Objective
Benzoylphenylurea (BPU) insect growth regulators have been used in a wide range of insect control applications. The number of BPU pesticides in use has rapidly expanded, due largely to their high degree of selectivity and low mammalian toxicity.
Predictive software utilizing algorithms developed using molecular structure and physical property correlations can be an extremely useful tool for environmental chemists. A number of commercial products for physical property prediction exist. These include boiling point, vapor pressure, LogP (Kow), pKa, LogD and toxicity predictors. Such programs have distinct educational advantages as they can demonstrate graphically and intuitively the impact of structural changes on these individual properties. Experimental determination of such properties can be time consuming and tedious as well as, in some cases, being subject to large experimental variation and errors. LogP can be calculated directly from the molecular structure using various mathematical models that utilize molecular fragmentation schemes and other structure-based indexes.
Materials and Methods
LogP for 1-(4-chlorophenyl)-3-(2,6-difluorobenzoyl)urea (diflubenzuron, DFB),
1-(3,5-dichloro-2,4-difluorophenyl) -3-(2,6-difluorobenzoyl)urea (teflubenzuron, TFB),
1-(3,5-dichloro-4-(1,1,2,2,-tetra-fluoroetoxy)phenyl) urea (hexaflumuron, HFM),
1-(2,6 difluorobenzoyl)-3-(2-fluoro-4(2-chloro-4-trifluoromethylphenoxy [ring-U-14C]phenyl)urea (flufenoxuron, FFX) and
N-[[[3,5-dichloro-4- [[3-chloro-5-(trifluoromethyl-2- pyridinyl]oxy]phenyl]amino] carbonyl] -2,6-difluorobenzamide
(chlor--flua-zuron, CFA) were measured using reverse phase HPLC. LogP for 1-(4-fluorophenyl)-3- (2,6-difluorobenzoyl)urea
(FFB) was taken from Sotomatsu (1987). For LogP determinations tefluthirin, DDT, chlorpyrifos, trifluralin, atrazine,
fluorometuron and carbaryl standards were prepared in 7:3 methanol:water (concentration 1ug/mL).
A mixed standard of DFB, TFB, HFM, CFA and FFX (2ug/mL) was prepared in AcCN. All 7 individual pesticide
standards as well as five BPU mixed standard were run at three mobile phase strengths (methanol:water:
90:10, 80:20, 75:25, n=2). Extrapolation back to a 100% water mobile phase was performed as described by
Valko, 1987 using data obtained by RP-HPLC.
We have used Advanced Chemistry Development's ACD/LogP 3.5 to predict the octanol-water
partition coefficients of selected BPU's. The ACD/LogP algorithm was developed using a training set of over 5000
experimentally determined LogP values. The algorithm includes well-characterized LogP contributions of separate atoms,
structural fragments(over 500) and intramolecular interaction contributions (over 2000 types) between different fragments.
The ACD/LogP program is the first program on the market, to the best of our knowledge, whereby the user has the ability to
customize and train the algorithm by including "new" fragments into the calculations. Using experimentally determined LogP
values for structures not previously available within the training set, the system training provides new opportunities for
LogP predictions and enables the program to improve the accuracy of predictions for specific structures of interest to the
user. It is enough to enter LogP experimental value for only one compound into User Data Base from a class of compounds to
use the system training for this particular class. The program splits the entered structure into fragments (which could
be seen in the database window) for which additive increments are generated and used for LogP calculations with regard of
entered experimental LogP values. These generated increments could be eventually used for later LogP calculations of new
compounds, depending on their statistical significance compared to the program's internal database.
Results and Discussion
The LogP experimental value of DFB was input into the "Users system training" (Fig.1,Table 1).
The User Data Base, containing the DFB structure shares a common substructure fragment with TFB, HFM, CFA, FFX and FFB:
We obtained St.D=0.38, R=0,9638 (N=5) without inclusion of the CFA. For logP>5.0 ACD/LogP program
gives the following warning: "Note, that the calculated LogP is close to or above the upper limit for the logP values that
can be precisely measured. The correct value may be underestimated due to solute aggregation/solubilization or limited
instrumental sensitivity".
Fig.1. Experimental (exp.) vs. calculated (calc.*) LogP values after system training for DFB, TFB, HFM, FFX and FFB
References
- T.Sotomatsu, Y.Nakagawa and T.Fujita. 1987. Quantitative Structure-Activity Studies of Benzoylphenylurea Larvicides, Pesticide Biochem. and Physiol., 27, 156-164.
- K.Valko. 1987. RP-HPLC Retention Data for Measuring Structural Similarity of Compounds for QSAR Studies. J.Liquid Chromatogr., 10 (8&9), 1663-1686.
|
