctane (second fraction). The first fraction contains the PCBs and some organic chlorinated pesticides (OCPs), and the second fraction OCPs only. Both fractions were concentrated to 1 mL and an internal standard (PCB #112) was added, and the fractions were analyzed with a gas chromatograph and electron capture detection (GC-ECD) using a 50-m CP Sil 8 column (i.d. 0.15 mm, film thickness 0.3 µm) and a 50-m CP Sil 19 column (i.d. 0.15 mm, film thickness 0.2 µm) (Varian, CA) using H2 as the carrier gas. PCB #174 could not be determined properly in soil due to interference and was therefore not used in our calculations. Quality assurance was performed by including a local reference material (LRM), a blank, a duplicate measurement, and a recovery test within every sequence of samples (n ) 11). The limit of determination was set by the lowest concentration of the multi-level (6 point) calibration curve. Chemical Analysis of Fibers. Following the exposure in soil, fibers were cleaned with a wetted tissue and transferred to glass thermal desorption tubes, 1 µL hexane containing 100 pg internal standard (PCB #155) was added and the tubes with the fibers were thermally desorbed at 275 °C using a Gerstel TDS A thermodesorption autosampler and a Gerstel TDS 2 thermodesorption chamber (Gerstel GmbH & Co, K G Mu ¨lheim, Germany). During desorption, the CIS-3 injector (Gerstel GmbH & Co, K G Mu ¨ lheim, Germany) was cooled with liquid nitrogen and held at -50 °C. Fibers were analyzed using a GC (Carlo Erba 5300 series, Milan, Italy) with the same analytical parameters as used for the earthworm samples (see below). VOL. 38, NO. 18, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4843
TABLE 2. Concentrations of Contaminants in Soil (µg/kg Dry Soil, Single Values) in the Soils Used in This Study concentration in soil (µg/kg dry soil) contaminants
log Kowa
m/z ratiosb
site 1
site 2
site 3
HCB telodrin dieldrin PCB #95 PCB #110 PCB #149 PCB #153 PCB #138 PCB #179 PCB #174 PCB #180
5.73c
282 + 284 + 286 309 + 311 + 313, 373e 261 + 263 + 265 324 + 326 + 328 324 + 326 + 328 358 + 360 + 362 358 + 360 + 362 358 + 360 + 362 392 + 394 + 396 392 + 394 + 396 392 + 394 + 396
0.90 3.6 39 nmf 0.30 0.40 0.50 0.50 0.60 g 0.40
360 4100 38 000 nmf 53 120 54 89 14 g 71
32 64 550 nmf 58 110 64 92 15 g 64
5.2d 5.4c 6.69d 6.84d 7.3d 7.53d 7.68d 7.45d 7.83d 8.06d
a Log K b m/z ratios used for detection and quantification ow values are based on experimental data and calculated values by de Bruijn et al. (45). of contaminants. cMeasured values by de Bruijn et al. (45). d Calculated values according to de Bruijn et al.(45). e m/z ratio used when overload of the MS-detector occurred and conventional m/z ratios could not be used. f nm ) not measured. g Could not be measured.
Desorbed fibers were checked for carryover of contaminants by reinserting and analyzing them a second time. No contaminants were detected after reinsertion. Matrix effects were observed resulting in higher detection limits for the more polluted soil from site 2. However, this did not affect the measurements. For only HCB and PCB #138 in the site 1 soil, and for PCB #110 in the site 2 soil, were the concentrations in fibers too low to be detected. Extraction and Chemical Analysis of Earthworms. Frozen worms were homogenized with 7 times their wet weight of dry Na2SO4, using a pestle and mortar, and extracted for 16 h with a mixture of acetone:hexane (1:3, v/v) using a Soxhlet apparatus. After evaporation of the acetone:hexane mixture, the samples were weighed to determine the lipid content of the earthworms, after which the sample was dissolved in cyclohexane. Cleanup was performed using a column with 2 g deactivated silica (5% H2O) and 0.5 g dry Na2SO4. After pre-elution with 5 mL cyclohexane, the sample was added and eluted using 2 times 5 mL cyclohexane/dichloromethane (1:1, v/v). Sample volumes were reduced to approximately 0.5 mL using a nitrogen flow, transferred to GC vials and frozen (at -20 °C) until analysis. Internal standards were added to the samples prior to Soxhlet extraction (PCB #44) and in the GC vials prior to the analysis (PCB #155) via GC using mass spectrometry (GC-MS). Recoveries of the extraction and cleanup procedure were tested by analyzing earthworms spiked with the test chemicals. Cyclohexane/dichloromethane fractions were kept at -20 °C until analysis using GC. The samples were injected in splitless mode at an injection temperature of 275 °C. The oven temperature program was 80 °C for 1 min, increased by 25 °C/min to 250 °C, then increased by 2 °C/min to 265 °C, and finally increased by 15 °C/min to 310 °C with a hold time of 10 min. The analytical column used was a 30-m × 0.25-mm DB-5 MS (J&W Scientific, Folsom, CA) with a 0.25-µm film thickness and a 2-m retention gap (0.32 mm i.d.). Helium was used as a carrier gas at 75 kPa. After separation, compounds were detected using a Carlo Erba QMD 1000 mass spectrometer in the EI+-mode (70 eV, full scan 140-400 amu, scan time 0.5 s., source temp 200 °C, and interface temp 275 °C). The m/z ratios used in detecting and quantifying the contaminants are shown in Table 2. All concentrations in earthworms are expressed on a lipid weight basis (µg/kg lipid). Recoveries of the earthworm extractions varied between 83.4% and 100.3% for all test chemicals, except for HCB with a recovery of 58.7%. The concentrations were corrected for recovery. 4844
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 18, 2004
Results and Discussion Concentrations in Soil, SPME Fibers and Earthworms. Concentrations in Soil. Characteristics of the selected soils and total concentrations of contaminants are given in Tables 1 and 2. Soil characteristics were similar for all soils except for organic carbon content, which was approximately 2 times lower in site 1 than in sites 2 and 3 (Table 1). Site 1 contained the lowest concentrations of soil contaminants (Table 2). Site 2 was the most contaminated site with very high concentrations of telodrin and dieldrin, which were approximately 60-70 times higher than in site 3 and 10 times higher concentrations of HCB compared to site 3. Overall the soil concentrations differed a factor of 100 000, ranging from 0.3 µg/kg (site 1, PCB #110) to 38 000 µg/kg (site 2, dieldrin). Concentrations in Fibers and Kinetics of the Uptake Process. Concentration time profiles, determined in 30-µm PDMSSPME fibers exposed for up to 43 days to soil of site 3, show that almost all chemicals have reached steady state within 20 days (see Figure 1). For the most hydrophobic chemical (PCB #180), at least 87% of steady state is reached. Based on these results, a fixed exposure time (20 days) was chosen to expose fibers to all three soils and concentrations in the fibers after 20 days exposure are shown in Table 3 as mean values from the three triplicate samples. Because each sample contained three fibers, which were analyzed separately, and because the mean concentrations in the three samples did not differ significantly from each other (one-sample t-test, group means compared to overall mean), an overall mean value was calculated from nine measurements. The average coefficients of variation are 17.2% for fiber measurements in site 1 soil, and 13.3% and 10.5% for site 2 and site 3 soil measurements, respectively. Estimated steady-state concentrations obtained from the kinetic study of site 3 are also reported in Table 3. Because the deviations between steady-state concentrations in the fibers and the concentrations measured after 20 days in site 3 soil are small for all compounds ((7% on average), the concentrations in the fiber after a 20-day exposure were used for comparisons with the earthworm data. Figure 1 shows that the time to reach equilibrium increases with the hydrophobicity of the chemicals. Estimated log k values (eq 2) are plotted against log Kow in Figure 2. Data for HCB, telodrin, and dieldrin are excluded because these chemicals reach equilibrium too fast to estimate reliable k values. The slope of this relationship is -0.63 ( 0.08. Linear relationships between log k values and hydrophobicity (with a slope of around -1.0) have been observed in uptake studies in fish, semipermeable membrane devices (SPMDs), and
FIGURE 1. Uptake curves of hydrophobic contaminants into 30-µm PDMS-coated SPME fiber after exposure to site 3 soil. SPME fibers with water exposure only (21, 22, 27-29). The lower slope of -0.63 observed in our study may be due to the presence of soil particles or dissolved organic carbon (DOC) in the aqueous diffusion layer. Desorption from particles or DOC-bound chemicals in the aqueous diffusion layer may result in an extra flux to the SPME fiber leading to shorter equilibration times (30). This phenomenon will
be more important as hydrophobicity increases and this leads to a decrease in the slope of the relation of log k versus log Kow. Concentrations in Earthworms at Steady State. Both earthworm species were exposed to site 2 and site 3 soils. Toxicity was observed in the site 2 soil, especially with E. andrei. These earthworms became inactive and swollen, and mortality was even observed. Toxicity to A. caliginosa manifested itself as loss of weight, with a maximum weight loss of 35% after 28 days. Because of the severe toxic effects on E. andrei in the site 2 soil, these data are not used in the comparisons. Concentrations in earthworms (mean values from at least three measurements at each time) against time are reported by Jager et al. (26) and only the estimated steadystate concentrations are reported in Table 4. Concentrations of PCBs in site 2 and 3 soil and in A. caliginosa exposed to both soils are very similar. Also the BCFs in both soils are very similar (see Figure 5 and the related discussion about BCF). This observation is a strong suggestion that the effects observed in A. caliginosa exposed to the site 2 soil did not affect the uptake. Lipid content was not significantly different between both species. E. andrei had a lipid content of 2.3% ( 0.6 (site 3, n ) 24) and A. caliginosa 2.2% ( 0.5 (site 3, n ) 32) and 2.1% ( 0.5 (site 2, n ) 25) of the wet weight. Comparison of Accumulation in Earthworms with Partitioning to the SPME Fiber. The data were analyzed as follows: (i) A direct comparison of concentrations on the SPME fiber with concentrations in earthworms, and (ii) A comparison of observed and calculated concentrations in earthworms based on equilibrium partitioning theory applied to: (a) Concentrations in soil (b) Concentrations in the SPME fiber Approach ii-a is the standard approach in risk assessment where estimated pore water concentrations (using soil-water partition coefficients, Koc) and bioconcentration factors (BCF) are used to calculate concentrations in soil and sediment organisms. Both BCF and Koc are often estimated from octanol-water partition coefficients (Kow). In ii-b, a similar approach is applied but now with pore water concentrations measured via SPME analyses. Comparison Cworm versus CSPME. Concentrations in the earthworms (Cworm) for both species exposed to site 3 soil and for A. caliginosa exposed to site 2 soil are highly correlated to the concentrations in the SPME fiber (CSPME) (see Figure 3). This relationship is valid over a 1000-fold difference in Kow as well as a 100 000-fold difference in concentrations of the test compounds. The ratio Cworm /CSPME, for both species exposed to site 3 soil and A. caliginosa exposed to site 2 soil, is constant and does not depend on log Kow. The mean values of Cworm /CSPME (n ) 11) for the different groups are 4.5 ( 1.7 (site 3, E. andrei), 4.8 ( 2.2 (site 3, A. caliginosa) and 3.7 ( 2.1 (site 2, A. caliginosa), respectively. Comparison of Concentrations in Earthworms Using Equilibrium Partitioning (EqP) Theory. In Figure 4A,B, concentrations in earthworms are plotted against calculated concentrations derived with classical equilibrium partitioning theory using concentrations in soil (Csoil), estimated soilwater partition coefficients (Koc), and bioconcentration factors (BCF). Two equations for soil sorption coefficients were used: eq 3 (reported by Sabljic´ et al. (31)) and the classical eq 4 from Karickhoff et al. (32). BCF on a lipid basis was estimated based on a model for earthworms (eqn 5) derived by Jager et al. (33).
log Koc ) 0.81 log Kow + 0.10
(3)
log Koc ) log Kow - 0.21
(4)
log BCF ) log Kow
(5)
VOL. 38, NO. 18, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4845
TABLE 3. Mean ((s.d.; N)9) Concentrations of Contaminants (µG/L Hydrophobic Phase) Measured in SPME Fibers Exposed for 20 days to Different Soils and Estimated Steady-state Concentrations from the Uptake Experiment (see Figure 1). The Last Column Shows the Difference between Steady State Concentrations and Concentrations Measured after T)20 days Exposure in Site 3 Soil site 1 (t ) 20 days)
site 2 (t ) 20 days)
site 3 (t ) 20 days)
site 3 (steady state)
% difference
