Theoretical design of novel, 4 base pair selective derivatives of mitoxantrone

Mitoxantrone (MTX) is a recently synthesized antitumor intercalative molecule, currently in use in chemotherapy. Previous theoretical computations showed that the base pair selectivity of MTX is limited to the sole two base-pair sequence making up the intercalation site. In order to further extend the recognition site, we undertook, by means of theoretical computations, the design of novel MTX derivatives, in which the terminal hydroxyl group of each side chain is esterified with oligopeptides. We compare in the present study the binding affinities of two derivatives, depsiGly-Lys(D) and depsiGly-Gly-Orn(L), for the palindromic sequences d(CCCGGG)2, d(GCCGGC)2, d(GGCGCC)2, and d(CGCGCG)2. Major groove binding of the oligopeptide arms was shown to be significantly more favourable than either minor groove binding, or binding to the sole phosphate groups. With the two arms adopting two antiparallel directions, two distinct arrangements were investigated in the major groove: (a) the two oligopeptides are brought closer together by means of two hydrogen bonds involving the backbone of their second residue in a beta-sheet like arrangement; (b) the two arms are remote from each other so as to reduce their mutual electrostatic repulsion. Whatever the disposition, the optimal binding configurations were invariably found to be those in which the cationic side chains of the terminal residues chelate N7/O6 of two successive guanines, whenever present on a given strand. A distinct energetical preference for arrangement (a) was obtained with the depsiGly-Gly-Orn(L) derivative. Replacement of the central Gly residue by a Cys one, as in the sequence depsiGly-Cys-Orn(L), was proposed subsequently, so as to further stabilize such a beta-sheet arrangement by means of a disulfide bridge between the two Cys residues. The two investigated compounds were shown to preferentially bind sequences d(CCCGGG)2 and d(GCCGGC)2, with a tetrameric core CCGG rather than sequences d(GGCGCC)2 and d(CGCGCG)2, with a tetrameric core GCGC.