Chad A. Lewis

Scalable Total Syntheses of N-Linked Tryptamine Dimers by Direct Indole−Aniline Coupling: Psychotrimine and Kapakahines B and F

This report details the invention of a method to enable syntheses of psychotrimine (1) and the kapakahines F and B (2, 3) on a gram scale and in a minimum number of steps. Mechanistic inquiries are presented for the key enabling quaternization of indole at the C3 position by electrophilic attack of an activated aniline species. Excellent chemo-, regio-, and diastereoselectivities are observed for reactions with o-iodoaniline, an indole cation equivalent. Additionally, the scope of this reaction is broad with respect to the tryptamine and aniline components. The anti-cancer profiles of 1-3 have also been evaluated.

Total Synthesis of Palau’amine

Polycyclic dimeric pyrrole-imidazole alkaloids such as palau’amine (1, Figure 1),1 axinellamine A (2),2 and massadine chloride (3)3 possess daunting structural and physical attributes, including nine or more nitrogen atoms, eight contiguous stereogenic centers, reactive (hemi)aminal moieties, oxidation-prone pyrroles, and highly polar, non-crystalline morphologies. Their unique structures have been the focus of numerous publications from many groups worldwide, and have led to notable advances in synthetic methodology.4 Among the more complex members of this class, only the axinellamines (e.g. 2)5 and the massadines (e.g. 3)6 have succumbed to total synthesis, aided by the invention of a highly chemoselective and controllable late-stage oxidation reaction. Figure 1 Selected pyrrole-imidazole alkaloids, and retrosynthetic analysis of palau’amine (1). Ar = 2-(4,5-dibromopyrrole). In contrast to its siblings (2 and 3), palau’amine (1) possesses a unique chemical challenge: one of the pyrrole-amide sidechains is embedded in an exquisite, hexacyclic core architecture which contains a highly strained trans-azabicyclo[3.3.0]octane substructure (unprecedented among natural products). This is undoubtedly a central reason why the synthesis of palau’amine (1) has thus far eluded organic chemists despite the dozens of Ph.D. theses7 and studies towards publications8 that have appeared since its isolation in 1993 and structural reassignment in 2007.1 Many well-founded and logical plans to secure the idiosyncratic trans-5,5 core of 1 in our laboratory resulted in unfortunate empirical realities. Presumably, the high degree of strain implicit in the hexacyclic architecture thwarted all attempts at a biomimetic closure (N14-C10 and N1-C6 simultaneously)4 or a stepwise closure (N14-C10 followed by N1-C6).9 The lessons learned during those initial attempts inspired an alternative strategy that ultimately led to the total synthesis of 1 presented herein. As depicted in Figure 1, our retrosynthetic analysis relied upon a speculation that hypothetical macrocycle 4, dubbed “macro-palau’amine”, would be a kinetically stable isomer of the natural product core found in 1. It was predicted that an irreversible transannular ring-chain tautomerization would convert 4 into its consitutional isomer 1 via a dynamic equilibrium involving amidine tautomer 4′. Handheld molecular models suggested that 4 might adopt a folded conformation wherein N14 and C10 would be in close proximity to facilitate such a ring closure. A conceptually related late-stage shift of topology between constitutional isomers through dynamic equilibration was a key design element of our recent synthesis of the kapakahines.10 As with 1, “macro palau’amine” (4) exhibits a high level of strain and was believed to be accessible via macrolactamization of the diamine derived from diazide 5. This intermediate was envisioned to arise from the SNAr of a pyrrole (or surrogate thereof) to the bromo-aminoimidazole 6. The total synthesis of 1, outlined in Scheme 1, commences with the readily-available cyclopentane core 7, an intermediate enlisted in the synthesis of the massadines and available in 19 steps from commercially available materials in 1% overall yield.6 Treatment of 7 with aqueous TFA unveiled aminoguanidine 8, which was directly converted in unprotected form to the hemiaminal 10 in 64% isolated yield (along with 17% recovered 8, 130 mg scale)11 using silver(II)-picolinate (9). It is notable that this oxidation reaction takes place with precise regioselectivity – no oxidation of the primary amine is observed under these acidic reaction conditions. Construction of the remaining 2-aminoimidazole took place in 65% yield (251 mg scale)11 to afford 11 using cyanamide in brine (sat. aq. NaCl), a solvent that minimizes displacement of the highly labile chlorine atom.3,6 Subsequent bromination using Br2 in a 1:1 mixture of TFA:TFAA delivered the desired 2-amino-4-bromoimidazole 6 in 54% yield (150 mg scale).11 The introduction of the pyrrole moiety proved challenging, as standard conditions to couple amines to aryl halides using transition metal catalysis failed to produce any detectible amounts of product (even on the Boc-shielded 2-amino-4-bromoimidazole derivatives). In principle, the inherent ambiphilicity of the 2-aminoimidazole could lend itself to a unique reactivity pattern, one that would allow for uncatalyzed nucleophilic attack on the 2-amino-4-bromoimidazole as a possible direct route to the pyrrole-acid intermediate 5. Scheme 1 Total synthesis of palau’amine (1). Counterions are CF3CO2− and are omitted for clarity. Reagents and conditions: a) TFA/H2O (1/1), 50 °C, 12 h, then silver(II)picolinate (2.4 equiv), TFA/H2O (1/9), 23 °C, 5 min, 64% + ... In the event, the nucleophilic pyrrole surrogate 1212 was reacted with 2-amino-4-bromoimidazole 6 buffered with AcOH, followed by treatment with TFA, to deliver the desired N-coupled pyrrole-2-carboxylic acid 5 in a one-pot operation in 44% yield (91 mg scale).11 Presumably, facile N–C bond formation is observed due to the high reactivity of its tautomeric amidine form (6′). This reaction appears to be general and its scope will be reported in the full account of this work. The pyrrole-forming step, mediated by TFA and traversing through oxonium 14, involves no less than five chemical transformations occurring in tandem to deliver 5. In preparation for the key macrolactamization step, the azide groups of 5 were reduced to afford highly polar diamine 15 (4.0 mg scale). The synthesis of “macro-palau’amine” 4 was effected using EDC and HOBt. Heating of the crude reaction mixture in TFA (70 °C) elicited the crucial transannular cyclization (presumably proceeding via amidine tautomer 4′) that fastened the remaining two stereocenters and cemented the hallmark trans-5,5 ring system to deliver palau’amine (1) in 17% overall yield from 5 (one-pot, average of 55% per operation) after repeated purification with reverse phase HPLC (spectroscopically identical to that reported for 1 with the exception of optical rotation).13 Optimization and mechanistic investigation of this final sequence (5 Π 1) is currently underway.9 The journey to 1 (25 steps from commercial material, 0.015% overall yield with current procedures)9 has led not only to useful strategies and methods, but also to an empirical demonstration of numerous guiding principles for synthesis design at the frontiers of chemical complexity.14 Over six years ago our lab embarked on the synthesis of dimeric pyrrole-imidazole alkaloids by methodically applying the logic of biosynthesis where appropriate during the syntheses of sceptrin, oxysceptrin, nakamuric acid, ageliferin, nagelamide, the axinellamines (e.g. 2), and the massadines (e.g. 3).5,6,15 The synthesis of 1 benefited from a tremendous amount of chemical reactivity learned during those endeavours. Our 2004 biosynthetic hypothesis15b led us to pursue the true structure of 1 prior to the realization of its revised structure.1 In an effort to apply redox economic principles16 to this chemical synthesis program, a late-stage, chemoselective, silver-mediated oxidation was invented to circumvent laborious routes to the key hemiaminal unit expressed in 1–3 (C–20, Figure 1). Cascade reactions were incorporated to rapidly assemble complexity (e.g. 6 Π 5 Π 1). Finally, innate reactivity was embraced so as to minimize the use of redundant and orthogonal protecting group operations,17 and instead maximize the discovery of interesting chemical reactivity such as the direct coupling of nucleophiles to unprotected 2-amino-4-bromoimidazoles. An enantioselective, scalable variant of the current synthesis, as well as a full account of this work will be forthcoming.

Enantiospecific Total Syntheses of Kapakahines B and F

Gram-scale, enantiospecific total syntheses of the anti-leukemia marine natural products kapakahines B and F have been completed. Two powerfully simplifying transformations have enabled the gram-scale synthesis of this natural product family: (1) an oxidative N-C bond-forming reaction that sets a key quaternary center with complete stereocontrol and (2) a unique strategy for the construction of twisted macrocycles via dynamic equilibration of complex late-stage intermediates.

A Case of Remote Asymmetric Induction in the Peptide-Catalyzed Desymmetrization of a Bis(phenol)

We report a catalytic approach to the synthesis of a key intermediate on the synthetic route to a pharmaceutical drug candidate in single enantiomer form. In particular, we illustrate the discovery process employed to arrive at a powerful, peptide-based asymmetric acylation catalyst. The substrate this catalyst modifies represents a remarkable case of desymmetrization, wherein the enantiotopic groups are separated by nearly a full nanometer, and the distance between the reactive site and the pro-stereogenic element is nearly 6 A. Differentiation of enantiotopic sites within molecules that are removed from the prochiral centers by long distances presents special challenges to the field of asymmetric catalysis. As the distance between enantiotopic sites increases within a substrate, so too may the requirements for size and complexity of the catalyst. The approach presented herein contrasts enzymatic catalysts and small-molecule catalysts for this challenge. Ultimately, we report here a synthetic, miniaturized enzyme mimic that catalyzes a desymmetrization reaction over a substantial distance. In addition, studies relevant to mechanism are presented, including (a) the delineation of structure-selectivity relationships through the use of substrate analogs, (b) NMR experiments documenting catalyst-substrate interactions, and (c) the use of isotopically labeled substrates to illustrate unequivocally an asymmetric catalyst-substrate binding event.

Direct, chemoselective N-tert-prenylation of indoles by C-H functionalization.

Prenylated indole alkaloids have long been targets for total synthesis, possessing a broad range of medicinal properties and intriguing architectures.1 Our interest in this family began with the stephacidin family of indole alkaloids2 during which the fortuitous finding shown in Figure 1A was made. Thus, in 2003, during an attempt to convert N-Boc-tryptophan methyl ester (1) to the C-2 prenylated tryptophan 2 directly using 2-methyl-2-butene via electrophilic palladation3 and olefin capture, we instead observed small amounts (<10%) of a non-polar compound identified as N-tert-prenylated indole 3. Whereas many elegant methods have been invented for accomplishing the direct prenylation of indoles,4–7 no methods currently exist for the direct N-tert-prenylation of indoles (Figure 1B).8 Inspired by our initial findings (Figure 1A), this communication delineates a mild, highly chemoselective, scalable, and one-step route to these biologically relevant motifs via C–H functionalization. Figure 1 Inspiration from a failed campaign in the stephacidin total synthesis (A), currently known direct prenylation modalities (B), and the known route to N-tert-prenyl indoles compared to the C–H functionalization alternative (C). Boc = tert-butoxycarbonyl, ... The only known route to N-tert-prenylated indoles requires a four step sequence, three of which involve non-strategic redox-fluctuations9 (Figure 1C): 1) reduction of the indole to the indoline, 2) propargyl substitution via CuI-catalysis, 3) oxidation back to the indole, 4) and finally Lindlar reduction of the alkyne to the olefin. This chemistry has been successfully incorporated into a number of total syntheses.9 Building off of our initial observations (Figure 1A), we envisioned a direct, one-step procedure to synthesize N-tert-prenylated indoles without the use of pre-functionalized starting materials and superfluous redox steps – well known tenets of C–H functionalization logic.10 Such a strategy would be orthogonal to routes that involve nucleophilic prenylation that in this case would not be applicable.11 Specifically, C–H activation of indoles are known to occur at C-2 or C-3, involving the direct coupling of arenes12 and electron-deficient olefins13 and annulations,2,9,14 even in the presence of free N-H indoles.15 Pd-catalyzed allylic C–H aminations16 have been accomplished; however, there have been only a few reports involving the intermolecular coupling of amines and allylic olefins.17 The fundamental mechanistic insights and reaction designs of Stahl18a and Yu19 were instrumental in transforming our esoteric 2003 observations into a useful method for synthesis. Selected results of extensive optimization with the N-Phth-tryptophan methyl ester 4 are outlined in Table 1. The work of Stahl pointed us to the use of CH3CN as a solvent while research from the Yu lab inspired the use of Cu(OAc)2 and AgOTf in concert. Ultimately, we found that 40 mol% of a Pd source (either 40 mol% Pd(OAc)2 or 20 mol% Pd2dba3•CHCl3) with 30 eq of 2-methyl-2-butene in the presence of Cu(OAc)2 and an AgI source (AgOTf or AgTFA) as the co-oxidants in CH3CN was optimal for this transformation. Table 1 Optimization of the direct indole N-tert-prenylation. With these conditions in hand, the synthetic utility could be immediately demonstrated by applying it to known intermediates in total synthesis (Figure 2). For example, compound 3, an intermediate in the okaramine N synthesis9c requiring the aforementioned four step sequence (50% overall yield), could be obtained in 66% yield on a gram-scale and in a single step with no other regioisomers observed under these conditions. Similarly, N-Cbz-tryptophan methyl ester (6) was converted to 7, an intermediate towards the synthesis of the rufomycins,9b in 61% yield (gram-scale) as compared to 60% over 4 steps. Indole 3-carboxaldehyde (8) was prenylated to give 9 on a gram-scale, an intermediate in a cyclomarin synthesis,9d in 68% yield versus 70% (4-step sequence from indoline). The use of methyl acrylate18a and 3-NO2-pyridine12b,14 (possibly as stabilizing ligands of Pd0) was needed when an electron-withdrawing group was occupied at C-3. The natural product 11, isolated from Aporpium caryae,20 has been synthesized previously starting from indoline9a and indole9e in 60% yield over 5 steps and 34% over 7 steps, respectively. The current approach starts from commercially available methyl indole 3-carboxylate (10) and leads to 11 in a single step in 83% yield (gram-scale). Figure 2 Scalable, one-step routes to previously employed N-tert-prenyl indole intermediates. As delineated in Table 1, this mild reaction exhibits broad functional group tolerance. For example, tryptophan derivatives with various protecting groups can be prenylated (Table 2, 13a–i). Peptides containing tryptophan also undergo prenylation (13a, 13h), including a tripeptide (13b). The presence of amides, particularly at the tryptophan nitrogen (13d, 13h), reduces the reactivity, possibly due to ligation of the amides onto electrophilic PdII. However, a sterically hindered amide substrate does lead to an increased yield (13e) compared to other amide substrates. A tryptamine derivative (13j) is also prenylated, albeit in lower yield, but starting material can be recovered. Free alcohol, acid, and protected phenols substrates are well tolerated under the reaction conditions (13i, 13k, and 13m respectively). Halogenated substrates (13f–g), including those incorporated in the indole ring (13l) work well, and are not oxidatively cleaved under these conditions. Table 2 Scope of N-tert-prenylation. To gain insight into the mechanism of this transformation, we studied the interactions that Pd may have with both the indole and the olefin. We initially believed that the indole was being palladated at C-2, thus providing proximal delivery of the prenyl group to nitrogen.3,13 When a methyl group occupied the C-2 position, there was less than 5% conversion to the desired product (Figure 3A). However, when deuterium was placed on C-2 (16), we found that the deuterium was fully incorporated in product 17. We also found that 1,1,1-d3-3-methyl-2-butene reacted at the same rate as its protio isomer with PdII, consistent with Bercaw’s mechanistic studies of allylic C–H activation (the C-H activation step is not rate-determining).21 Figure 3 Mechanistic probe experiments (A), and postulated mechanism (B). Based on these results, we have proposed the following mechanism (Figure 3B). First, C–H activation of the olefin with PdII gives intermediate 18. The palladated olefin reacts with indole in three possible modes. The first involves direct coordination of the indole nitrogen to Pd leading to 19. Alternatively, indole palladation at C-3 takes place to give 20 which rearranges to the product by a metallo-Clasien rearrangement.22 Finally, a Wacker-type mechanism (21) may also be operative.16a All three intermediates are plausible, although intermediate 20 may explain the regioselectivity more convincingly. The Pd0 species then undergoes oxidation with AgI and CuII to close the cycle. In order to rule out initial allylic oxidation of the olefin to prenyl acetate a control experiment was performed using prenyl acetate in place of 2-methyl-2-butene (following Method A). The reaction did not provide the desired product, but rather produced 22 in ca. 10% yield (tentative assignment). The use of cis-butene substituted for 2-methyl-2-butene did afford desired products in decent yield (23, 24). This method is obviously not without limitations. For example, terminal olefins provided only enamine products. Lowering the loading of palladium to 10–20% or reducing the equivalents of olefin added (5 equiv.) led to diminished yields (<30%). With regard to indole substrate scope, substitution at C–3 is required, diketopiperazine ring systems give lower yields of product, and pyrroles are too reactive under these conditions. While the current method is by no means atom-economic, it is direct, scalable, and selective even on complex substrates. It compares favorably to the only other method known for the synthesis of these compounds.23 Finally, our attempts to utilize prenyl acetate (or similar derivatives) in concert with various transition metals has not led to indole N-tert-prenylation.24 In conclusion, a simplified, redox-conserving route to N-tert-prenylated indoles using PdII-mediated C–H functionalization has been developed. Contrary to the known reactivity of indoles, prenylation occurs exclusively at N-1. Although substitution at C-2 hindered reactivity and the use of stoichiometric AgI and CuII salts are required, this method is amenable to gram-scale synthesis using a variety of indoles, including formal syntheses of a number of natural products and the synthesis of antifungal natural product 11 in a single step. The high level of chemoselectivity exhibited in this reaction bodes well for further applications in both the early and advanced stages of prenylated indole total synthesis endeavors.

An approach to the site-selective diversification of apoptolidin A with peptide-based catalysts.

We report the application of peptide-based catalysts to the site-selective modification of apoptolidin A (1), an agent that displays remarkable selectivity for inducing apoptosis in E1A-transformed cell lines. Key to the approach was the development of an assay suitable for the screening of dozens of catalysts in parallel reactions that could be conducted using only microgram quantities of the starting material. Employing this assay, catalysts (e.g., 11 and ent-11) were identified that afforded unique product distributions, distinct from the product mixtures produced when a simple catalyst (N,N-dimethyl-4-aminopyridine (10)) was employed. Preparative reactions were then carried out with the preferred catalysts so that unique, homogeneous apoptolidin analogues could be isolated and characterized. From these studies, three new apoptolidin analogues were obtained (12-14), each differing from the other in either the location of acyl group substituents or the number of acetate groups appended to the natural product scaffold. Biological evaluation of the new apoptolidin analogues was then conducted using growth inhibition assays based on the H292 human lung carcinoma cell line. The new analogues exhibited activities comparable to apoptolidin A.

Functional Characterization of the Cyclomarin/Cyclomarazine Prenyltransferase CymD Directs the Biosynthesis of Unnatural Cyclic Peptides⊥

In vitro and in vivo characterization of the cyclomarin/cyclomarazine prenyltransferase CymD revealed its ability to prenylate tryptophan prior to incorporation into both cyclic peptides by the nonribosomal peptide synthetase CymA. This knowledge was utilized to bioengineer novel derivatives of these marine bacterial natural products by providing synthetic N-alkyl tryptophans to a prenyltransferase-deficient mutant of Salinispora arenicola CNS-205.

Catalytic site-selective synthesis and evaluation of a series of erythromycin analogs.

The generation of a series of analogs of erythromycin A (EryA, 2) is described. In this study, we compared two peptide-based catalysts-one originally identified from a catalyst screen (5) and its enantiomer (ent-5)-for the selective functionalization of EryA. The semi-synthetic analogs were subjected to MIC evaluation with two bacterial strains and compared to unfunctionalized EryA.