Persistent, Bioaccumulative, and Toxic Chemicals II - American

Chapter 8

Estimating Accumulation and Baseline Toxicity of Complex Mixtures or Organic Chemicals to Aquatic Organisms

Downloaded by UNIV MASSACHUSETTS AMHERST on July 28, 2012 | http://pubs.acs.org Publication Date: January 15, 2000 | doi: 10.1021/bk-2001-0773.ch008

The Use of Hydrophobicity Dependent Analytical Methods E r i c M. J. Verbruggen and Joop L. M. Hermens Research Institute of Toxicology, Utrecht University, P.O. Box 80176, 3508 TD Utrecht, The Netherlands

Many chemical products, such as oils, are complex mixtures. Also in the environment itself chemicals always occur as complex mixtures. Physicochemical parameters however are mostly single-compound properties. One such a parameter that has a crucial role in many environmental processes is the hydrophobicity of organic chemicals. Two methods are presented for measuring the hydrophobicity of complex mixtures: a) hydrophobicity-based fractionation (hydrophobicity distribution profile) and b) hydrophobicity dependent extraction (biomimetic extraction). For two petroleum products the estimated toxicity by these two approaches was in good agreement with toxicity to Daphnia magna.

Introduction: risk assessment of complex mixtures Risk assessment is usually based on the evaluation of the fate and toxic effects of single chemicals. However, many chemical products are not pure compounds but complex mixtures, of which the precise composition is unknown. Typical examples of these mixtures are petroleum derived products (7). Once emitted into the environment, all chemicals in effluents and surface waters are part of complex mixtures, of which the composition even after chemical analysis is largely unknown (2).

© 2 0 0 1 American Chemical Society

In Persistent, Bioaccumulative, and Toxic Chemicals II; Lipnick, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2000.

89

90 Physicochemical properties that are defined for single compounds are often used in quantitative structure activity relationships (QSARs) to predict the behavior in the environment. A very important compound property for organic chemicals is hydrophobicity. The parameters that are mostly used to reflect differences in hydrophobicity are the w-octanol-water partition coefficient (K ), aqueous solubility (S) or retention factors in reversed-phase high-performance liquid chromatography (k). The use of these hydrophobicity parameters in QSARs to estimate partition behavior is manifold. Some examples are the partition coefficient for sorption to soil (3), the bioconcentration factor (BCF) (4), and narcosis or baseline-toxicity (5, 6). Baseline toxicity is related to the concentration of compounds in the cell membranes (7) and it has been shown that the internal effect concentrations are almost constant. This level is often called the critical body residue (CBR) (8, 9). Consequently external effect concentration directly reflect the differences in bioconcentration factors. Another important feature of narcosis is that toxicity exerted by different chemicals is completely concentration additive (10-12). Because of the additivity of narcotic effects and the constant internal effect concentrations, analytical methods to estimate bioconcentration of complex organic mixtures would be very useful. It is not necessary to identify and quantify all individual compounds, but to have information about total molar concentrations. In both approaches in this paper, the hydrophobicity distribution profile and the biomimetic extraction, total molar concentrations are estimated using either vapor pressure osmometry (VPO) (13-15) or gas chromatographymass spectrometry (GC-MS) (75).


ow

Approaches to measure hydrophobicity and baseline toxicity of complex mixtures

Hydrophobicity distribution profiles (HDP) In order to determine a hydrophobicity distribution profile (HDP) of a complex mixture, it is first separated on a reversed-phase high performance liquid chromatography (RP-HPLC) column (14, 16). The retention time in gradient elution R P - H P L C (16) is a good measure for separation of a complex organic mixture according to hydrophobicity, even for micropollutants with very diverse chemical structures (see Figure 1). Individual compounds are grouped together in



91 fractions with a small range of hydrophobicity. In each fraction the total molar concentrations is determined. The resulting hydrophobicity distribution profile is presented as a limited number of blocks that can be considered to be homogeneous with respect to hydrophobicity. A n example of such a profile for kerosene is given in Figure 2. The H D P approach is very similar to the hydrocarbon block approach by C O N C A W E (18). However, here hydrophobicity is the discriminating factor, while the hydrocarbon block approach distinguishes for example between aromatic and aliphatic hydrocarbons and on carbon chain length. The results of the hydrophobicity distribution profile can be used in QSARs on basis of hydrophobicity, to estimate the fate of the mixture in the aquatic environment and baseline toxicity. The hydrophobicity parameters K and solubility of liquid compounds are strongly correlated. Both aromatic and aliphatic liquid compounds from petroleum derived substances, show almost the same relationship between K and solubility (19). The H D P determined for such complex mixtures can thus be seen as a (subcooled) liquid solubility distribution as well as a K distribution. Also other types of compounds show almost the same correlation between solubility and A: (20). If a mixture does not dissolve completely, for example i f a layer of oil is equilibrated with water, the data from a H D P can also be used to calculate to what extent the compounds in each fraction will be dissolved. For nearly ideal mixtures of organic compounds, the ratio of the equilibrium concentration of a compound in the aqueous phase and the aqueous solubility of the pure compound approaches the mole fraction of that compound in the organic phase of the mixture (21). For the partitioning of some fuel components between gasolines and water, this assumption of ideal behavior gave very good results (22). From this equation it follows that for a mixture that is just above its solubility in water, the compounds with the lowest (subcooled) liquid solubility will aggregate first. The aqueous concentration of the most hydrophobic fractions is strongly reduced, if a mixture is not completely dissolved. QW

OVf

ow

ow

Biomimetic extraction The biomimetic extraction procedure (13) is a method to determine the total amount of chemicals that will accumulate in organisms after exposure to an aqueous sample. In this procedure the bioconcentration process is simulated by a surrogate hydrophobic phase with similar partition behavior. The procedure is based on an extraction with negligible depletion and selective for compounds that will concentrate in biota. After that the total amount of organic chemicals on the hydrophobic phase is determined. The procedure was first developed making use of extraction disks coated with octadecylsilica (13, 15, 23) or solid phase microextraction (SPME) fibers (24). Polyacrylate S P M E fibers also appear to be



30

20 Βο ο Ν©

10

Ί

i

Γ

Figure 2: Hydrophobicity distribution profile of kerosene determined by GCMS.


93 suitable for biomimetic extraction purposes (25). The partition coefficients to these fibers (K ) are equally dependent on hydrophobicity as the partition coefficients to artificial membranes (K ) (see Figure 3), which make them a very good surrogate to simulate partitioning to the target site for narcosis, i.e. the cell membranes. Other advantages of the SPME procedure compared to the procedure with the extraction disks are the much shorter period to obtain equilibrium (one day instead of two weeks) and the much smaller volumes to achieve negligible depletion (250 mL instead of 10 L ) . SVME


m

a, . on OS)

Ο

Figure 3: Polyacrylate (SPME)-water (+) and membrane-water (O) partition coefficients as a function oflogK ,; data from (25). ov

Deriving internal concentrations from HDP and biomimetic extraction Bioconcentration factors of compounds that are not strongly metabolized, are strongly correlated with their hydrophobicity (4, 26-29). However, at high hydrophobicity (log K , > 5.5) constant or even decreasing BCFs have been observed (30-32). Gobas et al. (33) showed the same effect for the partition coefficient between water and artificial membranes, which is confirmed by additional data from Dulfer and Govers (34) for compounds with high log K and data from Vaes et al. for compounds of moderate hydrophobicity (35, 36). The following relationship between K and K can be established: ov

ow

m

\ogK

m

2

= -0.017 • l o g ^

r = 0.961,

w

ow

+ 0.131 ·log j £ + 0.797 ·\ogK

OVf

- 0.119

s.e.= 0.437 (n = 86)



94 A possible explanation for this effect is that for large solutes the solubility in membranes decreases as much as or faster than the solubility in water with increasing molecular volume. This can be explained by a relative fast increase in the activity coefficient of the highly structured membrane bilayers in comparison with the activity coefficients in the bulk phases water and octanol for compounds with large molecular volumes (29, 37). Also membrane permeation, which is important for the uptake of compounds into an organism e.g. via the gills of fish, is affected by the size of the molecule (38, 39). However, for compounds that are not biotransformed, slower membrane permeation will only influence the kinetics but not the equilibrium constant (38). To estimate the toxicity of a mixture due to narcosis, the total concentration at the target site, i.e. the membranes, has to be calculated from the HDP. Based on the considerations above, it is probably more accurate to use a relationship between log K and log ^ that is not linear at high hydrophobicity, especially for complex mixtures, containing many very hydrophobic compounds such as petroleum products. In the case of the biomimetic extraction, the membrane concentration is directly estimated from the concentration on the hydrophobic phase. The partitioning to this hydrophobic phase has to be similar to the membrane-water partitioning. The few available data on very hydrophobic compounds indicate that partitioning to polyacrylate SPME fibers is also not linear with hydrophobicity in the range above log ^ 5.5 (25). m

o w

o w

Application of HDP and biomimetic extraction to predict baseline toxicity of a petroleum products Petroleum products mainly consist of aliphatic and aromatic hydrocarbons (18). For this type of compounds acute toxicity can probably be attributed to narcosis, e.g. (40, 41). Therefore, this type of complex mixtures is most suitable to test whether the two methods do provide good estimates for baseline toxicity. The biomimetic extraction procedure will give a direct estimate of the internal concentration, which is compared to an internal effect concentration (Figure 4). As internal effect concentrations, a critical target concentration (CTC, i.e. concentration in the membranes) of 125 m M can be used for median lethality due to narcosis (25). To estimate toxicity from the hydrophobicity distribution profile, aqueous concentrations, resulting from dosing in the toxicity test have to be calculated first (Figure 4). For this purpose, the hydrophobicity of the different fractions in the H D P can be used to estimate the aqueous solubility of each fraction. For nearly ideal mixtures of organic compounds, the equilibrium concentration of a compound in the aqueous phase can be estimated as the product of its aqueous solubility and the mole fraction in the organic phase of the mixture (21). Because



95 the H D P fractions are considered homogeneous, all compounds in it have the same aqueous solubility and can be treated as one. Then, the different fractions are the components, to which the ideal mixture theory can be applied, using a mass balance that accounts for the part of each fraction in the organic phase of the mixture and the part that is dissolved in the aqueous phase. From the aqueous concentrations of each fraction and its hydrophobicity, the contribution to the total internal concentration can be calculated. The hydrophobicity distribution profile is needed to estimate the toxicity after emission or dosing to water for complex organic chemical products, such as petroleum products. The biomimetic extraction procedure is suitable to estimate toxicity from a water sample that contains a complex mixture of organic chemicals, which is common for environmental water samples, for example water contaminated by oil (Figure 4).

Figure 4: Overview of analytical methods to assess mixture toxicity.

Comparison of the analytical methods with toxicity for two petroleum products For two petroleum products (a gas oil and kerosene), the two approaches are applied and compared with a 48-h toxicity test with water fleas (Daphnia magna). Details are described in (19). The observed median lethal loadings (LL ) are displayed in Table 1. The somewhat lower lethal loading of kerosene can probably be explained from the relative higher amount of compounds with log K 4-6 in comparison with the heavier gas oil. These compounds are relatively soluble and probably have almost equal membrane-water partition coefficients as the highly hydrophobic compounds. From the biomimetic extractions with polyacrylate SPME fibers the internal concentrations (total target concentrations = concentration in the cell membranes) were estimated. By comparing these TTC values with the critical target concentration of 125 m M for lethality, median lethal loadings of the petroleum 50

ow



96 products could be estimated (Table 1). B y expressing the percentage lethality that was observed in the toxicity experiment as a dose-response function of the estimated TTC values, the critical target concentration can also be estimated (Table 1). For the calculations of lethal loadings from the hydrophobicity distribution profiles, the internal concentrations at equilibrium are calculated from the membrane-water partition coefficients and the aqueous concentrations after dosing. B y combining with the results of the toxicity experiment, the CTC values are derived (Table 1). Kinetic behavior of the uptake process can be estimated from a Q S A R on basis of log AT for the first order elimination rate constant for daphnids (42), log k . However, with the assumption of ideal mixture behavior with regard to solubility and membrane-water partition coefficients that level off at high hydrophobicity, non-equilibrium plays a minor role in a 48-h toxicity test with water fleas. 0W

2

Table 1: Lethal loading rates (LL ) and estimated critical target concentration (CTC ) determined by a 48-h toxicity test with Daphnia magna, the SPME biomimetic extraction procedure or hydrophobicity distribution profiles (HDP). SQ

est

LL

[μΙ/L] Toxicity test Daphnia magna Biomimetic extraction HDP C T C . [mM] Biomimetic extraction HDP

Gas oil

Kerosene

Parameter 5 0

5.1 3.6 2.7

1.7 0.8 5.7

166 149

206 78

est

In the calculations, it is assumed that the petroleum products are ideal mixtures with respect to mixture solubility. Further, it is supposed that membrane-water partitioning can be estimated from hydrophobicity in the case of the HDPs or can be mimicked by polyacrylate S P M E fibers in the case of the biomimetic extractions. These assumptions will lead to small differences between the methods. Still, both methods, biomimetic extraction and hydrophobicity distribution profile, provide good estimates of the toxicity of these mixtures. It can be concluded that 1. Acute toxicity of complex petroleum products can be explained entirely by narcosis. 2. The S P M E biomimetic extraction procedure is a very useful tool in estimating the total internal concentrations in aquatic organisms after exposure to complex mixtures of organic chemicals that are present in the aquatic environment.


97


3. The hydrophobicity distribution profiles provide useful information for the risk assessment of chemical products that are complex mixtures of organic chemicals, not only for predicting environmental concentrations but also for estimation of baseline toxicity. The advantage of both methods is that they provide information that can be expressed directly in terms of concentrations and toxic effects. Moreover, in both methods it is not necessary to identify individual compounds. Still, with G C - M S as molar detection technique it is possible to identify major compounds.

References (1) C O N C A W E Ecology group The chemistry and grouping of petroleum -derived substances. Paper for the DG XI/CONCAWE workshop on the environmental risk assessment of petroleum substances to be held at the EU Joint Research Centre, Ispra, on 6/7 December 1994; C O N C A W E : Brussels, Belgium, 1994. (2) Hendriks, A. J.; Maas-Diepeveen, J. L.; Noordsij, Α.; van der Gaag, M. A . Water Res. 1994, 28, 581-598. (3) Karickhoff, S. W.; Brown, D . S.; Scott, T. A. Water Res. 1979, 13, 241-248. (4) Neely, W. B . ; Branson, D. R.; Blau, G . E. Environ. Sci. Technol. 1974, 8, 1113-1115. (5) Könemann, H . Toxicology 1981, 19, 209-221. (6) Veith, G. D.; Call, D. J.; Brooke, L . T. Can. J. Fish. Aquat. Sci. 1983, 40, 743-748. (7) van Wezel, A . P.; Opperhuizen, A . Crit. Rev. Toxicol. 1995, 25, 255-279. (8) McCarty, L . S.; Mackay, D.; Smith, A. D.; Ozburn, G. W.; Dixon, D . G . Sci. Total Environ. 1993,109/110, 515-525. (9) McCarty, L . S.; Mackay, D . Environ. Sci. Technol. 1993, 27, 1719-1728. (10) Könemann, H . Toxicology 1981, 19, 229-238. (11) Hermens, J.; Canton, H . ; Janssen, P.; de Jong, P. Aquat. Toxicol. 1984, 5, 143-154. (12) Broderius, S.; Kahl, M. Aquat. Toxicol. 1985, 6, 307-322. (13) Verhaar, H . J. M.; Busser, F. J. M.; Hermens, J. L. M. Environ. Sci. Technol. 1995, 29, 726-734. (14) Verbruggen, Ε. M. J.; van Loon, W. M. G. M.; Hermens, J. L . M. Environ. Sci. Poll. Res. Int. 1996, 3, 163-168. (15) van Loon, W. M. G. M.; Wijnker, F. G.; Verwoerd, M. E.; Hermens, J. L . M. Anal. Chem. 1996, 68, 2916-2926. (16) Verbruggen, Ε. M. J.; Klamer, H. J. C.; Villerius, L . ; Brinkman, U. A. T.; Hermens, J. L. M. J. Chromatogr. A 1999, 835, 19-27. (17) Biobyte Corp.; MedChem: Claremont, C A , U S A , 1994.



98 (18) C O N C A W E Ecology group Environmental risk assessment of petroleum substances: the hydrocarbon block approach; C O N C A W E : Brussels, Belgium, 1996. (19) Verbruggen, Ε. M. J. PhD, Utrecht University, Utrecht, 1999. (20) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D . M. Environmental organic chemistry; John Wiley & Sons: New York, 1993. (21) Banerjee, S. Environ. Sci. Technol. 1984, 18, 587-591. (22) Cline, P. V.; Delfino, J. J.; Rao, P. S. C. Environ. Sci. Technol. 1991, 25, 914-920. (23) van Loon, W. M. G . M.; Verwoerd, M. E.; Wijnker, F. G.; van Leeuwen, C. J.; van Duyn, P.; van de Guchte, C.; Hermens, J. L. M. Environ. Toxicol. Chem. 1997, 16, 1358-1365. (24) Parkerton, T. F.; Stone, M. A. SETAC 17th annual meeting, Washington, D C , U S A , 17-21 November 1996; Abstract book, S E T A C Press; 150. (25) Verbruggen, Ε. M. J.; Vaes, W. H . J.; Parkerton, T. F.; Hermens, J. L. M. Environ. Sci. Technol. 2000, 34, 324-331. (26) Kenaga, Ε. E. Environ. Sci. Technol. 1980, 14, 553-556. (27) Mackay, D . Environ. Sci. Technol. 1982, 16, 274-278. (28) Chiou, C. T. Environ. Sci. Technol 1985, 19, 57-62. (29) Banerjee, S.; Baughman, G. L . Environ. Sci. Technol. 1991, 25, 536-539. (30) Opperhuizen, Α.; van de Velde, E. W.; Gobas, F. A . P. C.; Liem, D . Α. Κ.; van de Steen, J. M. D . Chemosphere 1985, 14, 1871-1896. (31) Anliker, R.; Moser, P. Ecotoxicol. Environ. Saf. 1987, 13, 43-52. (32) Oliver, B . G.; Niimi, A . J. Environ. Sci. Technol. 1985, 19, 842-849. (33) Gobas, F. A. P. C.; Lahittete, J. M.; Garofalo, G.; Shiu, W. Y.; Mackay, D . J. Pharm. Sci. 1988, 77, 265-272. (34) Dulfer, W. J.; Govers, H . A. J. Environ. Sci. Technol. 1995, 29, 2548-2554. (35) Vaes, W. H . J.; Urrestarazu Ramos, E.; Hamwijk, C.; van Holsteijn, I.; Blaauboer, B . J.; Seinen, W.; Verhaar, H . J. M.; Hermens, J. L . M. Chem. Res. Toxicol. 1997, 10, 1067-1072. (36) Vaes, W. H . J.; Urrestarazu Ramos, E.; Verhaar, H . J. M.; Cramer, C. J.; Hermens, J. L . M. Chem. Res. Toxicol. 1998, 11, 847-854. (37) Gobas, F. A . P. C.; Shiu, W. Y.; Mackay, D . In QSAR in Environmental Toxicology - II; Kaiser, K . L . E., Ed.; D . Reidel Publishing Company: Dordrecht, The Netherlands, 1987, pp 107-123. (38) Gobas, F. A. P. C.; Opperhuizen, Α.; Hutzinger, O. Environ. Toxicol. Chem. 1986, 5, 637-646. (39) Opperhuizen, Α.; Sijm, D . T. H . M. Environ. Toxicol. Chem. 1990, 9, 175186. (40) Hutchinson, T. C.; Hellebust, J. Α.; Tam, D.; Mackay, D.; Mascarenhas, R. Α.; Shiu, W. Y . In Hydrocarbons and halogenated hydrocarbons in the aquatic environment; Afghan, Β. K . ; Mackay, D., Eds.; Plenum Press: New York, N Y , U S A , 1980, pp 577-586. (41) Bobra, A . M.; Shiu, W. Y.; Mackay, D . Chemosphere 1983, 12, 1137-1149. (42) Hawker, D . W.; Connell, D . W. Ecotoxicol. Environ. Saf. 1986, 11, 184-187.


Persistent, Bioaccumulative, and Toxic Chemicals II - American

Recommend Documents