Synthesis of the Biologically Active Natural Product Cyclodepsipeptides Apratoxin A and Its Analogues

Abstract

This paper describes the synthetic studies conducted on a marine natural product, cyclodepsipeptide apratoxin A. Total synthesis of the oxazoline analogue of apratoxin A was achieved. The conversion of oxazoline to thioamide, as well as thioamide formation from a serine-derived compound, were both unsuccessful. However, thiazoline formation from a cysteine-derived compound led to the total synthesis of apratoxin A. An in vivo study on synthetic apratoxin A revealed that it has potent antitumor activity, but with significant toxicity. Solid-phase synthesis of apratoxin A was accomplished using a preformed thiazoline derivative as a coupling unit. This method was used to synthesize several azido-containing analogues as precursors of molecular probes, and these analogues exhibited potent biological activity.

1. Introduction

Several naturally occurring cyclodepsipeptides isolated from marine sponges, bacteria, and fungi exhibit unique biological activities.^1–3) The cyclodepsipeptides contain a hydroxy acid that bears one or more chiral centers, and also often contain N-methylamino acids, which significantly increases the structural diversity of these compounds. The ester linkage between a hydroxy acid and an amino acid is also a unique structural feature. In this paper, we focus on the rapid synthesis of natural products and their derivatives, and describe the synthetic study of the cyclodepsipeptide apratoxin A.

2. Apratoxin A

Apratoxin A (1) was isolated by Luesch et al. in 2001 from the marine cyanobacterium Lyngbya majuscula HARVEY ex GOMONT. They concluded that apratoxin A exhibits potent cytotoxicity for cancer cells, with IC₅₀ values of 0.52 nM against KB and 0.36 nM against LoVo cancer cells^4,5) (Fig. 1). The first total synthesis of this cyclodepsipeptide was achieved by Chen and Forsyth.⁶⁾ They succeeded in synthesizing thiazoline 3 by the formation of thioester 2, followed by a Staudinger–aza-Wittig condensation as a key step. Macrolactamization between proline and N-methylisoleucine residues provided apratoxin A (1) (Chart 1).

Fig. 1. Structure of Apratoxin A (1)

Chart 1. Total Synthesis of 1 Accomplished by Chen and Forsyth⁶⁾

The position of macrocyclization is very important in the synthesis of cyclodepsipeptides. Considering some cases in the synthesis of apratoxin A, one can see those positions which should not be chosen as a cyclization site. For example, if macrolactam formation between tyrosine and N-methylisoleucine was selected, as shown in Fig. 2, macrocyclization of 4A would be slower because of the poor nucleophilicity of the N-methylamino group as a result of steric hindrance. As a consequence, epimerization would take place at the α position of the C-terminus via the formation of azlactone, or diketopiperazine formation at the N-terminus would cleave the ester moiety. Similarly, linear precursors 4B and 4C are not suitable for macrocyclization due to poor nucleophilicity of the N-methyl amino group in 4B and steric hindrance of the neopentyl alcohol in 4C. On the other hand, 4D may be a good linear precursor, unless diketopiperazine is formed through the favorable s-cis form of the N-methylamide bond between Tyr(Me) and MeAla residues. Before it was reported that the macrocyclization of a similar precursor to 4D did lead to the oxazoline-alanogue of apratoxin A,⁷⁾ we had planned macrolactamization between proline and N-methylisoleucine, as reported by Chen and Forsyth.⁶⁾ Because it will be suitable to synthesize various analogues of apratoxin A on solid-phase afterwards, synthetically valuable 3,7-dihydroxy-2,5,8,8-tetramethylnonanoic acid (Dtena) can be introduced at the late stage of the synthesis.

Fig. 2. Conceivable Linear Precursors 4 for Macrocyclization (X=Leaving Group)

Another important issue to consider is the method for constructing a thiazoline ring. 2,4-Disubstituted thiazolines can be prepared from serine amide via thioamide formation, followed by intramolecular thioether formation using the hydroxy group as a leaving group⁸⁾ (Fig. 3). Another approach is from the cysteine amide by dehydrative cyclization using a thiol group.^9–12) As the thiol-protected amino acid residue is labile for acid, base, oxidation, and reduction, it would be preferable to avoid using the thiol-protected residue in any total synthesis that involves long reaction steps. Wipf and Uto elegantly demonstrated that the oxazoline moiety was ring-opened to the corresponding thioamide on a cyclic peptide by treatment with H₂S, and then dehydrative cyclization using N,N-dimethylaminosulfur trifluoride (DAST) led to the total synthesis of the thiazoline-containing cyclic peptide trunkamide A.¹³⁾ We initially synthesized the oxazoline analogue of apratoxin A (16)^7,14,15) and attempted to convert the oxazoline moiety to a thiazoline ring.

Fig. 3. Formation of a Thiazoline Ring

3. Total Synthesis of the Oxazoline Analogue of Apratoxin A and Attempted Conversion of Oxazoline to Thiazoline

The coupling of serine-derived compound 5 and H-Tyr(Me)-OMe (6) (1-ethyl-(3-(3-dimethylamino)propyl)-carbodiimide hydrochloride (EDCI)/1-hydroxybenzotriazole (HOBt)) provided dipeptide 7 (Chart 2). Hydrolysis of the methyl ester, followed by condensation with dipeptide 8 (2-(7-aza-1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU)/N,N-diisopropylethylamine (DIEA)), gave undesired azlactone 10 in 49% yield, along with the desired tetrapeptide 9 in only 21% yield. This result was due to the poor nucleophilicity of the N-methyldipeptide 8. Therefore, we performed sequential coupling of dipeptide 8 with mono amino acid 11, followed by treatment of 12 with 5 to afford tetrapeptide 9 in 73% overall yield (Chart 3). Sequential removal of the Boc group using trimethylsilyl triflate (TMSOTf)/2,6-lutidine and the TBS group using tetrabutylammonium fluoride (TBAF), condensation with acid 13¹⁶⁾ (HATU/DIEA), and removal of the TBS group using TBAF provided 14 in 93% yield. Formation of oxazoline 15 was achieved by treatment of 14 with DAST at −78°C (70%).

Chart 2. Attempted Condensation of Dipeptide 7 and N-Methylated Dipeptide 8

Chart 3. Formation of Oxazoline Derivative 15

The oxazoline analogue of apratoxin A (16) was successfully synthesized by (i) cleavage of the Boc group of 15 (TMSOTf/2,6-lutidine), (ii) removal of the TMS ether formed in the reaction (i) (TBAF), (iii) removal of the allyl ester (cat. Pd(PPh₃)₄/morpholine), and (iv) macrolactamization (HATU/DIEA) in 52% overall yield¹⁷⁾ (Chart 4). On the other hand, ring opening of the oxazoline ring in 15 with H₂S failed to provide thioamide 17.

Chart 4. Synthesis of the Oxazoline Analogue of Apratoxin A (16) and Attempted Ring Opening of Oxazoline 15

We subsequently attempted the formation of thioamide 20 from 19 (Chart 5). Coupling of 13 and 18 was performed using HATU to provide 19 in 90% yield according to the preparation of 14, as shown in Chart 3. Treatment of 19 with Lawesson’s reagent initially afforded the desired thioamide 20 (90%). However, 20 gradually underwent intramolecular Michael addition, which led to the formation of undesired thiazoline 21. In particular, the use of TBS cleavage conditions (TBAF or HF-pyridine) quickly led to the formation of 21.¹⁷⁾ Therefore, we changed the synthetic plan from thioamide formation from serine derivatives to thioamide formation from a cysteine-derived compound (see Fig. 3).

Chart 5. Attempted Thiazoline Formation of Thioamide 20

4. Total Synthesis of Apratoxin A and in Vivo Evaluation

Cysteine-derived compound 25 was used for thiazoline formation in the total synthesis of apratoxin A (1) (Chart 6). Amine 23 was prepared from Boc-Cys(Trt)-OH (22) as follows: (i) Weinreb amide formation, (ii) DIBAL reduction, (iii) Wittig olefination using Ph₃P=C(CH₃)CO₂Et, (iv) basic hydrolysis of the ethyl ester (LiOH), (v) allyl ester formation (K₂CO₃/allyl bromide), and (vi) cleavage of the Boc group (TMSOTf/2,6-lutidine) (73% overall yield). Coupling of 23 and 24a¹⁷⁾ (EDCI/1-hydroxy-7-azabenzotriazole (HOAt)/DIEA) afforded 25 in 81% yield. Thiazoline formation was performed by Kelly’s method (Tf₂O/Ph₃P=O),¹¹⁾ and immediate removal of the 2,2,2-trichloroethoxycarbonyl (Troc) group (Zn/NH₄OAc) provided the desired thiazoline in 90% yield. When we purified the Troc-protected derivative, dehydration was observed during silica gel column chromatography. The use of AcOH, instead of NH₄OAc, also induced dehydration because of the presence of the acidic proton of thiazoline-attached methine. After removal of the ally ester (cat. Pd(PPh₃)₄/N-methylaniline), coupling of 26 with the amine prepared from tripeptide 12 by Fmoc cleavage afforded 27 (HATU/DIEA) in 75% yield. Removal of the allyl ester from 27 (as described above), Fmoc cleavage (Et₂NH/CH₃CN), and macrolactamization (HATU/DIEA) (under 1 mM concentration) furnished apratoxin A (1) in 72% overall yield. The spectral data were identical to those of the natural product.^16,17)

Chart 6. Total Synthesis of Apratoxin A (1) from Cysteine-Derived Compound 23

As it was reported that the dehydration of apratoxin A occurred in CDCl₃,⁵⁾ we also observed the isomer of apratoxin A by LC/MS analysis after recording the NMR spectrum in CDCl₃. We synthesized 34-epi apratoxin A (29) to confirm the structure of the isomer, because it might be a rotamer, although it was not likely observed (Chart 7). Cyclization precursor 28 was prepared from 23 and 24b using a method similar to that used for synthesizing 1. However, macrolactamization resulted in the formation of 29 in only 25% yield, and only one of the stereogenic centers was significant to adopt a suitable conformation for the macrocyclization. The retention time for 29 observed by HPLC analysis was identical to that of the isomer observed by NMR analysis of apratoxin A.¹⁷⁾

Chart 7. Synthesis of 34-epi Apratoxin A (29)

We then synthesized 80 mg of 1 according to the above method, and further studied this sample to evaluate the clonogenic survival of HCT-116 cells exposed to 1. It was found that the effective concentration of 1 would have to be chronically maintained above 1.6 nM.¹⁸⁾ Next, 12.5 µg of 1 per mouse (0.7 mg/kg) was utilized daily for 5 d to evaluate its therapeutic efficacy against HCT-116 human colon cancer in SCID mice. The size of the tumors in the treated mice was found to be only 20% of that of the control mice at 22 d (T/C, 20%), although four out of the five treated tumor-bearing mice died.¹⁸⁾ It should be noted that 1 exhibited significantly potent antitumor activity for HCT-116 cancer cells in vivo, but this was accompanied by significant toxicity.⁴⁾

5. Solid-Phase Assisted Synthesis of Apratoxin A and Its Analogues, and Evaluation of Their Cytotoxicities

The solid-phase-assisted synthesis of apratoxin A was investigated to prepare a variety of synthetic analogues. Our initial approach was as follows: (i) assembly of each amino acid residue on a polymer support; (ii) macrolactamization after cleavage from the polymer support; and (iii) formation of the thiazoline ring as a final step. This approach was used in the synthesis of apratoxin analogue 36 (Chart 8). Attachment of Fmoc-Melle-OH to trityl-linked Lanterns™ (30) (33 µmol/unit),¹⁹⁾ sequential coupling with Fmoc-MeAla-OH (bromotripyrrolidinophosphonium hexafluorophosphate (PyBroP)/DIEA, twice) and Fmoc-Tyr(Me)-OH (11) (PyBroP/DIEA, twice) by an Fmoc method provided the polymer-supported tripeptide 31. Coupling of 31 with 32 (N,N′-diisopropylcarbodiimide (DIC)/HOBt/DIEA) afforded the polymer-supported tetrapeptide 33 with 91% purity, as determined by UV spectroscopy (214 nm). Condensation of unsubstituted proline depsipeptide 34 to 33 (DIC/HOBt/DIEA), removal of the Fmoc group (20% piperidine/DMF), and cleavage from the polymer support with 30% hexafluoroisopropyl alcohol (HFIP) in CH₂Cl₂ provided the cyclization precursor with 83% purity. Macrolactamization was performed, as described above, to afford 35 in 50% overall yield from the initial loading on the polymer support. Finally, formation of thiazoline using Kelly’s method (−20°C) furnished the desired apratoxin analogue 36 in 50% yield. Several undesired products were also observed, which were probably derived from the activation of other amide moieties. The same problem was observed to a greater extent in the synthesis of apratoxin A (1) on using proline depsipeptide 24a instead of 34 (Chart 9). Thiazoline formation on cyclodepsipeptide 37 by Kelly’s method (0°C) and immediate removal of the Troc group (Zn/NH₄OAc) provided a complex mixture of products. Only 10% yield of 1 from 37 was isolated by preparative TLC. Therefore, it was necessary to modify the solid-phase-assisted synthesis of 1.

Chart 8. Solid-Phase Assisted Synthesis of Apratoxin Analogue 36

Chart 9. Attempted Thiazoline Formation of 1 from Macrocycle 37

The coupling of polymer-supported tripeptide 31 and the rather large acid 26, which has a preformed thiazoline ring, was investigated (Chart 10). The reaction proceeded smoothly using two equivalents of 26 and (7-azabenzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyAOP)/DIEA as condensation agents. Fmoc cleavage and acidic treatment (HFIP) provided cyclization precursor 39, which underwent macrolactamization, leading to a separable mixture of apratoxin A (1) and 34-epimer 29 in a ratio of 85 : 15. It was found that epimerization at the 34-position, which was adjacent to the 2-position of thiazoline, was suppressed to 15%.²⁰⁾

Chart 10. Solid-Phase-Assisted Total Synthesis of 1

The solid-phase method was employed to synthesize several analogues of apratoxin, 40–46 (Fig. 4). The cytotoxicities of the synthetic 1 and its analogues against HeLa cells were evaluated, and the results are summarized in Table 1. The activity of 34-epi apratoxin A (29) was as high as that of apratoxin A. The oxazoline analogue of apratoxin A (16) was also so potent that the thiazoline ring could be replaced by an oxazoline ring, as also reported by Ma’s group.¹⁵⁾ Protection of the hydroxy group with a TES group, or removal of all substituents from the hydroxy acid moiety, led to a loss of cytotoxicity. Azido-containing analogues 41–45 exhibited potent cytotoxicity. Therefore, these analogues can be used as precursors of molecular probes to identify target molecules. For example, compound 44 was successfully coupled with phenylacetylene (CuSO₄/sodium ascorbate/t-BuOH–H₂O) without decomposition of the thiazoline moiety to yield 46 (65%), which also retained potent cytotoxic activity.²⁰⁾

Fig. 4. Apratoxin A (1) and Its Synthetic Analogues

Table 1. Cytotoxicities of Synthetic 1 and Its Analogues for HeLa Cells

Compound	Substituenta)	IC50 (µM)
Apratoxin A (1)		0.19
34-Epimer (29)		0.19
16		0.38
36		28
40	X=OSiEt3	34
41	R1=(CH2)4N3	0.45
42	R2=(CH2)4N3	1.7
43	R2=CH2N3	2.1
44	R3=O(CH2)7N3	0.081
45	R3=(OCH2CH2)3N3	0.35
46	R3=O(CH2)7	0.030

a) Different substituent from 1 is noted in 40–46.

Recently, Robertson et al. achieved the total synthesis of apratoxin D.²¹⁾ Chen et al. carried out structure–activity relationship studies on several analogues of apratoxin A and reported that one of the analogues showed potent antitumor activity in vivo without evidence of toxicity.²²⁾ The mechanism of the action has been proposed to be the inhibition of cotranslational translocation. Further synthetic studies on apratoxin derivatives, including apratoxin C, elucidation of structure–activity relationships, three-dimensional structure analysis based on both theoretical calculation and NMR information, and protein network analysis using a molecular probe, are underway in our laboratories. The results will be reported in due course.

Acknowledgment

This study could not have been conducted without the extensive efforts of my colleague, Dr. Yoshitaka Numajiri (at present with Toray Industries, Inc.). I thank all co-workers cited in the references, especially Professor Takashi Takahashi (Yokohama College of Pharmacy), who supported my research at Tokyo Tech. I also thank Professor Kou Hiroya (at present with Musashino University), Dr. Kiyofumi Inamoto, Dr. Masahito Yoshida, Dr. Hirokazu Tsukamoto, Dr. Yuichi Masuda, and all students of the Hannou Laboratory at Tohoku University. This study was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (Nos. 14103013 and 23310145); a Grant-in-Aid for Scientific Research on Innovative Area (No. 2105); and the Naito Foundation. I deeply thank my friends at the Graduate School of Medicine, Dentistry and Pharmaceutical Sciences at Okayama Univ. and at the School of Pharmaceutical Sciences at the Univ. of Shizuoka for the voluntary food aid right after the Great East Japan Earthquake. I appreciate very kind personal support from the Tsuji–Yamamoto–Takahashi group members (Tokyo Tech). I am also grateful for the financial support not only from MEXT but also from many countries in the world for helping in the recovery from damage by the earthquake.

References

• 1) Lemmens-Gruber R., Kamyar M. R., Dornetshuber R., Curr. Med. Chem., 16, 1122–1137 (2009).
• 2) Stolze S. C., Kaiser M., Molecules, 18, 1337–1367 (2013).
• 3) Pelay-Gimeno M., Tulla-Puche J., Albericio F., Mar. Drugs, 11, 1693–1717 (2013).
• 4) Luesch H., Yoshida W. Y., Moore R. E., Paul V. J., Corbett T. H., J. Am. Chem. Soc., 123, 5418–5423 (2001).
• 5) Luesch H., Yoshida W. Y., Moore R. E., Paul V. J., Bioorg. Med. Chem., 10, 1973–1978 (2002).
• 6) Chen J., Forsyth C. J., Proc. Natl. Acad. Sci. U.S.A., 101, 12067–12072 (2004).
• 7) Gilles A., Martinez J., Cavelier F., J. Org. Chem., 74, 4298–4304 (2009).
• 8) Wipf P., Venkatraman S., Synlett, 1997, 1–10 (1997).
• 9) Walker M. A., Heathcock C. H., J. Org. Chem., 57, 5566–5568 (1992).
• 10) Raman P., Razavi H., Kelly J. W., Org. Lett., 2, 3289–3292 (2000).
• 11) You S., Razavi H., Kelly J. W., Angew. Chem. Int. Ed., 42, 83–85 (2003).
• 12) Sakakura A., Kondo R., Ishihara K., Org. Lett., 7, 1971–1974 (2005).
• 13) Wipf P., Uto Y., J. Org. Chem., 65, 1037–1049 (2000).
• 14) Zou B., Wei J., Cai G., Ma D., Org. Lett., 5, 3503–3506 (2003).
• 15) Ma D., Zou B., Cai G., Hu X., Liu J. O., Chemistry, 12, 7615–7626 (2006).
• 16) Doi T., Numajiri Y., Munakata A., Takahashi T., Org. Lett., 8, 531–534 (2006).
• 17) Numajiri Y., Takahashi T., Doi T., Chem. Asian J., 4, 111–125 (2009).
• 18) Tidgewell K., Engene N., Byrum T., Media J., Doi T., Valeriote F. A., Gerwick W. H., ChemBioChem, 11, 1458–1466 (2010).
• 19) Rasoul F., Ercole F., Pham Y., Bui C. T., Wu Z., James S. N., Trainor R. W., Wickham G., Maeji N. J., Biopolymers, 55, 207–216 (2000).
• 20) Doi T., Numajiri Y., Takahashi T., Takagi M., Shin-ya K., Chem. Asian J., 6, 180–188 (2011).
• 21) Robertson B. D., Wengryniuk S. E., Coltart D. M., Org. Lett., 14, 5192–5195 (2012).
• 22) Chen Q.-Y., Liu Y., Luesch H. ACS Med. Chem. Lett., 2, 861–865 (2011).