Biomimetic synthesis of 5,6-dihydro-glaucogenin C: Construction of the disecopregnane skeleton by iron(II)-promoted fragmentation of an α-alkoxy hydroperoxide

Article abstract of DOI:10.1002/anie.201101893

A skeleton key: A biomimetic synthesis of the title natural product was completed in 19 steps and 6.4% overall yield. Iron(II)-promoted fragmentation of α-alkoxy hydroperoxide and subsequent trapping of the resulting tertiary carbon radical by iodide enab

Full text of DOI:10.1002/anie.201101893

DOI: 10.1002/anie.201101893
Natural Product Synthesis
Biomimetic Synthesis of 5,6-dihydro-glaucogenin C: Construction of
the Disecopregnane Skeleton by Iron(II)-Promoted Fragmentation of
an a-Alkoxy Hydroperoxide
Jinghan Gui, Dahai Wang, and Weisheng Tian*
Natural products that have rearranged steroid skeletons and
promising biological activity have attracted significant atten-
tion in the field of synthetic chemistry in recent years, with the
most representative examples being cortistatins[1] and naki-
terpiosin.[2] Another representative example of such natural
products is the glaucogenin family (1–4), which is a collection
and biomimetic approach to 13,14:14,15-disecopregnane
steroids.[6]
The synthetic challenge posed by the glaucogenin family
lies, in great part, in the stereoselective construction of the
À
highly functionalized D/E rings and scission of the C13 C14
single bond to form the nine-membered lactone ring. As
shown in Scheme 1, we envisaged that the target molecule 5
and other glaucogenin members could be accessed from
unsaturated ester 6 through reduction and transposition of the
of unique 13,14:14,15-disecopregnane steroids. In the 1980s,
Mitsuhashi and co-workers isolated these unusual C21
steroids from the hydrolysates of crude glycosides in three
Chinese herbal medicines “Bai-Qian (Cynanchum glauces-
cens)”, “Bai-Wei (Cynanchum atratum)”, and “Xu-Chang-
Qing (Cynanchum paniculatum)”, which all belong to the
Asclepiadaceae plant family.[3] The structure and configura-
tion of this class of compounds were established unequiv-
ocally by X-ray crystallographic analysis of the diacetate of
glaucogenin C mono-d-thevetoside.[3a,b] Recent biological
studies on glaucogenin C and its four glycosides revealed
that they were effective and selective inhibitors of alpha-
virus-like positive-strand RNA viruses but not of other RNA
or DNA viruses, yet they showed no toxicity to host cells.[4]
The IC50 values of glaucogenin C (17 nm) and its four
glycosidal derivatives (each 25 nm) suggest that it is the
aglycon unit and not the linked oligosaccharide chains that is
in fact responsible for the inhibitory effect. Despite the
unique chemical structure and powerful antiviral activity of
the glaucogenin family, no chemical synthesis of these steroids
has been reported to date. Herein, we report the first
synthesis of 5,6-dihydro-glaucogenin C (5) from (16S,20S)-
5a-pregnane-3b,16,20-triol (12), which is a readily available
C21 steroid obtained from tigogenin[5] that provides a general
[*] J. Gui, Dr. D. Wang, Prof. Dr. W. Tian
Key Laboratory of Synthetic Chemistry of Natural Substances
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
345 Lingling Road, Shanghai 200032 (China)
Fax : (+86)21-5492-5176
E-mail: wstian@mail.sioc.ac.cn
Scheme 1. Retrosynthetic analysis of 5,6-dihydro-glaucogenin C and
other members of the glaucogenin family. TBDPS=tert-butyldiphenyl-
silyl.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201101893.
Angew. Chem. Int. Ed. 2011, 50, 7093 –7096
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7093

Communications
À
C7 C8 olefin, respectively. A biogenetic pathway proposed
ing a leaving group at C14, followed by an SN2’ reaction
involving H2O.[11] On the basis of this consideration and
previous work carried out within our research group,[12] we
were pleased to find that exposure of 16 to radical bromina-
tion condition (NBS, AIBN) afforded a major product, which
was converted into the allylic alcohol 11 with the required
a configuration at C16 during chromatography on silica gel
(53% yield). The stereoselectivity of this reaction is believed
to originate from attack at the Si face of water on the initially
formed allylic bromination product 17.[13] Unfortunately,
when this reaction was performed on a larger scale, the
yield was variable (24–48%). Efforts to optimize this reaction
eventually revealed that the use of 1,3-dibromo-5,5-dimethy-
limidazolidine-2,4-dione (DBDMH)[14] as the brominating
reagent and subsequent treatment with wet silica gel could
deliver 11 in reproducible 72% yield on a 14 g scale.
With the hydroxy group at the C16 stereogenic center now
introduced, the preparation of 10 was then investigated. The
two-step procedure of reduction of lactone 11 using LiAlH4 in
conjunction with subsequent oxidative olefin cleavage
resulted in the formation of 19 (93%, 17 g scale), the
structure of which was confirmed by NMR characterization
of its acetylation product 22. To access 10, an efficient and
chemoselective oxidation of the ether at C20 and reduction of
the acetal at C15 were required. To this end, 19 was subjected
to a second remote intramolecular free radical functionaliza-
tion[15] to afford the cage-like C20-oxidized product 20 (82%,
6 g scale) with an interesting consecutive ketal/acetal/ketal
linkage. The next task, that is, selective reduction of the acetal
by Mitsuhashi et al. suggested that the key nine-membered
lactone ring might be generated from hirundigenin through
hydroxylation at C18 and subsequent Grob-type fragmenta-
tion (Scheme 1, top equation).[3a,b] Inspired by this biogenesis,
we surmised that such bond scission could also be accom-
plished through b-fragmentation of the alkoxy radical 8
without preoxidation at C18. In this regard, the iron/copper-
promoted fragmentation reaction of alkoxy hydroperoxide[7]
appeared to be an attractive approach. The alkoxy hydro-
peroxide 9 in turn might be prepared from the key inter-
mediate 10 by the Schenck ene reaction.[8] Intermediate 10
could then be traced back to 12 through stereoselective
synthesis of allylic alcohol 11.
The first objective of our synthesis was to develop a
scalable route to 10 on the way to 5 and other glaucogenin
members. As depicted in Scheme 2, the synthesis commenced
with bromination/acetylation of 12 (69%, 200 g scale),[9] and
subsequent ester hydrolysis and selective protection of the
hydroxy group at C3 (quant., 50 g scale) to give 14 in large
quantities. Proximal functionalization of the unactivated
angular methyl group at C18 in 14 by using the method
developed by Meystre and co-workers[10a] provided the
desired lactone 15 (84%, 30 g scale),[10b–d] whose elimination
using DBU in toluene at reflux then led to the desired alkene
16 (quant., 50 g scale) with absolute regiocontrol.
It was postulated that the transformation involving the
stereoselective installation of the hydroxy group at C16 with a
concomitant double bond shift could be effected by introduc-
Scheme 2. Gram-scale synthesis of key intermediate 10. Reagent and conditions: a) 33% HBr/HOAc, 408C, 2 h, 69%; b) LiOH·H2O, THF/MeOH
(1:4), 808C, 1 h; c) TBDPSCl, imidazole, CH2Cl2, RT, 1 h, quant. (2 steps); d) Pd(OAc)4, I2, CaCO3, cyclohexane, hn, 2 h, then Jones oxidation,
84%; e) DBU, toluene, reflux, 1 day, 97%; f) DBDMH, AIBN, CCl4, cyclohexene oxide, reflux, 1 h, then wet silica gel, CH2Cl2, RT, 10 h, 72%;
g) LAH, THF, 508C, 1 h; h) O3, CH2Cl2, À788C, 40 min, then Zn-HOAc, RT, 1 day, 93% (2 steps); i) DIB, I2, cyclohexane, hn, 1 h, 82%; j) PPTS,
Ac2O, 1308C, 2 h, 72%; k) Et3N, MeOH, 508C, 9 h; l) MsCl, Et3N, CH2Cl2, 08C, 3 h; m) LAH, THF, RT, 1 h, 70% (3 steps); n) TBAF, THF, 608C,
quant. AIBN=azobisisobutyronitrile, DBDMH=1,3-dibromo-5,5-dimethylhydantoin, DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, DIB=(diacetoxy-
iodo)benzene, LAH=lithium aluminum hydride, Ms=methanesulfonyl, PPTS=pyridinium 4-toluenesulfonate, TBAF=tetra-n-butylammonium
fluoride, THF=tetrahydrofuran.
7094 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7093 –7096

at C15 in the presence of ketals at C20 and C14 proved to be
exceedingly challenging, as various direct reduction condi-
tions resulted in either no reaction or overreduction of the
ketal at C14. Ultimately, however, this problem was solved by
first converting 20 into the enol ether 21 (72%, gram scale)
having an inverted configuration at C15,[16] and then reducing
the acetal at C15 by a one-pot, three-step procedure involving
ester hydrolysis, mesylation, and reduction using LiAlH4 to
furnish 10 in 70% overall yield (gram scale). Removal of the
TBDPS protecting group of 10 (TBAF, THF) then afforded a
known compound, dihydro-anhydrohirundigenin (10a),
radical fragmentation were highly efficient. Meanwhile, Cu-
(OAc)2 showed little reactivity toward olefin production in
our system. To test this hypothesis, the initially formed
tertiary carbon radical 7 was trapped with TEMPO, which
indeed afforded the desired product 24 in 70% yield.
To continue the synthesis, TEMPO was substituted with I2
for ease of elimination during installation of the olefin at
À
C13 C18. In the event, treatment of 9 with FeSO4 and I2
followed immediately by highly regioselective HI elimination
afforded 6 in a striking 69% yield over two steps (Scheme 4).
1
whose H NMR spectrum was in good agreement with that
reported.[17]
As anticipated, the Schenck ene reaction[8] of 10 afforded
the alkoxy hydroperoxide 9 in quantitative yield and in a
highly regio- and stereoselective manner, and thus set the
stage for the key fragmentation reaction (Scheme 3). Treat-
Scheme 4. Completion of the synthesis of 5,6-dihydro-glaucogenin C.
Reagent and conditions: a) FeSO4, I2, MeOH, RT, 1 h; b) DBU, toluene,
808C, 1 h, 69% (2 steps); c) NiCl2·6H2O, NaBH4, THF/MeOH (2:3),
08C to RT, 73%; d) LAH, THF, 458C, 1.5 h, 66%; e) TBAF, THF, 608C,
13 h, 92%; f) NaH, THF, 808C, 2 h, 90%.
Then, hydrogenation of the electron-deficient alkene in 6
(73%),[18] and subsequent removal of the protecting group
(92%) and epimerization (90%) completed the synthesis of 5,
whose 13C NMR data of the D/E rings was in accordance with
the data reported for glaucogenin C.[3f] The configuration of
the hydrogenated product 25 was deduced from the NOE
interactions of its reduced derivative 26.
In conclusion, we have achieved the first biomimetic
synthesis of 5,6-dihydro-glaucogenin C in 19 steps from
(16S,20S)-5a-pregnane-3b,16,20-triol (12) in an overall yield
of 6.4%. The key features of our synthesis were: 1) radical
bromination (SN2’ reaction) to set the stereogenic center at
C16, 2) oxidation of Me and ether groups at C18 and C21,
respectively, by two remote intramolecular free radical
functionalizations, 3) high scalability of the synthetic route
to key intermediate 10; and most importantly, 4) Schenck ene
reaction and subsequent iron(II)-promoted regioselective
fragmentation reaction of a-alkoxy hydroperoxide to con-
struct the key nine-membered lactone ring of 6. Our modified
Schreiber procedure, that is, replacing Cu(OAc)2 with a free
radical trapping reagent (I2 or TEMPO), may be applicable to
the synthesis of other macrolides. Furthermore, the presence
Scheme 3. Construction of the key nine-membered lactone by iron(II)-
promoted regioselective fragmentation reaction of a-alkoxy hydroper-
oxide. Reagent and conditions: a) TPP, O2, CH2Cl2, hn, 08C, 1 h,
quant.; b) FeSO4, Cu(OAc)2, MeOH, RT, 1 h; c) FeSO4, MeOH, RT, 1 h;
d) FeSO4, TEMPO, MeOH, RT, 1 h. TEMPO=2,2,6,6-tetramethylpiper-
idin-1-yloxyl, TPP=5,10,15,20-tetraphenyl-21H,23H-porphine.
ment of 9 with known condition (FeSO4, Cu(OAc)2)[7a–d] gave
the desired disecopregane product 6 in only 26% yield,
together with 32% of the 13,18-dihydro derivative 23.
According to Schreiberꢀs proposed mechanism,[7d] the olefin
À
at C13 C18 is formed via b-hydrogen elimination of the alkyl
copper intermediate generated by the oxidative coupling of
carbon radical 7 with Cu(OAc)2 (Scheme 1). To elucidate the
role of Cu(OAc)2 in the reaction, a control experiment was
conducted in the absence of Cu(OAc)2, and surprisingly, 6 and
23 were also isolated in 27% and 30% yield, respectively. The
nearly 1:1 ratio of 6 and 23 suggested that they might be
produced through disproportionation of 7. On the basis of the
above results, we hypothesized that the ferrous-mediated
homolytic cleavage of the peroxide and subsequent alkoxy
À
of the olefin at C7 C8 in our synthesis may provide us the
À
opportunity to introduce the olefin at C5 C6 for the synthesis
of the natural product, glaucogenin C, as well as other
members of the glaucogenin family. This endeavor is the
subject of our current research and will be reported in due
course.
Angew. Chem. Int. Ed. 2011, 50, 7093 –7096
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7095

Communications
Received: March 17, 2011
Published online: June 17, 2011
[7] a) S. L. Schreiber, T. Sammakia, B. Hulin, G. Schulte, J. Am.
Chem. Soc. 1986, 108, 2106; b) S. L. Schreiber, B. Hulin, W. F.
Liew, Tetrahedron 1986, 42, 2945; c) S. L. Schreiber, W. Liew, J.
Am. Chem. Soc. 1985, 107, 2980; d) S. L. Schreiber, J. Am. Chem.
Soc. 1980, 102, 6163; for early reports on alkyl hydroperoxides,
see: e) Z. Cekoviꢁ, L. Dimitrijevic, G. Djokic, T. Srnic,
Tetrahedron 1979, 35, 2021; f) Z. Cekoviꢁ, M. M. Green, J.
Am. Chem. Soc. 1974, 96, 3000.
Keywords: alkoxy hydroperoxide · biomimetic synthesis ·
.
fragmentation · iron · natural products
[1] For reviews, see: a) D. Y.-K. Chen, C.-C. Tseng, Org. Biomol.
Chem. 2010, 8, 2900; b) A. R. Hardin Narayan, E. M. Simmons,
R. Sarpong, Eur. J. Org. Chem. 2010, 3553; c) Y. Shi, W. Tian,
Chin. J. Org. Chem. 2010, 30, 515, and references therein.
[2] a) T. Ito, M. Ito, H. Arimoto, H. Takamura, D. Uemura,
Tetrahedron Lett. 2007, 48, 5465; b) S. Gao, Q. Wang, C. Chen,
J. Am. Chem. Soc. 2009, 131, 1410; c) H. Takamura, Y.
Yamagami, T. Ito, M. Ito, H. Arimoto, I. Kadota, D. Uemura,
Heterocycles 2009, 77, 351; d) S. Gao, Q. Wang, L. J.-S. Huang, L.
Lum, C. Chen, J. Am. Chem. Soc. 2010, 132, 371.
[3] a) T. Nakagawa, K. Hayashi, H. Mitsuhashi, Tetrahedron Lett.
1982, 23, 757; b) T. Nakagawa, K. Hayashi, H. Mitsuhashi, Chem.
Pharm. Bull. 1983, 31, 870; c) T. Nakagawa, K. Hayashi, H.
Mitsuhashi, Chem. Pharm. Bull. 1983, 31, 2244; d) T. Nakagawa,
K. Hayashi, H. Mitsuhashi, Chem. Pharm. Bull. 1983, 31, 879;
e) T. Nakagawa, K. Hayashi, K. Wada, H. Mitsuhashi, Tetrahe-
dron 1983, 39, 606; f) Z. Zhang, J. Zhou, K. Hayashi, H.
Mitsuhashi, Chem. Pharm. Bull. 1985, 33, 1507; g) Z. X. Zhang,
J. Zhou, K. Hayashi, H. Mitsuhashi, Chem. Pharm. Bull. 1985,
33, 4188; h) K. Sugama, K. Hayashi, H. Mitsuhashi, K. Kaneko,
Chem. Pharm. Bull. 1986, 34, 4500; i) J. Dou, Z.-M. Bi, Y.-Q.
Zhang, P. Li, Chin. J. Nat. Med. 2006, 4, 192.
[8] a) H. H. Wasserman, R. W. Murray, Singlet Oxygen, Academic
Press, New York, 1979; b) A. A. Frimer, Chem. Rev. 1979, 79,
359; c) M. Prein, W. Adam, Angew. Chem. 1996, 108, 519;
Angew. Chem. Int. Ed. Engl. 1996, 35, 477; d) E. L. Clennan,
Tetrahedron 2000, 56, 9151; e) M. Stratakis, M. Orfanopoulos,
Tetrahedron 2000, 56, 1595.
[9] J. Lin, J. Wang, L. Chen, Y. Wang, X. Tang, Q. Xu, Z. Zhu, W.
Tian, Acta Chim. Sin. 2006, 64, 1265.
[10] a) K. Heusler, P. Wieland, C. Meystre, Org. Synth. 1965, 45, 57;
b) S. Bhandaru, P. L. Fuchs, Tetrahedron Lett. 1995, 36, 8347;
c) Y. Shi, L. Jia, Q. Xiao, Q. Lan, X. Tang, D. Wang, M. Li, Y. Ji,
T. Zhou, W. Tian, Chem. Asian J. 2011, 6, 786; d) G. Majetich, K.
Wheless, Tetrahedron 1995, 51, 7095.
[11] For such SN2’ reactions with H2O as the leaving group, see: a) W.
Li, P. L. Fuchs, Org. Lett. 2003, 5, 2849; b) J. S. Lee, P. L. Fuchs, J.
Am. Chem. Soc. 2005, 127, 13122.
[12] L. Jia, PhD thesis, Shanghai Institute of Organic Chemistry
(China), 2002.
[13] When this SN2’ reaction proceeded intramolecularly, the oppo-
site stereoselectivity would result, see Ref. [11b].
[14] L. Vanmaele, P. J. De Clercq, M. Vandewalle, Tetrahedron 1985,
41, 141.
[4] Y. Li, L. Wang, S. Li, X. Chen, Y. Shen, Z. Zhang, H. He, W. Xu,
Y. Shu, G. Liang, R. Fang, X. Hao, Proc. Natl. Acad. Sci. USA.
2007, 104, 8083.
[5] a) W. Tian, CN 1146457A, 1997; b) W. Tian, M. Li, H. Yin, X.
Tang, CN 1299821A, 2001; c) W. Tian, S. Liu, J. Shen,
CN 1341603A, 2002; d) W. Tian, S. Liu, B. Qiu, X. Wu,
CN 1475494A, 2004.
[6] For a biomimetic approach to C-nor-d-homo-steroids (such as
nakiterpiosin), see: P. Heretsch, S. Rabe, A. Giannis, J. Am.
Chem. Soc. 2010, 132, 9968.
[15] P. de Armas, J. I. Concepcion, C. G. Francisco, R. Hernandez,
J. A. Salazar, E. Suarez, J. Chem. Soc. Perkin Trans. 1 1989, 405.
[16] Modified conditions to those reported in the literature: K.
Ravindar, M. S. Reddy, L. Lindqvist, J. Pelletier, P. Deslong-
champs, Org. Lett. 2010, 12, 4420.
[17] a) K. Stoeckel, W. Stoecklin, T. Reichstein, Helv. Chim. Acta
1969, 52, 1403; b) K. Stoeckel, W. Stoecklin, T. Reichstein, Helv.
Chim. Acta 1969, 52, 1429; c) R. Imhof, F. Marti, B. P. Schaffner,
H. Wehrli, Helv. Chim. Acta 1973, 56, 1920.
[18] M. Seki, T. Shimizu, K. Matsumoto, J. Org. Chem. 2000, 65, 1298.
7096 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7093 –7096