Kiralj R., Ferreira M. M. C., “Chemometric and Molecular Modeling Study of 1H-Indol-3-acetid Acid Derivatives with Auxin Activity”. Mariehamn, Åland Islands, Finland, 14-18/06/2003: 8th Scandinavian Symposium on Chemometrics (SSC8). Book of Abstracts, P16 (2003). Poster P16.
Chemometric
and Molecular Modeling Study of 1H-Indole-3-acetic Acid Derivatives
with
Auxin Activity
Rudolf Kiralj and Márcia M. C. Ferreira
Instituto de Química, Universidade Estadual de Campinas, Campinas
SP, 13083-970, Brazil
1H-Indole-3-acetic
acid (IAA) and its derivatives are among
the most important plant-growth
regulation hormones.
Quantitative Structure-Activity Relationships (QSAR) study on auxins are
not
straitforward, mainly
due to the lack of 3D structure of auxin binding protein (ABP) which has
been
determined just
recently, and intrinsic auxin properties.
The skeleton of IAA consists of
heteroaromatic, planar
indole ring, and the acetate group which can adopt a few distinct orientations
with respect
to the ring. IAA auxins are limited
to derivatives with small substituents at
five
substitution
positions at the indole ring.
This work
deals with over twenty IAA auxins,
half of which has measured straight-growth
promoting activities
on Avena Sativa L. coleoptiles1: the optimal
coleoptile elongation
L, and the
half-optimum concentration
c,
both relative to IAA. Molecular geometries for all compounds were
optimized at ab
initio B3LYP 6-31G** level, and topological, electrotopological, steric,
electronic
and lipophilicity
molecular descriptors were calculated. Selected descriptors
were analyzed by
means of Hierarchical
Cluster Analysis (HCA) and Principal Component Analysis (PCA).
Partial
Least Squares
(PLS) regression models were built and validated based on these
descriptors, for
estimation of
the biological activities logL and
logc. ABP – auxin complexes for all
studied
molecules were
built from ABP – 1-naphthalenic acid complex
crystal structure2, and their
geometries optimized
by using molecular mechanics MMFF94 force
field.
Five descriptors
were sufficient for QSAR analysis of
activity logL: lipophilicity descriptor
(logKOW),
molar Gibbs free energy of formation, a radial distribution function descriptor,
molecular
electrostatic variation,
and a Broto-Moreau autocorrelation descriptor. The PLS model (Q2
= 0.73,
R2
> 0.87) with four principal components predicted well activities of non-halogenated
compounds.
Plant-growth
activity logc showed to be related only to three
descriptors in non-linear form: a
Randic molecular profile,
a 3D-MORSE descriptor and another radial distribution
function. The
PLS model
(Q2 = 0.91, R2 = 0.95) based
on only one PC could predict the activities even
for
halogenated compounds.
HCA and PCA analyses showed the clustering of the auxins into
classes
with different activities.
QSAR, HCA and PCA results are interpretable
in terms of molecular
graphics and
modeling of ABP – auxin complexes.
1 Nigovic,
B.; Antolic, S.; Kojic-Prodic, B.; Kiralj, R.; Magnus, V.; Salopek-Sondi,
B. Acta Cryst.,
2000, B56,
94-111.
2 Woo,
E.-J.; Marshall, J.; Bauly, J.; Chen, J.-G.; Venis,
M.; Napier, R. M.; Pickersgill, R. W.
EMBO J., 2002,
21,
2877-85.
Acknowledgement: The
authors acknowledge FAPESP for support.