Journal of the American Chemical Society
Article
a
withTsNHNH2, which was then treated with a base and
Rh2(Oct)4 to give 47 in 85% yield along with 10% yield of a
1,2-H shift alkene byproduct. Using copper salts as the catalyst
in substitute for Rh2(Oct)4, only the alkene product was
observed.17 Amide reduction of 47 with DIBAL-H gave 48 in
good yield. We envisioned that opening the cyclopropane ring
with a hydroxyl nucleophile would not only establish the B ring
but also introduce a hydroxyl group at C7. As shown in
Scheme 6, a cascade of hydrolysis of the enol ether and SN2-
Scheme 5. Failed Attempts to Construct the B Ring
Scheme 6. Nucleophilic Ring Opening of the
a
Cyclopropane
a
Reagents and conditions: 47 or 48, acid (5 equiv), THF, 60 °C, 1 h.
49a (48% HBF4 in H2O, 80% yield, >20/1 dr); 49b (37% HCl in
H2O, 86% yield, >20/1 dr); 49c (57% HI in H2O, 83% yield, >20/1
dr); 50 (48% HBF4 in H2O, 61% yield, >20/1 dr).
a
Reagents and conditions: (a) I2, NaHCO3, THF/H2O, rt, 1 h, 97%;
(b) DIBAL-H, CH2Cl2, −78 °C, 10 min, 90%; (c) DIBAL-H,
CH2Cl2, −30 °C, 10 min, 94%; (d) NBS, Ph3P, rt, 1 h, 86%; (e) Na,
t
liquid NH3, BuOH, THF, −78 °C, 2 h, 85%; (f) PhOCSCl, DMAP,
CH2Cl2, rt, 2 h, 95%; (g) SmI2, HMPA, THF, rt, 2 h, 73%; (h) Na,
liquid NH3, BuOH, THF, −78 °C, 2 h, 78%.
like ring opening of cyclopropane moiety by treating 47 or 48
with an aqueous solution of HBF4 formed the B ring and
stereospecifically installed the problematic equatorial C7−OH.
Although attempts to introduce fluoro, benzenesulfonyl, or
phenylthio groups at C7 were unfruitful, treatment of 47 with
an aqueous solution of HCl or HI generated 49b or 49c with a
halogen atom introduced, which would facilitate preparation of
natural product analogues with unnatural functionalities at
C7.13 The stereochemistry of the newly formed C7−OH of
49a was determined by X-ray crystallographic analysis, and the
same configuration was assigned to other products by
analogy.16,17
Construction of the D Ring. With the B ring established,
the next task was to construct the D ring. As shown in Scheme
7, our initial plan to construct the B ring was an intramolecular
alkylation. Mesylate 52 was prepared by a sequence of
desilylation, mesylation, cyclopropane ring opening, and
hydrogenation. To our surprise, attempts to construct the
bicyclo[2.2.2]octane motif by treatment of 52 with a base (e.g.,
LDA, tBuOK, NaH, and K2CO3) only resulted in generation of
54 with a four-membered ring. The structure of 54 was
confirmed by X-ray analysis of a single crystal of its derivative
56, which was prepared by sequential olefination and allylic
oxidation.16,17 Generally, formation of a six-membered ring is
favored over a four-membered ring. A possible rationale for
this unusual selectivity is the specific and rigid architecture of
compound 52. We next resorted to a reversible aldol reaction
with the expectation of generating the thermodynamically
favored six-membered ring product. As shown in Scheme 4C,
sequential reactions of acetylation, one-step alkene hydro-
genation/Bn-hydrogenolysis, and Boc-protection delivered 57,
which was then elaborated to aldehyde 58 via desilylation and
oxidation. Exposure of 58 to the classic aldol reaction
t
to prepare aldehyde 36 from ester 25 were unfruitful, and only
the suspected quaternary ammonium byproducts were
observed. This can be attributed to tendency toward
quaternization of the basic tertiary amine moiety of 25.
Accordingly, amine 25 was oxidized to amide 38 by treatment
with I2 (Scheme 5B). DIBAL-H reduction of the ester afford
aldehyde 39 in high yield. However, treatment of 39 with SmI2
under various conditions resulted in complex reaction mixtures
(Scheme 5B). To explore the feasibility of constructing the B
ring via dearomative addition with a primary free radical
intermediate, as we did previously, primary bromide 42 was
prepared from 38 by a sequential reduction and bromination
and was tested with SmI2 or Bu3SnH. However, only
unidentified products were generated (Scheme 5C). After
several unsuccessful dearomative additions, we then turned to
free radical addition to the alkene group after an arene
dearomatization. Thus, precursor 44 was prepared by a
sequence of Birch reduction and esterification and was then
subjected to SmI2. Notably, only 45 was isolated in a high yield
with formation of a five-membered ring (Scheme 5D).
Interestingly, Birch reduction of ester 38 led to compound
46 with a similar five-membered ring (Scheme 5E). We
hypothesized that formation of a five-membered ring is favored
over the six-membered ring without the tethering C6−N bond,
which is present in our hetisine-type alkaloid synthesis.
After unsuccessful attempts to construct the B ring via free
radical addition, we were ultimately drawn to a dearomative
cyclopropanation strategy. As indicated in Scheme 4B, the
diazo precursor was generated by condensation of aldehyde 39
7092
J. Am. Chem. Soc. 2021, 143, 7088−7095