Keywords

15.1 Introduction

As exemplified by halichondrins and bryostatins, marine macrolide natural products are an important source of chemotherapeutic lead compounds for human diseases [1,2,3,4,5]. However, most of, if not all, this class of natural products are only scarcely isolable from natural sources. The structural complexity of marine macrolides, commonly characterized by a macrolactone skeleton with multiple stereogenic centers, also hampers selective derivatizations for analogue synthesis. Accordingly, total synthesis is currently the most practical way to access marine macrolides for detailed investigations into their chemical reactivity and biological activity [6,7,8,9]. It should also be emphasized that total synthesis plays an indispensable role in the structure determination of marine macrolides, wherein NMR spectroscopic analysis sometimes results in incorrect configurational assignment of stereogenic centers [10].

The synthetic challenges in total synthesis of marine macrolide natural products basically include the construction of the stereochemically complex carbon chain and the closure of the macrocyclic backbone [6,7,8,9]. Owing to significant advances in synthetic organic chemistry over the past half century, nowadays a body of versatile methods for the synthesis of chiral building blocks is available, and a repertoire of powerful macrocyclization reactions is now in our hand. However, efficiency in total synthesis of marine macrolides still remains unsatisfactory; it is not uncommon to find cases where 25 steps or more are required to complete a total synthesis [11]. Thus, it seems necessary to formulate a new way of thinking in total synthesis of marine macrolides.

Motivated by the structural complexity and medicinal importance, our group has initiated synthetic campaigns toward anticancer marine macrolides more than 15 years ago (Fig. 15.1) [12,13,14,15,16,17,18,19,20,21,22,23]. Our first target was (+)-neopeltolide, which was originally isolated by Wright et al. from a Jamaican sponge of the family Neopeltidae [24].

Fig. 15.1
Chemical structures of neopeltolide, exiguolide, lyngbyaloside B, aspergillide A, aspergillide B, iriometeolide 2 a, enigmazole A, and enigmazole B.

Structures of marine macrolide natural products

Full size image

On the basis of detailed 2D-NMR experiments, the planar structure and relative stereochemistry of (+)-neopeltolide were initially determined to be that shown by structure 10 (Fig. 15.2). However, the structure of (+)-neopeltolide reported by Wright et al. was unfortunately misassigned and later corrected to be structure 1 by Panek [25] and Scheidt [26] through total synthesis. Notably, the structure of (+)-neopeltolide is closely similar to that of (+)-leucascandrolide A (11), a marine macrolide that was previously identified by Pietra et al. from the sponge Leucascandra caveolata collected in New Caledonia [27].

Fig. 15.2
Three chemical structures are as follows. Proposed structure of neopeltolide, correct structure of neopeltolide, and leucascandrolide A.

Structures of (+)-neopeltolide and (+)-leucascandrolide A

Full size image

(+)-Neopeltolide exhibited single-digit nanomolar in vitro antiproliferative activity in cancer cells. In addition, (+)-neopeltolide showed potent growth inhibition against fungal pathogen Candida albicans (MIC = 0.625 μg/mL in liquid culture). These biological activities of (+)-neopeltolide were quite similar to those of (+)-leucascandrolide A reported by Pietra [27]. Later, the Kozmin group revealed that neopeltolide and leucascandrolide A inhibited the complex III of the electron transport chain of the mitochondria to exert their potent biological activities [28]. We reported that 8,9-dehydroneopeltolide showed potent cytotoxic activity in cancer cells under energy stressed conditions [29, 30]. We also described the synthesis of fluorescent derivatives of 8,9-dehydroneopeltolide to demonstrate its rapid accumulation in the mitochondria and the endoplasmic reticulum in live cells [31].

Because of the unique structural and biological aspects, (+)-neopeltolide represents an intriguing target for synthetic organic chemists [32,33,34,35]. Since the first total synthesis of 1 [25], a number of research groups have demonstrated the total and formal synthesis of (+)-neopeltolide. Furthermore, synthesis-driven structure–activity relationship investigations of this natural product have been described by several groups.

This chapter will describe the first-, second-, and third-generation total synthesis of (+)-neopeltolide, achieved by our group. The step economy [36] of each synthesis will be analyzed to illuminate how we strived for achieving synthetic efficiency.

15.2 The First-Generation Synthesis of (+)-Neopeltolide: The Suzuki–Miyaura Coupling Approach (2008)

In 2007, we initiated our synthetic studies toward the proposed structure 10 of (+)-neopeltolide. Inspired by our past experience in the total synthesis of polycyclic ethers by capitalizing on Suzuki–Miyaura cross-coupling [37] of enol phosphates [38], it was envisioned that the tetrahydropyran ring of 10 could be accessible by means of a cross-coupling of alkylborate 13 and enol phosphate 14 followed by a ring-closing metathesis [39] (Fig. 15.3). The 14-membered macrocyclic framework of 10 was to be forged via a Yamaguchi macrolactonization [40] of seco acid 12.

Fig. 15.3
A series of chemical structures and reactions related to the proposed structure of neopeltolide involving compounds 12, 13 and 14.

First-generation synthetic blueprint toward the proposed structure 10 of (+)-neopeltolide

Full size image

We synthesized alkyl iodide 15 as the immediate precursor of alkylborate 13, as summarized in Fig. 15.4. Aldehyde 16, available in five steps from (R)-Roche ester, was allylborated with (−)-Ipc2Ballyl [41] to yield homoallylic alcohol 17 (94%). After O-methylation (MeOTf, 2,6-DTBP, 90%), ozonolysis of the double bond delivered aldehyde 18 quantitatively. Allylboration of 18 with (+)-Ipc2Ballyl afforded homoallylic alcohol 19 in 90% yield. Hydrogenation, PMB protection, and desilylation gave alcohol 20 (66%, three steps), which was iodinated to provide alkyl iodide 15 (quantitative).

Fig. 15.4
A detailed chemical synthesis diagram for alkyl iodide 15 through compounds 16 to 19. It features labeled chemical structures and reactions, with step-by-step process involved in the synthesis.

Synthesis of alkyl iodide 15

Full size image

The synthesis of enol phosphate 14, shown in Fig. 15.5, started with Keck asymmetric allylation (Ti(Oi-Pr)4, (R)-BINOL, allylSnBu3, 4 Å MS) [42] of aldehyde 21 to give homoallylic alcohol 22 (93%, > 95% e.e.). PMB protection of 22 (77%) and subsequent cross-metathesis (G-II, methyl acrylate) led to α,β-unsaturated ester 23 (89%, E/Z > 20:1), which was reduced with DIBALH to provide allylic alcohol 24 quantitatively. Asymmetric epoxidation under Sharpless conditions using (−)-DET [43] gave epoxy alcohol 25 (95%, >20:1 d.r.). Iodination of 25 followed by zinc reduction afforded allylic alcohol 26 (87%, two steps). Protection of 26 using BOMCl/i-Pr2NEt, cleavage of the PMB group with DDQ, and acetylation gave acetate 27 (98%, three steps). Finally, treatment of 27 with KHMDS and (PhO)2P(O)Cl [44] gave rise to enol phosphate 14. Because this compound was unstable, it was used directly in the subsequent Suzuki–Miyaura coupling without purification by silica gel chromatography.

Fig. 15.5
A synthesis reaction of enol phosphate 14 through compounds 21 to 27.

Synthesis of enol phosphate 14

Full size image

With the advanced intermediates 14 and 15 in hand, we proceeded to assemble the key tetrahydropyran ring and complete the total synthesis, as illustrated in Fig. 15.6. According to the procedure described by Marshall [45], alkyl iodide 15 (1 equiv.) was lithiated and immediately trapped with B-MeO-9-BBN, and the resultant alkylborate 13 was cross-coupled with acetate-derived enol phosphate 14 using aq. Cs2CO3 as a base and Pd(PPh3)4 complex as a catalyst to deliver enol ether 28. This was immediately subjected to ring-closing metathesis using G-II (10 mol%), leading uneventfully to dihydropyran 29 in 67% yield from 15. Hydrogenation of 29 proceeded from the β-face of the molecule to evade unfavorable steric contact with the benzyloxymethyl group and afforded tetrahydropyran 30 in 81% yield with >20:1 d.r. Desilylation of 30 (97%), a two-step oxidation of the derived alcohol, and subsequent treatment of the so obtained carboxylic acid with TMSCHN2 afforded methyl ester 31 in 91% over the three steps. Removal of the PMB group and TMSOK-mediated hydrolysis of the ester gave seco acid 12 (85%, two steps). Macrolactonization of 12 successfully closed the 14-membered macrocyclic skeleton to provide macrolactone 32 in 97% yield. Hydrogenolytic deprotection of the BOM group gave rise to alcohol 33 (quantitative). Mitsunobu esterification [46] of 33 with carboxylic acid 34 [47, 48] (DIAD, Ph3P) resulted in the proposed structure 10 of (+)-neopeltolide (86%).

Fig. 15.6
A reaction schematic to synthesize neopeltolide through compounds 15, 13, 14, 28, 29, 30, 31 and 12.

Total synthesis of the proposed structure 10 of (+)-neopeltolide

Full size image

During the course of the above investigation, the stereochemical reassignment of (+)-neopeltolide was disclosed by the Panek group [25], which prompted us to synthesize the correct structure 1 in a similar manner as that shown for 10 (Fig. 15.7).

Fig. 15.7
A reaction schematic to synthesize neopeltolide through compounds 16, 35, 36, 37, 38, 39, 14, 40, 41 and 42.

Total synthesis of (+)-neopeltolide (1)

Full size image

Alkyl iodide 38 with correct configurations at C11 and C13 was easily synthesized from aldehyde 16 by using suitable Brown’s chiral allylboration reagents. Suzuki–Miyaura coupling of alkylborate 39, prepared from 38, with enol phosphate 14 under the optimized conditions, followed by ring-closing metathesis using G-II complex, afforded dihydropyran 41 in 78% yield for the two steps. Subsequent hydrogenation of 41 delivered 2,6-cis-configured tetrahydropyran 42 (81%, >20:1 d.r.). The remainder of the synthesis proceeded in much the same way as that described for 10.

Our first-generation synthesis of (+)-neopeltolide was thus achieved in 25 linear steps from (R)-Roche ester (or 1,3-propanediol) and in 49 total steps [49, 50]. Analysis of the convergency of the present synthesis is shown in Fig. 15.8. The two advanced intermediates, alkyl iodide 38 and enol phosphate 14, were synthesized in 13 steps each. After the point of convergence [51] at the 14th step, completion of the total synthesis required 11 additional steps, seven of which were concession steps [52]. These step counts illuminate the following points: (1) the first-generation synthesis is only moderately convergent because the point of convergence was placed at the mid of the macrolactone synthesis; (2) many concession steps were required in between the tetrahydropyran construction and the macrocyclization. We considered that these inefficiencies should, at least in part, be ascribable to the anion chemistry for preparing alkylborate 39 and enol phosphate 14, where all the hydroxy groups (C1, C3, C5, C11, and C13) must be differentially protected. Extensive usage of protecting groups inevitably increases the number of concession steps. With this point in mind, we strived to develop a second-generation synthesis of (+)-neopeltolide as described in the following section.

Fig. 15.8
A detailed chemical synthesis pathway. It begins with Roche ester transforming into 1, 3 propanediol through 14 steps followed by 13 steps and 10 steps to obtain neopeltolide.

Summary of the first-generation synthesis of 1

Full size image

15.3 The Second-Generation Synthesis of (+)-Neopeltolide: The Ring-Closing Metathesis Approach (2010)

To maximize the convergency of the second-generation synthesis, we envisioned a synthetic blueprint in which the point of convergence was placed at late stage as possible (Fig. 15.9). Thus, the macrolactone skeleton of 1 would be accessed from diene 44 via a macrocyclic ring-closing metathesis, and the latter should be available from carboxylic acid 45 and alcohol 46 through a Yamaguchi esterification.

Fig. 15.9
A three-step chemical reaction in the conversion of neopeltolide and esterification to yield two distinct compounds, labeled 44 and 45.

Second-generation synthetic blueprint toward (+)-neopeltolide (1)

Full size image

First, we synthesized carboxylic acid 45 as shown in Fig. 15.10. Asymmetric aldol reaction [53] of trans-cinnamaldehyde (47) with thiazolidinethione 48 under Nagao conditions gave alcohol 49 (87%, 11:1 d.r.). The undesired minor diastereomer could be readily separable by silica gel flash column chromatography. After removal of the thiazolidinethione moiety (94%), the resultant amide 50 was reacted with allylMgCl to deliver β,γ-unsaturated ketone 51 (90%). Evans–Tishchenko reduction [54] of 51 (SmI2, EtCHO) afforded alcohol 52 quantitatively with > 20:1 d.r. Cross-metathesis of 52 with methyl acrylate catalyzed by G-II complex proceeded cleanly to afford α,β-unsaturated ester 54 without producing the corresponding ring-closing metathesis product. The remarkable chemoselectivity observed for the present cross-metathesis reaction would be ascribable to conformational locking of ruthenium alkylidene intermediate 53 by an intramolecular H-bonding [55,56,57], making the styryl group away from the reactive site. Note that exposure of relevant substrate 55 to G-II complex mainly gave ring-closing metathesis product 56 in 71% yield, along with cross-metathesis product 57 in 25% yield. Protection of the hydroxy group of 54 using BOMCl/i-Pr2NEt provided BOM ether 58 (68%, two steps from 52). Upon exposure of 58 to K2CO3 in methanol, removal of the propionyl group and concomitant intramolecular oxa-Michael addition [58,59,60] occurred to give 2,6-cis-configured tetrahydropyran 59 albeit with only moderate diastereoselectivity (cis/trans ca. 2:1). Accordingly, the diastereomer mixture was treated with DBU (toluene, 100 °C) to achieve a thermodynamic equilibration of a retro-oxa-Michael/oxa-Michael sequence, giving 2,6-cis-configured tetrahydropyran 59 (53%, > 20:1 d.r.). Hydrolysis of 59 provided carboxylic acid 45 quantitatively.

Fig. 15.10
A reaction schematic to obtain carboxylic acid 45. It begins with trans cinnamaldehyde followed by compounds 49, 50, 51, 52, 53, 54, 58 and 59. Another schematic presents the conversion of 55 to 56.

Synthesis of carboxylic acid 45

Full size image

Next, we synthesized alcohol 46 from (R)-epichlorohydrin (60) as depicted in Fig. 15.11. Nucleophilic attack of 2-lithio-1,3-dithiane to 60 gave epoxide 61 (90%). Addition of an organocuprate derived from EtMgBr/CuI to 61 afforded alcohol 62 (92%). PMB protection of 62 (92%) and hydrolytic removal of the dithiane (91%) led to aldehyde 63. Chelate-controlled addition of methallyltrimethylsilane to 63 under the influence of MgBr2•OEt2 delivered homoallylic alcohol 64 in 73% yield with 15:1 d.r. Methylation of 64 and PMB deprotection then afforded alcohol 46 (91%, two steps).

Fig. 15.11
A reaction schematic begins with R epichlorohydrin followed by compounds 61 to 64 to obtain alcohol 46.

Synthesis of alcohol 46

Full size image

Now the stage was set for assembly of the advanced intermediates (Fig. 15.12). Carboxylic acid 45 and alcohol 46 were esterified according to Yamaguchi conditions to deliver ester 44 (94%). Macrocyclic ring-closing metathesis of 44 required extensive optimization efforts owing to the moderate reactivity of the styryl group and the methallyl (2-methyl-2-propenyl) group; eventually it was found that a syringe pump addition of a solution of G-II complex in toluene over 6 h to a mixture of 44 and 1,4-benzoquinone [61] in toluene (3 mM, 100 °C) provided macrocycle 65 in 85% yield. The ring-closing metathesis of 44 could be carried out at higher concentration without significant decline of the product yield (82% at 10 mM and 75% at 40 mM). Stereoselective hydrogenation of the olefin and in situ deprotection of the BOM group by hydrogenolysis led to neopeltolide macrolactone (43) (93%, > 20:1 d.r.). Mitsunobu coupling of 43 with carboxylic acid 34 furnished (+)-neopeltolide (1) in 85% yield.

Fig. 15.12
A reaction schematic begins with the reaction of compounds 45 and 46 to yield 44 followed by compounds 65, 43 and neopeltolide.

Second-generation synthesis of 1

Full size image

Our second-generation total synthesis of (+)-neopeltolide was achieved in 13 linear steps from trans-cinnamaldehyde (47) and in 31 total steps [62, 63]. The outline of the second-generation synthesis is summarized in Fig. 15.13. The advanced intermediates, carboxylic acid 45 and alcohol 46, were prepared in nine and seven steps, respectively. The point of convergence appeared at the tenth step. After assembly of 45 and 46, the synthesis was finished in just three steps. Thus, it is clear that the convergency of the second-generation synthesis is much higher than that of the first-generation synthesis. Another key feature of the second-generation synthesis is the minimization of concession steps throughout the synthesis. The styryl group of trans-cinnamaldehyde (47) was carried through most of the synthesis and used in the macrocyclization step. The α,β-unsaturated ester group of 58, served as a Michael acceptor to forge the tetrahydropyran ring, was used for subsequent esterification without oxidation state adjustments, thereby enabling expedient access to carboxylic acid 45. The two-fold use of nucleophilic epoxide-opening reactions facilitated short synthesis of alcohol 46.

Fig. 15.13
A reaction schematic begins with trans-cinnamaldehyde transitioning into compounds 45 and 44 that further yield 46 and 43. R epichlorohydrin yields 46. Compound 43 yields 34 followed by neopeptolide.

Summary of the second-generation synthesis of 1

Full size image

Notably, our second-generation synthesis of 1 was applied to the construction of a 16-member (−)-8,9-dehydroneopeltolide stereoisomer library to investigate the structure–activity relationship in detail [63] and also to the synthesis of fluorescent-labeled analogues to examine the cellular target of 1 [31].

15.4 The Third-Generation Synthesis of (+)-Neopeltolide: The Tandem Macrocyclization/Transannular Pyran Cyclization Approach (2022)

In 2015, Hoveyda and co-workers disclosed a synthesis of (+)-neopeltolide (1), in which catalyst-controlled stereoselective olefin metathesis reactions were utilized extensively [64]. The Hoveyda synthesis of 1 proceeded in 11 linear steps (28 total steps). Motivated by this elegant work, we embarked on a third-generation synthesis of 1, which was based on the macrocyclization/transannular pyran cyclization strategy developed within our group [22, 65].

The third-generation synthesis blueprint toward 1, summarized in Fig. 15.14, featured not only an expedient access to neopeltolide macrolactone (43) on the basis of the macrocyclization/transannular pyran cyclization strategy but also a convergent synthesis of side chain carboxylic acid 34 through a two-fold application of palladium-catalyzed cross-coupling reactions. Thus, 43 would be available from propargylic alcohol 66, and the latter was traced back to carboxylic acid 67 and alcohol 68. Meanwhile, 34 was planned to be synthesized from iodooxazole 69 and alkyne 70.

Fig. 15.14
A reaction schematic begins with neopeltolide transitioning into compound 66 through macrocyclization or transannular pyran cyclization. Suzuki Miyaura reaction yields compounds 69, 70, 67 and 68.

Third-generation synthetic blueprint toward (+)-neopeltolide (1)

Full size image

The synthesis of carboxylic acid 67 is depicted in Fig. 15.15. Alcohol 71 was prepared from (R)-epichlorohydrin through a one-pot, sequential exposure to 2-lithio-1,3-dithiane and vinylMgBr/CuBr•SMe2 [66]. Benzylation of 71 (93%) followed by hydrolytic deprotection of the dithioacetal of 72 afforded aldehyde 73 (91%). Asymmetric Kiyooka aldol reaction of 73 and enol silane 74 (N-Ts-L-Val, BH3•THF) [67] provided alcohol 75 in 90% yield with 93:7 d.r. After protection of 75 (99%), saponification of the ester moiety of 76 delivered carboxylic acid 67 (88%).

Fig. 15.15
A reaction schematic. It begins with compound 71 followed by 72 to 76 to obtain carboxylic acid 67.

Synthesis of carboxylic acid 67

Full size image

Alcohol 68 was synthesized from (S)-epichlorohydrin (77) as shown in Fig. 15.16. Regioselective epoxide-opening of 77 with alkyne 78 (n-BuLi, BF3•OEt2) [68] gave chlorohydrin 79 (98%). After treatment of 79 with NaH, in situ regioselective ring-opening of the resultant epoxide with (vinyl)2Cu(CN)Li2 delivered homoallylic alcohol 80 (82%). Methylation of 80 (quantitative) and subsequent epoxidation of 81 with m-CBPA gave the corresponding epoxide in 94% yield with 54:45 d.r. Hydrolytic kinetic resolution using (R,R)-CoII-salen complex [69] afforded epoxide 83 as a single diastereomer (52%, > 95:5 d.r.), along with 1,2-diol 82 (40%, 82:18 d.r.), after separation by silica gel flash column chromatography. Regioselective ring-opening of 83 with EtMgBr/CuCN furnished alcohol 68 (89%).

Fig. 15.16
A reaction schematic. It begins with S epichlorohydrin followed by compounds 79 to 83 to obtain compound 68.

Synthesis of alcohol 68

Full size image

Carboxylic acid 34 was synthesized as illustrated in Fig. 15.17. Iodooxazole 69 was prepared from ethyl oxazole-4-carboxylate (LHMDS, 1,2-diiodoethane) in one step [70]. Oshima–Yorimitsu cross-coupling [71] of 69 with alkyne 70 proceeded cleanly by treatment of 70 with InCl3/DIBALH in the presence of Et3B/air, followed by cross-coupling of the generated alkenylindium with 69 under the catalysis of Pd(PPh3)4, giving alkenyloxazole 84 in 80% yield. Half reduction of the ester of 84 provided aldehyde 85 (74%, rsm 24%), whose methylenation under Takai conditions [72] delivered vinyl oxazole 86 (90%). Site-selective hydroboration of the vinyl group of 86, followed by cross-coupling [37] with ethyl cis-β-iodoacrylate (87) under Suzuki–Miyaura conditions, led to (Z)-α,β-unsaturated ester 88 (57%). Hydrolysis of 88 furnished carboxylic acid 34 (73%).

Fig. 15.17
A reaction schematic begins with reaction between compounds 69 and 70 to obtain compound 84 followed by compounds 85 to 88 to obtain carboxylic acid 34.

Synthesis of carboxylic acid 34

Full size image

The third-generation total synthesis of 1 was achieved as shown in Fig. 15.18. Yamaguchi esterification of carboxylic acid 67 and alcohol 68, followed by in situ treatment with acidic ethanol to remove the THP groups, gave propargylic alcohol 66 (89%). Meyer–Schuster rearrangement of 66 under Au/Mo combo catalysis (IPrAuCl, AgOTf, MoO2(acac)2, toluene) [73] generated intermediary vinyl ketone 89, whose macrocyclic ring-closing metathesis under the catalysis of Zhan-1B complex [DCE (20 mM), 40 °C] proceeded with spontaneous transannular oxa-Michael addition of the intermediary α,β-unsaturated ketone 90, giving rise to 2,6-cis-substituted tetrahydropyran 91 (70%, single stereoisomer), after separation of undesired minor diastereomers by silica gel flash column chromatography. The diastereoselectivity of the transannular reaction was determined to be approximately 91:9 through careful inspection of the reaction mixture. A separate experiment on macrocyclic ring-closing metathesis of chromatographically purified 89 under Zhan-1B catalysis (DCE, 80 °C) resulted in macrocyclic α,β-unsaturated ketone 90 (87%, sole product). While in our previous work tandem olefin cross-metathesis/intramolecular oxa-Michael addition of olefinic alcohol derivatives was catalyzed by Ru species [74, 75], it appears that the cationic Au species derived from IPrAuCl/AgOTf was responsible for the transannular oxa-Michael addition of 90 to give tetrahydropyran 91 in the present tandem reaction. Takai methylenation of 91 gave exo-olefin 92 (84%). Hydrogenation of the exo-olefin and hydrogenolytic removal of the benzyl group gave rise to neopeltolide macrolactone (43) (93%, 79:21 d.r.). Assembly of 43 with carboxylic acid 34 (DIAD, Ph3P) furnished (+)-neopeltolide (1) in 94% yield. The minor C9 diastereomer, i.e., 9-epi-neopeltolide (9-epi-1) [76], was separated by preparative reverse-phase HPLC. Notably, a single batch experiment provided 40 mg of spectroscopically pure 1 after HPLC purification.

Fig. 15.18
A reaction schematic begins with reaction between compounds 67 and 68 to obtain compound 66 followed by compounds 89, 90 and 43 to obtain neopeltolide and 9 epi neopeltolide.

Third-generation synthesis of (+)-neopeltolide (1)

Full size image

The third-generation synthesis of (+)-neopeltolide (1) proceeded in 11 linear and 23 total steps from inexpensive commercially available materials [77, 78]. The outline of our third-generation synthesis is summarized in Fig. 15.19. The advanced intermediates, carboxylic acid 67 and alcohol 68, were synthesized from (R)- and (S)-epichlorohydrin, respectively, in six steps each. After coupling of these intermediates, propargylic alcohol 66 was advanced to neopeltolide macrolactone (43) in three steps. Meanwhile, alkenyloxazole 84 was available from ethyl 4-oxazolecarboxylate in two steps, and the former was transformed into carboxylic acid 34 in four steps.

Fig. 15.19
A reaction schematic begins with R epichlorohydrin followed by compounds 67, 68, 66, 43, 34 and 84 to obtain compounds 69, 70 and neopeltolide.

Summary of the third-generation synthesis of 1

Full size image

The present synthesis represents the shortest access to 1 in terms of the longest linear sequence and the total number of steps. The third-generation synthesis has three significant features. First, the synthesis of key intermediates 67 and 68 was shortened as much as possible by exploiting regioselective epoxide ring-opening chemistry. Second, neopeltolide macrolactone (43) was constructed from propargylic alcohol 66 in an expedient manner by capitalizing on the macrocyclization/transannular pyran cyclization strategy. Third, the synthesis of side chain carboxylic acid 34 was achieved in just six steps from ethyl 4-oxazolecarboxylate by exploiting two Pd-catalyzed cross-coupling reactions. Note that, in our first- and second-generation synthesis of 1, the synthesis of 34 was built on previous works by Leighton [47] and Kozmin [48] and required 11 steps from a commercially available material.

15.5 Summary

This chapter delineated our first-, second-, and third-generation total synthesis of (+)-neopeltolide (1). The synthetic efficiency in terms of step count was improved significantly during the course of these synthetic campaigns (Table 15.1).

Table 15.1 Synthetic efficiency of our total syntheses of (+)-neopeltolide (1)
Full size table

In the first-generation synthesis, it was evident that multiple concession steps throughout the synthesis and lengthy transformations after the point of convergence made the synthesis inefficient. Such superfluous steps were avoided as much as possible upon planning the second- and third-generation synthesis. Nucleophilic epoxide-opening was the basis of short syntheses of key fragments. Macrocyclic ring-closing metathesis served as a powerful means to forge the macrocycle with high chemoselectivity, thereby minimizing extra functional group interconversions. Palladium-catalyzed cross-coupling reactions enabled an expedient access to side chain carboxylic acid fragment.

As implemented in the third-generation synthesis, tandem reactions were effective for increasing step economy. In particular, our macrocyclization/transannular pyran cyclization strategy enabled an expedient construction of the 14-membered macrocyclic skeleton and the engrafted 2,6-cis-configured tetrahydropyran ring in just one step, thereby minimizing extra transformations after the point of convergence. Recently, we have also disclosed a 13-step synthesis of (−)-exiguolide (2) by taking advantage of the macrocyclization/transannular pyran cyclization strategy [22], further underscoring the validity of our synthetic planning. Because the benefits of tandem reactions, for example, reduction of labor, time, and wastes by omitting isolation and/or purification of intermediates, are widely accepted, such reactions should be appropriately reflected to metrics of synthetic efficiency, including step count. Moreover, the development of tandem reactions provides rich opportunities for the discovery of new reactivities and methods [79].

As a beneficial consequence of improving the efficiency in total synthesis of (+)-neopeltolide, we were able to synthesize not only the natural product itself but also various structural analogues that were useful for investigating the structure–activity relationship [63] and biological functions [29,30,31]. Pursuing step economy in total synthesis of marine macrolides will contribute to future advances in the chemical biology and medicinal chemistry of this promising class of natural products.