A new series of asymmetrically N,N'-substituted ureas 20–25 was synthesised using solvent free reactions. The synthetic procedure was performed in three steps starting with Schiff bases derived from cinnamaldehyde with subsequent reduction reactions affording the corresponding allyl amines, which was followed by coupling with phenyl isocyanate, as outlined in Scheme 1. The syntheses of all of the compounds, Schiff bases, corresponding amines and ureas were performed under solvent free conditions.

First, the Schiff bases were prepared from cinnamaldehyde (1) and 4-substituted-anilines 2–7, which were mixed using a mortar and pestle for 15 min. The products 8–13 were obtained in good yield and were recrystallised from ethanol. The Schiff bases were also prepared with the same reactants, silica gel and ethanol as solvent using an ultrasonic bath for 20 min. The products were obtained after silica gel filtration, solvent evaporation and recrystallisation from ethanol. Table 1 shows the yields and the corresponding melting points of the Schiff bases prepared under both reaction conditions.

The allyl amines 14–19 were obtained by reduction of the Schiff bases using NaBH₄ as the reducing agent via two different methodologies. Under solvent free conditions, the reduction products were obtained by vigorous mixing of the Schiff bases with NaBH₄ and H₃BO₃ in a mortar and pestle for 30 min. Next, the solids were washed with a NaHCO₃ solution, extracted with CH₂Cl₂ and dried under vacuum affording the corresponding allyl amines in quantitative yields. The second methodology involved equimolar quantities of the Schiff bases and NaBH₄ with refluxing for 8 to 10 h in ethanol, which was then evaporated under vacuum. Then, the allyl amines were extracted using the method described above to afford the products in good yields (45–75%). Table 2 shows the yields obtained for the allyl amines via both methodologies.

The asymmetrically N,N'-substituted ureas 20–25 were synthesised from allyl amines 14–19 in the presence of phenyl isocyanate via two different methods. The traditional methodology afforded the target compounds after refluxing for 4 to 24 h in toluene. Under solvent free conditions, the reactants were mixed vigorously for 30–45 min with a mortar and pestle. The solid products were recrystallised from ethanol, and the oil products were purified by column chromatography using CHCl₃ as the solvent. The ureas were obtained in excellent yield (98–100%) after remarkably low reaction times (30 min) compared to those obtained with traditional methodology (25–64% and 4–24 h). Table 3 shows the yields for the asymmetrically N,N'-substituted ureas obtained from both methodologies.

The formation of the ureas 20–25 was observed via infrared spectra. The ν_C=O stretching vibration observed at 1654 to 1674 cm⁻¹ (Experimental section) indicates the formation of the products because the secondary amines (14–19) do not absorb in this region. These values are in good agreement with the literature values (1623-1647 cm⁻¹) [8].

The ¹H-NMR spectra of compounds 20–25 was employed to identify the three aromatic rings. However, other chemical shifts did not change compared to the amines 14–19. The ¹³C-NMR spectra were of fundamental importance in the characterisation because the chemical shifts of the downfield signals with a δ of 154.29 to 154.43 could be attributed to the carbon atom of the urea carbonyl groups (Experimental section).

The ability of compounds 20–25 to inhibit human DNA topoisomerase II-α (topo II-α) activity was observed in relaxation assays using supercoiled pBR322 plasmid DNA in the presence of ATP. The reaction products were analysed by electrophoretic mobility and revealed in ethidium bromide in the presence of UV light. As shown in Figure 1, all compounds inhibited the enzyme activity at 10 μM; however the urea 24 presented a weak supercoilled DNA band indicating a smaller inhibitory effect. Etoposide, a known topo II-α inhibitor, was used as a positive control at 100 μM.

To analyse the molecular details of the mode of interaction with topo II-α, compounds 20–25 were docked into the enzyme’s DNA and ATP binding sites. Initially, all of the scoring functions available in the GOLD 5.1 program (CCDC Software Ltd., Cambridge, UK) were tested through a redocking procedure with a co-crystallised ligand (i.e., etoposide and phosphoaminophosphonic acid-adenylate ester, an ATP analogue, the PDB codes for the crystallographic structures are 3QX3 and 1ZXM, respectively). Although all of the available scoring functions were able to dock both ligands into their respective binding sites, the best results were obtained with the ChemPLP function [23,24], which resulted in etoposide and the ATP analogue exhibiting the lowest RMSD values (i.e., 0.250 Å and 0.645 Å, respectively) compared to the crystal structures.

Several poses were obtained for each urea with ChemPLP, and the best-ranked pose for each one was chosen for analysis of the interactions with the enzyme. In the GOLD docking program, the docking functions yield fitness scores, which are dimensionless values. In each case, the scale of the score is a guide of how good the docking pose is with a higher score indicating a better docking result. The fitness scores for each ligand in topo II-α are presented in Table 4.

All of the tested ureas were predicted to successfully dock in both binding sites of topo II-α, but the score was higher for the DNA binding site, indicating that this site was the most probable interaction site for these compounds. The differences observed between the scores from the ureas and the ATP analogue also indicated that it would be more difficult for these compounds to effectively compete with ATP by its binding site. In fact, the ATP analogue established a number of interactions that the ureas were not able to establish. Therefore, the docking results suggest that the DNA site in topo II-α is the more probable binding site for this type of molecule, which results in isosteric inhibition.

Etoposide is known to prevent re-ligation of the DNA strands, and by doing so causes DNA strands to break. Analysis of the co-crystal structure of a fragment of human topoisomerase complexed to DNA and to the anticancer drug etoposide (PDB code 3QX3) revealed that the inhibitor forms H bonds with Asp479 residue and also with a guanine fragment (DG13) of DNA; the inhibitor’s aglicone moiety is inserted between DNA base pairs. Compound 22 was predicted as the best urea ligand for the DNA binding site; although not all interactions observed for etoposide could be matched by compound 22 in the best-ranked docking pose, its overall structure is able to occupy at least partially the same subsites occupied by the co-crystallized ligand molecule. As presented in Figure 2, compound 22 is inserted between the same base pairs as etoposide, which is expected to abolish stacking interactions, effectively blocking religation of the cleaved phosphodiester bond. The major difference is observed in the etoposide’s glycosidic group subsite, but this group is not involved in drug-DNA interactions. The 4-hydroxy-3,5-dimethoxyphenyl and the fused four-ring system of the co-crystallized ligand are mimicked by the methoxy-phenyl and the N-[(2E)-3-phenylprop-2-en-1-yl]urea groups, respectively, of compound 22, so a mechanism of action similar to that of etoposide can be proposed for topo II-α inhibition by the urea ligands.

The ¹H- and ¹³C-NMR spectra were obtained on a Bruker Avance II 400 spectrometer (¹H, 400 MHz; ¹³C, 100 MHz) using tetramethylsilane (TMS) as the internal standard and CDCl₃ as the solvent, and the abbreviations to indicate the multiplicity of ¹H-NMR spectra signals were: s (singlet), sl (singlet large), d (doublet), dd (double doublet), t (triplet) and m (multiplet). The melting points were determined with a Meltemp II apparatus and were uncorrected. The infrared spectra (KBr pellets) were recorded on a Bruker Vertex 70 spectrophotometer. The ultrasonic-assisted reactions were performed in a Unique ultracleaner bath, model 1400A.

The DNA-topoisomerase II-α (topo II-α) drug screening kits were provided by TopoGEN [27] and contained supercoiled (form I) plasmid substrate DNA (25 μg in 10 mL TE buffer). TE buffer (10 mM Tris–HCl at pH 7.5 and 1 mM EDTA) and the assay buffer (50 mM Tris–HCl, pH 8, 120 mM KCl, 10 mM MgCl₂, 100 mM EDTA, 3 mg/mL bovine serum albumin, 0.5 mM dithiothreitol, and 0.5 mM ATP) were used. The supercoiled pBR322 plasmid DNA was purchased from Gibco (Grand Island, NY, USA). The loading buffer contained 25% bromophenol blue, 50% glycerol and 10% SDS. The agarose and the substances utilised in these assays were purchased from Sigma® (St. Louis, MO, USA).

The topo II-α inhibition assay was performed as described in the TopoGEN screening kit. Briefly, the reaction mixture (10 μL) contained the drug, DNA, assay buffer, 2U of topo II-α, and water. In the topo II-α assay, two units were utilised to relax 0.125 μg/mL pBR322 (Gibco). The mixture was incubated at 37 °C for 30 min, and the reaction was terminated by addition of 1 μL of a dye solution containing 25% bromophenol blue, 50% glycerol and 10% SDS (sodium dodecyl sulfate). The products were submitted to electrophoresis using 1% agarose gel in TAE buffer (50× stock: 242 g Tris base, 57.1 mL glacial acetic acid, and 100 mL of 0.5 M EDTA) at 15 V for 3.5 h. The gels were stained with ethidium bromide (0.5 μg/mL) for 30–45 min, washed, and photographed under UV light.

Compounds 20–25 were constructed and energy-minimised with the PM3 semi-empirical method [28] as implemented in the Spartan’08 program (Wavefunction, Inc., Irvine, CA, USA). The resulting structures were used for the docking study of the DNA and the ATP binding sites. Due to the significant structural modifications of the enzyme after ATP binding (and also of its N-substituted analogue) to its specific allosteric site, it was necessary to use different crystal structures for the study of these two binding sites. Docking solutions with the DNA binding site were obtained with the human topo II-α co-crystallised with DNA and etoposide (PDB code 3QX3, resolution 2.16 Å) [29]. For the ATP binding site, human topo II-α co-crystallised with phosphoaminophosphonic acid-adenylate ester, an ATP analogue, in the ATPase domain (PDB code 1ZXM, resolution 1.87 Å) was used [30]. For the docking studies of compounds 20–25, the etoposide and water molecules were deleted from the 3QX3 structure, and the ATP analogue and water molecules were deleted from the 1ZXM crystal structure. Both TOPO-II structures contain Mg²⁺ ions, and the 3QX3 structure contains a co-crystallised DNA segment, which was not removed.

The docking procedure was accomplished with the GOLD 5.1 program (CCDC Software Ltd.), and the hydrogen atoms were added to the protein structures based on ionisation and tautomeric states defined by the program. The number of genetic operations (crossover, migration, mutation) for each run was set to 100,000 in the searching procedure. All of the scoring functions available in the program, such as GoldScore [31], ChemScore [32], ASP [33] and ChemPLP [34], were evaluated and compared for the docking calculations.

The internal binding site for docking was defined with a 20 Å radius from residue Asp479 in the DNA binding site and with a 10 Å radius from the magnesium atom in the ATP binding site. Both are located at the centre of each site. For the docking evaluation and adequate scoring function selection, the co-crystallised structures (etoposide and the ATP analogue, in the respective binding site) were redocked to each original site. The resulting poses of compounds (20–25) were compared and analysed to identify structural characteristics of the complexes formed by these molecules and topo-II.