Design, Synthesis and Biological Evaluation of a Structurally Simplified Syringolin A Analogues

Abstract

In this study, we designed and synthesized a structurally simplified syringolin A analogue 4, which could have a switched hydrogen bonding interaction with the β5 subunit of 20S proteasome. This analogue exhibits potent β5 proteasome inhibitory activity with an IC₅₀ value of 107 nM. It also shows cytotoxicity against a range of human cancer cells at submicromolar level (109–254 nM). This analogue is expected to be a lead compound as a next generation proteasome inhibitor because of its simple structure.

Multiple myeloma is a hematological malignancy, where clonogenic mature plasma cells accumulate in bone marrow. These malignant cells multiply rapidly and interfere with the production of normal blood cells in bone marrow, and produce abnormal monoclonal immunoglobulins called M proteins in large amount. Bortezomib (Velcade®)¹⁾ and carfilzomib (Kyprolis®)²⁾ have been approved by U.S. Food and Drug Administration (FDA) for the treatment of multiple myeloma. They are inhibitors of proteasome,³⁾ which is a complex machinery composed of many subunits. Amongst these subunits, β1, β2, and β5 subunits have caspase-like, trypsin-like and chymotrypsin-like catalytic activities, respectively. Proteasome degrades ubiquitin-labeled proteins, and is responsible for the maintenance of an intracellular protein expression.⁴⁾ Inhibition of proteasome results in accumulation of unnecessary proteins that ultimately causes cell death. Bortezomib reversibly inhibits proteasome to cause apoptosis of malignant plasma cells. The clinical introduction of bortezomib has changed the treatment paradigm of multiple myeloma resulting prolonged survival of patients. However, there are some drawbacks in clinical use of bortezomib such as peripheral neuropathy as well as drug resistance, which is a common issue to be concerned in cancer chemotherapy.^4–6) Although several proteasome inhibitors are currently under clinical trials,³⁾ new proteasome inhibitors are still necessary to be developed.

Syringolin A (1, Fig. 1) is a naturally occurring 12-membered macrolactam isolated from a strain of the plant pathogen Pseudomonas syringae pv. syringae in 1998.⁷⁾ It irreversibly inhibits proteasome by a 1,4-addition of the hydroxyl group of the N-terminal threonine (Thr) residue on the β5 subunit to the α,β-unsaturated carboxamide moiety embedded in 1.⁸⁾ Since the mode of inhibition of 1 is different from that of bortezomib and carfilzomib, 1 is expected to be a new lead as an anti-multiple myeloma drug. The fact that 1 can exhibit a moderate β5 subunit inhibitory activity (K_i′ value of 0.8 µM for β5) along with high cytotoxicity (IC₅₀ value of 8.5 µM for human myeloma MM1.S cells) has motivated several groups to improve its biological activity.^9–13) The X-ray crystal structure of the yeast 20S proteasome covalently bound to 1 (PDB code: 2ZCY) has been elucidated.⁸⁾ The structure indicates that in the β5-subunit, the isopropyl side chain directly attached to the 12-membered ring of 1 is recognized by an S1 subsite and that of the internal valine (Val) residue was recognized by an S3 subsite, which constitutes a hydrophobic pocket that can accept a larger hydrophobic substituent than the isopropyl group (Fig. 2). Moreover, a hydrogen bonding between the NH of the internal Val residue and the Asp114 residue of the neighboring β6 subunit can be observed. Although X-ray crystallography provides valuable information about a crystal complex, it is difficult to predict the association state of the covalent inhibitor to the target protein, because the X-ray crystal structure of covalent inhibitor–protein complexes does not always reflect the correct association state.¹⁴⁾ Therefore, a systematic structure–activity relationship study is necessary to confirm whether the interactions observed in the complex structure of covalent inhibitor–protein are crucial in the association state. Our group has investigated a systematic structure–activity relationship of 1 and found that compound 2, where the ureadipeptide moiety was replaced with an N-decanoyl-L-phenylalanine, can exhibit a strong β5 subunit inhibitory activity with a K_i′ value of 0.14 nM as well as strong cytotoxicity with an IC₅₀ value of 2.2 nM against human myeloma RPMI8226 cells.¹⁵⁾ The acyl side chain of 2 plays an important role in improving cell membrane permeability and makes an additional contact with the hydrophobic region of the β6 subunit of the 20S proteasome. This additional interaction might have enhanced the inhibitory activity. Of note, the analogue 3, where the nitrogen at the L-Phe moiety of 2 is methylated, exhibited a large decrease in the β5 subunit inhibitory activity (IC₅₀ 1000 nM) and completely diminished the cytotoxicity.¹⁶⁾ This result indicates that the hydrogen bonding between the NH and the Asp114 of the neighboring β6 residue is a critical interaction between 2 as well as 1 and the β5 subunit in their association state. There exists Ala49 residue near the Asp114 residue in the β6 subunit. In the complex structure of proteasome with bortezomib (PDB code: 2F16), the backbone NH of Ala49 interacts with the carbonyl oxygen of bortezomib (3.23 Å),¹⁷⁾ and it is presumed that the interaction is important in the association state of bortezomib to the β5 subunit. On the other hand, the carbonyl oxygen of internal Val residue of 1 is close to the NH of Ala49, but its distance is 3.63 Å, indicating that the interaction between these functional groups is not likely in the complex with 1. We hypothesized that removing the substituent recognized by the S3 subsite and the NH group that interact with the Asp114 would increase the flexibility of the side chain. This could induce a reorganization of the structure to switch the hydrogen bonding partners between the oxygen of the exocyclic carboxamide of 1 and the Ala49 of the β5 subunit in the association state. This molecular design allows us to generate a new structurally simplified inhibitor 4, where the L-phenylalanine residue of 2 is truncated and the acyl side chain is directly attached to the core 12-membered macrolactam. Here, we report the synthesis and biological evaluation of 4 in order to confirm the hypothesis and develop a guide to design new type of syringolin analogues, which are expected to be a next generation proteasome inhibitor.

Fig. 1. Structure and Biological Activities of Syringolin A and Its Analogues

Fig. 2. Expected Binding Mode of Syringolin A and Bortezomib with β5-Subunit of 20S Proteasome and Design of Simplified Analogue by Switching a Key Hydrogen Bonding in Ligand/Target Association

Results and Discussion

We have previously reported the total synthesis of 1 featuring intramolecular Ugi-three component reaction with a formylisonitrile, amine, and carboxylic acid.¹⁵⁾ The synthetic strategy was followed for the synthesis of 4 (Chart 1). The known alcohol 5,¹⁵⁾ which was prepared from N-Boc-L-valine in 10 steps, was oxidized by 2-iodoxybenzoic acid (IBX), and the formamide group of the resulting α,β-unsaturated aldehyde was dehydrated by triphosgene in the presence of Et₃N in CH₂Cl₂ at −78°C to provide the formylisonitrile 6, which was used for the intramolecular Ugi-three component reaction without further purification. The formylisonitrile 6, 2,4-dimethoxybenzylamine, and trifluoroacetic acid were heated in toluene at 70°C in the presence of NH₄Cl gave the desired 12-membered macrolactam 7. After removing the tert-butyldimethylsilyl (TBS) group of 7 by 3HF–Et₃N in MeCN–CH₂Cl₂, the hydroxylactam 8 was isolated in 15% yield over four steps from 5. The stereochemistry of the newly formed stereogenic center was found to be in S-configuration, and its diastereomer was not detected at all in the reaction mixture. The hydroxyl group of 8 was mesylated and the resulting mesylate was treated with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to promote β-elimination to afford the E-α,β-unsaturated carboxamide 9 selectively. The 2,4-dimethoxybenzyl group at the exocyclic nitrogen of 9 was removed by BCl₃ in CH₂Cl₂ at 0°C gave 10 in 75% yield. Next, an acyl side chain was introduced into the 12-membered maclolactam. The acyl side chain is found to be important in increasing the inhibitory activity, presumably by interacting with the β6 subunit of the 20S proteasome (Fig. 2). The length of the acyl side has been optimized by our group, revealing that C8–11 acyl groups were favored and the difference in the biological property among these analogues was very subtle.¹⁶⁾ Thus, a decanoyl group was chosen as the acyl substituent. The trifluoroacetyl group of 10 was removed by saturated NH₃ in MeOH, and the liberated amine 11 was acylated with decanoyl chloride to afford the target analogue 4 in 91% yield over two steps. In order to study the impact of the carbonyl oxygen as a hydrogen acceptor on the inhibitory effect, the N-decyl analogue 12, which lacks a carbonyl oxygen at the side chain, was designed. This was prepared by the reductive alkylation of 11 with decanal in the presence of NaBH(OAc)₃ and AcOH in CH₂Cl₂ in 18% yield over two steps from 11.

Chart 1. Synthesis of Syringolin A Analogues

For the evaluation of biological activity of the synthesized analogues, the proteasome inhibitor bortezomib was used as a positive control. Compound 4 exhibited much stronger inhibitory activity than the parent natural product 1 with an IC₅₀ value of 107 nM. Its inhibitory activity was 9.3 fold stronger than that of 3 (1000 nM). The improved activity of 4 may be due to the switching of a hydrogen bond inside the binding pocket of the β5 subunit. The lack of an NH group in 4 that interacts with the Asp114 residue would allow the oxygen of the exocyclic carboxamide to interact with the Ala49 residue in the association state (Fig. 2). The hydrogen bond switching together with the simplified structure of 4 would form a strong complex between 4 and proteasome. This kind of information may not be obtained from a crystal structure of the covalent complex because the X-ray crystal structure does not always provide structural information that is amenable to the structure-based approach. To form a covalent complex, a covalent inhibitor first associates with its target protein via non-covalent interactions to form an inhibitor–protein complex. The chemical reaction then takes place between the inhibitor and the protein to form a covalent complex with a conformational change in the complex. Finally, the conformation of the covalent complex re-organizes to attain an energetically stable state. Therefore, the X-ray crystal structure of covalent inhibitor–protein complexes does not always reflect the association state. The existence of the interaction between the oxygen of the exocyclic carboxamide of 4 and the Ala49 of the β5 subunit in the association state was supported by the fact that analogue 12, which lacks the carbonyl oxygen that is present at the side chain of 4, exhibited large decrease in the inhibitory activity showing only 23% inhibition of the β5 proteasome activity at 1000 nM. The trifluoroacetate 10, which lacks the lipophilic side chain, completely diminished the inhibitory activity. This indicates the lipophilic side chain attributes to interact with the hydrophobic region of the β6 subunit of the 20S proteasome and this additional interaction plays an important role in enhancement of the inhibitory activity.

Cytotoxicity against a range of human cancer cells including epidermal (A431), colon (HCT-116 and RKO), and myeloma (RPMI8226) cancer cells was also evaluated (Table 1). Overall, the cytotoxicity of the analogues is quite in accordance with the β5 proteasome inhibitory activity. 4 exhibited a potent cytotoxicity against all cells tested in this study with the 50% growth inhibitory concentration (GI₅₀) value of 254 nM for A431, 166 nM for HCT-116, 109 nM for RKO and 124 nM for RPMI8226. Although these values are higher than those of 2, 4 still exhibited a submicromolar cytotoxicity. There was a large decrease in the cytotoxicity (15–69 fold) in the case of 12 indicating the importance of exocyclic carbonyl moiety. The trifluoroacetate 10 was also tested, however it did not show any cytotoxicity, presumably because the short acyl group did not interact with the β6 subunit and 10 has poor cell membrane permeability.

Table 1. Biological Activities of Syringolin A Analogues

Compound	IC50 (nM)a)	GI50 (nM)
		A431b)	HCT-116c)	RKOc)	RPMI8226d)
Bortezomib	—	28.2	27	20.8	14.4
Syringolin A (1)	680	3240	2830	n.d.e)	1160
3	10000	>10000	>10000	>10000	>10000
4	107	254	166	109	124
10	>1000	>10000	>10000	>10000	>10000
12	23% Inhibition at 1000 nM	3913	7290	7502	6381

a) Chymotrypsin-like activity of 20S proteasome. b) Human epidermal cancer cells. c) Human colon cancer cells. d) Human myeloma cells. e) Not determined.

Conclusion

We designed and synthesized a structurally simplified syringolin A analogue 4, which could interact with the β5 subunit of 20S proteasome through a switched hydrogen bonding. This analogue exhibits potent inhibitory activity (IC₅₀ 107 nM) as well as cytotoxicity against a range human cancer cells at a submicromolar level (109–254 nM). The chemical structure of 4 is very simple compared to the parent natural product with lower molecular weight and the number of chiral center. This analogue is expected to be a promising lead compound and further optimization of 4 would generate a next generation proteasome inhibitor.

Experimental

(4R,5S,8S,9E)-1,6-Diaza-8-[N-(2,4-dimethoxybenzyl)-N-trifluoroacetyl]amino-4-hydroxy-5-isopropyl-2,7-dioxocyclododec-9-en (8)

A solution of 5¹⁵⁾ (2.23 g, 5.76 mmol) in MeCN (58 mL) was treated with IBX (3.23 g, 11.5 mmol) at room temperature, and the mixture was stirred at 80°C for 1 h. The mixture was cooled to 0°C, the insolubles were filtered off through a Celite pad, and the filtrate was concentrated in vacuo to afford crude aldehyde. A solution of the crude aldehyde and Et₃N (2.79 mL, 20.2 mmol) in CH₂Cl₂ (58 mL) was treated with triphosgene (855 mg, 2.88 mmol) at −78°C for 30 min. The mixture was quenched with MeOH and partitioned between AcOEt and saturated aqueous NaHCO₃. The organic phase was washed with water and brine, dried (Na₂SO₄) and concentrated in vacuo to afford a crude isonitrile 6. A mixture of the crude isonitrile 6, 2,4-dimethoxybenzylamine (1.73 mL, 11.5 mmol), NH₄Cl (3.08 g, 57.6 mmol) and trifluoroacetic acid (856 µg, 11.5 mmol) in toluene (5.8 L) was stirred at 50°C for 24 h. The whole mixture was concentrated in vacuo. The residue was partitioned between AcOEt and 1 M aqueous HCl. The organic phase was washed with saturated aqueous NaHCO₃, water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo to afford Ugi-product 7 as a yellow oil. A solution of the Ugi-product in MeCN–CH₂Cl₂ (1 : 1) (58 mL) was treated with 3HF·Et₃N (9.4 mL, 57.6 mmol) at room temperature for 2 d. The reaction mixture was partitioned between AcOEt and saturated aqueous NaHCO₃. The organic phase was washed with water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo. The residue was purified by column chromatography (2% MeOH–CHCl₃) to afford 8 (436 mg, 15%) as a yellow oil. ¹H-NMR (CDCl₃, 400 MHz, a mixture several rotamers at 20°C) Selected data for the major rotamer δ: 7.13 (d, 1H, Ph-H, J=8.2 Hz), 6.40 (m, 2H, Ph-H), 6.27 (d, 1H, NH, J=7.3 Hz), 5.65 (m, 2H, H-9, 10), 5.50 (m, 1H, NH), 4.83–4.52 (m, 3H, H-8, Bn), 4.21 (m, 1H, H-5), 3.92–3.69 (m, 8H, H-12, OMe), 2.83 (br s, 1H, H-4), 2.52 (m, 2H, H-11), 2.19 (d, 1H, H-3, J=12.8 Hz), 2.12 (m, 1H, H-1′), 1.91 (m, 1H, H-3), 1.01–0.70 (m, 6H, Me); ¹³C-NMR (CDCl₃, 100 MHz) δ: 173.2, 169.9, 160.7 (d, J_F,C=43.9 Hz), 158.8, 157.4, 135.7, 128.9, 125.6, 123.6 (d, J_F,C=347.8 Hz), 116.6, 104.1, 98.2, 69.2, 68.5, 65.5, 60.3, 55.5, 45.3, 40.0, 37.5, 34.6, 29.7, 20.0, 19.0; low resolution-electrospray ionization (LR-ESI)-MS m/z 538 [(M+Na)⁺]; high resolution (HR)-ESI-MS Calcd for C₂₄H₃₂F₃N₃NaO₆ 538.2141. Found 538.2135.

(4R,5S,8S,9E)-1,6-Diaza-8-[N-(2,4-dimethoxybenzyl)trifluoroacetamido-5-isopropyl-2,7-dioxocyclododeca-3,9-diene (9)

A solution of 8 (290 mg, 0.40 mmol) and Et₃N (333 µL, 2.40 mmol) in CH₂Cl₂ (4 mL) was treated with MsCl (93.0 µL, 1.20 mmol) at 0°C for 10 min. The mixture was partitioned between AcOEt and 1 M aqueous HCl. The organic phase was washed with saturated aqueous NaHCO₃, water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo to afford crude mesylate. A solution of the crude mesylate in CH₂Cl₂ (4 mL) was treated with DBU (359 µL, 2.40 mmol) at room temperature for 2 d. The mixture was partitioned between AcOEt and 1 M aqueous HCl. The organic phase was washed with saturated aqueous NaHCO₃, water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo. The residue was purified by column chromatography (3% MeOH–CHCl₃) to afford 9 (105 mg, 38%) as a colorless solid. ¹H-NMR (DMSO-d₆, 400 MHz, a mixture several rotamers at 20°C). Selected data for the major rotamer δ: 8.27 (d, 1H, NH, J_N,5=7.8 Hz), 7.49 (m, 1H, NH), 7.05 (d, 1H, Ph-H, J=8.7 Hz), 6.70 (dd, 1H, H-4, J_4,5=5.0, J_4,3=15.6 Hz), 6.45 (br s, 2H, Ph-H), 6.08 (d, 1H, H-3, J_3,4=15.6 Hz), 5.49 (m, 1H, H-10), 5.23 (d, 1H, H-8, J_8,9=7.8 Hz), 5.10 (dd, 1H, H-9, J_9,8=7.8, J_9,10=15.6 Hz), 4.72–4.40 (m, 2H, PhCH₂), 4.12 (br s, 1H, H-5), 3.73 (s, 6H, OMe), 3.13 (br s, 1H, H-12), 3.03 (br s, 1H, H-12), 2.02–1.64 (m, 2H, H-11), 1.73 (dt, 1H, H-1′, J_1′,Me=6.9, J_1′,5=14.2 Hz), 0.90 (d, 6H, Me, J_Me,1′=6.9 Hz); ¹³C-NMR (DMSO-d₆, 100 MHz) δ: 168.2, 166.1, 159.3, 156.0 (d, J_F,C=40.2 Hz), 143.2, 137.3, 127.3, 126.3, 121.6, 121.0, 119.5 (d, J_F,C=369.8 Hz), 117.7, 104.4, 97.9, 60.2, 55.4, 55.1, 43.2, 42.6, 35.2, 31.6, 19.8, 19.4, 19.2; LR-ESI-MS m/z 520 [(M+Na)⁺]; HR-ESI-MS Calcd for C₂₄H₃₀F₃N₃NaO₅ 520.2035. Found 520.2030; [α]_D²⁰ −62.9 (c=0.48, dimethyl sulfoxide (DMSO)).

(4R,5S,8S,9E)-1,6-Diaza-8-trifluoroacetamido-5-isopropyl-2,7-dioxocyclododeca-3,9-diene (10)

A solution of 9 (162 mg, 0.31 mmol) in CH₂Cl₂ (10 mL) was treated with BCl₃ (1 M CH₂Cl₂ solution) (3.10 mL, 3.10 mmol) at 0°C for 2 h. The mixture was partitioned between AcOEt and saturated aqueous NaHCO₃. The organic phase was washed with water and brine, dried (Na₂SO₄), filtrated, and concentrated in vacuo. The residue was purified by column chromatography (3% MeOH–CHCl₃) to afford 10 (80.6 mg, 75%) as a colorless solid. ¹H-NMR (DMSO-d₆, 400 MHz) δ: 9.75 (d, 1H, NH, J_NH,8=6.2 Hz), 8.19 (d, 1H, NH, J_NH,5=6.9 Hz), 7.50 (t, 1H, NH, J_NH,12=7.1 Hz), 6.69 (dd, 1H, H-4, J_4,5=6.9, J_4,3=15.6 Hz), 6.08 (d, 1H, H-3, J_3,4=15.6 Hz), 5.72 (dt, 1H, H-10, J_10,11=7.8, J_10,9=15.8 Hz), 5.46 (dd, 1H, H-9, J_9,8=6.2, J_9,10=15.8 Hz), 4.92 (t, 1H, H-8, J_8,9=J_8,NH=6.2 Hz), 4.08 (dt, 1H, H-5, J_5,4=J_5,NH=6.9, J_5,1′=14.2 Hz), 3.18 (m, 2H, H-12), 2.32 (m, 1H, H-11), 1.97 (m, 1H, H-11), 1.73 (dt, 1H, H-1′, J_1′,Me=7.0, J_1′,5=14.2 Hz), 0.94 (d, 3H, Me, J_Me,1′=7.0 Hz), 0.90 (d, 3H, Me, J_Me,1′=7.0 Hz); ¹³C-NMR (CDCl₃, 100 MHz) δ: 167.9, 166.2, 156.1 (d, J_F,C=36.4 Hz), 143.3, 134.9, 123.6, 121.6, 115.8 (d, J_F,C=289.3 Hz), 55.6, 54.8, 42.3, 35.2, 31.5, 19.7, 19.2; LR-ESI-MS m/z 370 [(M+Na)⁺]; HR-ESI-MS Calcd for C₁₅H₂₀F₃N₃NaO₃Si 370.1354. Found 370.1349.

(4R,5S,8S,9E)-1,6-Diaza-8-decanamido-5-isopropyl-2,7-dioxocyclododeca-3,9-diene (4)

Compound 10 (5.5 mg, 15.8 µmol) was treated with saturated ammonia in methanol solution (1 mL) at room temperature for 13 h. The mixture was concentrated in vacuo to afford crude amine 11. A solution of the crude amine and Et₃N (6.6 µL, 47.4 µmol) in MeCN (1 mL) was treated with decanoyl chloride (3.8 µL, 19.0 µmol) at 0°C for 10 min. The mixture was partitioned between AcOEt and 1 M aqueous HCl. The organic phase was washed with saturated aqueous NaHCO₃, water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo. The residue was purified by column chromatography (3% MeOH–CHCl₃) to afford 4 (5.8 mg, 91%) as a white solid. ¹H-NMR (CD₃OD, 400 MHz) δ: 6.96 (dd, 1H, H-4, J_4,5=4.8, J_4,3=15.7 Hz), 6.21 (d, 1H, H-3, J_3,4=15.7 Hz), 5.64 (m, 2H, H-9, 10), 4,18 (br s, 1H, H-8), 3.43–3.18 (m, 2H, H-12), 2.32 (br s, 1H, H-11), 2.22 (br s, 3H, H-11, H-2″), 1.82 (m, 1H, H-1′), 1.59 (br s, 2H, H-13″), 1.40–1.21 (br s, 12H, CH₂), 1.06 (d, 3H, Me, J=6.4 Hz), 1.01 (d, 3H, Me, J=8.1 Hz), 0.89 (s, 3H, Me); ¹³C-NMR was not able to be analyzed because of the solubility of the compound; LR-ESI-MS m/z 428 [(M+Na)⁺]; HR-ESI-MS Calcd for C₂₃H₃₉N₃NaO₃ 428.2889. Found 428.2873.

(4R,5S,8S,9E)-1,6-Diaza-8-(1-decyl)amino-5-isopropyl-2,7-dioxocyclododeca-3,9-diene (12)

Compound 10 (10.0 mg, 28.8 µmol) was treated with saturated ammonia in methanol solution (2 mL) at room temperature for 13 h. The mixture was concentrated in vacuo to afford crude amine 11. A mixture of the crude amine, decanal (5.4 µL, 28.8 µmol) and AcOH (8.2 µL, 144 µmol) in MeCN (1 mL) was stirred at room temperature for 10 min. NaBH(OAc)₃ (12.0 mg, 57.6 µmol) was added to the mixture and stirred at room temperature for 1 d. Decanal (2.7 µL, 14.4 µmol) and NaBH(OAc)₃ (6.0 mg, 28.8 µmol) was added to the mixture and stirred at room temperature for 1 d. The reaction was quenched with saturated aqueous NaHCO₃ and the organic phase was washed with water and brine, dried (Na₂SO₄), filtered, and concentrated in vacuo. The residue was purified by column chromatography (4% MeOH–CHCl₃) to afford 12 (2.0 mg, 18%) as a white solid. ¹H-NMR (DMSO-d₆, 400 MHz) δ: 7.88 (d, 1H, NH, J_NH,5=7.4 Hz), 7.44 (t, 1H, NH, J_NH,12=7.1 Hz), 6.67 (dd, 1H, H-4, J_4,5=7.3, J_4,3=15.4 Hz), 6.08 (d, 1H, H-3, J_3,4=15.4 Hz), 5.46 (dt, 1H, H-10, J_10,11=7.3, J_10,9=16.4 Hz), 5.27 (dd, 1H, H-9, J_9,8=7.5, J_9,10=16.4 Hz), 4.09 (br s, 1H, H-5), 3.70 (d, 1H, H-8, J_8,9=7.5 Hz), 3.14 (m, 2H, H-12), 2.45–2.29 (m, 2H, H-1″), 2.22 (br s, 1H, H-11), 1.96 (br s, 1H, H-11), 1.74 (br s, 1H, H-1′), 1.34 (br s, 1H, NH), 1.23 (m, 16H, CH₂), 0.97–0.81 (m, 9H, Me); ¹³C-NMR (DMSO-d₆, 125 MHz) δ: 166.4, 143.3, 131.4, 125.1, 120.7, 111.8, 55.5, 55.4, 47.2, 42.7, 35.1, 31.3, 29.6, 29.6, 29.0, 28.8, 28.7, 26.8, 22.1, 19.8, 19.8, 19.3, 14.0; LR-ESI-MS m/z 414 [(M+Na)⁺]; HR-ESI-MS Calcd for C₂₃H₄₁N₃NaO₂ 414.3096. Found 414.3091; [α]_D²³ +27.8 (c=0.10, DMSO).

Evaluation of in Vitro Chymotrypsin-Like (CT-L) Activity of 20S Proteasome

In vitro proteasome assays were performed in 96-wel microtiter plates with human erythrocyte 20S proteasome (R&D Systems, Inc.) using the AK-740 Assay Kit for Drug Discovery (Biomol). Reactions were performed at 37°C in 100 µL volumes containing a serial dilution of test samples, 2 µg/mL 20S proteasome, and 100 µM Suc-LLVY-AMC for assaying chymotrypsin-like activity, according to the manufacturer’s instructions. Fluorescence was monitored with infinite M200 microplate reader (Tecan) equipped with 360 nm excitation and 465 nm emission filters.

Measurement of Cytotoxicity

Antiproliferative activities of the compounds against A431, HCT-116, RKO, TOV21G, CCRF-CEM, RPMI8226 cells were measured by quantifying the amount of ATP as follows. Briefly, cells (1–2×10³ cells/well) in a 96 well plate were cultured in RPMI medium containing 10% fetal bovine serum in the presence of test compounds at 37°C for 72 h under 5% CO₂ atmosphere. Then, a solution of CA-1000 reagents (TOYO INK Co., Ltd.) was added, and plates were shaken and luminescence of oxyluciferin was monitored with the plates held at 23°C.

Acknowledgments

We thank Ms. S. Oka and Ms. A. Tokumitsu (Center for Instrumental Analysis, Hokkaido University) for the measurement of mass spectra. This research was supported by JSPS Grant-in-Aid for Challenging Exploratory Research (SI, Grant Number 22659020), Scientific Research (B) (SI, Grant Number 25293026), Scientific Research on Innovative Areas “Chemical Biology of Natural Products” (SI, Grant Number 24102502).

Conflict of Interest

The authors declare no conflict of interest.

References

• 1) Bross P. F., Kane R., Farrell A. T., Abraham S., Benson K., Brower M. E., Bradley S., Gobburu J. V., Goheer A., Lee S.-L., Leighton J., Liang C. Y., Lostritto R. T., McGuinn W. D., Morse D. E., Rahman A., Rosario L. A., Verbois S. L., Williams G., Wang Y.-C., Pazdur R., Clin. Cancer Res., 10, 3954–3964 (2004).
• 2) Herndon T. M., Deisseroth A., Kaminskas E., Kane R. C., Koti K. M., Rothmann M. D., Habtemariam B., Bullock J., Bray J. D., Hawes J., Palmby T. R., Jee J., Adams W., Mahayni H., Brown J., Dorantes A., Sridhara R., Farrell A. T., Pazdur R., Clin. Cancer Res., 19, 4559–4563 (2013).
• 3) Kisselev A. F., van der Linden W. A., Overkleeft H. S., Chem. Biol., 19, 99–115 (2012).
• 4) Murata S., Yashiroda H., Tanaka K., Nat. Rev. Mol. Cell Biol., 10, 104–115 (2009).
• 5) Lü S., Wang J., Biomarker Res., 1, 13 (2013).
• 6) Alsina M., Trudel S., Furman R. R., Rosen P. J., O’Connor O. A., Comenzo R. L., Wong A., Kunkel L. A., Molineaux C. J., Goy A., Clin. Cancer Res., 18, 4830–4840 (2012).
• 7) Wäspi U., Blanc D., Winkler T., Rüedi P., Dudler R., Mol. Plant Microbe Interact., 11, 727–733 (1998).
• 8) Gröll M., Schellenberg B., Bachmann A. S., Archer C. R., Huber R., Powell T. K., Lindow S., Kaiser M., Dudler R., Nature (London), 452, 755–758 (2008).
• 9) Clerc J., Groll M., Illich D. J., Bachmann A. S., Huber R., Schellenberg B., Dudler R., Kaiser M., Proc. Natl. Acad. Sci. U.S.A., 106, 6507–6512 (2009).
• 10) Clerc J., Schellenberg B., Gröll M., Bachmann A. S., Huber R., Dudler R., Kaiser M., Eur. J. Org. Chem., 2010, 3991–4003 (2010).
• 11) Clerc J., Li N., Krahn D., Groll M., Bachmann A. S., Florea B. I., Overkleeft H. S., Kaiser M., Chem. Commun., 47, 385–387 (2011).
• 12) Archer C. R., Groll M., Stein M. L., Schellenberg B., Clerc J., Kaiser M., Kondratyuk T. P., Pezzuto J. M., Dudler R., Bachmann A. S., Biochemistry, 51, 6880–6888 (2012).
• 13) Totaro K. A., Barthelme D. B., Simpson P. T., Sauer R. T., Sello J. K., Bioorg. Med. Chem., 23, 6218–6222 (2015).
• 14) Kawamura S., Unno Y., Tanaka M., Sasaki T., Yamano A., Hirokawa T., Kameda T., Asai A., Arisawa M., Shuto S., J. Med. Chem., 56, 5829–5842 (2013).
• 15) Chiba T., Hosono H., Nakagawa K., Asaka M., Takeda H., Matsuda A., Ichikawa S., Angew. Chem. Int. Ed., 53, 4836–4839 (2014).
• 16) Chiba T., Matsuda A., Ichikawa S., Bioorg. Med. Chem. Lett., 25, 4872–4877 (2015).
• 17) Gröll M., Berkers C. R., Ploegh H. L., Ovaa H., Structure, 14, 451–456 (2006).