Stereoselective total synthesis of hainanolidol and harringtonolide via oxidopyrylium-based [5 + 2] cycloaddition

Article abstract of DOI:10.1021/ja406255j

The tetracyclic carbon skeleton of hainanolidol and harringtonolide was efficiently constructed by an intramolecular oxidopyrylium-based [5 + 2] cycloaddition. An anionic ring-opening strategy was developed for the cleavage of the ether bridge in 8-oxabicyclo[3.2.1]octenes derived from the [5 + 2] cycloaddition. Conversion of cycloheptadiene to tropone was realized by a sequential [4 + 2] cycloaddition, Kornblum-DeLaMare rearrangement, and double elimination. The biomimetic synthesis of harringtonolide from hainanolidol was also confirmed.

Full text of DOI:10.1021/ja406255j

Article
pubs.acs.org/JACS
Stereoselective Total Synthesis of Hainanolidol and Harringtonolide
via Oxidopyrylium-Based [5 + 2] Cycloaddition
Min Zhang,† Na Liu,† and Weiping Tang†,‡,*
†School of Pharmacy, University of Wisconsin, Madison, Wisconsin 53705, United States
‡Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706, United States
S
* Supporting Information
ABSTRACT: The tetracyclic carbon skeleton of hainanolidol
and harringtonolide was efficiently constructed by an intra-
molecular oxidopyrylium-based [5 + 2] cycloaddition. An
anionic ring-opening strategy was developed for the cleavage
of the ether bridge in 8-oxabicyclo[3.2.1]octenes derived from
the [5 + 2] cycloaddition. Conversion of cycloheptadiene to
tropone was realized by a sequential [4 + 2] cycloaddition,
Kornblum−DeLaMare rearrangement, and double elimination. The biomimetic synthesis of harringtonolide from hainanolidol
was also confirmed.
oxidation mediated by lead tetraacetate.4 This biomimetic
transformation was validated by identical IR and MS of the
1. INTRODUCTION
Harringtonolide, hainanolidol, and fortunolides A and B are
semisynthetic and natural 1. Although the structure of 2 was
representative members of Cephalotaxus norditerpenes (1−4,
assigned mainly based on this critical reaction,4 its yield has
Figure 1). They have complex architectures featuring a fused
never been reported.
Fortunolides 3 and 4 together with hainanolidol 2 were also
isolated from C. fortunei by Chiu’s group in 1999.5 The first
synthesis of hainanolidol 2, and thus a formal synthesis of
harringtonolide 1, was realized by Mander’s group in 1998,
featuring an elegant arene cyclopropanation followed by ring
expansion strategy for the construction of the tropone moiety.6
Their attempts toward the preparation of harringtonolide 1 by
forming the THF ring at earlier stages have not been fruitful.7
Recently, significantly more synthetic efforts were devoted to
harringtonolide 1,8 after the discovery of its remarkably potent
and selective anticancer activity by Nay’s group in 2008 (IC50
=
43 nM for KB tumor cells).9
2. RESULTS AND DISCUSSION
Figure 1. Representative Cephalotaxus Norditerpenes.
To evaluate the therapeutic potential of harringtonolide 1 and
related natural products as anticancer agents, we embarked on a
synthetic program toward their synthesis. We envisioned that
the tropone and lactone in natural products 1 and 2 could be
derived from intermediate 6 (Scheme 1). We proposed to
construct the tetracyclic carbon skeleton of 6 by an intra-
molecular [5 + 2] cycloaddition of intermediate 7, where the
oxidopyrylium ion could be prepared from oxidative ring
expansion of the furan ring in decalin 8.10 The six contiguous
stereogenic centers and their associated functional groups in
decalin 8 would be derived from known compound 9, available
in two steps diastereoselectively.11
tetracyclic carbon framework 5, a cyclohexane ring A bearing
five or six contiguous stereogenic centers, an unusual tropone
ring D, and a bridged lactone. The cage-like harringtonolide 1
and fortunolide B 4 have an additional tetrahydrofuran (THF)
ring.
Harringtonolide 1 was first isolated in 1978 from C.
harringtonia, and its structure was assigned unambiguously by
X-ray crystallography.1 In 1979, both 1 (named as hainanolide)
and 2 were isolated from C. hainanensis.2 Interestingly, the
former was found to possess antineoplastic and antiviral
activities, while the latter was inactive.2,3 This suggests that
the THF ring in 1 is important for its bioactivity. Hainanolidol
2 was proposed to be the precursor of harringtonolide 1 as the
former could be converted to the latter by transannular
Received: June 20, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja406255j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Retrosynthetic Analysis of 1 and 2
Scheme 2. Preparation of a Decalin Derivative with Six
Contiguous Stereogenic Centers for [5 + 2] Cycloaddition
a
Our synthesis began with oxidation of enone 9 (Scheme 2).12
A mixture of diastereomeric allylic alcohols in a 4:1 (α/β) ratio
was obtained and both of them were converted to ketone 15
eventually. The hydroxyl group in 11 could be installed highly
diastereoselectively by dihydroxylation of the corresponding
silyl enol ether. Initially, we tried to reduce the ketone in enone
11 to a trans-diol by delivering the hydride intramolecularly
using NaBH(OAc)3.13 The fact that the resulting diol could be
protected as an acetonide in compounds 12 and 13 suggested
that a cis-diol was formed.
Reduction of ketone 11 by NaBH4, on the other hand,
afforded a diol that could not be protected as an acetonide. This
diol was assigned as a trans-1,2-diol.14 Protection of the
resulting diol with TBSOTf followed by selective removal of
the TES group by TFA afforded allylic alcohol 14. The desired
β-stereochemistry in alcohol 14 could be realized by a sequence
of oxidation and diastereoselective reduction,14 through enone
15.
A highly stereoselective Claisen rearrangement of vinyl ether
16 yielded bicyclic compound 17 with five contiguous
stereogenic centers. An olefin isomerization was required for
the conversion of this intermediate to desired compound 8 in
addition to the appendage of a furan. We envisaged that a
stereoselective [3,3]-sigmatropic rearrangement of intermediate
20 might afford a cis-fused decalin derivative with an olefin in
the desired position.15 To this end, aldehyde 17 was converted
to enone 19 through allylic oxidation of amide 18.16 Treatment
of this enone with tosylhydrazine followed by diastereoselective
reduction15c of the corresponding hydrazone gave us
intermediate 20 (R = Weinreb amide), which underwent
stereoselective rearrangement to form cis-decalin 21.15a,b
Ketone 22 was then obtained after the addition of furanyl
Grignard reagent to amide 21. In the absence of MgBr2 salt, the
furanyl lithium reagent also reacted with the angular ester
group. Reduction of ketone 22 by NaBH4 then afforded a
mixture of diastereomeric alcohols 8. The synthesis of alcohol 8
from aldehyde 17 would be much more efficient if the olefin
isomerization could be performed on intermediate 20 where R
is a furanyl alcohol. The furan moiety, however, could not be
tolerated under the allylic oxidation conditions.
a
(a) TMSCl, NaI, Ac2O, 0 °C to rt, 85%; (b) oxone, NaHCO3, 70%;
(c) LDA, TESCl; then OsO4, NMO, THF/H2O (10:1), 72% over 2
steps, dr >20:1; (d) NaBH(OAc)3, AcOH, MeCN, 72%, dr >20:1; (e)
acetone, TsOH, 74%; (f) Dess−Martin periodinane, 76%; (g) NaBH4;
then TBSOTf, 2, 6-lutidine, 84% over two steps, dr >20:1; (h) TFA,
76%; (i) Dess−Martin periodinane, 84%; (j) NaBH4, 93%, dr >20:1;
(k) Hg(TFA)2, vinylbutyl ether, Et3N, 86%; (l) toluene, reflux, 85%, dr
>20:1; (m) NaClO2, NaH2PO4, 2-methyl-2-butene; then MeONHMe·
HCl, PyBOP, Et3N, 76% over two steps; (n) Mn(OAc)3, TBHP, 76%;
(o) TsNHNH2; then catecholborane and NaOAc, 75% over two steps,
dr >20:1; (p) furan, BuLi, MgBr2, 92%; (q) NaBH4; then VO(acac)2,
TBHP, DCM, 87% over two steps; (r) Ac2O, DMAP, pyridine, 91%;
(P = TBS).
Substrate 23 was then obtained as a mixture of four
diastereomers after VO(acac)2-catalyzed oxidative ring ex-
pansion17 of furanyl alcohol 8 followed by esterification. The
stereochemistry on the dihydropyran ring of 23 is incon-
sequential as an achiral oxidopyrylium moiety is formed in the
subsequent [5 + 2] cycloaddition. After screening different
solvents and bases, we found that the intramolecular
Hendrickson [5 + 2] cycloaddition18 occurred smoothly in
refluxing chloroform in the presence of DBU (Scheme 3).10
Only one stereoisomer was observed for the cycloaddition
product 24. Its structure and relative stereochemistry are
B
dx.doi.org/10.1021/ja406255j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 3. [5 + 2] Cycloaddition and Attempts for the
Synthesis of Tropone by Dehydration
Scheme 4. Opening of Ether Bridge
a
(a) PhSH, BF3·OEt2; (b) LDA, HMPA, THF; (c) MeMgBr; (P =
TBS).
The introduction of the phenylthio group not only solved the
reproducibility issue of the 1,3-transposition but also provided a
handle for the opening of the oxygen bridge. To test the
generality of this strategy, we prepared allylic thio ether 31
from known compound 3021 using the same sequence of
addition of methyl Grignard reagent and SN1′ displacement by
thiophenol. The anionic opening worked smoothly to yield
bicyclic product 32. In related anionic ring-opening reactions
reported by Lautens and his co-workers,22 oxabicyclo[3.2.1]-
octene was opened by a SN2′ process22a,b or a base-mediated β-
elimination.22c
confirmed by X-ray analysis as shown in the Supporting
Information. It is worth to mention that the synthesis of
product 24 via the key [5 + 2] cycloaddition could be carried
out on gram scale.
Diene 32 was then used as a model system for the synthesis
of tropone. We envisioned that a sequence of hetero-Diels−
Alder cycloaddition of a cycloheptadiene with a nitrosoarene,23
reductive cleavage of the N−O bond to an amino alcohol, and
double elimination of amine and water might provide the
tropone moiety. After protecting the free alcohol in 32, the
resulting diene was treated with 2-nitrosopyridine (Scheme 5).
Addition of methyl Grignard reagent, selective removal of
one of the two silyl protecting groups under acidic conditions,
and lactonization in the presence of potassium carbonate
afforded hexacyclic compound 25. Although the steric environ-
ments of the two silyl ethers in 24 are quite different, strong
acids or fluoride salts removed both of them nonselectively.
High chemoselectivity was achieved by using trichloroacetic
acid. The tertiary allylic alcohol in 25 could undergo a PCC-
mediated tandem 1,3-transposition/oxidation sequence to
afford enone 26. The yield of this reaction, however, varied
from 10% to 40%.
a
Scheme 5. Synthesis of Model Tropone 34
A dehydration process involving the removal of two Hs and
one O highlighted in red in 26 would furnish product 27, which
is just one step away from hainanolidol 2. We were not able to
prepare the tropone moiety directly, however, under various
dehydration conditions, regardless the thermodynamic prefer-
ence for the formation of a nonbenzenoid aromatic tropone
ring. We then turned our attention to the reductive cleavage of
the ether bridge in 26 using SmI2. This reductive process would
produce an enone, which may be converted to tropone by
elimination of water and dehydrogenation. However, treatment
of enone 26 with SmI2 or other reducing agents led to either no
reaction or a complex mixture.19
After many trials, we developed a two-step protocol to open
the ether bridge (Scheme 4). A phenylthio group was first
introduced to 28 through a Lewis acid-mediated SN1′
substitution of 25.20 Only one diastereomer was observed for
thio ether 28. The thiophenol likely approached intermediate
25 from the less sterically hindered β-face based on the X-ray
structure of compound 24. The α-proton of the phenyl sulfide
was then removed by LDA in the presence of HMPA, and the
oxygen bridge was cleaved under this condition. Two potential
isomeric dienes can be generated in this anionic ring-opening
reaction by cleaving either C(β′)−O or C(δ)−O bonds. HSQC
spectra indicated that diene 29 was formed by the cleavage of
the C(δ)−O bond.
a
(a) TBSOTf, 2,6-lutidine, DCM; (b) 2-nitrosopyridine, DCM, 90%
over two steps; (c) SnCl2, EtOAc, 50%.
The hetero-Diels−Alder cycloaddition occurred and afforded
adduct 33, which was directly converted to tropone 34 by the
treatment of SnCl2. Elimination of both 2-aminopyridine and
water indeed occurred after the reductive cleavage of the N−O
bond. The structure of intermediate 33 was assigned based on
the regioselectivity observed in a similar hetero-Diels−Alder
reaction involving nitroso compounds.24 With this method in
hand, we then treated diene 29 or its derivatives with 2-
nitrosopyridine in the absence or presence of different Lewis
acids.25 Unfortunately, no desired cycloaddition product was
observed.
Several conditions have been reported for the hydrolysis of
vinyl phenylthio ethers to ketones.26 Under these conditions,
however, the thio ethers in 29, 32, or their derivatives could not
be hydrolyzed to the corresponding enone. We then decided to
remove the phenylthio group in 29 and examine conditions for
the conversion of the resulting diene 35 (Scheme 6) to a
C
dx.doi.org/10.1021/ja406255j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
2] cycloaddition to construct the tetracyclic carbon skeleton, an
anionic ring opening of the ether bridge derived from [5 + 2]
cycloaddition, and the formation of a tropone through a
sequence of [4 + 2] cycloaddition, Kornblum−DeLaMare
rearrangement, and double elimination. This new synthetic
route for hainanolidol and harringtonolide offers the flexibility
to access other members of Cephalotaxus norditerpenes and
various simplified analogues. Evaluation of the anticancer
activity of harringtonolide and its analogues will be reported
in due course.
Scheme 6. Formation of Tropone and Completion of the
Synthesis of 1 and 2
a
ASSOCIATED CONTENT
■
S
* Supporting Information
Detailed experimental procedures, characterization data, and
1
spectra (IR, H, 13C NMR, and HRMS). This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
a
Corresponding Author
wtang@pharmacy.wisc.edu
(a) TESCl, DMAP, TEA, DCM; (b) MMPP on silica gel, DCM; (c)
SmI2, DMPU, MeOH, THF, 70% over three steps; (d) O2, TPP, light,
CH3CN, 40%; (e) DBU, DCM; (f) TsOH, CDCl3, 85% over two
steps; (g) Pb(OAc)4, benzene, 90 °C, 52%. (P = TBS).
Notes
The authors declare no competing financial interest.
tropone. We could not obtain diene 35 by treating 29 with
Raney-Ni directly. Although reductive cleavage of a carbon−
sulfone bond by SmI2 has been reported,27 we were not able to
prepare the corresponding sulfone from 29 in a good yield.
Interestingly, the corresponding sulfoxide can be synthesized
efficiently, and the reductive cleavage of a carbon−sulfoxide
bond worked well. After silyl protection, the phenylthio group
in 29 was removed by a two-step sequence: oxidation by
magnesium monoperoxyphthalate (MMPP) to its sulfoxide28
and reduction with SmI2.
Inspired by the strategy of [4 + 2] cycloaddition, N−O bond
cleavage, and double elimination to access tropone 34 in
Scheme 5, we envisioned a sequence of [4 + 2] cycloaddition of
diene 35 with singlet oxygen to afford peroxide 36, Kornblum−
DeLaMare rearrangement of 36 to ketone 37,29 and double
elimination to prepare tropones in natural products 1 and 2.
Indeed, peroxide 36 was formed by the cycloaddition between
diene 35 and singlet oxygen together with a byproduct derived
from an ene reaction, using tetraphenylporphyrin (TPP) as the
photosensitizer. DBU-promoted Kornblum−DeLaMare rear-
rangement provided ketone 37. In the presence of acid,
removal of silyl groups and elimination of two water molecules
occurred to yield hainanolidol 2. The 1H and 13C NMR spectra
of our synthetic hainanolidol are in accordance with natural
product as shown in the Supporting Information. 2D NMR
data (COSY, HMBC, HSQC, and NOE) further confirmed the
structure and stereochemistry of hainanolidol.
ACKNOWLEDGMENTS
■
We are grateful for financial support from NIH
(R01GM088285) and the University of Wisconsin. We thank
Prof. Mander (Australian National University, Australia), Prof.
Chiu (Kunming Institute of Botany, China), and Prof. Nay
(CNRS, France) for generously sharing their NMR spectra with
us.
REFERENCES
■
(1) Buta, J. G.; Flippen, J. L.; Lusby, W. R. J. Org. Chem. 1978, 43,
1002.
(2) Sun, N.-J.; Xue, Z.; Liang, X.-T.; Huang, L. Acta pharm. Sin. 1979,
14, 39 The X-ray structure of 1 was also reported in this reference..
(3) Kang, S. Q.; Cai, S. Y.; Teng, L. Acta Pharmacol. Sin. 1981, 16,
867.
(4) Xue, Z.; Sun, N.-J.; Liang, X.-T. Acta Pharmacol. Sin. 1982, 17,
236.
(5) Du, J.; Chiu, M. H.; Nie, R. L. J. Nat. Prod. 1999, 62, 1664.
(6) (a) Frey, B.; Wells, A. P.; Rogers, D. H.; Mander, L. N. J. Am.
Chem. Soc. 1998, 120, 1914. (b) Rogers, D. H.; Frey, B.; Roden, F. S.;
Russkamp, F. W.; Willis, A. C.; Mander, L. N. Aust. J. Chem. 1999, 52,
1093. (c) Frey, B.; Wells, A. P.; Roden, F.; Au, T. D.; Hockless, D. C.;
Willis, A. C.; Mander, L. N. Aust. J. Chem. 2000, 53, 819.
(7) (a) Zhang, H. B.; Appels, D. C.; Hockless, D. C. R.; Mander, L.
N. Tetrahedron Lett. 1998, 39, 6577. (b) Mander, L. N.; O’Sullivan, T.
P. Synlett 2003, 1367. (c) O’Sullivan, T. P.; Zhang, H.; Mander, L. N.
Org. Biomol. Chem. 2007, 5, 2627.
(8) (a) Evanno, L.; Deville, A.; Dubost, L.; Chiaroni, A.; Bodo, B.;
Nay, B. Tetrahedron Lett. 2007, 48, 2893. (b) Abdelkafi, H.; Evanno,
L.; Herson, P.; Nay, B. Tetrahedron Lett. 2011, 52, 3447. (c) Hegde,
V.; Campitelli, M.; Quinn, R. J.; Camp, D. Org. Biomol. Chem. 2011, 9,
4570. (d) Li, W. Asian J. Chem. 2012, 24, 1411. (e) Abdelkafi, H.;
Herson, P.; Nay, B. Org. Lett. 2012, 14, 1270. (f) Abdelkafi, H.; Nay, B.
Nat. Prod. Rep. 2012, 29, 845.
(9) Evanno, L.; Jossang, A.; Nguyen-Pouplin, J.; Delaroche, D.;
Herson, P.; Seuleiman, M.; Bodo, B.; Nay, B. Planta Med. 2008, 74,
870.
(10) For recent reviews on [5 + 2] cycloadditions and their
applications, see: (a) Singh, V.; Krishna, U. M.; Vikrant; Trivedi, G. K.
Tetrahedron 2008, 64, 3405. (b) Pellissier, H. Adv. Synth. Catal. 2011,
353, 189. (c) Ylijoki, K. E. O.; Stryker, J. M. Chem. Rev. 2013, 113,
Treatment of hainanolidol 2 with lead tetraacetate in
refluxing benzene finally provided harringtonolide 1 in a 52%
4
1
isolated yield. The H and 13C NMR spectra of our synthetic
harringtonolide are in agreement with those reported for
natural harringtonolide.1,2,9 The key biomimetic transformation
of biologically inactive hainanolidol 2 to bioactive harringto-
nolide 1 is thus confirmed for the first time by total synthesis.
3. CONCLUSION
In summary, the total synthesis of natural products hainanolidol
and harringtonolide was realized featuring two stereoselective
[3,3]-sigmatropic rearrangements, an oxidopyrylium-based [5 +
D
dx.doi.org/10.1021/ja406255j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
2244. (d) Nguyen, T. V.; Hartmann, J. M.; Enders, D. Synthesis 2013,
45, 845.
(11) (a) Marshall, J. A.; Faubl, H.; Warne, T. M. Chem. Commun.
1967, 753. (b) Scanio, C. J. V.; Starrett, R. M. J. Am. Chem. Soc. 1971,
93, 1539. (c) Takatori, K.; Nakayama, M.; Futaishi, N.; Yamada, S.;
Hirayama, S.; Kajiwara, M. Chem. Pharm. Bull. 2003, 51, 455.
(12) Wijnberg, J.; Jongedijk, G.; Degroot, A. J. Org. Chem. 1985, 50,
2650.
(13) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93,
1307.
(14) (a) Cherest, M.; Felkin, H. Tetrahedron Lett. 1971, 383.
(b) Huet, J.; Maronibarnaud, Y.; Anh, N. T.; Seydenpenne, J.
Tetrahedron Lett. 1976, 159.
(15) (a) Hutchins, R. O.; Milewski, C. A.; Maryanoff, B. E. J. Am.
Chem. Soc. 1973, 95, 3662. (b) Hutchins, R. O.; Kacher, M.; Rua, L. J.
Org. Chem. 1975, 40, 923. (c) Kabalka, G. W.; Yang, D. T. C.; Baker, J.
D. J. Org. Chem. 1976, 41, 574.
(16) Shing, T. K. M.; Yeung, Y.-Y.; Su, P. L. Org. Lett. 2006, 8, 3149.
(17) (a) Achmatowcz, O.; Bukowski, P.; Szechner, B.;
Zwierzchowska, Z.; Zamojski, A. Tetrahedron 1971, 27, 1973.
(b) Ho, T. L.; Sapp, S. G. Synth. Commun. 1983, 13, 207. (c) Wender,
P. A.; Rice, K. D.; Schnute, M. E. J. Am. Chem. Soc. 1997, 119, 7897.
(18) (a) Hendrickson, J. B.; Farina, J. S. J. Org. Chem. 1980, 45, 3359.
(b) Sammes, P. G.; Street, L. J. J. Chem. Soc., Chem. Commun. 1982,
1056.
(19) Williams, D. R.; Benbow, J. W.; McNutt, J. G.; Allen, E. E. J. Org.
Chem. 1995, 60, 833.
(20) Radix-Large, S.; Kucharski, S.; Langlois, B. R. Synthesis 2004,
456.
(21) Burns, N. Z.; Witten, M. R.; Jacobsen, E. N. J. Am. Chem. Soc.
2011, 133, 14578.
(22) (a) Lautens, M.; Abdelaziz, A. S.; Lough, A. J. Org. Chem. 1990,
55, 5305. (b) Lautens, M. Synlett 1993, 177. (c) Lautens, M.; Ma, S. H.
Tetrahedron Lett. 1996, 37, 1727.
(23) For a recent review on this topic, see: Bodnar, B. S.; Miller, M. J.
Angew. Chem., Int. Ed. 2011, 50, 5629.
(24) Kirby, G. W.; Sweeny, J. G. J. Chem. Soc., Chem. Commun. 1973,
704.
(25) Yamamoto, Y.; Yamamoto, H. J. Am. Chem. Soc. 2004, 126,
4128.
(26) (a) Corey, E. J.; Shulman, J. I. J. Org. Chem. 1970, 35, 777.
(b) Mukaiyam., T.; Kamio, K.; Kobayash., S.; Takei, H. Bull. Chem.
Soc. Jpn. 1972, 45, 3723. (c) Grieco, P. A.; Dai, Y. J. Tetrahedron Lett.
1998, 39, 6997.
(27) Keck, G. E.; Savin, K. A.; Weglarz, M. A. J. Org. Chem. 1995, 60,
3194.
(28) Ali, M. H.; Stevens, W. C. Synthesis 1997, 764.
(29) Kornblum, N.; DeLaMare, H. J. Am. Chem. Soc. 1951, 73, 880.
E
dx.doi.org/10.1021/ja406255j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX