Article | Published:

Enantioselective cyclizations and cyclization cascades of samarium ketyl radicals

Nature Chemistry volume 9, pages 11981204 (2017) | Download Citation

Abstract

The rapid generation of molecular complexity from simple starting materials is a key challenge in synthesis. Enantioselective radical cyclization cascades have the potential to deliver complex, densely packed, polycyclic architectures, with control of three-dimensional shape, in one step. Unfortunately, carrying out reactions with radicals in an enantiocontrolled fashion remains challenging due to their high reactivity. This is particularly the case for reactions of radicals generated using the classical reagent, SmI2. Here, we demonstrate that enantioselective SmI2-mediated radical cyclizations and cascades that exploit a simple, recyclable chiral ligand can convert symmetrical ketoesters to complex carbocyclic products bearing multiple stereocentres with high enantio- and diastereocontrol. A computational study has been used to probe the origin of the enantioselectivity. Our studies suggest that many processes that rely on SmI2 can be rendered enantioselective by the design of suitable ligands.

  • Compound

    methyl 2-acetyl-2-(but-3-en-1-yl)hex-5-enoate

  • Compound

    methyl (1S,2R,3S)-1-(but-3-en-1-yl)-2-hydroxy-2,3-dimethylcyclopentane-1-carboxylate

  • Compound

    methyl (1S,2R,3S)-1-((E)-hex-3-en-1-yl)-2-hydroxy-2-methyl-3-propylcyclopentane-1-carboxylate

  • Compound

    methyl (1S,2R,3S)-1-((Z)-hex-3-en-1-yl)-2-hydroxy-2-methyl-3-propylcyclopentane-1-carboxylate

  • Compound

    methyl (1S,2R,3R)-2-hydroxy-3-isopropyl-2-methyl-1-(4-methylpent-3-en-1-yl)cyclopentane-1-carboxylate

  • Compound

    methyl (1S,2R,3R)-2-hydroxy-2-methyl-1-(penta-3,4-dien-1-yl)-3-vinylcyclopentane-1-carboxylate

  • Compound

    methyl (1S,2R,3R)-1-((E)-5-acetoxypent-3-en-1-yl)-2-hydroxy-2-methyl-3-vinylcyclopentane-1-carboxylate

  • Compound

    methyl (1R,2R)-1-(buta-2,3-dien-1-yl)-2-hydroxy-2-methyl-3-methylenecyclopentane-1-carboxylate

  • Compound

    isopropyl (1S,2R,3S)-1-(but-3-en-1-yl)-2-hydroxy-2,3-dimethylcyclopentane-1-carboxylate

  • Compound

    (1S,2R,3S)-1-(but-3-en-1-yl)-2-hydroxy-N,N,2,3-tetramethylcyclopentane-1-carboxamide

  • Compound

    methyl (1S,5S,8R)-8-hydroxy-8-methylbicyclo[3.2.1]octane-1-carboxylate

  • Compound

    methyl (1S,3aR,5S,6aR)-1-(but-3-en-1-yl)-6a-hydroxy-5-methyloctahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,5S,6aR)-1-(but-3-en-1-yl)-6a-hydroxy-5-isopropyloctahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,5S,6aR)-1-(but-3-en-1-yl)-6a-hydroxy-5-(3-phenylpropyl)octahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,4S,5R,6aR)-4-ethyl-1-((E)-hex-3-en-1-yl)-6a-hydroxy-5-methyloctahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,6aR)-1-(but-3-en-1-yl)-6a-hydroxy-5,5-dimethyloctahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,5S,6aR)-1-(but-3-en-1-yl)-6a-hydroxy-5-vinyloctahydropentalene-1-carboxylate

  • Compound

    methyl (1S,3aR,4R,5S,6aR)-6a-hydroxy-5-methylhexahydro-1,4-ethanopentalene-1(2H)-carboxylate

  • Compound

    methyl (1S,3aR,4R,5R,6aR)-6a-hydroxy-5-isopropylhexahydro-1,4-ethanopentalene-1(2H)-carboxylate

  • Compound

    methyl (1S,3aR,4R,5S,6aR)-6a-hydroxy-5-(3-phenylpropyl)hexahydro-1,4-ethanopentalene-1(2H)-carboxylate

  • Compound

    (1R,2R)-1,2-diphenylethane-1,2-diol

  • Compound

    (1R,1'R)-2,2'-(1,2-phenylenebis(oxy))bis(1-phenylethan-1-ol)

  • Compound

    (1R,1'R)-2,2'-((1,2-phenylenebis(methylene))bis(oxy))bis(1-phenylethan-1-ol)

  • Compound

    ((2S,2'S)-ethane-1,2-diylbis(pyrrolidine-1,2-diyl))bis(diphenylmethanol)

  • Compound

    (1R,1'R)-2,2'-(ethane-1,2-diylbis(benzylazanediyl))bis(1-phenylethan-1-ol)

  • Compound

    (1R,1'R)-2,2'-(benzylazanediyl)bis(1-phenylethan-1-ol)

  • Compound

    (R)-N-benzyl-2-methoxy-N-((R)-2-methoxy-2-phenylethyl)-2-phenylethan-1-amine

  • Compound

    (1R,1'R)-2,2'-(neopentylazanediyl)bis(1-phenylethan-1-ol)

  • Compound

    (1R,1'R)-2,2'-(benzylazanediyl)bis(1-(naphthalen-1-yl)ethan-1-ol)

  • Compound

    (1R,1'R)-2,2'-(benzylazanediyl)bis(1-(3,5-dimethylphenyl)ethan-1-ol)

  • Compound

    ((1R,5S,8R)-8-hydroxy-8-methylbicyclo[3.2.1]octan-1-yl)methyl 4-bromobenzoate

Main

The enantioselective construction of all-carbon cyclic and polycyclic arrays is of great importance due to the ubiquity of such motifs in bioactive natural products and the role that these three-dimensional scaffolds play in inspiring modern drug design and the provision of molecular probes for biology1,2,3. Due to the difficulty of forming C–C bonds between sterically crowded reacting sites and the need to control the stereochemical outcome of such processes, accessing densely functionalized all-carbon molecular architectures is a major challenge in synthetic science. As reactions of open-shell reactive intermediates are often exothermic and proceed through early transition states, long incipient C–C bonds can mean that steric impedance can be overcome in difficult C–C couplings, and methods involving radicals have thus emerged for the construction of targets bearing contiguous stereocentres in crowded acyclic and cyclic settings4,5,6. On the other hand, this high reactivity makes the development of enantioselective radical reactions difficult, and creative attempts to surmount this challenge have led to key breakthroughs in synthesis7,8. For example, the use of chiral Lewis acids was established in seminal studies by Porter and Sibi and constitutes a robust strategy for selective radical conjugate additions and atom transfer reactions involving two-point-binding substrates9,10,11. More recently, enantioselective radical approaches exploiting various ingenious reactivity manifolds have been devised, including the synergistic combination of Lewis acid catalysis12 or organocatalytic methods13,14,15 with photoinduced electron transfer, the use of chiral thiols as radical16 or hydrogen atom-donating17 catalysts, and transition metal-catalysed electron transfer processes18,19,20,21. In particular, Yoon22, Streuff 23 and Knowles24 have recently described reductive ketyl generation and enantioselective C–C bond formation.

Despite these exciting advances, multiple-ring-forming, enantioselective cascade reactions triggered by radical generation remain rare. To our knowledge, the only three examples were reported in 2006 by Yang and involve atom transfer, radical–radical cyclization cascades initiated by Et3B/O2 (Fig. 1a)25. Related to the chiral Lewis acid approach of Porter and Sibi, the method grants access to all-carbon bicyclic scaffolds from organoselenium precursors with moderate to high enantioselectivity. In Yang's seminal study, a stoichiometric amount of chiral reagent was found to deliver optimal enantioselectivities. Thus, the great synthetic potential of enantioselective radical cyclization cascades remains largely unexplored. The exploitation of new activation modes in the development of enantioselective radical cyclization cascades is therefore an important and timely goal that promises to greatly expand the palette of synthetic methods available for the streamlined preparation of complex molecules.

Figure 1: Reagent-controlled enantioselective radical C–C bond-forming cyclizations.
Figure 1

Shown is an example of an enantioselective radical cascade reaction, an attempt to control the enantioselectivity of a SmI2-mediated intramolecular C–C bond-forming reaction and our strategy to achieve enantiocontrol in both radical cyclization and cyclization cascade processes mediated by SmI2. a, Enantioselective group transfer radical C–C bond-forming cyclization cascades of phenylselenyl ketoesters mediated by a chiral Lewis acid (ref. 25). b, An attempted enantioselective radical C–C bond-forming pinacol-type cyclization mediated by SmI2 (ref. 35). c, This work involves the SmI2-mediated enantioselective desymmetrizing radical C–C bond-forming cyclizations and cyclization cascades of unsaturated ketoesters using a readily available and recyclable chiral ligand. The two-electron processes convert simple substrates to complex products containing up to two new rings and five new stereocentres. Computational studies have been used to probe the origin of the selectivity observed.

A mild and widely used method for radical initiation relies on the use of single electron transfer (SET) reductants to generate nucleophilic ketyl radical anions from carbonyl compounds, thus allowing a formal umpolung (polarity reversal) of the carbonyl26,27. Of particular note, such ketyls readily add to olefins intramolecularly28, forging new C–C bonds and delivering cycloalkanols rich in stereochemistry. The commercially available reagent samarium diiodide (SmI2, Kagan's reagent29) is arguably the most efficient promoter of such reductive C–C couplings, as evidenced by its pivotal use in numerous high-profile total syntheses30,31. The many outstanding features of SmI2 include its high reducing ability, allowing many classes of carbonyl substrate to be converted to ketyls, its high oxophilicity, often leading to exquisite control of relative stereochemistry via chelated transition states, and the potential to fine-tune reactivity and selectivity using various additives28,32. Unfortunately, in the four decades since its introduction to organic synthesis, attempts to use SmI2 in reagent-controlled enantioselective radical C–C couplings have met with little success. In fact, only one isolated study by Mikami33 describes the enantioselective intermolecular C–C coupling of samarium ketyls, derived from aryl ketones, with acrylates using SmI2 and the ligand (R)-BINAPO34. Unfortunately, the process was limited in scope and gave low diastereoselectivities and yields. Examples of enantioselective intramolecular C–C coupling are even more elusive; for example, an attempted samarium ketyl cyclization by Skrydstrup using an enantiopure bisphosphoramide ligand proceeded with little enantiocontrol (Fig. 1b)35. No examples of enantioselective C–C bond-forming cascade cyclizations using SmI2 have been described.

Here, we describe enantioselective desymmetrizing ketyl-alkene radical cyclizations and cyclization cascades of dienyl β-ketoesters mediated by an in situ generated chiral Sm(II) reagent. Key to the success of this process is the use of a readily available and recyclable, enantiopure tripodant aminodiol and an achiral alcohol additive, in conjunction with SmI2 (Fig. 1c). The resulting samarium(II) reagent effectively triggers radical cyclizations via chelated Sm(III) ketyl intermediates I, desymmetrizing simple starting materials and delivering complex mono- and polycyclic all-carbon scaffolds containing up to five stereocentres and versatile alkenyl units for further elaboration, with high enantio- and diastereocontrol. A computational study has been used to explore the origin of enantioselectivity in the process.

Results and discussion

Reagent design

Given the lack of precedent for enantioselective SmI2-mediated transformations, several factors guided our reaction design. First, drawing on the seminal work of Molander36, a neighbouring Lewis basic ester group was used to coordinate to Sm(II), thus facilitating reduction and helping to control the stereochemical course of samarium(III) ketyl cyclizations through the provision of chelated transition states. Thus, readily available dienyl β-ketoesters, such as 1a, were selected as benchmark two-point-binding substrates for the development of enantioselective desymmetrizing ketyl radical cyclizations. Attractively, desymmetrization of dienyl β-ketoesters 1 allows complex products bearing multiple stereocentres and an alkenyl unit for further functionalization to be assembled rapidly. Second, and with regard to the choice of chiral ligand, to maximize the interaction of the substrate with the chiral Sm(III) species we chose to avoid the use of Lewis basic phosphoramide-type chiral ligands (cf. Fig. 1b), as Flowers has suggested that the cyclization of HMPA-bound samarium ketyls takes place via a solvent-separated ion pair37. Third, protonation of both a Sm(III)–O and a Sm(III)–C bond in intermediate III, formed after cyclization and further SET reduction of II, is required (cf. Fig 1c). Protonation of the Sm(III)–O bond prevents detrimental retro-aldol pathways of the product, but should be sufficiently slow as to not compromise the transfer of stereochemical information from the Sm(III)-ligand moiety in I in the desymmetrization step. We proposed that either a chiral or achiral protic additive could affect this task. The affinity of ethyleneglycol for SmI2 is well known38, and chiral C2-symmetrical diols have been used successfully for the enantioselective protonation of Sm(III) enolates39. Thus, we hypothesized that a flexible multidentate chiral diol would bind Sm(II), accommodate the change in ionic radius from Sm(II) to Sm(III) following SET to the substrate, potentially control asymmetry during cyclization of the Sm(III) ketyl radical, and act as a proton source in quenching the anions formed during the process.

The feasibility of the enantioselective samarium ketyl cyclization of 1a was assessed through the extensive screening of chiral diols 3 in conjunction with a slight excess of SmI2 in THF (2.2 equiv.), as summarized in Table 1 (see the e.r. values associated with the structures and entries 1–4). Early studies suggested that a 1:1 ratio between the chiral ligand and SmI2 gave the best enantioselectivities. From the outset, the crucial influence of temperature was clear, and significant enantioselectivity (up to 65:35 e.r. with 3f) was only obtained when the reaction was conducted at −40 °C (entry 2). At this temperature, acceptable conversion to 2a (≥70% in most cases) was also observed. Evans et al. have shown that the samarium(III) bis-alkoxide derived from 3f is a highly efficient catalyst for the enantioselective Meerwein–Ponndorf–Verley reduction of aryl ketones40, but there are no reported applications of the neutral form of the aminodiol in synthesis. Employing its dimethyl ether analogue 3g afforded racemic 2a, confirming the expected higher affinity of SmI2 for free hydroxyl-containing chiral ligands (entry 3). Similarly, replacing the N-benzyl moiety in the aminodiol with a neopentyl unit (ligand 3h, entry 4) resulted in a low-yielding cyclization with no asymmetric induction, thus underlining the importance of sterics in ligand binding to Sm(II).

Table 1: Optimization of the enantioselective, SmI2-mediated desymmetrizing ketyl cyclization of 1a.

We next investigated the impact of small amounts of achiral protic additives on the samarium ketyl cyclization of 1a. We reasoned that the additive would act as a proton source, in place of the chiral aminodiol, in the quenching of anionic intermediates, thus leaving the crucial chiral ligand and its coordination chemistry unaltered. Although adding non-coordinating alcohol additives was inefficient (entries 5 and 6), addition of MeOH (equimolar to SmI2) resulted in an enhancement of enantioselectivity and efficiency in the formation of 2a (entry 7). An erosion of asymmetric induction was observed when excess MeOH was present (entry 8), while the use of H2O as an additive resulted in near racemic 2a (entry 9). The negative impact of H2O probably arises from displacement of the chiral ligand from Sm(III) or fast protonation41 of the Sm(III) ketyl species involved in the enantiodetermining step. It is known that H2O exhibits a high affinity for SmI2 even at low concentrations42. Similarly, coordination of MeOH to Sm(III) ketyls has been used to rationalize the outcome of several reactions requiring fast proton transfer43,44. On the basis of these results, it is unlikely that the achiral additive acts solely as a proton source and may act as an additional ligand for samarium. Thus, the role of MeOH in improving enantioinduction in the ketyl cyclization is probably twofold: (1) the achiral alcohol acts as a sacrificial proton donor, thus preserving the integrity of the chiral aminodiol ligand; (2) MeOH binds to Sm(II) and/or Sm(III), affecting the coordination chemistry and the environment around the metal and producing a species that gives rise to higher enantioinduction.

A second round of ligand optimization was undertaken. Most analogues of 3f (readily prepared by epoxide aminolysis) displayed lower efficiency in the reaction (for example, entry 10), with the exception of 3j, bearing bulkier 3,5-dimethylphenyl substituents in place of the phenyl groups in 3f (entry 11). However, the use of 3j was detrimental to the diastereoselectivity of the cyclization. Focusing on the use of 3f, subtle adjustments to the protocol (for example, adding SmI2 then MeOH to the chiral ligand before slow cooling) prevented precipitation and resulted in improved yield and enantioselectivity (89:11 e.r.) (entry 12). Attempts to use co-solvents in conjunction with THF typically led to lower enantiocontrol; for example, performing the reaction in THF/toluene resulted in the efficient formation of 2a but with significantly lower enantio- and diastereocontrol. Crucially, ligand 3f is easy to prepare in one step on a multigram scale (65% yield) from inexpensive (R)-styrene oxide and benzylamine, and can be readily isolated after use and recycled (>90% recovery). In contrast to Evans’ use of the aminodiolate derived from 3f in a Sm(III)-catalysed asymmetric hydride transfer process, our studies employ 3f in its diol form in a Sm(II)/Sm(III)-mediated radical, C–C bond-forming process.

Enantioselective radical cyclizations

Having identified optimal conditions for enantioselective samarium ketyl cyclization using ligand 3f, we set out to assess the scope of the process using a range of unsaturated ketoester derivatives (Table 2). Pleasingly, enantioselective cyclization of disubstituted diene E-1b at −45 °C afforded cyclopentanol E-2b in 75% yield and 91:9 e.r. Interestingly, analogous diene Z-1b only underwent cyclization at −40 °C, giving Z-2b in 72% yield but with significantly lower selectivity (90:10 d.r., 87:13 e.r.). This probably arises from unfavourable steric interactions in the transition state of the cyclization (vide infra). Pleasingly, trisubstituted diene 1c underwent cyclization to give 2c with high enantioselectivity (91:9 e.r.) and as a single diastereoisomer. bis-Allenyl substrates 1d and 1f delivered vinyl-substituted cyclopentane 2d (78:22 e.r.) and methylene cyclopentane 2f (65:35 e.r., 2:1 mixture of exo and endo alkene regioisomers) in good yield but with modest enantiocontrol. A successful alternative approach for the formation of a vinyl-substituted cyclopentane product employed bis-allylic acetate 1e and a process terminated by anionic elimination; 2e was formed in 80% yield with high enantiocontrol (92:8 e.r.) and complete diastereocontrol. Surprisingly, switching from the methyl ester group in 1a (2a obtained in 89:11 e.r.) to the larger isopropyl ester group (in 1g) resulted in a significant drop in enantiocontrol (2g obtained in 79:21 e.r.): the large isopropyl group probably disrupts the all-important binding of the ester group and the chiral ligand to Sm(III). Furthermore, the attempted use of a more Lewis basic amide moiety in 1h resulted in the formation of racemic product 2h. The amide has a higher affinity for Sm(III) than the corresponding ester and its presence may also prevent or alter the coordination of the chiral aminodiol to Sm(III)36. We next examined the feasibility of an enantioselective transannular cyclization of cycloheptenyl methyl ketoester 1i. Pleasingly, bicyclo[3.2.1]octane 2i was obtained in 83% yield, high enantiocontrol (93:7 e.r.) and as a single diastereoisomer. To our knowledge, this reaction constitutes the first example of an enantioselective transannular radical reaction proceeding under reagent control. X-ray crystallographic analysis of a derivative allowed the absolute and relative stereochemistry of 2a to be assigned. X-ray crystallographic analysis of the 4-bromobenzoate benzoate 4 allowed the absolute and relative stereochemistry of 4 and 2i to be assigned. The absolute stereochemistry of 2b–h (and 2jr, vide infra) was inferred based on the assignment of 2a and 2i.

Table 2: Scope of the enantioselective, SmI2-mediated desymmetrizing ketyl olefin cyclization.

Model for stereoinduction

A model for the enantio- and diastereocontrol observed in the desymmetrizing samarium-ketyl cyclizations is shown in Fig. 2a. SmI2, aminodiol 3f and substrate 1a form a 1:1:1 complex giving model 5a. SET from Sm(II) in 5a gives a Sm(III)-ketyl (cf. I in Fig. 1c) that can react via its Re face (transition structure anti-6a) or its Si face (transition structure anti-6a′). It is well established that anti-modes of addition are typically favoured in ketyl-alkene cyclizations28. We propose that transition structure anti-6a′ is disfavoured due to steric interactions between the alkenyl side chain and the phenyl ring of the aminodiol ligand and also between the methyl group at the radical centre and the proximal hydroxyl of the aminodiol ligand (Supplementary Fig. 6). Ketyl-alkene cyclization therefore proceeds through anti–transition state 6a to give 2a with high control. Computational studies have been used to probe the enantioselective process and to validate the proposed model. To our knowledge, the study of Sm(II)-mediated C–C bond formation using computational chemistry has not previously been reported (Supplementary Information pages 70–74). Focusing on the enantioselective cyclizations of 1a and using 3f′ (N-Me) as a simplified model for ligand 3f (N-Bn), density functional calculations were performed to elucidate the nature of the electronic and geometric transformations involved in the cyclization process. The optimized structure of the Sm(III)-ketyl radical, 1a-rad, and the transition structures for the cyclizations anti-6a and anti-6a′, denoted as anti-[6a] and anti-[6a′], are shown in Fig. 2b. As expected, the anti-modes of cyclization (alkene orientated anti to the C–O bond of the ketyl) are favoured; for example, compare the energy of anti-[6a] (22.2 kJ mol−1) to that of syn-[6a] (31.2 kJ mol−1). As can be seen from the reaction profile for the cyclization of 1a-rad, there is a significant energetic preference for the formation of 2a (rather than ent-2a) via the transition structure anti-[6a]. Having shown the feasibility of modelling complex samarium-bound radicals and predicting the absolute stereochemistry of the products arising from enantioselective cyclization, future studies will develop new computational models that probe the important role of MeOH in the process and predict enantioselectivities in line with those observed experimentally. Crucially, these computational models will facilitate the rational design of new enantioselective processes involving low-valent lanthanides.

Figure 2: Origin of enantioselectivity in the cyclizations.
Figure 2

Shown is a working model for the active Sm(III)-ligand–substrate complex and the origin of asymmetric induction. a, The model suggests that anti-transition structure 6a is favoured over anti-transition structure 6a′ due to steric interactions between the alkenyl side chain and the phenyl ring of the aminodiol ligand and also between the methyl group at the radical centre and the proximal hydroxyl of the aminodiol ligand in the later transition structure. b, Density functional calculations exploring the nature of the electronic and geometric transformations involved in the enantioselective cyclization of 1a and supporting the proposed model for the origin of selectivity. Activation barriers (kJ mol−1) are given for two anti-transition structures and one syn-transition structure formed from 1a-rad. c, Optimized structures obtained at the PBE0/Def2-SVP level for 1a-rad, anti-[6a], anti-[6a′] and syn-[6a]. Hydrogen atoms have been omitted in all structures for clarity.

Enantioselective radical cascade cyclizations

We next sought to develop enantioselective desymmetrizing samarium-ketyl cyclization cascades that would deliver higher returns in terms of molecular complexity. Thus, allyl ketones 1j–r were prepared and exposed to SmI2, ligand 3f and MeOH in THF (Table 3). In the case of 1j, cis-octahydropentalene 2j bearing four stereocentres was formed in 75% yield. No monocyclization product was detected. Complete control of relative stereochemistry was observed in the first radical cyclization of the cascade, while the second radical ring closure proceeded with good selectivity (80:20 d.r., major diastereoisomer shown). Running the reaction at −50 °C resulted in full conversion of 1j, and 2j was obtained with enhanced enantioselectivity (93:7 e.r.) when compared to the related monocyclization (conversion of 1a to 2a). Radical cyclization of prenyl-substituted analogue 1k was next examined and was found to afford by-products arising from bimolecular disproportionation of the tertiary radical formed upon cascade cyclization45. Addition of excess 1,4-cyclohexadiene as a hydrogen atom source completely suppressed this unwanted pathway and isopropyl-substituted 2k was isolated in 73% yield with excellent diastereo- and enantiocontrol (92:8 d.r. and 98:2 e.r.). ((E)-5-phenylpent-2-enyl) precursor 1l gave 2l with similarly high control (94:6 d.r. and 95:5 e.r.), while efficient control over the formation of five contiguous stereocentres was achieved in the cascade cyclization of 1m (the major diastereoisomeric product 2m was obtained with 97:3 e.r.). The generation of a third fully substituted centre in the cascade proved more challenging, and higher temperature was required (−25 °C) for the transformation of ketoester 1n: gem-dimethyl product 2n was obtained in good yield but with moderate enantioselectivity (82:18 e.r.). Finally, 1o bearing an allylic leaving group gave vinyl-substituted product 2o after a highly selective cascade terminated by acetate elimination (94:6 d.r. and 93:7 e.r.). It should be noted that high sequence integrity was observed in all the radical cascade cyclizations, with only minor amounts of monocyclization by-products (<5%) detected by 1H NMR analysis of some crude product mixtures. Finally, an even higher degree of molecular complexity can be accessed from cycloheptene substrates 1p–r. For these substrates, tricyclic products 2pr were obtained in good yields and uniformly high enantioselectivity (94:6 e.r.), albeit with moderate control of relative stereochemistry in the second cyclization event in the cascade.

Table 3: Scope of the enantioselective, SmI2-mediated desymmetrizing ketyl-olefin cyclization cascade.

Conclusion

We have developed a new enantioselective SmI2-mediated radical cyclization. The combination of SmI2 with a simple and inexpensive, recyclable chiral aminodiol promotes the enantioselective desymmetrizing 5-exo ketyl-alkene cyclization of unsaturated ketoesters. Cyclizations typically proceed with high enantioselectivity (up to 92:8 e.r.) and diastereoselectivity (up to >99:1 d.r.). An analogous transannular variant delivers enantiomerically enriched bicyclic tertiary alcohol 2i (93:7 e.r. and >99:1 d.r.). Enantioselective, desymmetrizing cascade cyclizations mediated by SmI2 proceed with high sequence integrity and deliver even more complex molecular architectures: two carbocyclic rings and up to five contiguous stereocentres are formed with high enantiocontrol (>98:2 e.r.). A computational study has been used to probe the origin of enantioselectivity in the cyclizations.

Data availability

The X-ray crystallographic coordinates for a derivative of 2a (2a″) and 4 have been deposited at the Cambridge Crystallographic Data Centre (CCDC) under deposition nos. CCDC 1543525 and 1531051, respectively. These data can be obtained free of charge from the CCDC (www.ccdc.cam.ac.uk/data_request/cif). The authors declare that all other data supporting the findings of this study are available within this article and its Supplementary Information.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    & Pharmaceuticals that contain polycyclic hydrocarbon scaffolds. Chem. Soc. Rev. 44, 7737–7763 (2015).

  2. 2.

    et al. Natural product-inspired cascade synthesis yields modulators of centrosome integrity. Nat. Chem. Biol. 8, 179–184 (2012).

  3. 3.

    & Medicinal chemistry of polycyclic cage compounds in drug discovery research. Med. Chem. Res. 17, 137–151 (2008).

  4. 4.

    & Radicals in Organic Synthesis (Wiley, 2001).

  5. 5.

    , & Stereochemistry of Radical Reactions: Concepts, Guidelines, and Synthetic Applications (Wiley, 2008).

  6. 6.

    & Encyclopedia of Radicals in Chemistry, Biology and Materials (Wiley, 2012).

  7. 7.

    & Stereoselective radical reactions. Chem. Soc. Rev. 32, 251–263 (2003).

  8. 8.

    , , & Progress in enantioselective radical cyclizations. Chem. Eur. J. 23, 6225–6236 (2017).

  9. 9.

    & Enantioselective free radical reactions. Acc. Chem. Res. 32, 163–171 (1999).

  10. 10.

    , & Enantioselective radical processes. Chem. Rev. 103, 3263–3296 (2003).

  11. 11.

    , , & Enantioselective cascade radical addition–cyclization–trapping reactions. Angew. Chem. Int. Ed. 45, 5863–5866 (2006).

  12. 12.

    & Enantioselective Lewis acid catalysis of intramolecular enone [2+2] photocycloaddition reactions. Science 342, 840–843 (2013).

  13. 13.

    & Organocatalysis in radical chemistry. Enantioselective α-oxyamination of aldehydes. J. Am. Chem. Soc. 129, 4124–4125 (2007).

  14. 14.

    & Merging photoredox catalysis with organocatalysis: the direct asymmetric alkylation of aldehydes. Science 322, 77–80 (2008).

  15. 15.

    , , & Photochemical activity of a key donor–acceptor complex can drive stereoselective catalytic α-alkylation of aldehydes. Nat. Chem. 5, 750–756 (2013).

  16. 16.

    , & An organic thiyl catalyst for enantioselective cyclization. Nat. Chem. 6, 702–705 (2014).

  17. 17.

    , & Enantioselective synthesis of an ophiobolin sesterterpene via a programmed radical cascade. Science 352, 1078–1082 (2016).

  18. 18.

    , , & A catalytic enantioselective electron transfer reaction: titanocene-catalyzed enantioselective formation of radicals from meso-epoxides. Angew. Chem. Int. Ed. 38, 2909–2910 (1999).

  19. 19.

    et al. Enantioselective cyanation of benzylic C–H bonds via copper-catalyzed radical relay. Science 353, 1014–1018 (2016).

  20. 20.

    & Enantioselective functionalization of radical intermediates in redox catalysis: copper-catalyzed asymmetric oxytrifluoromethylation of alkenes. Angew. Chem. Int. Ed. 52, 12655–12658 (2013).

  21. 21.

    & Catalytic enantioselective alkene aminohalogenation/cyclization involving atom transfer. Angew. Chem. Int. Ed. 51, 3923–3927 (2012).

  22. 22.

    , & Enantioselective photocatalytic [3+2] cycloadditions of aryl cyclopropyl ketones. J. Am. Chem. Soc. 138, 4722–4725 (2016).

  23. 23.

    , , , & Enantioselective titanium(III)-catalyzed reductive cyclization of ketonitriles. Angew. Chem. Int. Ed. 51, 8661–8664 (2012).

  24. 24.

    , , , & Enantioselective photoredox catalysis enabled by proton-coupled electron transfer: development of an asymmetric aza-pinacol cyclization. J. Am. Chem. Soc. 135, 17735–17738 (2013).

  25. 25.

    , , , & Enantioselective PhSe-group-transfer tandem radical cyclization reactions catalyzed by a chiral Lewis acid. Angew. Chem. Int. Ed. 45, 255–258 (2006).

  26. 26.

    Carbonyl-coupling reactions using low-valent titanium. Chem. Rev. 89, 1513–1524 (1989).

  27. 27.

    Application of lanthanide reagents in organic synthesis. Chem. Rev. 92, 29–68 (1992).

  28. 28.

    , , & Cross-coupling reactions using samarium(II) iodide. Chem. Rev. 114, 5959–6039 (2014).

  29. 29.

    , & Divalent lanthanide derivatives in organic synthesis. 1. Mild preparation of samarium iodide and ytterbium iodide and their use as reducing or coupling agents. J. Am. Chem. Soc. 102, 2693–2698 (1980).

  30. 30.

    , & Samarium diiodide mediated reactions in total synthesis. Angew. Chem. Int. Ed. 48, 7140–7165 (2009).

  31. 31.

    , & Samarium(II)-iodide-mediated cyclizations in natural product synthesis. Chem. Rev. 104, 3371–3404 (2004).

  32. 32.

    , & Organic Synthesis using Samarium Diiodide (RSC Publishing, 2009).

  33. 33.

    & Chiral ligand control in enantioselective reduction of ketones by SmI2 for ketyl radical addition to olefins. Tetrahedron Lett. 39, 4501–4504 (1998).

  34. 34.

    , & Diastereo- and enantioselective hydrodimerization of β-monosubstituted acrylic acid amides induced by chiral samarium(II) complexes. Tetrahedron Lett. 40, 7497–7500 (1999).

  35. 35.

    , & Studies on the SmI2-promoted pinacol-type cyclization: synthesis of the hexahydroazepine ring of balanol. J. Org. Chem. 65, 5382–5390 (2000).

  36. 36.

    . & Intramolecular reductive coupling reactions promoted by samarium diiodide. J. Am. Chem. Soc. 111, 8236–8246 (1989).

  37. 37.

    , , & Mechanistic study of samarium diiodide-HMPA initiated 5-exo-trig ketyl–olefin coupling: the role of HMPA in post-electron transfer steps. J. Am. Chem. Soc. 130, 7228–7229 (2008).

  38. 38.

    , , & . Mechanistic studies of proton-donor coordination to samarium diiodide. Angew. Chem. Int. Ed. 46, 8160–8163 (2007).

  39. 39.

    et al. Enantioselective protonation of samarium enolates derived from α-heterosubstituted ketones and lactone by SmI2-mediated reduction. Tetrahedron 55, 4595–4620 (1999).

  40. 40.

    , , & A chiral samarium-based catalyst for the asymmetric Meerwein–Ponndorf–Verley reduction. J. Am. Chem. Soc. 115, 9800–9801 (1993).

  41. 41.

    & Reactions of SmI2 with olefins: mechanism and complexation effect on chemoselectivity. J. Am. Chem. Soc. 118, 261–262 (1996).

  42. 42.

    & . Mechanistic impact of water addition to SmI2: consequences in the ground and transition state. J. Am. Chem. Soc. 127, 18093–18099 (2005).

  43. 43.

    , & . The role of proton donors in SmI2-mediated ketone reduction: new mechanistic insights. J. Am. Chem. Soc. 126, 44–45 (2004).

  44. 44.

    , & Switching between novel samarium(II)-mediated cyclizations by a simple change in alcohol cosolvent. Org. Lett. 5, 4811–4814 (2003).

  45. 45.

    , , , & Mixed disproportionation versus radical trapping in titanocene(III)-promoted epoxide openings. Tetrahedron 65, 10837–10841 (2009).

Download references

Acknowledgements

This work was partially supported by The Leverhulme Trust (Postdoctoral Fellowship to N.K.; RPG-2012-761) and the EPSRC (DTA Studentship to M.P.) (Established Career Fellowship to D.J.P.; EP/M005062/1).

Author information

Affiliations

  1. The School of Chemistry, University of Manchester, Oxford Road, Manchester M13 9PL, UK

    • Nicolas Kern
    • , Mateusz P. Plesniak
    • , Joseph J. W. McDouall
    •  & David J. Procter

Authors

  1. Search for Nicolas Kern in:

  2. Search for Mateusz P. Plesniak in:

  3. Search for Joseph J. W. McDouall in:

  4. Search for David J. Procter in:

Contributions

N.K. and D.J.P. conceived the study and co-wrote the manuscript. N.K. designed and performed experiments and M.P.P. performed experiments. J.J.W.M. performed the computational study.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to David J. Procter.

Supplementary information

PDF files

  1. 1.

    Supplementary information

    Supplementary information

Crystallographic information files

  1. 1.

    Supplementary information

    Crystallographic data for compound 2a

  2. 2.

    Supplementary information

    Crystallographic data for compound 4

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/nchem.2841