Exploration of Oxidative Ritter-Type Reaction of α-Arylketones and Its Application for the Collective Total Syntheses of Erythrina Alkaloids

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE3 Oct 2022Exploration of Oxidative Ritter-Type Reaction of α-Arylketones and Its Application for the Collective Total Syntheses of Erythrina Alkaloids Meng-En Chen†, Shi-Zhong Tang†, Yue-Hong Hu, Qin-Tong Li, Zhang-Yan Gan, Jian-Wei Lv and Fu-Min Zhang Meng-En Chen† State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 †M.-E. Chen and S.-Z. Tang contributed equally to this work.Google Scholar More articles by this author , Shi-Zhong Tang† State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 †M.-E. Chen and S.-Z. Tang contributed equally to this work.Google Scholar More articles by this author , Yue-Hong Hu State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 Google Scholar More articles by this author , Qin-Tong Li State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 Google Scholar More articles by this author , Zhang-Yan Gan State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 Google Scholar More articles by this author , Jian-Wei Lv State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 Google Scholar More articles by this author and Fu-Min Zhang *Corresponding author: E-mail Address: [email protected] State Key Laboratory of Applied Organic Chemistry, College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.021.202101385 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Although the classical Ritter reaction has been widely applied to prepare sterically hindered amides since 1948, it has intrinsic problems, such as harsh reaction conditions, the multistep preparation of synthetic precursors, and the use of solvent quantities of nitrile. In particular, only a few examples of the total synthesis of natural products using the Ritter reaction as a key step have been reported to date. In this article, we report the development of an oxidative Ritter-type reaction of α-arylketones to efficiently construct a sterically hindered N-acyl aza-quaternary carbon moiety. The current transformation features the use of 10 equiv of nitriles, a broad substrate scope (with 81 examples), a short reaction time, and mild reaction conditions. Notably, both the use of a limited amount of nitriles and producing carbocation intermediates via the C–H bond oxidation strategy address two intractable problems of the classical Ritter reaction. Furthermore, based on an unprecedented synthetic strategy using this oxidative Ritter reaction to construct a C-5 aza-quaternary carbon center, the collective total syntheses of erythrina alkaloids, including erysotramidine, 11-α-methoxyerysotramidine, 11-β-hydroxyerysotramidine, erytharbine, the proposed 11-β-methoxyerysotramidine and 10,11-dioxoerysotramidine, and the unnatural 11-α-hydroxyerysotramidine, have been completed using a common precursor through a one-step chemical transformation. Download figure Download PowerPoint Introduction The Ritter reaction, in which a nitrile reacts with a reactive carbocation intermediate generally derived from alkyl alcohol, alkene, or their derivatives, provides a concise synthetic method for the preparation of amides,1–5 especially those featuring an N-acyl aza-quaternary carbon moiety from the tertiary carbocation intermediate.6–8 Such sterically hindered amides are widely found in functional nitrogen-containing organic molecules,9–15 including bioactive natural alkaloids,9,10 drugs,11 catalysts,12 DNA damage probes,13 and precursors of materials such as polymers14 and liquid crystal materials15 (Figure 1). Therefore, the Ritter reaction has attracted attention from the synthetic community, and diverse synthetic strategies have been developed in the past decade.16–21 Among them, the C–H bond oxidation strategy for producing reactive carbocation intermediates can avoid the preparation of synthetic precursors bearing an additional leaving group or a suitable C=C double bond, and thus dramatically improves the efficiency of the Ritter reaction.5,19–26 Although some pioneering oxidative systems, such as anodic oxidation,20–22 NOPF6,23N-hydroxyphthalimide/ceric ammonium nitrate (NHPI/CAN),24 and CAN/NaN3,25 have been developed, the reaction substrates are generally limited to adamantine and its analogs or structurally special substrates. Therefore, further expansion of the substrate scope to produce variously functionalized and synthetically valuable amides is highly desirable but underdeveloped. Retrospectively, when unbranched alkyl ketones were used as substrates, poor regioselectivities at the β/γ/δ/ε positions were observed under electrochemical oxidative conditions, while similar branched ketones yielded some rearranged acetamides (Scheme 1a).22 Recently, Baran and co-workers26 developed a Cu-catalyzed oxidative Ritter-type reaction using 1-(chloromethyl)-4-fluoro-1,4-diazoniabicyclo[2.2.2]octane bis(hexafluorophosphate) (F-TEDA-PF6) as an oxidant and CH3CN as both the reaction solvent and reaction partner. The transformation showed excellent selectivity at the β-position of ketones with acceptable yields (Scheme 1b). Unfortunately, CH3CN was used as the reaction solvent in this excellent work, resulting in only the corresponding acetamides; this drawback of the Ritter reaction is one of the most important issues to be addressed to improve both the diversity of the produced amides and the atomic economy of the Ritter reaction, although some Ritter reactions using other nitriles have also been reported.27,28 Surprisingly, the related Ritter reaction at the α-position of the carbonyl group has been unexplored. This may be due to the following: (1) although an umpolung at the α-position of the carbonyl group is a straightforward strategy to achieve this transformation,29 it is challenging because the generated carbocation intermediate is highly unstable due to the destabilizing effect of the adjacent carbonyl group; (2) the resulting carbocation intermediate could lead to undesired side reactions, such as oxidative dehydrogenation to produce enone byproduct and other competitive transformations at the α-position of the carbonyl group; and (3) the regioselectivity at the α'-position of the carbonyl group of unsymmetric ketones is another problem. To fill this gap in this type of Ritter reaction, herein we present an oxidative Ritter-type reaction of 2-arylketone with nitrile (10 equiv) and its application toward the collective total syntheses of erythrina alkaloids (Scheme 1c). Figure 1 | Selected functional organic molecules bearing N-acyl aza-quaternary carbon centers. Download figure Download PowerPoint Experimental Methods General procedure for oxidative Ritter-type reaction To a 10 mL sealed tube containing a magnetic stir bar were added sequentially α-arylketone (0.2 mmol), 2,3-dicyano-5,6-dichlorobenzoquinone (DDQ, 90.8 mg, 0.4 mmol), nitrile (2.0 mmol), HBF4 (50 wt % aqueous solution, 50.0 μL, 0.4 mmol), and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, 1.0 mL) at room temperature. Then the mixture was placed in a preheated oil bath at 60 °C and stirred at this temperature. After the indicated time (for details, please see Supporting Information), the reaction mixture was cooled to room temperature, diluted with dichloromethane (20.0 mL), washed with water and brine, dried over Na2SO4, filtered, and concentrated under reduced pressure. The residues were purified by flash column chromatography on silica gel to afford the desired amides. Scheme 1 | An overview of oxidative Ritter-type reactions involving carbonyl substrates. Download figure Download PowerPoint Results and Discussion Screening for the optimization of reaction conditions To achieve the expected oxidative Ritter-type reaction, the reactive carbocation intermediate generated at the α-position of the carbonyl group through an oxidative procedure needs to be stabilized. We opted for the installation of an aryl group as a carbocation stabilizing group at a suitable position on a substrate as an alternative strategy. Based on this logic, we selected 2-phenylcyclohexanone ( 2a)30,31 as a model substrate and 10 equiv of CH3CN ( 3a) as another reaction partner. After numerous solvent screenings, we found that the desired product 4aa could be obtained in 17%, 21%, and 57% yields when the reaction was performed in 1,2-dichloroethane (DCE), CH3NO2, and HFIP,32 respectively, using DDQ as an oxidant and p-toluenesulfonic acid monohydrate (PTSA) as an acidic promoter (entries 1–3, Table 1), and no product 4aa was isolated in other tested solvents (for crystallographic data for the product 4aa, please see Supporting Information Figure S2). Encouraged by these preliminary experimental results, we further screened other reaction parameters to improve the yield of product 4aa. Various Lewis and protonic acids were evaluated (entries 4–10, Table 1), and 2.0 equiv of HBF4 led to an increasing yield of desired product 4aa (entry 7, Table 1). Subsequently, a variety of oxidants were used to replace DDQ, resulting in unsatisfactory results (entries 11–15, Table 1). Finally, the reaction temperature was also examined (entries 16–18, Table 1), and both decreasing and increasing the reaction temperature was detrimental to the reaction outcome (for more details, please see Supporting Information Tables S1–S6). Therefore, the reaction parameters listed in entry 7 (Table 1) were selected as the optimal reaction conditions for further investigation. Table 1 | Screening for the Optimization of Reaction Conditionsa Entry Acid Oxidant Solvent Temperature (°C) Yieldb (%) 1c,d,e PTSA DDQ DCE 60 17 2 PTSA DDQ CH3NO2 60 21 3f PTSA DDQ HFIP 60 57 4 HCl DDQ HFIP 60 16 5 H3PO4 DDQ HFIP 60 14 6 H2SO4 DDQ HFIP 60 59 7 HBF4 DDQ HFIP 60 80 8g ZnCl2 DDQ HFIP 60 NR 9h Fe(OTs)3 DDQ HFIP 60 51 10i Fe(OTf)3 DDQ HFIP 60 59 11j HBF4 p-BQ HFIP 60 36 12k HBF4 CAN HFIP 60 8 13 HBF4 Na2IO4 HFIP 60 10 14 HBF4 K2S2O8 HFIP 60 7 15l HBF4 Selectfluor HFIP 60 52 16 HBF4 DDQ HFIP 40 72 17 HBF4 DDQ HFIP 50 79 18 HBF4 DDQ HFIP 70 76 aOtherwise specified, reactions were performed using 2-phenylcyclohexanone (0.2 mmol), acids (2.0 equiv), and oxidant (2.0 equiv) in 1.0 mL solvent at the noted temperature in a 10 mL sealed tube. bIsolated yield. cPTSA = p-Toluenesulfonic acid monohydrate. dDDQ = 2,3-Dicyano-5,6-dichlorobenzoquinone. eDCE = 1,2-Dichloroethane. fHFIP = 1,1,1,3,3,3-Hexafluoro-2-propanol. gNR = No reaction. hTs = p-Toluenesulfonyl. iTf = Trifluoromethanesulfonyl. jp-BQ = p-Benzoquinone. kCAN = Ceric ammonium nitrate. lSelectfluor = 1-(Chloromethyl)-4-fluoro-1,4-diazoniabicyclo[2.2.2]octane bis(tetrafluoroborate). Investigation of substrate scope With the optimal reaction conditions in hand, we started to investigate the scope of the substrates, and a variety of nitriles were first evaluated. With respect to the alkylic nitriles, all tested acyclic and cyclic nitriles reacted well under the optimal reaction conditions (Table 2). n-Butyronitrile, n-pentonitrile, cyclopropaneacetonitrile, benzeneacetonitrile, and 2-nitrophenylacetonitrile generated the desired amides 4ab– 4af, respectively, in good yields. Similarly, cyclopropanecarbonitrile, cyclopentonitrile, and cyclohexanecarbonitrile produced the expected amides 4ag– 4ai, respectively. Subsequently, a series of substituted alkylnitriles were investigated, and chloroacetonitrile, bromoacetonitrile, iodoacetonitrile, 4-chlorobutylnitrile, 3-hydroxylpropionitrile, 4-(1,3-dioxoisoindolin-2-yl)butanenitrile, 3-butenenitrile, and 5-hexyne-1-nitrile were proven to be viable substrates, affording the corresponding products ( 4aj– 4aq) in good yields. Interestingly, the O-bridging dinitrile ( 3r) also reacted well, and monoamide 4ar was isolated. Moreover, monoamide 4as was obtained in 63% yield when adiponitrile ( 3s) was applied; while the amount of ketone 2a was changed to 6 equiv, diamide 4at could also be isolated in 64% yield; remarkably, the resulting amides 4ar– 4at were the monomer analogs of nylon 6, 6.33 Notably, halogen atoms (Cl, Br, and even I), free OH, terminal alkene, terminal alkyne, and N-phthalyl protected functional groups in acyclic alkyl nitrile were tolerated under the optimal reaction conditions. Additionally, malononitrile and ethyl cyanoacetate, two commercially available nitriles bearing a strong electron-withdrawing group (EWG) at the α-position, also proved to be ideal substrates, delivering desired products 4au and 4av in 55% and 40% yields, respectively. With respect to the alkenylic nitriles ( 3w– 3az), the reaction also proceeded uneventfully to afford the corresponding products ( 4aw– 4az) in good yields. It is worth mentioning that 2-chloroacrylonitrile ( 3y) and benzylidenemalononitrile ( 3z) reacted efficiently, and chlorine atom and another CN group, respectively, were entirely retained in the resulting products; for nitrile 3z, E/Z mixtures were obtained in 74% total yield with a 4.2:1 ratio, and the major (E)-isomer was further confirmed by X-ray diffraction (see Supporting Information Figure S3). Overall, cyclic nitriles and acyclic nitriles, mononitriles and dinitriles, or conjugated and nonconjugated nitriles could generate sterically hindered amides; therefore, this transformation opened a new avenue to access diverse functional amides bearing an aza-quaternary carbon center. Table 2 | The Scope of Alkyl and Alkenyl Nitrilesa aOtherwise specified, the reactions were performed using 2-phenylcyclohexanone (0.2 mmol), HBF4 (2.0 equiv), DDQ (2.0 equiv) and nitrile (10.0 equiv) in 1.0 mL HFIP at the 60 °C in a 10 mL sealed tube. b5.0 equiv of nitriles was used. cThe reaction was performed using adiponitrile (0.1 mmol), 2-phenylcyclohexanone (6.0 equiv), HBF4 (9.0 equiv), and DDQ (7.2 equiv) in 2.5 mL HFIP. To further examine the flexibility of the current reaction toward various nitriles, we then sought to evaluate the aryl nitriles (Table 3). Monosubstituted substrates bearing either EWGs and electron-donating groups (EDGs) on the aryl rings or substituents at the o-, m-, or p-positions on the aryl rings did not obviously influence the reaction results, and the desired benzamides ( 5aa′– 5ap′) were isolated in yields ranging from 64% to 89%. Disubstituted aryl nitriles were also suitable substrates, giving the corresponding products 5aq′– 5at′. Moreover, terephthalonitrile ( 3u′) reacted well with excellent selectivity, providing monoamide 5au′ retaining another intact CN group. The reaction also proceeded smoothly with pharmaceutically important heteroaryl nitriles, such as 2-furonitrile ( 3v′), 2-thiophenecarbonitrile ( 3w′), 3-thiophenecarbonitrile ( 3x′), and three different substituted indole nitriles ( 3y′– 3aa′), providing the expected amides 5av′–5aaa′ in good yields. In these procedures, in addition to a few liquid arylnitriles, solid arylnitriles delivered the desired benzamides in 1 mL HFIP as solvent. Therefore, the current transformation has solved a long-existing problem of the classical Ritter reaction, which is generally performed in liquid nitrile (as both solvent and reagent); hence, limited examples of applying solid aryl nitriles in the Ritter reaction have been reported.34 Remarkably, not only were various functional groups (CF3, OMe, free OH, etc.) on the aryl ring tolerated under the optimal reaction conditions, but also the resulting benzamides bearing diverse substituents (such as OH, F, Cl, Br, I, CO2Me, and NO2) and different skeletons (aryl ring and heteroaryl ring) could be easily converted into other valuable and versatile molecules in synthetic, medical, and material chemistry.35 Table 3 | The Scope of Aryl- and Heteroaryl-Nitrilesa aOtherwise specified, the reactions were performed using 2-phenylcyclohexanone (0.2 mmol), HBF4 (2.0 equiv), DDQ (2.0 equiv) and nitrile (10.0 equiv) in 1.0 mL HFIP at the 60 °C in a 10 mL sealed tube. b5.0 equiv of nitriles was used. Next, the scope of another reaction partner, namely α-aryl ketones, was investigated under the optimal reaction conditions, and the reaction results are shown in Table 4. With respect to cyclic ketones, the tested six-membered ring substrates bearing either EWGs or EDGs on the aryl ring reacted smoothly with acetonitrile ( 3a), affording desired amides 6ba– 6na in good yields. Notably, although two reaction sites existed in the ketone 2o, monoamide 6oa was still isolated in a moderate yield, whose structure was further confirmed by X-ray analysis (see Supporting Information Figure S4). Moreover, 2-phenyl-3,4-dihydronaphthalen-1(2H)-one ( 2p) and 1-methyl-3,4-dihydronaphthalen-2(1H)-one ( 2q), two six-membered ring variants, were proven to be ideal substrates, resulting in products 6pa and 6qa in 85% and 71% yields, respectively. The ring-size effects of α-arylketones were then investigated, and 2-phenyl-2,3-dihydro-1H-inden-1-one ( 2r) and cycloheptanone derivative ( 2s) reacted smoothly, affording the corresponding amides 6ra and 6sa, respectively. Similarly, with respect to the linear ketones, desired products 6ta– 6xa could also be isolated in satisfactory yields under the optimal reaction conditions. Table 4 | The Scope of Aryl Substituted Ketones and Some Deuterium Amidesa aThe reactions were performed using ketones (0.2 mmol), HBF4 (2.0 equiv), DDQ (2.0 equiv) and CH3CN or CD3CN (10.0 equiv) in 1.0 mL HFIP at the 60 °C in a 10 mL sealed tube. Recently, the replacement of hydrogen with deuterium has been widely applied in different disciplines.36 Therefore, the preparation of deuterium-labeled organic molecules has received considerable interest from synthetic chemists. In terms of synthetic chemistry, the replacement of CH3CN with CD3CN could conveniently produce the corresponding deuterium amides using our developed synthetic method. To our delight, five anticipated deuterium products, namely, 4aa-D3, 6ga-D3, 6ia-D3, 6qa-D3, and 6sa-D3, were obtained by simply switching 10 equiv of acetonitrile to the same amount of CD3CN correspondingly, and the results are listed in Table 4. These experimental results indicate that the current transformation provides an alternative approach to access highly sterically hindered deuterium acetamides in good yields by using the readily available reagent CD3CN. As evidenced by the abovementioned substrate scope studies (Tables 2–4), we could conclude that our developed oxidative Ritter-type reaction tolerates a wide substrate scope, including various nitriles or diverse α-arylketones, providing a concise approach toward synthetically challenging amides. Mechanistic investigations After the expansion of the substrate scope of the reaction, we turned our attention to understanding the reaction mechanism by designing and performing control experiments (Scheme 2). First, when 2-phenylcycloketone ( 2a) was replaced by cyclohexylbenzene ( 7) under the optimal reaction conditions, no desired amide 7a was isolated, indicating that the carbonyl group is crucial and that enolization may be a vital process in this oxidative Ritter-type reaction (Scheme 2, eq 1). Second, when the model reaction was performed for 10 min, the hydroxyl product 7a was isolated in 18% yield (along with amide 4aa in 67% yield and trace enone 8b), demonstrating that the carbocation intermediate could participate in this transformation (Scheme 2, eq 2). Under the optimal reaction conditions, hydroxyl product 8a was transformed to the desired amide 4aa in 78% yield, and this result did not exclude that hydroxyl product 8a could be a precursor of this transformation, while enone 8b was excluded as a reaction intermediate because no desired amide 4aa was isolated (Scheme 2, eq 3). Moreover, when cyclopentanone derivative 2y was used as a substrate, the expected amide 2ya was isolated in 25% yield, albeit with byproduct enone 2yb (in 18% yield), further verifying that the carbocation intermediate should be involved in this transformation (Scheme 2, eq 4). Finally, when the radical scavenger (2,2,6,6-tetramethylpiperdin-1-yl)oxyl was additionally added under the optimal reaction conditions, the desired amide 4aa was isolated in 11% yield along with compound 8a (28% yield) during the same reaction time (90 min), implying that a radical in intermediate may be quickly oxidized to the corresponding carbocation intermediate by DDQ in this transformation (Scheme 2, eq 5). Scheme 2 | Some control experiments designed for investigation of the reaction mechanism. Download figure Download PowerPoint Based on these control experimental results and the reported literature,37 a possible reaction mechanism was proposed (Scheme 3). First, the enolization of substrate 2a could occur under acidic reaction conditions to generate intermediate I,38 which was oxidized through single electron transfer by DDQ to produce radical intermediate I-A, and DDQ was simultaneously reduced to release intermediate I-B. Further oxidation of radical intermediate I-A by intermediate I-B afforded the carbocation intermediate II-A, which was trapped by acetonitrile to furnish amide 4aa through the classical Ritter reaction. If intermediate II-A was not immediately trapped by acetonitrile, it was transformed to alcohol 8a through a hydroxylation process and/or another side-product enone 8b through a proton loss procedure. Alternatively, under acidic reaction conditions, the key intermediate II-A was regenerated from alcohol 8a, and enone 8b was obtained through the elimination reaction from alcohol 8a. Scheme 3 | The proposed reaction mechanism. Download figure Download PowerPoint Collective total syntheses of erythrina alkaloids Inspired by the above-mentioned reaction results and to further expand the synthetic application of our developed oxidative Ritter-type reaction, we targeted tetracyclic erythrina alkaloids bearing an N-acyl aza-quaternary carbon center. More than 140 erythrina alkaloids have been isolated and identified to date, which exhibit broad bioactivities, including hypotensive, sedative, central nervous system depressant, anti-HIV, and curare-like activity.10,39–41 Hence, diverse synthetic strategies toward erythrina alkaloids have been developed by synthetic communities39–52 from the early pioneers Belleau, Modndon, and Prelog in the 1950s to Tsuda in the 1980s. Recent contributions from Padwa, Tietze, Reisman, Kim, You, Tu, Zhang, and others further advance the research area39–41 and inspire continued synthetic interest, as amplified by the latest studies from the groups of Fukuyama and Kitamura,47 Booker-Milburn,48 Vital and Kristensen,49 Tong and Wang,50 Xu,51 and She.52 In contrast to the previously reported synthetic strategies, we speculated that the target molecules could be synthesized through a novel aza-quaternary-carbon-center-guided disconnected strategy.53 Namely, we anticipate first introducing the C-5 spiroamide moiety of target molecules and subsequently constructing and functionalizing the B, A, and C rings. Starting from two commercially available building blocks, 2-(3,4-dimethoxyphenyl)cyclohexan-1-one ( 2z) and bromoacetonitrile ( 3k), a key oxidative Ritter-type reaction not only installed a challenging C-5 aza-quaternary carbon center but also enlarged the two-carbon chain bearing a synthetically valuable bromine atom for the subsequent ring-closing transformation under the slightly modified reaction conditions (Selectfluor54 and PTSA were used to replace DDQ and HBF4, respectively; for details, please see the Supporting Information Tables S7 and S8), resulting in the desired racemic amide 9. With brominated amide 9 in hand, Friedel–Crafts alkylation,55 radical cyclization,56 and photo-Witkop reaction57,58 were applied to assemble a six-membered C-ring via an intramolecular approach, respectively, and numerous reaction conditions were examined; unfortunately, no expected C-ring-forming spirolactam was isolated under any of the conditions tried, and amide 9 was either degraded to produce a complex mixture or recovered intact. Therefore, the order of the ring-forming sequence was adjusted, and an intramolecular Reformatsky reaction was applied to first assemble the five-membered B ring.59 To our pleasure, the reaction proceeded smoothly to afford the expected tricyclic lactam 10 in 74% yield. Notably, the ABD tricyclic skeleton of erythrina alkaloids was rapidly assembled via just two steps from commercially available starting materials. Next, treatment of the resulting tertiary hydroxy group with thionyl chloride installed the entirely desired functionalities of the B ring to relieve α,β-unsaturated γ-lactam,60 which was further reacted with phenyl vinyl sulfoxide to yield aza-Michael adduct 11.61 We note that this conjugate adduction not only excluded the negative effect of the free N-H group in the subsequent chemical transformations but also introduced the two-carbon chain-bearing sulfoxide functional group to construct the C ring of target molecules. With compound 11 in hand, the adjustment of the oxidative states of the A ring and subsequent installation of the desired functionalities were carried out through a redox-economic strategy. The first oxidation of compound 11 with SeO246,62,63 and subsequent elimination of the resulting hydroxyl group introduced the desired C1–C2 double bond in the A ring of compound 12.60 Second stereospecific oxidation at the allylic position of the resultant product 12 installed the desired C-3 α-hydroxyl group, and the subsequent treatment of the resulting allylic alcohol (not shown) with MeI produced tricyclic intermediate 13,46 which possesses the entire functional groups of the ABD ring of the natural alkaloids. After assembly of the ABD tricyclic skeleton and further introduction of all desired functional groups, an acid-promoted Pummerer-type cyclization of sulfoxide 13 conveniently completed the construction of the C ring in the target products, affording key precursor 14 in 72% yield.61,64 Remarkably, the highly functionalized tetracyclic skeleton was concisely constructed through a linear nine-step transformation with a protecting-group-free synthetic procedure (Scheme 4). Scheme 4 | The concise syntheses of the key precursor 14. Download figure Download PowerPoint In the final stages of the syntheses of erythrina alkaloids, we selected diverse transformations to achieve collective total syntheses through a one-step procedure based on rich transformations of sulfur chemistry.65,66 First, the radical desulfurization of key precursor 14 using Bu3SnH/azobisisobutyronitrile (AIBN) in toluene61 afforded erysotramidine ( 1a) in 82% yield.46,67,68 Second, selective desulfurization of key precursor 14 using Hg(OTf)2 produced the carbocation intermediate, which was trapped by various nucleophiles to produce diverse erythrina alkaloids. When the intermediate was captured by MeOH, 11-α-methoxyerysotramidine ( 1d)69 and the proposed 11-β-methoxyerysotramidine ( 1c)70 were isolated in 32% yield and 38% yield, respectively; when it was trapped by H2O, natural 11-β-hydroxyerysotramidine ( 1e)68 was obtained in 38% yield, along with its isomers 11-α-hydroxyerysotramidine ( 1m, 42% yield). Furthermore, the resulting carbocation intermediate lost a proton in anhydrous tetrahydrofuran (THF) to complete the total syntheses of erytharbine ( 1b, 88% yield).67,68,71 Encouraged by these exciting results, we next focused on the more challenging transformations to install two carbonyl functional groups at the C ring in a one-step procedure. After a number of oxidative conditions were tested, we gratifyingly found that SeO2 oxidation of precursor 14 at 130 °C produced 10,11-dioxoerysotramidine ( 1f)72 in 53% yield, whose structure was further verified by single-crystal X-ray diffraction (see Supporting Information Figure S1). Unfortunately, the spectral data of the synthetic product did not match the reported one.72 Finally, a similar SeO2 oxidation of precursor 14 at 110 °C yielded erytharbine