Total Synthesis of Complex Diketopiperazine Alkaloids

Abstract

The 2,5-diketopiperazine (DKP) motif is present in many biologically relevant, complex natural products. The cyclodipeptide substructure offers structural rigidity and stability to proteolysis that makes these compounds promising candidates for medical applications. Due to their fascinating molecular architecture, synthetic organic chemists have focused significant effort on the total synthesis of these compounds. This review covers many such efforts on the total synthesis of DKP containing complex alkaloid natural products.

Introduction

The 2,5-diketopiperazine (DKP) substructure is found in a diverse family of biologically active natural products. Alkaloids featuring a DKP (Figure 1) core embedded in more elaborate molecular structures have been isolated from fungal, bacterial, and animal sources.^1,2,3 There has been significant interest to access these compounds through chemical synthesis due to their range of antitumor, antiviral, antifungal, and antibacterial functions as well as fascinating molecular architecture. Relative to acyclic peptides, the DKP ring, a cyclic dipeptide, confers additional rigidity and stability against proteolysis.¹ These favorable pharmacological properties, combined with their donor and acceptor capability for hydrogen bonding with biological targets have made naturally occurring DKPs or their derivatives promising candidates for development for various medicinal applications.^4,5

Figure 1.

A 2,5-diketopiperazine.

This review focuses on representative case studies in the total synthesis of complex DKP natural products with particular attention given to new strategies to form key bonds in these structures. The challenges that were addressed in these syntheses, namely introduction of difficult to secure stereocenters and bonds, including vicinal quaternary stereocenters, sensitivity of certain functional groups, and propensity for epimerization at α-stereocenters are highlighted. The selected syntheses are grouped by natural product family and presented in chronological order within each section. Other reviews have focused on the total synthesis of tryptophan based dimeric DKP natural products,^6,7 C3–C3′ linked bispyrroloindoline alkaloids including a subset of DKPs,⁸ and natural products with higher oxidation state of the DKP including sulfidation such as epidithiodiketopiperazines.^9,10,11 Due to their potent biological activity, many of these compounds have been the subject of biosynthetic studies,^12,13 and where relevant to the synthetic planning, that information is included as well.

Amauromine

The natural product (−)-amauromine (1, Figure 2) was isolated in 1984 from Amauroascus sp. No. 6237.¹⁴ It displayed vasodilating properties, determined by comparing relaxation activity of aortic strips treated with potassium chloride or norepinephrine to the known vasodilator papaverine. Its structure, containing two reverse prenylated groups with C2 symmetry, was determined in a follow-up report by spectroscopic methods and chemical modification of the natural extract.¹⁵ Findings from the early syntheses of this DKP informed later syntheses of other DKPs, offering an outstanding example to commence this review.

Figure 2.

Structure of (−)-amauromine.

In 1986, Aoki and coworkers published the first total synthesis of (−)-amauromine (1), taking advantage of a thio-Claisen rearrangement to form the key reverse-prenyl quaternary centers.¹⁶ The thio-Claisen rearrangement was proposed as a possible biosynthetic pathway for reverse prenylated natural products including the ergot alkaloids, echinulins, and ilamycins by Bycroft and Landon in 1970.¹⁷ Starting from 2-methylthio-L-tryptophan, DKP 2 was prepared via N,N-dicyclohexylcarbodiimide, N-hydroxysuccinimide (DCC-Hosu) coupling followed by cyclization with ammonia in methanol (Scheme 1). Treatment of this compound with prenyl bromide (3) and potassium carbonate over seven days afforded reverse prenylated DKP 4 (18%) and its epimer (15%) at the quaternary center, presumably via a thio-Claisen rearrangement. Activation of the thiomethyl group and cyclization to (−)-amauromine (1) was achieved via exposure to titanium tetrachloride and lithium aluminum hydride in low yield.

Scheme 1.

Total synthesis of (−)-amauromine (1). Conditions: (a) K₂CO₃, dioxane, 18%. (b) TiCl₄, LiAlH₄, THF, relfux, 15%. (c) N-phenylselenophthalimide, PTSA, Na₂SO₄, CH₂Cl₂, 78%. (d) MeOTf, prenyl tributylstannane, DTBMP, CH₂Cl₂, −78 °C to reflux, 60%, (e) TMSI, CH₃CN, 0 °C to rt, 58%. (f) 10, [Ir(COD)Cl]₂, (R)-13, 9-BBN-octyl, DBU, CH₂Cl₂, then piperidine, 11, 87% or 10, [Ir(COD)Cl]₂, (R)-13, Ph₃B, TBD, CH₂Cl₂, then piperidine,12, 77%. (g) CsOH, THF, MeOH; HBTU, 0 °C to rt, 35%.

In 1994, the Danishefsky group reported their synthesis of (−)-amauromine (1), relying on a diastereoselective, oxidative cyclization of tryptophan derivative 5 to selenoether 7 (Scheme 1).¹⁸ In this critical transformation, treatment of bis(Boc)tryptophan methyl ester (5) with N-phenylselenophthalimide and p-toluenesulfonic acid gave a 9:1 mixture of diastereomers (78% yield) favoring the desired exo isomer 6. Activation of the selenoether with methyl trifluoromethanesulfonate and trapping of the resulting carbocation with prenyl tributylstannane led to reverse-prenylated tricycle 7 with complete stereochemical control. Selective deprotection of the amine on one half of the molecule and the carboxylic acid on the other followed by peptide coupling gave amide 8, which upon removal of the carbamates spontaneously cyclized to (−)-amauromine (1). In a follow up report,¹⁹ the diastereoselectivity of their key oxidative cyclization was improved to 18:1 with the use of pyridinium p-toluenesulfonate instead of p-toluenesulfonic acid.

In 2016, the Stark group developed an iridium catalyzed method for the reverse prenylation of indole and tryptophan derivatives, allowing highly diastereoselective synthesis of both endo- and exo-pyrroloindolines 11 or 12, respectively.²⁰ For optimal endo selectivity in the reverse prenylation, tryptophan 9 was treated with allylic carbonate 10, the Ir-catalyst (2.5 mol%, [Ir(COD)Cl]₂), the phosphoramidite ligand 13 (10 mol%), 9-octyl-9-borabicyclo[3.3.1]nonane (9-BBN-Octyl), and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to give a 9:1 mixture of indolines 11 and 12 in 87% combined yield. Switching the boron reagent to triphenylborane and the base to 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) reversed the diastereoselectivity to give predominantly the exo-diastereomer 12 (exo 12:endo 11, >20:1) in 77% yield. Saponification of indoline 12 followed by peptide coupling gave (−)-amauromine (1) in 35% yield.

Ditryptophenaline and WIN 64821

The C3–C3′ linked dimeric DKPs constitute a large family of alkaloids bearing two quaternary carbons that fuse respective complex DKP fragments together (Figure 3). While their dimeric core is closely related to C3–C3′ linked bispyrroloindoline alkaloids derived from tryptamine,^21,22,23,24 the fusion of the DKP onto the hexahydropyrroloindoline substructure allows significant structural diversity.^1,4,8 Members of the dimeric DKP alkaloids, including the homodimer (−)-ditryptophenaline (14), have been isolated from both marine and terrestrial fungi species.¹ Compound 14, an alkaloid comprised of two N-methyl-L-phenylalanine and two L-tryptophan substructures, was the first C3–C3′-dimeric DKP alkaloid to be isolated and has since been identified in various samples from Aspergillus species.²⁵ Another example is the symmetrical dimer (+)-WIN 64821 (16), which was isolated from an Aspergillus sp. SC319 culture soil extract, and was identified as a potent competitive substance P antagonist.^26,27,28 In 1994, directed biosynthetic experiments led to isolation of dozens of symmetrical and unsymmetrical analogs of (+)-WIN 64821 (16) through feeding analogs of Phe, Trp, and other amino acids to intact Aspergillus sp. SC319 cells.²⁹ Woodward and Robinson hypothesized in the 1950s that the biogenesis of the structurally related cyclotryptamine family of natural products in addition to more complex cyclotryptophan-derived compounds arises from the oxidative radical dimerization of tryptamine or tryptophan monomers.^30,31 In 1981, Nakagawa, Hino and co-workers reported a thallium(III)-promoted oxidative dimerization of cyclo-L-N-methyl-Phe-L-Trp to give (−)-ditryptophenaline (14) in 3% yield.^32,33 In 2014, Watanabe reported the discovery of P450 DtpC as a biological catalyst for the oxidative dimerization leading to formation of the C3–C3′ bond.³⁴ Notably, higher diversification of these dimeric DKP alkaloids is observed via introduction of additional functional groups, including hydroxylation and sulfidation at the DKP rings, leading to distinct and large classes of alkaloids such as the epidithiodiketopiperazines, including (+)-11,11′-dideoxyverticillin (19)³⁵ and (+)-chaetocin A (20),³⁶ that have been reviewed separately.^9,10,11

Figure 3.

Representative C3–C3′ dimeric alkaloids.

In 2001, Overman and Paone reported an elegant total synthesis of (−)-ent-WIN 64821 (16) and (−)-ditryptophenaline (14) through a metal-chelated, stereocontrolled dialkylation approach to furnish the C3−C3′ vicinal quaternary centers (Scheme 2).³⁷ Alkylation of the bisoxindole 21 with the chiral pool-derived electrophile bistrifluoromethanesulfonate 22 proved to be a powerful method for the synthesis of various hexahydropyrrolindole natural products^38,39 including these two DKP targets. Oxidative cleavage of the cyclohexane diol 23 followed by Grignard addition and azidation of the resulting alcohols afforded the diazides 24 and 25 separately. Reduction of the oxindole with sodium bis(2-methoxyethoxy)aluminum hydride (Red-Al) and heating to 100 °C led to cyclization to the bispyrrolidinoindolines 26 and 27 in 52% and 71% yield, respectively. Having secured the critical C3–C3' linkage in dimeric intermediates 26 and 27, the stage was set for completion of the synthesis through introduction of the fused DKP ring. The bisamine 26 was then debenzylated, coupled with N-Me-phenylalanine, and cyclized to give (−)-ditryptophenaline (14) in 38% overall yield (6 steps from 26). The synthesis of (−)-ent-WIN 64821 (16) was accomplished using a similar strategy from diastereomeric bisamine 27 in 33% overall yield (7 steps from 27).

Scheme 2.

Overman’s synthesis of (−)-ditryptophenalanine (14) and (−)-ent-WIN 64821 (16). Conditions: (a) LHMDS, THF, DMPU, −78 °C, 55% (b) CSA, MeOH/CH₂Cl₂, 100% (c) Red-Al, PhMe, 26, 52%, 27, 71%.

In 2008, the Movassaghi group reported the syntheses of symmetrical dimeric alkaloids (+)-WIN 64821 (16), (−)-ditryptophenaline (14), and the unsymmetrical dimeric alkaloid (−)-N-(2-phenyl-ethylene)-ditryptophenaline (15).⁴⁰ Central to their synthetic strategy for these alkaloids was their use of a reductive cobalt(I) dimerization step to secure the critical C3–C3′ bond.⁴¹ Building on their earlier use of this biosynthesis-inspired radical-based dimerization strategy, which they demonstrated in the enantioselective total synthesis of dimeric cyclotryptamine alkaloids (+)-calycanthine, (+)-chimonanthine, and (+)-folicanthine,⁴¹ this approach allowed for the union of advanced and fully-formed DKP monomers. Inspired by earlier reports on the synthesis of bromohexahydropyrrolindoles,⁴² they found that bromocyclization of N-sulfonylated DKP 28 directly afforded the endo- and exo-tetracyclic bromides 29 and 30, respectively, in 86% combined yield (44:56 endo:exo).⁴⁰ Importantly, bromides 29 and 30 were the desired precursors for the alkaloids 16 and 14, respectively. They later demonstrated similar halocyclizations with high diastereocontrol through substrate and reaction condition design to utilize dipole minimization during bromide synthesis when a only one bromide diastereomer is required for the target, such as their application of this chemistry to the synthesis of chaetocins.^{11, 36} Treatment of the bromide 29 with tris(triphenylphosphine)cobalt chloride in acetone led to the dimeric DKP 32 in 48% yield. Desulfonylation using samarium iodide provided the first synthetic sample of (+)-WIN 64821 (16) with the correct absolute stereochemistry. In a similar approach, the exo-bromide 31, prepared by N-methylation of exo-bromide 30, was converted to (−)-ditryptophenaline (14) via a two-step sequence involving dimerization (52% yield) and desulfonylation (79% yield). Exposure of alkaloid 14 to phenylacetaldehyde dimethyl acetal and camphorsulfonic acid led to compound 34. Differentiation of the two halves⁴³ was achieved via partial hydrolysis to afford the heterodimer (−)-N-(2-phenyl-ethylene)-ditryptophenaline (15). The versatility of this biosynthesis-inspired approach, relying on dimerization of the transiently formed C3 radical, is highlighted by the efficiency of the dimerization using diastereomeric bromides with varying substitution. Importantly, their approach to access the radical intermediates (Scheme 3) for the critical dimerization provides complete stereochemical control in the C3−C3′ bond forming steps, using the appropriate bromide precursors in contrast to most oxidative approaches from indole precursors.⁴⁴ Indeed, the group’s successful application of this dimerization chemistry was critical to their synthesis of complex ETPs including (+)-11,11′-dideoxyverticillin (19)³⁵ and (+)-chaetocin A (20) (Figure 3).^11,36

Scheme 3.

Total synthesis of (+)-WIN 64821 (16), (−)-ditryptophenaline (14), and (−)-N-(2-phenyl-ethylene)-ditryptophenaline (15). Conditions: (a) Br₂, MeCN, 0 °C, 86%. (b) MeI, K₂CO₃, acetone, 93%. (c) CoCl(PPh₃)₃, acetone, 32, 48%, 33, 52%. (d) SmI₂, NMP, tBuOH, THF, 0 °C, 16, 75%, 14, 79%. (e) BnCH(OMe)₂, CSA, 81% (f) H₂O, C₆D₆, TFA, 48%.

In 2008, de Lera and coworkers reported the first synthesis of (+)-WIN 64745 (18, Scheme 4).⁴⁵ They applied the cobalt promoted dimerization of bromohexahydropyrrolindoles developed by Movassaghi^40,41 to access the dimer 35 via formation of the critical C3–C3' bond. Epimerization of the esters and deprotection of the amines afforded the C3–C3' linked intermediate 36, which could be subjected to DKP formation in a manner reminiscent of Overman’s synthesis of DKPs post C3–C3' bond formation en route to alkaloids 14 and 16 (Scheme 2). Since desymmetrization⁴³ is an effective strategy to access a subset of dimeric hexahydropyrrolindoles,^40,46 judicious choice of conditions for monoacylation of intermediate 36 provided an excellent opportunity to form two different DKPs. Thus, treatment of indoline 36 with one equivalent of N-Cbz-L-leucine and 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) at −15 °C followed by addition of N-Cbz-L-phenylalanine and additional HATU gave an unsymmetrical bisamide, which upon Cbz deprotection and cyclization afforded (+)-WIN 64745 (18). In a similar fashion de Lera and co-workers have accessed other natural and unnatural analogs of dimeric DKPs,⁴⁷ including (+)-asperdimin (17) that allowed reassignment of its structure.⁴⁸

Scheme 4.

Total synthesis of (+)-WIN 64745 (18), and (+)-aspermidin (17). Conditions: (a) LHMDS, THF, −15 °C then MeOH, −78 °C, 99%. (b) TMSI, MeCN, 0 °C, 93%. (c) N-Cbz-L-Leu, HATU, Et₃N, DMF, −15 °C; N-Cbz-L-Phe, HATU, Et₃N. (d) H₂, Pd/C, MeOH:H₂O (9:1), 74% over 2 steps. (e) 180 °C, 95%. (f) N-Fmoc-D-Leu, HATU, Et₃N, DMF, −15 °C; N-Fmoc-D-Val, HATU, Et₃N. (g) Et₂NH, MeOH, 64% over 2 steps.

Alternative conditions have been reported to prepare the C3–C3′ bond of dimeric hexahydropyrrolindoles. In 2013, Oguri and coworkers reported a reductive method similar to Movassaghi’s^40,41 use of cobalt(I) using catalytic NiI₂ in the presence of 1,2-bis(diphenylphosphino)ethane (DPPE) and manganese (Scheme 5).⁴⁹ This method generated endo-dimer 43 and exo-dimer 47 in 76% and 42% yield, respectively, from the corresponding endo-bromide 37 and exo-bromide 38. In 2014, the Peng group reported a nickel mediated reductive dimerization using zinc dust as reductant.⁵⁰ In 2013, Ishikawa reported an oxidative dimerization^{39, 44} of tryptophan 39 to prepare diastereomeric dimers 44 and 48 in 28% yield each.⁵¹ Their use of acidic conditions masks the primary amine allowing for the oxidation of indole by vanadium or manganese to give dimers 44 and 48, in addition to other regioisomeric dimers as discussed later in this review (see naseaseazines A (96) and B (97)). In 2015, Xia and coworkers reported the use of different protected forms of tryptophan to alter the selectivity of the product for the endo- or exo-diastereomer.⁵² Treatment of tryptophan 40 bearing an ortho-nitrobenzenesulfonyl group with superstoichiometric CuCl₂ favored the formation of the endo-diastereomer 45 in 37% (dr, 3:1) whereas the use of tryptophan 41 with a para-nitrobenzenesulfonyl group under identical conditions reversed the selectively and provided the exo-diastereomer 49 in 47% (dr, 3:1). In 2020, Tu reported a potassium iodide catalyzed system to access diastereomeric dimers 46 and 50 in 79% as a 1.2:1 mixture favoring endo-dimer 46.⁵³ The endo-dimers 43-46 and the exo-dimers 47-50 could be converted to (+)-WIN 64821 (16) and (−)-ditryptophenaline (14), respectively, using similar chemistry to that developed in the earlier syntheses described above.

Scheme 5.

Methods for construction of the C3−C3′ bond. Conditions: (a) 15 mol % NiI₂・6H₂O/DPPE, Mn, DMA, 43, 76% from 37, 47, 42% from 38. (b) V₂O₅, MeSO₃H, H₂O, −15 °C, 44, 28% 48, 28%. (c) DBU, CuCl₂, MeCN, 0 °C, 45:49 (3:1), 37% from 40, 49:45 (3:1), 47% from 41. (d) 10 mol % KI, NaBO₃, TFE, 46:50 (1.2:1), 79%.

Asperazine and Pestalazine A

In 1997, Crews and co-workers reported their isolation of asperazine (51) from the fungus Aspergillus niger extracted from a Caribbean marine sponge endosome (Figure 4).⁵⁴ This unique, dimeric DKP alkaloid possesses a C3–C8′ quaternary center binding two DKP tetracycles, and its absolute configuration was assigned by hydrolysis to unveil only (R)-phenylalanine. While not displaying any antibacterial or antifungal activity, asperazine (51) did exhibit selective in-vitro cytotoxicity against leukemia cell lines. The structurally related heterodimeric pestalazine A (52) was isolated in 2008 by Che and co-workers from cultures of the plant pathogenic fungus Pestalotiopsis theae, along with the N1′–C3 bound regioisomer pestalazine B (156, Figure 8).⁵⁵ Pestalazine A (52) inhibited HIV-1 replication in C8166 cells.

Figure 4.

Structure of +)-asperazine (51) and (+)-pestalazine A (52).

Figure 8.

(+)-asperazine A (155) and (+)-pestalazine B (156)

In 2001, Overman described the first total synthesis of (+)-asperazine (51) relying on a fascinating diastereoselective Heck reaction to set the key quaternary stereocenter.⁵⁶ The synthesis began with Stille coupling of iodoindole 53, which was synthesized in six steps from L-tryptophan methyl ester, and vinyl stannane 54 en route to vinyl indole 55 in 79% yield over three steps (Scheme 6). In the key step, exposure of the vinyl indole 55 to the Pd-catalyst and phosphine gave a single product 56 in 66% yield with the desired quaternary stereochemistry. Completion of the total synthesis of (+)-asperazine (51) was achieved over additional steps including the substrate-controlled hydrogenation (4:1 dr) of intermediate 56, partial reduction of the resulting oxindole and cyclization to the hexahydropyrroloindoline substructure, followed by condensation and formation of the DKPs.

Scheme 6.

Total synthesis of (+)-asperazine (51). Conditions: (a) Pd₂(dba)₃•CHCl₃, (2-furyl)₃P, PMP, DMA, 90 °C, 66%. (b) Bu₄NF, 85%.

In 2016, the Movassaghi group published their synthesis of (+)-asperazine (51) as well as the related natural product (+)-pestalazine A (52).⁵⁷ Of central interest to their final synthetic strategy was the controlled union of nucleophilic and electrophilic complex DKPs in nearly their final form. Their initial retrosynthetic analysis for (+)-asperazine (51) relied on a boronated indole nucleophile reminiscent of the type used for their synthesis of (+)-naseseazines A (96) and B (97) via a completely regioselective Friedel-Crafts union (Scheme 13).⁵⁸ While treatment of tetracyclic bromide 58 (Scheme 7) with silver promoter and boronate 57 gave the desired adduct 59 in 22% yield, leading to (+)-asperazine (51) through additional steps, the synthesis suffered from low yield of the key step and lengthy synthesis of boronate 57. They posited that increasing the nucleophilicity at the desired carbon by using an indoline would both streamline the synthesis and increase the efficiency in the critical bond formation. To that end, the tetracyclic nucleophile 60 and the electrophile 62 were prepared from a common phenylalanine–tryptophan DKP whereas the nucleophile 61 and the electrophile 63 were prepared from the corresponding leucine-tryptophan DKP precursor. Importantly, treatment of tetracyclic bromides 62 or 63 with silver(I) led to electrophilic activation and coupling with the desired indolines 60 or 61 to directly give the corresponding dimeric product (64-66, Scheme 7). The dimers 64-66 were converted to the desired targets through sequential hydrogenation and acid promoted reductive opening of the northern pyrroloindoline substructure to give alkaloids 51, 52, and 67. While the data for (+)-asperazine (51) matched those reported in the literature, they noted key differences in the data for their synthetic sample of the “pestalazine A” (later renamed iso-pestalazine (67)) and those reported in the isolation report for pestalazine A. However, their data for the synthetic sample having the phenylalanine and leucine amino acid positions switched was in complete agreement with the isolation data, leading to their revision of the structure of (+)-pestalazine A (52) as that shown in Figure 4 and Scheme 7.⁵⁷

Scheme 13.

Total synthesis of (+)-naseseazines A (96) and B (97). Conditions: (a) PyHBr₃, 2,2,2-trifluoroethanol, 101, 67%, 102, 51%. (b) AgSbF₆, 100, EtNO₂, 1:1.4 103:105, 53%, 1:1.4 104:106, 47%. (c) H₂, Pd/C, AcOH, 107, 52%, 108, 61%. (d) H₂, Pd/C, AcOH, 96, 80%, 97, 80%. (e) 2-aminobiphenyl-(XPhos)PdCl (5 mol %), XPhos (15 mol %), (Bpin)₂, K₃PO₄, DMSO, 60 °C, 65%. (f) KHF₂ (aq), MeOH, 88%. (g) AgSbF₆, 18-crown-6, EtNO₂, 105, 56%, 106, 50%. (h) H₂, Pd/C, AcOH, 96, 80%, 97, 80%. (i) (CuOTf)₂•PhMe (10 mol %), 117 (40 mol %), CH₂Cl₂, 118, 82%, 119, 92%. (j) Pd[P(o-tol)₃]₂, Na₂CO₃, DMF, 100 °C; 1N HCl/MeOH, 96, 56%, 97, 51%.

Scheme 7.

Total synthesis of (+)-asperazine (51) and (+)-pestalazine A (52). Conditions: (a) AgSbF₆, DTBMP, 18-crown-6, EtNO₂, 22%. (b) AgSbF₆, DTBMP, EtNO₂ 64, 41%, 66, 48%; AgSbF₆, DTBMP, 3:1 EtNO₂:CH₂Cl₂ 65, 32%.

In 2019, Gong, Peng, and coworkers published their synthesis of (+)-asperazine (51) and (+)-pestalazine A (52) via an interesting nickel-catalyzed reductive coupling to forge the quaternary center (Scheme 8).⁵⁹ Starting from N-Cbz-L-tryptophan, borylated tryptophan 68 was prepared using an approach reported by Movassaghi for a similar substrate.^60,61 After conversion of the carbon-boron bond of indole 68 to the corresponding carbon-iodide bond of intermediate 69, a nickel-catalyzed reductive coupling with chloride 70 gave the dimer 71 in 47% yield. Subsequent peptide coupling reactions and DKP synthesis allowed for completion of their synthesis of alkaloids 51, 52, and 67.

Scheme 8.

Total synthesis of (+)-asperazine (51) and (+)-pestalazine A (52). Conditions: (a) Cu₂O, NaI, aq NH₃/MeOH, 76%. (b) Zn, Ni(acac)₂, 3-F-pyridine, MgCl₂, LiCl, DMF, 47%.

Stephacidins A, B

In 2002, researchers at Bristol-Myers Squibb reported the isolation of biologically active metabolites from a fermentation broth of Aspergillius ochraceus WC76466.⁶² The stephacidins A (72) and B (73) (Figure 5) displayed potent, nanomolar inhibition of several human tumor cell lines, such as testosterone-dependent prostate LNCaP lymphoma. Studies of the biological activity suggest that the natural products 73 and avrainvillamiade (74) interact with a variety of cysteine-containing proteins, including heat-shock protein 60 (HSP60), exportin 1 (XPO1), glutathione reductase (GR), and peroxiredoxin 1 (PRX1).⁶³ Both the structures of 72 and 73 were deciphered through NMR experiments, with X-ray crystallography also required in the case of dimeric alkaloid 73. Inspection of the stephacidins suggests a plausible biogenesis of alkaloid 73 predicated on dimerization of avrainvillamide (74), which itself stems from oxidation of DKP 72.⁶⁴

Figure 5.

Stephacidins A (72), B (73), and avbrainvillamide (74).

In 2005, the Baran group published the first total synthesis of stephacidin A (72).⁶⁵ The coupling of the amino acid 75, prepared efficiently by application of a modified indole synthesis reported by Reider and coworkers at Merck,⁶⁶ with an enantiopure proline derivative gave DKP 76 (Scheme 9). Enolization of DKP 76 followed by oxidative C–C bond formation provided the intermediate 77 in 41% yield as a single diastereomer. Deprotection of the methoxymethyl ether followed by conversion of the methyl ester to an isobutenyl group gave indole 78, which upon heating (neat) to 200 °C led to formation of stephacidin A (72). The data for their synthetic stephacidin A (72) matched those reported in the isolation report securing its relative configuration.⁶⁵ Since no optical rotation data was provided in the isolation report, the absolute stereochemistry could not be confirmed at the time. The authors assigned their synthetic sample as (−)-ent-stephacidin A (72) in a later report in 2006.⁶⁷

Scheme 9.

Total Synthesis of (−)-ent-stephacidin A (72). Conditions: (a) LDA, Fe(acac)₃, 41%. (b) B-bromocatecholborane, CH₂Cl₂, 63%. (c) MeMgBr, PhMe then Burgess reagent, PhH, 88% (d) 200 °C, neat, 45%.

In 2005, Myers and Herzon reported the first enantioselective synthesis of (+)-stephacidin B (73) (Scheme 10).⁶⁸ Alkylation of cyclohexenone 79 with electrophile 80 gave the adduct 81 in 70% yield. Addition of hydrogen cyanide across the enamine, epimerization of the ketone, and hydration of the cyano group afforded amide 82, which was converted to bisamide 83. Heating of compound 83 in the presence of tert-amyl peroxybenzoate led to formation of the bridged DKP 84 in 62% yield. Conversion of DKP 84 to α-iodoenone 85 and coupling with aryl iodide 86 afforded enone 87. Reduction of nitroarene 87 by activated zinc powder gave (−)-ent-avrainvillamide (74) in 49% yield. Treatment of synthetic (−)-avrainvillamide (74) with triethylamine in acetonitrile led to formation of (+)-ent-stephacidin B (73) with remarkable efficiency (est. >95%). Indeed, interconversion between the alkaloids 73 and 74 was observed under a variety of conditions. Concentration of a pure sample of stephacidin B (73) from acetonitrile-water led to a 2:1 mixture of avrainvillamide (74):stephacidin B (73). Exposure of stephacidin B (73) to silica gel (2D TLC analysis) or addition of molecular sieves to a solution of stephacidin B (73) in a mixture of DMSO-d₆-CD₃CN (1:1) caused partial reversal of the dimerization. Notably, their synthetic sample of (+)-stephacidin B (73) clearly confirmed the relative stereochemistry of the alkaloid. The optical rotation of the natural stephacidin B (73) was not available at the time of this report to enable comparison of the absolute configuration.

Scheme 10.

Total synthesis of (+)-stephacidin B (73). Conditions: (a) KHMDS, 70%. (b) t-amylOOCOPh, t-BuPh, 62%. (c) 1. HF, 2. DMP, 3. I₂, DMAP, 72%. (d) Pd₂(dba)₃, Cu, 72%. (e) Zn, EtOH, 40 °C, 49%. (f) Et₃N, CH₃CN, est. 95%.

In 2006, Baran reported⁶⁷ that subjection of (+)-stephacidin A (72), the enantiomer of their earlier synthetic alkaloid (−)-ent-72 (Scheme 9), to Gribble reduction conditions yielded indoline 89 (93%), which, upon oxidation, gave (+)-avrainvillamide (74) in 27% yield (Scheme 11). Spontaneous dimerization of alkaloid (+)-74 to (−)-stephacidin B (73) was observed upon exposure to triethylamine and acetonitrile as observed by Myers.⁶⁸ Importantly, analysis of their synthetic samples along with natural samples of (+)-avrainvillamide (74) and (−)-stephacidin B (73) allowed assignment of their absolute configuration (Scheme 11).⁶⁷

Scheme 11.

Total synthesis of (−)-stephacidin B (73) establishing synthetic connection between (+)-stephacidin A (72) and (−)-stephacidin B (73). Conditions: (a) NaBH₃CN, AcOH, >95%. (b) SeO₂, H₂O₂, dioxane, 27%, 50% RSM. (c) Et₃N, CH₃CN, 15–20%, 70–80% recovered 74.

The Williams group had earlier proposed that natural products with bicyclo[2.2.2]diazaoctane substructures were biosynthetically accessed via a Diels-Alder reaction.^69,70 In 2007, they reported the total synthesis of stephacidin A (72) in racemic form via a Diels-Alder reaction (Scheme 12). Base promoted tautomerization of the DKP derivative 90 to azadiene 91 followed by a Diels-Alder reaction gave two diastereomeric products favoring diastereomer 92, which was converted to (±)-stephacidin A (72). A similar biomimetic Diels-Alder reaction was used by the same group to construct related DKP natural products paraherquamide A,⁷¹ VM55599,^72,73 notoamide B,⁶⁹ and the malbrancheamides.⁷⁴

Scheme 12.

Williams and Sarpong syntheses of stephacidin A (72). Conditions: (a) 20% aq KOH, MeOH, 86% (61% syn 92; 25% anti). (b) HCl, THF, 0 °C, then NaHCO₃, 96%. (c) DMP then K₂CO₃, acetone, 42%. (d) H₂NNH₂, KOtBu, ethylene glycol, (e) NaBH₄, 40% over 2 steps.

In 2015, Sarpong and coworkers reported their synthesis of (+)-stephacidin A (72).⁷⁵ The authors identified compounds analogous to amide 93 as a structure that could be useful for synthesis of stephacidin and related compounds including those without the bicyclo[2.2.2]diazaoctane substructure. Oxidation of the secondary alcohol 93, followed by exposure to potassium bicarbonate afforded the bridged DKP 95, presumably via the isocyanate intermediate 94, in 42% yield. Wolff-Kishner reduction of the 3-pyrrolidinone followed by reduction and elimination of the chromanone gave (+)-stephacidin A (72) in 40% yield over two steps. This approach offered a route to access (+)-notoamide I upon treatment of (+)-stephacidin A (72) with excess manganese dioxide in ethyl acetate (32% yield). The authors later explored dimerization of (+)-stephacidin A (72) and related derivatives under conditions described in the earlier syntheses of stephacidin B (73) described above.⁷⁶

Naseseazines A, B, and C

Chemical analysis of a Streptomyces sp. (CMB-MQ030) isolated from Fijian marine sediment by Capon and coworkers⁷⁷ in 2009 yielded two new DKPs, naseseazines A (96) and B (97), featuring a unique dimeric framework. Structures were determined by detailed spectroscopic analysis and C3 Marfey’s analysis. Naseseazine C (98) was isolated in 2016 and is the only member of the family to possess biological activity (3.5 μM against Plasmodium falciparum (3D7)).^78,79 Similar to the alkaloid asperazine (51, Figure 4), the naseseazines are members of a rare class of dimeric DKPs exhibiting a union between the C3 of one tetracyclic DKP and the indole of the other DKP. The naseseazines display added novelty in that they are bacterial not fungal in origin and possess a unique carbon framework defined by a C3 to C7′/C6′ linkage. Naseseazines A (96) and B (97) were determined to be noncytotoxic in antibacterial and antifungal assays. Biosynthetically, a cyclodipeptide synthase and a P450 enzyme together mediate the C−C bond formation step. Cyclodipeptide synthases nascA, nzn A, and nze A together with P450 nascB, nznB, and nzeB respectively have been identified as gene clusters responsible for these transformations.^80,81,82

In 2011, the Movassaghi group reported the first total synthesis of naseseazines A (96) and B (97) and revised the structure of these alkaloids to those shown in Figure 6.⁵⁸ Their synthesis relied on a Friedel-Crafts dimerization of late-stage diketopiperazine intermediates. Their synthetic strategy for securing the dimeric structure was based on their biosynthetic hypothesis that late-stage dimerization occurs via union of two complex DKPs. The strategic bond formation was envisioned to occur through interception of a cyclotryptophan-derived C3 electrophile.⁵⁸ The synthesis began with bromocyclization of alanine- and proline-based DKPs (99 and 100, respectively) (Scheme 13) by treatment with pyridinium tribromide to afford the tetracyclic bromides 101 and 102, respectively. Exposure of the alanine derived tetracyclic bromide 101 to silver hexafluoroantimonate generated the C3 carbocation, which, in the presence of DKP 100, gave a regioisomeric mixture of C6′ and C7′ linked heterodimers 103 and 105 (1:1.4) in 53% combined yield. Similarly, treatment of proline-derived DKP 102 with silver hexafluoroantimonate and DKP 100 provided a mixture of dimers 104 and 106 (1:1.4) in 49% combined yield. Hydrogenation of the indoline benzyloxycarbonyl (Cbz) group gave the first access to natural alkaloids naseseazines A (96) and B (97) in a highly concise route in addition to samples of isomers named (+)-iso-naseseazines A (107) and B (108). Later, the Movassaghi group used the same approach to access the first synthetic samples of the exo-cyclotryptophan based natural alkaloid (−)-naseseazine C (98) and a regioisomer (−)-iso-naseseazine C with the C3–C7' connectivity to confirm the structure of alkaloid 98 as that shown in Figure 6.⁷⁹

Figure 6.

Structure of Naseseazines A (96), B (97), and C (98).

The Movassaghi group also described their second-generation approach to naseseazines A (96) and B (97) that avoided the formation of the isomeric products 107 and 108. The completely regioselective syntheses of these alkaloids were achieved by introducing an aromatic directing group to enhance the nucleophilicity at the desired position of the indole ring. Starting from 6-bromoindole-3-carboxaldehyde, brominated DKP 109 was prepared in 70% yield over 5 steps. Palladium-catalyzed cross coupling using Buchwald’s precatalyst 2-aminobiphenyl-(XPhos)PdCl⁸³ and bis(pinacolato)diboron gave the desired pinacol boronate 110 without loss of the Cbz group or undesired aromatic bromide reduction. The potassium trifluoroborate 111 was used in the key dimerization step and allowed for regiochemical control via charge pairing when exposessd to tetracyclic carbocation 113. A final-stage hydrogenolysis gave the natural products (+)-naseseazine A (96) and B (97) exclusively as the desired regioisomers. Complete agreement of the synthetic compounds’ optical rotations with those of the natural alkaloids allowed for revision of their absolute stereochemistry, and clarified that these alkaloids are derived from l-amino acids instead of d-amino acids.

In 2013, the Reisman group reported their total synthesis of naseseazines A (96) and B (97) relying on a fascinating copper-catalyzed, diastereoselective arylation method to forge the key sp³−sp² bond.⁸⁴ Previous methods for direct synthesis of aryl pyrroloindolines^85,86 lacked the necessary substituents to access DKPs. Pyrroloindoline formation and arylation with diaryl iodonium salt precursor 116, prepared in two steps from 2-bromo-5-iodoaniline, gave DKPs 118 and 119 from the corresponding DKPs 114 and 115, respectively. A modified Larock indolization^87,88 was then employed to construct the indole ring and complete the syntheses.

In the same year, the Ishikawa group reported the direct oxidative dimerization of l-tryptophan methyl ester (121) to access dimer 122 in modest yield (28% yield). This intermediate was then converted to (+)-naseseazine B (97) via peptide coupling and heating to result in DKP formation (Scheme 14).⁵¹

Scheme 14.

Total synthesis of (+)-naseseazine B (97). Conditions: (a) Mn(OAc)₃, MeSO₃H, H₂O, 28%. (b) Boc-L-Proline, DMTMM, EtOH; neat, 230 °C, 70%.

Gliocladins

Gliocladins A-C were isolated together from a strain of Gliocladium found in the sea hare Aplysia kurodai (Figure 7).⁸⁹ Two of the compounds, gliocladins A (123) and B (124) are DKPs with sulfidation of the α-carbons while (+)-gliocladin C (125) is a triketopiperazine. All were tested for biological activity in P388 lymphocytic leukemia cells and gliocladin C (125) was more than twice as active as any other member of the family.

Figure 7.

Structure of (+)-gliocladins A (123), B (124), and C (125).

The first total synthesis of (+)-gliocladin C (125) was described by Overman and coworkers in 2007 (Scheme 15).⁹⁰ Construction of the key C3-quaternary center was achieved via a Mukaiyama aldol reaction the group first described in 2005.⁹¹ To this end, oxindole derivative 126 was treated with boron trifluoride diethyl etherate, Garner’s aldehyde (127),⁹² and 2,6-di-tert-butyl-4-methylpyridine (DTBMP) to give aldol product 128 in 89% yield with excellent diastereoselectivity (>80:1, dr). The pyrroloindoline portion of the intermediate was then elaborated over 14 steps to the secondary amide 129. While methods to generate the triketopiperazine in one step from this intermediate failed, treatment of the amide with ethyl chlorooxoacetate followed by heating in bis(trimethylsilyl)amine (HMDS) according to a method described by Mulliez⁹³ successfully formed (+)-gliocladin C (125) in 63% over the two steps. The absolute stereochemistry of the natural product was confirmed via single-crystal X-ray analysis of an intermediate.

Scheme 15.

Overman’s total synthesis of (+)-gliocladin C (125). Conditions: (a) BF₃•OEt₂, 127, DTBMP, DCM, −78 °C, 89%. (b) 2,2,2-trichloro-1,1-dimethylethyl chloroformate, Et₃N, (S)-(−)-4-pyrrolidinopyrindinyl(pentamethylcyclopentadienyl)iron, THF 88%. (c) NaBH₄, MeOH 81%. (d) HC(OMe)₃, PPTS, MeOH, 83%. (e) LiBH₄, MeOH, Et₂O, 84%. (f) Dess-Martin periodinane, pyridine, CH₂Cl₂, 95%. (g) (i) 134, LDA, THF; (ii) 133; AcOH, 75%. (h) Sc(OTf)₃, MeCN, 60%.

The Overman group reported their second-generation synthesis of (+)-gliocladin C (125) in 2011 (Scheme 15).⁹⁴ They found that exposure of oxindole 130 to Fu’s (S)-(−)-4-pyrrolidinopyrindinyl(pentamethylcyclopentadienyl)iron catalyst,⁹⁵ 2,2,2-trichloro-1,1-dimethylethyl chloroformate, and triethylamine gave oxindole ester 131 in 88% yield and excellent enantiomeric ratio (98:2) on decagram scale. The oxindole carbonyl was reduced to the O-methyl hemiaminal, and the ester was reduced to the aldehyde 133, a compound poised for direct condensation with triketopiperazine derivative 134. Aldol condensation of aldehyde 133 and compound 134 afforded the unsaturated intermediate 135, which, upon treatment with scandium trifluoromethanesulfonate, gave (+)-gliocladin C (125).

Since Overman’s synthesis of (+)-gliocladin C (125), other syntheses have been reported with notable differences in the methods used to form the C3-stereocenter. In 2011, Stephenson reported a visible light-mediated radical process for coupling bromopyrroloindolines such as 136 and indole nucleophiles (Scheme 16).⁹⁶ They found that exposure of bromopyrroloindoline 136 and indole-2-carboxaldehyde (137) to photocatalyst tris(bipyridyl)ruthenium(II) chloride along with blue LEDs gave the coupled pyrroloindoline 138 in 82% yield. Decarbonylation of the C2-aldehyde and synthesis of the triketopiperazine completed their synthesis of (+)-gliocladin C (125).

Scheme 16.

Stephenson’s total synthesis of (+)-gliocladin C (125). Conditions: (a) [Ru(bpy)₃Cl₂], Bu₃N, DMF, blue LEDs, 82%.

In 2012, Movassaghi and coworkers described their approach to (+)-gliocladin C (125) as well as the related sulfidated natural product (+)-gliocladin B (124) (Scheme 17).⁹⁷ This approach relied on Friedel-Crafts alkylation of a C3-carbocation with an indole nucleophile. Ionization of bromopyrroloindoline 139 with silver (I) followed by nucleophilic trapping with indole derivative 140 gave hexacycle 142 in 83% yield and excellent regioselectivity. Desilylation and and aryl bromide reduction provided DKP 143. Application of their DKP-hydroxylation methodology^98,99 led to stereoselective dihydroxylation³⁵ to afforded diol 144, a versatile intermediate for members of the gliocladin family. Diol 144 was converted to (+)-gliocladin C (125) through a sequence of steps ensuring controlled dehydration and oxidation, whereas its treatment with sodium thiomethoxide and trifluoroacetic acid gave dithioether 145 that upon desulfonylation provided the first total synthesis of (+)-gliocladin B (124) and allowed stereochemical assignment of the C15-methyl sulfide.

Scheme 17.

Total synthesis of (+)-gliocladin C (125) and (+)-gliocladin B (124). Conditions: (a) AgBF₄, 140, DTBMP, EtNO₂, 83%. (b) H₂, Pd/C, MeOH, EtOAc; Et₃N•HF, 100%. (c) nBu₄MnO₄, CH₂Cl₂, 41%. (d) MeSNa, TFA-MeNO₂ (1:1 v/v), 0→23 °C, 77%. (e) hν (350 nm), 1,4-dimethoxynaphthaline, ascorbic acid, sodium ascorbate, H₂O, MeCN, 88%.

In 2013, Gong reported¹⁰⁰ a catalytic enantioselective alkylation of the hydroxyacetaldehyde 147 (Scheme 18) to access a key intermediate that could be converted to (+)-gliocladin C (125) following similar conditions as those described earlier. In this approach to alkaloid (+)-125, the 2-((4-methoxybenzyl)oxy)acetaldehyde (147) was subject to asymmetric alkylation with the hydroxyoxindole 146 using quinidine-derived amine 148 and 1,1′-bi-2,2′napthol (BINOL) derived phosphoric acid 149 as catalysts to give aldehyde 150 in 80% yield, 94% ee (8:1 dr). Aldehyde 150 could be converted to indoline aldehyde 133, an intermediate used in Overman’s synthesis (Scheme 15),⁹⁴ in 54% yield over 8 steps. This was then subjected to the conditions described by Overman to afford (+)-gliocladin C (125) in 53% yield over 3 steps.

Scheme 18.

Approaches for construction of the C-3 quaternary center in gliocladin C (125). Conditions: (a) 10 mol% 148, 30 mol% 149, 80%, 94% ee, 8:1 d.r. (b) MsCl, (i-Pr)₂EtN, DMF; then DBU; then 152, 83%. (c) TMSOTf, DTBP, DCE, 0 °C; then TMSOTf, indole, 40 °C; then NH₄OH, 86%.

In 2017, Martin reported an approach to gliocladin C (125) in racemic form that is based on the use of a didehydrodiketopiperazine derivative as a nucleophile in a key carbon−carbon bond formation.¹⁰¹ Double dehydration of DKP 151 followed by nucleophilic addition into isatin derivative 152 furnished tertiary alcohol 153 (Scheme 18). Ionization of the alcohol 153 and trapping of the carbocation with indole in the presence of excess trimethylsilyltrifluoromethanesulfonate (5 equiv) led to DKP 154. Intermediate 154 was converted to (±)-gliocladin C (125) through a series of steps including the partial reduction of the oxindole, a dehydrative cyclization, and oxidative cleavage of the exocyclic methylene.

Asperazine A and Pestalazine B

A C3−N1′ linked regioisomer of asperazine (51), (+)-asperazine A (155), was identified in 2015 by the Lou group from the endophytic fungus Aspergillus niger (Figure 8).¹⁰² This natural product displayed weak cytotoxicity versus human cancer cell lines PC3, A2780, K562, MBA-MD-231, and NCI-H1688. A structurally related natural alkaloid (+)-pestalazine B (156) was isolated from Pestalotiopsis theae in 2008 along with (+)-pestalazine A (52).⁵⁵

In 2010, the de Lera group reported the first total synthesis of (+)-pestalazine B (156).¹⁰³ To forge the key nitrogen−carbon bond, they relied on the method first described by Rainier and Espejo¹⁰⁴ for base promoted coupling of indole with bromopyrroloindolines. Accordingly, treatment of DKP 157 with potassium tert-butoxide in the presence of bromopyrroloindoline 158 yielded dimer 159 in 31% yield (Scheme 19). Under the basic conditions employed in this coupling, the ester enolate is thought to displace the benzylic bromide, and the resulting cyclopropane facilitates the addition of the nucleophile at the benzylic position. Despite the modest yield for this transformation, the expedient formation of the advanced intermediate 159 allowed access to key intermediate en route to the natural product. N-Boc deprotection, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU) promoted peptide coupling, and cyclization to the DKP completed the total synthesis of (+)-pestalazine B (156). This total synthesis led to revision of the originally proposed structure, a regioisomer with the position of the phenylalanine and leucine residues exchanged.

Scheme 19.

Total synthesis of (+)-pestalazine B (156). Conditions: (a) KOt-Bu, MeCN, 31%. (b) 161, Cu(OAc)₂, Na₂CO₃, DMF, 45%. (c) Pd(OAc)₂, DTBPF, Na₂CO₃, NMP, 57%.

In 2015, the Liao group reported a total synthesis of (+)-pestalazine B (156).¹⁰⁵ The authors employed a diastereoselective azo-coupling of a tryptophan-leucine cyclodipeptide to prepare the southern DKP 160, similar to reports the year prior using diazonium compounds as electrophilic aminating reagents by Toste¹⁰⁶ and Larionov¹⁰⁷ that were applied to tryptamine substrates (Scheme 19). This was followed by a copper mediated coupling of DKP 160 and the iodonium salt 161 to obtain the aniline-DKP 162 in 45% yield. Application of a Larock indole synthesis¹⁰⁸ using o-bromoaniline 162 and alkyne 163 resulted in (+)-pestalazine B (156) in 57% yield.

The Movassaghi group published the synthesis of (+)-pestalazine B (156) and the first synthesis of related alkaloid (+)-asperazine A (155) in 2018 (Scheme 20).¹⁰⁹ Consistent with their earlier biosynthesis-inspired synthesis of the related alkaloids (+)-asperazine (51) and (+)-pestalazine A (52),⁵⁷ they focused on the convergent assembly of complex DKPs through late-stage formation of the strategic C–N linkage. They discovered that an exo-tetracyclic DKP such as indoline 164, selectively intercepts electrophiles at the N1' locus. This selectivity could be switched to interception by the C8' carbon of the same electrophile through deployment of the corresponding endo-diastereomer, as demonstrated by the use of indoline 60 (Scheme 7) in their synthesis of (+)-asperazine (51) and (+)-pestalazine A (52). This divergent C–N vs. C–C bond formation using exo- vs. endo-tetracyclic DKPs, respectively, was consistent with their unified biosynthetic hypothesis¹⁰⁹ for both modes of connectivity and offered a plausible rational for the observed divergent bond formation based on the conformation of the substrates in the putative active site of the enzyme responsible for the dimerization step.^79,82 Treatment of bromide 165 with silver(I) hexafluoroantimonate in the presence of exo-tetracyclic DKP 164 led to exclusive formation of the C–N coupled dimer 167 in 50% yield. Reductive opening of the aminal of the exo-tetracyclic DKP moiety (83% yield) followed by hydrogenation (73% yield) afforded the first synthetic sample of (+)-asperazine A (155), leading to the revision of the sign and magnitude of the optical rotation for this alkaloid. A similar synthetic sequence rapidly provided (+)-pestalazine B (156) starting with exo-tetracyclic DKP 164 and bromide 166, thus demonstrating the versatility of the synthetic approach.

Scheme 20.

Total synthesis of (+)-pestalazine B (156) and (+)-asperazine A (155). Conditions: (a) AgSbF₆, DTBMP, EtNO₂, 167, 50%, 168, 46%. (b) TASF, MeCN; MsOH, Et₃SiH, CH₂Cl₂, 171, 83% 172, 65%. (c) 10% Pd/C, HCO₂NH₄, Et₃N, MeOH, 155, 75%, 156, 73%,.

Nocardioazines A, B and Lansai B

In 2011, two C3-prenylated, DKP based natural products, nocardioazines A (173) and B (174), were isolated from a culture of Nocardiopsis sp. (CMB-M0232) (Figure 9).¹¹⁰ Nocardioazine A (173) exhibited strong inhibition of P-glycoprotein, a transmembrane protein that is often overexpressed in multi-drug resistant tumor cells, even reversing doxorubicin resistance in MDR colon cancer cells. The structurally related lansai B (175) was isolated from a fungal strain Streptomyces sp. SUC1 on the aerial roots of Ficus benjamina.¹¹¹ While lansai B (175) shares much of the structural features of nocardioazines A (173) and B (174), it harbors less biological activity.¹¹¹ Despite the structural similarity of these alkaloids, their significant differences include their oxidation, the site of prenylation, and the relative stereochemistry of the cyclotryptophan substructures.

Figure 9.

Structure of (+)-nocardioazine A (173), B (174), and (−)-lansai B (175).

Xu and Ye reported the first total synthesis and structural revision of nocardioazine B (174) in 2012.¹¹² Their retrosynthetic analysis traced nocardioazine B (174) back to hexahydropyrroloindoline building blocks 176 and 177 (Scheme 21). Amine 177 was prepared by allylation of the corresponding bromo-hexahydropyrroloindoline followed by oxidative double bond cleavage and a Wittig reaction to afford the prenylated hexahydropyrroloindoline 177 in 56% over 3 steps. The carboxylic acid 176 was prepared via the application of Rainier’s methodology¹¹³ for cyclopropanation and nucleophilic ring opening of the corresponding bromo-hexahydropyrroloindoline. Coupling of the acid 176 and amine 177 using 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) afforded the amide 178 in 60% yield. Intermediate 178 was converted to (−)-ent-nocardioazine B (174) through a series of steps including reductive amination and DKP cyclization. Importantly, while the NMR data for the synthetic sample of nocardioazine B (174) were identical to those reported in the isolation report, the optical rotation bore the opposite sign. This led to the revision of the absolute configuration of the natural product to that shown in Figure 9.

Scheme 21.

Total synthesis of (+)-nocardioazine B (174), (−)-lansai B (175), and (+)-nocardioazine A (173). Conditions: (a) HATU, HOAt, DMF, NEt₃, 60%. (b) 180, (R)-3,3′-Cl₂-BINOL, SnCl₄, 2,6-Br-phenol, CH₂Cl₂, 85%. (c) AcCl, MeOH. (d) LiOH, THF/H₂O, 76% over 2 steps. (e) BOPCl, DIPEA, CH₂Cl₂, 38% (+20% each homodimer). (f) 2-methyl-2-butene, HG-II, 83%. (g) PyBroP, DIPEA, DMF, 63%. (h) CuI (10 mol%), PhMe, LiHMDS, 60 °C; then MeI, 0 °C, 91%.

The Reisman group sought to synthesize alkaloids 173-175¹¹⁴ through application of their versatile enantioselective pyrroloindoline synthesis¹¹⁵ from C3-substituted indole derivatives.¹¹⁶ Treatment of the C3-substituted indole 179 with amidoacrylate 180, a 1,1′-bi-2-napthol (BINOL) derivative, and tin tetrachloride gave pyrroloindole 181 with high dr and ee (14:1 and 92% respectively) via a formal 3+2 cycloaddition. Notably, a similar approach was used to prepare the carboxylic acid 184. Coupling of the amino acids 183 and 184 yielded a statistical mixture of heterodimer (−)-lansai B (175, 38%) and the corresponding homodimeric side products (20% each). They prepared DKP 185 in an analogous fashion to the core of (−)-lansai B (175) through synthesis of each hexahydropyrroloindoline followed by peptide coupling and DKP formation. Cross metathesis of the allyl group of DKP 185 with 2-methyl-2-butene provided (+)-nocardioazine B (174). Interestingly, while attempts to epoxidize the olefin in DKP 186 failed to provide (+)-nocardioazine A (173), altering the order of DKP formation and epoxidation allowed for the synthesis of (+)-nocardioazine A (173) from epoxide 187.

In 2018, Yang and Zhang described a new approach to (+)-nocardioazine B (174) based on an enantioselective, copper-catalyzed arylation-alkylation methodology using o-haloanilides.¹¹⁷ Starting from o-iodoanilide 188, sequential arylation and alkylation afforded oxindole 189 in 91% yield as a single detectable diastereomer (Scheme 21). Deprotection of the amine followed by partial reduction of the oxindole gave the hexahydropyrroloindoline 190. Preparation of the other coupling partner via a similar route followed by DKP synthesis using the protocol described by Reisman¹¹⁴ led to the synthesis of (+)-nocardioazine B (174).

Azonazine

The alkaloid (+)-azonazine (191) was isolated from a Hawaiian marine sediment-derived fungus Aspergillus insulicola in 2010.¹¹⁸ It was found to show anti-inflammatory activity by inhibiting NF-κB luciferase and nitrite production. (+)-Azonazine (191) contains a benzofuranoindoline ring system with a quaternary stereocenter at the C10 position. A 10-membered ring connects the benzofuranoindoline and the DKP substructures of (+)-azonazine (191) and contributes to its rigid structure.

In 2013, the Yao group reported the first total synthesis of (−)-ent-azonazine (191).¹¹⁹ Exposure of DKP 192 to (diacetoxyiodo)benzene led to the formation of the diastereomeric cyclization products 193 and 194 in 16% and 12% yield, respectively (Scheme 22). The authors initially advanced the cyclization product 193 toward azonazine through a sequence of steps including C14 ketal hydrolysis, reduction to the desired indoline, and N-acetylation. However, the ¹H NMR data was not consistent with those from the isolation report of (+)-azonazine (191). Subsequently, the authors applied the same multi-step synthetic sequence to advance the diastereomeric cyclization product 194 and successfully gained access to (−)-ent-azonazine (191). Based on their results both the relative and absolute stereochemistry of (+)-azonazine (191) were revised.

Scheme 22.

Synthesis of (−)-ent-azonazine (191). Conditions: (e) PIDA, LiOAc, TFE, −15 °C, 193, 16%, 194, 12%.

Aspergilazine A

Aspergilazine A (195) is a dimeric DKP that was isolated in 2012 from the marine-derived fungus Aspergillus taichungensis ZHN-7-707 (Figure 11).¹²⁰ The authors report the use of Marfey’s analysis and NOESY experiments to establish the presence of l-proline and a cis-relationship between the α-methines of the DKPs, hence assigning it as a dimer of the alkaloid brevianamide F (196) with a unique N1′-C6 bond. Biological evaluation by the cytopathic effect (CPE) inhibition assay indicated that aspergilazine A (195) displayed weak activity against influenza A (H₁N₁) virus. A P450 enzyme catalyzes the dimerization of brevianamide F (196) to aspergilazine A (195) as well as C–C linked naseseazine alkaloids.^80,79,81,82

Figure 11.

Structure of (−)-Aspergilazine A (195) and (−)-brevianamide F (196).

In 2014, Sperry reported the first total synthesis of (−)-aspergilazine A (195) utilizing a palladium catalyzed N-arylation for the formation of the N1′–C6 bond.¹²¹ Initial attempts at coupling DKPs 196 and a deprotected version of 197 resulted in extensive epimerization (Scheme 23). The use of dicyclohexyl[2′,4′,6′-tris(propan-2-yl)[1,1′-biphenyl]-2-yl]phosphane (XPhos)¹²² as a ligand was beneficial and provided the desired coupling product (−)-aspergilazine A (195) in 9% yield with 80% recovered DKP 196. The N-Boc protected bromide 197 proved a more reliable coupling partner with (−)-brevianamide F (196), affording the desired dimer 198 in 79% yield. Deprotection of dimer 198 with trifluoracetic acid gave (−)-aspergilazine A (195) in 51% yield.

Scheme 23.

Total synthesis of (−)-aspergilazine A (195). Conditions: (a) Pd₂(dba)₃, X-Phos, K₃PO₄, PhMe, 100 °C, 79%. (b) TFA, CH₂Cl₂, 51%. (c) Pd₂(dba)₃, rac-BINAP, NaOtBu, PhMe, 70 °C, 75%. (d) Pd[P(t-Bu)₃], Cy₂NMe, 1,4-dioxane, 80 °C, 62%. (e) HCl/MeOH, 74%. (f) 5 mol % CuI, DMEDA, K₂CO₃, dioxane, 130 °C, 100%. (g) SiO₂, PhMe, 110 °C, 100%

In 2016, the Reisman group reported a fascinating synthesis of (−)-aspergilazine A (195) that relied on the early formation of the N1′–C6 bond (Scheme 23).¹²³ Buchwald-Hartwig coupling of 1-bromo-2-iodobenzene (199) with diamine 200 gave dibromide 201, which was poised for bidirectional Larock indole synthesis.¹⁰⁸ Under optimal reaction conditions, the coupling of dibromide 201 with alkyne 120 with a palladium-catalyst gave dimer 202 in 62% yield. Protodesilylation of the dimer 202 afforded (−)-aspergilazine A (195) in 74% yield.

In 2017, de Lera reported an interesting bis-indole based copper-based strategy for the synthesis of (−)-aspergilazine A (195).¹²⁴ The coupling of 6-bromo-N-tosyl-indole (203) and indole (204) with copper iodide (5 mol%) as catalyst under optimal conditions led to quantitative formation of bisindole 205 (Scheme 23). However, attempts to apply these conditions to more advance coupling partners such as amino acid derivatives and DKPs led to epimerization. Hence, conversion of bisindole 205 to the diiodide 206 in four steps followed by Negishi coupling of the zinc reagent derived from N-Boc-iodoalanine and condensation with N-Boc-L-proline gave tetrapeptide 207. Heating of intermediate 207 led to deprotection and DKP formation to quantitatively afford (−)-aspergilazine A (195).

Indotertine A and Drimentines A, F, G

The drimentine alkaloids represent a class of natural products biosynthesized through the union of terpenoid and nonterpenoid moieties. While leucine- and proline-derived drimentines A-E were reported in 1998,¹²⁵ valine-based (−)-drimentine F (210) and G (209), along with a related variant (+)-indotertine A (211), were isolated in 2012 from the soil-derived actinomycete Streptomyces sp. CHQ-64.¹²⁶ The structures of these DKP alkaloids, including absolute stereochemistry, were elucidated by spectroscopic methods, X-ray single crystal diffraction analysis, and time-dependent density functional theory electronic circular dichroism (TDDFT ECD) calculations. Among alkaloids 208-210, only (−)-drimentine G (209) exhibited any notable cytotoxicity against the in-vitro human tumor cell lines HCT-8, Bel-7402, A549, and A2780, with IC₅₀ values of 2.81, 1.38, 1.01, and 2.54 μM, respectively.

In 2013, the Li group disclosed the first total synthesis of (−)-drimentines A (208), F (210), and G (209) along with (+)-indotertine A (211).¹²⁷ The key C3-alkyl bond was forged through a visible-light mediated photoredox reaction, taking advantage of the C3-radical generated from bromopyrroloindoles such as tricycle 38 (Scheme 24). In their initial studies, Li noted that the use of tris(triphenylphosphine)cobalt chloride (Co(PPh₃)₃Cl) for attempted coupling of substrates 212 and 38 led to exclusive formation of a homodimeric product, consistent with Movassaghi’s strategy for the synthesis of homodimeric hexahydropyrroloindole alkaloids.^40,41 However, the application of conditions described by Crich,¹²⁸ namely the slow addition of tributyltin hydride, allowed for isolation of the desired product 213 in 58% yield. Under optimal conditions, the coupling of bromide 36 with the (+)-sclareolide-derived sesquiterpene 212 gave the adduct 213 in 91% yield using blue LEDs in the presence of photocatalyst [Ir(ppy)₂(dtbbpy)]PF₆. Coupling with the corresponding amino acids, DKP synthesis and ketone methylenation afforded the alkaloids 208-210 from intermediate 213. Treatment of (−)-drimentine F (210) with bismuth(III) trifluoromethanesulfonate and potassium hexafluorophosphate gave (+)-indotertine A (211) in 78% yield.

Scheme 24.

Total synthesis of (−)-drimentines A (208), G (209), F (210), (+)-indotertine A (211), and synthesis of drimentine isomer (216). Conditions: (a) [Ir(ppy)₂(dtbbpy)]PF₆, blue LEDs, NEt₃, 91%. (b) Bi(OTf)₃, KPF₆, 78%. (c) t-BuOK, BEt₃, 14%.

In 2016, Evanno and Poupon reported a biosynthesis-inspired synthesis of the drimentine A isomer 216 (Scheme 24).¹²⁹ Treatment of DKP 215 with potassium tert-butoxide and triethylborane in the presence of sclareolide-derived bromide 214 afforded the drimentine A isomer 216 in 14% yield. While compound 216 was not elaborated to a specific natural drimentine, its biological activity was evaluated and it was found to exhibit similar cytotoxic activity against three human tumor cell lines (HCT-116, A549 and K562) as that of the natural drimentines.

Penicimutanins A

Penicimutanin A (219) and B (220) are two marine fungus-derived DKP natural products isolated by Cui and coworkers in 2014 from Penicillium purpurogenum G59 through chemical mutagenesis.¹³⁰ The penicimutanin A (219) and B (220) and penicimutanolone (218) were found to possess a range of anti-fungal, anti-inflammatory, and anti-tumor activities. The authors reported the concurrent isolation of penicimutanolone (218) and fructigenine A (217) alongside the penicimutanins, indicating a potential biosynthetic interrelationship (Figure 13). Penicimutanin C (221) was isolated in 2020 by the same group and showed similar anti-cancer activity.¹³¹

Figure 13.

(+)-fructigenine A (217), penicimutanolone (218), (−)-penicimuanin A (219), B (220), and C (221).

In 2020, Xu and coworkers reported the first total synthesis of (−)-penicimutanin A (219).¹³² The authors hypothesized that (−)-penicimutanin A (219) is biosynthetically made from the union of penicimutanolone (218) and deacetyl-fructigenine A (226, Scheme 25). Treatment of tryptophan derivative 222 with N-chlorosuccinimide and 1,4-diazabicyclo[2.2.2]octane (DABCO) gave the C3-chlorinated intermediate 223, which, upon exposure to aluminum chloride and prenol, gave the reverse prenylated oxindole 225 in 54% yield (2.5:1, dr). Oxindole 225 was converted to deacetyl-fructigenine A (226) over five steps. Acetylation of indoline 226 furnished the natural product (+)-fructigenine A (217). To prepare penicimutanolone (218), the authors began with an electro-oxidative dearomatization reaction of N-Cbz-tyrosine gave spirocycle 228 in 85% yield. A series of oxidation, reduction, and epoxidation reactions allowed the conversion of spirocycle 228 to bisoxirane 229. Addition of an acetone enolate to ketone 229 resulted in alcohol 230 in 72% yield, which, upon hydrogenation, condensation with carboxylic acid 231, and desilylation, provided penicimutanolone (218) in 47% yield overall. The condensation of deacetyl-fructigenine A (226) with penicimutanolone (218) led to the convergent synthesis of (−)-penicimutanin A (219) in 26% yield.

Scheme 25.

Total synthesis of (−)-penicimutanin A (219). Conditions: (a) NCS, DABCO; AlCl₃, prenol, 54%. (b) PTSA, 26% (77% BRSM). (c) Ac₂O, 56%. (d) C(+):Pt(−), PhI, I=10mA, TFE, LiClO₄, 85%. (e) LDA, CeCl₃, acetone, 72%. (f) Pd/C, H₂; (COCl)₂, 231; TBAF, AcOH, 47% over 3 steps.

Conclusion

As illustrated by the representative studies discussed above, there have been significant advances in the total synthesis of DKP natural products in recent years. Many of the syntheses highlighted in this review have looked to nature and biosynthetic hypotheses for inspiration and identification of exciting new synthetic strategies. These innovative syntheses have enabled access to precious samples of these alkaloids, allowing their detailed study and analysis, in many cases leading to structure refinement or revision. Given the range of potent biological activities observed from complex DKP natural products, total chemical synthesis of complex DKPs and their derivatives continues to be important in identification of lead compounds with translational potential.

Figure 10.

(+)-azonazine (191).

Figure 12.

(−)-drimentines A (208), G (209), F (210) and (+)-indotertine A (211).