Concise Stereoselective Synthesis of (−)‐Podophyllotoxin by an Intermolecular Iron(III)‐Catalyzed Friedel–Crafts Alkylation

Without further ado, the building blocks 1–3 were combined in three CC bond-forming reactions to provide the enantiomerically pure natural product (−)-podophyllotoxin (4). The stereogenic center at C1 was generated in the key reaction, a diastereoselective iron(III)-catalyzed intermolecular Friedel–Crafts alkylation. Owing to their diverse biological activities, in particular their cytotoxicity and antiviral properties, podophyllotoxin (1) and its derivatives are important members of the lignane class of natural products.1 The biological profile of these compounds, which has inspired the design of new drugs, has been investigated intensively. The biological activity of podophyllotoxin has generated strong interest in the development of synthetic routes to this natural product.2 Since the first synthesis of enantiomerically pure (−)-podophyllotoxin (1) by Meyers and co-workers3 in a 24-step sequence, another five syntheses of the enantiomerically pure natural product have been reported.4, 5 Many additional studies have been concerned with the development of formal total syntheses and syntheses of racemic podophyllotoxin.2 In all synthetic approaches to podophyllotoxin, the modular generation of the tetracyclic backbone and the brevity of the sequence are of central importance. We report herein a six-step total synthesis of enantiomerically pure (−)-podophyllotoxin with an iron(III)-catalyzed Friedel–Crafts alkylation as a key step. Apart from the use of a terminal alkene as an aldehyde equivalent, no protecting-group manipulations were required in the entire synthetic sequence. Our retrosynthesis involved the disconnection of the target molecule into the three similarly sized fragments 2, 3, and 4 (Scheme 1).6 We planned to use an intramolecular arylation to complete the tetracyclic core. Possible cyclization methods included a Heck reaction, a Nozaki–Hiyama coupling, and a classical Lewis acid mediated hydroxyalkylation. A diastereoselective intermolecular Friedel–Crafts alkylation was designed as the pivotal step, which should establish the stereogenic center at C1. Retrosynthetic disconnection of (−)-podophyllotoxin (1) into fragments 2, 3, and 4. In accordance with the outlined plan, the synthesis commenced with the Taniguchi lactone (3), which is accessible in enantiomerically pure form from 2-butyne-1,4-diol in two steps with a subsequent conventional resolution7 or in six steps through an enantioselective iridium-catalyzed allylation.8 An aldol reaction with aldehyde 4 afforded 5 with excellent stereoselectivity with respect to the stereogenic center at the α position of the lactone (Scheme 2). The fact that the simple diastereoselectivity was low, as expected, was of no relevance, as the hydroxy group was substituted with an aryl group in the next step in a SN1-type alkylation. To this end, 5 was treated with 1,3-benzodioxole (2 a, X=H) and other derivatives 2 under acid catalysis. On the basis of our previous studies on diastereoselective Friedel–Crafts alkylation reactions with chiral benzylic carbenium ions,9 we expected the stereoisomer 6 to be the predominant product. Aldol reaction und subsequent intermolecular Friedel–Crafts alkylation: a) LDA (1.1 equiv), THF, −78 °C, 30 min, then 4 (1.1 equiv), −78 °C, 3 h, 94 % (d.r. 52:48); b) see Table 1. LDA=lithium diisopropylamide. Whereas the reaction of 2 a with 5 proceeded smoothly under the usual reaction conditions (HBF4⋅OEt2 in CH2Cl2) with good diastereoselectivity to give product 6 a (d.r. 85:15; Table 1, entry 1), the analogous reactions of the substituted derivatives 2 b (X=Br), 2 c (X=OTf), and sesamol (2 d, X=OH) failed (Table 1, entries 2–4). Only O-allyl-protected sesamol 2 e underwent the desired reaction to afford products 6 e and 7 e (d.r.=84:16; Table 1, entry 5). The discovery that this last reaction was also promoted by a catalytic amount of Bi(OTf)3 (Table 1, entry 6)10 prompted us to evaluate Bi catalysis with substrates 2 b–2 d. Under these conditions, sesamol (2 d) was transformed diastereoselectively into the desired product 6 d (d.r. 90:10; Table 1, entry 9); no conversion was observed with 2 b and 2 c (entries 7 and 8). Further optimization studies led to the identification of FeCl3,11 which outperformed even AuCl3 (Table 1, entry 10), as the optimum catalyst for the transformation of 5 into 6 d (entry 11).12 The product 6 d was formed with high diastereoselectivity in nearly quantitative yield (d.r. 94:6; Table 1, entry 11). Entry X 2 Acid (mol %) Solvent T [°C] t [min] Yield [%][c] d.r.[d] 6/7 1[a] H 2 a HBF4 (400) CH2Cl2 20 45 76 85:15 2[a] Br 2 b HBF4 (400) CH2Cl2 20 45 – – 3[a] OTf 2 c HBF4 (400) CH2Cl2 20 45 – – 4[a] OH 2 d HBF4 (400) CH2Cl2 20 45 – – 5[b] OAllyl 2 e HBF4 (125) CH2Cl2 −78→20 15 97 84:16 6[b] OAllyl 2 e Bi(OTf)3 (10) MeNO2 20 80 94 77:23 7[b] Br 2 b Bi(OTf)3 (10) MeNO2 20 80 – – 8[b] OTf 2 c Bi(OTf)3 (10) MeNO2 20 120 – – 9[b] OH 2 d Bi(OTf)3 (10) MeNO2 20 80 95 90:10 10[b] OH 2 d AuCl3 (10) MeNO2 20 60 98 90:10 11[b] OH 2 d FeCl3 (5) CH2Cl2 20 60 99 94:6 For the completion of the synthesis, different cyclization methods were evaluated. Unfortunately, an attempted Lewis acid catalyzed ring closure, which would have given rise to the shortest possible sequence, did not afford the desired product. Although oxidative cleavage of the terminal alkene to provide the aldehyde proceeded smoothly, electronic and conformational factors13 prohibited the envisaged cyclization onto the 1,3-benzodioxole substituent. Instead, under a variety of conditions (one example is shown in Scheme 3), cyclization always occurred by electrophilic attack at the trimethoxyphenyl substituent to generate the podophyllotoxin isomer 8. Undesired regioselectivity in the BF3-catalyzed cyclization of 6 a: a) OsO4 (5 mol %), NMO (3 equiv), CH2Cl2, 20 °C, 4 h, then NaIO4 (2 equiv), 30 min, 92 %; b) BF3⋅OEt2 (10 equiv), CH2Cl2, −78 °C, 3 h, 41 %. NMO=N-methylmorpholine N-oxide. Reactions of triflate 6 c, which can be prepared readily from alcohol 6 d (Scheme 4), were more successful. A Heck reaction14 led to the desired cyclized product 9.15 An attempted Nozaki–Hiyama coupling reaction16 investigated as an alternative cyclization procedure (after oxidative cleavage of the terminal alkene to give the corresponding aldehyde) did not lead to podophyllotoxin. Olefin 9 was converted smoothly by dihydroxylation and periodate cleavage into podophyllotoxon,17 which was reduced diastereoselectively to 1 by a previously reported procedure.18 Completion of the total synthesis of (−)-podophyllotoxin (1): a) Tf2O (1.5 equiv), NEt3 (2 equiv), CH2Cl2, 0 °C, 1 h, 89 %; b) Pd(OAc)2 (10 mol %), PPh3 (0.3 equiv), K2CO3 (3 equiv), MeCN, 80 °C, 20 h, 58 %; c) OsO4 (5 mol %), NMO (3 equiv), CH2Cl2, 20 °C, 4 h, then NaIO4 (2 equiv), 30 min, 95 %; d) LiAlH(OtBu)3 (10 equiv), Et2O, −78→20 °C, 18 h, 79 % (d.r. 98:2). Tf=trifluoromethanesulfonyl. All physical properties of synthetic (−)-podophyllotoxin (1), which was obtained from the Taniguchi lactone (3) in 35 % overall yield, were identical to those of the natural product. The synthesis illustrates that a stereogenic center in the β position to an ester or lactone moiety can be constructed highly diastereoselectively through a Lewis acid catalyzed SN1 reaction if a stereogenic center is already present in the α position. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2002/2008/z802611_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.