Objective: The objective of this study was to provide experimental chemists with useful guidance for
proper choice of a reasonable cyclization precursor in macrolactonization in which there are several options in
Method: A combination of DFT calculations and MD simulations was used to achieve the aim. In the former
method, (1) structures having a local minimum energy near the macrocyclic structure are located by DFT calculations
starting from coordinates generated by breaking the corresponding ester C–O bonds (2) among the structures
obtained, the one with the shortest distance between the carbon and oxygen atoms and the carboxylic carbon
atom is selected as the most likely candidate for macrolactonization. The latter approach involves that (1) global
minimum structures for the possible macrocyclization precursors are located by molecular mechanics calculations
(2) each of the structures is subjected to MD simulations and a structure maintaining the shortest distances between
the two reacting centers during the simulations is selected as a reasonable cyclization precursor.
Results: The utility of the computation-assisted disconnection approach was verified by the synthesis of prunustatin
A, whose synthesis has already been reported by us, and demonstrated by the application to the synthesis of
neoantimycin, a 15-membered macrocyclic depsipeptide.
Conclusion: It was demonstrated that the approach provides useful guidance for designing synthetic strategies for
synthetic chemists because of the relatively high speed and simplicity of the calculations involved.