3-indoleacetic acid (IAA) and 3-indolepropionic acid (IPA), 3-indolebutyric acid (IBA) were purchased from J&K Scientific Ltd. (Beijing, China). 1,10-Phenanthroline, DMSO and copper acetate were obtained from Aladdin Co. The complexes IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen were synthesized in Jining Medical University Pharmaceutical College (Rizhao, China) and were dissolved by DMSO to 25 mM. 3-[4,5-dimethylthiazol2-yl]-2.5-diphenyl-tetrazolium bromide (MTT) were obtained from Sigma-Aldrich (St. Louis, MO, USA). The fluorogenic peptide substrate Suc-LLVY-AMC was obtained from Calbiochem (San Diego, CA, USA). Mouse monoclonal antibody to human poly(ADP-ribose) polymerase (PARP), mouse monoclonal antibody against ubiquitin (Ub), goat polyclonal antibody against β-actin (C-11) and all secondary antibodies were purchased from Cell Signaling Technology, Inc. (Shanghai, China).

Infrared spectra were performed on a Nicolet 170-SX spectrophotometer using KBr pellets. ¹H NMR spectra were collected at a Bruker AVANCE III (600-MHz) spectrometer. ¹³C NMR spectra were carried out with a Bruker AV600 (600-MHz) spectrometer. Thermal analyses were recorded on a NETZSCH thermal analyzer. Elemental chemical analysis (C, H and N) were obtained on a Perkin-Elmer 2400 analyzer. Cellular changes were observed using a Zeiss Axiovert 25 microscope.

The complexes were synthesized in two steps by following a previously described procedure (22). In the first step, ligand IAA (0.351 mg, 2 mmol), IPA (0.378 mg, 2 mmol), or IBA (0.406 mg, 2 mmol) were dissolved in ethyl alcohol to give a solution, respectively. Cu(CH₃COO)₂. H₂O (0.199 mg, 1 mmol), dissolved in ethyl alcohol, was added dropwise to the above mentioned solution and reacted for 4 h to give a green solution. In the second step, C₁₂H₈N₂ (0.180 mg, 1 mmol) dissolved in ethyl alcohol, was added into the solution of the first step and then reacted for 3 h to give a green precipitate, which was filtered to give the product.

[Cu(C₁₀H₈O₂N)₂(C₁₂H₈N₂)], FW=592.10 g·mol⁻¹. Yield, 72%. IR (KBr, cm⁻¹): 3389.63 υ(-NH-); 1576.24 υ_as(COO⁻); 1408.06 υ_s(COO⁻); 528.27 υ(M-N); 476.85 υ(M-O); ¹H NMR (DMSO, 600 MHz): d (ppm) 10.736 (s, -NH-, indole ring); 8.931 (d, H-2, 9, 1,10-phenanthroline); 7.187 (s, H-5, 6, 1,10-phenanthroline); 7.352, 7.560 (d, H-4, 7, indole ring); 6.813, 6.975 (t, H-5, 6, indole ring); 7.113 (d, H-2, indole ring); 3.367 (s, -CH₂-indole ring); ¹³C NMR (DMSO): d (ppm) 184.35 (-COO); 155.28, 142.10, 133.14, 125.29 (1,10-phenanthroline carbons); 138.82, 127.41, 122.56, 121.27, 118.18, 117.93, 116.37 and 109.25 (indole ring carbons); 20.80 (-CH₂). TG analysis: residue 14.05% (calculated 13.43%, CuO). Anal. Calc. for Cu (C₁₀H₈O₂N)₂ (C₁₂H₈N₂): C, 64.91; H, 4.09; N, 9.46. Found (%): C, 64.84; H, 4.21; N, 9.32. Molar conductivity: Λ_m (S•cm⁻¹•mol⁻¹): 13.22.

[Cu(C₁₁H₁₀O₂N)₂(C₁₂H₈N₂)], FW=620.16 g·mol⁻¹. Yield, 69%. IR (KBr, cm⁻¹): 3405.72 υ(-NH-); 1586.96 υ_as(COO⁻); 1417.58 υ_s(COO⁻); 546.34 υ(M-N); 497.64 υ(M-O); ¹H NMR (DMSO, 600 MHz): d (ppm) 10.473 (s, -NH-, indole ring); 8.722 (d, H-2, 9, 1,10-phenanthroline); 7.262 (s, H-5, 6, 1,10-phenanthroline); 7.606, 7.398 (d, H-4, 7, indole ring); 6.877, 7.044 (t, H-5, 6, indole ring); 7.039 (d, H-2, indole ring); 4.372 (s, -CH₂-indole ring); 3.420 (t, -CH₂-indole ring). ¹³C NMR (DMSO): d (ppm) 179.63(-COO); 157.32, 145.10, 140.81, 133.05 (1,10-phenanthroline carbons); 137.15, 129.38, 124.87, 119.04, 118.23, 117.65, 112.21 and 110.52 (indole ring carbons); 25.19, 22.66 (-CH₂). TG analysis: residue 13.26% (calculated 12.83%, CuO). Anal. Calc. for Cu (C₁₁H₁₀O₂N)₂ (C₁₂H₈N₂): C, 65.85; H, 4.55; N, 9.03. Found (%): C, 65.79; H, 4.60; N, 8.97. Molar conductivity: Λ_m (S•cm⁻¹•mol⁻¹): 8.29.

[Cu(C₁₂H₁₂O₂N)₂(C₁₂H₈N₂)], FW=648.21g·mol⁻¹. Yield, 66%. IR (KBr, cm⁻¹): 3417.19 υ(-NH-); 1592.77 υ_as(COO⁻); 1423.03 υ_s(COO⁻); 551.20 υ(M-N); 483.12 υ(M-O); ¹H NMR (DMSO, 600 MHz): d (ppm) 10.559 (s, -NH-, indole ring); 8.806 (d, H-2, 9, 1,10-phenanthroline); 7.224 (s, H-5, 6, 1,10-phenanthroline); 7.587, 7.390 (d, H-4, 7, indole ring); 6.846, 7.012 (t, H-5, 6, indole ring); 7.156 (d, H-2, indole ring); 4.437 (s, -CH₂-indole ring); 3.382 (t, -CH₂-indole ring); 3.053 (t, -CH₂-indole ring). ¹³C NMR (DMSO): d (ppm) 180.76 (-COO); 153.27, 141.01, 139.84, 129.69 (1,10-phenanthroline carbons); 137.92, 127.58, 123.30, 120.16, 118.88, 117.43, 112.87 and 109.53 (indole ring carbons); 26.44, 24.60, 21.55 (-CH₂). TG analysis: residue 12.80% (calculated 12.27%, CuO). Anal. Calc. for Cu (C₁₂H₁₂O₂N)₂ (C₁₂H₈N₂): C, 66.70; H, 4.98; N, 8.64. Found (%): C, 66.64; H, 5.08; N, 8.55. Molar conductivity: Λ_m (S•cm⁻¹•mol⁻¹): 11.04.

MDA-MB-231 and MCF-7 breast cancer cells were grown at 37°C and 5% CO₂ in DMEM/F-12 (1:1) or RPMI-1640 (Invitrogen), respectively, supplemented with 10% fetal bovine serum (FBS). A whole cell extracts were prepared and applied to assess the CT-like activity and western blot analysis (26).

MTT assay was applied to evaluate the antiproliferative effect of each complex on breast cancer cells. Cells were seeded in a 96-well plate and grown to 70–80% confluency, then treated with the indicated concentration of each complex at 37°C. After cultured for 24 h, cell proliferation was assessed as previously described (26).

Purified 20S human proteasome (35 ng) or whole-cell extract (10 μg) of human breast cancer cells were incubated with a series of concentrations of three copper complexes in 100 μl of assay buffer (20 mM Tris-HCl, pH 7.5) with 20 μM fluorogenic peptide substrate Suc-LLVY-AMC. After 2-h incubation at 37°C, the production of hydrolyzed AMC groups was measured as previously described (26).

MAD-MB-231 human breast cancer cells were treated with the complexes as already described (27). Proteins (40 μg) were separated by SDS-PAGE and transferred to a nitrocellulose membrane. The expression of ubiquitin, PARP and β-actin were examined by western blot analysis.

Sequence of human 20S proteasome β subunits were retrieved from Protein KnowledgeBase (UniProtKB). For each subunit sequence, a BLAST search was performed on PDB library and bovine 20S proteasome [PDB code 1IRU (21)] was found to be the most homologous structure to human and was chosen for this docking study after some minor modifications. The small molecules were sketched manually and energy minimized and then docked to all three catalytic sites of the proteasome. Molecular docking was performed with the CCDC Gold 5.2 (28). Scoring function was set to ChemPLP, and other docking parameters were set to their default values. Accelrys Discovery Studio Visualizer 4.0 was used for result analysis and figure generation.

Statistical analysis was performed using Graphpad Prism 5 software. Differences between groups were analyzed using the Student's t-test.

3-indolecarboxylic acid copper complexes could exhibit potent anticancer activity through inhibiting proteasome activity (22). Whereas the exact mechanism is still unknown, there is a hypothesis that the copper complexes interact with proteasome β5 catalytically active subunit. At the same time, the length of linker between indole ring and carboxylic acid group has a certain impact on the activity of these complexes. In the present study, we synthesized three copper complexes with similar structure: IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen (Fig. 1), among which IBA-Cu-phen is novel, and then compared their anticancer activity and investigated the potential mechanism of action. The obtained complexes were characterized by IR, ¹H NMR, ¹³C NMR, TG analysis and elemental analysis. Because these compounds were not crystallized, X-ray diffraction was not performed.

According to the IR spectra data of ligands and complexes (IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen), the strong absorption peaks at 3389.63, 3405.81 and 3417.19 cm⁻¹, respectively, belong to the -NH- group. No significant shifts were observed on the peaks among these ligands, suggesting that nitrogen atom in the indole ring did not coordinate with copper. Compared with the IR spectra of the ligands, two new peaks in the range of 1576.24 cm⁻¹–1592.77 cm⁻¹ and 1408.06 cm⁻¹–1423.03 cm⁻¹ were assigned to υ_as(COO⁻) and υ_s(COO⁻), respectively. In addition, the magnitude between υ_as(COO⁻) and υ_s(COO⁻) was <200 cm⁻¹, indicating that the oxygen in -COO⁻ group formed a coordination bond with copper ion, which is consistent with the literature (29).

The absorption vibration frequencies appeared at 476.85, 497.64 or 483.12 cm⁻¹ in these complexes were attributed to the formation of Cu-O bond. A new peak at 528.27, 546.34 or 551.20 cm⁻¹, respectively, appeared in these complexes, suggesting the formation of Cu-N bond.

The ¹H NMR spectra of IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen showed that there were small shifts in the signals among these ligands, suggesting the interaction between ligands and Cu(II) ion. In addition, the signal for -NH- proton in the spectra (10.736, 10.473 and 10.559 ppm) were still present, while the signal for the -COOH proton of the ligands were absent in these complexes, indicating the hydrogen atom of carboxyl group were displaced by Cu(II) ion. In comparison with the ¹H NMR spectra of ligands, the signal of 1,10-phenanthroline did appear in these complexes, indicating there was coordination of 1,10-phenanthroline and Cu (II) ion. The ¹H NMR spectra analysis results were further supported by ¹³C NMR spectra data (¹H NMR and ¹³C NMR data of the ligands IAA, IPA and IBA were not shown).

Thermal analysis (TG) of complexes IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen were recorded in the range from 25 to 800°C. These copper-containing compounds were all decomposed actually in one step and the residue rates of them were 14.05, 13.26 and 12.80%, respectively, which were consistent with the calculated values (13.43, 12.83 and 12.27%).

Elemental analysis was applied to further confirm the composition of these copper complexes. The contents of carbon, hydrogen and nitrogen in each complex were consistent with the calculated values, which supported the proposed structure of these complexes.

The ligands and complexes were stable in air. Furthermore, all the complexes were soluble in DMSO and their molar conductivities (Λ_m) were 13.22, 8.2, and 11.04 S·cm²·mol⁻¹, respectively, which were <35 S·cm²·mol⁻¹ (30).

Consequently, the complexes IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen were considered as non-electrolytes and quite stable in culture medium.

We first studied effects of IAA-Cu-phen, IPA-Cu-phen, IBA-Cu-phen and their ligands on cell growth (Fig. 2). Human breast cancer MDA-MB-231 and MCF-7 cells were treated with 1, 5, 10 and 20 μM (IAA-Cu-phen, IPA-Cu-phen, IBA-Cu-phen) as well as their ligands for 24 h, followed by an MTT assay. We found that these three complexes possessed similar growth-inhibitory activity with >90% inhibition at 20 μM in both cell lines in dose-dependent, compared to a control treated with DMSO (Fig. 2). However, IAA-Cu-phen was most potent, IPA-Cu-phen came next, followed by IBA-Cu-phen while none of the ligands were active. The IC₅₀ values for cell death induction by all the ligands and copper complexes are given in Table I.

In this part of the present study, we investigated whether these three complexes could act as potent proteasome inhibitors and possess similar effects of proteasome inhibition as on cell proliferation. Purified human 20S proteasome were kept with various concentrations of the three complexes for 2 h and DMSO was used as a control. This result was consistent with MTT assays. The three complexes inhibited the 20S proteasomal chymotrypsin-like in vitro with rather similar IC₅₀ (3.5–3.75 μM), and the inhibition ability increased along the concentration of these complexes (Fig. 3).

Chemical structures of several complexes could play a key role in inhibiting proteasome activity to some degree. Therefore, we studied the effects on cellular proteasomal CT-like inhibition and apoptosis induction of these three copper complexes with similar structures. Human breast cancer MDA-MB-231 cells were treated with three copper complexes at the concentrations of 1, 5, 10 and 20 μM for 24 h (DMSO as control), and cellular proteasomal CT-like activity and morphological changes were measured. The results indicated that all three complexes at 1 μM showed no less than 5% inhibition of proteasomal CT-like activity (Fig. 4A) and no significant apoptosis-associated cellular morphology changes were observed (Fig. 4B). However, the abilities on the proteasome inhibition of these three copper complexes were obviously different at the concentrations of 5, 10 and 20 μM. We found that IAA-Cu-phen was the most potent, inhibiting 48, 83 and 98% (Fig. 4A), followed by IPA-Cu-phen, inhibiting 36%, 66 and 89 (Fig. 4A), and the last was IBA-Cu-phen, with the degrees of inhibition being 28, 59 and 79% (Fig. 4A), respectively. Consistently, cellular morphology changes (shrunken and rounded up) were found at concentrations of 5, 10 and 20 μM (Fig. 4B). Our results suggested all three copper complexes possessed the ability to inhibit the proteasome activity and induce apoptosis in a dose-dependent manner with the order IAA-Cu-phen >IPA-Cu-phen >IBA-Cu-phen.

Previous studies have shown that the decrease of CT-like activity in proteasome is associated with induction of tumor cell apoptosis (5,17–20). Herein, we performed a kinetic experiment to investigate whether the new complex IBA-Cu-phen possessed the same effect. MDA-MB-231 cells were treated with 20 μM compound for 0, 4, 8, 12 or 24 h, followed by measurement of proteasomal inhibition and cell apoptosis (Fig. 5). The result showed that IBA-Cu-phen inhibited ~12% of proteasome activity after 4-h treatment (Fig. 5A). As time went on, the chymotrypsin-like activity further decreased. The inhibition ratios were 43, 77 and 86% at 8, 12 and 24 h, respectively. Consistently, the accumulation of ubiquitinated proteins were also detected at 4 h and the levels increased gradually at later time-points (Fig. 5B). However, cellular apoptosis morphological changes (Fig. 5C) and apoptosis-specific PARP cleavage (Fig. 5B) both appeared at 12 h and later time-point of treatment. These results demonstrate that proteasome inhibition by IBA-Cu-phen triggers the occurrence of tumor cell apoptosis.

To further explain how and why these three copper complexes could serve as proteasomal CT-like inhibitors, we investigated the possible binding mode of the three compounds with proteasome by molecular docking simulation. The small molecules IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen were docked to all the three catalytic site of the above receptor structure using CCDC GOLD 5. These three small molecules were found to have the best binding with catalytic site of subunit β5, which locates near the adjacent subunit β1 and also consists of a number of residues from subunit β1, like SER123, ASP125, PRO126, VAL127 and GLN131 (Fig. 6).

The binding mode of IAA-Cu-phen, which has the highest activity among all three complexes, was analyzed. As shown in Fig. 6A, IAA-Cu-phen forms a few hydrogen bonds and hydrophobic interactions with the receptor. Briefly, it forms a hydrogen bond with THR1 sidechain by its indole nitrogen atom; also, its carboxylic oxygen atom (which coordinates with copper), forms a hydrogen bond with ALA49, which plays an important role in IAA-Cu-phen binding. Residues ALA20, VAL31, LYS33 and MET45 form a large hydrophobic cavity with their sidechains, to accommodate one indole ring of IAA-Cu-phen and bind it with hydrophobic interactions. The naphthisodiazine group of IAA-Cu-phen binds to ALA50 and two residues from subunit β1 (PRO126 and VAL127) through hydrophobic interactions. Furthermore, binding modes of three complexes were compared. As displayed in Fig. 6B, IAA-Cu-phen, IPA-Cu-phen and IBA-Cu-phen in proteasome catalytic pocket have similar binding conformations, but differ in the position and orientation of the two terminal indole rings. IAA-Cu-phen binds to the pocket tightly, while IBA-Cu-phen shows a little weaker binding to the pocket than IAA-Cu-phen and IPA-Cu-phen. These results agree with the biological study results.