Ferreira M. M. C., Kiralj R., "CHEMOMETRIC INVESTIGATIONS OF MULTIDRUG RESISTANCE IN STRAINS OF THE PHYTOPATHOGENIC FUNGUS PENICILLIUM DIGITATUM". Moscow, Russia, 01-05/09/2007: Fourth International Symposium on Computational Methods in Toxicology and Pharmacology Integrating Internet Resources, Book of Abstracts (2007) 67. Oral OC-22.

OC-22
_____________________________________________________________________________________

CHEMOMETRIC INVESTIGATIONS OF  MULTIDRUG RESISTANCE  IN STRAINS
OF THE PHYTOPATHOGENIC FUNGUS PENICILLIUM DIGITATUM

Márcia M. C. Ferreira, Rudolf Kiralj

Instituto de Química, Universidade Estadual de Campinas,  Campinas 13083-862,  SP,  Brazil;
E-mail: marcia@iqm.unicamp.br

Demethylation   inhibitor  (DMI)  resistance   by   pathogenic  fungi   is  a  serious  problem  in
agriculture and medicine. P. digitatum (the green mold) causes important postharvest  diseases
of  citrus fruits.  The  present work  studies  P. digitatum  strains  (DMI-resistant,  moderately
resistant  and  sensitive)  and  4  DMIs  and  3  non-DMIs  by means of  Principal Component
Analysis  (PCA)  and  Hierarchical Cluster Analysis  (HCA)  and  Partial Least Squares  (PLS).
Novel types of relationships between toxicant structure and fungal resistance,  and between the
resistance and fungal genome,  were established.
Biological  activity  datasets  for P. digitatum   strains  with   respect   to  DMIs  (triflumizole,
fenarimol,  bitertanol,  pyrifenox)   and   non-DMIs  (cycloheximide,  4-nitroquinoline-N-oxide,
acriflavine)  were  generated  from  experimental  EC₅₀  (effective inhibitory concentration  for
50% radial growth inhibition) and  MIC  (minimal inhibitory concentration)  from literature [1].
The  datasets  contained:   pEC₅₀ = -log(EC₅₀ / mol dm^-3);   pECr₅₀ = pEC₅₀(standard) – pEC₅₀;
descriptors (a, b, c, |a|, |c|)  from regressions pMIC=a+b pEC₅₀ and c=a/b for each toxicant;  8
and 16 morphological  descriptors  (radii, circumferences and areas of fungal cultures)  for  the
growth   of   39   strains   without   and   with   toxicant,  respectively;   pEC₅₀   from   multiple
measurements for  DMIs,  non-DMIs  and   all  7  substances.  Genome  structure  descriptors
related to constitutive and toxicant-induced expression levels of  CYP51  and  PMR1  genes  in
diverse  P. digitatum  strains   were   generated,   as  well  as  their  products  with   molecular
descriptors  for  4 DMIs.  The  descriptors  were  correlated   with   the  corresponding  pEC₅₀
activity  at  PLS  level,   applying    leave-one-out   crossvalidation   and   external   validation.
Novel  Activity-Structure  Relationships  (ASRs)  resulted  from  exploratory  analyses  of  the
activity  datasets.   Relationships  between  toxicants  structure   and  strain  features  (baseline
resistance, morphology, origin/target)  are  visible,  can be rationalized and used in  predictions.
Quantitative  Genome-Activity Relationship   (QGAR)   and   Quantitative  Genome/Structure-
Activity Relationship (QGSAR)  are novel relationships between  pEC₅₀  and  genome/toxicant
descriptors,  with s atisfactory  PLS  statistics   (QGAR: 3PCs,  R=0.90,  Q=0.89,  SEV=0.34;
QGSAR:  5PCs,R=0.93, Q=0.92, SEV=0.286).  The primary  contribution to  DMI resistance
comes  from  the  CYP51  gene,  secondary  from the  PMR1  gene  and   its  interaction  with
toxicants.  QGSAR  and  QGSAR  can aid in detecting resistance strains of  P. digitatum  and
development of novel antifungals.
The authors thank to FAPESP.

[1] Nakaune, R. et al., Mol. Genet. Genom. 2002, 267, 179 and references cited therein.
 

_____________________________________________________________________________________
CMTPI-200767