Recent Progress in Steroid Synthesis Triggered by the Emergence of New Catalytic Methods

Abstract

The rich biology associated with steroids dictates a growing demand for the new synthetic strategies that would improve the access to natural and unnatural representatives of this family. The recent advances in the field of catalysis have greatly impacted the field of natural product synthesis including the synthesis of steroids. This article provides a short overview of the recent progress in the synthesis of steroids that was enabled by the advances in catalysis.

Graphical Abstract

This mini review discusses the recent progress in the synthesis of steroids that was enabled by the advances in the transition metal catalysis, Lewis acid catalysis and organocatalysis.

1. Introduction

Due to the fact that steroidal hormones are of great importance for the regulation of a wide range of cellular functions in eukaryotic organisms, and humans in particular, this large and diverse family of natural products has played a central role in the fields of medicine and drug discovery.^[1] Historically, chemists have played an important role in both helping to understand the vital biological processes regulated by steroids as well as in developing steroid-based medicines for the treatment of diseases or improvement of the quality of human life. Needless to say, these advances would not be possible without breakthroughs in synthetic chemistry and catalysis, including transition metal mediated catalysis, asymmetric catalysis and organocatalysis. Synthesis of steroids requires addressing many challenges including the installation of all-carbon quaternary stereocenters,^[2] multiple redox manipulations,^[3] and assembly of polycyclic ring systems with defined stereochemistries at the ring junctions.^[4] In this review, we summarize some of the recent advances in the synthesis of steroids that have been enabled by the advances in catalysis. It is primarily focused on the studies that have emerged since 2014 and have not been reviewed elsewhere,^[5] and does not cover some of the important advances in closely related areas of asymmetric non-steroidal terpene natural product syntheses.^[6,7]

2. Syntheses Enabled by Transition Metal Catalysis

2.1. Enantioselective Palladium-Catalyzed Dearomatizative Cyclization for the Synthesis Steroid Boldenone Core[8]

Chiral phenanthrenone derived tricyclic cores bearing all-carbon quaternary centers are present in numerous complex terpenes and steroid natural products.^[9] An asymmetric intramolecular Heck reaction^[10] is a conventional approach to construct these polycyclic cores bearing all-carbon quaternary centers. Alternatively, in 2015, Tang et.al.^[8] reported an efficient palladium-catalyzed asymmetric intramolecular dearomatizative cyclization^[11] based on the earlier studies of the Buchwald group^[12] to afford various chiral phenanthrenones, some of which containing many key features of the steroidal and terpenoid frameworks (Scheme 1). They envisioned that bromoaryl-tethered phenol 1 would undergo dearomatizative cyclization in the presence of a chiral palladium catalyst to provide chiral product 2 and its achiral congener 3. The tentative mechanism for the formation of products 2 and 3 is depicted in Scheme 1. Initially formed through the oxidative addition to aryl bromide, intermediate II could potentially undergo two competing pathways in the presence of a base: 1) nucleophilic attack via the C4 position of phenol leading to the formation of palladacycle IIIa and eventually resulting in product 2 and 2) nucleophilic attack via the C2 position of phenol providing the regioisomeric achiral product 3. The intramolecular dearomatizative cyclization of 1 was optimized by screening various chiral palladium catalysts. The use of novel P-chiral ligand containing diphenylpyrrole substituent provided the best outcome with 94% yield and 92% ee of the desired product 2. The substrate scope study of this useful transformation revealed that the reaction is compatible with the fluoro-, chloro-, and methoxy- substituents in the aryl-bromide portion of the substrate.

Scheme 1.

Competitive cyclization pathways for oxidative dearomatization^[8]

The reaction was also compatible with the presence of biaryl and heteroaryl groups such as naphthalene, quinoline, and furan in aryl bromide end. The variation in tether length allowed to construct both 5- as well as 7-membered rings. The reaction outcome was not affected by varying the alkyl substituent at the C4 position as both ethyl and butyl substituted substrates provided favorable outcome; however, the phenyl substituted at the C4 (a) position substrate did not give any desired product. The versatility and efficiency of this method was also used to quickly set the core of the anabolic steroid boldenone (Scheme 2). The key precursor 7 was prepared from Hajosh-Parrish ketone-derived ketal 4 and bromide 5 following a three-step sequence consisted of alkylation, triflation, and debenzylation. The subsequent asymmetric dearomatization reaction with [{Pd(cinnamyl)Cl}₂]/(R)-8 smoothly provided boldenone skeleton 9 in high yield (90%) and high diastereoselectively (>99:1). Remarkably, the use of [{Pd(cinnamyl)Cl}₂]/(S)-8 granted access to diastereomeric skeleton 10 (90% yield, 98:2 d.r.) demonstrating the efficiency of this catalyst-controlled transformation.

Scheme 2.

Dearomatizative cyclization in the synthesis of boldenone skeleton^[8]

2.2. Enantioselective Total Syntheses of Furanosteroids (–)-Viridin and (–)-Viridiol by Guerrero’s Group^[13]

Recently in 2017, the Guerrero group reported an elegant approach to furanosteroids, (−)-viridin and (−)-viridiol based on an enantioselective intramolecular Heck reaction approach^[13]. While the racemic synthesis of viridin^[14] and viridiol^[15] was accomplished by Sorensen in 2004^[16], this late stage fragment coupling based strategy may be amenable to the synthesis of viridin analogs with modifications in D ring.

The synthesis began with the access of thioester 13 and stannane 16 (Scheme 3). A known aryl triflate^[17] was subjected to a Heck alkenylation to get compound 11. Hydrogenation of alkene and concomitant removal of benzyl group followed by treating the resulting cinnamic acid with neat chlorosulfonic acid led to indanone 12. Finally, the functionalized indanone 13 was obtained after thioesterification and triflation of 12. On the other hand, a known furan derivative^[18] 14 was converted to the A-ring furan fragment 16 in a 4-step sequence involving 1) chlorination of a primary alcohol using methanesulfonyl chloride and triethylamine, 2) displacement of chloride by allyl group using allyl magnesium chloride to get 15, 3) ring closing metathesis using Grubb’s second-generation catalyst and 4) lithiation followed by stannylation.

Scheme 3.

Synthesis of the key intermediates for the (–)-viridin and (–)-viridiol synthesis^[13]

Next, the fragment coupling between 13 and 16 was executed under Liebeskind stannane-thioester coupling conditions^[19] remarkably to get diketone 17 with aryl triflate intact in good yields and on multigram scale (Scheme 4). The coupled product 17 was then subjected for the stereodefining enantioselective intramolecular Heck reaction using Pd(0)-S-^tBuPHOX complex in presence of 1,2,2,6,6-pentamethylpiperidine (PMP) to furnish (+)-18 in 75% yield and high enantioselectivity (>99% ee). This reaction proceeds via the classical Heck reaction mechanism (Scheme 5), and the facial selectivity for the conversion of 17A into 17B is determined by the chiral ligand S-^tBuPHOX. It is noteworthy that the use of PMP was required to complete the conversion and reduce the alkene isomerization which was otherwise observed with diphosphine ligands.

Scheme 4.

Fragment assembly and synthesis of the precursor 23 to (–)-viridin and (–)-viridiol^[13]

Scheme 5.

Mechanism of the Heck reaction leading to (+)-18

Compound 18, though contains the complete carbon skeleton of viridin, still lacks the required oxidation of the A-ring. To install it, Upjohn dihydroxylation was carried out to get diol 19 diastereoselectively from α-face biased by the angular methyl group. The following Swern oxidation and methylation led to the methoxyketone 20 providing the required oxidation at C-3 via indirect means as the direct means of oxidation of this position on 18 were unsuccessful. The TBS group in 20 was cleanly removed by stirring in TFA to get compound 21 which was then subjected for a three-step sequence involving 1) a temporary protection of the D-ring ketone as a TBS-silyl-enol ether, 2) stereoselective reduction of the A-ring ketone using BH₃•THF, and 3) desilylation by using Et₃N•HF to reveal the D-ring ketone, to deliver keto alcohol 22 in 72 % isolated yields for three steps. Even though allylic alcohol 22 may potentially be advanced to (–)-viridin by hydroboration and selective alcohol oxidation, the former step was unsuccessful. Consequently, enol-ether 22 was epoxidized using AcOOH in methanol to give a mixture of diastereomers (–)-23 and (+)-23.

The seemingly challenging diastreoselective monodemethoxylation of dimethoxyacetals (–)-23 and (+)-23 was then successfully performed using TMSOTf and hydride donor to respectively give viridiol and epi-viridiol, albeit in low yields due either to decomposition of reactant or product (Scheme 6). Both epimers were then converted to viridin after TEMPO oxidation completing an 18-step synthesis starting from commercial materials.

Scheme 6.

Elaboration of (–)-23 and (+)-23 into (–)-viridin and (–)-viridiol^[13]

2.3. Synthesis of natural and enantiomeric steroids via metallacycle-mediated annulative cross-couplings^[20]

In 2017, Micalizio group reported an interesting and concise approach based on the metallacycle-mediated cross coupling to access a series of natural and enantiomeric steroids.^[20c] The strategy involves the early construction of the C/D ring system via a metallacycle-mediated annulative cross coupling^[21] between an alkyne and a suitably functionalized chiral enyne followed by a strategic formation of the C5-C6 bond after a suitable functionalization and activation of ACD tricycle. The enyne 24 was synthesized from (–)-epichlorohydrin either via a low yielding two-step protocol involving 1) S_N2 displacement of the chloride by alkynyl lithium derived from phenylpropargyl ether and 2) opening of the epoxide by propenyl cuprate or a more efficient three-step protocol involving 1) opening of the epoxide by alkynyl lithium in presence of BF₃•Et₂O, 2) treating with KO^tBu and 3) S_N2 displacement with propenyl cuprate. The enyne 24 was then subjected for a unique titanium-mediated annulative cross coupling with trimethylsilyl-phenylacetylene (25) to deliverhydrindane 26 along with the exo-diene 27 (Scheme 7). This powerful transformation constructs three C-C σ bonds and two stereocenters including one quarternary center leading to an angularly substituted trans-fused hydrindane from acyclic precursors. Mechanistically, this transformation was proposed to proceed via a series of cascade events as depicted in Scheme 7, which include 1) formation of titatanocycle intermediate 28; 2) an alkoxide-directed formation of metallacycle 29 followed by alkoxide exchange to generate 30 in high regioselectivity (>20:1 r.s.);^[22] 3) diastereoselective intramolecular [4+2] cycloaddition to give bridged metallacyclopentene 31 (>20:1 d.r.), 4) elimination to furnish tertiary allylic metal species 32, 5) isomerization to provide primary allylic metal species 33 and 6) stereoselective protonation via an allylic transposition to furnish hydrindanes 26 and 27. After the key cyclization transformation, the introduction of the C6 carbon and subsequent construction of the B ring was investigated (Scheme 8). Thus, the exocyclic methylene in 27 was cyclopropanated to give vinylcyclopropane intermediate 34. It should be noted that the yields of this reaction could be improved when the C-16 alcohol is protected.

Scheme 7.

Titanium(IV)-mediated formation of the steroidal C/D-ring system^[20]

Scheme 8.

Elaboration of 27 into steroidal core 36^[20c]

The initial attempts to ionize 34 with the expectation that it would undergo an electrocyclic ring opening followed by an intramolecular Friedel-Crafts alkylation were deemed unsuccessful. However, it was found that the treatment of 34 with TiCl₄ in nitromethane led to the steroidal product 36 in 68% yield. This process presumably proceeds via homoallylic cationic intermediate 35, which is formed after protodesilylation, protonation of resulting double bond and regioselective cyclypropane fragmentation^[23] triggered by the in situ generated protic acid arising from the reaction between the C-16 alcohol and TiCl₄. The subsequent intramolecular Friedel-Crafts alkylation reaction and the loss of HBr led to 36.^[24]This strategy was later adopted to the synthesis of terpenoid euphane analogs and investigated them as potential selective agonists of the estrogen receptor beta (ERβ).^[20a] The homopropargylic alcohol 37 underwent Ti(IV)-mediated annulation with alkyne 38 to provide the functionalized hydrindane 39 (Scheme 9).

Scheme 9.

Application to the synthesis of euphane analogs by Micalizio and coworkers^[20a]

This intermediate was then subjected to various Lewis and Br⊘nsted acids to obtain diastereomeric steroid cores 40a and 40b. While the diastereomer 40a is favored under the achiral Lewis acid catalysis with BF₃•OEt₂ (entry 1), the authors decided to further improve the selectivity by utilizing combinations of chiral Br⊘nsted acids obtained by complexation of Lewis acids and BINOL or its methylation of the C16 hydroxyl group of the D-ring did not lead to the erosion in the reaction dr. Such catalysts were previously explored by the Yamamoto^[27] and Corey^[28] to achieve enantioselective polyene cyclization reactions, and turned out to be of particular utility in enhancing the formation of the diastereomer 40a. Thus, the combination of (S)-BINOL and SnCl₄ at –78 °C resulted in the conversion of 39 to 40a in 50% yield and >20:1 dr. This transformation proceeds through the protonation of 41a and follows by the intramolecular Friedel-Crafts reaction through the intermediacy of 41b (Scheme 9). This remarkable reaction is dependent on the chirality of BINOL, and the use of the (R)-BINOL as the ligand instead led to 1:1 dr. In addition, the methylation of the C16 hydroxyl group of the D-ring did not lead to the erosion in the reaction dr.

The diastereomer 40a and related compounds were subjected to the oxidative dearomatizative rearrangement leading to the steroidal cores containing the C19 methylation. Thus, the treatment of 40a led to the formation of the deprotected C3 hydroxyl (Scheme 10). The reaction of this resultant intermediate with phenyliodine(III) diacetate (PIDA) results in intermediate 42 that undergoes subsequent rearrangement to provide the dearomatized carbocation 43, deprotonation of which leads to 44.

Scheme 10.

Dearomatizative rearrangement leading to the formation of steroidal skeleton with C19 methylation by the Micalizio group^[20a]

2.4. Pd-catalyzed intramolecular alkenylation to form 4,5-spirocyclic skeleton of phainanoid A^[29]

In 2017, Dong and coworkers published a palladium-catalyzed intramolecular alkenylation approach to access the strained cyclobutane-containing 4,5-spirocycle of the western part of phainanoids.^[29] This approach featured Pd-catalyzed intramolecular alkenylation of an enolate that provided the spirocyclic cyclobutane-containing portion of phainoids.

These model studies began by first developing a scalable route to allylic epoxide 52 (Scheme 11). This compound was generated from the commercially available geranyl acetate 45. Thus, 45 was subjected to a three-step sequence involving copper-catalyzed allylic coupling, chemoselective epoxidation, and oxidative cleavage of the epoxide that resulted in aldehyde 46 in 64% yield. Subsequently, the Still-Gennary variant of the Horner-Wadsworth-Emmons olefination was employed to afford Z-olefin-containing 48 in 79% yield. The reduction of the ester moiety in 48 with DIBAL-H and subsequent two step oxidation of the resultant alcohol 49 afforded aldehyde 50 in 70% yield over 3 steps. Finally, basic-alumina-promoted aldol condensation between 50 and 3-coumaranone (51) afforded allylic epoxide 52 with Z-selectivity.

Scheme 11.

Synthesis of precursor 52 for the model studies toward the synthesis of phainanoid A by the Dong group^[29]

With the efficient, scalable synthesis of 52 in hand, the synthesis of the western part of phainanoids was completed (Scheme 12). Lewis acid mediated polyene cyclization with SnCl₄ was employed to obtain tricyclic alcohol 53 with a trans-decaline core in 70% yield. Subsequent DMP oxidation and triflation of ketone 54 with Comins’ reagent afforded vinyl triflate 55 in 46% yield over 2 steps. Previous studies showed that the challenging selective reduction of a trisubstituted olefin in the presence of a vinyl triflate could be obtained via Pd/C-catalyzed hydrogenation with an H₂-balloon. This method was utilized to obtain vinyl triflate 56 in 80% yield. Finally, they attempted to form the benzofuranone-based 4,5-spirocyclic motif with and exocylic olefin. Previous work by Helquist and coworkers indicated that tert-butyl substituted phosphine ligands could promote palladium-catalyzed intermolecular alkenylation of ketones.^[30] Based on these findings Dong attempted the intramolecular alkenylation of vinyl triflates with Pd(OAc)₂ and QPhos. They observed the base had a significant effect on the reaction, and LiO^tBu was superior compared to LiHMDS, KHMDS, NaHMDS, KO^tBu, NaO^tBu, and Cs₂CO₃. They also found that the reaction performed best under moisture free conditions. Utilizing the optimized conditions from the model studies, they successfully performed the intramolecular alkenylation of 56 to afford the hexacyclic western part of the phainanoids (57).

Scheme 12.

Completion of the synthesis of the western portion of phainanoid A by the Dong group^[29]

2.5. Application of the transformations involving transition metal-catalyzed hydrogen atom transfer (HAT) to the synthesis of steroids

Transition metal catalyzed reactions involving HAT have been of great utility to the synthesis of steroids and terpenoids. While many applications of HAT reactions are focused on introduction of oxygenation via olefin hydration under mild conditions, several recent applications feature the application of HAT reaction for the C-C bond constructions that result in the formation of challenging motifs present in various steroids. Below we provide a brief overview of the recent applications of such transformations and highlight their use in the synthesis of aplysiasecosterol by the Li group^[31] and construction of (–)-nodulisporic acid C by the Pronin group.^[32]

2.5.1. Mukayiama’s hydration for the diastereoselective introduction of tertiary alcohols in the syntheses of steroids cortistatin A,^[35] ouabagenin^[36] and linckosides A and B^[37]

Transition metal-catalyzed transformations resulting in the net Markovnikov’s hydration of the alkenes represent powerful and mild methods for the introduction of hydroxyl groups into complex substrate.^[33] These transformations typically proceed in the presence of Co(II) or Mn(II) salts and require molecular oxygen and silane. The mechanism of Mukayiama hydration^[34] is not completely understood; however, it is believed to proceed through the formation of transition metal hydrides, that react with alkenes to accomplish hydrogen atom transfer (HAT) that results in the species with radical-like properties (Scheme 13A). These species undergo further reaction with molecular oxygen and the resultant peroxide species are being reduced by silane and Co(II). Importantly, these conditions may result in different stereoselectivities then what is observed in more traditional hydration reactions that proceed through the carbocationic intermediates.

Scheme 13.

A. The mechanism of Mukayiama hydration. B. C. Application of the Mukayiama hydration in the synthesis of cortistatin A by Baran and coworkers^[35]

This feature has been explored by the Baran group for the installation of the C5 oxygenation in their synthesis of cortisatin A (Scheme 13B and C).^[35] In order to introduce the α-C5 oxygen required for the synthesis of cortistatin A, Baran group first explored the hydration of the intermediate 58. However, the undesired β-C5 hydroxylated product 59 was formed instead. The authors rationalized this selectivity by proposing that the reaction mechanism involves a C5 radicals or a related species denoted by 60 and 61 (Scheme 13B). These radical intermediates differ by the configuration of the radical containing C5 carbon.

Due to the presence of the α-C1/ α-C2 epoxide moiety, the intermediate 61 containing pseudoaxial –NHCHO group is less favored and the reaction happens through 60 that contains cis-A/B ring junction. This mechanistic hypothesis prompted the authors to examine other substituents at the A ring. As a result of this studies, the hydration of compound 62 leading to intermediate 63 with the desired α-C5 configuration was identified (Scheme 13C). The formation of 63 may potentially happen through different conformers 64 and 65. The absence of the α-C1/ α-C2 epoxide allows now to achieve trans-AB ring junction in 65, which is the reactive intermediate, through which the formation of 63 happens. Importantly, the facile formation of 65 enabled the completion of the synthesis of cortistatin A.

Subsequently, the Baran group successfully implemented this strategy in the synthesis of polyoxygenated cardiotonic steroid ouabagenin from progesterone.^[36] As progesterone does not contain the β-C14 hydroxyl group, Baran and coworkers installed this moiety using Δ¹⁴-alkene-containing intermediate 66 (Scheme 14). Thus, treatment of 66 with molecular oxygen, phenylsilane and Co(acac)₂ as the reaction promoter resulted in the formation of 67 as the 8:1 mixture of β : α isomers. The selective formation of the β-isomer may be explained by the higher stability of the cis-hydrindane conformer of the C14-radical intermediate than the corresponding trans-hydrindane one.This strategy for the diastereoselective installation of the tertiary alcohols at the ring junction of the steroid skeleton was recently utilized by the Yu and coworkers in their synthesis of linckosides A and B (Scheme 15).^[37] Surmising that the Δ⁷-alkene of precursor 68 would preferentially react to form the β-C8 hydroxyl group, Yu and coworkers successfully carried the hydration of 68 to form advanced intermediate 69 that was later elaborated into linckosides A and B.

Scheme 14.

Application of the Mukayiama hydration in the synthesis of ouabagenin by Baran and coworkers^[36]

Scheme 15.

Application of the Mukayiama hydration in the synthesis of linckosides A and B by Yu and coworkers^[37]

2.5.2. Application of the transformations involving iron-catalyzed hydrogen atom transfer (HAT) to the synthesis of aplysiasecosterol by Li and coworkers^[31]

Aplysiasecosterol is a natural steroid derivative with a highly reorganized skeleton. While this molecule features many functionalities present in steroids, its assembly required the development of new synthetic strategy, in particular, for the construction of the highly substituted cyclopentane ring. Li’s group performed the first asymmetric total synthesis of the 9,11-secosteroid aplysiasecosterol in 2018^[31]. A convergent synthesis was accomplished to establish the stereocenters shown in Scheme 16 prior to a HAT based radical cyclization to construct the cyclopentaine ring of aplysiasecosterol in a selective manner.

Scheme 16.

Retrosynthesis of aplysiasecosterol^[31]

The Li group found that a desymmetrizing lactolization reaction could be performed using the Roush enantioselective allylation of symmetric aldehyde 73 with boronic ester 74 to bypass the traditional Corey-Bakshi-Shibata reduction and streamline the synthesis providing 75 in 85% yield and 9:1 er as determined by Mosher esterification followed by the ¹⁹F NMR analysis (Scheme 17). After the subsequent enantioenrichment of 76, the benzyl acetal formation was accomplished with BnOH and MsOH. This intermediate was then converted into α,β-unsaturated enone 77 using Nicolaou’s protocol (silyl enol ether formation then IBX/MPO oxidation) in 77% yield from 76. The left-hand segment synthesis was finished with annulation initiated with V-40 and(TMS)₃SiH (78% yield), subsequent silyl enol ether formation, NBS bromination, and ozonolysis to give 71. (91% yield over 3 steps).

Scheme 17.

Synthesis of the key intermediate 71 in the synthesis of aplysiasecosterol^[31]

For the synthesis of the right-hand segment 72, Li and coworkers initially utilized Myer’s alkylation; however, the process was tedious thus decreasing the efficiency. Therefore, a different route commencing with the formation of 80 from 79 using Sharpless asymmetric dihydroxylation was developed (cf. Scheme 18). This followed by the Grieco elimination/ozonolysis sequence to derive one carbon shorter homologue 81 (85%, 3 steps). The primary alcohol moiety of 81 was esterified under the Mitsunobu conditions to provide 82 (98% yield), which set stage for the Aggarwal’s lithiation-borylation chemistry that involved a stereoselective lithiation with (+)-sparteine. The resultant organolithium species was alkylated with alkyl pinacolborane to provide 83 with the desired configuration of the C17 stereocenter in 93% yield. Treating 83 with t-BuLi allowed for a Zweifel-Evans olefination to give 84 (93%, 15:1 dr at C17). This compound was further elaborated to the corresponding aldehyde by a sequence involving silyl deprotection and Dess-Martin oxidation to provide aldehyde 72 for the coupling with 71.

Scheme 18.

Synthesis of the intermediate 72 by Li and coworkers^[31]

Subsequently, 71 and 72 were linked by a radical Reformatsky type of aldol addition reaction that provided the anti-aldol product in 70% yield (Scheme 19A). This aldol addition product was converted to the corresponding aldol condensation product 70 by the elimination of water using Burgess reagent (58%, 2 steps). From here, the Li group envisioned that the key highly functionalized cyclopentane moiety of aplysiasecosterol could be installed using Fe(III)-catalyzed radical cyclization developed by the Baran group. The tentative mechanism of this transformation is depicted in Scheme 19B and is likely to involve a HAT resulting in a radical intermediate.^[38] Upon screening various ligands for the Fe(III) catalyst (acac, ox, dibm, and dpm, etc.) the optimal conditions were identified to provide 85 in 56% yield. This compound was subsequently subjected to deprotection that provided aplysiasecosterol in 92% yield (2 steps). It is of note that the cyclization conditions were extended to produce analogues with similar yield. From this work, Li and coworkers have provided an advantageous route that can be used to synthesize 9,11-secosteroids in a convergent fashion.

Scheme 19.

A. Fragment coupling and completion of the synthesis of aplysiasecosterol. B. Proposed mechanism for the iron(III)-promoted HAT cyclization.^[31]

2.5.3. Application of the transformations involving iron-catalyzed hydrogen atom transfer (HAT) to the synthesis of aplysiasecosterol by Pronin and coworkers^[39]

Pronin’s group recently applied iron-catalyzed HAT cyclization/aldol addition sequence to establish the eastern portion of indole diterpenoid, (–)-nodulisporic acid C (Scheme 20).^[39] The synthesis began with the copper-catalyzed asymmetric conjugate addition to 86 using the JosiPhos derivative SL-J015–1 to prepare the silyl-enol ether 87 in 95% yield and 73% ee. This silyl enol ether underwent an indium (III) bromide catalyzed addition to the TBS-protected pent-4-yn-1-ol,^[40] and the following acid work up provided access to 88 containing a quaternary stereocenter. This intermediate was subjected to a three-step sequence that included the formation of cyanohydrin, Sharpless allylic oxidation and primary alcohol oxidation by Dess-Martin periodane to provide the dialdehyde 89 in 40% yield over 3 steps. Subjecting 89 to Fe(acac)₃ with PhSi(OiPr)H₂^[38] resulted in diastereoselective formation of two vicinal quaternary centers and concomitant aldol addition.This followed by the acidic work up with HCl and then base to cleave the cyanohdrine moiety and provided 90 in 41% yield and 10:1 d.r. The high diastereoselectivity of this reaction is attributed to the pseudoaxial substituent at the C2 position. The substrate 90 was esterified (Piv₂O, Et₃N) and then subjected to Horner-Wadsworth-Emmons olefination to provide the fully functionalized eastern portion of (–)-nodulisporic acid C (91), which was subsequently elaborated to the natural product itself.

Scheme 20.

Application of iron(III)-promoted HAT cyclization/aldol addition cascade in the synthesis of (–)-nodulisporic acid C by Pronin and coworkers^[39]

2.5. Application of the directed copper-catalyzed C-H oxidation in the synthesis of pergularin, utendin and tomentogenin by the Baran group^[41]

While many steroids carry significant degree of skeletal oxidation, synthesis of steroids with high degree of oxidation represents a significant challenge. Recently, new approaches based on the late stage selective C-H oxidation started to emerge and be applied to complex steroid synthesis. Thus, based on the promising results obtained by the Schonecker group, Baran and coworkers have developed a powerful method for the introduction of the β-C12 hydroxylation at steroid skeleton under aerobic conditions (Scheme 21). This approach requires the presence of the C17 ketone that is then being converted to an imine functionalized with a 4-methylpyridyl moiety. This auxiliary is required for achieving chelation with Cu to form dimer 94 and for directing the oxidation to the β-C12 position resulting in 95. The imine moiety of 95 could be subsequently removed by the work up with saturated solution of Na₄TMEDA. The Baran group was able to demonstrate that this oxidation protocol is rather general and could be applied to a variety of substrates to obtain the corresponding oxidized products 93A-93H in preparatively useful yields (40–90%).

Scheme 21.

Cu-catalyzed C-H auxiliary-directed C-H oxidation of the C12 position of steroidal skeleton by the Baran group^[41]

This protocol was subsequently applied to the synthesis of steroids utendin, pergularin and tomentogenin that feature β-C12 oxidation of the C ring and contain highly oxidized D-ring (cf. Scheme 22). This synthesis commenced with inexpensive DHEA (96) that was subjected a 2 step sequence to provide auxiliary-containing product 97 (35% yield). This substrate was subjected to the aerobic oxidation in the presence of Cu(MeCN)₄PF₆ and sodium ascorbate to provide β-C12 oxidized product 93D in 40% yield. This trans-C/D ring-containing intermediate was converted to 98 in 34% yield via a 3 step sequence that involved the formation of silyl enol ether (TMSOTf, Et₃N), and subsequent Saegusa oxidation (Pd(OAc)₂, FeCl₃) followed by the silica gel promoted deconjugation of the resultant enone moiety. Intermediate 98 was next subjected to Mn(II)-promoted Mukayiama hydration conditions that afforded 99 in 67% yield.

Scheme 22.

Directed C-H oxidation in the synthesis of pergularin, utendin and tomentogenin by the Baran group^[41]

It is noteworthy that despite the fact that substrate 98 contains two alkenes, only the trisubstituted Δ¹⁴-alkene moiety reacted under these conditions. The diastereoselectivity of this reaction was consistent with the observations made by the Baran group in their studies on the synthesis of ouabagenin (cf. Scheme 14).^[38] The resultant product 99 was reacted with lithiated ethyl vinyl ether in the presence of LaCl₃ to provide 100 upon hydrolysis of the vinyl ether moiety (51% yield). The allylic cyclopropane functionality of 100 was cleanly converted to the homoallylic bromide by the treatment with HBr. The subsequent silver(I)-assisted solvolysis of this intermediate resulted in the corresponding homoallylic trifluoroacetate that was hydrolyzed to form pergularin by the treatment with aqueous trifluoroacetic acid (60% yield). This followed by the stereoselective reductions that were used to convert pergularin to utendin (NaBH₄, 75% yield) and then utendin to tomentogenin (Pd/C, H₂, 80% yield).

2.6. Rh-catalyzed cyclopropanation with quinone diazides in the synthesis of cycloartenol core by the Baran group^[45]

Cyclopropane-containing steroids such as cycloartenols represent challenging synthetic targets. In 2014, Baran and coworkers reported a new approach to the synthesis of such steroids that is based on an intramolecular cylopropanation with quinone diazides using Rh(II)-based catalysts and provided steroidal core 109 that may serve as the intermediate for the synthesis of cycloartenols and other classes of steroids.^[43]

These studies commenced with the preparation of the fragments 103 and 106 (Scheme 23). Thus, bromophenol derivative 101 was subjected to a 4 step sequence that involved Sonagashira coupling with TMS-protected acetylene (PdCl₂(PPh₃)₂, CuI), reduction of the nitro group (Zn, FeSO₄), and subsequent installation of the triazene moiety via the intermediacy of the diazonium salt (NaNO₂, HCl then HNiPr₂). This sequence culminated by base-promoted cleavage of the silane protection (K₂CO₃, MeOH) to provide triazene 102 in 68% yield (4 steps). Compound 102 was then subjected to Ni(0)-catalyzed hydroalumination of the alkyne moiety (NiCl₂(PPh₃)₂, DIBAL-H) that followed by the conversion of the organoaluminum intermediate to vinyl iodide 103 (51% yield) by its reaction with NIS. The synthesis of intermediate 106 commenced with 2-methyl-2-cyclopentenone 104 (Scheme 23). This compound was subjected to conjugate addition of vinyl cuprate followed by the capture of the enolate as the TMS-enol ether. After purification, this silyl enol ether was converted to the corresponding lithium enolate, which was cross-coupled with β-methallyl chloride ([PdCl₂(PPh₃)₂], MeLi, β-methallyl chloride) resulting in compound 105 (83% yield, 10:1 dr). This intermediate was subjected to protection to install 1,3-dioxalane moiety (ethylene glycol, p-TSA), ozonolysis (O₃, PPh₃) and aldol condensation (LiOH, iPrOH followed by PPh₃, CBr₄) to provide key intermediate 106 in 47% yield over 4 steps.

Scheme 23.

Rh(II)-catalyzed intramolecular cyclopropanation in the synthesis of the cycloartenol core by the Baran group^[43]

With both intermediates in hand, their coupling was accomplished next. Thus, 103 was subjected to Mg/halogen exchange with iPrMgCl•LiCl, and the resultant Grignard reagent was employed as a nucleophile for the Cu(I)-catalyzed 1,4-conjugate addition to 106 followed by the capture of the resultant enolate with TMSCl to form the corresponding silyl enol ether. The subsequent reduction of the styrene moiety (Pd/C, H₂) provided intermediate 107 in 69% yield over 2 steps. This compound reacted with Eschenmoser’s salt and Li₂CO₃ to install the exocyclic enone moiety and subjected to deprotection of 1,3-dioxalane (SiO₂). The resultant compound was treated with trifluoroacetic acid, which resulted in the deprotection of the MOM group and formation of the quinone diazide 108. Quinone diazide was then treated with various Rh(II) salts to accomplish intramolecular diastereoselective cyclopropanation proceeding through Rh-based carbenoid and leading to steroidal core 109. The evaluation of various catalysts helped to identify [Rh₂(esp)₄] as the best catalyst of this transformation that afforded 109 in 80% yield.

2.7. Pd/Ni-promoted Ulman coupling in the synthesis of the batrachotoxin core by the Inoue group^[44]

Batrachotoxin is a steroidal alkaloid with neurotoxic properties isolated in 1968 from the skin of Columbian poison-arrow frogs.^[45] Batrachotoxin features multiple fused rings and high level of skeletal oxidation, and represents a formidable synthetic target.^[7k,46] In 2018 Inoue’s group disclosed a new approach that allowed for the synthesis of batrachotoxin’s steroidal core 110 (Scheme 24), which could potentially be expanded to the synthesis of batrachotoxin.^[44] Intermediate 110 was generated from 111 through the intramolecular Pd/Ni-catalyzed Ullman coupling reaction. Compound 111, in turn, was prepared by the radical addition of 113 to 112, which represents another the key step in the synthesis of 110.

Scheme 24.

Retrosynthetic analysis of Inoue’s approach to batrachotoxin ^[44]

The synthesis of intermediate 112 commenced with the asymmetric Noyori transfer hydrogenation of prochiral substrate 114 resulting in the desymmetrized reduction product in 96% ee (Scheme 25). The following TBS protection resulted in 115 (59% yield, 2 steps). This product was enolized in the pesence of the Commins’ reagent to provide 116 (84%), which was converted to 112 in 3 steps via the oxidative cleavage of the terminal alkene followed by the aldol condensation with malononitrile (80%, 3 steps). The synthesis of 113 commenced with 117 that was derived from Wieland-Miescher ketone. This compound was reacted with lithiated vinyl ethyl ether, and the resultant product was subjected to 1,3-dioxolane hydrolysis/intramolecular acetalization (CH(OMe)₃, MeOH, CSA) followed by the oxidative cleavage of the vinyl ether moiety (RuCl₃, NaIO₄) to provide 118 in 58% yield.

Scheme 25.

Synthesis of the key intermediates^[44]

With both fragments 112 and 113 in hand, the subsequent coupling was attempted (Scheme 26) using Et₃B and molecular oxygen. The mechanism of this transformation is depicted in Scheme 26 and commences with Et₃B reacting with oxygen and producing ethyl radical. Ethyl radical then abstracts Te from the C-Te bond of 113. The resultant acyl radical undergoes a subsequent decarbonylation to provide an α-alkoxy radical with fixed stereochemical configuration.^[47] The following addition to the electron deficient alkylidenemalonitrile 112 provides a stabilized radical that is subsequently trapped with Et₃B to form boron enolate, which hydrolyses upon work up and provides 111 (α-C11-H, 41% yield; β-C11-H, 27% yield).

Scheme 26.

Coupling of intermediates 112 and 113 leading to 111^[44]

With this in hand, the intermediate 111 was elaborated to batrachotoxin core 110 (Scheme 27A). These studies commenced with the oxidation of the malononitrile moiety with monoperoxyphthalate to provide 120, which was subjected to ethanolysis under the basic conditions (NH₃, EtOH) leading to 121 (α-C11-H, 60% yield; β-C11-H, 61% yield). This substrate contains both vinyl bromide and vinyl triflate moieties that provide handles for the subsequent reductive cyclization.

Scheme 27.

A. Synthesis of batrachotoxin core 110. B. Potential mechanism for the Pd/Ni-catalyzed Ullman coupling^[44,48]

The Inoue group first employed Ni-catalyzed Ullman coupling (NiCl₂, 2,2’-bipyridine, Zn, CH₃CN, Py, 50 °C); however, this transformation did not lead to high yields of 110 due to the slow oxidative insertion into C-OTf bond. Subsequently, a modified protocol developed by the Weix group and utilizing dual Pd/Ni catalytic system was employed.^[48] Low yields were observed when the transformation was attempted catalytically; however, the use of stoichiometric metal complexes tripled the yields of 110. While the actual reaction mechanism is yet to be clarified, based on the Weix group proposal, a catalytic cycle depicted in Figure 27B and potentially involving a bimetallic intermediate could be envisioned.

3. Syntheses Enabled by the Lewis Acid Catalysis

3.1. Cu(II)-catalyzed enantioselective tandem Michael/Aldol reactions in the synthesis of cardiotonic steroids by Nagorny and coworkers^[49]

Cardiotonic steroids represent a large family of steroids with unique structural features such as characteristic β–C14 alcohol, β-C17 heterocyclic ring, and an unusual cis-C/D ring junction pattern, which imparts a rigid “U” shape to the molecule. Around a thousand of different natural cardenolides have been isolated to date from various plant and animal sources. Considering that cardenolides are involved in regulating vital biological processes and for centuries have been used as therapeutic agents, the total synthesis of cardenolides has attracted considerable attention. Despite these numerous studies, the synthesis of highly oxygenated cardenolides, such as ouabagenin, represents a significant challenge, and only recent advances in catalysis and synthetic methodology allowed to overcome some of these challenges. Such landmark efforts include the recently disclosed syntheses of ouabagenin and 19-hydroxysarmentogenin by the Deslongchamps,⁵⁰ Baran³⁶ and Inoue⁵¹ groups as well as the studies by the Nagorny group that will be the focus of this mini-review.^[49]

The tandem Michael/aldol reaction approach to cardiotonic steroids is summarized in Scheme 28. This approach is based on the presumption that enones 123 and β-ketoesters 122 could undergo a stereoselective Michael reaction to provide 124. Subjecting 124 to tandem aldolization under acidic or basic conditions would lead to 125, which represents a fully functionalized cardenolide skeleton that could be further elaborated into steroids 126 featuring various oxidation patterns.

Scheme 28.

Cu(II)-catalyzed tandem Michael/aldol reaction approach to the synthesis of cardiotonic steroids^[49]

Michael reactions are sensitive to the substrate sterics, and the transformations leading to products like 124 that contain vicinal quaternary/tertiary all-carbon stereocenters are very rare. After screening a variety of conditions, Nagorny and coworkers identified Cu(II) salts as unique catalysts for this transformation under no solvent conditions. Based on these findings, in 2015 Nagorny group demonstrated that the formation of 130 is possible in a highly enantioselective and diastereoselective fashion from simple and readily available building blocks 127 and 128 (Scheme 29).^[49g] Thus, it was demonstrated that readily available chiral bis(oxazoline) copper(II) complex 129 could promote

Scheme 29.

Enantioselective Cu(II)Box-catalyzed Michael/aldol cascade approach to cardenolide core. A. Overall transformation. B. Substrate scope. C. Tentative reaction mechanism^[49c,g]

Michael reactions with high levels of stereocontrol and the stereochemistry required for their subsequent elaboration to natural cardiotonic steroids (Scheme 29A). These Michael adducts could be subjected to various acidic or basic conditions that promote the intramolecular aldolization and provide various steroid-like cores 130A-J. The choice of the cyclization conditions was of particular importance for establishing the stereochemistry of the CD ring junction. By using DBU as base in THF, the Michael adducts were converted into products 130A-F with unnatural α-configuration of the C13/C14 stereocenters in good to excellent yields and selectivities. Alternatively, treating the substrate with Cs₂CO₃ in DMF at 140 °C led to the product 130G possessing natural β-C13/β-C14 configuration. Similarly, subjecting the Michael adducts to 2 step conditions involving pre-cyclization of the B-ring using pyrolidinium acetate followed by the formation of the C-ring with LiHMDS or NaHMDS led to products 130I-J. Finally, the reaction with β-ketocarboxylic acid instead of ester (i.e. 142, R₁=H), led to the formation of nor-steroid skeleton 130H in good yield and selectivity (85%, 8:1 dr, 93% ee, gram scale).^[51c] The tentative mechanism of this transformation is depicted in Scheme 29C. Unlike the majority of other Cu(II)Box-catalyzed reactions, 129 is proposed to bind and activate nucleophile 131 rather than enone electrophile. The chelation of 129 and 131 leads to the formation of complex 132 that then transfers a proton to the enone and results in Cu(II)-based chiral enolate 133. The subsequent Michael reactions presumably happens through an open transition state 134 and results in the corresponding product 136.

These studies were subsequently translated into the synthesis of more complex cardenolide cores 146 and 147 (Scheme 30) that were later elaborated to a variety of natural cardenolides.^[49a,d,e] The modified approach required adjusting the oxidation state of both the β-ketoester and enone fragments in order to install the C3 and, optionally, C11 oxygenation. To address the problems associated with the installation of the C3 stereochemistry, vinylchloride-containing β-ketoester 137 was employed. β-Ketoester 137 reacted with either the achiral enone 138 or prepared in 3 steps OBz-containing chiral enone 139 (Scheme 30). In the former case, the reaction was catalyzed by chiral catalyst 129 that promoted the formation of Michael adduct 140 in 92% yield, 92%ee, >20:1 dr. This compound was then cyclized with p-TSA to provide functionalized steroidal core 142 (57% yield, >20:1 dr).

Scheme 30.

Enantioselective and diastereoselective Cu(II)-catalyzed Michael/aldol cascade reactions for the synthesis of functionalized cardenolides cores 146 and 147^[49a,d,e]

In the case of chiral enone 139, the Michael reaction with 137 was catalyzed by Cu(OTf)₂ (50 mol%), and resulted in the diastereoselective formation of product 141 in 70% yield as well as a monocyclized aldol product resulting from 141 (20% yield). Subjection of both of these compounds to p-TSA in the subsequent step resulted in the formation of aldol adduct 143 (59% yield from 139). Alternatively, the formation of 143 from 139 and 137 could be executed in a single operation (45% yield, 20:1 dr) by carrying both steps in one pot. In both cases the stereoselective formation of all 5 new stereocenters was imposed by the C11-OBz group-containing stereocenter.

The aforementioned protocols enabled concise formation of the steroidal cores 142 and 143 on multigram scale. Both of these intermediates contained the desired stereochemistry and oxygenation with the exception of the α-C13/α-C14 stereocenters at the CD ring junction. However, our DFT calculations indicated that the natural β-C13/β-C14 configuration present in 144 and 145 has higher thermodynamic stability. Therefore, 142 and 143 were subjected to retro-Aldol/Aldol sequence resulting in the net epimerization of the C13/C14 stereocenters and providing 144 and 145 in 64% and 68% (brsm) respectively. These products underwent the global reduction with DIBAL-H followed by the solvolysis of vinyl chlorides to generate unsaturated C3 ketones in an efficient one pot transformation (71% yield, 8:1 dr and 72% yield, 8:1 dr for 146 and 147, correspondingly). Both advanced intermediates were produced in only 6 steps (LLS) on a multigram scale and contained all of the necessary functionalization to be advanced to various cardiotonic steroids (Schemes 31 and 32).

Scheme 31.

Elaboration of core 147 into cardenolide ouabagenin^[49a]

Scheme 32.

Elaboration of cores 146/147 into various cardiotonic steroids^[49]

More advanced intermediate 147 contains oxygenations at the C3, C11, C14, C17, and C19 positions as well as the unsaturation at the C-5 position, which is critical for its rapid elaboration to various highly oxygenated cardenolide ouabagenin (Scheme 31).^[51a] The presence of the additional C1/C5 oxygenation in ouabagenin makes this important natural product to be one of the most complex targets among the cardiotonic steroids. The synthetic studies towards ouabagenin commenced with the reduction of the Δ⁷-olefin of 147 with LiDBB, followed by a basic work-up with NaHCO₃ to provide 148. Polyol 148 was then subjected to a 3 step sequence that included acetylation of the C11, C17, and C19 positions (Ac₂O, Py, DMAP), diastereoselective reduction of the C3 ketone with K-Selectride®, and C3-directed epoxidation of the Δ⁴-olefin with m-CPBA (6:1 dr for epoxidation step) to afford 149 in 62% yield. The following oxidation of 149 with Pd(TFA)₂ and NaOAc under molecular oxygen atmosphere afforded 150 in 90% yield. Enone 150 was subjected to another 3 step sequence to furnish 151 (62% yield) as a single diastereomer. This was accomplished by deacetylation followed by the Δ¹-epoxidation with LiOH and H₂O₂ to generate the C1/C2 epoxide, as a single diastereomer. The C19 hydroxyl of this bis-epoxide underwent selective protection as the TIPS ether (TIPSCl, 2,6-lutidine) that followed by the oxidation of the C11 and C17 hydroxyls to afford 151. The reduction of the epoxide moieties of 151 represented a significant challenge due to the propensity of the resultant product to undergo dehydration leading to a complex mixture of enones. After extensive screening, the conditions that involved the reduction of 151 with PhSeSePh (20 mol%) and N-Ac-L-Cys-OH under basic conditions resulted in the desired product 152 in 87% yield. Careful control of the N-Ac-L-Cys and NaOH stoichiometry was required to avoid water elimination and subsequent aromatization of the A ring. Then 152 was subjected to a single pot diastereoselective reduction of the C3 ketone with K-Selectride® followed by the protection of the C3 and C17 alcohols as the TBS ether and enol ether, respectively (TBSOTf, 2,6-lutidine). The resultant product was subjected to another single pot procedure that involved the C11 ketone reduction under dissolving metal conditions to obtain the equatorial C11 alcohol, followed by work up with TBAF that provided ouabagenin core 153 in 74% yield over two steps. Vinyl iodide moiety at C17 was then installed using Barton’s protocol to give 154 in 96% yield. The final installation of the C-17 butenolide was achieved via a four-step sequence involving: 1) a Stille coupling of vinyl iodide 154 with stannane 155, 2) TMS protection of the C11 and C14 alcohols, 3) diastereoselective hydrogenation of Δ¹⁶- double bond, and 4) global deprotection of silyl-protecting groups with aqueous HF to afford ouabagenin in 42% over the final four steps.

The aforementioned strategy for the assembly of ouabagenin could be generally applied to a variety of other cardiotonic steroids (Scheme 32).^[49c,d] Thus, key intermediate 147 was successfully elaborated to diastereomeric cardenolides trewianin aglycone (12 steps, 14% yield), ^[49a,d] 19-hydroxysarmentogenin (12 steps, 8% yield), ^[49a,d] panogenin (9 steps, 5% yield) ^[49a] and 5-epi-panogenin (9 steps, 23% yield)^[49a,d] that have different configurations at the C5 and C11 stereocenters. In addition, these efforts provided sarmentologenin (14 steps, 7% yield) that lacks C1-oxygenation in comparison to ouabagenin.

Similar strategy was used to convert the intermediate 146 lacking the C11 stereocenter into glycosylated cardiotonic steroid cannogenol-3-O-α-L-rhamnoside (12 steps, 11% yield) and its analogs with modified sugar.^[49d]

4. Syntheses Enabled by the Organocatalysis

The seminal studies of Hajos, Parrish, Eder, Sauer and Wiechert revolutionized the synthesis of steroids and terpenoids and helped to reveal the power of organocatalysis for the synthesis of complex natural products.^[52] Not surprisingly, the development of new organic catalysts and organocatalyzed tandem reactions is still of great importance to the synthesis of steroids. While many modern studies are focused on improving the synthesis and formation of functionalized Hajosh-Parrish and Wieland-Miescher ketones, the developments in organocatalysis allow to achieve rapid formation of significantly more complex intermediates. This subsection of the mini review highlights some of such studies that emerged from the groups of Hayashi,^[53] Hong^[54] and List.^[55]

4.1. Pot-economical total synthesis of estradiol methyl ether by the Hayashi group^[53]

Interested in developing rapid and economical ways for the synthesis of pharmaceutically important natural products,^[56] Hayashi and co-workers recently reported a highly pot-economical synthesis^[53] of estradiol methyl ether using proline-derived organocatalyst 158 (cf. Scheme 33).^[57] The highlight of their synthesis involves an expedient stereo-controlled construction of the intermediate 159 with all carbons necessary for installing the steroid framework within the first pot. The reaction involves diphenylprolinol trimethylsilylether (158) catalyzed asymmetric reaction between nitroalkane 156 and 3-(p-methoxyphenyl)-propenal 157. The mechanism of this transformation is depicted in Scheme 34A. The condensation of 157 and 158 leads to the formation of iminium ion 162, which undergoes an asymmetric Michael reaction with nitronate anion 163 derived from 156. This results in an enamine-containing product 164 that undergoes highly selective addition to one of the ketone moieties of the 1,3-cyclopentanedione linked to the enamine. Upon hydrolysis of the resultant iminium ion 165, the bicyclic structure 166 containing five contiguous stereocenters is obtained. The selectivity of this domino reaction is remarkable and provides essentially a single diastereomer with very high enantioselectivity (>99% ee). This reaction was successfully telescoped with two other subsequent reactions: 1) stereoselective addition of cyanide followed by the formation of xanthate ester 167 (cf. Scheme 34A), and 2) dehydration using SOCl₂ and pyridine to afford intermediate 159 in 78% yield from 156 and 157 in the first pot (cf. Scheme 33). The simultaneous removal of nitro-group and xanthate in 159 was carried out reductively using Bu₃SnH and AIBN under microwave condition and resulted in 160. This intermediate was subjected a two-step one pot sequence involving diastereoselective reduction of the ketone and nitrile via sequential addition of LiBHEt₃ at −78 °C and DIBAL at 0°C. The resulting hydroxyl aldehyde after TIPS protection (TIPSCl, imidazole) affords aldehyde 161 in 59% yield from 160. What follows next is a remarkable six-step one-pot sequence depicted in Scheme 33 that includes 1) Krauss-Pinnick oxidation yielding 168 (Scheme 34B), 2) trans-hydrindane-selective hydrogenation controlled by OTIPS group and leading to 169, 3) acyl chloride 170 formation, 4) Friedel-Crafts acylation 5), deprotection of the silyl group by the addition of MeOH leading to 171, and 6) reduction of benzyl ketone to give estradiol methyl ether in 55% yield. The synthesis highlights the idea of compatibility in one-pot transformations and accomplishes the synthesis of estradiol methyl ether in five reaction pots involving four purification steps.

Scheme 33.

Hayashi’s pot-economical synthesis of estradiol methyl ether^[53]

Scheme 34.

A. Mechanism for the formation of intermediate 167. B. Synthetic intermediates for the pot-economical synthesis of estradiol methyl ether^[53]

The key domino reaction of diphenylprolinol silyl-ether mediated Michael reaction of nitroalkane and intramolecular aldol reaction is quite general in terms of aryl group at 3-position of propenal as various electron rich (p-methoxy-, p-methyl-) as well as electron deficient (p-fluoro-, p-bromo-, p-chloro-, o-fluoro-) aryl group containing enals afford bicycle [3,3,0] nonane frameworks with excellent enantioselectivity.

4.2. Organocatalytic enantioselective Michael-Michael-Aldol-Henry reaction cascade for the one pot synthesis of nor-steroid skeleton by Hong and coworkers^[54]

In 2014, Hong and coworkers reported a one pot organocatalytic enantioselective double Michael/aldol/Henry cascade reactions to assemble enantiomerically enriched nor-steroid 181 from simple precursors 172 and 173.^[54] This approach is based on the organocatalytic union of 172 and 173 catalyzed by the organocatalyst 158 (Scheme 35). This transformation represents the key step in the sequence leading to 181 and proceeds through the initial formation of iminium ion 175 from 173 and prolinol derivative 158.^[57] Iminium 175 undergoes a subsequent Michael reaction with nitronate anion 176 derived from 172. This transformation proceeds with high levels of stereocontrol at the newly formed carbon stereocenter; however, a ~1:1 mixture of diastereomers at the stereocenter containing the nitro- group is formed. The chiral prolinol portion of the resultant adduct 177 is of great importance for controlling the stereochemistry of the subsequent intramolecular Michael addition of enamine into enone to provide intermediate 178, which subsequently undergoes hydrolysis to form aldehyde 174 as a 1:1 mixture of epimers at the nitro- group containing stereocenter. This transformation was followed by in situ addition of p-TSA that led to intramolecular aldol condensation to form the mixture of 179 and 180 epimeric at the NO₂-containing stereocenter. The one pot sequence leading to 181 was then completed by the sequential addition of DBU followed by TBAF. Both reagents promoted the intramolecular Henry reaction by deprotonation leading to nitronate intermediate 182; however, the addition of DBU triggered the cyclization of 179 while TBAF was required for the reaction of 180 to happen. The cyclization presumably proceeds through the transition state 182B (rather than 182A) and leads to the natural configuration of the CD ring junction. The desired product 181 was obtained in 47% yield over the entire single pot sequence as a single isomer. Its absolute and relative configurations corresponded to the stereochemistry observed for the natural cardiotonic steroids.

Scheme 35.

One pot construction of nor-cardenolide core 181 by Hong and coworkers^[54]

4.3. Organocatalytic chiral Br⊘nsted acid-catalyzed synthesis of estrogen by List and coworkers^[55]

While many applications of organocatalysis in synthesis of steroids are based on the exploration of chiral amine catalysis, other applications started to immerge in the recent years. Thus, in 2014, Benjamin List’s group reported an asymmetric synthesis of (+)-estrone that was enabled by the chiral Br⊘nsted acid catalysis (cf. Scheme 36).^[55] This synthesis was based on the racemic approach by Torgov, who in 1963, described an acid catalyzed cyclization of easily available diketone 184 into a steroidal Δ^8,14-dienone 186, which is a useful precursor to various steroids including estrone.^[58] Even though Torgov’s cyclization-represents an efficient way to generate racemic steroidal scaffolds, the development of the corresponding enantioselective variants has always been elusive.^[59] Mechanistically, the developed Torgov’s cyclization involves four sequential acid catalyzed steps: 1) isomerization of exocyclic Δ^9,11-olefin in compound 184 to the endocyclic Δ^8,9-isomer 187; 2) intramolecular Prins-type cyclization through 188 leading to a stabilized carbocation 189; 3) deprotonation of carbocation 189 to give isomeric mixture of olefins 190, and 4) isomerization and dehydration of 190 to furnish Torgov’s diene 186. List’s group realized that the likely stereo-determining step, the cyclization of 188 to 189, could potentially be catalyzed by using chiral Br⊘nsted acid.^[60] In fact, upon screening of various chiral Br⊘nsted acids, a unique chiral disulfonamide (DSI) catalyst 185 containing SF₅- and NO₂- substituents was found to catalyze the cyclization in high yield and high enantioselectivity (95% yields, 93% ee). A single recrystallization essentially provided enantiopure compound 186 (>99.8% ee). It is noteworthy that precious DSI catalyst 185 could be recovered in 88% yield and reused following the acidification with similar efficacy. Diene 186 was then successfully converted to (+)-estrone by using two-step diastereoselective reduction protocol developed by E.J. Corey.^[61] This sequence included stepwise reduction of first Δ¹⁴- (Pd/C, H₂, Et₃SiH) and then Δ⁸- (CF₃CO₂H, TBAI, Et₃SiH) alkene moieties followed by the deprotection of (+)-estrone methyl ether with BBr₃.

Scheme 36.

Synthesis of (+)-estrone by enantioselective Torgov’s cyclization using chiral acid 201^[55]

5. Conclusion and Outlook

This mini review has highlighted some recent progress in the synthesis of steroids that was enabled by the advances in catalysis. The rapid progression of the field of catalysis has greatly expanded the toolbox of a synthetic organic chemist and provided new powerful transformations that allow to install complex functionalities in a highly selective and mild manner. This has resulted in many new and creative approaches to complex steroidal skeletons that feature high efficiency and allow to accomplish multiple steps in a single reaction pot. Such approaches provide robust platforms for the subsequent medicinal chemistry exploration of steroids and other related natural products and their derivatives.

Biographies

Dr. Hem Raj Khatri has obtained his B.S. degree in chemistry from Tri-Chandra Campus in Nepal in 2002, and his M.Sc. Degree in organic chemistry from Tribhuvan University in Nepal in 2005. In 2010 he joined the group of Professor Jianlong Zhu at the University of Toledo where he worked on the synthesis of derhodinosylurdamycin A and synthetic methodology. In 2015 he completed his Ph.D. and joined the group of Pavel Nagorny at the University of Michigan as a postdoctoral fellow. His postdoctoral work has been focused on the synthesis of cardiotonic steroids ouabagenin, 19-sarmentologenin and trewianin aglycone. In 2019, he joined Astex Pharmaceuticals as a process chemist.

Nolan Carney obtained a B.S. in Chemistry and Biology from Jacksonville University, Florida in 2018. Since then he has moved to Ann Arbor, Michigan to pursue a Ph.D. in the Program of Chemical Biology at University of Michigan under the supervision of Professor Pavel Nagorny. His research interests include total synthesis of natural products and analogs with a focus on cardiotonic steroids.

Ryan Rutkoski obtained his B.S. degree in chemical engineering from Rowan University, New Jersey at 2018. Upon completion of these studies, he joined the Medicinal Chemistry Program at the University of Michigan. In 2019 he joined the Nagorny group where he currently explores the synthesis and medicinal chemistry of cardiotonic steroids and develops new methods for the glycosylation of natural products.

Bijay Bhattarai obtained B.S. in chemistry from Southeastern Louisiana University in 2013 where he conducted research in organic synthesis and methodology. In 2013 he joined the Nagorny group at the University of Michigan where he carried research on hydrogen bond catalysis and total synthesis of cardiotonic steroid cannogenol and its derivatives. After completing his Ph.D. in 2018, he joined Wave Life Sciences as a medicinal chemist.

Dr. Pavel Nagorny received his B.S. degree in chemistry in 2001 from the Oregon State University where he conducted research in the laboratory of Professor James. D. White. After earning his Ph.D. degree in chemistry under the mentorship of professor David A. Evans from Harvard University in 2007, he spent three years as a postdoctoral fellow with Professor Samuel J. Danishefsky at the Memorial Sloan-Kettering Cancer Center. In 2010, Pavel joined the faculty of the University of Michigan as a Robert A. Gregg Assistant Professor in Chemistry. From 2014–2017 he was appointed as a William R. Roush Assistant Professor in Chemistry, and in 2017 he was promoted to the rank of Associate Professor. Pavel’s research group interests range from natural product synthesis to asymmetric catalysis, organocatalysis and carbohydrate chemistry.