Modern synthetic pathways towards eribulin and its subunits
Sebastian Dominik Graf
TL;DR
This review discusses recent advances in the synthesis of eribulin, a drug used to treat breast cancer and liposarcoma.
Contribution
The paper summarizes new and efficient synthetic pathways for eribulin and its subunits developed in the last decade.
Findings
Recent methods focus on assembling the four key fragments of eribulin more efficiently.
New total synthesis pathways for eribulin have emerged to meet clinical demand.
The review highlights advancements in scalable techniques for eribulin production.
Abstract
Eribulin is a synthetic analog of halichondrin B, a natural product derived from marine sponges, and has gained significant importance in oncology (as its commercial mesylate salt, Halaven) due to its unique mechanism of action as a microtubule dynamics inhibitor. It is primarily used in the treatment of metastatic breast cancer and liposarcoma, offering a new therapeutic option for patients with advanced disease. To meet the increasing clinical demand, the research on new synthetic approaches is rigorously ongoing. Recent procedures mainly focus on more efficient and scalable techniques for the assembly of the 4 key fragments of eribulin. But also new pathways for the total synthesis have emerged in the last decade. In this review the latest advancements towards the construction of eribulin are summarized.
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Scheme 1
Scheme 2
Scheme 3
Scheme 4
Scheme 5
Scheme 6
Scheme 7
Scheme 8
Scheme 9
Scheme 10
Scheme 11
Scheme 12
Scheme 13
Scheme 14
Scheme 15
Scheme 16
Scheme 17
Scheme 18
Scheme 19
Scheme 20
Scheme 21
Scheme 22
Scheme 23
Scheme 24
Scheme 25
Scheme 26
Scheme 27
Scheme 28
Scheme 29
Scheme 30
Scheme 31
Scheme 32
Scheme 33
Scheme 34
Scheme 35
Scheme 36
Scheme 37| Authors, Date | Fragment of | No. of stepsa | Scalesb | References |
|
| ||||
| Konda, 2016 | C27–C35 | 16 and 12 | mg–g | [ |
| Lee, 2016 | total synthesis | 71 (LLSc: 33) | mg | [ |
| Choi, 2017 | C14–C35 | 51 (LLS: 29) | mg–g | [ |
| Gaddam, 2018 | C1–C13 with C28–C35 | 27 | mg–g | [ |
| Kim, 2018 | C1–C15 | 29 (from | mg–g | [ |
| Kim, 2019 | C1–C14 | 15 | g | [ |
| Lee, 2019 | C1–C13 | 23 | g–kg | [ |
| Kathravath, 2019 | C1–C11 | 17 | mg–g | [ |
| Krishna, 2020 | C14–C19‘ | 5 | g | [ |
| Mallurwar, 2021 | C14–C21 | 16 and 15 | mg–g | [ |
| Kim, 2021 | C1–C13 | 21 | kg | [ |
| Senapati, 2021 | C14–C28 | 16 | mg–g | [ |
| Nicolaou, 2021 | C1–C26‘ | 35 (LLS: 17) | mg–g | [ |
| Nicolaou, 2022 | total synthesis | 62 (LLS:28) | mg–g | [ |
| Kim, 2022 | C14–C23 | 13 | kg | [ |
| Yu, 2023 | C14–C23 | 16 | g | [ |
| Kaghad, 2023 | C14–C35 | 30 (LLS: 20) | mg–g | [ |
| Nasam, 2023 | C14–C26 | 17 (LLS: 14) | mg–g | [ |
| Kim, 2024 | C14–C35 | 29 (LLS: 27) | g–kg | [ |
| Kim, 2025 | total synthesis | 74 (LLS: 32) | mg–g | [ |
| Kathravath, 2025 | C22–C28 | 15 | mg–g | [ |
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCancer Treatment and Pharmacology · Microtubule and mitosis dynamics · Microbial Natural Products and Biosynthesis
Introduction
Eribulin (1) is a truncated derivative of halichondrin B (2), a complex natural product originally isolated from the marine sponge Halichondria okadai (Figure 1) [1–5]. Already within their isolation study on halichondrin B, in 1986, Hirata and Uemura showed its promising activity against murine cancer cells [6], which led to a great interest in the pharmaceutical society [7–20]. Only 6 years later, Kishi and co-workers first described the total synthesis of the marine natural product [19] and shortly thereafter, also its simplified structure, 1, was assembled and showed similar anticancer behavior [19,21]. Since 2010, the mesylate salt of 1 is approved by the U.S. Food and Drug Administration (FDA) for the treatment of patients with locally advanced breast or metastatic cancer and has evolved to a commonly used agent for this type of cancer in nowaday’s medicine (commercial name: Halaven) [22–30]. Therefore, the discovery of 1 also marked a significant milestone in the field of medicinal chemistry, as it exemplifies the successful translation of marine natural products into effective therapeutic agents.
Eribulin with common synthetic precursor fragments and halichondrin B.
The clinical importance of 1 primarily stems from its efficacy in treating aggressive and refractory cancers, notably metastatic breast cancer and liposarcoma [31–35]. Historically, treatment options for advanced breast cancer have been limited, especially after patients have progressed on initial therapies such as anthracyclines and taxanes [36–40]. 1 has demonstrated a significant survival benefit in this setting. Similarly, in liposarcoma, a rare but challenging soft tissue sarcoma, 1 has shown promise in prolonging overall survival and improving quality of life. The key feature that underpins 1's clinical importance is its unique mechanism of action [41–50]. Unlike other microtubule inhibitors such as taxanes and vinca alkaloids, 1 binds to a specific site on tubulin, inhibiting microtubule growth without affecting its disassembly. This results in the suppression of mitotic spindle formation, leading to cell cycle arrest at the G2/M phase and subsequent induction of apoptosis. Its distinctive mode of action not only enhances its therapeutic efficacy but also helps in overcoming resistance mechanisms that limit the effectiveness of other microtubule-targeting agents.
Research continues to explore additional applications of 1 in various cancer types, including non-small cell lung cancer, ovarian cancer, and other soft tissue sarcomas. Moreover, ongoing studies aim to optimize combination therapies involving 1 with targeted agents, immunotherapies, and other chemotherapeutics to enhance its efficacy and reduce adverse effects [51–65].
Given the challenging structure of 1, its rapid development from first total synthesis to large scale production also highlights advances in the realm of synthetic chemistry. This progress was indispensable to ensure todays broad accessibility, since only smallest quantities would be obtainable through isolation form its natural source (12.5 mg from 600 kg *Halichondria okadai*) [[6](#R6)]. The commercial manufacture of **1** is carried out by the company Eisai and involves multiple linear and convergent synthesis paths ([Scheme 1](#C1), 67 steps in total with the longest linear sequence of 32 steps from **8**→**9**→**12**→**1**), which aim towards the merger of C1–C13 fragment **11** with C14–C35 fragment **12** [[66](#R66)–[68](#R68)]. Despite these great advances, still, the research on improving synthetic efficiency, reducing production costs, omitting toxic chemicals, as well as on new pathways towards **1**s 4 heterocyclic precursor fragments is rigorously ongoing [69–71]. In 2016, Bauer already reported on the current state of research, focusing on contributions from Kishi and co-workers and the Eisai process [71]. However, due to the great demand of 1, this research field continues to grow. In this context, the following review should summarize and explain modern approaches towards the key fragments and total synthetic strategies for 1 in recent years.
Overview of the industrial process pathway for the large-scale production of the mesylate salt of 1 by Eisai.
Review
In 2016, Konda and co-workers reported two approaches for the assembly of the tetrasubstituted tetrahydrofuran unit of 1 (Scheme 2 and Scheme 3) [72]. For the first path, (S,S)-tartaric acid (13) was used as a starting material and was protected as acetonide within the first step to enable the reduction of both acid moieties towards 14 (Scheme 2). Bn-protection, followed by oxidation and olefination yielded sulfone 15, which was vinylated leading to 16 as a single diastereomer. Further Grubbs metathesis with ethyl acrylate, acidic cleavage of the diol protecting group and addition of NaH induced the oxy-Michael reaction towards 18 in 4:1 dr. Reduction of the ester moiety, subsequent protection with TBDPSCl and methylation of the secondary alcohol furnished 19. After Bn-deprotection, iodination and addition of vinylmagnesium bromide, 21 was received and dihydroxylated towards 22 in 5:1 dr.
Synthesis of 22. (a) i. 2,2-dimethoxypropane, p-TsOH, MeOH, 65 °C; ii. NaBH4, MeOH, rt; (b) i. NaH, BnBr, THF, rt; ii. iodobenzoic acid, MeCN, 80 °C; iii. PhOOSCH2PO(OEt)2, NaH, THF, 0 °C; (c) vinylmagnesium bromide, CuI, THF, −78 °C; (d) i. ethyl acrylate, Grubbs cat 2nd generation, toluene, 110 °C; ii. p-TsOH, H2O, THF, 60 °C; (e) NaH, THF, 0 °C; (f) i. LiAlH4, THF, 0 °C to rt; ii. TBDPSCl, imidazole, DCM, 0 °C to rt; iii. MeI, Ag2O, DMF, 0 °C to rt; (g) i. Pd/C, H2, EtOAc; ii. I2, PPh3, imidazole, DCM, 0 °C to rt; (h) vinylmagnesium bromide, CuI, HMPA, THF, −30 °C; (i) OsO4, (DHQ)2PYR, 0 °C; (DHQ)2PYR: hydroquinine 2,5-diphenyl-4,6-pyrimidinediyl diether.
The second path towards 27 commenced with the assembly of 23 from 14 via Bn-protection, following oxidation and Horner–Wadsworth–Emmons (HWE) reaction (Scheme 3). Stereospecific vinylation with a Gilman cuprate and acidic treatment afforded 25 in 13:1 dr. After protection with 2,6-DCBCl, reduction with LiAlH_4_, TBDPS-protection and methylation, 26 was received in 35% yield. Eventually, the addition of I_2_ triggered an iodocyclization towards 27. While in the first sequence only mg amounts of 22 were received, 1.45 g of 27 were obtainable during the second one from one batch (Scheme 3, step e) showing the scalability of this path. Moreover, besides 22 and 27, Konda and co-workers also accomplished the stereoselective syntheses of other diastereomers of the target tetrahydrofuran unit starting from 13.
Synthesis of 27. (a) i. NaH, BnBr, THF, rt; ii. iodobenzoic acid, MeCN, 80 °C; iii. (EtO)2POCH2COOEt, NaH, THF, 0 °C; (b) vinylmagnesium bromide, CuI, THF, TMSCl, HMPA, −78 °C to rt; (c) 2 M HCl, THF, H2O, 65 °C; (d) i. Ag2O, 2,6-DCBCl, DMF, TBAI, rt, darkness; ii. LiAlH4, THF, 0 °C to rt; iii. TBDPSCl, imidazole, DCM, 0 °C to rt; iv. NaH, MeI, DMF, 0 °C; (e) I2, MeCN, −20 °C; 2,6-DCBCl: 2,6-dichlorobenzyl chloride.
The Kishi group has made major contributions in the area of 1 synthesis over the last three decades [69,71]. Also recently, the group established a versatile protocol for the macrocyclization towards precursor 36 (Scheme 4 and Scheme 5) [73]. Herein, fragments 31 and 33 were fused via Nozaki–Hiyama–Kishi (NHK) coupling and Pd-mediated cyclization. Fragment 31 was synthesized from known precursor 28 [74] in 9 steps via MMTr-protection, replacement of the Bn- with TBDMS-protecting groups, hydroxylation of sulfone 29 to alcohol 30, tosylation, bromide substitution, acidic MMTr-cleavage and DMP-oxidation (Scheme 4, above). For the assembly of 33, only the hydrolysis of previously reported 32 [75] with Me_3_SnOH and thioesterification using EtSH and DCC were necessary (Scheme 4, below).
Synthesis of 31 and 33. (a) i. MMTrCl, iPr2NEt, DCM, rt; ii. K2CO3, MeOH, DCM, rt; iii. TBDMSCl, imidazole; (b) i. n-BuLi, THF, −78 °C, Sia2BH, −10 °C to rt; ii. H2O2, 3 M NaOH, 0 °C; (c) i. TsCl, DMAP, NEt3, DCM, rt; ii. NaBr, n-Bu4NBr, acetone, reflux; iii. HFIP, H2O, rt; iv. DMP; (d) i. Me3SnOH, DCE, 80–85 °C, then 0.1 M HCl; ii. EtSH, DCC, DMAP, DCM, rt; MMTrCl: 4-methoxytriphenylmethyl chloride; Sia2BH: disiamylborane; HFIP: 1,1,1,3,3,3-hexafluoroisopropanol.
Both fragments (31 and 33) were fused together via NHK coupling to furnish 34 in 86% yield. The addition of SrCO_3_ proved to be suitable for the cyclization towards 35 and the eventual macrocyclization was achieved via coupling of the alkyl bromide unit with the thioester. Mechanistically, this reaction is enabled by the formation of an intermediate alkylzinc halide, which is produced by single electron transfer using CrCl_3_ and NbCpCl_4_. The last steps towards 1 are known procedures [68]. By comparison to former methods, this technique does not require further desulfonylation after the macrocyclization [19,68].
Synthesis of 1. (a) CrCl2, 37, 38, 39 (proton sponge), LiCl, Mn, ZrCp2Cl2, MeCN, EtOAc; (b) SrCO3, t-BuOH, H2O, open air; (c) Pd2dba3, PCyp3, CrCl3, NbCpCl4, Zn(0), DMI, THF; PCyp3: tricyclopentylphosphine; DMI: 1,3-dimethyl-2-imidazolidinone.
Choi and co-workers used a previously reported protocol on stereo- and regioselective allene-Prins reactions [76–77] for the assembly of fragment 45 (Scheme 6) [78]. Here, 40 and 42 served as the substrates for the allene-Prins reaction towards 43. Notably, the Bz-derivative of 40, 41, served as a starting point for a corresponding halichondrin B analog. The stereoselective course of this cyclization is described in Scheme 6 below (R^1^ and R^2^ in equatorial position). From 43, the methylene unit in 44 was formed by Pd(0)-mediated generation of a π-allyl–palladium intermediate, followed by reductive termination (Tsuji-reduction), alongside the substrate-controlled alignment of the adjacent methyl substituent. Eventual change of the two PNB- with TBDMS-protecting groups yielded 45 in 58%. Notably, by this technique, the central pyran motif was assembled in one step and required only the stereoinformation of 42`s alcohol unit, while in former works two individual steps and the use of expensive metal catalysts were necessary [66–6873,79].
Synthesis of 45. Above: Reaction conditions: (a) methoxyacetic acid, BF3·OEt2, DCM, −30 °C; (b) Pd(PPh3)4, PPh3, HCOOH, NEt3, THF, 60 °C; (c) i. Mg(OMe)2, THF, MeOH, rt; ii. TBDMSCl, imidazole, DMF, rt. Below: Mechanism of the allene-Prins reaction.
In early 2018, Gaddam and co-workers assembled the northern fragment of 1 from ᴅ-mannose (48, Scheme 7) [80]. The synthesis commenced with the protection of 48 as acetonide, vinylation and cyclization to 49 by the treatment with TsCl according to a procedure from Liu and co-workers [81]. Regioselective deprotection, oxidative cleavage, reduction with NaBH_4_ and Bn-protection afforded 50, which was subsequently transformed to aldehyde 51 via ozonolysis and treatment with SMe_2_. Next, HWE reaction and acidic treatment triggered the cyclizations towards enone 53. Protection of the free alcohol unit enabled the transformation towards 55, which involved the reduction of lactone to lactole, protection of the alcohol as acetate, BF_3_·Et_2_O-mediated C-allylation of the aldehyde (via oxocarbenium intermediate) and cyclization (oxy-Michael reaction). Hydroboration–oxidation, stereoselective dihydroxylation with AD-mix β and diol-protection yielded acetonide 57. Another DMP-oxidation, followed by HWE reaction, and deprotection of the diol motifs enabled the cyclization towards 60. After TBDMS-protection and reduction to the respective aldehyde (62), an alkylation with the α-sulfonyl carbanion of 65 (intermediate from the Eisai process [68]) via interrupted Julia olefination was performed to furnish 63. Eventually, 64 was received by oxidation and SmI_2_-induced desulfonylation. Herein, Gaddam and co-workers enable the synthesis of 1 by the merger of northern fragment 64 with a potential southern fragment (ongoing research by the same group) and thereby provide an alternative approach to the current one by Eisai [66].
Synthesis of 64. Reaction conditions: (a) i. acetone, I2, rt; ii. vinylmagnesium bromide, THF, −20 °C to rt; iii. TsCl, pyridine, 65 °C; (b) i. AcOH, rt; ii. NaIO4, MeOH, 0 °C to rt; iii. NaBH4, MeOH, 0 °C to rt; iv. NaH, BnBr, 0 °C to rt; (c) i. O3, DCM, −78 °C; ii. SMe2; (d) (PhO)2POCH2COOMe, KHMDS, 18-crown-6, THF, −78 °C; (e) p-TsOH, toluene, H2O, 100 °C; (f) MOMCl, DIPEA, DCM, 0 °C to rt; (g) i. DIBAL-H, DCM, −78 °C; ii. Ac2O, pyridine, DMAP, DCM, rt; iii. allyl-TMS, BF3·Et2O, DCM, −78 °C; (h) 9-BBN, THF, H2O2, 0 °C to rt; (i) i. AD-mix β, MsNH2, t-BuOH, H2O, rt; ii. 2,2-dimethoxypropane, (±)-CSA, DCM, rt; (j) i. DMP, DCM, rt; ii. Ph3PCH2COOEt, toluene, rt; (k) p-TsOH, EtOH, reflux; (l) i. DBU, toluene, reflux; ii. TBAF, THF, rt; (m) TBDMSCl, imidazole, DCM, 50 °C; (n) DIBAL-H, toluene, −78 °C; (o) 36, n-BuLi, −78 to −50 °C; (p) i. DMP, DCM, rt; ii. SmI2, MeOH, THF, −78 to −50 °C; 18-crown-6: 1,4,7,10,13,16-hexaoxacyclooctadecane, 9-BBN: 9-borabicyclo[3.3.1]nonane, (±)-CSA: camphorsulfonic acid.
Shortly thereafter, Kim and co-workers made use of 66, a byproduct formed during the multigram synthesis of Halaven, to assemble 79 (Scheme 8) [82]. 66 was received within a filtrate during this process and isolated as a diastereomeric mixture of 2.6:1. At first, the benzyl moieties of 66 were cleaved off and the diol was reprotected as acetonide to afford 67. The remaining secondary alcohol was PMB-protected and the olefin was oxidatively cleaved, before addition of a lithiated furanyl unit took place. The so-obtained 1:1 diastereomeric mixture of 68 was treated with (PhO)3_PMeI in dimethylacetamide and formed an intermediate trans-olefin, which subsequently underwent an asymmetric Sharpless dihydroxylation. The acetonide was cleaved with AcOH and the addition of NBS under basic conditions triggered an Achmatowicz rearrangement (shown in Scheme 8, below) to assemble a hydropyranone ring [83]. Eventual acetylation of the free alcohol units afforded 70 in 3:1 dr. C-Glycolysation with allyl-TMS and oxidative cleavage of the PMB moiety yielded 71. The stereochemistry of the secondary alcohol unit of 71 was changed via oxidation and stereospecific reduction, then 1,4-reduction of the enone and treatment with camphorsulfonic acid led to fully cyclized intermediate 72 as a single isomer. During this process the acetate groups were cleaved off and had to be reinstalled using acetic anhydride. Oxidative cleavage of olefin 73, followed by treatment with ʟ-proline led to an intermediate β-aldehyde with inverted stereochemistry at the β-C (4:1 dr). Subsequent reduction with NaBH_4 yielded alcohol 74. Piv-protection of the primary alcohol, reductive deprotection of the Bn-moiety and reprotection with (iPr)2_SiHCl afforded 75. The addition of BF_3·OEt_2_ at low temperatures induced the fluorination of the silane protecting group and a follow-up intramolecular hydride transfer (via 76). Afterwards, the introduced fluoride was substituted with Mg(OMe)2 in MeOH and the stereochemistry of the central protected alcohol unit was inverted within a sequence involving deprotection of all acetates, protection of the 1,2-diol as an acetonide, oxidation of the central alcohol with DMP and stereospecific reduction, which yielded alcohol 78. Finally, the silyl protecting group was cleaved with TBAF leading to 79. Although the synthesis of 79 through this method is quite laborious, especially with regard to all necessary adjustments of stereocenters, the idea of recycling byproducts of Halaven production clearly shows the advancements in this process. Therefore, this route should not be valued solely for its total yield, but rather as a starting point for improving atom economy.
Synthesis of 79. Above: Reaction conditions: (a) i. K2CO3, MeOH, 60 °C; ii. 2,2-dimethoxypropane, H2SO4 (aq), acetone, rt; (b) i. PMBCl, t-BuOK, TBAI, DMF, THF, rt; ii. OsO4, NaIO4, 2,6-lutidine, 1,4-dioxane, H2O, −20 °C; iii. N-BuLi, furan, THF, 0 °C; (c) i. (PhO)3PMeI, dimethylacetamide, rt; ii. AD-mix-α, t-BuOH, H2O, rt; (d) i. AcOH, H2O, rt; ii. NBS, NaHCO3, NaOAc, THF, H2O, 0 °C; iii. acetic anhydride, pyridine, DCM, 0 °C; (e) i. allyl-TMS, BF3·OEt2, MeCN, −10 °C; ii. DDQ, DCM, H2O; (f) i. DMP, NaHCO3, DCM, rt; ii. NaBH4, DCM, MeOH, −78 °C; iii. Stryker’s reagent, toluene, rt; iv. (±)-CSA, MeOH, 60 °C; (g) acetic anhydride, DMAP, DCM, rt; (h) i. OsO4, NaIO4, 2,6-lutidine, 1,4-dioxane, rt; ii. ʟ-proline, MeOH, −10 °C; iii. NaBH4, MeOH, 0 °C; (i) i. PivCl, pyridine, DMAP, DCM, rt; ii. H2, Pd/C, MeOH, EtOAc, rt; iii. (iPr)2SiHCl, imidazole, DMAP, DCM, rt; (j) BF3·OEt2, DCM, −20 to −10 °C; (k) i. Mg(OMe)2, MeOH, rt; ii. 2,2-dimethoxypropane, pyridinium p-toluenesulfonate, acetone, rt; iii. DMP, NaHCO3, DCM, rt; iv. NaBH4, MeOH, 0 °C; (l) TBAF, THF, rt. Below: Mechanism of the Achmatowicz rearrangement during step (d).
The same group also reported on a metal-free synthesis of 92 starting from an intermediate occuring during the current large-scale production of Halaven (Scheme 9) [84]. Initially, the diol motif of 85 was protected with TESCl, then regioselective oxidation of the primary TES ether and addition of vinyl grignard led to 87. Another oxidation afforded α,β-unsaturated carbonyl 88 and acidic treatment with BnOH triggered the oxa-Michael reaction and transketalization towards tetracyclic 90. Following Kishi reduction [85], epimerization of the secondary alcohol was effected via DMP-oxidation, and stereospecific reduction with Li(t-BuO)_3_AlH yielded 92. In comparison to the current pathway from 85 to 92, which involves the use of heavy metals adding considerable amounts of cost [19,66], this technique stands out for the employment of cheaper and more eco-friendly reagents.
Synthesis of 92. Reaction conditions: (a) TESCl, imidazole, DCM, 0 °C to rt; (b) i. oxalyl chloride, DMSO, NEt3, DCM, −78 °C; ii. vinylmagnesium bromide, THF, −70 °C; (c) DMP, NaHCO3, DCM, rt; (d) BnOH, p-TsOH·H2O, toluene, 70–75 °C; (e) BF3·OEt2, Et3SiH, DCM, 0 °C to rt; (f) i. DMP, NaHCO3, DCM, 0 °C; ii. Li(t-BuO)3AlH, THF, 0−10 °C.
In 2019, Lee and co-workers developed a technique towards the octahydropyrano[3,2-b]pyran fragment of 1 from ᴅ-gluconolactone (93) (Scheme 10) [86]. After protection of 93, the lactone unit was reduced with DIBAL-H and the emerging aldehyde motif (equilibrium of lactol to aldehyde and alcohol) was trapped by HWE reaction to obtain 95. The secondary alcohol unit of 95 was protected and the ester reduced to the respective alcohol, before Sharpless epoxidation, oxidation of the alcohol and subsequent Wittig reaction yielded allylic epoxide 98. Deprotection of the alcohol motif of 98 enabled the cyclization towards tetrahydropyran 99. Next, 99 underwent an esterification with acrylic acid, was cyclized via Grubbs metathesis and the remaining double bond was hydrogenated leading to bicyclic core structure of 101. Notably, during the cyclization also considerable amounts of dimerized product were observed. Again, reduction with DIBAL-H led to the regioselective deprotection of the alcohols and reduction to an intermediate lactol unit, which is in equilibrium with its acyclic aldehyde and alcohol. The aldehyde was trapped by HWE reaction and the so obtained α,β-unsaturated carbonyl was reattacked by the adjacent hydroxy moiety via oxy-Michael reaction to form 102. This sequence is shown in detail in Scheme 10 below. From here, a sequence involving the oxidation of the unprotected 1,2-diol moiety towards an intermediate aldehyde, Ni(II)/Cr(II)-mediated coupling of 1-bromo-2-trimethylsilylethene, acidic cleavage of the remaining cyclohexylidene ring, TBDMS-protection of the three alcohol units and electrophilic substitution of the silyl moiety to afford vinyl iodide 103 was applied. Eventually, ester 103 was reduced to target aldehyde 104, which contains the necessary functional group pattern to be used as a building block for the assembly of 1. By the design of this novel synthetic pathway, Lee and co-workers showed that the commonly used NHK reaction involving dual chromium/nickel catalysis can be circumvented [86]. Here, intramolecular ring opening of epoxide 98 and metathesis led to target 104 with high regio- and stereoselectivity.
Synthesis of 104. Above: Reaction conditions: (a) cyclohexanone, p-TsOH, toluene, 110 °C, crystallization. (b) i. DIBAL-H, THF/toluene, −10 °C; ii. Ph3PCHCO2Et, benzoic acid, DCM, 50 °C; (c) i. TESOTf, 2,6-lutidine, DCM, 0 °C; ii. 5 mol % Pn-Bu3 (5 mol %), THF, 50 °C; (d) DIBAL-H, DCM, 20 °C; (e) i. (−)-DET, Ti(OiPr)4, DCM, 0 °C; ii. DMP, DCM, 0 °C; iii. MePPh3Br, NaHMDS, THF, 0 °C; (f) i. TBAF, THF, 0 °C; ii. pyridinium p-toluenesulfonate, DCM, 0 °C; (g) acrylic acid, DIC, DMAP, DCM, rt; (h) i. Grubbs cat 2nd generation, toluene, reflux; ii. Pd/C, H2, EtOAc; (i) i. DIBAL-H, THF/toluene, 10 °C ; ii. triethyl phosphonoacetate, t-BuOK, THF, 60 °C; iii. AcOH, 40 °C; (j) i. NaIO4, EtOAc, 15 °C; ii. CrCl2, NiCl2, 1-bromo-2-trimethylsilylethene, DMSO, MeCN, 30 °C; iii. AcOH, H2O, 95 °C, crystallization; iv. TBDMSOTf, 2,6-lutidine, MTBE, 30 °C, crystallization; v. NIS, MeCN, toluene, TBDMSCl, 35 °C; (k) DIBAL-H, 2,6-di-t-Bu-4-hydroxytoluene, toluene, 65 °C. Below: Reaction sequence towards 102; DIC: N,N′-diisopropylcarbodiimide.
A novel approach towards C1–C10 fragment of 1 was reported by Kathravath and co-workers (Scheme 11) [87]. From ʟ-ascorbic acid (108), protection of the diol motif and treatment with H_2_O_2_, then EtI yielded ester 109. The alcohol motif was protected with BnBr, before the ester unit was completely reduced to the respective alcohol and selectively oxidized to aldehyde 110. After HWE reaction, the dihydroxylation of (Z)-111 furnished diol 112 in 7:3 dr*.* The shown major diastereomer of 112 was cyclized towards 113 upon acidic treatment and TBDMS-protection. Reduction, cleavage of the silyl-protecting groups and acetylation led to tetrahydropyran 114. Lewis acid-catalyzed C-allylation, cross metathesis, basic cleavage of the acetate motifs and transesterification enabled the DBU-induced isomerization of 116’s double bond and following oxy-Michael reaction to the target compound 117. In total, 117 was obtained in a yield below 1% via the 17-step sequence, mainly due to mediocre yields (partially due to unselective reactions) in steps c→g.
Synthesis of 117. (a) i. acetone, CuSO4, rt; ii. H2O2, K2CO3, H2O, rt; iii. EtI, MeCN, 70 °C; (b) i. Ag2O, BnBr, toluene, rt; ii. LiAlH4, THF, rt; iii. (COCl)2, NEt3, DMSO, DCM, −78 °C; (c) (PhO)2POCH2COOEt, DBU, THF, −78 °C; (d) K3(FeCN)6, OsO4, K2CO3, (DHQ)2PHAL, t-BuOH, H2O, 0 °C; (e) i. TFA, H2O, MeCN, 70 °C; ii. TBDMSCl, imidazole, DMAP, DMF, 55 °C; (f) i. DIBAL-H, toluene, −78 °C; ii. TBAF, THF, rt; iii. Ac2O, NaOAc, 90 °C; (g) BF3·OEt2, MeCN, 80 °C, allyltrimethylsilane; (h) allyl-COOBn, Grubbs cat 2nd generation, DCM, 40 °C; (i) i. K2CO3, MeOH, rt; ii. DBU, toluene, 100 °C; (DHQ)2PHAL: hydroquinine 1,4-phthalazinediyl diether.
In the following year, Krishna and co-workers used an enzymatic transformation for the continuous flow production of acetate 121 (C14–C19‘ fragment), which is a starting material needed in large quantities for the total synthesis of 1 (Scheme 12) [88]. Herein, pent-4-en-1-ol (118) was protected with TBDPSCl, before ozonolysis, followed by reduction with PPh_3_ yielded aldehyde 119. The addition of propargyl bromide led to racemic 120. By the use of a flow setup involving a column packed with Amano lipase, an enzyme from the bacterium Pseudomonas fluorescens, a kinetic resolution of 120 was performed leading to the continuous production of acetate 121 and free alcohol 122. Here, 121 bears the right configuration needed for Halaven synthesis, but the authors also showed that 122 was easily converted to 121 via Mitsunobu inversion. Although 121 only represents a small building block for the total assembly of 1, this method especially stands out for its cost-efficiency and the continuous production in bigger scales (100 mg/mL (272 mM) at 0.1 mL/min).
Synthesis of 121. Reaction conditions: (a) i. TBDPSCl, imidazole, DMF, rt; ii. O3, DCM, −78 °C; iii. PPh3, rt; (b) propargyl bromide, Zn, NH4Cl, THF, −15 °C; (c) enzyme, vinyl acetate, MTBE, continuous flow reactor, rt; enzyme: Amano lipase produced by Pseudomonas fluorescens.
In 2021, Mallurwar and co-workers used (R,R)-tartaric acid (123) for the assembly of fragment 131 (Scheme 13) [89]. After diol-protection as acetonide and reduction of the acid motifs to receive diol 124, one alcohol moiety was Bn-protected, while the other one was iodinated. From 125, the addition of vinylmagnesium bromide and acidic deprotection furnished diol 126, which underwent an iodocyclization towards a diastereomeric mixture of 127 (4:1 dr). The shown main diastereomer of 127 could be isolated and underwent TBDPS-protection, which enabled the malonic ester synthesis towards 128. Reduction with subsequent Piv-protection and TBDPS-deprotection was followed by DMP-oxidation. Finally, the olefination of 129 with PPh_3_MeBr and addition of TiCl_4_ for the cleavage of the Bn-moiety furnished target structure 131.
Synthesis of 131. (a) i. 2,2-dimethoxypropane, p-TsOH, MeOH, 60 °C; ii. LiAlH4, THF, 0 °C to rt; (b) i. NaH, BnBr, THF, rt; ii. PPh3, imidazole, I2, THF, 50–60 °C; (c) i. vinylmagnesium bromide, CuI, HMPA, −30 °C; ii. p-TsOH, THF, H2O, 23 °C; (d) I2, NaHCO3, Et2O, H2O, 0 °C; (e) i. TBDPSCl, imidazole, DCM, 0 °C; ii. NaH, diethyl malonate, TBAI, DMF, 120 °C; iii. DMSO, NaCl, 160 °C; (f) i. LiAlH4, THF, 0 °C to rt; ii. PivCl, NEt3, DCM, 0 °C to rt; iii. TBAF, THF, 0 °C to rt; iv. DMP, DCM, 0−20 °C; (g) PPh3MeBr, n-BuLi, THF, 0 °C; (h) TiCl4, DCM, 0 °C.
Within the same work, a similar fragment (143) was synthesized using an oxy-Michael reaction as the key step (Scheme 14). Starting material 132 was obtained from natural ʟ-ascorbic acid (108) by a sequence involving the protection of diol motif, oxidative cleavage, reduction and regioselective Bn-protection. 132 was further iodinated and reacted with vinylmagnesium bromide to afford 133 in 48% yield. Consecutive acidic deprotection, TBDMS-protection of the secondary alcohol, dihydroxylation, oxidative cleavage and olefination with Ph_3_PCHCOOEt yielded acrylic acid ester 135. The addition of benzyltrimethylammonium hydroxide induced the oxy-Michael reaction towards an unseparable mixture of 137 and 138, and 136. A plausible explanation for the formation of 138 could be a [1,2]-rearrangement of the deprotonated alcohol motif and the OTBDMS-moiety of 137 via an epoxide intermediate (indicated in arrows) and subsequent oxy-Michael reaction. The mixture (137 + 138) was treated with p-TsOH leading to alcohols 139 and 140, which were separated via chromatography. From 140, Bn-deprotection of the primary, followed by Bn-protection of the secondary alcohol unit afforded 141 in 52% yield. Eventually, the ester motif of 141 was reduced to the respective alcohol and Piv-protected, and the secondary alcohol oxidized to an intermediate ketone, then olefinated towards 143. Although both sequences require multiple applications and removals of protecting groups, they represent attractive pathways towards the C14–C21 fragment of 1 due to the use of inexpensive starting materials.
Synthesis of 143. (a) i. I2, PPh3, imidazole, DCM; ii. HMPA, CuI, vinylmagnesium bromide, THF, −20 °C; (b) i. p-TsOH, MeOH, rt; ii. TBDMSCl, imidazole, DCM; (c) i. OsO4, NMO, THF, H2O, 0 °C; ii. NaIO4, THF, H2O, 0 °C; iii. Ph3PCHCOOEt, DCM, rt; (d) benzyltrimethylammonium hydroxide, EtOH, 0 °C to rt; (e) p-TsOH, MeOH, rt; (f) i. Pd/C, H2, EtOAc; ii. Ag2O, BnBr, toluene, rt; (g) i. LiAlH4, THF, 0 °C; ii. PivCl, NEt3, DMAP, DCM, rt; (h) i. DMP, DCM, 0 °C; ii. PPh3CH2Br, n-BuLi, THF, 0 °C to rt.
Kim and co-workers enhanced the procedure from Lee and co-workers by changing few intermediate steps (Scheme 15) [86,90]. From 144, the authors performed a HWE reaction towards 145 in 80%. Cleavage of the TES-protecting group and Rh(I)-induced cyclization afforded 147 in good yield, which was hydrogenated subsequently. Thereafter, DIBAL-H reduction led to lactol formation towards 149, which intercepted the former route (Scheme 10, step (i) i.) With these changes, the synthesis was enhanced with regard to scalability of the process. Scaling up the Grubbs metathesis of the first route (Scheme 10, step (h) i.) led to an increasement of dimer yield derived from cross metathesis reactions. Therefore, by the replacement with the Rh(I)-catalyzed cyclization performed herein, this drawback is circumvented and the reactions were performed with kg amounts.
Modified synthesis of 104. Reaction conditions: (a) (EtO)2POCH2COOEt, KOt-Bu, THF, 15 °C; (b) TBAF, imidazole·HCl, THF, 5 °C to rt; (c) [Rh(CO2)Cl]2 (0.5 mol %), THF, 60 °C; (d) H2, Pd/C, EtOAc, rt; (e) DIBAL-H, toluene, −65 °C.
In 2021, Senapati and co-workers developed a route towards key fragment 161 containing the 3-methylenetetrahydrofuran and the 3-methylenetetrahydro-2H-pyran motifs of 1 (Scheme 16) [91]. As a cheap and commerically available starting material, ᴅ-glyceraldehyde (150) was chosen and allylated with crotyl bromide according to a procedure from Loh and co-workers [92]. The diastereomeric mixture thus obtained was separated via chromatography and (R)-151 was protected using BnBr. After hydroboration–oxidation and DMP-oxidation to the respective aldehyde, Ohira–Bestman reaction was applied to afford 154. Allylation and acidic deprotection led to diol 156 in excellent yields. Next, Au(I)-catalyzed cyclization and reduction led to tetrahydropyran 157, which was TBDMS-protected subsequently. The cross metathesis of 158 and 162 (obtained from ʟ-glutamic acid) yielded 159 in 81%. After mesylation of the free alcohol moiety and dihydroxalation using AD-mix α, a diastereomeric mixture of tetrahydropyrans (R)-160 and (S)-160 was obtained and separated for analytical purposes. The Bn-moiety of (S)-160 was cleaved using DDQ and the intermediate diol was further transformed to the respective dicarbonyl via Swern oxidation, and finally olefinated via Wittig reaction. Hence, by the use of one-pot Au(I)-catalyzed cyclization/Kishi reduction for the assembly of the tetrahydropyran and cross metathesis/Sharpless dihydroxylation/etherification for the tetrahydrofuran motifs, this pathway describes a novel and sustainable, but not fundamentally improved alternative in terms of yield towards the structure of 161 (7.2% yield over 14 steps).
Synthesis of 161. Reaction conditions: (a) crotyl bromide, Sn, TBAI, NaI, DMF/H2O, rt; (b) NaH, BnBr, DMF, rt; (c) 9-BBN, THF, NaOH·H2O2, EtOH, rt; (d) i. DMP, NaHCO3, DCM, rt; ii. Ohira–Bestmann reagent, MeOH, rt; (e) allyl bromide, K2CO3, CuI, Na2CO3, DBU, DMF, rt; (f) AcOH (60%), rt; (g) i. Au(PPh3)Cl (1 mol %), AgSbF6 (1 mol %), DCM, rt; ii. Et3SiH, BF3·OEt2, 0 °C; (h) TBDMSCl, imidazole, DMF, rt; (i) 162, Grubbs cat 2nd generation, CuI, Et2O, 40 °C; (j) i. MsCl, NEt3, DCM, 0 °C; ii. AD-mix α, MsNH2, t-BuOH, H2O, 0 °C; (k) i. DDQ, DCM, 45 °C; ii. TFAA, DMSO, DIPEA, DCM, −78 °C; iii. Ph3PCH2, toluene, 40 °C.
In the same year, Nicolaou and co-workers achieved the total synthesis of halichondrin B, which included the assembly of C1–C26‘ fragment 186 (Schemes 17–19) [93]. To afford 3-methylene tetrahydrofuran 169, both starting materials, 163 and 164, were prepared within few steps from commercially available sources [93]. The etherification of 163 and 164 (Nicholas reaction) led to a diastereomeric mixture, which was separated (Scheme 17). (S)-165 was converted to 166 via radical cyclization, then the TBDPS-protecting group was cleaved and the obtained alcohol oxidized to aldehyde 167. The Cr(II)/Co(II)-induced asymmetric NHK coupling mediated by 172 with vinyl iodide 171 led to tetrahydrofuran 168. Protection of the secondary alcohol enabled the conversion to ketophosphonate 169 using MePO(OMe)2 and n-BuLi. Notably, building block 171 used herein can be synthesized from 170 within 6 steps including a kinetic resolution with Amano lipase PS-800. [94]
Synthesis of 169. Reaction conditions: (a) i. Co2(CO)8, BF3·Et2O, DCM, 23 °C; ii. CAN, acetone, 0 °C; (b) AIBN, n-Bu3SnH, toluene, 100 °C; (c) i. TBAF, THF, 0–23 °C; ii. DMP, DCM, 0–23 °C; (d) 171, CrCl2, CoPc, 172, LiCl, Mn, 39 (proton sponge, see Scheme 5), ZrCp2Cl2, DME, 23 °C; (e) i. TBDMSOTf, 2,6-lutidine, DCM, −78–0 °C; ii. N-BuLi, MePO(OMe)2, THF, −78 °C; (f) i. TMSCl, NaI; ii. m-MePhCOOH, NaHCO3; iii. ClPO(OEt)2; iv. AlMe3, CuBr; v. Amano lipase PS-800; vi. I2, PPh3, imidazole; CAN: ceric ammonium nitrate.
Another fragment, 181, was synthesized from alcohols 173 and 174, which were perpared from tri-O-acetyl-ᴅ-glucal and (S)-methyl-2,3-dihydroxypropanoate, respectively (Scheme 18) [93]. Again, Nicholas etherification of both starting materials was conducted to afford 175 in 5:1 dr. The major isomer was further transformed to 176 in a microwave oven via Kornblum oxidation. Radical cyclization and DMP-oxidation yielded intermediate 178, which was reduced with K-selectride, before acidic treatment was applied to remove the TBDPS-protecting group and the tin moiety. From 179, ozonolysis and subsequent reduction afforded 180 stereoselectively, and eventual protection of the 1,2-diol motif followed by oxidation yielded aldehyde 181.
Synthesis of 181. Reaction conditions: (a) i. Co2(CO)8, BF3·Et2O, DCM, 23 °C; ii. (NH4)2Ce(NO3)6, acetone, 0–23 °C; (b) DMSO, 2,4,6-collidine, microwave, 180 °C; (c) AIBN, n-Bu3SnH, toluene, 100 °C; (d) DMP, NaHCO3, DCM, 23 °C; (e) i. K-selectride, THF, −20–0 °C; ii. HCl (aq.); (f) i. O3, MeOH, −78 °C; ii. NaBH4, 23 °C; (g) i. 2,2-dimethoxypropane, p-TsOH·H2O, acetone, 0–23 °C; ii. DMP, DCM, 0–23 °C.
Both building blocks from Scheme 17 and Scheme 18 were fused together via HWE reaction to construct 182 (Scheme 19). Treatment with HF·pyridine led to cleavage of the acetonide- and TBDMS-protecting groups, which induced the oxy-Michael cyclization towards 183. Notably, during this cyclization also considerable amounts of 183`s epimer were formed (2:1 dr). Oxidation with DDQ removed the Bn-moiety (184) and triggered ketalization towards 185. Eventual mesylation formed key fragment 186 in 95% yield. For the assembly of both heterocyclic subunits of 186, Nicholas etherification and radical cyclization proved to be suitable, but only moderately yielding, procedures. Further, this pathway, with the exception of the coupling reaction between 167 and 171, involves only simple and mild transformations using standard reagents. As mentioned above, 186 served as an intermediate for the total synthesis of halichondrin B, which was assembled in a total of 25 steps.
Synthesis of 186. Reaction conditions: (a) NEt3, LiCl, MeCN, 0–23 °C; (b) HF·pyridine, MeCN, 23 °C; (c) DDQ, hν, MeCN, 23 °C; (d) MsCl, NEt3, DMAP, DCM, 0 °C.
One year later, Nicolaou and co-workers focused specifically on the total synthesis of 1 and therefore optimized few steps from their previous halichondrin B synthesis (Schemes 20–23) [95]. Hence, for the assembly of fragment 181 the route shown in Scheme 18 was intercepted at 176. Instead of the previous radical cyclization, here, a reductive Ni(II)-induced cyclization afforded 188 (Scheme 20). Oxidation and ozonolysis with subsequent addition of NaBH_4_ yielded diol 190, whose TBDPS-group was cleaved under acidic conditions, before the 1,2-diol motif was protected as acetonide. Eventually, DMP-oxidation afforded aldehyde 181.
Modified synthesis of 181. Reaction conditions: (a) i. Ni(cod)2, P(n-Bu)3, Et3SiH, THF, 23 °C; ii. HCl (aq), 23 °C; (b) DMP, DCM, 0–23 °C; (c) i. O3, MeOH, −78 °C; ii. NaBH4, −78–23 °C; (d) HCl (aq), acetone, MeOH, 0–23 °C; (e) DMP, DCM, 0–23 °C.
The synthesis of fragment 200 was accomplished starting from substrates 192 and 193 (Scheme 21). After Nicholas etherification of both substrates, which led to a diastereomeric mixture, (S)-194 was isolated and further treated with Lindlar catalyst and DIBAL-H to afford the alkene and aldehyde motifs of 195, respectively. Oxime formation and oxidation yielded an intermediate nitrile oxide, which underwent an intramolecular [3 + 2]-cycloaddition with the adjacent ethene substituent towards isoxazoline 196. Reductive N–O bond cleavage and stereospecific reduction with Me_4_NBH(OAc)3 yielded diol 198, whose primary alcohol unit was PMB-protected, while the secondary one was methylated. Dihydroxylation and protection with 2,2-dimethoxypropane yielded 199. Then, desilylation with TBAF and DMP-oxidation led to aldehyde 200.
Synthesis of 200. Reaction conditions: (a) i. Co2(CO)8, DCM, 23 °C; ii. BF3·Et2O, 0 °C; iii. (NH4)2Ce(NO3)6, acetone; (b) i. Lindlar cat., quinoline, EtOAc, 23 °C; ii. DIBAL-H, DCM, −78 °C; (c) i. NH2OH·HCl, pyridine, EtOH, 23 °C; ii. NaOCl, DCM, 23 °C; (d) Mo(CO)6, MeCN, H2O, 80 °C; (e) Me4NBH(OAc)3, MeOH, 0 °C; (f) i. PMB-trichloroacetimidate, (±)-CSA, DCM, hexane, 0–23 °C; ii. NaH, MeI, THF, DMF, 0–23 °C; iii. AD-mix α, NH2Ms, t-BuOH, H2O, 5 °C; iv. 2,2-dimethoxypropane, p-TsOH·H2O, acetone, 23 °C; (g) i. TBAF, THF, 23 °C; ii. DMP, DCM, 0–23 °C.
Another optimization was achieved during the formation of the central ketal motif of 186 (Scheme 22). Here, the diastereomeric mixture of 183 (see Scheme 19) was exposed to light-mediated DDQ-oxidation within a continuous flow setup, which yielded epimers (R)-184 and (S)-184. This mixture was further treated with p-TsOH·H_2_O leading to cyclized 185 from (R)-184 and leftover (S)-184. While 185 was mesylated within the last step towards 186, (S)-184 was epimerized upon addition of base, and another fraction of (R)-184 was recovered and also taken for the assembly of 186.
Modified synthesis of 186. Reaction conditions: (a) DDQ, 2,6-di-t-Bu-4-hydroxytoluene, hv, MeCN, 23 °C, continuous flow process; (b) p-TsOH·H2O, 23 °C; (c) MsCl, NEt3, DMAP, DCM, 0 °C; (d) NaOMe, MeOH, 23 °C.
With 186 and 200 in hand, Nicolaou and co-workers merged both building blocks within a Ni(II)/Cr(II)-catalyzed NHK reaction using chiral ligand 206 (Scheme 23) [95]. Upon treatment with DBU, 3-methylenetetrahydrofuran 201 was obtained in 56% yield over two steps. PMB-deprotection, iodination and selective reduction of the ester moiety with DIBAL-H furnished 203. The cyclization procedure towards 204 was based on a CoPc/CrCl_2_ coupling protocol by Takai and co-workers, and was optimized to a yield of 67% [96]. DMP-oxidation and acidic cleavage of the acetonide moiety yielded 205 in 81%. Eventual tosylation of the primary alcohol with Ts_2_O and following amination yielded 1. In comparison to their former route [93], here, the bicyclic motif of 181 was not assembled via tin-mediated radical cyclization reaction, but with Ni-catalyzed cyclization (see Scheme 18 and Scheme 20). This innovation led to a remarkable increasement of overall yield (17.6% vs 9.3%). Also, the ketalization from 183 to 184 was enhanced. Here, the application of a continuous flow process and the recovery of (R)-184 led to an increasement of yield in 6.7% in comparison to the former approach. It is also important to mention that Nicolaou and co-workers paved the way for further derivatives of 1 with this synthesis.
Synthesis of 1. Reaction conditions: (a) i. CrCl2, NiCl2, 206, NEt3, THF, 23 °C; ii. DBU, toluene, 110 °C; (b) DDQ, DCM, buffer (pH 7), 23 °C; (c) i. imidazole, PPh3, I2, THF, 0–23 °C; ii. DIBAL-H, DCM, −78 °C; (d) CoPc (see Scheme 10), KI, CrCl2, DMF, 23 °C; (e) i. DMP, DCM, 23 °C; ii. AcOH·H2O, 50 °C; (f) i. Ts2O, 2,4,6-collidine, pyridine, DCM, −10–0 °C; ii. NH4OH, iPrOH, 0–30 °C.
A large-scale synthesis of 217 was achieved by Kim and co-workers using a Rh(II)-induced carbene formation and subsequent [2,3]-Wittig rearrangement [97] (Scheme 24). Herein, alcohol 207 was TBDPS-protected, before epoxidation yielded 209. Hydrolytic kinetic resolution of 209 afforded enantioenriched epoxide (S)-209, which was opened regioselectively with lithium acetylide. The allylation and [3 + 2]-cycloaddition with tosyl azide led to 212. Treatment with Rh_2_(OAc)4 induced a [2,3]-Wittig rearrangement and subsequent hydrolysis afforded 213 in 71% yield. A detailed mechanism for this transformation is shown in Scheme 24 below [98]. Here, the Rh(II)-induced elimination of N_2_ generates carbene 218, which is attacked by the adjacent allyl ether moiety to form unstable zwitterion 219. Sigmatropic rearrangement and acidic workup yielded 213 stereoselectively. Further methylenation of the keto group with MePh_3_PBr and exchange of the alcohol protecting group were used to form 215. Finally, regioselective hydroboration–oxidation and DMP-oxidation afforded aldehyde 217. Notably, all reactions were performed in a (multi-)kg scale in consistently high yields, with exception of step (c), showing the applicability of this sequence for production.
Synthesis of 217. Above: Reaction conditions: (a) TBDPSCl, imidazole, DCM, 0–5 °C. (b) m-CPBA, DCM, 0 °C; (c) Jacobsen cat. (0.5 mol %), H2O, rt; (d) ethylenediamine, lithium acetylide, DMSO, 10–15 °C; (e) NaHMDS, allyl bromide, TBAI, DMF, 0 °C; (f) n-BuLi, TsN3, THF, −60 °C; (g) i. Rh2(OAc)4 (5 mol %), toluene, 60–90 °C; ii. acidic work-up; (h) MePh3PBr, t-BuOK, THF; (i) i. TBAF, THF, rt; ii. PivCl, DMAP, DCM, rt; (j) BH3·THF, 2,3-dimethyl-2-butene, cyclopentane, −20 °C; (k) DMP, NaHCO3, DCM, rt. Below: Mechanism of the [2,3]-Wittig rearrangement in step (g).
As a cost-efficient approach towards fragment 231, Yu and co-workers applied ʟ-arabinose (220) as a starting material (Scheme 25) [99]. Herein, 220 was first treated with AcCl in MeOH, then protected with TrCl to yield 221 in 68%. The diol was converted stereospecifically to epoxide 222 under Mitsunobu conditions, then reopened with LiAlH_4_ and protected with BzCl. Deprotection of the primary alcohol was necessary for the consecutive DMP-oxidation towards an intermediate aldehyde and HWE-reaction to assemble 226 as an E/Z mixture. Hydrogenation of alkene 226 and treatment with allyltrimethylsilane were used to form 228. The ester moiety of 228 was reduced and the Bz-group was cleaved simultaneously to afford the respective diol (229), then TIPS-protection of the primary alcohol unit enabled the oxidation towards an intermediate ketone, which was further methylenated with MePPh_3_Br. Eventually, the regioselective hydroboration–oxidation and DMP-oxidation led to aldehyde 231. Herein, both stereocenters of 231 were adopted from the naturally given structure of 220 and were maintained over 14 steps with a remarkable total yield of 24%.
Synthesis of 231. Reaction conditions: (a) i. AcCl, MeOH, 0 °C to rt; ii. TrCl, pyridine, 50 °C; (b) PPh3, DIAD, THF, rt; (c) LiAlH4, THF, relfux; (d) BzCl, pyridine, 0 °C to rt; (e) 60% AcOH (aq), 50 °C; (f) i. DMP, DCM, rt; ii. Ph3PCHCO2Me, toluene, rt; (g) Pd/C, MeOH, H2, rt; (h) allyltrimethylsilane, BF3·Et2O, DCM, 0 °C to rt; (i) LiAlH4, THF, rt; (j) TIPSCl, imidazole, DMF, rt; (k) i. DMP, DCM, rt; ii. MePPh3Br, t-BuOH, THF, rt; iii. Sia2BH, THF, −20 °C; iv. DMP, NaHCO3, DCM, rt; TrCl: triphenylmethyl chloride.
In 2023, Kaghad and co-workers designed a novel route towards C14–C35 fragment 255 involving three stereoselective α-chlorinations of aldehydes as the key steps (Schemes 26–28) [100]. From tartrate salt 232 (obtained from tetrahydrofurfurylamine and ᴅ-(−)-tartaric acid), Boc-protection and Ru(III)-induced oxidation led to lactone 233 (Scheme 26). Treatment with NaBH_4_ triggered the cleavage into an intermediate diol, which was cyclized upon addition of 2,2-dimethoxypropane and oxidized towards aldehyde 234. The first α-chlorination yielded a racemic mixture ((R)-235 and (S)-235), which epimerized upon the addition of ᴅ-proline. The subsequent stereoselective aldol reaction with tetrahydrothiopyran-4-one went significantly slower than the epimerization and was directed as well by the amino acid. Therefore, 236 was produced diastereoselectively via dynamic kinetic resolution favoring the (S)-enantiomer of 235. Next, the reduction of the keto moiety of 236 induced a stereospecific cyclization towards 237, whose stereocenter of the alcohol moiety was inverted under Mitsunobu conditions. Methylation of the newly arranged alcohol group and addition of p-Tol_2_IOTf led to the formation of salt 239.
Synthesis of 239. Reaction conditions: (a) i. Boc2O, K2CO3, THF, rt; ii. Ru(acac)3, NaBrO3, EtOAc, H2O, rt; (b) i. NaBH4, MeOH, THF, rt to 60 °C; ii. 2,2-dimethoxypropane, p-TsOH, acetone, rt; iii. PCC, celite, DCM, rt; (c) i. ᴅ-proline, NCS, DCM, 0 °C; ii. tetrahydrothiopyran-4-one, DMSO; (d) i. DIBAL-H, THF, −78 °C; ii. SrCO3, H2O, MeOH, 75 °C; (e) i. p-nitrobenzoic acid, DIAD, PPh3, THF, 0 °C to rt; ii. NaOH, MeOH, rt; (f) i. NaOt-Bu, MeI, THF, 0 °C to rt; ii. p-Tol2IOTf, Cu(OBz)2, DCE, 110 °C.
Additionally, building block 247 was synthesized from commercial 240 (Scheme 27). Again, stereoselective chlorination using NCS and MacMillan catalyst 248 was perfomed to yield 241 [101]. Subsequently, ketone 249 was added to aldehyde 241 within a stereoselective aldol reaction. Reduction with DIBAL-H and addition of AgOTf triggered the cyclization of 242 and the untouched secondary alcohol moiety was TBDMS-protected. 243 underwent dihydroxylation and subsequent oxidative cleavage reaction to furnish 244, which served as a substrate for the third α-chlorination to afford 245. HWE-reaction with 250 towards 246 and reduction with PhSiH_3_ led to 247 as a separable mixture of epimers (247:*epi-*247 = 1.5:1). Notably, by the treatment of *epi-*247 with DBU, again an epimeric mixture was formed (247:*epi-*247 = 1:1) and another fraction of 247 was isolated.
Synthesis of 247. Reaction conditions: (a) NCS, 248, MeCN, 0 °C to rt; (b) LDA, 249, THF, −78 °C; (c) i. DIBAL-H, THF, −78 °C; ii. AgOTf, Ag2O, THF, 0 °C to rt; iii. TBDMSOTf, NEt3, DCM, 0 °C; (d) OsO4, NaIO4, THF, H2O, rt; (e) NCS, ent-248, MeCN, 0 °C to rt; (f) 250, Ba(OH)2, THF, rt; (g) i. Mn(dpm)3, PhSiH3, iPrOH, rt; ii. DBU, MeCN, 5 °C; during step g (i) also the epimer of 247 (epi-247) was formed (247:epi-247=1.5:1), clean 247 was obtained fractionwise via chromatographical separation of the mixture (247+epi-247) after both reaction steps.
The products from Scheme 26 and Scheme 27 were coupled via Corey–Chaykovsky reaction and the so-obtained epoxide was treated with m-CPBA yielding sulfone 251 as the major diastereomer (Scheme 28). Regioselective deprotonation and epoxide opening with in situ-formed Ti(III)-species from Cp_2_TiCl_2_ and zinc led to the formation of 252. Next, conditions reported by Kishi and co-workers were applied for the nucleophilic substitution of the chloride by the oxygen from the tetrahydrofuran ring towards 253 [94], before an adjacent secondary alcohol group reopened the bicyclic cationic intermediate. Cleavage of the TBDMS-protecting group and DMP-oxidation afforded 254, which was finally olefinated forming 255. In this route especially the application of stereoselective α-chlorinations was well demonstrated. Also, the dynamic kinetic resolution towards 236 ressembles a key step of this strategy. With these innovations, 1 can be synthesized formally in 52 steps, which therefore strongly competes with the route from Eisai (67 steps) [66]. Incidentally, by the exchange of starting material 232 (Scheme 26) with ʟ-glutamic acid, Eisai intermediate 12 and an analogous Kishi intermediate were synthesized [19,68].
Synthesis of 255. Reaction conditions: (a) i. LiHMDS, THF, −78 °C to rt; ii. m-CPBA, DCM, −78 °C to rt; (b) Cp2TiCl2, Zn, THF, rt; (c) AgBF4, 2,6-di-t-Bu-4-methylpyridine, t-BuOAc, rt; (d) i. HF·pyridine, THF, rt; ii. DMP, NaHCO3, DCM, rt; (e) MePPh3Br, t-BuOK, THF, 0 °C to rt.
Shortly thereafter, Nasam and co-workers developed a novel synthesis of 1`s fragment 272 using multiple asymmetric catalytic transformations (Schemes 29–31) [102]. 256 was synthesized from trans-crotonic acid and served as a starting material for the first building block (261, Scheme 29). Hosomi–Sakurai allylation of 256 was used to afford 257 [103], which was subsequently treated with LiBH_4_ to yield alcohol 258. TBDPS-protection and ozonolysis led to aldehyde 259. The stereoselective allylation was conducted according to a procedure from Maruoka and Kii, and yielded 260 in 9:1 dr [104]. After protection of the secondary alcohol, hydroboration–oxidation of the terminal alkene and treatment with DMP produced aldehyde 261.
Synthesis of 261. Reaction conditions: (a) allyltrimethylsilane, TiCl4, DCM −78 °C; (b) LiBH4, EtOH, Et2O, 0 °C; (c) i. TBDPSCl, imidazole, DMAP, THF, 0 °C; ii. O3, PPh3, DCM, −78 °C; (d) allyltributylstannane, TiCl4, Ti(OiPr)4, Ag2O, (R)-BINOL, 0 °C; (e) i. TBDMSOTf, 2,6-lutidine, DCM, 0 °C; ii. BH3·SMe2, 30% H2O2, NaOH (aq.), THF; iii. DMP, NaHCO3, DCM, 0 °C to rt.
The second building block (265) was synthesized via stereoselective Noyori reduction of 262 [105], followed by TMS cleavage and Bz-protection of the free alcohol unit (Scheme 30).
Synthesis of 265. Reaction conditions: (a) (R,R)-Ru-cat (0.2 mol %), DCM, NEt3, HCOOH, rt; (b) TBAF, THF, 0 °C; (c) BzCl, DMAP, NEt3, 0 °C to rt.
Both products were fused together by the nucleophilic attack of deprotonated alkyne 265 to aldehyde 261 (Scheme 31). The obtained diastereomers 266 were oxidized to the respective ketone 267, and again, Noyori reduction was performed to access 268 [105]. Treatment with AgBF_4_ at elevated temperature induced a rearrangement via allene 269 and cyclization towards 2,5-dihydrofuran 270. Cleavage of the Bz-moiety and tautomerization to ketone 271 enabled the unmasking of the methylene motif of 272 via Wittig reaction. In this convergent synthesis, a remarkable overall yield of 17.6% for 272 was obtained, which makes it a competitive reaction sequence for the large-scale process currently applied by Eisai [66]. Furthermore, this method stands out for its cost-efficiency, since only an enantiomeric pair of ruthenium catalysts is needed in addition to other standard chemicals.
Synthesis of 272. Reaction conditions: (a) LDA, THF, −78 °C; (b) DMP, NaHCO3, DCM, 0 °C to rt; (c) (S,S)-Ru-cat (0.2 mol %), DCM, NEt3, HCOOH, rt; (d) AgBF4, cyclohexane, 80 °C; (e) MeLi·LiBr, Et2O, −78 °C; (f) MePh3PBr, n-BuLi, 0 °C.
A similar building block as from Kaghad and co-workers (see Scheme 32) was assembled in a kg-scale by Kim and co-workers intercepting the total synthesis of 1 from Eisai (Schemes 32–34) [100,106]. The synthetic route design could be guided by the patent landscape. This paper is the first to demonstrate that the complex web of legal exclusivities held by an originator company should be a primary consideration when developing a synthetic route for the earliest possible market entry by generic API manufacturer. This strategic approach directly led to the design of non-infringing intermediate 292 for the key NHK coupling with fragment 296 (see Scheme 34). For the assembly of 292, commercial xylose acetonide (273) was Ts-protected, before treatment with K_2_CO_3_ led to formation of oxetane 275 (Scheme 32). Allylation, methylation of the alcohol moiety and dihydroxylation with subsequent Bz-diprotection yielded 279. After another stereospecific allylation, 280 was isolated as a single diastereomer upon recrystallization. The dihydroxylation, oxidative cleavage and reduction with NaBH_4_ afforded diol 281, whose primary alcohol unit was protected enabling the selective oxidation towards ketone 283. Olefination with 293 (Nysted reagent) and stereoselective hydroboration with texylborane and oxidation yielded primary alcohol 285. Another DMP-oxidation, isomerization under basic conditions, reduction and Ms-protection were conducted to assemble 289. Substitution of the mesylate with PhSNa, replacement of both Bz- with TBDMS-protecting groups and eventual deprotection of the Piv-moiety, followed by DMP-oxidation led to the formation of the target fragment 292.
Synthesis of 292. Reaction conditions: (a) TsCl, NEt3, DCM, rt; (b) K2CO3, MeOH, 45 °C; (c) vinylmagnesium bromide, THF, reflux; (d) NaH, MeI, TBAI, THF, rt; (e) (DHQ)2AQN, K2OsO4·2H2O, K3Fe(CN)6, K2CO3, t-BuOH, H2O, 0 °C; (f) N-methylmorpholine, BzCl, DMAP, toluene, 75 °C; (g) allylTMS, Ti(OiPr)4, TiCl4, toluene, rt; (h) i. K2OsO4·2H2O, rt; ii. NaIO4, NaHCO3, DCM, rt; iii. NaBH4, MeOH, rt; (i) PivCl, pyridine, DCM, rt; (j) DMP, NaHCO3, DCM, rt; (k) 293, TiCl4, THF, 5 °C to rt; (l) i. texylborane, THF, −20 °C; ii. NaBO3·4H2O, THF, H2O, rt; (m) DMP, NaHCO3, rt; (n) NEt3, DCM, rt; (o) NaBH4, EtOH, 0 °C; (p) Ms2O, NEt3, DCM, 0 °C; (q) PhSNa, THF, 0 °C; (r) i. Mg(OMe)2, MeOH, rt; ii. TBDMSCl, imidazole, DCM, rt; (s) i. DIBAL-H, THF, 0 °C; ii. DMP, NaHCO3, rt; (DHQ)2AQN: hydrochinin-(anthrachinon-1,4-diyl) diether.
In parallel, 296 was synthesized via Cr(III)-catalyzed coupling of aldehyde 294 with diiodide 171 (see Scheme 17) and subsequent mesylation (Scheme 33).
Synthesis of 296. Reaction conditions: (a) 171 (see Scheme 17), Cr-cat, CoPc (see Scheme 17), Mn, NEt3·HCl, LiCl, TMSCl, DME, rt; (b) Ms2O, NEt3, DCM, −5–0 °C.
Stereoselective Cr(II)-catalyzed coupling between aldehyde 292 and iodide 296 with ligand 172 (see Scheme 17) was used for the assembly of 297 (Scheme 34). The addition of KHMDS induced the cyclization towards tetrahydropyran 298. Finally, oxidation of the sulfide moiety afforded sulfone 299. Using this synthetic strategy, Kim and co-workers assembled 276 via simple ring opening of 275`s oxetane moiety and no cooling and intermediate purification were necessary. These practical improvements are particularly important for the large-scale synthesis of 1. Furthermore, every step up to 292 was carried out in kg amounts using only operationally simple standard procedures.
Synthesis of 299. Reaction conditions: (a) 172 (see Scheme 17), CrCl2, NEt3, NiCl2, THF, rt; (b) KHMDS, THF, −20 °C; (c) (NH4)2MoO4, urea·H2O2, EtOH, rt.
Employing their recent protocol on allene-Prins reactions (see Scheme 6), Kim and co-workers developed a completed route towards 1, which included no chromium reagents (Scheme 35 and Scheme 36) [107]. Therefore, 300 was synthesized from ᴅ-quinic acid according to a known procedure [79]. Acidic treatment in MeOH and subsequent addition of MeLi furnished full acetal 301 (Scheme 35). Treatment with Tf_2_NPh yielded an intermediate enol triflate, which underwent a β-hydride elimination towards allene 302 upon addition of Pd_2_(dba)3. Next, ring opening of the full acetal to obtain the respective aldehyde and alcohol motifs were accomplished by acidic treatment, then HWE reaction and 1,4-reduction with LiBH_4_ afforded sulfone 303. The iodination and addition of Zn in AcOH triggered a Bernet–Vasella-type fragmentation (indicated by arrows) towards 304 [108]. Finally, Mitsunobu inversion of the secondary alcohol and TES-protection yielded allene 305 in 48%.
Synthesis of 305. Reaction conditions: (a) i. p-TsOH, MeOH, 40 °C; ii. MeLi, LiBr, THF, −25 °C; (b) i. Tf2NPh, NaHMDS, THF, −78 °C; ii. Pd2(dba)3, (S)-MOP, iPr2NEt, heptane, 55 °C; (c) i. HCl (aq), THF, rt; ii. PhSO2CH2PO(OEt)2, LiCl, iPr2NEt, MeCN, 0 °C to rt; iii. LiBH4, THF, 0 °C; (d) i. Tf2O, iPr2NEt, NaI, DMF, −4 °C to rt; ii. Zn, AcOH, THF, 0 °C; (e) i. p-nitrobenzoic acid, PPh3, DIAD, toluene, 0 °C; ii. LiOH·H2O, THF, H2O, MeOH, rt; iii. TESCl, imidazole, DCM, 0 °C to rt; (S)-MOP: (S)-2-diphenyphosphino-2'-methoxyl-1,1'-binaphthyl.
The assembly of all building blocks is shown in Scheme 36. 306 was derived from the TBDPS-protection and DIBAL-H reduction of 92 (see Scheme 9). Deprotonated 41 (see Scheme 6) was added to 306s aldehyde motif, before DMP-oxidation yielded ketone **307**. SmI_2_-induced defunctionalization of the sulfone unit afforded **308** in 88%. Afterwards, the Bn-moiety was cleaved off and the compound was oxidized to the respective aldehyde **309**, which enabled the reaction with building block **305** towards **310** via interrupted Julia coupling. The Prins-cyclization via **311** proceeded upon addition of BF_3_·OEt_2_ and methoxyacetic acid. From **312**, Tsuji-reduction and, again, defunctionalization with SmI_2_ led to the formation of **314**. Desilylation followed by ketalization under acidic conditions furnished diol **316**. The final transformation towards **1** is known from previous works and can be accomplished by regioselective tosylation of **316**s primary alcohol unit and subsequent amination [66]. Notably, in this sequence, 3 fragments are synthesized individually (41, 305 and 306) and finally merged in a short sequence (306→310) to the target macrocycle. Hence, this pathway enables the quick assembly of multiple derivatives. The late-stage allene-Prins reaction also facilitates the application of other substituents on the central 3-methylenetetrahydropyran unit. On top of that, the commonly used NHK reaction involving a Cr(II)-catalyst is bypassed [66], which makes the pathway cheaper and eco-friendlier.
Synthesis of 1. Reaction conditions: (a) i. 41 (see Scheme 6), LDA, THF, −78 °C; ii. DMP, NaHCO3, DCM, rt; (b) SmI2, THF, MeOH, −78 °C; (c) i, Pd/C, H2, EtOAc, MeOH, rt; ii. DMP, NaHCO3, DCM, rt; (d) i. 305, n-BuLi, t-BuOK, THF, −78 °C; ii. DMP, NaHCO3, pyridine, DCM, rt; (e) methoxyacetic acid, BF3·OEt2, DCM, –25 to −10 °C; (f) Pd(PPh3)4, PPh3, HCOOH, NEt3, THF, 60–65 °C; (g) SmI2, THF, MeOH, −78 °C; (h) i. TBAF, imidazole·HCl, N,N-dimethylazetimide, THF, rt; ii. pyridinium p-toluenesulfonate, DCM, rt; (i) Mg(OMe)2, MeOH, rt.
Very recently, Khatravath and co-workers used ᴅ-isoascorbic acid (317) as a starting material for the synthesis of tetrahydropyran fragment 324 (Scheme 37) [109]. First, 317 was protected as acetonide, before oxidative cleavage and esterification with EtI were conducted to form 318. TBDPS-protection and reduction yielded alcohol 319 in 58%, which was further oxidized to the respective aldehyde and olefinated leading to a 4:1 (Z/E) mixture of 320. The isomers were separated via chromatography, then, (Z)-320 was deprotected under acidic conditions, which led to a cyclization, and benzylated with BnBr towards 321. Stereoselective Michael addition of a Gilman reagent yielded 322. Reduction of the lactone unit yielded a lactol, which is in equilibrium with acyclic aldehyde and alcohol. The aldehyde was trapped via HWE reaction and the obtained α,β-unsaturated carbonyl was reattacked by the alcohol (oxy-Michael reaction, see reaction sequence in Scheme 10) leading to a diastereomeric mixture of 323 (4:1 dr). 323 was deprotected, oxidized and methylenated to afford 324 in 28% yield. Overall, 324 was obtained in a total yield of 3% over a sequence of 12 steps, which only included the use of cheap materials and preparatively simple transformations. Therefore, if the overall yield can be increased even further, this synthesis represents a useful alternative to existing pathways.
Synthesis of 324. Reaction conditions: (a) i. acetone, CuSO4, rt; ii. H2O2 (30%), K2CO3, rt; iii. EtI, MeCN, 70 °C; (b) i. TBDPSCl, DCM, rt; ii. LiBH4, THF, rt; (c) i. 2-iodoxybenzoic acid, MeCN, 90 °C; ii. NaH, (PhO)2POCH2COOEt, THF, −78 °C; (d) i. p-TsOH·H2O, benzene, rt; ii. Ag2O, BnBr, toluene, rt; (e) Me2CuLi, TMSCl, Et2O, −20 °C; (f) i. DIBAL-H, DCM, −78 °C; ii. triethyl phosphonoacetate, NaH, THF, rt; (g) i. TBAF, THF, rt; ii. DMP, DCM, rt; iii. methyltriphenylphosphonium bromide, n-BuLi, THF, 0 °C.
Conclusion
In summary, numerous advancements were made in the field of the synthesis of 1. This is especially observable for the synthesis of the 4 heterocyclic key fragments or intermediates obtained during the Eisai process. Here, the major progresses can be registered in terms of enhancing the sustainability of pathways through the removal of metal catalysts, use of cheap and readily available substrates and reagents, application of mild reaction conditions, and improving step economy and scalability of the process. Moreover, 3 new completed total syntheses were developed by Lee, Nicolaou and Kim, which show improvements in terms of overall yield or enable the targeted derivatization of 1 for future works [110]. An overview of methods, including the respective fragments, steps and scales, is shown in Table 1. Given the importance of 1 for medicine and the reasearch interest in 1`s derivatives to potentially enhance its potency, this research field will undoubtably continue to grow in the following years.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Shigemori H Bae M A Yazawa K Sasaki T Kobayashi J J Org Chem 1992574317432010.1021/jo 00041 a 053 · doi ↗
- 2Ohno O Chiba T Todoroki S Yoshimura H Maru N Maekawa K Imagawa H Yamada K Wakamiya A Suenaga K Chem Commun 2011471245310.1039/c 1cc 15557 a 22042320 · doi ↗ · pubmed ↗
- 3Ledford H Nature 201046860860910.1038/468608 a 21124423 · doi ↗ · pubmed ↗
- 4Kikuchi S Okada K Cho Y Yoshida S Kwon E Yotsu-Yamashita M Konoki K Tetrahedron 2018743742374710.1016/j.tet.2018.05.049 · doi ↗
- 5Jackson K L Henderson J A Phillips A J Chem Rev 20091093044307910.1021/cr 900016 w 19534494 · doi ↗ · pubmed ↗
- 6Hirata Y Uemura D Pure Appl Chem 19865870171010.1351/pac 198658050701 · doi ↗
- 7Gitlitz B J Davies A M Belani C P Argiris A Ramalingam S S Hoffman P C Koczwas M Groshen S G Gandara D R J Clin Oncol 200927805610.1200/jco.2009.27.15_suppl.8056 · doi ↗
- 8Dabydeen D A Burnett J C Bai R Verdier-Pinard P Hickford S J H Pettit G R Blunt J W Munro M H G Gussio R Hamel E Mol Pharmacol 2006701866187510.1124/mol.106.02664116940412 · doi ↗ · pubmed ↗