Key words
aphanorphine - total synthesis - alkaloids -
tert-butanesulfinimine - cycloaddition - hydrogen-atom transfer
Figure 1 Representative benzomorphan alkaloids
In 1988, an alkaloid named aphanorphine (1) was isolated by Shimizu and Clardy and their co-workers during their studies on the biosynthesis of the neurotoxic alkaloid neosaxitoxin in the freshwater blue-green alga Aphanizomenon flos-aquae.[1] Aphanorphine has a tricyclic benzazepine core and is structurally similar to the natural and synthetic analgesic benzomorphan alkaloids morphine (2), pentazocine (3), and eptazocine (4) (Figure [1]). Its intriguing structure and its potential analgesic biological activity made aphanorphine an attractive target for organic synthesis. Many elegant strategies have been developed to construct the tricyclic benzazepine motif, such as Lewis acid-promoted Friedel–Crafts or tin hydride-mediated radical cyclization of the 2-benzylpyrrolidine intermediate to construct the ring B,[2] transannular enolate or radical cyclization of 3-benzazepine derivatives to form both rings B and C,[3] or intramolecular nucleophilic cyclization of tetralin or dihydronaphthalene substrates to build ring C.[4] Grainger developed a unique approach including a carbamoyl-radical cyclization to close ring C and a late-stage formation of aromatic ring A through an inverse-electron-demand Diels–Alder reaction.[5] Here, we report a concise total synthesis of (+)-aphanorphine (5) based on transition metal-catalyzed cyclization reactions.
The metal-catalyzed hydrogen-atom transfer (MHAT) reaction has emerged as a powerful tool in organic synthesis.[6]
[7] As shown in Scheme [1], we envisioned that the ring B and C1 quaternary carbon center of (+)-aphanorphine (5) might be obtained by a radical cyclization initiated by MHAT of the 2-benzylpyrrolidine 6, which, in turn, could be assembled by intermolecular trimethylenemethane (TMM) [3+2]-cycloaddition[8] of the known chiral imine 7 with 2-[(trimethylsilyl)methyl]allyl acetate (8) (Scheme [1]).
Scheme 1 Retrosynthetic analysis of (+)-aphanorphine (5)
Our total synthesis of (+)-aphanorphine (5) commenced with the TMM [3+2]-cycloaddition of 2-[(trimethylsilyl)methyl]allyl acetate (8) with the chiral imine 7 (Scheme [2]),[9] which can be prepared from (4-methoxyphenyl)acetaldehyde (9) and (R)-(+)-tert-butylsulfinamide (10) in 66% yield by a known procedure.[10] Stockman and co-workers previously investigated the TMM [3+2]-cycloadditions of chiral aryl and alkyl tert-butanesulfinimines to yield enantiopure pyrrolidine products.[11] Unfortunately, when we followed Stockman’s method, none of the desired cycloaddition product was detected when 7 and 8 were stirred with Pd(PPh3)4 in THF for 18 hours at 25 °C. Instead, the unexpected alkylation product 11 was isolated in 42% yield (Scheme [2a]). We surmised that 11 might be formed by proton transfer from the C5 atom of 7 to the Pd–TMM intermediate 12. The C5 position of 7 is activated by both an electron-withdrawing inductive effect of the imine group and by the conjugate effect of the phenyl group; consequently, instead of the expected cycloaddition of the TMM intermediate 12 with the imine, proton transfer from the C5 atom of 7 to the Pd-TMM intermediate 12 becomes the favored pathway to give methallyl complex 13, which is attacked by the resulting anion 14 to deliver the alkylation product 11.[12]
Scheme 2 Investigation of the [3+2]-cycloaddition
Reports by Trost and co-workers[12a]
[c] suggested that increasing the temperature might enhance the nucleophilicity of TMM. Pleasingly, when the reaction mixture was stirred under reflux for 19 hours, our desired cycloaddition products 15a and 15b were obtained in 1:3 dr with a combined yield of 52%, along with the mono- and dialkylation products 11 and 16, respectively, in yields of 9 and 19%. For the synthesis of (+)-aphanorphine (5), the tert-butylsulfinyl group of 15b was removed by treatment with 2 M HCl in MeOH, and the resulting secondary amine was treated with ClCO2Me in the presence of NEt3 to give the methyl carbamate 17 in 80% yield over the two steps (Scheme [3]).
Scheme 3 Synthesis of benzylpyrrolidine 17
According to our synthetic plan, the next work was to construct the tricyclic benzazepine core of (+)-aphanorphine (5) through MHAT-based radical cycloaddition. We began our study by evaluating a catalytic system previously used by Shigehisa et al. for the hydroarylation of nonactivated alkenes (Table [1]).[13] Treatment of 17 with 1,1,3,3-tetramethyldisiloxane (TMDSO), N-fluoro-2,4,6-trimethylpyridinium triflate (O1, Figure [2]), and the ethylenediamine-containing salen Co-catalyst C1 in PhCF3 gave the desired tricyclic benzazepine 18 in only 6% yield (Table [1], entry 1). To our delight, the use of the 1,3-diaminopropane-containing catalyst C2 (Figure [2]) improved the yield to 72% (entry 2). The longer 1,4-butanediamine gave a much lower yield (entry 3). Replacing the tert-butyl group on the 5-position of the aromatic ring of C2 with H, Me, or OMe (C4–C6) led to no conversion (entries 4–6). Further catalyst screening showed that C7 was the best catalyst, affording a 76% yield of the desired product (entries 7 and 8). Next, a series of oxidants including N-fluoro-2,4,6-trimethylpyridinium tetrafluoroborate (O2), N-fluoropyridinium triflate (O3), N-fluoropyridinium tetrafluoroborate (O4), and (diacetoxyiodo)benzene (O5) were evaluated, but all proved inferior to N-fluoro-2,4,6-trimethylpyridinium triflate (O1) (entries 9–12). Finally, we examined various silanes and we found that poly(methylhydrosiloxane) (PMHS) was superior to TMDSO, PhSiH3, or Ph(i-PrO)SiH2,[14] giving an improved yield of 83%[15] (entries 13–15).
Table 1 Optimization of the MHAT-Based Radical Cycloaddition
|
|
Entry
|
Catalysta
|
Silane
|
Oxidanta
|
Yieldb (%)
|
|
1
|
C1
|
TMDSO
|
O1
|
6
|
|
2
|
C2
|
TMDSO
|
O1
|
72
|
|
3
|
C3
|
TMDSO
|
O1
|
12
|
|
4
|
C4
|
TMDSO
|
O1
|
NDc
|
|
5
|
C5
|
TMDSO
|
O1
|
ND
|
|
6
|
C6
|
TMDSO
|
O1
|
ND
|
|
7
|
C7
|
TMDSO
|
O1
|
76
|
|
8
|
C8
|
TMDSO
|
O1
|
13
|
|
9
|
C7
|
TMDSO
|
O2
|
58
|
|
10
|
C7
|
TMDSO
|
O3
|
ND
|
|
11
|
C7
|
TMDSO
|
O4
|
36
|
|
12
|
C7
|
TMDSO
|
O5
|
38
|
|
13
|
C7
|
PhSiH3
|
O1
|
32
|
|
14
|
C7
|
PMHS
|
O1
|
83
|
|
15
|
C7
|
PhSiH2(O-i-Pr)
|
O1
|
27
|
a For catalyst and oxidant structures, see Figure [2].
b Isolated yield.
c ND = not detected.
Figure 2 Catalyst structures C1–C8 and oxidants O1–O5
With 18 in hand, the remaining transformations of the synthesis were N-methylation and O-demethylation. Reduction of 18 with excess LiAlH4 afforded (–)-8-O-methylaphanorphine (19) in 88% yield. On following the procedure of Fuchs and Funk,[3a] treatment of 19 with BBr3 in DCM at a low temperature effected the expected O-demethylation, giving (+)-aphanorphine (5) in 50% yield (Scheme [4]). The physical and spectroscopic data of the synthetic (+)-aphanorphine (5) {[α]D
25 +20.8 (c 0.4, MeOH)} agreed with those reported previously.[1]
[2l]
Scheme 4 Completion of the total synthesis of (+)-aphanorphine (5)
In summary, a concise total synthesis of (+)-aphanorphine (5) was achieved, starting from the known chiral tert-butanesulfinimine 7, in six steps and 11% overall yield. The transition-metal-catalyzed intermolecular TMM [3+2]-cycloaddition and a MHAT-based radical cyclization were used in a rapid construction of the tricyclic benzazepine core of the natural product. In addition, methyl carbamate was used as a latent methylamine, avoiding additional steps involving manipulation of N-substituent group, as required in the previous synthesis, thereby improving the overall synthetic efficiency.
