Dedicated to Professor Reinhard Brückner (Albert-Ludwigs University Freiburg) on the occasion of his 64th birthday
3
Applications of the Dötz Benzannulation Reaction in Natural Products Synthesis over the Last Decade
The Dötz benzannulation has been steadily used over the last decade in the synthesis of various natural products (Figures [1] and 2). While there have been previous reviews covering different aspects of this reaction, in this review, its strategic use in natural products synthesis has been abstracted from 2008 onwards.
3.1
Synthesis of (+) and (–)-Juglomycin A (1)[20]
[21]
Scheme 2 Synthesis of (+)- and (–)-juglomycin A
We started working in the area of the strategic use of the Dötz benzannulation reaction in natural products synthesis back in 2008, with the first installment of our work disclosing the synthesis of both enantiomers of juglomycin A (1).[20] This compound was isolated from the culture filtrate of the fungus Streptomyces sp. 190–2 and shows antitumor as well as antibacterial activities against both Gram-negative and Gram-positive bacteria.[22] Both enantiomers of juglomycin A were synthesized utilizing the Dötz benzannulation reaction (Scheme [2]). To begin with, the condensation of Fischer carbene complex 28
[11d] with alkyne 29
[23] afforded the Dötz benzannulated product 30 in a good yield of 78%. Methylation of the phenolic hydroxy group with MeI and TBS deprotection provided 31 in 81% yield (two steps). Swern oxidation of 31 to the corresponding aldehyde and condensation with the half ester of malonic acid under decarboxylative–deconjugative–Knoevenagel conditions[24] gave the β,γ-unsaturated ester 32 in 72% yield (over two steps). Sharpless asymmetric dihydroxylation[25] of 32 with the (DHQD)2PHAL ligand afforded β-hydroxy-γ-lactone 33 in 84% yield and an excellent enantioselectivity of 99.5% ee. Treatment of 33 with cerium(IV) ammonium nitrate (CAN) gave the quinone (94%) and demethylation with AlCl3 afforded (–)-juglomycin A (1) in 92% yield. Similarly, the unnatural enantiomer ent-1 was synthesized from 32 through asymmetric dihydroxylation using the (DHQ)2PHAL ligand to give ent-33 in 83% yield and 98.5% ee. Quinone formation and demethylation gave (+)-juglomycin A (ent-1).
Later in 2011, we considered an alternative synthesis of (+)- and (–)-juglomycin A (1) using a functionalized alkyne (Scheme [3]).[21] The reaction of Fischer carbene 28 with alkyne 34 gave the naphthol 35 in 68% yield. Methylation of the naphthol and TBS group removal furnished the alcohol 36. Oxidation of the latter to the acid and lactonization then gave the lactone 33. Finally, quinone formation and demethylation provided (–)-juglomycin A (1). Similarly, the use of the alkyne ent-34 led to (+)-juglomycin A (ent-1).
Scheme 3 An alternative synthesis of (+)- and (–)-juglomycin A
3.2
Synthesis of (+)-Eleutherin (2a), (+)-allo-Eleutherin (2b)and a Formal Synthesis of (+)-Nocardione B (3)[26]
[27]
We have also considered concise enantioselective syntheses of (+)-eleutherin (2a) and (+)-allo-eleutherin (2b) as well as the formal synthesis of (+)-nocardione B (3) (Scheme [4]). (+)-Eleutherin (2a) and (–)-isoeleutherin (ent-2b) were isolated from the bulbs of Eleutherin bulbosa
[28] in 1950. Eleutherin shows activity against Bacillus subtilis.[29] Extracts of Eleutherin americana, of which eleutherin and isoeleutherin are the major constituents, have been used to treat the heart disease angina pectoris.[30] (–)-Nocardione B (ent-3), showing moderate antifungal and cytotoxic activities, was isolated by Otani et al. in 2000 as a new tyrosine phosphate inhibitor.[31] Toward their synthesis, the Fischer carbene complex 28 was condensed with the chiral alkyne 37 via the Dötz benzannulation reaction to give the naphthol 38 in 69% yield (Scheme [4]). Deprotection of the TBS group and an oxa-Pictet–Spengler reaction with (MeO)2CHMe gave the 7-membered cyclic acetal in 93% yield with the phenolic OH group involved. This on reaction with CAN resulted in the quinone 41 in 83% yield. The conversion of quinone 41 into (+)-nocardione B (3) has already been reported in the literature.[32] Hence, this work describes the formal synthesis of (+)-nocardione B (3). For the synthesis of eleutherin and allo-eleutherin, the phenolic OH in 38 was protected to give the methyl ether and TBS group removal then gave 39. The subsequent oxa-Pictet–Spengler reaction provided a mixture of 40a/40b in a 36:64 ratio. This mixture was separated by preparative TLC and oxidation with CAN produced (+)-eleutherin (2a) and (+)-allo-eleutherin (2b), respectively.
Scheme 4 Synthesis of (+)-eleutherin (2a), (+)-allo-eleutherin (2b) and (+)-nocardione B (3)
An improved synthesis of (+)-eleutherin (2a) and (+)-allo-eleutherin (2b) was subsequently reported by us (Scheme [5]).[27] The intermediate 39 was prepared as reported by us earlier (Scheme [4]). With the success in obtaining the cis-isomer in the synthesis of demethoxycardinalin-3 using dry HCl gas[33] (this work is discussed later), we considered similar conditions for the synthesis of eleutherin, aiming for either diastereomer to be obtained as the major product. Thus, the reaction of 39 with (MeO)2CHMe/BF3·OEt2 in CH2Cl2 at room temperature for 10 hours gave a mixture of 40a/40b in a 21:79 ratio and isolated yields of 16% and 70%, respectively. However, the reaction by bubbling dry HCl gas through the mixture of 39 and (MeO)2CHMe in Et2O at 0 °C for 2.5 hours provided a switched diastereoselectivity for 40a/40b with a 78:22 ratio and isolated yields of 66% and 18%, respectively. The separated diastereomers on CAN oxidation gave (+)-eleutherin (2a) and (+)-allo-eleutherin (2b) in 86% and 89% yields, respectively.
Scheme 5 An improved synthesis of (+)-eleutherin (2a) and (+)-allo-eleutherin (2b)
3.3
Synthesis of (–)-Hongconin (4)and (–)-1-epi-Hongconin (4′)[34]
(–)-Hongconin (4) was isolated from the rhizomes of Eleutherine americana by Chen and co-workers.[35] Pharmacological studies showed that hongconin enhanced the blood flow of coronary arteries and also reduced chest pain.[36] Our group has completed a concise enantioselective synthesis of both (–)-hongconin (4) and (–)-epi-hongconin (4′) (Scheme [6]).[34] The Fisher carbene complex 28 was reacted with the alkyne 42 in a Dötz benzannulation reaction to provide the corresponding substituted naphthol, which on methylation of the phenolic OH followed by benzyl deprotection gave 43. An oxa-Pictet–Spengler reaction of 43 with (MeO)2CHMe/BF3·OEt2 furnished a mixture of 44a/44b in a 3:2 ratio that was separated in 48% and 31% yields, respectively. Advantageously, the ketal group also underwent deprotection in this reaction. Subsequent CAN oxidation to the quinone and reduction with sodium dithionate gave (–)-hongconin (4). Similarly, CAN oxidation of 44b and sodium dithionate reduction produced non-natural (–)-epi-hongconin (4′).
Scheme 6 Synthesis of (–)-hongconin (4) and (–)-epi-hongconin (4′)
3.4
Synthesis of the Regioisomeric Core Structure of Cardinalin-3 (49)[37] and Demethoxycardinalin-3 (5)[33]
The cardinalins are a series of cytotoxic dimeric pyranonaphthoquinones isolated from the New Zealand toadstool Dermocybe cardinalis.[38] An ethanolic extract of the latter inhibits the growth of P388 murine leukemia cells (IC50 = 0.47 μg/mL).[38b] We executed a bidirectional Dötz benzannulation and oxa-Pictet–Spengler strategy toward the dimeric core structure of cardinalin-3 (49′) (Scheme [7]). The dimeric Fischer carbene 46 (prepared from 45) on reaction with alkyne 37 gave the dimer naphthol, which on methylation and TBS deprotection provided 47. A subsequent oxa-Pictet–Spengler reaction using (MeO)2CHMe/BF3·OEt2 gave a mixture of anti/anti and syn/anti dimeric pyran products, while the syn/syn diastereomer was the desired intermediate. Fortunately, bubbling dry HCl gas into the oxa-Pictet–Spengler reaction mixture gave the desired syn/syn-pyran 48 along with the syn/anti diastereomer (not shown), which were obtained in 56% and 16% isolated yields, respectively. CAN oxidation of the former gave the dimeric core structure 49 (75%) of cardinalin-3 (49′).
Scheme 7 Synthesis of the core structure 49 of cardinalin-3
The bidirectional Dötz benzannulation strategy was then extended to complete the total synthesis of demethoxycardinalin-3 (5) (Scheme [8]).[33] The Fischer carbene 50 and alkyne 37 reacted to give the dimeric naphthol, which on methylation and TBS removal gave 51. Subsequent oxa-Pictet–Spengler reaction under our successful HCl(g)[37] bubbling conditions gave the desired syn/syn-pyran product 52 along with the syn/anti diastereomer (not shown), obtained in 55% and 22% isolated yields, respectively. CAN oxidation of 52 and demethylation with AlCl3 furnished demethoxycardinalin-3 (5). This structure represents the closest chiral analogue synthesized for cardinalin-3 (49′) in the literature.
Scheme 8 Total synthesis of (+)-demethoxycardinalin-3 (5)
3.5
Total Synthesis of Kendomycin (19)[39]
Scheme 9 Nakata’s total synthesis of kendomycin (19)
Kendomycin (19), isolated from Streptomyces species, possesses an interesting ansa-type quinone methide framework which is directly fused with a substituted tetrahydropyran moiety and shows promising antibacterial and cytotoxic activities.[40] Nakata and co-workers have described a challenging synthesis of kendomycin using an intramolecular Dötz reaction (Scheme [9]). The synthesis commenced with the preparation of intermediate 54 from known alcohol 53
[41] in 12 steps. Meanwhile, compound 55 was prepared from ent-53
[42] in 10 steps.[39a] The Pd-catalyzed Suzuki–Miyaura cross-coupling of 54 with 55 followed by alcohol oxidation generated the ynone 56. Next, the ynone was subjected to acetal exchange followed by deoxygenation, then hydroxy protection as the TES ether and TMS/TBS removal gave the terminal alkyne 57. In the next step, the carbene 58 was anchored to the alcohol 57 and then subjected to an intramolecular Dötz reaction to produce the phenol 59 in 58% yield. Silyl protection of the phenol and aryl-Claisen rearrangement in the presence of Ac2O/DMAP gave the aryl acetate that was reduced back to the phenol and then protected as MOM ether 60. The electron-rich internal double bond was masked as a diol and subsequent cleavage of the terminal double bond using O3 gave the ketone 61. Regeneration of the internal olefin and silyl deprotection furnished the phenol that was oxidized to the 2-hydroxy-1,4-quinone with IBX resulting in 62. Finally, formation of the quinone methide on silica gel treatment and desilylation gave kendomycin (19). The synthesis was completed in 32 steps from known intermediate 53.
Scheme 10 Nakata’s second-generation total synthesis of kendomycin (19)
In a subsequent paper, Nakata and co-workers[39b] studied the simultaneous macrocyclization and intramolecular Dötz benzannulation reaction for construction of the ansa structure, thereby preparing some analogues and completing a new synthesis of kendomycin (19) (Scheme [10]). The previous synthesis (Scheme [9]) required masking of the more reactive double bond for the oxidative cleavage of the terminal double bond. The new synthesis varied at this point and at the end. Compound 63, obtained in the first two steps from 56 (see Scheme [9]), on TBS protection gave the bis-TBS compound and selective removal of the primary TBS and the TMS groups gave 64. A similar intramolecular Dötz benzannulation as before led to phenol compound 65 (this differs from 59 in having a TBS group). Subsequent TBS protection of the phenol and aryl-Claisen rearrangement in the presence of Ac2O/DMAP gave the aryl acetate that was reduced back to phenol 66. The latter on hydroxy-directed epoxidation also opened the epoxide to give a 2,3-dihydrobenzofuran with a primary alcohol that was oxidized to the acid 67. Removal of the TBS group and ortho oxidation with IBX gave the unstable orthoquinone 68 through decarboxylation. Finally, treatment with HF gave kendomycin (19). This completed a second-generation synthesis with an improved overall yield (0.74%) compared to the previous version (0.026%).
3.6
Stereoselective Synthesis of the Tetracyclic Naphthoquinone, (–)-Isagarin (6)[43]
Isagarin, a new type of tetracyclic naphthoquinone, was isolated by Van Puyvelde and co-workers from the roots of Pentas longiflora Oliv. (Rubiaceae),[44a] which possess a broad range of medicinal properties.[44`]
[c]
[d] A concise synthesis of (–)-isagarin by our group is depicted in Scheme [11].[43] The alkyne 70 was prepared from d-mannitol (69) as reported in the literature.[45] Addition of 70 to propylene oxide and acetylation resulted in the acetate 71. The Dötz benzannulation with carbene 72 and in situ methylation then gave naphthalene 73 in 56% yield. The in situ methylation helped to avoid the formation of regioisomers. Further, the sequence of acetate reduction, hydroxy oxidation and transketalization provided 74 in 91% yield (three steps) from 73. Finally, quinone formation produced (1S,4R)-(–)-isagarin (6) in 99% yield.
Scheme 11 Synthesis of (–)-isagarin (6)
3.7
Total Synthesis of (+)- and (–)-Arizonins B1 (12a) and C1 (12b)[46]
Brückner and co-workers reported a concise synthesis of arizonin C1 (12b) (Scheme [12]),[46a] a member of six anti-Gram-positive antibiotics isolated from Actinoplanes arizonaensis sp. nov., strain AB660D-122, by Hochlowski et al.[47] The Dötz benzannulation of Fischer carbene 75 with alkyne 29 furnished the naphthol that was methylated in the same-pot to give 76 in 42% yield. The latter, on TBS removal, hydroxyl oxidation and olefination under modified Knoevenagel condensation,[24] furnished the β,γ-unsaturated ester 77 (as an 88:12 mixture with the α,β-regioisomer). Asymmetric dihydroxylation then gave the lactone 78 in 79% yield and >99% ee. The oxa-Pictet–Spengler reaction with acetaldehyde furnished a mixture of C5-epimers in 86% yield and in a 66:34 ratio. Subsequent CAN oxidation afforded a mixture of 12b and 5-epi-12b in a 67:33 ratio and 80% yield. The epimer ratio of this mixture increased to 94:6 toward arizonin C1 (12b) on treatment with concentrated H2SO4.
Scheme 12 Total synthesis of (–)-arizonin C1 (12b) by Brückner and co-workers
We also achieved an efficient total synthesis of arizonins B1 (12a) and C1 (12b) by employing functionalized alkyne substrate 79 (Scheme [13]).[46b] The Dötz benzannulation of carbene 75 with alkyne 79 gave the corresponding naphthol (48%), which was methylated (85%) and then subjected to acid-mediated cyclization to give the lactone 78 in 87% yield. Subsequent oxa-Pictet–Spengler reaction and CAN oxidation furnished 12b′ and arizonin C1 (12b) in a 23:77 ratio. The AlCl3-mediated demethylation of the mixture gave the expected quinone mixture in 80% yield. Treatment of this mixture with concentrated H2SO4 resulted in C-5 epimerization giving 12a′/12a in a 6:94 ratio and a single recrystallization gave arizonin B1 (12a) in 52% yield from the mixture. On the other hand, stirring the mixture of 12b′/12b with concentrated H2SO4 provided directly arizonin B1 (12a) in 51% yield via a remarkable regioselective C-7 demethylation and C-5 epimerization. Ag2O/MeI-based methylation of 12a gave arizonin C1 (12b) in 77% yield. Alternatively, treatment with concentrated H2SO4 over a shorter time of 25 min gave only the C-5 epimerized product (12b′/12b = 6:94 ratio). A single recrystallization then gave arizonin C1 (12b) in 48% yield. This work described the first asymmetric synthesis of arizonin B1.
Scheme 13 Our syntheses of (+)-arizonin B1 (12a) and (+)-arizonin C1 (12b)
In our work on the synthesis of (+)-arizonins B1 and C1, we observed discrepancies in the signs of the optical rotations for both molecules compared to that reported for the natural isolate and that by Brückner.[46a] We also synthesized the enantiomer of arizonin C1, i.e., ent-12b, by following the literature method and found it to have the same value[46b] as the natural isolate and Brückner’s compound. Although at that stage we could not resolve these differences, a subsequent paper by Brückner and co-workers[46c] indicated that our report was correct and that probably their samples corresponding to lactone 78 had been exchanged between the two enantiomers.
Brückner’s alternative syntheses of (–)-arizonin B1 and (–)-arizonin C1 are depicted in Scheme [14]. The required substituted naphthalene 80 was prepared from isovanillin in six steps and a further 3 steps were required to arrive at bromo derivative 81. The latter on Heck coupling with methyl 3-butenoate gave 82 in 82% yield, containing 13% of the α,β-unsaturated ester isomer. Asymmetric dihydroxylation gave the lactone 83 (60%, 98.6% ee), which on oxa-Pictet–Spengler reaction provided the trans-pyran compound 84 (dr = 77:23) that was purified by flash chromatography to give the pure diastereomer. The Boc group was cleaved during this reaction and the free phenol compound 84 was found to decompose during the CAN oxidation for quinone formation. Hence Boc reprotection was executed followed by CAN oxidation to produce 85. TFA-mediated Boc removal then gave (–)-arizonin B1 (ent-12a). The latter was methylated to give (–)-arizonin C1 (ent-12b). The obtained optical rotations were in agreement with respect to the sign as obtained by us (Scheme [13]),[46b] and thereby confirming that their earlier synthesis (Scheme [12]) had sample of 78 that was actually of its enantiomer.
Scheme 14 Brückner’s alternative syntheses of (–)-arizonin B1 and (–)-arizonin C1 and confirmation of the correctness of the chemistry reported by our group in Scheme [13]
3.8
Synthesis of (+)-Ventiloquinone L (7) and (–)-1-epi-Ventiloquinone L (7′)[48]
Ventiloquinone L (7), with a syn-configured 1,3-dimethypyran, was isolated from the root bark of Ventilago goughii
[49a] and shown to inhibit specific topoisomerase II.[49b] In 2012, we completed a short synthesis of 7 (Scheme [15]).[48] Reaction of Fischer carbene 86 with alkyne 37 gave the benzannulated product, which on methylation and TBS removal afforded the alcohol 87. Similar to our previous work on cardinalin-3, treatment of 87 under normal oxa-Pictet–Spengler conditions gave the anti-1,3-dimethylpyran product 88 in a 72:28 diastereomer ratio. On the other hand, the reaction of 87 with (MeO)2CHMe under dry HCl gas bubbling gave the syn-1,3-dimethylpyran product 89 with a dr of 88:12. The separated diastereomer 89 on quinone formation under phenyliodine bis(trifluoroacetate) (PIFA) conditions and regioselective demethylation with BCl3 gave (+)-ventiloquinone L (7) in 73% yield from 89. Similarly, quinone formation and demethylation using 88 furnished (–)-1-epi-ventiloquinone L (7′) in identical yield.
Scheme 15 Synthesis of (–)-1-epi-ventiloquinone L (7) and (+)-ventiloquinone L (7′)
3.9
Synthesis of (–)-Juglomycin A (1), (+)-Deoxyfrenolicin (8), (+)-Kalafungin (9) and (+)-Frenolicin B (10)[50]
Our group again utilized the potential of the Dötz benzannulation reaction for the simple and efficient chiral-pool-based synthesis of (–)-juglomycin A (1),[21] (+)-kalafungin (9), (+)-frenolicin B (10) and (+)-deoxyfrenolicin (8) (Scheme [16]).[50] Kalafungin was first isolated by Bergy[51] from Streptomyces tanashiensis, while frenolicin B and deoxyfrenolicin were isolated by Omura and co-workers[52] from the fermentation of Streptomyces roseofulvus, strain No. AM-3867. The synthesis commenced with the Dötz benzannulation reaction of 28 with the alkyne 79 (prepared in six steps from d-glucono-δ-lactone) to give the naphthol, which on methylation produced 90. Acetonide deprotection and lactonization then afforded the lactone 33 in 85% yield. Oxidation of 33 with CAN and AlCl3-mediated demethylation gave (–)-juglomycin A (1). The lactone 33 on oxa-Pictet–Spengler reaction (to the syn-pyran diastereomer only) and CAN oxidation resulted in C-5 epi-methylkalafungin (91). Our earlier described protocol of demethylation with BBr3 also induced C-5 isomerization, which was increased further by H2SO4 treatment and a final recrystallization afforded (+)-kalafungin (9) in 56% overall yield from 91.
Scheme 16 Synthesis of (–)-juglomycin A (1), (+)-kalafungin (9), (+)-frenolicin B (10) and (+)-deoxyfrenolicin (8)
Similarly, the oxa-Pictet–Spengler reaction of 33 with butyraldehyde using the Lewis acid TMSOTf followed by CAN oxidation gave a mixture of diastereomers with the syn-isomer as the major product. This mixture on BBr3-based demethylation also underwent C-5 epimerization resulting in the anti-diastereomer 10 as the major product. These diastereomers were separated in 58% and 24% yields, respectively, resulting in (+)-frenolicin B (10) and 5-epi-frenolicin B (10′). The latter minor diastereomer was isomerized by H2SO4 treatment and recrystallized to give (+)-frenolicin B (10) in 60% yield. Hydrogenation of 10 resulted in lactone opening giving (+)-deoxyfrenolicin (8) in 80% yield.
3.10
Synthesis of (+)-Astropaquinone B (11a) and (+)-Astropaquinone C (11b)[53]
Astropaquinones B and C were isolated by Wang and co-workers from the cultures of the freshwater fungus Astrosphaeriella papuana YMF 1.01181, and were found to demonstrate moderate antagonistic activity against fungi and bacteria.[54] Since in our previous synthesis of ventiloquinone L (7) (Scheme [15]), the Dötz benzannulation using Fischer carbene 86 gave a lower yield of 35%, we considered the use of the different Fischer carbene 92 (Scheme [17]) for the synthesis of the astropaquinones. The carbene 92 on reaction with alkyne 37 gave the naphthol product in 48% yield. This on methylation and TBS removal furnished the alcohol 87. Subsequent oxa-Pictet–Spengler reaction of 87 with (MeO)3CH and PIFA-based oxidation gave the anti-pyran acetal (+)-astropaquinone B (11a) in 80% yield. During PIFA oxidation the dimerized compound 93 was isolated in 10% yield. Partial acetal hydrolysis of compound 11a gave (+)-astropaquinone C (11b) in 75% yield.
Scheme 17 Synthesis of (+)-astropaquinone B (11a) and (+)-astropaquinone C (11b)
3.11
Synthesis of (–)-Thysanone (13)[55]
(–)-Thysanone (13), a pyranonaphthoquinone antibiotic, was isolated from the solid-state fermentation of the fungus Thysanophora penicilloides (MF 5636, Merck Culture Collection)[56] and shows inhibition against human rhinoviruses (HRVs) 3C-protease (IC50 = 13 μg/mL), which are responsible for afflictions such as polio, hepatitis A and foot and mouth diseases.[57] We considered a concise synthesis of compound 13 as shown in Scheme [18].[55] The Dötz benzannulation of 92 with the alkyne 37 gave the corresponding naphthol, which on methylation provided compound 94. TBS deprotection and oxa-Pictet–Spengler reaction with CH(OMe)3 then gave the trans-acetal 95 in 84% yield over two steps. Subsequent PIFA-mediated quinone formation unfortunately gave the ortho-quinone 96 in 73% yield. Hence protection of the phenol as an iso-propyl ether and quinone formation resulted in a mixture of 97 and the hemiacetal 98 in 54% and 36% yields, respectively. Subsequent dealkylation of 97 or 98 with AlCl3 gave O-methylthysanone (99) without demethylation. Other dealkylating agents (BBr3, BCl3 or TiCl4) also gave similar results. In our ventiloquinone L synthesis, the C9-OMe group was demethylated quite selectively in the presence of the C7-OMe group using BCl3.[48] Hence an alternative strategy using different protecting groups was considered (Scheme [19]). The Fischer carbene 100, prepared in four steps, on Dötz benzannulation with the alkyne 37 gave the expected naphthol, which on subsequent methylation provided compound 101. This on TBS deprotection and oxa-Pictet–Spengler reaction with CH(OMe)3 gave the trans-acetal 102 in 75% yield over two steps. Benzyl protection of the phenol in 102 (86%) and quinone formation provided the quinone 103 in 83% yield. Debenzylation of the latter under hydrogenolysis conditions unfortunately also reduced the acetal giving the pyran 104 in 79% yield. The two-step conversion of 104 into thysanone is known in the literature[58] and thus our route constitutes a formal synthesis of thysanone (13).
Scheme 18 Fernandes’ synthesis of O-methylthysanone (99)
Scheme 19 Fernandes’ formal synthesis of (–)-thysanone (13)
3.12
Synthetic Studies on γ-Actinorhodin, Actinorhodin and Crisamicin A[4]
[59]
Actinorhodin and γ-actinorhodin are interesting dimeric pyranonaphthoquinones isolated from Streptomyces coelicolor
[60] (soil-dwelling bacteria) with promising bioactivity against Staphylococcus aureus bacteria[60b] located in the human respiratory tract and on the skin. Crisamicin A, isolated from Micromonospora purpureochromogenes,[61] shows activity against B16 murine melanoma cells and the herpes simplex and vesicular stomatitis viruses.[62] A promising bidirectional Dötz benzannulation approach was attempted toward the syntheses of these compounds (Scheme [20]).
Scheme 20 Synthetic studies on actinorhodin
Reaction of the dimeric Fischer carbene 105 with alkyne 29 gave the expected naphthol (70%) and subsequent methylation and TBS removal provided the dimeric alcohol 106 (86%). The latter on alcohol oxidation to the aldehyde followed by an allyl Grignard reaction gave 107. All attempts to construct the pyran ring by an oxa-Pictet–Spengler reaction failed to give the pyran 108. When the reaction was attempted on a monomeric compound, participation of the homoallylic double bond in a Prins-type reaction was observed. Alternatively, the Dötz benzannulation of 105 with alkyne 109 gave the corresponding naphthol (52%) and subsequent methylation and TBS deprotection provided the dimeric alcohol 110. The latter was subjected to an oxa-Pictet–Spengler reaction to give a syn/anti mixture of pyran 111. This on CAN oxidation gave a complex mixture. Other conditions using Ag2O, PIFA and CrO3 resulted in either decomposition or delivered regioisomeric and differently oxidized quinone mixtures arising from multiple 1,4-dimethoxy aryl units and/or possible quinone isomerizations.
Alternatively, the decarboxylative–deconjugative Knoevenagel reaction on the aldehyde derived from 106 gave the ester 113 (Scheme [21]). This on asymmetric dihydroxylation furnished the bis-lactone 114. The latter resisted the oxa-Pictet–Spengler reaction to install the pyran ring of 115. Similar reactions were quite successful on monomeric molecules in the arizonin synthesis.[46] Hence the failure was attributed to the electron density on the naphthyl rings having four methoxy groups on each ring.
Scheme 21 Synthetic studies on γ-actinorhodin
In another approach, we considered the synthesis of deoxy-γ-actinorhodin (Scheme [22]).[4] The Dötz benzannulation of 50 with alkyne 109 gave the expected naphthol, which on methylation provided 116. Removal of the TBS group followed by an oxa-Pictet–Spengler reaction gave a syn/anti-pyran mixture that was oxidized successfully to the quinone mixture. Separation of the latter furnished the anti/anti-pyran 117a and the anti/syn-pyran 117b in 62% and 18% yields, respectively. Demethylation of 117a to 118 (79%) and subsequent ester hydrolysis and air oxidation resulted in cyclization, possibly through the quinone methide intermediate, to furnish deoxy-γ-actinorhodin (isocrisamicin A) (119). The undesired isomer 117b on demethylation was subjected to concentrated H2SO4 treatment for benzylic epimerization giving 118 with 35% recovery.
Scheme 22 Synthetic studies on γ-actinorhodin, actinorhodin and crisamicin A
Considering the possible oxidative homocoupling of monomeric units to the dimeric molecules actinorhodin and γ-actinorhodin, we planned to synthesize the former monomers (Scheme [23]).[59] Dötz benzannulation of the Fischer carbene 121 (prepared from 120) with alkyne 79 gave the naphthol 122 in 52% yield. Conversion of the phenol group into a methyl ether and lactonization provided 123 (74% over two steps). A subsequent oxa-Pictet–Spengler reaction gave the syn-pyran 124 in 76% yield. Various oxidative coupling methods using DDQ,[63a] FeCl3,[63b]
[c] CAN,[63d] PIFA and Ag2O[63e] failed to dimerize the monomer 124. Under CAN or PIFA conditions, a mixture of quinone regioisomers 126a and 126b was obtained (49% and 38% yields, respectively, with PIFA). The separated quinone 126b on demethylation to 127 (68%) and isomerization mediated by concentrated H2SO4 gave the trans-epimer in a 93:7 ratio. A single recrystallization furnished hemi-γ-actinorhodin (17) in 62% yield from 127. Benzylic hydrogenation then opened the lactone to give hemiactinorhodin (16) in 83% yield. The undesired terminal quinone on BBr3-mediated demethylation also underwent quinone isomerization and C-5 epimerization to give a mixture of 17 and 127 in a 62:38 ratio and 45% yield. The obtained syn-isomer was epimerized fully by H2SO4 treatment and recrystallization to give hemi-γ-actinorhodin (17).
Scheme 23 Fernandes’ synthesis of hemi-γ-actinorhodin (17) and hemiactinorhodin (16)
3.13
Synthesis of the Core Structure of Medermycin (20)[64]
Medermycin was first isolated in 1976 by Takano et al.[65a] from Streptomyces sp. and has a unique β-C-glycoside linkage of an aminosugar, d-angolosamine. It shows prominent activity against Gram-positive bacteria and also inhibits human leukemic K-562 cells and platelet aggregation.[65b] Hong and co-workers[64] developed a Dötz benzannulation based strategy for the preparation of the C-arylglycoside moiety of medermycin (Scheme [24]). The glycoside-based alkyne 130 was prepared from 3,4,6-O-acetyl d-glucal (129) in eight synthetic steps. The reaction of the latter with the Fischer carbene 128 gave the corresponding phenol that was methylated to afford the C-arylglycoside 131 in 65% yield. Subsequent benzyl deprotection, O-carbamate formation and aryl bromination gave the intermediate 132. Displacement of Br with CN and carbamate hydrolysis provided the phenol 133. Next, phenol deoxygenation via the triflate followed by DIBAL-H reduction of the nitrile to an aldehyde and oxidation gave the acid 134, which represents the core structure of medermycin (20).
Scheme 24 Hong’s synthesis of the medermycin core structure
3.14
Total Synthesis of (±)-Naphthacemycin A9 (24)[66]
Naphthacemycin A9 (24) is a member of a series of new antibiotics isolated by Ōmura et al. from the culture broth of Streptomyces sp. KB-3346-5.[67] This compound demonstrated anti-MRSA activity and could be used to treat bacteria showing β-lactam resistance. Compound 24 contains a naphthacene structure with the E-ring having atropchirality. The total synthesis of (±)-naphthacemycin A9 (24) by Ōmura et al. is depicted in Scheme [25].[66] The Fischer carbene 136 was prepared in three steps from 135 and then reacted with the alkyne 137 under microwave irradiation to give the phenol, which on methylation provided 139 as the minor product. The major product was the cyclobutenone 138. Following the report of Moore and Perri,[68] compound 138 was thermally rearranged into the intermediate ketene that then cyclized to give the phenol, which on methylation furnished 139. The subsequent Suzuki–Miyaura coupling of the latter with bromide 140 provided compound 141. Removal of both Ts groups, methylation and acid-mediated Friedel–Crafts cyclization led to the spirocyclic dienone 142 as an inseparable 1:1 mixture of diastereomers. Next, CAN-mediated p-quinone formation, TiCl4-based dienone–phenol rearrangement, acetylation of the phenol and benzylic oxidation with CAN gave the alcohol 143 as an inseparable diastereomer mixture. Alcohol oxidation to the ketone and deacetylation gave the methyl-naphthacemycin that resisted demethylation by boron trihalides. However, demethylation was achieved by reaction with CeCl3·7H2O and NaI to furnish the target molecule (±)-naphthacemycin A9 (24) in 64% yield.
Scheme 25 Ōmura’s synthesis of (±)-naphthacemycin A9 (24)
3.15
Total syntheses of Anhydrolandomycinone (21), Landomycinone (22) and Tetrangulol (23)[69]
Landomycins, with potent antiproliferation and antibiotic activities,[70] consist of an angular tetracyclic unit fused with a deoxyoligosaccharide framework and were isolated from Streptomyces bacteria.[71] Mong et al.[69] have synthesized anhydrolandomycinone (21), tetrangulol (23) and landomycinone (22) using the Dötz benzannulation and C–H activation (Schemes 26 and 27). The reaction of Fischer carbene 144a with alkyne 145 afforded the naphthol 146 in 55% yield. This was then subjected to CAN oxidation, hydroquinone formation and CBz protection to give 147. Subsequent Pd-catalyzed C–H-activation-based intramolecular cyclization, CBz removal and quinone formation led to 148. The latter on aromatization, demethylation and MOM group removal gave anhydrolandomycinone (21). Similarly, toward the synthesis of tetrangulol, the Dötz benzannulation of Fischer carbene 144b (a regioisomer of 144a) with alkyne 145 gave the naphthol 149 that on CAN oxidation, quinone reduction and CBz protection afforded 150. The latter on C–H-activation-based cyclization gave 151. Subsequent CBz removal, quinone formation, MOM deprotection and conversion of the phenol into OTf gave 152. Finally, deoxygenation of the triflate, aromatization and demethylation furnished tetrangulol (23).
Scheme 26 Mong’s total synthesis of anhydrolandomycinone (21) and tetrangulol (23)
The synthesis of landomycinone (22) is shown in Scheme [27]. The benzannulation of carbenes 153 with alkyne 154 gave the naphthol 155, which on quinone formation, reduction and acetylation afforded the diacetate 156. Subsequent intramolecular cyclization, acetate hydrolysis and careful quinone formation (no aromatization of the B ring) led to compound 157. Finally, benzyl group removal under hydrogenolysis conditions and MOM deprotection gave landomycinone (22).
Scheme 27 Mong’s total synthesis of landomycinone (22)
3.16
Synthesis of (–)-Juglomycin C (18a) and (–)-NHAB (18b)[72]
Scheme 28 Fernandes’ syntheses of (–)-juglomycin C (18a) and
(–)-NHAB (18b)
We recently completed step-economic and efficient syntheses of (S)-(–)-juglomycin C (18a) and (S)-(–)-NHAB (18b) employing the Dötz benzannulation reaction (Scheme [28]).[72] The former was isolated from Streptomyces sp. 815 and 3094,[73a] while the latter is a shunt product from disruption of the act-VI-ORFA gene in Streptomyces coelicolor A3(2), which produces actinorhodin biosynthetically.[73`]
[c]
[d] In our synthesis, the chiral alkyne 159 was prepared from epoxide (±)-158 by first resolving the racemic epoxide to obtain (+)-158 in high enantiomeric excess (99% ee). Opening of the epoxide with TMS-acetylene, TMS removal and OTBS ether formation gave alkyne 159 in 64% overall yield from 158. The subsequent Dötz benzannulation of Fischer carbene 28 with alkyne 159 gave the naphthol 160. CAN-mediated quinone formation also caused TBS group removal giving 161 (88%) and demethylation with AlCl3 also effected t-Bu group removal providing (–)-juglomycin C (18a) in a step-economic and convergent deprotection sequence. The acetylation of 161 and demethylation with concomitant t-Bu group removal gave (S)-(–)-NHAB (18b).
