Synthesis and Structure-Activity Relationship of Boronic Acid Bioisosteres of Combretastatin A-4 as Anticancer Agents

Abstract

The boronic acid group plays an important role in drug discovery. Following our discovery of a boronic acid analog of combretastatin A-4 (CA-4), a series of analogs featuring a boronic acid group on the C phenyl ring of CA-4 was synthesized and evaluated for cytotoxicity, as well as for their ability to inhibit tubulin polymerization, inhibit the binding of [³H]colchicine to tubulin and cause depolymerization of cellular microtubules. Modifications on the C ring of CA-4, either eliminating the methoxy group or replacing the C phenyl ring with a pyridine ring, resulted in a reduced potency for inhibiting tubulin polymerization, colchicine binding and cytotoxic activities as compared to CA-4. Replacing the phenol group with a boronic acid group on the C ring of phenstatin led to a slight increase in cytotoxic potency but a decreased potency for inhibition of tubulin assembly and colchicine binding. Moreover, there was a significant decrease in activity by replacing the C phenyl ring with a pyridine ring. Our results indicate the critical importance of the methoxy group on the C ring as well as the importance of the C phenyl ring compared to a pyridine ring, despite the latter providing a nitrogen atom as a hydrogen bond donor/acceptor, which was predicted by molecular modeling to enhance interaction with the target. The decreased activities of our modified CA-4 boronic analogs may be attributed to weakened hydrogen bonding in our docking model based on the crystal structure of colchicine bound to αβ-tubulin. Notably, even though their effectiveness in inhibiting tubulin polymerization and colchicine binding and causing microtubule depolymerization in cells, the majority of these boronic acid analogs exhibited substantial cytotoxicity. This suggests that they may have additional cellular targets that contribute to their cytotoxicity, and this warrants further evaluation of these unique boronic acid compounds as potential anticancer agents.

Graphical Abstract

Introduction

Combretastatin A-4 (CA-4), a natural product isolated from the South Africa bush willow tree by Pettit and coworkers, has demonstrated potent antiproliferative activity across a diverse spectrum of cancer cell lines.¹ The water soluble phosphate prodrug of CA-4, known as CA-4P, entered phase II/III clinical trials,^{2,3, 4} but it failed due to lack of efficacy. This has led to the synthesis and development of numerous of CA-4 analogues and description of structure-activity relationships (SARs) (Figure 1).^5–9 The SAR studies have revealed that the cis-configuration of CA-4 and an appropriate distance between two aromatic rings are considered to be the essential structural requirements for the targeting of tubulin and associated cytotoxicity.^10,11 The functional groups on ring C of CA-4 play a key role in interacting with tubulin and inhibiting its assembly. However, modifications to the functional group on ring C, involving various smaller substituents such as fluorine, chlorine and amino groups have been tolerated, leading to increasing the potency for the interaction with tubulin and subsequent cytotoxic effects.^12,13

Figure 1.

Structures of microtubule targeting agents.

We discovered that the boronic acid analogue 1 (Figure 2) of CA-4 was a potent inhibitor of tubulin polymerization with low-nanomolar IC₅₀ values in a panel of human cancer cell lines.¹⁴ Boronic acid-derived inhibitors are of considerable interest for drug development due to their stability and physical properties (pKa of 9-10).^15–17 Boronic acid inhibitors remain largely protonated under physiological conditions. Hydrogen bonds as well as boron-nitrogen bonds can be formed to enhance the interactions between the ligand and the protein receptor. To further investigate the features of this group and to obtain SAR information, we designed and synthesized a series of boronic acid analogues of CA-4 and phenstatin, based on our first generation boronic acid analog 1 (Figure 2). To improve water solubility, we eliminated the methoxy group in the C phenyl ring, as well as replaced the C phenyl ring with a pyridine ring. We calculated the physicochemical properties of these molecules using the Molinspiration database (https://www.molinspiration.com) (Table 1). These structural modification align well with Lipinski’s rule of five for drug-like properties. PSA values are suggested below 90 as consideration for an aromatic inhibitor to be further developed. Compound 22 with a PSA value of 107 may explain its lower activity in our biological evaluation (Table 3). We evaluated these new analogues for their ability to inhibit tubulin polymerization, [³H]colchicine binding to tubulin, and cytotoxicity in MCF-7 breast cancer cells, as well as for their ability to cause microtubule depolymerization in A-10 cells (Table 3).

Figure 2.

Design of boronic acid analogues of CA-4 and phenstatin.

Table 1.

Physical property calculations for 8a-c, 13, 18 and 22.

Compound	Mol. Weight (Daltons)	Log P	TPSA (Å2)	Volume (Å3)	H-bond donor	H-bond acceptor
8a	314.15	2.94	68.16	313.58	2	5
8b	314.15	3.15	68.16	313.58	2	5
8c	314.15	3.17	68.16	313.58	2	5
13	345.16	2.08	90.28	334.97	2	7
18	346.14	2.05	94.49	330.69	2	7
22	347.13	1.16	107.36	326.54	2	8
1	344.17	3.16	77.39	339.13	2	6
CA-4	316.35	3.47	57.16	293.07	1	5
Phenstatin	318.32	2.36	74.23	284.64	1	6

Table 3.

SAR of Boronic Acid Analogs

Compound	Tubulin Assemblya IC50 (μM) ± SD	MCF-7b IC50 (μM) ± SD	[3H]Colchicine Bindingc (% ± SD)		A-10 EC50 (μM)d
			5 μM	50 μM
8a	>40	>2.5	N/D	N/D	51
8b	27 ± 4	0.85 ± 0.1	N/D	N/D	9
8c	10 ± 0.2	0.48 ± 0.1	6.4 ± 9	77 ± 4	4
13	30 ± 0.4	0.39 ± 0.1	N/D	N/D	9
18	12 ± 0.6	0.30 ± 0.03	21 ± 2	59 ± 5	5
22	>40	>2.5	N/D	N/D	>50
1	1.5 ± 0.2e	0.017 ± 0.005e	79 ± 2e	99 ± 2e	0.058e
Phenstatin	1.0 ± 0.2f	0.44 ± 0.07g	86f	N/D	N/D
CA-4	2.0 ± 0.1e	0.032 ± 0.02e	99 ± 2e	N/D	<0.025e

N/D, no data.

^aInhibition of tubulin polymerization. Tubulin was at 10 μM.

^bInhibition of growth of MCF-7 human breast carcinoma cells.

^cInhibition of colchicine binding. Compounds were tested at 5 and 50 μM. The tubulin concentration was 1 μM, and the [³H]colchicine concentration was 5 μM. All experiments were performed in triplicate.

^dThe percentage of cellular microtubule loss was estimated visually over a range of concentrations. Dose response curves were generated, and EC₅₀ values were calculated. The values represent the results of two independent experiments.

^eReference¹⁴.

^fReference⁵.

^gReference²¹.

Results and Discussion

Chemistry

The boronic acid analogues 8a-c and 13 were synthesized using the Wittig reaction methodology (Scheme 1). The preparation of phosphonium bromide 5 started from 3,4,5-trimethoxybenzaldehyde 2 through a sequence involving reduction, bromination, and reaction with triphenylphosphine to provide phosphonium salt 5 in quantitative yields. The Wittig reaction of 5 with commercially available aldehyde 6a-c yielded mixtures of cis- and trans-stereoisomers of stilbene 7a-c, in which the cis-isomer possesses a higher R_f value than the trans-isomer.¹⁴ These stereoisomers were subsequently separated via column chromatography to afford the cis-stereoisomers 7a-c in 40-50% yields. The cis-stereoisomers 7a-c were treated with n-butyllithium and trimethylboronate, followed by acidification, resulting in the formation of the final products 8a-c in 43-53 % yield. Boronic acid compound 13, featuring a pyridine ring in place of the phenyl ring as ring C was synthesized from 5-bromo-2-methoxypyridine 9. The preparation involved sequential lithiation,¹⁸ bromination and a Wittig coupling reaction, followed by a procedure identical to the one used for the preparation of 8a-c (Scheme 2). The configuration of the cis-isomer was confirmed by the J coupling value of double bond proton NMR, which was 11-12 ppm, in the final products 8a-c and 13. Phenstatin analogues 18 and 22 were obtained starting from 3,4,5-trimethoxybenzene bromide 14, through a Grignard reaction with 3-bromo-4-methoxybenzaldehyde. This led to the formation of alcohol 15, and the alcohol group of 15 was protected with a TBS group, yielding compound 16. Through lithiation and acidification using 3 N HCl, compound 17 was obtained in a low yield. The final product 18 was achieved by oxidation of 17 via pyridinium dichromate (PDC) as illustrated in Scheme 3. Pyridine boronic acid phenstatin analog 22 was synthesized in a similar manner to compound 18, in moderate yields (Scheme 4). It is important to note that the use of a strong acid (3 N HCl) for deprotecting both TBS and borate ester protecting groups to generate 17 and 21 often leads to the formation of dimers or polymer-like materials, which subsequently complicates purification. As an alternative approach, crude preparations of 17 and 21 can be directly used up to the oxidation step, with final purification achieved through column chromatography, providing the final boronic phenstatin analogs 18 and 22 in good yields.

Scheme 1.

Synthesis of boronic acid analogues of CA-4

Scheme 2.

Synthesis of boronic acid analogue with pyridine ring as ring C of CA-4

Scheme 3.

Synthesis of boronic acid analogue of phenstatin

Scheme 4.

Synthesis of boronic acid analogue with pyridine ring of phenstatin

Molecular modeling

AutoDock Vina 1.2.0 was used to perform molecular docking of compounds into a high-resolution tubulin heterodimer structure (PDB ID: 6XER). Docking boron-containing compounds remains a challenge due to a lack of reasonable parameters for boron in most docking software. Therefore, all boron atoms were changed to carbon atoms in our docking model.^{19, 20} The docking position with the lowest dock score (or lowest binding energy) and orientated in a similar manner as colchicine was chosen for each compound. Docking score, number of hydrogen bonds, hydrogen bonding position and hydrogen bonding length are summarized in Table 2.

Table 2.

Molecular docking results of compound 8a-c, 13, 18 and 22.

Compound	Dock Score	H-Bond Number	H-Bond Position	H-Bond Length (Å)
Colchicine	−10.4	2	Carbonyl of Ring C to Val 181 α chain	1.972
			Second methoxy of Ring A to Cys 239 ß chain	2.737
CA-4	−7.0	1	Second methoxy of Ring A to Cys 239 ß chain	3.154
1	−7.2	2	Second methoxy of Ring A to Cys 239 ß chain	2.993
			Third methoxy of Ring A to Cys 239 ß chain	3.547
8a	−6.9	1	Second methoxy of Ring A to Cys 239 ß chain	2.639
8b	−7.2	1	Second methoxy of Ring A to Cys 239 ß chain	3.140
8c	−7.2	1	Second methoxy of Ring A to Cys 239 ß chain	2.972
13	−7.0	1	Third methoxy of Ring A to Cys 239 ß chain	3.582
Phenstatin	−7.4	2	Second methoxy of Ring A to Cys 239 ß chain	2.509
			Carbonyl to Lys 252 ß chain	3.067
18	−8.0	3	Second methoxy of Ring A to Cys 239 ß chain	2.640
			Carbonyl to Asn 256 ß chain	2.650
			Hydroxyl of Ring C to Ala 248 ß chain	2.368
22	−7.3	2	Second methoxy of Ring A to Cys 239 ß chain	2.608
			Carbonyl to Asn 256 ß chain	2.761

Colchicine was re-docked to confirm the docking model that resulted in the lowest score of −10.4 (or a binding energy of −10.4 kJ/mol) among all the compounds (Supplementary S1). Docking compound 1 formed two hydrogen bonds between the A ring and cysteine 239 of ß-tubulin (Figure 3A), while only one hydrogen bond was observed for compounds 8a-c and 13 (Figure 3B: 8c. Supplementary S2–4: 8a, 8b and 13). The presence of one less hydrogen bond in 8a-c may contribute to their reduced activity as inhibitors of tubulin assembly and MCF-7 cell growth as compared to compound 1. The weaker hydrogen bond for compound 13 compared to 8a-c is consistent with its lower activity as shown in Table 3. The docking model for phenstatin had two hydrogen bonds (Figure 3C), one between cysteine 239 of ß-tubulin and the second methoxy group of the A ring, the other between lysine 252 of ß-tubulin and the carbonyl group (Figure 3C). This indicated the stability of phenstatin in the binding site and was consistent with its observed activity in inhibiting tubulin polymerization and MCF-7 proliferation (Table 3). The docked model of compound 18 overlapped with phenstatin (Figure 3D), having three hydrogen bonds, which increased its stability with a lower docking score of −8.0 as compared to −7.4 for phenstatin. However, the indicated increased stability in the model of compound 18, due to a greater number of hydrogen bonds, was not reflected in improved inhibition of tubulin polymerization as compared to phenstatin. Similarly, the model of compound 22 had two hydrogen bonds between cysteine 239 of ß-tubulin and the methoxy group of the A ring and another between asparagine 256 of the ß chain and the carbonyl group (Supplementary S5), but compound 22 had substantially reduced potency when compared to compounds 18 and phenstatin, indicating the importance of the C phenyl as compared with the pyridine ring. Nevertheless, generally, the docking scores generated in our modeling demonstrated a good correlation between the ability of compounds to inhibit tubulin polymerization and MCF-7 cell proliferation (Figure 4).

Figure 3.

Molecule docking model of boronic acid analog in α-tubulin (cyan) and β-tubulin (orange). Protein originates from the X-ray structure of colchicine complex with tubulin (PDB ID: 6XER). A. Docking of compound 1 (blue). B. Docking of compound 8c (red) overlap with compound 1 (blue). C. Docking of phenstatin (gray). D. Docking of compound 18 (tan) overlap with phenstatin (gray).

Figure 4:

Correlation between the ability of compounds to inhibit tubulin polymerization (Tubulin IC₅₀) and inhibit MCF7 growth (MCF7 IC₅₀).

Biological Data

All the synthesized novel boronic acid analogues of CA-4 and phenstatin were evaluated for these four activities: (i) inhibition of tubulin polymerization, (ii) inhibition of [³H]colchicine binding to tubulin, (iii) effects on cellular microtubules using immunofluorescence in A-10 cells, and (iv) cytotoxicity in MCF-7 breast cancer cells (Table 3).

Our original compound 1 exhibited a somewhat higher potency than CA-4 as an inhibitor of tubulin polymerization and displayed a low nanomolar inhibition of proliferation against a wide range of human cancer cells. Considering that a boronic acid group is larger than a phenol group and that the 4’-methoxy group on ring C of compound 1 does not participate in the interaction with tubulin,¹⁴ we synthesized 8a-c to eliminate the 4’-methoxy and to substitute a boronic acid group at different positions of the C phenyl ring. Tubulin polymerization inhibition evaluation showed compounds 8a-c were less potent than the lead compound 1, with approximately 6-18-fold lower potency (Table 3). Compared to 8a and 8b, compound 8c with the boronic acid group at the para position is the best inhibitor of tubulin polymerization of this series (IC₅₀, 10 μM). At 50 μM, 8c inhibited colchicine binding by 77%. Both 8b and 8c demonstrated inhibition of MCF-7 human breast carcinoma cell proliferation with IC₅₀ values below 1 μM. The data indicate the importance of the methoxy group on the C ring in maintaining the appropriate molecular orientation within the binding site. Compound 13 containing a pyridine ring was designed to replace the C phenyl ring to enhance the interaction with tubulin and improve solubility,¹⁴ but compound 13 had 20-fold lower potency compared to 1 in inhibiting tubulin polymerization, and 23-fold lower activity in the inhibition of MCF-7 cell proliferation. This could be caused by the larger size of the pyridine ring and higher binding energy compared to the phenyl ring, potentially altering the molecular orientation and resulting in reduced activity. A similar trend was observed with compounds 18 and 22, confirming the importance of the phenyl ring over the pyridine ring. Phenstatin analogue 18, with the boronic acid group replacing a hydroxy group, exhibited a slightly increased toxicity against the MCF-7 breast cancer cell line²¹ relative to phenstatin. It was less potent than phenstatin in inhibiting tubulin polymerization, colchicine binding and cellular microtubule depolymerization.⁵ Compound 22, with the pyridine ring as ring C, displayed no activity, further indicating that incorporating a pyridine ring in this series was not well-tolerated. Despite the decreased activity of 13 and 18 in inhibiting tubulin polymerization and [³H]colchicine binding, both compounds exhibited mid-nanomolar cytotoxicity in MCF-7 cells. This implies the involvement of another cellular target beyond the interaction with tubulin and warrants further mechanistic investigations.²²

Indirect immunofluorescence was used to visualize the effects of analogues 8c, 13 and 18 on intracellular microtubules in A-10 cells (Figure 5). The results yielded estimates of the effects of these compounds, as shown in Table 3. In the vehicle control group (Figure 5A), the microtubule network was clearly visible. Upon treating the cells with compounds 8c, 13 or 18, disrupted microtubules and loss of the microtubule network in A10 cells were observed (Figure 5A–D, respectively).

Figure 5.

Effects of 8c, 13 and 18 on interphase microtubules in A-10 cells. (A) vehicle control; (B) 20 μM compound 8c; (C) 15 μM compound 13; (D) 2.5 μM compound 18.

We further evaluated the cytotoxic selectivity of these compounds in another two cancer cell lines and a normal cell line using Promega’s CellTiter assay kit (Supplementary material S6). In normal human pancreatic ductal epithelial cells (HPDE), all of the compounds showed low toxicity (IC₅₀ > 50 μM). In colon cancer cell line HCT-16, compounds 8a-c and 13 showed low micromolar inhibition (IC₅₀ < 5 μM), with slightly decreased toxicity compared to lead compound 1 and CA-4. The tested compounds were less toxic in pancreatic cancer cells (MiaPaca-2) than in HCT-16, with only compound 8a and 18 exhibiting potency comparable to that of CA-4 and phenstatin and less potent than compound 1. Moreover, compound 22, with the phenyl ring replaced with a pyridine ring exhibited dramatically reduced activity in both HCT-16 and MiaPaca-2 cells, consistent with its inactivity in the tubulin-based assays. The tested compounds (except compound 22) demonstrated a positive in vitro cytotoxicity selectivity index (SI) window (Supplementary material S6). The cytotoxicity data obtained from these three cell lines showed a similar trend as was observed with the data obtained with the MCF-7 cells (Table 3).

In conclusion, these boronic acid analogs of CA-4 inhibited proliferation of MCF-7 cells, but their potency as microtubule targeting agents was compromised in the absence of a methoxy group on ring C or by replacing the C phenyl ring with a pyridine ring. Compound 18, a boronic acid analog of phenstatin, demonstrated a modest increase in cytotoxicity compared to phenstatin, but 18 was less effective in inhibiting tubulin polymerization and colchicine binding than was phenstatin. Replacing the phenyl ring with a pyridine ring as ring C of boronic acid analogues of CA-4 and phenstatin proved to be not well-tolerated. These modifications caused a decreased activity compared to CA-4 and our lead compound 1, although they increased the water solubility due to their physical structural properties. To our knowledge, this is the first report to explore boronic acid CA-4 analogues and to use a pyridine ring to replace the C phenyl ring. These findings provide valuable insights for the future design of more potent tubulin inhibitors, and they provide a basis for further investigation of the unique boronic acid group in drug discovery.

Experimental

Chemistry General Methods.

NMR spectra were recorded using a Varian-300 spectrometer for ¹H (300 MHz) and ¹³C (75 MHz). Chemical shifts (δ) are given in ppm downfield from tetramethylsilane as internal standard, and coupling constants (J values) are in hertz. Mass spectra data were collected on a Finnigan MAT LC-Q mass spectrometer system using electrospray ionization (ESI). Elemental analyses were performed by Atlantic Microlabs. Melting points were obtained using an electrothermal digital melting point apparatus and are uncorrected. All purifications by flash chromatography were performed using Geduran silica gel 60, 35-75 μm (VWR). All solvents used were purified by pressure filtration through activated alumina. Reactions were run at ambient temperature and under a nitrogen atmosphere unless otherwise noted and were monitored by TLC used Merck 60 F₂₅₄ silica gel aluminium sheets (20 × 20 cm).

(3,4,5-Trimethoxyphenyl)methanol (3).

Sodium borohydride (1.06 g, 28.0 mmol) was added to a solution of 3,4,5-trimethoxybenzaldehyde (5.0 g, 25.5 mmol) in absolute methanol (50 mL) and stirred for 30 min. After completion of the reaction, water (15 mL) was added, and the solution was extracted with ethyl acetate (EtOAc) (3 × 30 mL). The combined extracts were washed with brine (30 mL) and dried over sodium sulfate. Rotary evaporation afforded alcohol 3 as a clear oil (4.73 g, 94 %): ¹H NMR (CDCl₃, 300 MHz) δ 6.47 (s, 2H), 4.48 (s, 2H), 3.72 (s, 9H), 3.24 (s, 1H).

5-(Bromomethyl)-1,2,3-trimethoxybenzene (4).

The alcohol 3 (4.7 g, 23.7 mmol) was dissolved in dry CH₂C1₂ (40 mL) and cooled to 0 °C. Phosphorus tribromide (2.25 mL, 23.7 mmol) was added dropwise, and the reaction mixture was warmed to room temperature and stirred overnight. The reaction was quenched with 5% NaHCO₃ (15 mL), and the solution was extracted with CH₂Cl₂ (3 × 15 mL). The combined extracts were washed with brine (30 ml) and dried over sodium sulfate. The CH₂Cl₂ layer was filtered and concentrated by rotary evaporation, and the residue was purified by flash column over silica using 5% EtOAc/hexanes to give a white solid 4 (3.7 g, 60 %): ¹H NMR (CDCl₃, 300 MHz) δ 6.62 (s, 2H), 4.47 (s, 2H), 3.87 (s, 6H), 3.85 (s, 3H).

3,4,5-Trimethoxybenzyltriphenylphosphonium bromide (5).

Triphenylphosphine (4.1 g, 15.6 mmol) was added to a solution of 4 (3.1 g, 12.0 mmol) in dry THF (30 mL). The mixture was refluxed with stirring for 6 h. The resulting white solid was filtered and washed with ether/hexanes to afford the product 5 in quantitative yield: ¹H NMR (CDCl₃, 300 MHz) δ 7.78 (s, 15H), 6.23 (s, 2H), 5.05 (d, 2H, J = 15 Hz), 3.59 (s, 3H). 3.36 (d, 6H, J = 15 Hz).

General procedure (A) for the stilbene syntheses (7a-c, 12).

To the suspension of phosphonium bromide 5 (2-4 mmol) in dry THF (10-20 mL) at −78 °C was added sodium bis(trimethylsilyl) amide (1.0 M in THF, 1.1 equiv.), and the resulting mixture was stirred for 2-3 h. A solution of aldehyde (1.1 equiv.) in THF was added dropwise, and the mixture was stirred for 5 h. The resulting solution was poured into a saturated NH₄Cl solution and extracted with ethyl acetate. The combined organic layers were washed with water and brine and dried over sodium sulfate. Removal of the solvent in vacuo afforded a sticky oil. The oil was separated by flash column chromatography (hexanes/ethyl acetate). The cis-stilbene eluted first as the major peak followed by the trans-isomer.

General procedure (B) for the boronic acid syntheses (8a-c, 13, 17, 21).

To the solution of the corresponding cis-stilbene (7a-c, 12) or bromo-substituted aromatic compound (16, 20) in dry THF (20 mL) at −78 °C was added n-butyllithium (1.6 M in hexane, 1.5 equiv). The resulting mixture was stirred for 1-2 h, followed by the dropwise addition of trimethyl borate (1.5 equiv.). The mixture was warmed to room temperature and stirred overnight. The reaction was quenched with 3 N HCl (10 mL) and stirred at room temperature for 30 min, followed by washing with ether (3 × 30 mL). The combined ether layers were extracted with 1 N NaOH (2 × 20 mL), and the basic layers were re-acidified with 1 N HCl to pH = 4. After re-acidification, the mixture was extracted with ether (2 × 30 mL), and the ether layers were combined. The combined ether layers were dried over Na₂SO₄, filtered, and concentrated to afford a pale yellow oil. The oil was purified by flash column chromatography (CH₂Cl₂/MeOH).

5-[(Z)-2-(2-Bromophenyl)vinyl]-1,2,3-trimethoxybenzene (7a).

General procedure A: Eluent ethyl acetate / hexanes 1:20; pale yellow oil, 0.49 g (54 %); ¹H NMR (CDCl₃, 300 MHz) δ 7.61 (dd, 1H, J = 1.4, 1.4 Hz), 7.26 (m, 1H), 7.12 (m, 2H), 6.58 (s, 2H), 6.39 (s, 1H), 6.35 (s, 1H), 3.84 (s, 3H), 3.81 (s, 3H), 3.60 (s, 3H); ¹³C NMR (75 MHz) δ 152.7, 138.1, 137.2, 133.4, 132.5, 131.4, 131.1, 130.9, 128.8, 128.6, 127.0, 123.6, 106.6, 105.7, 60.6, 55.9, 55.5.

5-[(Z)-2-(3-Bromophenyl)vinyl]-1,2,3-trimethoxybenzene (7b).

General procedure A: Eluent ethyl acetate/hexanes 1:20; pale yellow oil, 0.40 g (44 %); ¹H NMR (CDCl₃, 300 MHz) δ 7.61 (s, 1H), 7.34 (d, 1H, J = 8.1), 7.22 (d, 1H, J = 7.8), 7.13 (m, 1H). 6.52 (d, 1H, J = 4.8), 6.50 (s, 2H), 6.40 (s, 1H), 3.84 (s, 3H), 3.81 (s, 3H), 3.68 (s, 3H); ¹³C NMR (75 MHz) δ 152.8, 139.4, 137.4, 133.4, 131.6, 131.3, 129.9, 129.7, 128.0, 127.4, 122.0, 106.0, 105.8, 105.6, 61.0, 55.8, 55.7.

5-[(Z)-2-(4-Bromophenyl)vinyl]-1,2,3-trimethoxybenzene (7c).

General procedure A: Eluent ethyl acetate/hexanes 1:20; pale yellow oil, 0.29 g (40 %); ¹H NMR (CDCl₃, 300 MHz) δ 7.38 (d, 2H, J = 8.7 Hz), 7.17 (d, 2H, J = 8.4), 6.52 (dd, 2H, J = 12.3, 12.3 Hz ), 6.45 (s, 1H), 6.40 (s, 1H), 6.35 (s, 1H), 3.85 (s, 3H), 3.83 (s, 3H), 3.69 (s, 3H); ¹³C NMR (75 MHz) δ 152.9, 136.1, 133.4, 132.0, 131.2, 130.8, 130.5, 128.4, 120.8, 105.8, 60.7, 55.8, 55.71.

{2-[(Z)-2-(3,4,5-trimethoxyphenyl)vinyl]phenyl}boronic acid (8a).

General procedure B: Eluent CH₂Cl₂/MeOH 100:1.5; sticky oil, 0.19 g (43 %); ¹H NMR (CDCl₃, 400 MHz) δ 7.98 (d, 1H, J = 7.3 Hz), 7.37 (m, 3H), 6.96 (d, 1H, J = 12.0 Hz, cis-double bond H), 6.68 (d, 1H, J = 12.0 Hz, cis-double bond H), 6.31 (s, 2H), 5.07 (s, 2H), 3.79 (s, 3H), 3.55 (s, 6H). ¹³C NMR (CDCl₃, 100 MHz) δ 152.7, 143.6, 135.8, 132.9, 131.4, 130.5, 128.8, 127.0, 106.6, 60.8, 55.6. ESI m/z 315 (M + H)⁺. Anal. Calcd. for C₁₇H₁₉BO₅.0.3MeOH: C, 64.1; H, 6.2. Found: C, 64.2; H, 6.0.

{3-[(Z)-2-(3,4,5-trimethoxyphenyl)vinyl]phenyl}boronic acid (8b).

General procedure B: CH₂Cl₂/MeOH 100:1.5; sticky oil, 0.14 g (45 %); ¹H NMR (CDCl₃, 400 MHz) δ 8.10 (s, 1H), 7.98 (d, 1H, J = 7.0 Hz), 7.54 (d, 1H, J = 7.2 Hz ), 7.39 (m, 1H), 6.71 (d, 1H, J = 12.1 Hz, cis-double bond H), 6.59 (d, 1H, J = 12.0 Hz, cis-double bond H), 6.47 (s, 2H), 3.81 (s, 3H), 3.67 (s, 6H). ¹³C NMR (CDCl₃, 100 MHz) δ 152.9, 137.2, 137.0, 136.1, 134.4, 133.2, 132.4, 130.4, 129.8, 127.8, 106.0, 60.9, 55.8. ESI m/z 315 (M + H)⁺. Anal. Calcd. for C₁₇H₁₉BO₅.0.3MeOH: C, 64.13; H, 6.24. Found: C, 64.30; H, 6.24.

{4-[(Z)-2-(3,4,5-trimethoxyphenyl)vinyl]phenyl}boronic acid (8c).

General procedure B: Eluent ethyl acetate/hexanes 1:4; white solid, 0.11 g (53 %): mp 130 - 132 °C; ¹H NMR (CDCl₃, 400 MHz) δ 8.11 (d, 2H, J = 7.8 Hz), 7.65 (d, 1H, J = 7.9 Hz), 7.42 (d, 2H, J = 7.7 Hz), 7.34 (d, 1H, J = 7.8 Hz), 6.66 (d, 1H, J = 12.2 Hz, cis-double bond H), 6.61 (d, 1H, J = 13.8 Hz, cis-double bond H), 6.50 (s, 2H), 3.85 (s, 3H), 3.66 (s, 6H). ¹³C NMR (CDCl₃, 100 MHz) δ152.9, 141.9, 137.3, 135.5, 133.5, 132.3, 131.3, 129.8, 128.6, 106.0, 60.9, 55.9. ESI m/z 315 (M + H)⁺. Anal. Calcd. for C₁₇H₁₉BO₅: C, 65.00; H, 6.10. Found: C, 65.14; H, 6.07.

6-Methoxynicotinaldehyde (10).

²³ To a stirred solution of 5-bromo-2-methoxypyridine 9 (6.9 g, 36.7 mmol) in dry THF (30 mL) at −78 °C was added dropwise n-butyllithium (24.1 mL, 38.5 mmol, 1.6 M in hexane). The resulted mixture was stirred for 1 h, then a solution of DMF (5.38 g, 73.4 mmol) in THF (5 mL) was added slowly at −78 °C and stirred for a further 1 h. The reaction was quenched with aqueous NaHCO₃ (5%, 150 mL) and extracted with ether (3 × 50 mL). The combined organic layers were washed with brine and dried over Na₂SO₄. Removal of the solvent afforded a yellow solid, and the solid was purified by flash column chromatography (5% ethyl acetate/hexanes) to give the product 10 (3.8 g, 76%) as a white solid. ¹H NMR (CDCl₃, 300 MHz) δ 9.96 (s, 1H), 8.64 (d, 1H, J = 2.4 Hz), 8.07 (dd, 2H, J = 2.4, 2.4 Hz), 6.85 (d, 1H, J = 8.7 Hz), 4.04 (s, 3H); ¹³C NMR (75.5 MHz) δ 189.59, 167.75, 152.96, 137.48, 126.75, 121.15, 54.37.

5-Bromo-6-methoxynicotinaldehyde (11).

To a stirred suspension of AcONa (0.60 g, 7.29 mmol) and 6-methoxynicotinaldehyde 10 (1.0 g, 7.29 mmol) in AcOH (5 mL) was added bromine (1.22 g, 7.61 mmol). The reaction mixture was stirred at room temperature for 1 h, and then refluxed for 2 h. The resulted solution was cooled, poured into ice-water, neutralized with NaOH, and extracted with ether. The combined ether layer was washed by saturated Na₂S₂O₃ and brine and dried over Na₂SO₄. Removal of the solvent gave a dark brown solid, and the solid was purified by flash column chromatography to yield product 11 (0.22 g, 14%) as a white solid: mp 84 - 85 °C; ¹H NMR (CDCl₃, 300 MHz) δ 9.93 (s, 1H), 8.57 (d, 1H, J = 1.8 Hz), 8.3 (d, 1H, J = 1.8 Hz), 4.12 (s, 3H); ¹³C NMR (75.5 MHz) δ 188.36, 163.60, 150.58, 140.32, 127.90, 108.59, 55.70.

3-Bromo-2-methoxy-5-[(Z)-2-(3,4,5-trimethoxyphenyl)vinyl]pyridine (12).

General procedure A: Eluent ethyl acetate/hexanes 1:10; pale yellow solid, 0.23 g (26%); ¹H NMR(CDCl₃, 300 MHz) δ 8.01 (d, 1H, J = 1.8 Hz), 7.80 (d, 1H, J = 2.1 Hz), 6.57 (d, 1H, J = 12.3 Hz), 6.49 (s, 2H), 6.38 (d, 1H, J = 12.3 Hz), 3.99 (s, 3H), 3.86 (s, 3H), 3.73 (s, 6H).

{2-Methoxy-5-[(Z)-2-(3,4,5-trimethoxyphenyl)vinyl]pyridin-3-yl}boronic acid (13).

General procedure B: Eluent ethyl acetate/hexanes 1:4; sticky yellow oil, 0.13 g (81%); recrystalized from hexanes/ethyl acetate to yield white needle crystals: mp 104 - 105 °C; ¹H NMR (CDCl₃, 400 MHz) δ 8.17 (d, 1H, J = 2.0 Hz), 8.12 (d, 1H, J = 2.0 Hz ), 6.57 (d, 1H, J = 12.1 Hz, cis-double bond H), 6.48 (s, 2H), 6.46 (d, 1H, J = 12.2, cis-double bond H), 5.75 (s, 2H), 4.02 (s, 3H), 3.84 (s, 3H), 3.71 (s, 6H). ¹³C NMR (CDCl₃, 100 MHz) δ 166.3, 153.1, 150.1, 146.7, 137.5, 132.2, 130.9, 127.1, 125.7, 105.9, 60.9, 56.0, 53.9. ESI m/z 346 (M + H)⁺. Anal. Calcd. for C₁₇H₂₀BNO₆: C, 59.16; H, 5.84; N, 4.06. Found: C, 59.00; H, 5.91; N, 4.01.

(3-Bromo-4-methoxyphenyl)-(3,4,5-trimethoxyphenyl)methanol (15).

To one fourth of a solution of 5-bromo-1,2,3-trimethoxybenzene 14 (1.0 g, 4.05 mmol) in THF (20 mL) was added magnesium turnings (0.098 g, 4.05 mmol) with a small piece of iodine. As soon as the solution became colorless, the remaining solution of 14 was added dropwise and was further stirred at room temperature for 1 h to prepare the Grignard reagent. This Grignard reagent was slowly added to the solution of 3-bromo-4-methoxybenzealdehyde (0.73 g, 3.38 mmol) in THF (2.5 mL) at 0 °C. The reaction mixture was stirred at room temperature for 1 h. A saturated NH₄Cl solution (10 mL) was added slowly, and the mixture was stirred for 30 min. The reaction mixture was extracted with ether (3 × 20 mL). The combined organic layers were washed with brine and dried over Na₂SO₄. After removal of the solvent under reduced pressure, the residue was purified by flash column chromatography (hexanes/ethyl acetate 3:1) to yield product 15 as a sticky oil (0.83 g, 64%). ¹H NMR (CDCl₃, 300 MHz) δ 7.57 (d, 1H, J = 2.1 Hz), 7.24 (d, 1H, J = 2.1 Hz), 6.86 (d, 1H, J = 8.5 Hz), 6.57 (s, 2H), 5.69 (s, 1H), 3.89 (s, 3H), 3.83(s, 9H); ¹³C NMR (75.5 MHz) δ 154.59, 152.71, 139.45, 137.50, 136.43, 131.07, 126.49, 111.37, 111.04, 102.98, 74.49, 60.43, 55.87, 55.65.

[(3-Bromo-4-methoxyphenyl)-(3,4,5-trimethoxyphenyl)-methoxy]-tert-butyl-dimethyl silane (16).

To a solution of 15 (0.5 g, 1.3 mmol) in CH₂Cl₂ (10 mL) at 0 °C was added 2,6-lutidine (1.11 g, 10.4 mmol) and tert-butylydimethylsilyl trifluoromethanesulfonate (TBSOTf, 2.06 g, 7.8 mmol). The reaction mixture was stirred at room temperature for 1 h, quenched with saturated NH₄Cl (10 mL) and extracted with CH₂Cl₂. The combined extracts were washed with brine and dried over Na₂SO₄. After removal of the solvent in vacuo, the residue was purified by flash column chromatography (hexanes/ethyl acetate 10:1) to yield product 16 (0.54 g, 83%) as a yellow oil. ¹H NMR (CDCl₃, 300 MHz) δ 7.51 (d, 1H, J = 2.1), 7.23 (dd, 1H, J = 1.8, 1.8 Hz), 6.83 (d, 1H, J = 8.4 Hz), 6.56 (s, 2H), 5.59 (s, 1H), 3.85 (s, 3H), 3.82 (s, 9 H), 0.92 (s, 9H), 0.007 (s, 3H), −0.02 (s, 3H); ¹³C NMR (75.5 MHz) δ 154.97, 153.17, 140.63, 138.92, 131.43, 126.53, 111.68, 103.09, 75.68, 60.98, 56.34, 56.17, 25.97, 18.44, −4.66. ESI m/z 498 (M + H)⁺.

{5-[Hydroxy(3,4,5-trimethoxyphenyl)methyl]-2-methoxyphenyl}boronic acid (17).

General procedure B: Eluent MeOH/CH₂Cl₂ 1:20; pale yellow solid, 0.21 g (20%); ¹H NMR (CDCl₃, 300 MHz) δ 8.49 (m, 1H), 7.45 (m, 1H), 7.15 (m, 1H), 6.83 (m, 1H), 6.59 (m, 1H), 3.76 (m, 12H), 3.86 (s, 3H), 3.73 (s, 6H). ESI m/z 347 (M-H)⁺.

[2-Methoxy-5-(3,4,5-trimethoxybenzoyl)phenyl]boronic acid (18).

To a stirred solution of 17 (0.2 g, 0.57 mmol) was added pyridinium dichromate (0.44 g, 1.17 mmol) and powered 4 Å molecular sieves (0.1 g). The mixture was stirred at room temperature overnight and filtered through a pad of Celite. The filtrate was concentrated, and the residue was purified by flash column chromatography (MeOH/CH₂Cl₂ 1:100) to yield product 18 as a white solid (0.18 g, 83%): mp 150 - 151 °C; ¹H NMR (CDC1₃, 400 MHz) δ 8.35 (d, 1H, J = 2.0), 8.02 (dd, 1H, J = 2.2, 2.1 Hz ), 7.06 (s, 2H), 7.03 (s, 1H), 5.75 (s, 2H), 4.03 (s, 3H), 3.95 (s, 3H), 3.88 (s, 6H); ¹³C NMR (CDC1₃, 100 MHz) δ 194.6, 167.5, 152.8, 141.8, 139.7, 135.3, 132.9, 130.7, 110.0, 107.6, 61.0, 56.3, 56.0. MSI m z 347 (M + H)⁺. Anal. Calcd. for C₁₇H₁₉BO₇: C, 58.99; H, 5.53. Found: C, 58.78; H, 5.59.

(5-Bromo-6-methoxypyridin-3-yl)(3,4,5-trimethoxyphenyl)methanol (19).

To one fourth solution of 5-bromo-1,2,3-trimethoxybenzene 14 (0.69 g, 2.78 mmol) in THF (20 mL) was added magnesium turnings (0.068 g, 2.8 mmol) with a small piece of iodine. As soon as the solution became colorless, the remaining solution of 14 was added dropwise. The resulting mixture was stirred at room temperature for 1 h, and then this Grignard reagent was slowly added to the solution of 5-bromo-6-methoxynicotialdehyde (0.50 g, 2.31 mmol) in THF (2.5 mL) at 0 °C. The reaction mixture was stirred at room temperature for 1 h. A saturated NH₄Cl solution (10 mL) was added slowly and stirred for 30 min. The resulting mixture was extracted with ether (3 × 20 mL). The combined organic layers were washed with brine and dried over Na₂SO₄. The solvent was removed under reduced pressure, and the residue was purified by flash column chromatography (hexanes/ethyl acetate 4:1) to yield product 19 as a sticky oil (0.26 g, 29 %). ¹H NMR (CDC1₃, 300 MHz) δ 8.05 (m, 1H), 7.81 (m, 1H), 6.56 (s, 2H), 5.71 (s, 1H), 3.99 (s, 3H), 3.86 (s, 9H), 2.64 (bar, 1H).

3-Bromo-5-[(tert-butyldimethylsilanyloxy)-(3,4,5-trimethoxyphenyl)methyl]-2-methoxypyridine (20).

To a solution of 19 (0.2 g, 0.52 mmol) in CH₂C1₂ (10 mL) at 0 °C was added 2,6-lutidine (0.45 g, 4.16 mmol) and TBSOTf (0.83 g, 3.12 mmol). The reaction mixture was stirred at room temperature for 1 h, the reaction was quenched with saturated NH₄Cl (10 mL), and the mixture was extracted with CH₂C1₂. The combined extracts were washed with brine and dried over Na₂SO₄. After removal of the solvent under reduced pressure, the residue was purified by flash column chromatography (hexanes/ethyl acetate 10:1) to yield product 20 (0.16 g, 62%) as a pale yellow oil. ¹H NMR (CDC1₃, 300 MHz) δ 8.07 (d, 1H, J = 2.1 Hz), 7.71 (d, 1H, J = 2.1 Hz), 6.54 (s, 2H), 5.62 (s, 1H), 3.99 (s, 3H), 3.82 (s, 9H), 0.92 (s, 9H), 0.006 (s, 6H).

5-[Hydroxy(3,4,5-trimethoxyphenyl)methyl]-2-methoxypyridin-3-yl}boronic acid (21).

General procedure B: Pale yellow oil, 0.08 g (76 %). It was used directly for the next step without further purification.

[2-Methoxy-5-(3,4,5-trimethoxybenzoyl)pyridin-3-yl]boronic acid (22).

To a stirred solution of 21 (0.08 g, 0.23 mmol) was added pyridinium dichromate (0.26 g, 0.69 mmol) and powdered 4 Å molecular sieves (0.06 g). The mixture was stirred at room temperature overnight and filtered through a pad of Celite. The filtrate was concentrated, and the residue was purified by a preparative TLC plate to afford product 22 as a white solid (0.04 g, 50%): mp 147 - 149 °C; ¹H NMR (CDC1₃, 400 MHz) δ 8.73 (d, 1H, J = 2.4 Hz), 8.60 (d, 1H, J = 2.3 Hz ), 7.26 (s, 2H), 5.76 (s, 2H), 4.15 (s, 3H), 3.95 (s, 3H), 3.89 (s, 6H); ¹³C NMR (100 MHz) δ 193.4, 169.6, 153.3, 152.8, 148.3, 142.5, 132.6, 128.2, 107.7, 61.3, 56.6, 54.8. ESI m z 348 (M + H)⁺. Anal. Calcd. for C₁₆H₁₈BNO₇: C, 55.36; H, 5.23; N, 4.04. Found: C, 55.33; H, 5.29; N, 4.01.

Biological Testing

Tubulin Polymerization

Tubulin polymerization was followed turbidimetrically at 350 nm in Beckman model DU-7400 and DU-7500 spectrophotometers equipped with electronic temperature controllers, as described in detail previously.²³ The tubulin concentration was 10 μM (1.0 mg/mL).

MCF-7 Cell Proliferation Assay

IC₅₀ values for inhibition of cell growth were obtained by measuring the amount of total cell protein with the sulforhodamine B assay.²⁴ The MCF-7 cells were grown for 24 h without drug, followed by 48 h of growth in the presence of varying concentration of compound.

[³H] Colchicine Binding Assays

The binding of [³H]colchicine to tubulin was measured by the DEAE-cellulose filter method, as described in detail previously.²⁵ The tubulin concentration was 1.0 μM (0.1 mg/mL), and the [³H]colchicine concentration was 5.0 μM.

Indirect Immunofluorescence

Drug effects on the microtubule network of A-10 cells were evaluated by indirect immunofluorescent techniques as previously described.²⁶ Cells were plated onto glass coverslips and treated with a range of concentrations of compounds for 18 h. The cells were fixed, and microtubules visualized using a β-tubulin antibody and nuclei stained with 4,6-diamidino-2-phenylindole. Cells were examined with a Nikon ES800 fluorescence microscope, and images were captured with a Photometries Cool Snap FX3 camera and compiled using Metamorph^® software.

Molecule Docking

AutoDock Vina 1.2.0 was used to perform molecular docking of compounds into a high-resolution tubulin heterodimer structure (PDB ID: 6XER). Colchicine (CID: 2833), phenstatin (CID: 9948888), and CA-4 (CID: 5351344) structures were obtained to perform docking. SMILES files for the synthesized compounds were converted to mol files with 3D coordinates using OpenBabel. The tubulin crystal structure, which contains two heterodimers was edited to remove the second heterodimer, the stathmin domain, and other compounds and ions (water, guanosine-5′-triphosphate, guanosine-5′-diphosphate, sulfate ions, and magnesium ions). Docking was performed using a box of x = 22.62383, y = 20.6082, z = 22.8357 size with a center x = −6.64202, y = −9.81822, z = 40.4171, capturing the colchicine binding pocket. The docking position with the lowest dock score and orientated in a similar manner as colchicine was chosen for each compound.

Boronic acid analogs of combretastatin A-4 were designed and synthesized. The cytotoxicity study identified that several boronic acid analogs exhibited low nanomolar cytotoxicity in MCF-7 cells.

Highlights:

• Nature product combretastatin A-4 effectively inhibits tubulin polymerization

• Structure exploration of boronic acid analogs of combretastatin A-4

• Boronic acid analogs of combretastatin A-4 exhibit potent inhibition of cancer cell proliferation