Total Synthesis of Isodaphlongamine H by Iridium‐Catalyzed Reductive [3 + 2] Cycloaddition of N‐Hydroxylactam

Abstract

The total synthesis of isodaphlongamine H based on a lactam strategy, which enables quick access to complex cyclic amines, is described. The strategy begins with alkylation of a chiral lactam and subsequent N‐oxidation via an imino ether to afford the N‐hydroxylactam. For the key transformation to functionalize the amide carbonyl, an iridium‐catalyzed reductive [3 + 2] cycloaddition of the N‐hydroxylactam provides a tricyclic isoxazolidine in a one‐pot process. After the coupling reaction with an allylic silane fragment, the total synthesis is accomplished through intramolecular Hosomi–Sakurai allylation to construct a pentacyclic core. The deoxygenated pentacyclic intermediate shows higher cytotoxicity against HeLa and U937 cell lines than isodaphlongamine H, and might become a lead compound for further biological study.

The total synthesis of isodaphlongamine H was accomplished by a lactam strategy. This strategy started with alkylation and N‐oxidation of a readily available chiral lactam. For the key functionalization of the amide carbonyl, an iridium‐catalyzed reductive [3 + 2] cycloaddition of the N‐hydroxylactam afforded the tricyclic isoxazolidine, which was successfully transformed to isodaphlongamine H. Bn = benzyl, TBAF = tetrabutylammonium fluoride.

The Daphniphyllum alkaloids are one of the largest families belonging to the genus of Daphniphyllaceae, and consist of over 300 polyazacyclic natural products.^{[ 1} , ² , ³ , ⁴ , ^{5 ]} The large family of these alkaloids is further divided into fourteen classes. Their diverse structures have encouraged extensive investigation of their biosynthesis, total synthesis, and biological activities.^{[ 1} , ² , ³ , ⁴ , ^{5 ]} Heathcock and coworkers proposed a biosynthetic pathway and achieved a landmark total synthesis based on biosynthesis.^{[ 6 ]} In 2003, the Kobayashi group reported isolation of calyciphylline B (1) as a new class of the Daphniphyllum alkaloids from the leaves of Daphniphyllum calycnum (Figure 1).^{[ 7 ]} Structurally, it is comprised of a hexacyclic framework including a quaternary carbon center and a tertiary amine N‐oxide. After Kobayashi's report, isolations of several calyciphylline B‐type alkaloids were reported. The Yue group isolated deoxycalyciphylline B (2) from the stem of Daphniphyllum subverticillatum, and unambiguously confirmed its structure by X‐ray crystallographic analysis.^{[ 8 ]} The Yue group also reported the isolation of oldhamiphylline A (3) from the leaves of Daphyniphyllum oldhami.^{[ 9 ]} Oldhamiphylline A (3) possesses an additional β‐hydroxy group at C11 on the 5‐epi carbon framework of deoxycalyciphylline B (2). The Hao group documented isolation of daphlongamine H (4), which is the diastereomer of deoxycalyciphylline B (2) involving the C5 and C6 stereocenters.^{[ 10 ]} Following the isolation of these unique natural products, a number of synthetic chemists started their synthetic programs.^{[ 11} , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ^{19 ]} In 2016, the Hanessian group achieved the first total synthesis of isodaphlongamine H (5), which is an unnatural product, but the biogenetically related 5‐epi isomer of daphlongamine H (4).^{[ 12 ]} This compound exhibited antiproliferative effects comparable to natural deoxycalyciphylline B (2) against several human cancer cell lines such as HOP‐92 (lung), SNB‐75 (central nervous system), MDA_MB‐435 (melanoma), and UO‐31 (renal) cell lines in a panel screening of NCI. Hugelshofer and Sarpong developed a flexible synthetic route and accomplished the total synthesis of both daphlongamine H (4) and isodaphlongamine H (5).^{[ 15} , ^{16 ]} In their report, the structure of deoxyisocalyciphylline B, which was originally assigned as 6‐epi‐daphlongamine H, was revised by their synthesis, and found to be identical to the structure of daphlongamine H (4).

Figure 1.

Representative calyciphylline B‐type alkaloids.

An overarching objective of our research group is the development of a general synthetic strategy for biologically active complex alkaloids. We have been exploring a lactam strategy to gain quick access to highly functionalized cyclic amines (Scheme 1a).^{[ 20} , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ^{37 ]} First, the strategy begins with the use of easily available lactam 6. The second stage is functionalization of lactam framework 6, such as by alkylation at α‐position, to give highly substituted lactam 7. The third stage is the functionalization of the amide carbonyl group in lactam 7 to provide highly substituted cyclic amine 8. The salient feature of this strategy is the use of lactam intermediates as cyclic amine surrogates. Amine functional groups are often faced with issues involving high reactivity such as nucleophilicity and basicity. In contrast, lactams are highly stable, allowing for the use of various reaction conditions. Therefore, direct C─C and C─N bond formation of simple lactams is possible by taking advantage of amide carbonyls. The key to the success of this strategy is the functionalization of inert amide carbonyl groups, which can be achieved by an iridium‐catalyzed reductive functionalization.^{[ 23} , ²⁵ , ³⁶ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ^{76 ]}

Scheme 1.

a) Lactam strategy for synthesis of complex alkaloids. b) Synthetic plan to calyciphylline B‐type alkaloids. Boc = t‐butoxycarbonyl, Tf = trifluoromethanesulfonyl.

Our ultimate goal is the development of a general synthetic route to calyciphylline B‐type alkaloids based on the lactam strategy (Scheme 1b).^{[ 20} , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ^{37 ]} The strategy starts from easily available chiral lactam 9. Alkylation of lactam 9 with triflate 10 would provide cis‐substituted lactam 11. Subsequent N‐oxidation of the resulting lactam 11 would form chiral N‐hydroxylactam 12.^{[ 21} , ^{54 ]} For the functionalization of the amide carbonyl, we envisioned iridium‐catalyzed reductive [3 + 2] cycloaddition of N‐hydroxylactam 12.^{[ 50} , ^{52 ]} The iridium‐catalyzed reduction of 12 would form transient nitrone 13, which could undergo intramolecular [3 + 2] cycloaddition to give isoxazolidine 14. After cleavage of the N─O bond in 14, if carboxylic fragment 15 and allylic silane fragment 16 are installed, the intramolecular Hosomi–Sakurai allylation^{[ 77} , ⁷⁸ , ^{79 ]} of 17 would provide the pentacyclic intermediate, which can be converted to various calyciphylline B‐type alkaloids. As an initial communication of our synthetic program on the Daphniphyllum alkaloids, we report the total synthesis of isodaphlongamine H (5) based on the lactam strategy.

Our synthesis commenced with preparation of chiral lactam 9 in three steps from commercially available diester 18 (Scheme 2). Enantioselective hydrolysis of dimethyl ester 18 with a pig liver esterase provided carboxylic acid 19 in 97% yield (97% ee).^{[ 80} , ⁸¹ , ^{82 ]} The Curtius rearrangement of 19 with DPPA,^{[ 83} , ⁸⁴ , ^{85 ]} followed by addition of tBuOH in the presence of CuCl^{[ 86 ]} at 80 °C, gave carbamate 20, which was converted to lactam 9 ^{[ 87 ]} with Me₃Al. The next step is installation of an aliphatic substituent to chiral lactam 9. Deprotonation of 9 with KHMDS in toluene/HMPA = 2 at −78 °C and subsequent addition of freshly prepared triflate 10 gave cis‐ and trans‐11 in 69% combined yield (cis:trans = 1:14). Although the long aliphatic substituent was successfully installed by alkylation, undesired stereoselectivity was observed favoring the trans configuration. After extensive study, we found epimerization at C2 was feasible under kinetic conditions. Thus, after deprotonation of 11 (100 mg scale, cis:trans = 1:14) with KHMDS in toluene at −78 °C in the presence of iBu₃Al, dimethyl malonate was added as a proton source, giving a mixture of 11 in 83% combined yield favoring cis‐11 (cis:trans = 4.1:1).^{[ 88} , ^{89 ]} Addition of iBu₃Al was crucial probably due to the formation of the six‐membered chelated intermediate through the Boc carbamate (w/o iBu₃Al: 67%, cis:trans = 1.0:1). This chelation was also supported by the fact that the use of THF instead of toluene resulted in lower cis‐selectivity (78%, cis:trans = 2.7:1). Dimethyl malonate proved to be the best proton source. For example, use of AcOH showed poor selectivity (81%, cis:trans = 1.1:1). The developed conditions were scalable (6.9 g scale), giving cis‐11 and trans‐11 in 49% and 14% yields, respectively. Cleavage of the Boc group in cis‐11 with formic acid provided 21 in 89% yield.

Scheme 2.

Synthesis of chiral cis‐lactam 21. DPPA = diphenylphosphoryl azide, HMDS = hexamethyldisilazide, HMPA = hexamethylphosphoric triamide.

With the cis‐configuration successfully established, we turned our attention to N‐oxidation of chiral secondary lactam 21 (Scheme 3). Compared with oxidation of amino groups, secondary amides are more challenging substrates due to their poor electron density. However, the two‐step method via the imino ether proved to be effective.^{[ 21} , ^{54 ]} First, secondary amide 21 was converted to imino ether 22 with the Meerwein reagent.^{[ 90} , ⁹¹ , ⁹² , ^{93 ]} Treatment of 22 with mCPBA and NaOAc generated the unstable oxaziridine, which underwent ring opening with oxalic acid in MeOH/H₂O to give methyl ester 23. Addition of K₂CO₃ to 23 in a one‐pot process caused recyclization, providing N‐hydroxylactam 12 in 57% yield.

Scheme 3.

N‐Oxidation of secondary lactam 21 and Iridium‐catalyzed reductive [3 + 2] cycloaddition of N‐hydroxylactam 12 via nitrone 13. Ac = acetyl, mCPBA = meta chloroperbenzoic acid.

The stage was set for the functionalization of the amide carbonyl by iridium‐catalyzed reductive [3 + 2] cycloaddition (Scheme 3).^{[ 50} , ^{52 ]} Treatment of N‐hydroxylactam 12 with 0.5 mol % of IrCl(CO)(PPh₃)₂ and (Me₂HSi)₂O^{[ 94 ]} initiated both dehydrosilylation of the N‐hydroxy group and hydrosilylation of the amide carbonyl group, generating N,O‐acetal 24. Subsequent addition of benzyl alcohol, TBAF, and Et₃N produced nitrone 13 through desilylation and elimination of the N,O‐acetal. The resulting solution was then heated to 90 °C to promote intramolecular [3 + 2] cycloaddition, affording isoxazolidine 14 in 91% yield as a single diastereomer. The iridium‐catalyzed reaction was completely chemoselective without causing hydrosilylation of the exo‐olefin. Benzyl alcohol was added to quench the remaining silane, which could become a reducing agent when nitrone 13 was heated at 90 °C (14: 56% without benzyl alcohol). Although the role of Et₃N has yet to be elucidated, addition of Et₃N after formation of N,O‐acetal 24 improved the yield (14: 78% without Et₃N). Thus, reductive [3 + 2] cycloaddition of N‐hydroxylactam 12 was successfully demonstrated in a one‐pot process to give tricyclic isoxazolidine 14 including a C8 quaternary carbon center.

With tricyclic isoxazolidine 14 in hand, we turned our attention to the construction of the pentacyclic framework (Scheme 4). The cleavage of the N─O bond in 14 with zinc at 90 °C gave 1,3‐amino alcohol 25. Chemoselective condensation between amine 25 and carboxylic acid 15 with DMT‐MM^{[ 95} , ^{96 ]} provided amide 26 in 81% yield without formation of the ester (2 steps). After TPAP‐oxidation of the remaining primary alcohol, the intramolecular aldol reaction^{[ 97} , ^{98 ]} of aldehyde 27 with LiHMDS and subsequent Dess–Martin oxidation^{[ 99} , ^{100 ]} of 28 gave cyclic β‐ketoamide 29 in 74% yield (3 steps). Cyclic β‐ketoamide 29 was converted to enaminone 30 ^{[ 101} , ¹⁰² , ^{103 ]} by a slightly modified Sarpong method.^{[ 15} , ^{16 ]} Thus, treatment of cyclic β‐ketoamide 29 in THF with 3 mol% of IrCl(CO)(PPh₃)₂ and (Me₂HSi)₂O initiated amide‐selective hydrosilylation and subsequent elimination, providing 30 in 68% yield without affecting the ketone. Michael addition of the cuprate, which was prepared from allylic silane fragment 16 afforded tetracyclic intermediate 17α in 64% yield as a single trans isomer.^{[ 12} , ¹⁵ , ^{16 ]} The next stage was intramolecular Hosomi–Sakurai allylation^{[ 77} , ⁷⁸ , ^{79 ]} of 17α to construct the pentacyclic intermediate. In this event, use of aqueous acidic conditions was essential because hydrolysis of the cyclic acetal to produce the aldehyde must be faster than cyclization. We found that 6 M HCl/THF = 2 at room temperature was optimal to achieve both hydrolysis and cyclization in a single operation, providing pentacyclic intermediates 31, 32, and 33 in 59% combined yields (31:32:33 = 1:2:1).^{[ 104 ]} The dimethylphenyl silane of 33 was removed with TBAF in DMSO at 70 °C in quantitative yield to give 32. All three products 31–33 possessed the correct stereochemistry at C10 because Hosomi–Sakurai allylation is known to proceed from the opposite face of the silyl group.^{[ 78 ]} In contrast, the C6 stereocenter underwent the complete epimerization, resulting in the cis‐configuration between C6 and C7. It is unclear whether the epimerization at C6 via the enol occurred before or after cyclization. Based on these stereochemical outcomes, we set our first goal to be the total synthesis of isodaphlongamine H (5). A mixture of 31 and 32 was converted to xanthate 34, which underwent Barton–MaCombie deoxygenation with AIBN and nBu₃SnH,^{[ 105} , ^{106 ]} giving 35 in 76% yield (2 steps). Although the cleavage of the benzyl group was not trivial, treatment of 35 with TMSOTf and Ac₂O in CH₂Cl₂ at 0 °C provided the corresponding acetate 36. Nucleophilic addition of methyl lithium to 36 in the presence of CeCl₃ proceeded in a complete stereoselective manner from the less hindered β‐side, associated with removal of the acetate. Finally, Jones’ oxidation of the resulting primary alcohol and spontaneous lactonization^{[ 15} , ^{16 ]} accomplished the total synthesis of isodaphlongamine H (5) in 58% yield (3 steps).

Scheme 4.

Total synthesis of isodaphlongamine H (5). AIBN = azobis(isobutyronitrile), DMSO = dimethyl sulfoxide, DMT‐MM = 4‐(4,6‐dimethoxy‐1,3,5‐triazin‐2‐yl)‐4‐methyl‐morpholinium chloride, NMM = N‐methylmorphiline, NMO = N‐methylmorpholine oxide, TMS = trimethylsilyl, and TPAP = tetrapropylammonium perruthenate.

The Hanessian group documented that the natural product deoxycalyciphylline B (2) showed cytotoxicity comparable to unnatural isodaphlongamine H (5) in a panel screening of NCI human cancer cell lines.^{[ 11 ]} Encouraged by their report, the growth inhibitory activities of our synthetic intermediates were evaluated against HeLa (cervical cancer) and U937 (histiocytic lymphoma) cell lines (Table 1). As the Hanessian group reported, we also observed that isodaphlongamine H (5) showed cytotoxicity (IC₅₀ = 12.8 µM against HeLa cells, IC₅₀ = 4.9 µM against U937 cells). Interestingly, most tested synthetic intermediates after reductive [3 + 2] cycloaddition were found to be cytotoxic even if they were intermediates in the middle stage of the synthesis. For example, the IC₅₀ value for tricyclic isoxazolidine 14 was approximately half that for isodaphlongamine H (5). Pentacyclic intermediates including 31 and 32 exhibited cytotoxic activity comparable to that of isodaphlongamine H (5). The most potent intermediate was found to be deoxygenated pentacyclic compound 35, which showed better activity than isodaphlongamine H (5) (IC₅₀ = 1.8 µM against HeLa cells, IC₅₀ = 1.4 µM against U937 cells).

Table 1.

Cytotoxicity of isodaphlongamine H and synthetic intermediates against HeLa and U937 cell lines (IC₅₀ values in µM). ^a)

	IC50 (µM)
Cell Line	14	26	30	31	32	35	5
HeLa	19.6	62.5	32.2	8.0	10.4	1.8	12.8
U937	15.4	27.1	6.6	5.6	9.2	1.4	4.9

^a) Antiproliferative effects of tested compounds for 72 h against human cancer (HeLa and U937) cell lines were assessed. The IC₅₀ values were calculated by three independent MTT assays.

The present synthesis of isodaphlongamine H (5) by the lactam strategy was accomplished in 23 steps with 0.8% total yield from commercially available diester 18. Our present synthesis of isodaphlongamine H demonstrated the utility of the lactam strategy in the total synthesis of complex alkaloids. The synthesis commenced with alkylation of a readily available chiral lactam. The resulting trans‐lactam was converted to the desired cis‐lactam by kinetic epimerization with dimethyl malonate as a proton source in the presence of iBu₃Al. N‐Oxidation of the secondary lactam was efficiently achieved by a two‐step procedure via the imino ether. The key transformation in the lactam strategy was iridium‐catalyzed reductive [3 + 2] cycloaddition of the N‐hydroxylactam. A sequence including N‐oxidation and iridium‐catalyzed reduction proved to be useful for the synthesis of chiral cyclic nitrones, which are not quickly accessible by other conventional methods. After Michael addition of an allylic silane fragment, intramolecular Hosomi–Sakurai allylation constructed a pentacyclic framework accompanied by epimerization at the C6 carbon center. We expect that the deoxygenated pentacyclic intermediate could become a lead compound for further biological study.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting information

Iwamoto S., Nakano R., Sasaki K., Kobayashi S., Taira Y., Takei K., Kawakita R., Tokuyama A., Nakamura H., Tomoike M., Kawahara R., Murase A., Simizu S., Chida N., Okamura T., Sato T., Angew. Chem. Int. Ed. 2025, 64, e08062. 10.1002/anie.202508062

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.