Organic Letters
Letter
Table 1. Comparison of 1H and 13C NMR Spectroscopic Data of the Side Chain of Synthetic Material with Reported Data for
Fortisterol and Herbarulide
1H shifts
herbarulide
1H shifts
synthetic 2
1H shifts
synthetic 1
1H shifts
fortisterol
13C shifts
herbarulide
13C shifts
synthetic 2
13C shifts
synthetic 1
13C shifts
c
a
b
b
c
a
b
b
no.
fortisterol
20 2.03 m
21 1.02 d 6.6
2.04 m
1.02 d 6.6
2.04 m
1.02 d 6.6
2.04 m
1.01 s
40.2
21.1
40.2
21.0
40.3
21.0
40.2
21.0
22 5.15 dd 15.2, 8.5 5.15 dd 15.2, 8.6 5.14 dd 15.2, 8.6 5.25 dd 15.2,
134.7
134.7
134.9
134.7
7.7*
23 5.25 dd 15.2, 7.8 5.25 dd 15.2, 7.9 5.23 dd 15.2, 8.1 5.15 dd 15.2,
132.9
132.8
132.9
132.8
8.4*
24 1.86 m
25 1.48 m
26 0.84 d 6.8
27 0.82 d 6.8
28 0.91 d 6.8
1.86 m
1.48 m
0.84 d 6.8
0.82 d 6.8
0.91 d 6.8
1.84 m
1.47 m
0.84 d 6.8
0.82 d 6.7
0.92 d 6.8
1.84 m
2.44 m*
0.82 d 6.8
0.82 d 6.8
0.87 d 6.8
42.9
33.1
20.0
19.7
42.8
33.0
19.9
19.6
17.6
43.1
33.1
20.1
19.6
42.8
33.0*
19.9
19.6
17.5
17.6
18.0
a
b
1
1
Data collected in CDCl3 at 600 MHz for H and 151 MHz for 13C; see ref 4. Data collected in CDCl3 at 600 MHz for H, referenced to the
c
residual solvent peak at δH 7.26, and 151 MHz for 13C, referenced to the solvent peak at δC 77.00. Data collected in CDCl3 at 400 MHz for 1H and
100 MHz for 13C; see ref 5a. Misassigned signals are marked with *.
accessed from triol 1018 by the chemoselective oxidation of the
6-hydroxyl moiety (MnO2), followed by the selective
acetylation of the equatorial 3-hydroxy group. Recrystallization
of the crude material furnished 8 with high purity and in an
overall yield of 62% on a 5 g scale.
us.12c As a result, only (Z)-configurated material was obtained,
but partial isomerization was achieved through a dibromina-
tion/radical debromination sequence to give the (E)-
diastereomer 14 together with its (Z)-diastereomer 13 (d.r.
1.2:1) in excellent yield. Radical rearrangement was performed
on the (E)-/(Z)-mixture and gave, after HPLC separation, 1
and (22Z)-1. The NMR spectral data of 1 were strikingly
similar to our synthetic 2. (See Table 1 and the Supporting
Information.) The 13C NMR data did not match the reported
data for fortisterol, however, with the indicative signals C22,
C24, C26, and C28 being 0.2 to 0.5 ppm more downfield than
the reported values. Additionally, comparing diastereomer
(22Z)-1 with the reported data showed even greater deviation,
especially for C20 and C24. A comparison of the optical
activity values to the reported ones also did not provide any
In a first try, Burawoy’s ketone (8) was treated with HgO
and I2 to indeed provide the 5,6-epoxy-5,6-secosteroid 11 in a
yield of 60%. To gain access to the desired structure 2, we
found intermediary triol 10 even better suited for directly
reaching the alkoxy radical precursor 7 through oxidation with
Dess−Martin’s periodinane. When applying the conditions for
radical rearrangement to dione 7, 5,6-epoxy-5,6-secosteroid 2
could be directly obtained in 48% yield (four steps, 45% overall
yield from ergosterol).
Whereas the 13C NMR shifts of synthetic 2 were in close
1
agreement with the reported data, the H NMR data revealed
25
hint: Whereas for herbarulide, [α]D = +55.0 (c 0.185 in
three small deviations (Table 1): 24-H, 26-H, and 28-H were
detected 0.02 to 0.04 ppm more downfield compared with the
reported values. Comparing our synthetic material with the
data for herbarulide instead revealed a much closer agreement
(for both 1H and 13C shifts). This result led us to reassign the
structure of herbarulide to that of compound 2 in Scheme 1B
but left us with uncertainty regarding the true structure of
fortisterol. In the absence of any copies of the spectra in the
original publication, we can only suspect that there were
several mistakes in the original assignment of fortisterol, that is,
confusing 2-H with 25-H, as well as 2-C with 25-C,
furthermore 11-H with 15-H, as well as 11-C with 15-C, and
eventually 22-H with 23-H. We tentatively assumed fortisterol
to rather be campestane 1 which was originally assigned as
herbarulide. To provide unequivocal structural proof, we
decided to engage in the synthesis of structure 1. Thus, dione 7
was ozonolytically cleaved to an aldehyde (structure not
shown), which was reacted with sulfone 12 in a Julia−
Kocienski olefination under conditions previously employed by
CH2Cl2) was reported, and for fortisterol, [α]D20 = +1.8 (c 0.2
in CHCl3) was reported, our synthetic material showed
absolutely no agreement with either of the two, independently
of the concentration and the solvent employed (see the
22
in CH2Cl2); synthetic 2: [α]D = +177.2 (c 0.22 in CHCl3).
As a result, the true structure of fortisterol remains elusive,
although the original report raises questions regarding the
correct assignment of the originally obtained NMR data.19
In conclusion, we have achieved a potentially biomimetic
synthesis of the alleged structures of 5,6-epoxy-5,6-secosteroids
herbarulide and fortisterol through an alkoxy radical-mediated
process. A careful comparison of the obtained NMR spectral
data with the reported data led us to correct the originally
proposed structure of herbarulide, whereas the true structure
of fortisterol remains ambiguous. Toward this goal, we
revisited the preparation of Burawoy’s ketone (8) and
developed it into a scalable and sustainable process.
C
Org. Lett. XXXX, XXX, XXX−XXX