Chemical and toxicological characterization of the bricks produced from clay/sewage sludge mixture

CHEMICAL AND TOXICOLOGICAL CHARACTERIZATION OF AN UNRESOLVEDCOMPLEX MIXTURE-RICH BIODEGRADED CRUDE OIL

ALF G. MELBYE,{ ODD G. BRAKSTAD,{ JORUNN N. HOKSTAD,{ INGER K. GREGERSEN,{ BJØRN H. HANSEN,{ANDY M. BOOTH,*{§ STEVEN J. ROWLAND,§ and KNUT E. TOLLEFSEN{I

{Department of Marine Environmental Technology, SINTEF Materials and Chemistry, Brattørkaia 17B, TrondheimN-7465, Norway

{Department of Ecotoxicology and Risk Assessment, Norwegian Institute for Water Research (NIVA), Gaustadalleen 21,Oslo N-0349, Norway

§School of Earth, Ocean and Environmental Sciences, University of Plymouth, Drake Circus, Plymouth PL4 8AA, United KingdomIDepartment for Plant and Environmental Sciences, University of Life Sciences (UMB), P.O. Box 5003, N-1432 As, Norway

(Received 29 October 2008; Accepted 11 March 2009)

Abstract—Chemical and toxicological characterization of unresolved complex mixtures in the water-soluble fraction of an artificiallyweathered Norwegian Sea crude oil was determined by a combination of chemical analysis and toxicity testing in fish in vitrobioassays. The water-soluble fraction of the crude oil was separated into 14 increasingly polar fractions by preparative high-pressureliquid chromatography. The in vitro toxicity (7-ethoxyresorufin O-deethylase activity, estrogenicity, and metabolic inhibition) of thesefractions was characterized in a primary culture of liver cells (hepatocytes) from rainbow trout (Oncorhynchus mykiss). The maincontributor to toxicity was one of the most polar fractions, accounting gravimetrically for more than 70% of the organic material in thewater-soluble fraction and dominated by an unresolved complex mixture. Chemical analysis by gas chromatography–massspectrometry and comprehensive two-dimensional gas chromatography–time of flight–mass spectrometry identified a large number ofcyclic and aromatic sulfoxide compounds and low amounts of benzothiophenes (,0.1% of total mass) in this fraction. Commonlymonitored toxic components of crude oil (e.g., naphthalenes, polycyclic aromatic hydrocarbons, and alkylated phenols) eluted in lesspolar fractions, characterized by somewhat lower toxicity. Normalization of in vitro responses to the mass in each fractiondemonstrated a more even distribution of toxicity, indicating that toxicity in the individual fractions was related to the amount ofmaterial present. Although polar and nonpolar compounds contribute additively to crude oil toxicity, the water-soluble fraction wasdominated by polar compounds because of their high aqueous solubility and the high oil–water loading. Under these conditions, thepolar unresolved complex mixture–rich fraction might account for a large portion of crude oil toxicity because of its high abundance inthe water-soluble fraction.

Keywords—Petroleum hydrocarbons Polar compounds Unresolved complex mixture Bioassay Toxicity

INTRODUCTION

In addition to well-known hydrocarbons, crude oils often

contain poorly characterized mixtures of organic compounds,

sometimes referred to as unresolved complex mixtures

(UCMs) [1]. Such UCMs typically dominate the gas chro-

matograms of in-reservoir biodegraded crude oils and petro-

leum-contaminated environmental samples. Previous studies

have shown that mussels from the United Kingdom coast that

had bioaccumulated high concentrations of aromatic hydro-

carbon UCMs exhibited impaired health [2–4]. Analysis of the

mussel tissue extracts using comprehensive two-dimensional

gas chromatography–time of flight–mass spectrometry

(GC3GC-ToF-MS) led to the identification of thousands of

highly branched alkylaromatic hydrocarbons, including ben-

zenes, tetralins, naphthalenes, and more polycyclic aromatic

hydrocarbons (PAHs) [5]. Toxicity studies exposing mussels to

model UCM compounds have clearly shown that the impaired

health observed in wild populations is at least partially

attributable to the bioaccumulation of these chemicals [6,7].

Toxicological effects resulting from exposure to aromatic

UCM hydrocarbons have also been reported in other

organisms (e.g., the amphipod Corophium volutator) [8].

Furthermore, UCMs are widespread in the environment

mainly because of their resistance to natural weathering

processes. The UCM can originate from the weathering of

oils from natural and anthropogenic sources or from direct

inputs of refined oil fractions such as lubricants [1].

Biodegradation studies using model UCM hydrocarbons

(e.g., branched alkylbenzenes and tetralins) have shown this

resistance is at least partly due to the highly branched nature of

the alkyl components [1,9,10]. The combination of persistence,

bioavailability, and toxicity of aromatic UCM compounds

raises potential environmental concerns that warrant further

investigation, as addressed in this study.

In addition, crude oils contain a varying proportion of

more polar organic compounds that contain nitrogen, sulfur,

and oxygen atoms in their structures (NSO compounds).

Those polar compounds present in low concentrations are

expected to contribute to the UCMs [11]. The high polarity of

these compounds results in increased dissolution in the water

column when compared with their equivalent hydrocarbon

homologues (reviewed by Booth [12]). Following oil spills at

sea, these compounds have the potential to reach high aqueous

concentrations because of their high solubility, and they

potentially exert a more specific toxic mechanism of action

than hydrocarbon homologues that lack NSO heteroatoms.

Despite recent advances in analytical chemical techniques (e.g.,

GC3GC-ToF-MS), the vast majority of polar organic

compounds within UCMs still remain unidentified.

* To whom correspondence may be addressed([email protected]).

Published on the Web 5/4/2009.

Environmental Toxicology and Chemistry, Vol. 28, No. 9, pp. 1815–1824, 2009’ 2009 SETAC

Printed in the USA0730-7268/09 $12.00 + .00

1815

The use of in vitro bioassays has allowed high-throughput

screening and detailed studies to quantify the relative potency

of single chemicals, chemical mixtures, and environmental

extracts [13–18]. These bioassays, typically based on cellular

components, cells, and organs of small animals, have the

advantage of being highly sensitive, rapid, and reproducible

and require only minute amounts of sample material. In

addition to determining a wide variety of different toxicity

endpoints, these bioassays are able to integrate the toxicolog-

ical activity of multiple contaminants acting through a

common toxic mechanism through use of the toxicity

equivalent approach [19]. This makes it possible to assess the

potential of biological effects in complex samples without prior

knowledge of the exact chemical composition of the samples

tested. Development of bioassay-directed analysis (i.e., the

integration of chemical fractionation, bioassay testing, and

high-resolution chemical analysis) has recently provided links

between in vitro toxicological effects and exposure to oil-

related pollutants of environmental concern [18,20].

In this study, 14 high-pressure liquid chromatography

(HPLC) fractions (nonpolar to polar) were isolated from the

WSF of an in-reservoir biodegraded Norwegian Sea crude oil

after artificial weathering in the laboratory. In vitro bioassay

testing was used to determine the contribution of the fractions

to various toxic responses in rainbow trout (Oncorhynchus

mykiss) primary hepatocytes. Measurement of cytotoxicity

(metabolic inhibition), 7-ethoxyresorufin O-deethylase

(EROD) activity, and the estrogenic biomarker vitellogenin

were performed to characterize the toxic potential in the

individual fractions. Results from the bioassays and chemical

characterization were then used to design synthetic mixtures

and determine the contribution of key oil-related pollutants to

the observed in vitro toxicity of the UCM fractions. Although

in vitro toxicity assessment of a single oil–water loading does

not permit direct comparison with standardized in vivo toxicity

assessments and extrapolation to other oils and UCM composi-

tions, in this paper, we report the toxicity and chemical

characterization of the WSF from a UCM-rich oil and its

potential for toxic effects in aquatic animals such as fish.

MATERIALS AND METHODS

Preparation of water-soluble fractions

An in-reservoir biodegraded Norwegian Sea crude oil was

artificially weathered by atmospheric distillation at 200uC. The

crude oil was transferred to a round-bottom flask and heated

to 200uC for 2 h using a heating mantle. The nitrogen content

of the fresh crude oil was 770 mg/kg, and sulfur content

accounted for 0.211% of the mass. Seawater (112 L) was

collected from Trondheimsfjorden (Trondheim, Norway) and

sterile-filtered (0.2 mm exclusion limit). The seawater was

transferred to 2 3 45 L and 4 3 10 L PyrexH bottles with a

tap and valve slightly above the bottom of the bottle. The

weathered oil was carefully applied to the surface (oil:seawater

1:100; 10 g/L), leaving a headspace to water volume ratio of

approximately 1:4. The bottles were sealed, and a magnetic

stirrer used to mix the water without creating a vortex. The

WSF was prepared at room temperature (,20uC) and in

darkness according to the method of Singer et al. [21]. After

3 d equilibration time, the resulting WSF was removed

through the tap and collected in 2-L separating funnels. The

WSF was extracted in batches with 90 + 30 + 30 ml of

dichloromethane (DCM). Anhydrous sodium sulfate (Na2SO4)

was added to the DCM extracts to remove any remaining

water. The extracts were then filtered to remove the Na2SO4

before being concentrated (TurboVap, Caliper Life Sciences)

and combined in 25 ml of DCM. The total organic extract

(TOE) was split into two identical samples: one for gravimetric

analyses and one for fractionation and subsequent character-

ization and toxicity testing. The extracts were transferred to

preweighed vials and dried (N2). The mass of the TOE was

determined by weighing the vial on a five-figure balance scale.

Fractionation of WSFs

Fractionation of the WSF was performed on a semipre-

parative scale with the use of an Agilent LC1100 HPLC setup

with two columns in series (Zorbax NH2, 9.4 mm 3 25 cm;

Zorbax Si, 9.4 mm 3 25 cm). Elution was performed with a

gradient of hexane, dichloromethane, and methanol (see

Fig. 1). The flow rate was 5 ml/min. The fractionation method

was developed with the use of a mixture of standards

representing saturate, aromatic, and resin-type compounds.

Chromatographic retention times and recovery efficiencies

were determined for the standard compounds, and these were

used to identify the compound classes expected in each of the

HPLC fractions. The WSF extract was split (eight vials) and

fractionated in a series of consecutive injections (16 injections,

two from each vial). Injection volume was 900 ml. A total of 14

separate fractions of increasing polarity (each 32 ml [6.4 min

3 5 ml/min] from each run) were collected by a Gilson

FC203B fraction collector modified to receive larger volumes.

The fractions were transferred to preweighed vials and dried

(TurboVap and N2). The mass of each fraction was determined

by weighing the vials on a five-figure balance before storage in

the cold and dark.

Synthetic mixtures

On the basis of the chemical characterization and toxico-

logical analysis of the individual fractions, three of the most

interesting fractions were selected for in-depth studies. These

were fraction (F)1 (least polar, dominated by aromatic

hydrocarbons), F6 (medium polar, containing phenols), and

F11 (polar, UCM-dominated and accounting for most of the

WSF material). These fractions were chosen because they

exhibited differing polarity and chemical composition, while

also containing higher amounts of organic compounds

(determined gravimetrically). Synthetic mixtures (Supporting

Information, Table S1; http://dx.doi.org/10.1897/08-545.S1)

representing the concentrations of identifiable and commonly

studied compounds in these fractions were produced so that

their toxicity could be assessed. Forty-seven percent of the

compounds in F6 and 14% of the compounds in F1 were

identifiable and were included in the corresponding synthetic

mixture. In contrast, less than 0.1% of the compounds in F11

could be identified, the rest being highly complex UCMs. The

chemicals used in the synthetic mixtures were prepared in stock

solutions of 1 mg/ml. All stock solutions were prepared in

DCM. These synthetic mixtures were analyzed by gas

chromatography–mass spectrometry (GC-MS) and tested for

their toxicity in several fish in vitro bioassays.

Chemical analyses

The WSF extracts and HPLC fractions were analyzed by

gas chromatography with flame ionization detector (GC-FID),

GC-MS, and comprehensive GC3GC-ToF-MS. The GC-FID

1816 Environ. Toxicol. Chem. 28, 2009 A.G. Melbye et al.

analysis was performed on a Hewlett-Packard 5890 Series II

gas chromatograph fitted with a flame ionization detector. The

column was a HP-5 fused silica capillary column (30 m 3

0.25 mm internal diameter [i.d.] 3 0.25 mm film thickness).

The carrier gas was hydrogen at a constant flow of 2.0 ml/min

and the make-up gas was helium at a constant flow of 30 ml/

min. A 1.0-ml sample was injected into a 275uC splitless

injector. The oven temperature was programmed from 40uC(held 5 min) to 310uC at 6uC/min and held for 10 min. The

flame ionization detector was operated at a temperature of

325uC and supplied with air (360 ml/min) and hydrogen

(30 ml/min). Data and chromatograms were monitored and

recorded by ChemStation software (Agilent Technologies).

Analysis by GC-MS in selected ion monitoring mode was

used to identify target compounds in the fractions (described

in the Oljeindustriens Landsforening Guideline for Produced

Water Analysis) [22]. The GC-MS analysis was performed on a

Hewlett-Packard 6890 gas chromatograph fitted with a HP5973

quadrupole mass selective detector. The column was a HP-5MS

fused silica capillary column (60 m 3 0.25 mm i.d. 3 0.25 mm

film thickness). The carrier gas was helium at a constant flow of

1.0 ml/min. A 1.0-ml sample was injected into a 300uC pulsed

splitless injector. The oven temperature was programmed from

40uC (held 1 min) to 300uC at 6uC/min and held for 20 min. The

quadrupole mass spectrometer used ionization energy of 70 eV

and an ion source temperature of 230uC. It was operated in full

scan mode, with a mass range of 50 to 600 daltons monitored.

Data and chromatograms were monitored and recorded with

ChemStation (with EnviroQuant) software (Agilent).

Selected sample analysis was performed on a Pegasus 4D

(Leco Corporation) GC3GC-ToF-MS system, based on an

Agilent 6890 Gas Chromatograph interfaced to a Pegasus III

time-of-flight mass spectrometer (Leco). Differences from the

methods previously reported [5,6] are outlined here. The

system used the following parameters: injector 300uC, transfer

line 280uC. First-dimension column was DB-5 (30 m 3 320 mm

3 0.25 mm; J&W Scientific), and the second-dimension column

was DPX-50 (2.0 m 3 100 mm 3 0.1 mm; SGE Analytical

Science). The first-dimension oven was raised from 70 to

240uC at 5uC/min, then to 270uC at 20uC/min (5-min

isothermal period). The second-dimension oven was raised

from 85 to 245uC at 5uC/min, then to 285uC at 20uC/min (5-

min isothermal period). A second-dimension cryogenic mod-

ulation period of 4 s was employed. The carrier gas was

helium, and 1 ml of the sample was injected (splitless) into the

GC3GC-ToF-MS system via a Gerstel Multipurpose Sam-

pler. A ToF-MS was used as the detector and operated at a

spectrum storage rate of 100 Hz (100 spectra/s). The ion source

was operated at 250uC, the electron multiplier at 1,750 V, and

the mass range monitored was from 40 to 500 Daltons. The

automated data processing was achieved with LECOHChromaToFTM software (version 2.01, Leco).

Fish in vitro bioassays

Fourteen internal standard–free HPLC fractions from the

WSF of Norwegian Sea crude oil were solvent changed to

ultrapure anhydrous dimethyl sulfoxide (DMSO, Sigma-

Aldrich) under a gentle stream of nitrogen before testing in

the bioassays. The hepatocytes were isolated from male

rainbow trout (200–500 g) obtained from Killi Oppdrettsan-

legg (Dombas, Norway), seeded as a monolayer culture, and

exposed to test chemicals and extracts as previously described

[23]. Briefly, hepatocytes were isolated by a two-step perfusion

Fig. 1. Gas chromatograms of the total water soluble fraction (WSF) and individual high-pressure liquid chromatography (HPLC) isolated fraction(F)1, F6, and F11. The solvent polarity gradient used in the HPLC fractionation is shown for each of the 14 fractions.

Characterization and toxicity of the WSF of crude oil UCMs Environ. Toxicol. Chem. 28, 2009 1817

method and diluted in serum-free L-15 medium (Biowhit-

taker); 200-ml cell suspensions (0.5 million cells/ml) were seeded

in 96-well PrimariaH plates (Becton-Dickinson Labware) and

kept in ambient atmosphere at 15uC. The cells were then

cultured for 1 d in growth medium before replacement of half

the initial volume by growth medium containing the vehicle

DMSO (final concentration ,1%, v/v), different concentra-

tions of the bioassay standards 17b-estradiol (E2), 2,3,7,8-

tetrachlorodibenzo-p-dioxin (TCDD), copper sulfate (CuSO4),

and the HPLC fractions. The EROD activity was determined

after 48 h exposure, and cytotoxicity (metabolic activity) as

well as vitellogenin (VTG) induction was determined after 96-h

exposures. In the case of cells receiving 96 h of exposure, half

of the medium was changed after 2 d, and the cells were re-

exposed for an additional 2-d period before the cells were

subjected to analysis of metabolic activity. The cell medium

was transferred directly to 96-well MaxisorpTM microtiter wells

(Nunc) and frozen at 280uC for subsequent analysis of VTG.

The cytochrome P450 1A–mediated EROD activity was

determined directly in the cell culture media as a measure of

aryl hydrocarbon receptor (AhR) agonists, essentially as

described by Ganassin et al. [24]. The EROD activity was

expressed as TCDD toxicity equivalents (TCDD-TEQ),

obtained by comparing the EROD response of the fractions

with that of the standard inducing agent TCDD. Cytotoxicity

(metabolic inhibition) was determined directly in the cell

culture with the fluorescent probes Alamar blue, originally

proposed by Schirmer et al. [25] and modified for the current

test system [16]. The fluorescence of cells exposed to the

fractions was expressed relative to the DMSO control (no

effect) and the maximum toxicity obtained for CuSO4

(10 mM). The concentration of the extract in growth media

causing 50% effect (EC50) and toxicity units (TU, 100/EC50)

was calculated for the fractions causing toxicity. Vitellogenin,

a biomarker for estrogen receptor (ER)–mediated cellular

responses, was measured directly in the cell culture media by a

capture enzyme-linked immunosorbent assay (ELISA) as

described by Tollefsen et al. [23]. The VTG production was

expressed relative to maximum VTG production after 96 h of

exposure to 100 nM E2, and the concentration of ER agonists

in the fractions was expressed in estrogen equivalents (E2-EQs)

on the basis of the dose–response curve for E2. Protein content

in the cells was determined by the Lowry method [26] with the

use of immunoglobulin G (IgG) as the protein standard.

The vehicle DMSO did not lead to significant alterations in

any of the endpoints measured compared with cells grown in

cell culture media, and high-quality dose–response curves were

routinely obtained for all bioassay standards. All bioassay

exposures were conducted in triplicate, and a minimum of two

experiments was conducted independently with cells from

individual fish.

Regression and statistical analysis

Regression (dose–response curve fitting) and statistical

analyses were performed with GraphPad Prism 4.0 software.

Statistical differences between groups were analyzed by t test

(pairwise analysis). Groups were considered significantly

different from control at the p , 0.05 level.

RESULTS

Chemical characterization of the WSF HPLC fractions

Figure 1 shows the gas chromatogram of the extracted

WSF. The chromatogram contained a number of resolved

compounds in addition to a significant UCM. The GC-MS

analysis of the WSF permitted identification of many of the

major resolved compounds and calculation of the total

extractable organic material (nC10–nC40). The crude WSF

extracted from 56 L of filtered seawater yielded 120 mg of

organic material, giving a WSF concentration of 2.1 mg/L.

The major classes identified were alkyldiaromatic hydrocar-

bons (including C0–4 naphthalenes), PAHs (C0–4), benzo- and

dibenzothiophenes, and alkylphenols (C0–4). However, no

compounds in the UCM region of the chromatogram could

be determined because of the limited separation power of such

highly complex samples by GC-MS.

High-pressure liquid chromatography permitted separation

of the WSF into 14 individual fractions of increasing polarity

(F1–F14; Fig. 1). The two peaks present on top of the UCM in

the total WSF extract chromatogram were identified as

extraction standards (o-terphenyl and 5-a-androstane). The

total amount of material in each fraction was determined

gravimetrically, and the percent contribution to the WSF was

calculated (Table 1). Table 1 shows a significant proportion

(10.2%) of the WSF compounds eluted in F1, which

corresponds to a mobile phase comprising 100% hexane.

Fractions 2 to 4 (up to 50% DCM in hexane) contained

virtually no material (,1%; Table 1), indicating that most

low-polarity WSF compounds eluted in F1. The midpolar (50–

70% DCM) F5 to F7 accounted for approximately 2 to 5% of

the WSF (Table 1) and were dominated by GC-resolvable

components (Fig. 1). Fractions 8 to 10 (100% DCM)

accounted for 1 to 2% of the WSF, but the chromatograms

were dominated by small UCMs (results not shown). Fraction

11 (methanol 0–20% in DCM) accounted for 71.1% of the

total WSF compounds. The GC-MS chromatogram of F11

(Fig. 1) was dominated by a broad UCM with very few

resolved peaks. The final F12 to F14 (methanol 20–50% in

DCM) contained very little material (,2%). Gravimetrically,

therefore, more than 70% of the organic material present in the

crude oil WSF was accounted for as polar UCM compounds

in F11.

The toxicity of crude oil WSFs is typically associated with

readily identifiable compounds such as naphthalenes, PAHs,

and phenols. Figure 2 shows the percent contribution of

identified compounds relative to the estimated masses of F1

and F5 to F11 (fractions that accounted for .95% of the WSF

Table 1. Estimated concentration of organic material in each of the 14high-pressure liquid chromatography–separated fractions (F1–F14)and their percent contribution to the total organic extract of the water-

soluble fraction (WSF)

WSF fraction Concentration (mg/L) % WSF

1 0.21 10.22 0.01 0.43 0.01 0.44 0.01 0.35 0.04 1.86 0.06 3.07 0.11 5.18 0.03 1.29 0.04 1.810 0.03 1.611 1.49 71.112 0.04 1.813 0.01 0.614 0.01 0.7

Total 2.1 100


mass and toxicity). The least polar fraction (F1) was dominated

by C0–3 naphthalenes and three-ring PAHs and their C0–3

alkylated homologues. The midpolar fractions (F5–F7) con-

tained further C0–3 alkylated PAHs (three- to five-ring), C0–4

alkylphenols, and some alkylbenzothiophenes. Of the remaining

polar fractions, F8 to F10 contained only small UCMs and no

identifiable compounds, whereas F11 contained a large UCM, of

which a small contribution (,1%) was identified as C1–3

dibenzothiophenes by GC-MS.

The GC3GC-ToF-MS analysis of the UCM-rich F11

indicated the presence of more than 3,000 compounds, of

which a few provided excellent mass spectral matches to

sulfoxide compounds in the National Institute of Standards

and Technology (NIST) mass spectral library. This informa-

tion, together with chromatographic behavior and the

diagnostic mass spectra generated for each peak, permitted

the identification of series of sulfoxide compounds. These were

monocyclics, bicyclics, bicyclenes (single double bond), bicy-

clic-dienes (two double bonds), and monocyclicaromatics

(Supporting Information, Fig. S1; http://dx.doi.org/10.1897/

08-545.S1). In each series both five- and six-membered cyclic

rings were identified, and increasing degrees of alkylation were

observed. The large number of isomeric compounds also

indicated a significant degree of branching on these alkyl

chains. Because the GC3GC-ToF-MS analysis is qualitative

rather than quantitative, a mass contribution for these

sulfoxide compounds was beyond the scope of this study.

Toxicity of HPLC-separated fractions to fish hepatocytes

The fish in vitro bioassay testing was performed on the 14

HPLC fractions. The toxicity values were determined for each

HPLC fraction isolated from the crude extract, in that this was

assumed to represent the natural distribution of the crude oil

components in the water column (Fig. 3A, 3D, and 3G). With

the use of the toxicity data and the mass of each fraction (F1–

F14), the toxicity values were then normalized per milligram of

each fraction (Fig. 3B, 3E, and 3H). Because the log scale in

Figure 3B, 3E, and 3H might underrepresent the differences

between each fraction, an additional set of graphs (Fig. 3C,

3F, and 3I) has been included. Figure 3C, 3F, and 3I show the

mass-normalized data expressed as a percentage of the total

WSF toxicity (i.e., those fractions that contribute the most are

more readily identified).

The presence of AhR agonists, measured as an increase in

EROD activity (Fig. 3A), was observed in most fractions, with

those containing low-polarity aromatic hydrocarbons (F1) and

polar UCMs (F10–F12) inducing the highest responses.

Interestingly, fractions with relatively little material also

contained AhR agonists (e.g., F2–F4, F8, and F10). After

mass normalization (Fig. 3B), a more even distribution of

toxicity across all of the fractions was observed, although F10

accounted for almost 40% of the total WSF bioactivity

(Fig. 3C). Cellular metabolic inhibition (Fig. 3D) was ob-

served after exposure to F1 and F5 to F11, with apparently the

most pronounced cytotoxicity in F1 (least polar) and F11

(polar UCM). However, mass normalization of the data

(Fig. 3E) shifted the toxicity toward the midpolar fractions

(F5–F10). Figure 3F shows that individually these fractions

accounted for 5 to 30% of the total WSF toxicity and together

they account for approximately 90% of the toxicity.

Fraction 1 and F6 to F12 caused significant cellular ER-

mediated induction of the estrogenic biomarker VTG

Fig. 2. Percent contribution of identified compounds relative to the estimated masses of fraction (F)1 and F5 to F11 (A) total compounds, (B) C0–C3

naphthalenes, (C) C0–C3 polycyclic aromatic hydrocarbons (PAHs; including benzothiophenes), and (D) C0–C4 phenols.


(Fig. 3G). As with EROD activity and metabolic inhibition,

the highest VTG induction was observed in the polar UCM-

rich F11. This suggests that polar UCM components

contribute to the ER agonists present in crude oil WSFs.

However, mass normalization of the data resulted in a more

even contribution from all fractions, albeit with a slight shift

toward less polar components (Fig. 3H). Components in F8 to

F10 accounted for approximately 90% of the total ER agonists

in the WSF fractions when normalized to mass, with

considerably less contribution from F1, F6, F7, F11, and

F12 (Fig. 3I). Comparison of the measured and mass-

normalized data for each of the toxicity endpoints indicated

that toxicity was dependent on the individual amount of

compounds in each fraction.

Toxicity of synthetic chemical mixtures to primary hepatocytes

The EROD activity, metabolic inhibition, and estrogenicity

were determined for each synthetic fraction (SF) mixture and a

composite mixture of these (SF1 + SF6 + SF11) to investigate

the contribution of well-known oil-related pollutants to the

overall toxicity in the complex WSF fractions (Fig. 4A to C).

None of the synthetic mixtures resulted in induction of EROD

activity (Fig. 4A), indicating that these compounds were not

responsible for the EROD activity exhibited by the WSF in F1,

F6, and F11 (Fig. 3A and B). Synthetic fraction 1 and SF11

did not cause metabolic inhibition or estrogenicity (Fig. 4B

and C). In contrast, SF6 produced metabolic inhibition and

estrogenic responses in the bioassay (Fig. 4B and C), although

with somewhat higher potency than that reported for F6 of the

WSF (Fig. 3D, E, G, and H). In all cases, the composite

mixture produced a response that was statistically comparable

to that observed for SF6. The results indicate that only

compounds present in SF6 contributed to the observed toxicity

of the crude oil WSF.

DISCUSSION

WSF fractionation and chemical analyses

Gas chromatographic analysis of the WSF total extract

(Fig. 1) revealed an extremely complex mixture of chromato-

graphically resolvable compounds and a large UCM or hump.

Fractionation by HPLC into 14 separate fractions of

increasing polarity (F1–F14) led to 70% of the material eluting

in the polar organic F11, which also contained most of the UCM.

Polar organic compounds exhibit increased aqueous solubilities

and will typically dissolve into the water column in higher

amounts than their less polar hydrocarbon homologues [27].

In many studies of crude oil WSFs, observed toxicological

responses are frequently attributed only to those compounds

that are readily resolved and identified by GC-MS (e.g.,

naphthalenes, PAHs, and phenols). These compounds are

indeed toxic, and their quantification in petroleum-contami-

nated environmental samples can be used to assess the

potential health effects on marine organisms [16,17,28,29].

These compounds were also monitored in our study, and their

gravimetric and toxicological contributions to the WSF F1 to

F14 were determined (Fig. 2). However, it has been recognized

for some time that other hydrocarbon and organic constituents

are likely contributors to crude oil toxicity [30].

Fig. 3. In vitro toxicity of the 14 high-pressure liquid chromatography (HPLC)–separated fractions of the water-soluble fraction (WSF) of crude oil:(A) aryl hydrogen receptor (AhR) agonists (2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents, TCDD-EQs) measured by induction of CYP1A-mediated activity of 7-ethoxyresorufin O-deethylase (EROD), (D) metabolic inhibition expressed as toxic units (TU 5 100/EC50), and (G) estrogenreceptor (ER) agonists (17b-estradiol equivalents, E2-EQ) measured by ER-mediated induction of the estrogenic biomarker vitellogenin.Gravimetrically normalized toxicity for the 14 HPLC-separated WSF fractions: (B) AhR agonists (TCDD-EQ), (E) metabolic inhibition (TU), and(H) ER agonists (E2-EQ). Mass normalized numbers were also expressed as percentage of the sum of all fraction toxicity: (C) AhR agonists, (F)metabolic inhibition, and (I) ER agonists. All data are expressed as mean 6 standard error of the mean (n 5 2–3). Metabol. inhibit. 5metabolic inhibition.


The resolved peaks in F1 to F7 chromatograms permitted

identification and mass determination of some compounds on

the basis of their mass spectra (Fig. 2). The observed

distributions are similar to those reported previously for other

oils (e.g., [31] and references therein). The least polar fraction

(F1) was dominated by C0–3 naphthalenes and three-ring

PAHs and their C0–3 alkylated homologues. Aromatic

hydrocarbons exhibit a much lower polarity than homologous

compounds containing NSO atoms [27], and their elution in

the least polar fractions (e.g., F1; Fig. 2) is consistent with

their physicochemical properties. The polarity of the aromatic

ring system explains why the smaller two- to four-ring

aromatics occur in the WSF [27,32]. The midpolar fractions

(F5–F7) contained further three- to five-ring alkylated PAHs,

alkylphenols, and some benzothiophenes. The occurrence of

the alkylphenols and benzothiophenes in these fractions is

consistent with the higher polarity of these compounds.

Of the remaining polar fractions, F8–F10 contained only

UCMs and no identifiable compounds, and F11 contained an

abundant UCM, of which a small contribution (,0.1% by

mass) was identified as C1–3 dibenzothiophenes. However, the

presence of a polar UCM in the chromatograms of F11 and

other polar fractions made characterization of these unknown

compounds virtually impossible in that no ‘‘clean’’ mass

spectra could be attained. A preliminary GC3GC-ToF-MS

analysis of F11 identified five different classes of monocyclic,

bicyclic, and aromatic compounds containing a sulfur–oxygen

functional group (Supporting Information, Fig. S1; http://dx.

doi.org/10.1897/08-545.S1). Within each class, a homologous

series of increasingly alkylated compounds was identified from

mass spectra exhibiting molecular ions increasing by 14 mass

units. The large numbers of isomers present for each molecular

weight indicated the presence of isomeric compounds that

exhibited different degrees of branching on the alkyl chains

(Supporting Information, Fig. S1; http://dx.doi.org/10.1897/08-

545.S1). Such complex mixtures of structurally similar

compounds exhibiting high degrees of branched alkylation

have previously been reported for the aromatic hydrocarbon

fraction of petroleum-derived UCMs [5,6,33,34]. However,

further studies are required to fully characterize and quantify

the different compounds contributing to the polar UCM [7]

and establish a structure–activity relationship for their toxic

effects.

The bioassay approach with primary fish hepatocytes

determined multiple mechanisms of toxic action and thereby

improves our understanding of the cellular toxicity of

compound classes occurring in crude oil WSFs. Alterations

in levels or activities of proteins and enzymes or both are often

regarded as the most sensitive biomarkers of exposure to

environmental pollutants [35]. Such bioassay approaches have

previously been used for assessing effects of organic toxicants

in laboratory studies with the use of single oil-related

compounds [16,17] and complex samples such as produced

water [14,15] and WSFs from oil [36]. Although such bioassays

are ideal tools for deriving toxicity potencies because of high-

throughput capacity and minute sample requirements, these

bioassays lack the complexity of intact organisms. Further-

more, the partition between the cells, the bioassay gas and

aqueous phase, and the culture wells might not necessarily

predict the behavior of organic compounds in vivo. This could

potentially lead to limited in vitro to in vivo predictability for

compounds that are volatile (e.g., phenol and naphthalenes) or

highly hydrophobic (e.g., high–molecular weight PAHs) [16].

However, fish cells have been used successfully to determine

the effects of oil-related compounds elsewhere [16,17,37].

Most fractions, even those containing very small amounts

of material (e.g., F2–F4 and F8–F10), appeared to contain

AhR agonists. The more even distribution of toxicity across all

of the fractions after normalization (Fig. 3B) verified that

most fractions contained AhR agonists in similar quantities.

Interestingly, F10 accounted for almost 40% of the total WSF

bioactivity (Fig. 3C), although the presence of a UCM meant

that less than 0.2% of the compounds could be identified. This

suggests that a common structural feature (likely to be the

polyaromatic rings of PAHs and similar compounds) was

present in most of the fractions. On the basis of previous

studies showing induction of AhR-mediated EROD activity in

fish ([35] and references therein) and fish cell cultures [15,36],

the aromatic hydrocarbons present in F1 could be expected to

cause the observed increased EROD activity in this study. Less

information exists for polar NSO compounds, but a study by

Jung et al. [38] observed CYP1A induction in fish hepatoma

cells by nitro-PAHs and nitro-heterocyclic aromatic hydrocar-

bons. Such compounds could be present in the polar fractions

(e.g., F10 and F11), but they could not be identified with the

current chemical analytical approach because of the unre-

solved nature of the UCMs. Interestingly, the synthetic

mixture comprising known aromatic compounds (SF1;

Fig. 4A) did not cause increased EROD activity, even though

certain compounds (e.g., methylanthracene and chrysene) were

Fig. 4. Toxicity of synthetic fraction (SF) mixtures (SF1, SF6, and SF11) in a fish in vitro bioassay when exposed individually and in combination:(A) aryl hydrogen receptor (AhR) agonists (2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents, TCDD-EQ) measured by induction of CYP1A-mediated activity of 7-ethoxyresorufin O-deethylase (EROD), (B) metabolic inhibition expressed as toxic units (TU 5 100/EC50), and (C) estrogenreceptor (ER) agonists (17b-estradiol equivalents, E2-EQ) measured by ER-mediated induction of the estrogenic biomarker vitellogenin. Data areexpressed as mean 6 standard error of the mean (n 5 3). ND 5 not detected.


present at sufficient concentrations to theoretically cause

EROD activation in the bioassay. Therefore we suggest that

the lack of induction could be caused by the simultaneous

presence of EROD inducers (e.g., low–molecular weight PAHs

such as methylated phenanthrenes) and potential EROD

inhibitors (e.g., high molecular weight PAHs and alkylphe-

nols). As expected, SF11 did not generate a response in this or

the other bioassay studies. The low concentration of the

dibenzothiophenes in the synthetic mixture was probably

below the concentrations required to elicit a response in the in

vitro bioassay.

Metabolic inhibition (measured as reduction in cellular

metabolic activity) was detected in F1 and F5 to F11

(Fig. 3D), with F1 and F11 giving the greatest responses.

Mass normalization of the data (Fig. 3E) revealed that the

responses of F1 and F11 were lower than some of the other

fractions (e.g., F5–F10) when present at equal concentrations.

This indicates that a high concentration of less toxic

compounds in the WSF can be of larger toxicological relevance

than low concentrations of presumably more toxic com-

pounds. The change in response when the data are normalized

also suggests that toxicity is the result of a more unspecific

mode of action, consistent with the suggestions that most

organic chemicals act through narcosis or polar narcosis [39].

As expected, the synthetic mixture SF6, but not SF1, induced

metabolic inhibition (Fig. 4B), which is in agreement with

findings that polar compounds such as alkylphenols elicit

cytotoxicity in the fish in vitro bioassay [16]. Interestingly, the

more hydrophobic aromatic hydrocarbons in SF1 do

not appear to be responsible for the metabolic inhibition

observed in F1 of the WSF, thus suggesting that other

(probably unresolved) compounds present in these fractions

were responsible [5,6], that the concentrations of the com-

pounds occurring in the synthetic mixture were not high

enough to cause toxicity, or that bioassay-specific partitioning

factors underestimated the toxic potency of these constituents

in SF1.

Estrogenicity (assessed in terms of VTG induction) was

detected in F1 and F6 to F12 (Fig. 3C), with F11 apparently

leading to the greatest responses. Following mass normaliza-

tion, a more even distribution of toxicity across all of the

fractions was observed, but with the highest responses

corresponding to the polar fractions (F8–F10; Fig. 3F). In

fact F8 to F10 accounted for approximately 90% of the total

ER agonists in the WSF fractions when normalized to mass

(Fig. 3I). These findings are consistent with the presence of

polar ER agonists such as alkylphenols in these fractions.

Structure–activity relationship studies suggest that most

alkylphenols with straight and branched alkylated chains

ranging from C1 to C8 can bind to the rainbow trout estrogen

receptor and activate the production of VTG in vitro [17]. The

association of normal and branched C1 to C8 alkylphenols

with estrogenicity of oil-related discharges has been amply

demonstrated [20]. However, it is suggested that the significant

responses observed in the other fractions indicate that other

compounds in the oil could also elicit estrogenic effects. The

contribution of polar alkylphenols to estrogenicity was further

supported by the induction of VTG in hepatocytes after

exposure to the synthetic mixture SF6 containing C1 to C5

alkylphenols. A lack of response in less polar fractions, such as

the synthetic mixture SF1 (Fig. 4C), indicated that compounds

other than those identified in the apolar F1 of the WSF might

be responsible.

Comparison of the measured and mass normalized data for

each of the toxicity endpoints indicated that toxicity was

dependent on the individual amount of compounds in each

fraction. These findings are consistent with a nonspecific mode

of action across all the fractions. Although the synthetic

mixtures SF1 and SF11 did not produce toxic responses in any

of the endpoints studied, it is still possible that the compounds

present contribute additively to the observed toxicity in the oil

fractions. Because the total concentration of synthetic com-

pounds in SF1 and SF11 was only 16 and 0.1% of the total

hydrocarbon mass, respectively, this amount might have been

insufficient to elicit the response observed with the oil samples.

We suggest that the remaining uncharacterized compounds in

the oil fractions were largely responsible for the observed

toxicity.

The source of the crude oil and the method of WSF

preparation dictate the component composition of WSFs [40].

As a result, the high oil–water loading used might favor

increased dissolution of the more polar constituents, thus

potentially giving rise to enrichment of the polar compounds in

the WSF as proposed by Shiu et al. [40]. Although feasible as

screening tools, use of in vitro bioassays might not accurately

predict the potential effects to fish in vivo, and extrapolation

to fish population effects might be uncertain. However, this

study offers a novel approach for performing chemical and

toxicological characterization of complex mixtures. Future

studies should focus on studying the different factors affecting

WSF composition and the extrapolation of in vitro data to in

vivo bioactivity in order to perform thorough studies to

resolve complex samples such as WSFs from UCM-rich oils.

CONCLUSIONS

In this study, a single in-reservoir biodegraded Norwegian

Sea crude oil was used to generate a WSF at a high oil–water

loading ratio (10,000 mg/L) representing a large spill event.

Under these conditions, the data indicate that the WSF was

dominated by the more polar petroleum compounds. Al-

though the high oil–water loading ratio used might have

favored increased dissolution of the more polar constituents,

they accounted gravimetrically for more than 70% of the WSF.

Furthermore, these polar compounds are present as an UCM

that makes identifying and characterizing them extremely

difficult; comprehensive GC3GC-ToF-MS promises to ease

this situation somewhat, but interpretation of the mass spectra

of unknowns can still be time consuming and requires

considerable expertise [5,6,33,34]. Although polar compounds

exert significant in vitro effects, normalization of the data to

equal concentrations showed that they were not necessarily the

most toxic compounds present in the WSF. For a crude oil

such as that investigated here, a large spill event could result in

the concentration of polar compounds (including polar UCM

compounds), reaching much higher levels in the water column

than that of other crude oil compounds (e.g., apolar

hydrocarbons). Driven by a combination of high oil–water

loading and increased aqueous solubility (compared with

hydrocarbon homologues), polar UCM compounds could

become the most toxicologically important in the WSF. Until

now, most environmental monitoring has focused on aromatic

hydrocarbons (e.g., PAHs), and this study shows such

compounds might be less abundant in the WSF than the

polar UCM components. Future research should be directed

toward the chemical characterization, environmental fate and

persistence, and toxicological effects on marine organisms of


UCMs originating from a variety of commercially relevant

crude oils and different oil–water loadings.

SUPPORTING INFORMATION

Table S1. Concentrations (mg ml21) of decalins, naphtha-

lenes, polycyclic aromatic hydrocarbons (PAHs), and phenols

in synthetically prepared mixtures representing the identifiable

compounds in fractions 1, 6, and 11 from North Sea crude oil

water-soluble fraction (WSF). SF 5 synthetic fraction.

Fig. S1. Structures of the five alkylated sulfoxide compound

groups identified in an in-reservoir biodegraded North Sea

crude oil water-soluble fraction (WSF) with the use of

comprehensive two-dimensional gas chromatography–time of

flight–mass spectrometry (GC3GC-ToF-MS). All groups

contain both hexa- and penta-cyclic rings and exhibit

increasing degrees of alkylation.

All found at DOI: 10.1897/08-545.S1 (72 KB PDF).

Acknowledgement—The authors thank the Norwegian ResearchCouncil for providing financial support to this project (173451). Wethank A.C. Lewis (University of York) for access to GC3GC-ToF-MS and Olav Bøyum, Eivind Farmen Finne, Oscar Fogelberg, Anja J.Nilsen, and Ase Bakketun for their assistance in bioassay testing.Finally, we thank Kjersti Almas and Inger Steinsvik for theirassistance with the chemical analyses.

REFERENCES

1. Gough MA, Rowland SJ. 1990. Characterisation of unresolvedcomplex mixtures of hydrocarbons in petroleum. Nature 344:648–650.

2. Rowland S, Donkin P, Smith E, Wraige E. 2001. Aromatichydrocarbon ‘‘humps’’ in the marine environment: Unrecognizedtoxins? Environ Sci Technol 35:2640–2644.

3. Smith E, Wraige E, Donkin P. Rowland S. 2001. Hydrocarbonhumps in the marine environment: Synthesis, toxicity and aqueoussolubility of monoaromatic compounds. Environ Toxicol Chem 20:2428–2432.

4. Donkin P, Smith EL, Rowland SJ. 2003. Toxic effects ofunresolved complex mixtures of aromatic hydrocarbons accumu-lated by mussels, Mytilus edulis, from contaminated field sites.Environ Sci Technol 37:4825–4830.

5. Booth AM, Sutton PA, Lewis CA, Lewis AC, Scarlett A, Chau W,Widdows J, Rowland SJ. 2007. Unresolved complex mixtures ofaromatic hydrocarbons: Thousands of overlooked persistent,bioaccumulative, and toxic contaminants in mussels. Environ SciTechnol 41:457–464.

6. Booth AM, Scarlett AG, Lewis CA, Belt ST, Rowland SJ. 2008.Unresolved complex mixtures (UCMs) of aromatic hydrocarbons:Branched alkyl indanes and branched alkyl tetralins are present inUCMs and accumulated by and toxic to, the mussel Mytilus edulis.Environ Sci Technol 42:8122–8126.

7. Scarlett A, Rowland SJ, Galloway TS, Lewis AC, Booth AM.2008. Chronic sublethal effects associated with branched alkyl-benzenes bioaccumulated by mussels. Environ Toxicol Chem 27:561–567.

8. Scarlett A, Canty MN, Smith EL, Rowland SJ, Galloway TS.2007. Can amphipod behavior help to predict chronic toxicity ofsediments? Human Ecol Risk Assess 13:506–518.

9. Booth AM, Aitken C, Jones DM, Lewis CA, Rowland SJ. 2007.Resistance of toxic alkylcyclohexyltetralins to biodegradation byaerobic bacteria. Organic Geochem 38:540–550.

10. Holowenko FM, MacKinnin MD, Fedorak PM. 2002. Charac-terization of naphthenic acids in oil sands wastewaters by gaschromatography–mass spectrometry. Water Res 36:2843–2855.

11. Burwood R, Speers GC. 1974. Photo-oxidation as a factor in theenvironmental dispersal of crude oil. Estuar Coast Mar Sci 2:117–135.

12. Booth AM. 2004. Biodegradation, aqueous solubility and charac-terization studies of unresolved complex mixtures (UCMs) ofaromatic hydrocarbons. PhD thesis. University of Plymouth,Plymouth, Devon, UK.

13. Thomas KV, Balaam J, Barnard N, Dyer R, Jones C, Lavender J,McHugh M. 2002. Characterisation of potentially genotoxiccompounds in sediments collected from United Kingdom estuar-ies. Chemosphere 49:247–258.

14. Tollefsen K-E, Finne EF, Romstad R, Sandberg C. 2006. Effluentsfrom oil production activities contain chemicals that interfere withnormal function of intra- and extra-cellular estrogen bindingproteins. Mar Environ Res 62(Suppl):191–194.

15. Tollefsen K-E, Goksøyr A, Hylland K. 2006. Assessment ofcytotoxic, CYP1A inducing and oestrogenic activity in watersfrom the German Bight and the Statfjord area of the North Sea bya suite of fish in vitro bioassays. In Hylland K, Lang T, VethaakD, eds, Biological Effects of Contaminants in Marine PelagicEcosystems, 1st ed. SETAC, Brussels, Belgium, pp 385–395.

16. Tollefsen K-E, Blikstad C, Eikvar S, Finne EF, Gregersen IK.2008. Cytotoxicity of alkylphenols and alkylated non-phenolics ina primary culture of rainbow trout (Oncorhynchus mykiss)hepatocytes. Ecotoxicol Environ Saf 69:6–73.

17. Tollefsen K-E, Eikvar S, Finne EF, Fogelberg O, Gregersen IK.2008. Estrogenicity of alkylphenols and alkylated non-phenolics ina rainbow trout (Oncorhynchus mykiss) primary hepatocyteculture. Ecotoxicol Environ Saf 71:370–383.

18. Tollefsen K-E, Harman C, Smith A, Thomas KV. 2007. Estrogenreceptor (ER) agonists and androgen receptor (AR) antagonists ineffluents from Norwegian North Sea production platforms. MarPollut Bull 54:277–283.

19. Safe SH. 1998. Hazard and risk assessment of chemical mixturesusing the toxic equivalency factor approach. Environ HealthPerspect 106:1051–1058.

20. Thomas KV, Balaam J, Hurst MR, Thain JE. 2004. Bio-analyticaland chemical characterisation of offshore produced water effluentsfor estrogen receptor (ER) agonists. J Environ Monit 6:593–598.

21. Singer MM, Aurand D, Bragin GE, Clark JR, Coelho GM, SowbyML, Tjeerdema RS. 2000. Standardization of the preparation andquantification of water-accommodated fractions of petroleum fortoxicity testing. Mar Pollut Bull 40:1007–1016.

22. Norwegian Oil Industry Association. 2003. Recommended guide-lines for the sampling and analysis of produced water. Guideline085. Stavanger, Norway.

23. Tollefsen K-E, Mathisen R, Stenersen J. 2003. Induction ofvitellogenin synthesis in an Atlantic salmon (Salmo salar)hepatocyte culture: A sensitive in vitro bioassay for the oestrogenicand anti-oestrogenic activity of chemicals. Biomarkers 8:394–407.

24. Ganassin RC, Schirmer K, Bols NC. 2000. Cell and tissue culture.In Ostrander GK, ed, The Laboratory Fish. Academic, San Diego,CA, USA, pp 631–651.

25. Schirmer K, Chan AGJ, Bols NC. 2000. Transitory metabolicdisruption and cytotoxicity elicited by benzo[a]pyrene in two celllines from rainbow trout liver. J Biochem Mol Toxicol 14:262–276.

26. Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. 1951. Proteinmeasurement with the folin phenol reagent. J Biol Chem 193:265–275.

27. Schwarzenbach RP, Gschwend PM, Imboden DM. 2003. Envi-ronmental Organic Chemistry. John Wiley, Hoboken, NJ, USA.

28. Hutchinson TH, Field MDR, Manning MJ. 2003. Evaluation ofnon-specific immune functions in dab, Limanda limanda L.,following short-term exposure to sediments contaminated withpolyaromatic hydrocarbons and/or polychlorinated biphenyls.Mar Environ Res 55:193–202.

29. Page DS, Boehm PD, Stubblefield WA, Parker KR, Gilfillan ES,Neff JM, Maki AW. 2002. Hydrocarbon composition and toxicityof sediments following the Exxon Valdez oil spill in Prince WilliamSound, Alaska, USA. Environ Toxicol Chem 21:1438–145.

30. Neff JM, Ostazeski S, Gardiner W, Stejskal I. 2000. Effects ofweathering on the toxicity of three offshore Australian crude oilsand diesel fuel to marine animals. Environ Toxicol Chem 19:1809–1821.

31. Ali L. 1995. The dissolution and photodegradation of Kuwaitcrude oil in seawater. PhD thesis. University of Plymouth,Plymouth, Devon, UK.

32. McAuliffe C. 1966. Solubility in water of paraffin, cycloparaffin,olefin, acetylene, cycloolefin, and aromatic hydrocarbons. J PhysChem 70:1267–1275.

33. Ventura GT, Kenig F, Reddy CM, Frysinger GS, Nelson RK, VanMooy B, Gaines RB. 2008. Analysis of unresolved complexmixtures of hydrocarbons extracted from Late Archean sediments


by comprehensive two-dimensional gas chromatography(GC3GC). Organic Geochem 39:846–867.

34. Reddy CM, Eglinton TI, Hounshell A, White HK, Xu L, GainesRB, Frysinger GS. 2002. The West Falmouth oil spill after thirtyyears: The persistence of petroleum hydrocarbons in marshsediments. Environ Sci Technol 36:4754–4760.

35. Van der Oost R, Beyer J, Vermeulen NPE. 2003. Fishbioaccumulation and biomarkers in environmental risk assess-ment: A review. Environ Toxicol Pharmacol 13:57–149.

36. Navas JM, Babin M, Casado S, Fernandez C, Tarazona JV. 2006.The Prestige oil spill: A laboratory study about the toxicity of thewater-soluble fraction of the fuel oil. Mar Environ Res 62(Suppl):S352–S355.

37. Schirmer K, Chan AG, Greenberg BM, Dixon DG, Bols NC.1998. Ability of 16 priority PAHs to be photocytotoxic to a cellline from the rainbow trout gill. Toxicology 127:143–155.

38. Jung DKJ, Klaus T, Fent K. 2001. Cytochrome P450 induction bynitrated polycyclic aromatic hydrocarbons, azaarenes, and binarymixtures in fish hepatoma cell line PLHC-1. Environ Toxicol Chem20:149–159.

39. Jaworska JS, Schultz TW. 1994. Mechanism-based comparisons ofacute toxicities elicited by industrial organic chemicals in procaryoticand eucaryotic systems. Ecotoxicol Environ Saf 29:200–213.

40. Shiu WY, Bobra M, Bobra AM, Maijanen A, Sunito L, MackayD. 1990. The water solubility of crude oils and petroleumproducts. Oil Chem Pollut 7:57–84.


Chemical and toxicological characterization of the bricks produced from clay/sewage sludge mixture

Documents