Simple Synthesis of a Heterocyclophane Exhibiting Anti-c-Met Activity by Acting as a Hatch Blocking Access to the Active Site

Abstract

A simple approach to the synthesis of heterocyclophane consisting of two 4,4’-bithiazoles has been developed in mild conditions. The heterocyclophane with two short chains was conveniently prepared by Hantzsch thiazoles synthesis using the reaction of 3-tert-butoxycarbonyl-3-azapentanethiocarboxamide with 1,4-dibromobutane-2,3-dione in methanol under reflux for only 15 minutes. Amino groups at the linkers of this heterocyclophane can be functionalized to give acylated and carbamate derivatives. Their properties as protein kinase inhibitors were investigated, and one of the heterocyclophanes exhibited specific anti-activity for c-mesenchymal epithelial transition factor (IC₅₀ = 603 nM) among 7 types of protein kinases investigated. The computational site identification by ligand competitive saturation method was used to determine why the one heterocyclophane exhibited strong anti-activity for c-mesenchymal epithelial transition factor.

Graphical Abstract

Anti c-Met activity using cyclophane framework: Bithiazolophane 1 which was conveniently prepared under reflux for only 15 minutes by Hantzsch thiazoles synthesis exhibited anti-activity (IC₅₀ = 603 nM) against c-Met. The computational site identification by ligand competitive saturation (SILCS-MC) showed 1 blocked access to the active site of c-Met by acting as a hatch.

Introduction

In conventional cancer chemotherapies, cell growth inhibitory drugs are mainly used which cleave DNA,^[1] or inhibit DNA synthesis,^[2] or inhibit the mechanism of cell division.^[3] These drugs cause significant damage to DNAs in cancer cells and thus lead the cells to apoptosis. However, those drugs have side effects, including hair follicle, bone marrow, and digestive tract disorders, by targeting a common mechanism in both normal proliferating cells and cancer cells. Recently, drugs which target specific cancer cell types are being introduced. For example, it is reported that a mutation causes constitutive activation of a tyrosine kinase gene called epithelial growth factor receptor (EGFR) in lung cancers. The EGFR tyrosine kinase inhibitors (EGFR-TKI) such as Gefitinib (Iressa) and Erlotinib (Tarceva) are employed against the lung cancers having that mutation.^[4–6] c-mesenchymal epithelial transition factor (c-Met) is an oncogene that codes for a receptor tyrosine kinase that is regulated by hepatocyte growth factor (HGF).^[7,8] c-Met in complex with HGF generates signals to activate cell proliferation, survival, and motility. Since their activations play a pivotal role in tumor formation,^[9] c-Met inhibitors that block c-Met-induced cell line amplification have been developed and reported.^[10–12] Although the HGF/c-Met pathway is involved in acquired resistance for EGFR-TKI, the resistance may be overcome by using combinations of EGFR-TKI and c-Met inhibitors.^[13] Therefore, c-Met is a promising molecular target for cancer treatment.

In the present study, we develop a novel inhibitor of c-Met.^[14–16] In our previous work, we synthesized heterocyclophanes containing azole rings such as imidazole,^[17] oxazole,^[18,19] and bithiazole.^[20] In oxazolophane and bithiazolophane, those heterocyclophane frameworks were constructed by a convenient reaction based on the Hantzsch reaction^[21–23] using starting molecules without aromatic frameworks. These synthetic methods are very simple because the reactions proceed in mild conditions and finish generally faster than other cyclophane syntheses. In addition, the size of bithiazolophane can be controlled by changing the length of the linkers.

A large cyclophane 5 (Figure 1), 2,6-di(thiazol-4-yl)pyridinophane, was previously synthesized and tested as a tyrosine kinases inhibitor, though it didn’t show any inhibitory activity. While not a kinase inhibitor, we hypothesized that its large size may disallow binding. This motivated the present effort, and we have now succeeded in synthesizing bithiazolophanes 1–4 (Figure 1) bridged by short chains. Herein, we report the convenient Hantzsch thiazole synthesis of 1, that includes two 4,4’-bithiazoles. Additionally, amino groups at the short bridges of 1 were functionalized to give the derivatives 2–4. Their ability to inhibit a tyrosine kinase as molecular targeted drugs was then investigated. Of the compounds tested, only 1 showed strong anti-c-Met activity. To understand this observation, the site identification by ligand competitive saturation (SILCS) method was used to elucidate the molecular details of the inhibitor-enzyme interactions.

Figure 1.

Bithiazolophane derivatives 1–4 and 2,6-di(thiazol-4-yl)pyridinophane 5.

Results and Discussion

Synthesis of bithiazolophane 1

As shown in Scheme 1, bithiazolophane 1 with two short chains was synthesized through 4 steps. First, amino group of 6 was protected by Boc group to give 7. Next, aqueous ammonia was used to transform ester 7 to amide 8, which was converted into thioamide 9 by treatment with Lawesson’s reagent.^[24] Finally, the reaction of 9 with 1,4-dibromobutane-2,3-dione in methanol under reflux for 15 minutes to give the short chain-bridged bithiazolophane 1, together with trithiazolophane 10 as a side product.

Scheme 1.

Synthesis of 4,4’-bithiazolophane 1. i) 1.3 equiv Boc₂O, 1.5 equiv Me₃N, THF, r.t. 21 h; ii) excess NH₃, water, r.t., 3 h; iii) 1.0 equiv Lawesson’s reagent, THF, 60 °C, 3.5 h; iv) 1.0 equiv 1,4-dibromobutane-2,3-dione, MeOH, r.t.−60 °C, 15 min.

To confirm the structure of bithiazolophane 1, a single crystal was prepared. A colorless cubic crystal suitable for X-ray crystallographic analysis was obtained by slow evaporation of a chloroform/methanol solution (20:80, v/v) of 1. As shown in Figure 2, 1 has a cavity in the molecule, and the bithiazole rings are stacked side-by-side and held together by short bridges. Each nitrogen atom N(3)-N(3’) and N(3”)-N(3’“) in the bithiazole moieties points the opposite side of bithiazole plain. The distance between the distal bithiazole planes is 0.36 nm. Since the distances are almost equal to van der Waals distances, each bithiazole framework is stabilized by π-π stacking.

Figure 2.

ORTEP drawing of bithiazolophane 1 showing 50% probability displacement ellipsoids.

Synthesis of bithiazolophane derivatives

Bithiazolophane 1 was converted into two derivatives using the nitrogen atoms in the linker (Scheme 2). Boc groups of 1 were cleaved by HCl or CF₃COOH (TFA) to give the corresponding salts 2 and 11, respectively. After HCl salt 2 was treated by sodium hydroxide, the neutral 2 was reacted with n-heptanoyl chloride in pyridine to give acylated bithiazolophane 3. Amino groups of TFA salt 11 were reacted with isobutyl chloroformate in the presence of base to give isobutyl carbamate 4. New compounds 1, 2, 3, 4, 9, 10, and 11 were characterized by NMR, IR, MS, and elemental analysis. Furthermore, 1 was unambiguously characterized by X-ray crystallographic analysis.

Scheme 2.

Synthesis of bithiazolophane derivatives 3 and 4 using 1 as a starting material. i) 4.0 mol/L HCl, MeOH, r.t., overnight; ii) 4.0 mol/L NaOH; iii) 4.1 equiv n-heptanoyl chloride, pyridine, 60 °C, 21 h; iv) 3.0 equiv TFA, r.t., 2.5 h; v) 3.0 equiv isobutyl chloroformate, 3.0 equiv N, N-diisopropylethylamine, CHCl₃, r.t., overnight.

Inhibitory activity against protein kinases

Inhibitory activities of bithiazolophanes 1–4 were investigated against 7 protein kinases: c-Met, EGFR WT, EGFR L858R, EGFR T790M/L858R, Abl WT, Abl T315I, and FLT1.^[25] As summarized in Table 1, bithiazolophane 1 showed anti-c-Met activity (IC₅₀ = 603 nM). Similarly, 2 and 4 also showed anti-c-Met activity among these bithiazolophane derivatives. All the synthesized ligands did not show activity against the remaining kinases with the exception of partial inhibition of EGFR WT by 1.

Table 1.

Anti-protein kinase activities (IC₅₀, μM) of bithiazolophanes 1–4.

	c-Met	EGFR WT	EGFR L858R	EGFR T790M /L858R	Abl WT	Abl T315I	FLT1
1	0.603 (59%)	> 1 (44%)	-	-	-	-	-
2	> 1 (33%)	-	-	-	-	-	-
3	-	-	-	-	-	-	-
4	> 1 (5%)	-	-	-	-	-	-

The numbers in parentheses are the percentage of inactivated c-Met and EGFR WT when 1 μM of each compound was used as an inhibitor. “-“ represents no inhibition at > 1 μM.

These results demonstrated that the bithiazolophane framework had c-Met inhibitory activity while not being active against the other studied kinases.

As can be seen from Table 1, only 1 showed high anti-c-Met activity (IC₅₀ = 603 nM) among the bithiazolophane derivatives. When the tert-butoxy groups of 1 were changed into n-hexyl or iso-butoxy groups of 3 and 4, the anti-c-Met activities of those two compounds diminished though some activity was present with 4. To understand why only 1 showed high anti-c-Met activity, SILCS-MC docking of the inhibitors into the X-ray crystal structure of c-Mets was performed.^[26,27] In addition, docking was performed on the original ligands AM7 and SU11274 (Figure 3) because our compounds have long shapes and are thought to show similar binding behavior to AM7 and SU11274. Table 2 includes the SILCS Ligand Grid Free Energy (LGFE) scores for all the ligands. LGFE is an approximate estimate of the binding affinity of ligands as previously discussed.^[28] Of the new ligands, 1 showed the most favorable predicted binding affinity, consistent with it being the most active inhibitor, with 4 having a similar LGFE score. This is consistent with it having some activity. For the remaining two compounds, 2 had the least favorable LGFE score although it was more active then 3 and 4 in vitro (Table 1). However, 1, 3, and 4 were only slightly soluble in water, whereas 2 is almost freely soluble in water. Accordingly, solubility issues may contribute to 2 having higher anti-c-Met activity in vitro than 3 and 4.

Figure 3.

The structures of protonated original ligands AM7 and SU11274.

Table 2.

SILCS-MC simulation results (kcal/mol) on the complexes of c-Met with AM7, SU11274, and bithiazolophanes 1–4.

Ligand	LGFE
AM7	−5.68
SU11274	−5.20
1	−4.96
2	−4.13
3	−4.78
4	−4.95

Protonated forms of AM7, SU11274, and 2 used.

To understand the difference in the inhibitory activity of the compounds, visualization of the SILCS-MC docked orientations of 1–4 was undertaken along with that of previously published inhibitors AM7 and SU11274. ^[12,14] Understanding how different functional groups on the ligands overlap with the different types of SILCS FragMaps yields a molecular understanding of the types of interactions that contribute to the relative affinities of the c-Met inhibitors. Many c-Met ligands occupy region 3 which is part of the protein’s active site as seen in various X-ray structures^!14−16! of c-Met (Figure 4a). AM7 and SU11274 are analyzed as they have reported IC₅₀ values of 17 nM^[12] and 12 nM^[14] for inhibition of c-Met. This is consistent with their more favorable LGFE scores as compared to 1. Figure 4b shows the bound orientations of AM7 and SU11274 in the active site of c-Met along with the SILCS FragMaps. As may be seen, both ligands penetrate deep into the interior of the protein occupying a pocket, indicated as region 3 in the image. This region contains nonpolar, hydrogen-bond donor, and hydrogen-bond acceptor FragMaps that are occupied by the terminal phenyl ring and the adjacent hydrogen bond donor and acceptor groups. The overlap of these functional groups with the FragMaps leads to the more favorable interactions of AM7 and SU11274 versus the four compounds synthesized in the present study (Table 2). AM7 and SU11274 also interact with region 2, interactions that will also contribute to their binding based on the overlap with the FragMaps. The SILCS-MC predicted orientations of 1 and 2 are shown in Figure 4c. Neither ligand is able to access the deep pocket associated with region 3. Both inhibitors occupy the lower part of region 2, leading to their favorable LGFE scores due to overlap with nonpolar FragMaps in that region. However, the Boc groups on 1, which are absent on 2, occupy region 1 and the additional non-polar FragMap in the upper part of region 2. These additional interactions lead to the better inhibitor activity of 1 allowing it to bind in an orientation that blocks access to the enzyme active site.

Figure 4.

a) The active site of the c-Met crystal structure (PDB 3vw8). b) AM7 (ball and stick) and SU11274 (stick) SILCS-MC orientations overlaid on the active site of the c-Met crystal structure (PDB 3vw8). c) 1 (ball and stick) and 2 (stick) SILCS-MC orientations overlaid on the active site of the c-Met crystal structure (PDB 3vw8). d) 3 (ball and stick) and 4 (stick) SILCS-MC orientations overlaid on the active site of the c-Met crystal structure (PDB 3vw8). Included are SILCS FragMaps for generic nonpolar (NGEN, green), Hydrogen bond donor (DGEN, blue) and hydrogen-bond acceptor (GENA, red), functional groups at a contour level of −0.9 kcal/mol.

Analysis was also done on the SILCS-MC docked orientations of 3 and 4 (Figure 4d). Interestingly, 3 binds in an orientation similar to AM7 and SU11274. The extended linear alkyl chains of 3 are long enough allowing it to bind to region 3 of the active site. However, the less favorable LGFE score of 3 versus AM7 and SU11274 indicates that it does not fully occupy region 3 and the aliphatic chain does not take advantage of the hydrogen bond donor and acceptor FragMaps in that region that are occupied by AM7 and SU11274. While 3 also occupies region 2, this does not lead to improved affinity. In addition, the conformational flexibility as well as the hydrophobicity may lead to it being a poor inhibitor of c-Met. Comparison of the bound orientations of 1, 2, and 4 shows 4 to occupy both region 1 and the upper nonpolar FragMaps in region 2 similar to 1. However, the additional methylene group in the substituents in 4 versus 1 appears to lead to a small decrease in its ability to occupy the favorable regions indicated by the nonpolar FragMaps, leading to it being a poorer inhibitor as compared to 1. The increased conformational flexibility of 4 may also contribute to it being a less effective inhibitor than 1.

Conclusion

We demonstrated that short chain-bridged heterocyclophane 1 was yielded successfully in mild conditions by a convenient synthesis using starting molecules without having aromatic frameworks. Additionally, amino groups at the short bridges of 1 were functionalized, and bithiazolophane derivatives 2–4 were synthesized successfully. Only 1 exhibited selective high anti-c-Met activity (IC₅₀ = 603 nM). SILCS-MC docking analysis revealed that the Boc groups of 1 allowed it to effectively bind to the c-Met active site as indicated by their overlap with the nonpolar FragMaps. This binding leads to the improved binding as compared to 2, which lacks substituents at those sites while the substituents in 3 and 4 cannot optimally occupy the favorable binding region of the active site leading to their poor or lack of inhibition of c-Met. Interestingly, the SILCS-predicted binding orientation of 1 indicates that it acts like a hatch to block access to the active site leading to its anti-c-Met activity, with the cyclophane framework acting as the hatch. In addition, differential solubility and conformational flexibility of the tested compounds likely contribute to changes in their ability of inhibiting c-Met. The molecular understanding of the interactions leading to the differences in binding will facilitate the design of improved c-Met inhibitors in the context of the heterocyclophane scaffold. For example, the placement of long functional groups similar to those on 3 that better exploit the binding characteristics of region 3 may lead to improved affinity. In addition, it is expected that anti-c-Met inhibitory activity would increase by using bithiazolophanes with higher water solubility. Bithiazolophanes with amino acid residues aimed at improving water solubility have already been synthesized, and their anti-c-Met activities are being investigated in our laboratory.

Experimental Section

Melting points were determined on a Yanaco (MP-J3). Infrared radiation (IR) spectra were measured by Thermo scientific (NICOLET iS5). Proton nuclear magnetic resonance (¹H NMR) spectra were recorded on Bruker biospin (Asend 400) instruments. Chemical shifts are reported as δ values (ppm) downfield from internal tetramethylsilane of the indicated organic solution or sodium 3-(trimethylsilyl)propane-1-sulfonate of the indicated water. Peak multiplicities are expressed as follows: s, singlet; t, triplet; qnt, quintet; br s, broad singlet; m, multiplet. Coupling constats (J values) are given in hertz (Hz). Mass spectra (MS) were acquired using a Jeol MALDI-TOF system (JMS-S3000SpiralTOF), Bruker daltonics LC-MS system (microTOFQ) operating in electron spray ionization mode (ESI+) and were used to confirm the purity of each compound. Chromatographic purification was carried on silica gel columns 60 F254 plate (Merck). Chromatographic purification was carried on silica gel (SiliaFlash® F60, 40–63 μm, SiliCycle). Commercial reagents and solvents were used without additional purification. Abbreviations are used as follows: [D₆]DMSO, dimethyl sulfoxide-d6; DMF, N,N-dimethylformamide; Boc, tert-butoxycarbonyl; Boc2O, di-tert-butyl dicarbonate; THF, tetrahydrofuran; NH₄OH, ammonia solution; MeOH, methanol; HCl, hydrochlolic acid; CF₃COOH (TFA), trifluoroacetic acid; NaOH, sodium oxide; CHCl₃, chloroform; EtOAc, ethyl acetate; Na₂SO₄, sodium sulfate; NaHCO₃, sodium hydrogen carbonate; D₂O, deuterium oxide; KBr, potassium bromide.

Synthesis of dimethyl 3-tert-butoxycarbonyl-3-azapentanedioate (7):

Dimethyl 3-azapentanedioate (6) (98.8 g, 743 mmol) was dissolved in THF (1.0 L), and then trimethylamine (150 mL, 1.08 mol) was added to the solution at room temperature. To the mixture was added portionwise Boc₂O (219 g, 1.00 mol) at room temperature. After 21 hours, the organic solvent was removed under reduced pressure. After the residue was dissolved in a mixture of EtOAc and water, the organic layer was separated and washed with 10% aqueous citric acid and saturated NaCl aqueous solution. The solution was dried over anhydrous Na₂SO₄, and then the solvent was removed under reduced pressure to give 7 (224 g, quant) as a colorless viscous oil. IR (KBr): v^~= 3432 (m), 2977 (s), 2947 (s), 2795 (m), 2712 (m), 2610 (m), 2407 (m), 1757 (s), 1434 (s), 1342 (s), 1235 (s), 1070 (m), 981 (m), 888 cm⁻¹ (m).

Synthesis of 3-(tert-butoxycarbonyl)-3-azapentanedicarboxamide (8):

Dimethyl 3-tert-butoxycarbonyl-3-azapentanedioate (7) (112 g, 429 mmol) was added portionwise to a concentrated annonia water (250 mL) over 3 hours, and the mixture was stirred for overnight at room temperature. The obtained precipitate was collected by suction filtration to give 8 (79.0 g, 80%) as a colorless solid. m.p. 180–181 °C; ¹H NMR (400 MHz, [D₆]DMSO): δ=1.35 (s, 9H; ^tBu-H), 3.75 (s, 4H; CH₂), 7.09, 7.14 (each s, 2H; NH₂), 8.01, 8.12 ppm (each s, 2H; NH₂); ¹³C NMR (100 MHz, [D₆]DMSO): δ=27.9 (^tBu-CH₃), 52.1 (CH₂), 52.4 (CH₂), 79.4 (^tBu-C), 154.6 (OCON), 171.8 (NH₂CO), 172.1 ppm (NH₂CO); IR (KBr): v^~=3407 (m, b), 3354 (m), 3178 (m), 2978 (w), 2932 (w), 1697 (s), 1670 (s), 1478 (m), 1460 (m), 1390 (m), 1367 (w), 1325 (w), 1272 (m), 1254 (m), 1171 (m), 963 (w), 911 (w), 860 (w), 784 (w), 610 (m), 463 cm⁻¹ (w); MS (ESI-TOF): m/z calcd for C₉H₁₇N₃O₄+H⁺: 232.13 [M+H]⁺; found: 232.13; elemental analysis calcd (%) for C₉H₁₇N₃O₄: C 46.74, H 7.41, N 18.17; found: C 46.45, H 7.45, N 17.85.

Synthesis of 3-(tert-butoxycarbonyl)-3-azapentanethiocarboxamide (9):

3-(tert-Butoxycarbonyl)-3-azapentanedicarboxamide (8) (20.0 g, 86.5 mmol) was suspended in THF (300 mL) and then Lawesson’s reagent (35.0 g, 86.5 mmol) was added portionwise to the suspention at room temperature. The mixture was heated at 60 °C for 3.5 hours with stirring. After cooling to room temperature, the organic solvent was removed under reduced pressure to give light brown oil. The residual oil was dissolved in EtOAc. The solution was washed with 4.0 mol/L sodium hydroxide aqueous solution, saturated NaHCO₃, and brine, and then dried over anhydrous Na₂SO₄. The mixture was concentrated under reduced pressure until 2/3 of the total amount of the solvent was removed. The pale yellow precipitate was obtained from this solution. The procedure was repeated twice to yield the crude products. The crude product was purified by silica gel flash column chromatography (acetone : hexane = 2 : 1, v/v) to give 9 (16.0 g, 76%) as a yellow solid. m.p. 180–181 °C; ¹H NMR (400 MHz, [D₆]DMSO): δ=1.36 (s, 9H; ^tBu-H), 4.06, 4.12 (each s, 4H; CH2), 9.34, 9.40 (each s, 2H; NH2), 9.66, 9.68 ppm (each s, 2H; NH2); ¹³C NMR (100 MHz, [D₆]DMSO): δ=27.9 (^tBu-CH₃), 58.3 (CH₂), 58.4 (CH₂), 79.5 (^tBu-C), 154.7 (OCON), 203.2 (NH₂SO), 203.8 ppm (NH₂SO); IR (KBr): v^~=3333 (s), 3301 (s), 3154 (m), 2996 (w), 2974 (w), 2931 (w), 1671 (s), 1632 (s), 1459 (s), 1395 (m), 1370 (m), 1329 (w), 1282 (m), 1253 (m), 1160 (m), 984 (w), 890 (w), 774 (w), 697 (w), 669 cm⁻¹ (w); MS (ESI-TOF): m/z calcd for C₉H₁₇N₃O₂S₂+Na⁺: 286.07 [M+Na]⁺; found: 286.07; elemental analysis calcd (%) for C₉H₁₇N₃O₂S₂: C 41.04, H 6.51, N 15.95; found: C 41.26, H 6.54, N 15.73.

Synthesis of 2,15-diaza-2,15-bis(tert-butoxycarbonyl)[3.3](2,2’)(4,4’-bithiazolophane) (1):

1,4-dibromobutane-2,3-dione (490 mg, 2.00 mmol) was added to a solution of 9 (530 mg, 2.00 mmol) in boiled anhydrous MeOH (60 mL). After the solution was kept for 15 minutes with stirring by heating, the suspension was added trimethylamine (TEA, 2 mL) and adjusted to pH 11. The organic solvent was removed under reduced pressure. The residual oil was dissolved in EtOAc, and the solution was washed with saturated NaHCO₃ and brine, and then dried over anhydrous Na₂SO₄. The mixture was concentrated under reduced pressure to afford a concentrated solution. The crude product was purified by silica gel flash column chromatography (acetone : hexane = 1 : 4, v/v) to give 1 (102 mg, 16%) as a colorless solid. m.p. 228–229 °C; ¹H NMR (400 MHz, CDCl₃): δ=1.64 (s, 18H; ^tBu-H), 4.91 (s, 4H; CH₂), 4.94 (s, 4H; CH₂), 7.30, 7.31 ppm (each s, 4H; Ar-H); ¹³C NMR (100 MHz, CDCl₃): δ=28.6 (^tBu-CH3), 49.55 (CH2), 49.63 (CH2), 50.3 (CH2), 82.1 (^tBu-C), 116.76 (Ar-C), 116.85 (Ar-C), 117.09 (Ar-C), 117.16 (Ar-C), 149.02 (Ar-C), 149.07 (Ar-C), 149.13 (Ar-C), 154.7 (CO), 166.3 (Ar-C), 166.4 (Ar-C), 166.68 (Ar-C), 166.73 ppm (Ar-C); IR (KBr): v^~=3448 (m, b), 3121 (w), 2974 (w), 2932 (w), 1700 (s), 1696 (s), 1449 (w), 1399 (m), 1366 (w), 1252 (m), 1159 (s), 1059 (w), 872 (w), 794 (w), 760 cm⁻¹ (w); MS (MALDI-TOF): m/z calcd for C₂₆H₃₀N₆O₄S₄+Na⁺: 641.1104 [M+Na]+; found: 641.1109; elemental analysis calcd (%) for C₂₆H₃₀N₆O₄S₄: C 50.46, H 4.89, N 13.58, O 10.34; found: C 50.34, H 4.73, N 13.41, O 10.43. 2,15,28-tris(tert-butoxycarbonyl)-2,15,28-triaza[3.3.3](2,2’)(4,4’-bithiazolophane) (10): a pale yellow solid in the 7% yield. m.p. 228–229 °C; ¹H NMR (400 MHz, CDCl₃): δ=1.42, 1.46, and 1.49 (each s, 27H; ^tBu-H), 4.87 and 4.92 (each s, 12H; CH2), 7.35, 7.40, 7.47, and 7.53 ppm (each s, 6H; Ar-H); ¹³C NMR (100 MHz, CDCl₃): δ=28.4 (^tBu-CH3), 49.5 (CH2), 49.8 (CH2), 50.2 (CH2), 50.5 (CH2), 81.78 (^tBu-C), 81.86 (^tBu-C), 115.6 (Ar-C), 116.3 (Ar-C), 116.6 (Ar-C), 149.9 (Ar-C), 154.9 (CO), 167.0 (Ar-C), 167.2 (Ar-C), 168.1 (Ar-C), 168.6 (Ar-C), 168.9 ppm (Ar-C); IR (KBr): v^~=3432 (s, b), 3122 (w), 2974 (w), 2928 (w), 1696 (m), 1637 (w), 1449 (w), 1400 (w), 1367 (w), 1252 (m), 1160 (m), 1113 (w), 1060 (w), 874 (w), 763 (w), 568 (w), 409 cm⁻¹ (w); MS (MALDI-TOF): m/z calcd for C₃₉H₄₅N₉O₆S₆+Na⁺: 950.17 [M+Na]⁺; found: 950.43; elemental analysis calcd (%) for C₃₉H₄₅N₉O₆S₆: C 50.46, H 4.89, N 13.58; found: C 50.18, H 4.93, N 13.47.

Synthesis of 2,15-diaza[3.3](2,2’)(4,4’-bithiazolophanium) dichloride (2):

To a solution of 1 (610 mg, 10.0 mmol) in MeOH (50 mL) was added 4.0 mol/L HCl (6.5 mL). After the solution was stirred overnight at room temperature, the solvent was removed under reduced pressure. The residual oil was dissolved in MeOH, the crude product was purified by recrystallization from water to give 2 (460 mg, 94%) as a pale blue solid. m.p. >300 °C; ¹H NMR (400 MHz, D₂O): δ=4.87 (s, 8H; CH2), 7.60 ppm (s, 4H; Ar-H); IR (KBr): v^~=3449 (s, b), 3127 (w), 2959 (w), 2925 (w), 1637 (m), 1450 (m), 1127 (m), 1036 (w), 770 (w), 555 cm⁻¹ (w); MS (MALDI-TOF): m/z calcd for C₁₆H₁₆Cl₂N₆S₄-HCl₂⁻: 419.02 [M-HCl2⁻]⁺; found: 419.34; elemental analysis calcd (%) for C₁₆H₁₆Cl₂N₆S₄+2H₂O: C 36.43, H 3.82, N 15.93; found: C 36.17, H 3.84, N 15.68.

Synthesis of 2,15-diaza[3.3](2,2’)(4,4’-bithiazolophanium) bis(trifluoroacetate) (11):

Trifluoroacetic acid (TFA, 72 μL, 0.97 mmol) was added to a solution of 1 (200 mg, 0.323 mmol). After the solution was stirred for 2.5 hours at room temperature, the solvent was removed under reduced pressure. The residual solid was dried in vacuo to give 11 (146 mg, 70%) as a pale yellow solid. m.p. >300 °C; ¹H NMR (400 MHz, D₂O): δ=4.85 (s, 8H; CH₂), 7.59 ppm (s, 4H; Ar-H); ¹³C NMR (100 MHz, D₂O): δ=48.1 (CH₂), 120.2 (Ar-C), 149.5 (COO⁻), 160.7 ppm (Ar-C); IR (KBr): v^~=3441 (s, b), 3121 (w), 2927 (w), 1667 (s), 1584 (w), 1483 (w), 1437 (m), 1314 (w), 1205 (s), 1182 (m), 1132 (s), 1025 (w), 839 (w), 798 (m), 723 cm⁻¹ (m); MS (MALDI-TOF): m/z calcd for C₂₀H₁₆F₆N₆O₄S₄-[(CF₃COO)₂H]⁻: 419.02 [M-(CF₃COO)₂H]⁺; found: 419.02; elemental analysis calcd (%) for C₂₀H₁₆F₆N₆O₄S₄: C 37.15, H 2.49, N 13.00, F 17.63; found: C 36.95, H 2.59, N 12.82, F 17.47.

Synthesis of 2,15-Diaza-2,15-diheptanoyl[3.3](2,2’)(4,4’-bithiazolophane) (3):

4.0 mol/L sodium hydroxide was added to a solution of 2 (4.75 mM, 1.2 L) in water, and the pH was adjusted to 11. After the solution was cooled in ice bowel, white powder was precipitated. The powder was filtrated to give 1.08 g of white powder 2’. n-Heptanoyl chloride (126 μl, 0.814 mmol) was added to a solution of 2’ (83.6 mg, 0.200 mmol) in anhydrous pyridine (5.0 mL). After the solution was stirred at 60 °C for 21 hours, 5 ml of pyridine was removed under reduced pressure. The residual oil was dissolved in EtOAc. The solution was washed with 10% citric acid solution, 10% Na2CO3 solution, and brine, and then dried over anhydrous Na2SO4. The mixture was concentrated under reduced pressure to afford a concentrated solution. The crude product was purified by washing with hexane to give 3 (63.1 mg, 49%) as a colorless solid. m.p. 220–22 °C; ¹H NMR (400 MHz, CDCl₃): δ=0.89 (t, J_h,h=7.1 Hz, 6H; CH₃), 1.30–1.37 (m, 8H; CH₂CH₃), 1.38–1.46 (m, 4H; CH₂CH₂), 1.82 (qnt, J_h,h=7.5 Hz, 4H; CH₂CH₂), 2.60 (t, J_h,h=7.5 Hz, 4H; CH₂CO), 4.88 (s, 2H; CH₂Ar), 4.90 (s, 2H; CH₂Ar), 4.93 (s, 4H; CH₂Ar), 7.12, 7.13, 7.14, and 7.17 ppm (each s, 4H; Ar-H); ¹³C NMR (100 MHz, CDCl₃): δ=14.2 (CH3), 22.7 (CH2), 25.34 (CH2), 25.36 (CH2), 29.3 (CH2), 31.8 (CH2), 33.4 (CH2), 48.6 (CH2), 48.8 (CH2), 50.8 (CH2), 115.0 (Ar-C), 116.0 (Ar-C), 117.0 (Ar-C), 118.0 (Ar-C), 148.1 (Ar-C), 148.7 (Ar-C), 150.0 (Ar-C), 150.6 (Ar-C), 164.4 (Ar-C), 164.6 (Ar-C), 165.1 (Ar-C), 165.5 (Ar-C), 173.4 (CO), 173.5 ppm (CO); IR (KBr): v^~=3474 (m, b), 3118 (w), 3095 (w), 2955 (m), 2926 (m), 2857 (m), 1640 (s), 1497 (m), 1450 (w), 1408 (m), 1356 (m), 1251 (m), 1163 (m), 1113 ((w), 1070 (w), 799 (w), 743 cm⁻¹ (w); MS (ESI-TOF): m/z calcd for C₃₀H₃₈N₆O₂S₄+H⁺: 643.20 [M+H]⁺; found: 643.19; elemental analysis calcd (%) forC₃₀H₃₈N₆O₂S₄: C 56.04, H 5.96, N 13.07, O 4.98; found: C 55.82, H 5.96, N 13.04, O 5.20.

Synthesis of 2,15-Bis(isobutoxycarbonyl)-2,15-diaza[3.3](2,2’)(4,4’-bithiazolophane) (4):

Isobutyl chloroformate (78 μl, 0.600 mmol) was added to a suspension of 11 (130.0 mg, 0.200 mmol ) in anhydrous chloroform (30 ml) in the presence of W,W-diisopropylethylamine (105 μl, 0.600 mmol). After the suspension was stirred overnight at room temperature, the solvent was removed under reduced pressure. The residue was dissolved in EtOAc, the solution was washed with sat. NaHCO₃ and brine, and then dried over anhydrous Na₂SO₄. The mixture was concentrated under reduced pressure to afford a concentrated solution. The white crude product 4 (63.0 mg, 51%) was obtained. m.p. 254–255 °C; ¹H NMR (400 MHz, CDCl3): δ=1.028 (d, J_h,h=6.5 Hz, 6H; CH₃),1.034 (d, J_h,h=6.5 Hz, 6H; CH₃), 2.03–2.15 (m, 2H; CH), 4.12 (d, J_h,h=6.4 Hz, 4H; OCH₂), 4.96 (s, 8H; CH₂), 7.27–7.30 ppm (m, 4H; Ar-H); ¹³C NMR (100 MHz, CDCl₃): δ=19.4 (CH₃), 28.3 (CH), 49.88 (CH₂Ar), 49.94 (CH₂Ar), 72.9 (OCH₂), 116.8 (Ar-C), 117.1 (Ar-C), 149.2 (CO), 149.3 (CO), 155.9 (Ar-C), 165.5 (Ar-C), 165.7 (Ar-C), 166.1 (Ar-C), 166.4 ppm (Ar-C); IR (KBr): v^~=3448 (s, b), 3077 (w), 2957 (m), 2933 (w), 1711 (s), 1637 (w), 1493 (w), 1466 (w), 1421 (m), 1389 (w), 1296 (w), 1265 (s), 1182 (m), 996 (m), 794 (w), 772 (w), 753 cm⁻¹ (m); MS (ESI-TOF): m/z calcd for C₂₆H₃₀N₆O₄S₄+H⁺: 619.13 [M+H]⁺; found: 619.12; elemental analysis calcd (%) for C₂₆H₃₀N₆O₄S₄: C 50.46, H 4.89, N 13.58, O 10.34; found: C 50.26, H 4.92, N 13.32, O 10.64.

Single Crystal X-ray Diffraction of bithiazolophane 1

Single crystals X-ray diffraction data were collected on a Rigaku Saturn 724+ CCD diffractometer in omega and phi scan mode, γ (MoKα) = 0.71073 Ǻ at 173 K. All the intensities were corrected for Lorentzian, polarization and absorption effects using Rigaku CrystalClear-SM Expert software.^[29] The crystal structures were solved by Direct methods using program SIR-97^[30] and the full-matrix least squares refinements on F² were carried out by using SHELXL-2014/7.^[31] All non-hydrogen atoms were refined anisotropically, and hydrogen atoms were placed in idealized positions and constrained to ride on their parent atoms. All calculations were carried out using Yadokari XG2009.^[32] Molecular figures were generated using Mercury-3.10.3 software.^[33] Crystal data for bithiazolophane 1: C₂₆H₃₀N₆O₄S₄, M = 618.80, monoclinic, a = 24.783(1) Ǻ, b = 10.467(4) Ǻ, c = 11.279(4) Ǻ, α = 90°, β = 98.336(5)°, γ = 90°, V = 2895(2) Ǻ³, T = 173 K, space group C2/c, Z = 4, μ(MoKα) = 0.372 mm⁻¹, 11498 reflections measured, 3314 independent reflections (R_int = 0.0236). The final R₁ and wR(F²) values were 0.0321 (I > 2σ(I)) and 0.0999 (I > 2σ(I)), respectively. The final R₁ and wR(F²) values were 0.0398 (all data) and 0.1086 (all data), respectively. The goodness of fit on F² was 0.821. CCDC-1981855 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

Standard Assay

1) Remove the appropriate number of microtiter wells from the foil pouch and place them into the well holder. Return any unused wells to the foil pouch, refold, seal with tape and store at 4 °C. 2) Prepare the Kinase Assay Buffer containing test chemicals and tyrosine kinase inhibitor. All assays should be done in duplicate. 3) Add 10 μL of 0.1 unit/μL Met Positive Control or serial dilution of Met Positive Control to the wells of the assay plate on ice. 4) Begin the kinase reaction by addition of 90 μL kinase Reaction buffer per well, cover with plate sealer, and incubate at 30 °C for 60 minutes. 5) Wash wells five times with Wash Buffer making sure each well is filled completely. Remove residual Wash Buffer by gentle tapping or aspiration. 6) Pipette 100 μL HRP conjugated Detection Antibody PY-39 into each well, cover with a plate sealer and incubate at room temperature (ca. 25 °C) for 60 minutes. Discard any unused conjugate. 7) Wash wells five times with Wash Buffer making sure each well is filled completely. Remove residual Wash Buffer by gentle tapping or aspiration. 8) Add 100 μL of Substrate Reagent to each well and incubate at room temperature (ca. 25 °C) for 5–15 minutes. 9) Add 100 μL of Stop Solution to each well in the same order as the previously added Substrate Reagent. 10) Measure absorbance in each well using a spectrophotometric plate reader at dual wavelengths of 450/540 nm. Dual wavelengths of 450/550 or 450/595 nm can also be used. Read the plate at 450 nm if only a single wavelength can be used. Wells must be read within 30 minutes of adding the Stop Solution.

Docking Methodology

Site-identification by ligand competitive saturation (SILCS) method was used to characterize the binding of the inhibitors to the c-Met protein.^[26,27] The crystal structure of the c-Met kinase domain (PDB ID: 3VW8) was used in the setup of the SILCS simulations after the missing regions in the 3D structure were modelled using SWISS-Model.^[34] For the SILCS simulations the protein was immersed in a solvent box containing 8 probe solutes and TIP3P water as previously described.^[28] The box size was chosen such that the protein was separated from the edges of the box with a minimum of 15 Ǻ. SILCS simulations were performed with the default protocol using CHARMM36m force field.^[35] GROMACS^[36] was used for the Molecular Dynamics simulations and an in-house code used for the Grand-Canonical Monte-Carlo (GCMC) calculations. Ten replicates, each with a total of 100 ns accumulated simulation time, were used to calculate the fragment maps (FragMaps) representing different functional group types including generic nonpolar (NGEN), generic hydrogen-bond donor (DGEN), generic hydrogen-bond acceptor (AGEN), positive groups (DPOS), and negative groups (ANEG). SILCS-MC calculations for docking were performed using the generic apolar scoring scheme along with the exhaustive search algorithm with a 10 A radius cutoff from the center of the binding site that was identified in the crystal structure. Additional details of the SILCS-MC docking may be found in Ustach et al.^[36]