Synthesis and anticancer activities of diverse furo[2,3-d]pyrimidine and benzofuro[3,2-d]pyrimidine derivatives

Abstract A series of diverse furo[2,3-d]pyrimidines (2a–2b, 4a–4d and 8a–8c) and benzofuro[3,2-d]pyrimidines (12a–12c) were synthesized and screened for their antitumor effects against HepG2, Bel-7402 and HeLa cell lines in vitro. Representatively, 4a, with an IC50 of 0.70 μM, exhibited the best antitumor activity against the tested HepG2 cell lines. Molecular docking investigation further revealed the possible binding modes of compound 4a with receptor tyrosine kinase. Preliminary results indicated that the title compounds were helpful as leading structures for preparing a new antitumor drug. Graphical abstract


Introduction
Cancer is a serious disease worldwide that threatens human health and reduces patient quality of life. Although chemotherapy has some effects for cancer treatment, many patients still suffer from drug resistance and different degrees of side effects. [1,2] Therefore, the development of new anticancer drugs is very important.
In the presence of triethylamine, 7 reacted with morpholine and amino acid ester to give 8a-8c (Scheme 2). According to the effect of the molecular skeleton and substituent group on the biological activity of drugs, benzofuro[3,2-d]pyrimidine derivatives were synthesized by a process similar to that described above from the starting material of the ethyl 3-amino benzofuran-2-carboxylate and aza-wittig reactions (Scheme 3).
exposure of cells to the tested compounds for 48 h with 5-FU as a positive control experiment and solvent as a negative control experiment. All experiments were performed in triplicate.
As indicated in Table 1, all the compounds showed potential anticancer activity against the three cell lines. The most promising compound, 4a, exhibited the best antitumor activity against the HepG2 cell line with an IC 50 of 0.70 lM, and 2b, 4d, 8c, and 12c had a broad spectrum of cytotoxicity against the HepG2, HeLa and Bel-7402 cell lines with an IC 50 range of 4.5 À 38.5 lM. Meanwhile, the results also revealed that 2a, 8b, 12b, and 4b presented a certain protective effect on HepG2 and HeLa cell lines.

Molecular dynamics simulation
To obtain the stable binding mode of 4a with Abl, molecular dynamics (MD) simulation was performed based on the 4a-EGFR complex structure obtained from the docking results. We used the antechamber module of the Amber16 program [14] to assign bcc charges for the atoms of 4a. The topology and coordinate files of the 4a-EGFR complex were built with the Leap module in the Amber16 program. The AMBER ff14SB force field was used for amino acid residues, and the AMBER gaff force field was used for 4a [15] Clor Na þ was added to neutralize the system. All the molecules were solvated in a rectangular box of TIP3P waters extended at least 10 Å in each direction from the solute. [16] Energy minimization was performed on the system before the MD simulation. The MD simulation was carried out by employing the periodic boundary condition with the NPT ensemble to avoid edge effects. First, 10 picoseconds (ps) simulation was performed on water and ions to obtain an equilibrated solvent environment. Second, the temperature of the system was gradually heated from 10 to 298 K for 50 ps. Finally, to obtain a stable MD trajectory, the systems were run for 5.5 nanoseconds (ns) at 298 K and constant pressure. During the MD simulation, we used the SHAKE algorithm to constrain all covalent bonds involving hydrogen atoms. [16] The particle mesh Ewald (PME) algorithm was used to handle van der Waals (vdW) energy terms as well as the long-range electrostatic interactions with a cutoff distance of 10 Å. The 2.0 femtoseconds (fs) were used as the time step during the MD simulations, and the coordinates were collected every 1 ps.

Free energy calculation
The MM-PBSA method was used to calculate the binding free energy (DG bind ) between the receptor and ligand. [17] This value was obtained by calculating the differences in free energies between the ligand-receptor complex (G cpx ) and the unbound receptor (G rec ) and ligand (G lig ) as follows: DG bind consists of the molecular mechanical (MM) gas-phase binding energy (DE MM ), solvation free energy (DG sol ) and entropic contribution (-TDS): The DE MM includes two parts, the electrostatic energies (DE ele ) and van der Waals interactions (DE vdW ): The DG sol is made up of electrostatic contribution (DG PB ) and nonelectrostatic contribution (DG np ) to the solvation free energy. DG PB is calculated by the Poisson-Boltzman (PB) method using the MM_PBSA module in the amber16 program. DG np is determined by the solvent accessible surface area. [18] For the entropic contribution, an empirical method [19] was used, which consists of two subitems, the solvation entropy change (DS sol ) and conformational entropy change (DS conf ): The DS sol is obtained by the tendency of water molecules to minimize their contacts with hydrophobic groups in the protein, and DS conf is related to the change in the number of rotatable bonds during the binding process. The entropic contribution is evaluated, and the conformational entropy change is proportional to the number (DN rot ) of the lost rotatable bonds during binding: -TDS conf ¼ w DN rot ð Þ (6) in which w is a scaling factor that was set to be 1 Kcal/mol for the binding energy calculation. Thus, Eq. (1) can be written as: All other parameters in the energy calculation are the standard parameters or the default values of the Amber16 program.

Molecular simulation result
Approximately 5.5 ns of MD simulation was performed on the 4a-Abl complex. The Cpptraj tool in Amber16 was used to plot the root-mean-square displacement (rmsd) values (shown in Fig. 2A). The rmsd plots indicated that the systems of 4a-EGFR achieved equilibrium very quickly. The final conformations of the MD result were used to analyze the binding mode (shown in Fig. 2B). Compound 4a formed two H-bonds with residue D381, which belongs to the DFG motif of Abl. The phenyl ring at the end formed hydrophobic interactions with side chains of residues, including L370, T315 and M290, and at the same time formed p-p interactions with DFG-motif residue F382. The binding free energies were calculated by using the molecular mechanics-Poisson- Boltzmann surface area (MM/PBSA) [20][21][22] binding free energy calculation method (shown in Table 2).

Conclusion
In conclusion, we have developed an efficient synthesis for diverse furo [2,3-d]pyrimidine and benzofuro[3,2-d]pyrimidine derivatives. Bioassays indicated that these compounds possess potential antitumor activities, and promising compound 4a stood out as the most potent on HepG2 cell lines with an IC 50 value of 0.70 lM. A molecular simulation study predicted the possible binding mode of compound 4a with its potential target tyrosine kinase, which can help us gain a deeper understanding of the mechanism of action of our compounds to have drug effects. We hope our efforts will be helpful for further anticancer research and development.

Experimental section
General Melting points were recorded using an uncorrected X-4 digital melting point apparatus. All chemicals were commercially available, analytically pure and used without further purification. MS was measured by a high-performance liquid chromatography-triple quadrupole tandem mass spectrometer. NMR spectra were recorded in CDCl 3 (or DMSO-d6) using a Varian Mercury 400 spectrometer with resonances relative to tetramethylsilane (TMS) as an internal standard. Elemental analyses were recorded on a PerkinElmer CHN 2400 instrument. TLC analysis was carried out on silica gel plates GF254 (Wuhan, Geao Co.).
To a mixture of 6a (5 mmol), prepared as described in detail previously, in aqueous ethanol solution 50 mL, solid KOH (10 mmol) was added. The reaction mixture was stirred for 8 h at 50 C and then for an additional 1 h at room temperature after completion of the reaction, as indicated by TLC. The solution was concentrated and acidified to pH ¼ 1 with dilute hydrochloric acid in an ice bath. The solid obtained was filtered, washed with water and dried to provide 6a.
A mixture of 6a (3 mmol) and SOCl 2 (1 mL) was stirred at room temperature for 12 h and then concentrated under reduced pressure to remove excessive SOCl 2 to give 7a, which was used directly with further purification. Triethylamine (6 mmol) was slowly added to a stirred solution of 7a and amino acid ester (3 mmol) in 15 mL of dry CH 2 Cl 2 in an ice bath. The mixture was stirred at room temperature for 20 h, concentrated, and then poured into ice water. The solid obtained was filtered and recrystallized from DMF/CH 2 Cl 2 (v:v ¼ 1:4) to give 8a-8c.  13  Full experimental details, 1 H and 13 C NMR, and mass spectra data are provided in the supporting information. This material can be found via the "Supplementary Content" section of this article's webpage.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
We gratefully acknowledge the financial support of this work by the Natural Science