additional parameters in the nonlinear regression. The joint
PKPD model was used with nonlinear regression for fitting
all 92 observations. The former approach did not lead to
stable parameter estimates for our data when the k2 was
restrained to stay greater 0, while the latter approach con-
verged and produced reasonable parameter estimates.
Acknowledgments
R.A. acknowledges the receipt of an OECD research fellowship
to Roswell Park Cancer Institute, Buffalo. For constructive
criticism, we thank three anonymous reviewers. For skillful
technical support, thanks are due to Silke Aulhorn, Iris
Christmann, and Janet Krueger.
Table 2 provides the parameter estimates for the different
models with error estimates, skewness, and Akaike informa-
tion criterion, AICc, to characterize and compare the different
nonlinear model approaches. The parameter θ1 is the EC50
at infinite time, θ3,4,5 represent maximal and minimal frac-
tional effect and the slope of the concentration effect func-
tions, while the meaning of θ2 is model specific. The actual
estimates for θ3,4,5 are pretty similar for the different models
which shows that all data are used to drive the functions.
The skew measure indicates that the models vary greatly in
their ability to fit the observations, as Skew values greater
than (0.25 indicate apparent nonlinearity and greater than
(1 indicate that there is considerable nonlinearity in the
parameter (19). While the former indicates only problems
with the parameter estimates, the Akaike information cri-
terion (AICc) allows direct comparison between the models
(38). The lower the AICc value, the better the goodness of fit
of a model. While the critical target occupation model
achieves an AICc value similar to the empirical modified Hill
model, the critical body burden model is clearly less well
capturing the data.
Supporting Information Available
A summary of basic properties of the PNA derivatives
investigated, a more detailed description of the bioassays
used, and a figure displaying the findings on effect modifica-
tion with different co-exposure to UV light. This material is
Literature Cited
(1) Brack, W.; Altenburger, R.; Ensenbach, U.; Mo¨der, M.; Segner,
H.; Schu¨u¨rmann, G. Bioassay-directed identification of organic
toxicants in river sediment in the industrial region of Bitterfeld
(Germany) - A contribution to hazard assessment. Arch. Environ.
Contam. Toxicol. 1999, 37, 164-174.
(2) Winder, C.; Balouet, J. -C. The toxicity of commercial jet oils.
Environ. Res. A. 2002, 89, 146-164.
(3) BUA (Beratergremium fuer umweltrelevante Altstoffe (BUA) der
Gesellschaft Deutscher Chemiker). N-Phenyl-1-naphthylamin;
S.Hirzel Wissenschaftliche Verlagsgesellschaft, Stuttgart: 1993;
p 84.
(4) Jungclaus, G. A.; Lopez-Avila, V.; Hites, R. A. Organic Compounds
in an Industrial Wastewater: A Case Study of Their Environ-
mental Impact. Environ. Sci. Technol. 1978, 12, 88-96.
(5) Lopez-Avilla, V.; Hites, R. A. Organic compounds in an industrial
wastewater. Their transport into sediments. Environ. Sci.
Technol. 1980, 14, 1382-1390.
(6) IPCS (international Programme on Chemical Safety) N-Phenyl-
1-Naphthylamine; WHO: Geneva, 1998.
(7) Wang, H.; Wang, D.; Dzeng, R. Carcinogenicity of N-phenyl-
1-naphthylamine and N-phenyl-2-naphthylamine in mice.
Cancer Res. 1984, 44, 3098-3100.
(8) Greenhouse, G. A. Toxicity of N-phenyl-R-naphthylamine and
hydrazine to Xenopus laevis Embryos and Larvae. Bull. Environ.
Contam. Toxicol. 1977, 18, 503-511.
(9) Altenburger, R.; Walter, H.; Grote, M. What contributes to the
combined effect of a complex mixture? Environ. Sci. Technol.
2004, 38, 6353-6362.
(10) US EPA. EPI Suite V. 3.10, Syracuse Research Cooperation,
edl.htm, 2000.
(11) International Organization for Standardization. Water qualitys
Determination of the inhibitory effect of water samples on the
light emission of Vibrio fischeri (Luminescent bacteria test),
Part 1: Method using freshly prepared bacteria, ISO 11348-1;
ISO: Geneva, 1998.
(12) International Organization for Standardization. Water qualitys
Determination of the inhibition of the mobility of Daphnia magna
Straus (Cladocera, Crustacea)sAcute toxicity test, ISO 6341;
ISO: Geneva, 1996.
(13) Klu¨ttgen, B.; Du¨lmer, U.; Engels, M.; Ratte, H. T. ADaM, an
artificial freshwater for the culture of zooplankton. Water Res.
1994, 28, 743-746.
(14) Deutsches Institut fu¨r Normung e. V. German standard methods
for the examination of water, waste water and sludgesSub-
animal testing (group T) sPart 6: Toxicity to fish; Determination
of the non-acute-poisonous effect of waste water to fish eggs by
dilution limits (T 6), DIN 38415-6; DIN: Berlin, 2003.
(15) International Organization for Standardization. Water qualitys
Freshwater algal growth inhibition test with unicellular green
algae, ISO 8692; ISO: Geneva, 2004.
(16) Schu¨u¨rmann, G.; Somashekar, R. K.; Kristen, U. Structure-activity
relationships for chloro- and nitrophenol toxicity in the pollen
tube growth test. Environ. Toxicol. Chem. 1996, 15, 1702-1708.
(17) Jung, K.; Kaletta, K.; Segner, H.; Schu¨urmann, G. 15N Metabolic
test for the determination of phytotoxic effects of chemicals
and contaminated environmental samples. Environ. Sci. Pollut.
Res. 1999, 6, 72-76.
The inset of Figure 3 displays these findings by using the
models to simulate the time dependence of the EC50. The
good fit of the critical target occupation model as opposed
to the model of critical body burden suggests that, indeed,
irreversible activity may explain the time course of the
observed effects better than the assumption of an internal
critical threshold. One may think of this not only as irre-
versible receptor binding, which should lead to short-term
visible effects, but also in terms of continuously formed
reactants (21) that cumulatively lead to damage. Also, the
tendency that k2-values which were estimated to be sub-
stantially lower than those deduced from ref 37 greatly
improved the fitting capabilities of the critical body burden
model, point to a possible role of elimination mechanisms
in understanding PNA phytotoxicity.
From the results in different organism biotests shown
above we conclude that N-phenyl-2-naphthylamine exhibits
properties that lead to high phytotoxicity, which seem most
prominent for PNA but may be relevant for derivatives also.
Our investigation did not reveal a single target or plant-
specific process as specifically susceptible to PNA exposure.
Rather, primary photosynthetic reactions and subsequently
growth and reproduction cease slowly with time and fast
with an increase in concentration. Tentatively, we think of
it as a reactive toxicity exhibited by PNA, leading to irreversible
and thus cumulative damage. In mammals, PNA is reported
to be metabolized to epoxides and hydroxylated derivatives
as well as N-dephenylated to the carcinogenic compound
2-naphthylamine. The metabolism is catalyzed by the cyto-
chrome system and the prostaglandin endoperoxide syn-
thetase (39). The metabolites are known to attack nucleophilic
biomolecules such as DNA and proteins (40). Due to its high
lipophilicity, the intracellular distribution of PNA will be
favored in membranes, and therefore, first effects like those
observed for photosynthetic reactions might be expected to
occur in membrane-rich compartments such as the algal
chloroplast. If this holds, distinguishable responses for various
parameters in different organisms might depend on a species
capacity to metabolize the original compound into less toxic
ones or to repair damaged cellular components and functions.
Whether this actually happens should be checked in bioassays
using prolonged analytically monitored exposure.
(18) Scholze, M.; Boedeker, W.; Faust, M.; Backhaus, T.; Altenburger,
R.; Grimme, L.H. A general best-fit method for concentration-
response curves and the estimation of low-effect concentrations.
Environ. Toxicol Chem. 2001, 20, 448-457.
(19) SAS. SAS/STAT User’s Guide, V. 8; SAS Institute Inc.: Cary, NC,
1999; p 3884.
9
6168 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006