Ferreira M. M. C., “POLYCYCLIC AROMATIC HYDROCARBONS. PREDICTING PHOTOINDUCED TOXICITY AND PHYSICOCHEMICAL PROPERTIES”. Düsseldorf, Germany, 27/08-01/09/2000: 13th European Symposium on Quantitative Structure-Activity Relationships: Rational Approaches to Drug Design, Book of Abstracts, P.62 (2000). Poster P62. Section: New Developments in Chemometrics.

    POLYCYCLIC AROMATIC HYDROCARBONS. PREDICTING PHOTOINDUCED
                         TOXICITY AND PHYSICOCHEMICAL PROPERTIES

                                                    Márcia M. C. Ferreira
                   Instituto de Química - UNICAMP, 13083-970 ,Campinas – SP BRAZIL
                                                     marcia@iqm.unicamp.br

Polycyclic Aromatic Hydrocarbons (PAHs) have been for a long time a focus of great attention  by
the  scientific  community  due  to  their  impact  on  health   and  environment.   Some   of   these
compounds such as  benz[a]anthracene, chrysene, d ibenz[a,h]anthracene and  benzo[a]pyrene are
mutagens and carcinogens. Usually,  the  PAHs are introduced into the environment as a  result  of
anthropogenic activities which have increased dramatically in  the last  20 years.  PAHs  have  been
detected  in  the  atmosphere,  waters,soils, sediments and food.  Their  physicochemical properties
such  as aqueous solubility, volatility (e. g.  Henry’s  law  constant)  and   hydrophobicity  (e. g.  n-
octanol/water partition coefficient) vary widely. They are also quite involatile, show  relatively  low
vapor  pressure  and resistance  to  chemical  reactions.  As  a  result,  their  distribution  in aquatic
systems, the atmosphere  and soil  differ significantly.  Besides  that,  PAHs  show  a  tendency  to
accumulate in biota, soils, sediments, and to be transported through the atmosphere far away  from
their  source  discharge.
Consequently,  meaningful  health  and environmental  assessment  requires  reliable  data  on  their
physicochemical properties.  Unfortunately,  very often  these information  are  not available  in  the
literature,  and  frequently  the  reported  data  have  poor  precision  and accuracy  specially  when
dealing with hydrophobic and involatile chemicals. In  such conditions, multivariate  regression can
be  of  great  help since  relevant  properties  are modeled  and  predicted as  functions of  molecular
structure  parameters.   Methods  based  on   principal  component  analysis,   such   as   PLS,   are
advantageous  by  dealing  well  with  strongly  correlated  descriptors.
This  work  focuses  on  a subgroup  of  48 PAHs  composed  exclusively  by  nonnsubstituted six-
membered  aromatic rings. In  a first step, physicochemical  properties  relevant  for  environmental
studies   such  as   boiling   temperature  (Tb),   retention   index   (RI),   n-octanol/water   partition
coefficient  (Kow)  and   solubility  (S)  are  related  with   a  series  of  electronic,  geometric   and
topological  descriptors.  Among  them  are:  ionization  potential,  electron  affinity,  the  difference
between   electron  affinity  and   ionization  potential   (GAP),  Wiener  and  connectivity   indexes,
molecular volume, surface area, length-to-breath ratio and enthalpy of formation. In a second  step,
these   properties  were   incorporated  into   the  descriptor  matrix   to  build  several   quantitative
structure-property relationships and to obtain prediction rules for  the soil  sorption  coefficient  (log
Koc), octanol-air partitioning coefficient (KoA), bioconcentration factor (BCF) and Henry’s law
constant  (H)  or  air-water  partitioning  coefficient.
Finally, the  photoinduced  toxicity  for  the  freshwater  rganism  Daphinia-Magna  is modeled  by
using  the  electronic descriptors  calculated for  the  ground  states. (FAPESP)