Asymmetric synthesis of (?)- and (+)-neodichroine/hydrachine A from (+)- and (?)-febrifugine

Article abstract of DOI:10.1016/j.tetlet.2018.03.035

A new asymmetric approach to both enantiomers of the quinazolinone-containing natural product febrifugine is reported. Utilising a proline-mediated aminooxylation-Horner-Wadsworth-Emmons sequence provides the key optically active 2,3-disubstituted piperidine intermediate 7 in high enantioselectivity but poor overall yield (7 steps, 3%, 98% ee). This intermediate has been used to prepare both naturally occurring (+)-febrifugine (1) and its (?)-enantiomer. In turn, each were then used to synthesise, for the first time, both enantiomers of the claimed natural product neodichroine/hydrachine A. Spectroscopic data for the synthetic compound matched the claimed structure. However, the specific rotation differed in both magnitude and sign from the isolation work.

Full text of DOI:10.1016/j.tetlet.2018.03.035

Accepted Manuscript
Asymmetric synthesis of (–)- and (+)-neodichroine/hydrachine A from (+)- and
(–)-febrifugine
Shaun Smullen, Paul Evans
PII:
S0040-4039(18)30332-0
https://doi.org/10.1016/j.tetlet.2018.03.035
TETL 49804
DOI:
Reference:
Tetrahedron Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
1 February 2018
6 March 2018
14 March 2018
Please cite this article as: Smullen, S., Evans, P., Asymmetric synthesis of (–)- and (+)-neodichroine/hydrachine A
from (+)- and (–)-febrifugine, Tetrahedron Letters (2018), doi: https://doi.org/10.1016/j.tetlet.2018.03.035
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Asymmetric synthesis of (–)- and (+)-neodichroine/hydrachine A from (+)- and (–)-
febrifugine
Shaun Smullen and Paul Evans*
Centre for Synthesis and Chemical Biology, School of Chemistry, University College Dublin, Dublin,
Ireland. paul.evans@ucd.ie; ORCID: 0000-0002-0584-876X; Tel: 00353-(0)1-716-2291; Fax: 00353-
(0)1-716-2501
Abstract: A new asymmetric approach to both enantiomers of the quinazolinone-containing natural
product febrifugine is reported. Utilising a proline-mediated aminooxylation-Horner-Wadsworth-
Emmons sequence provides the key optically active 2,3-disubstituted piperidine intermediate 7 in
high enantioselectivity but poor overall yield (7 steps, 3%, 98% ee). This intermediate has been used
to prepare both naturally occurring (+)-febrifugine (1) and its (–)-enantiomer. In turn, each were
then used to synthesise, for the first time, both enantiomers of the claimed natural product
neodichroine/hydrachine A. Spectroscopic data for the synthetic compound matched the claimed
structure. However, the specific rotation differed in both magnitude and sign from the isolation
work.
Keywords: Natural product; organocatalysis; allylic alcohol; quinazolinone; quinolizidine;
stereoselective synthesis; intramolecular Mannich reaction
The quinazolinone natural product febrifugine (1), also known as -dichroine, has a long and
interesting history. It is found in the Asian medicinal plant Dichroa febrifuga and also more generally
in plants belonging to the hydrangea family. In its natural source it has been used as an agent to
reduce the fever (febrifuge) resulting from malarial infection1 and extractions of Dichroa febrifuga
roots and leaves have been shown to contain 1 - which evidence indicates is the plant’s principal
active ingredient.1,2 The determination of the structure of 1 has proven complex, in part because it
interconverts with an isomer termed isofebrifugine/-dichroine (2), which, unsurprisingly, can be
found in the same plants (Scheme 1).3 The precise structures of naturally occurring 1 and 2 were
eventually uncovered, primarily from the synthetic efforts of Baker,4 Barringer,5 Kobayashi6 and co-
workers.
CHCl3,
OH
N
O
O
N
OH

O
N
N
2
2
N
H
N
H
MeOH
(or H2O),

O
(+)-1
(+)-2
Scheme 1. Structures of the quinazolinone natural products (+)-febrifugine (1), (+)-isofebrifugine (2)
and their interconversion.
Although febrifugine possesses potent antimalarial action this activity has never translated into a
medicine due to a range of unacceptable side-effects. These problems led to the development of
various analogues in the hope of improving the activity to toxicity balance. The most widely studied
of which is halofuginone, a compound that in addition to possessing antiparasitic/coccicidial activity

has demonstrated human relevant activities and is currently the subject of a Duchenne muscular
dystrophy trial.7 An example of an additional structurally distinct analogue is bicycle 3 (Fig. 1), which
was prepared in 74% yield by an intramolecular Mannich reaction upon passing (+)-1 through a silica
gel column eluting with a mixture of acetone-chloroform-methanol. This compound proved to be an
only slightly less-potent antimalarial agent than (+)-1 (towards Plasmodium falciparum FCR-3) and
was of comparable activity to artemisinin.8
OH
H
OH
N
H
H
R
N
N
O
N
O
R
O
MeO
MeO
N
N
O
OH
N
5
6
3: R = Me
4: R = H
Figure 1. Structure of febrifugine analogue (3), neodichroine (4), initially proposed hydrachine A (5)
and (–)-lasubine II (6).
In 2000 Ye and co-workers reported that, in addition to 1 and 2, small amounts of a new
quinazolinone-containing material structurally related to 3 was present in an extraction of Dichroa
febrifuga leaves and termed this compound neodichroine (4).9 Since the absolute configuration of 1
had only just been corrected in 20006 the likely absolute configuration for their dextrorotatory
compound is as shown in Figure 1. Approximately a year later, Wu and co-workers reported an
isomeric dextrorotatory compound from the root parts of Hydrangea chinensis, which they termed
hydrachine A (5).10 The reported structure is related to the well-known lasubine family of
quinolizidine alkaloids [eg. (–)-lasubine II (6)].11 However, in 2003 the same group revised the
structure of hydrachine A, adjusting it to match with the previously reported Dichroa febrifuga
neodichroine structure (4) and now reporting that this compound possesses a laevorotatory optical
rotation value.12 In addition to the conflicting optical behaviour, the NMR spectroscopic data from
each isolation paper were recorded in different deuterated solvents and, therefore, were not
directly comparable.
Based on our interest in the asymmetric synthesis of febrifugine13 and due to ambiguity concerning
the reported data (optical rotation values etc.) for 4, we felt that its enantioselective synthesis would
enable us to unequivocally confirm the structure of the isolated material. As shown in Scheme 2, it
was envisaged that 4 would be accessed from (+)-1 via an intramolecular Mannich reaction, a
transformation which, in addition to the synthesis of the acetone derived 3,8 has some precedent for
the construction of the required ketone-containing quinolizidine core.14 In turn, we,13 and others15
have demonstrated that (+)-1 can be accessed from protected keto alcohol 7 which can be traced
back to optically active -hydroxy enone 8, via a diastereoselective intramolecular N-conjugate
addition. In our previous syntheses we have employed an asymmetric dihydroxylation of a vinyl
sulfone and also an enzymatic resolution of a racemic allylic alcohol followed by cross-metathesis.13
In the current work we envisaged that an optically active allylic alcohol of type 8 could be prepared

from an aldehyde using an -aminooxylation-Horner-Wadsworth-Emmons (HWE) olefination
sequence.16
OH
H
-amino-
oxylation-
HWE
OH
N
Mannich
OH
O
(+)-1
O
NH
O
N
Cbz
7
NBoc
Cbz
O
N
Cbz
4
O
9
8
N
Scheme 2. Intramolecular Mannich-based retrosynthetic analysis of 4.
Thus, doubly protected aldehyde 9 (accessed in five-steps from 5-aminopentanol, see ESI) was
treated under standard aminooxylation reactions conditions (Scheme 3).16 After consumption of
nitrosobenzene the reaction mixture was transferred into a second vessel containing the ethyl ester-
based phosphonate. Moderate yields of the resultant ethyl ester 10 were reproducibly encountered
and analysis by chiral HPLC indicated that high levels of stereochemical induction had taken place. It
was reasoned that the moderate yield encountered was associated with issues concerning the
efficiency of the HWE reaction with the oligomeric -hydroxy/aniloxy aldehyde intermediate.
Modest improvements in yield can be achieved using an excess of 9. However, in terms of the
overall sequence this is not a workable solution since the unfunctionalised -unsaturated ester,
stemming from the direct reaction of 9 with the phosphonate, required an arduous separation.
Notably, unlike some literature reports,16a we never detected the presumed intermediate aniloxy
ester adduct - and a formal reduction of the N-O bond clearly occurs during this process without the
addition of a separate reductant (such as a copper source).16
(1) D-Pro (20 mol%),
OH
PhNO (1 equiv.),
DMSO, rt, 15-35 min
9
BocN
Cbz
(2) (EtO)2POCH2COR,
DBU, LiCl, MeCN, 20 h
R
O
10: R = OEt, 35%, 95% ee
11: R = Me, 15%, 98% ee
(1) CAN, MeCN, 50 min
(2) BF3·OEt2 (0.5 equiv.),
MeCN, rt, 15 min
OH
3 steps13
O
(+)-1
N
Cbz
( )-7: R = Me: 55%
Scheme 3. Aminooxylation-Horner-Wadsworth-Emmons olefination of aldehyde 9 and the
synthesis of (–)-7.

On changing the identity of the phosphonate this sequence transfers to the methyl ketone 11.
However, whilst the enantioselectivity for the thus formed 11 remains excellent, unfortunately
changing the structure of the phosphonate in the HWE process has a negative impact on the
chemical yield. Nevertheless, the Boc group in 11 can be selectively removed with ceric ammonium
nitrate17 and the intermediate mono-carbamate (compound 8, Scheme 2) was primed to undergo a
diastereoselective13 intramolecular conjugate addition forming (–)-7. In three steps this protected
2,3-disubstituted piperidine has been used to prepare (+)-1. Use of L-proline, under an otherwise
identical sequence, gave (+)-7 with directly comparable yields and enantioexcess (see ESI).
With both enantiomers of febrifugine available their conversion into the corresponding
enantiomeric quinolizidines was considered (Scheme 4). We elected to use Eschenmoser’s salt as a
commercially available, solid form of formaldehyde.18 Thus, when (+)-113 was taken up in MeOH,
treated with Eschenmoser’s salt and heated to reflux overnight, two less polar compounds were
formed which proved isolable by flash column chromatography. Proton and carbon NMR
spectroscopy of the more polar, major component from this mixture, in both D5-pyridine9 and
10,11
CDCl3
matched the available literature data for neodichroine and hydrachine A respectively.
Additionally, NOE experiments concurred with the claimed stereochemistry for the new stereogenic
centre and are consistent with a preferred trans-decalin-like chair-chair structure (see ESI).
However, this synthetic compound, synthesised from naturally occurring (+)-febrifugine (1),
possesses a negative specific rotation in both chloroform and methanol (see Table 1). In order to
further confirm this finding, unnatural (–)-febrifugine (1) was similarly converted into (+)-4 and its
value is roughly equal and opposite to that obtained previously for (–)-4.
OH
H
OH
H
NMe2
I
N
N
O
O
(+)-1
O
N
O
N
MeOH,
reflux,
15 h
N
N
( )-4: 40%
OH
(+)-12: 8%
OH
H
H
NMe2
I
N
N
O
O
( )-1
O
N
O
N
MeOH,
reflux,
15 h
N
N
(+)-4: 36%
( )-12: 11%
Scheme 4. Use of (+)- and (–)-1 to prepare (–)- and (+)-4 and diastereomer 12.
Table 1. Comparison of specific rotation values for synthetic and isolated 4.

[]D values for isolated 4
[]D values for synthetic 4
[]D = +198.8 (c = 0.17, MeOH)9
[]D = +25.3 (c = 0.2, CHCl3)10
[]D = -25.3 (c = 0.2, MeOH)12
[]D20 = -34.2 (c = 0.5, MeOH)a
[]D20 = -44.4 (c = 0.5, CHCl3)a
[]D20 = +49.6 (c = 0.5, CHCl3)b
aFrom (+)-1; bfrom (–)-1
During these reactions a minor, slightly less-polar isomer, was also formed. This proved to be
compound 12 which presumably arises from the competitive isomerisation of febrifugine into
isofebrifugine (see Scheme 1) which then, following an intramolecular Mannich reaction in the keto
form, generates the diastereomer shown in Scheme 4. NOE experiments were consistent with the
overall epimerisation of the stereocentre adjacent to the quinolizidine nitrogen atom and also the
newly installed -keto centre (see ESI).
In summary, although hampered by a low yield for the key-stereoinducing step, we report a new
organocatalysis-based method (eleven steps) to prepare febrifugine. In addition, the synthesis of
both enantiomers of neodichroine/hydrachine A are outlined. On this basis we are able to state
unequivocally that neodichroine and hydrachine A are the same compound and that it is most likely
that the naturally occurring material possesses a negative specific rotation.
Supporting Information is available free of charge via the internet:
Acknowledgements: We thank UCD for support including the provision of a postgraduate
scholarship (SS).
References
1. (a) Jang, C. S.; Fu, F. Y.; Wang, C. Y.; Huang, K. C.; Lu, G.; Chou, T. C. Science 1946, 103, 59; (b) Koepfli, J. B.;
Mead, J. F.; Brockman, Jr., J. A. J. Am. Chem. Soc. 1947, 69, 1837; (c) Jang, C. S.; Fu, F. Y.; Huang, K. C.; Wang, C.
Y. Nature 1948, 161, 400; (d) Chou, T. Q.; Fu, F. Y.; Kao, Y. S. J. Am. Chem. Soc. 1948, 70, 1765; (e) Kuehl, Jr., F.
A.; Spencer, C. F.; Folkers, K. J. Am. Chem. Soc. 1948, 70, 2091; (f) Koepfli, J. B.; Mead, J. F.; Brockman, Jr., J. A.
J. Am. Chem. Soc. 1949, 71, 1048.
2. For reviews, see: (a) Takeuchi, Y.; Harayama, T. Trends Het. Chem. 2001, 7, 65; (b) McLaughlin, N. P.; Evans,
P.; Pines, M. Bioorg. Med. Chem. 2014, 22, 1993; (c) Pines, M.; Spector, I. Molecules 2014, 20, 573.
3. Solutions of 1 (as a free-base) gradually form 2. For example, heating in CHCl3 can be used to convert 1 into
2, see: Uesato, S.; Kuroda, Y.; Kato, M.; Fujiwara, Y.; Hase, Y.; Fujita, T. Chem. Pharm. Bull. 1998, 46, 1 and
reference 13 (b); whereas, heating 2 in MeOH (or H2O) has been used as a tactic to form equilibrium amounts
of 1, see: (a) Takeuchi, Y.; Azuma, K.; Takakura, K.; Abe, H.; Kim, H.-S.; Wataya, Y.; Harayama, T. Tetrahedron
2001, 57, 1213; (b) Huang, P.-Q.; Wei, B.-G.; Ruan, Y.-P. Synlett 2003, 1663.
4. (a) Baker, B. R.; Schaub, R. E.; McEvoy, F. J.; Williams, J. H. J. Org. Chem. 1952, 17, 132; (b) Baker, B. R.;
McEvoy, F. J.; Schaub, R. E.; Joseph, J. P.; Williams, J. H. J. Org. Chem. 1953, 18, 153; (c) Baker, B. R.; McEvoy, F.
J. J. Org. Chem. 1955, 20, 136.
5. (a) Barringer, Jr., D. F.; Berkelhammer, G.; Carter, S. D.; Goldman, L.; Lanzilotti, A. E. J. Org. Chem. 1973, 38,
1933; (b) Barringer, Jr., D. F.; Berkelhammer, G.; Wayne, R. S. J. Org. Chem. 1973, 38, 1937.
6. Kobayashi, S.; Ueno, M.; Suzuki, R.; Ishitani, H.; Kim, H. –S.; Wataya, Y. J. Org. Chem., 1999, 64, 6833.

7. Duchenne muscular dystrophy trial suspended using (±)-halofuginone NCT01978366. Safety, Tolerability and
Pharmacokinetics of Single and Multiple Doses of HT-100 in Duchenne Muscular Dystrophy.
8. Takaya, Y.; Tasaka, H.; Chiba, T.; Uwai, K.; Tanitsu, M.; Kim, H. –S.; Wataya, Y.; Miura, M.; Takeshita, M.;
Oshima, Y. J. Med. Chem. 1999, 42, 3163.
9. Deng, Y.; Xu, R.; Ye, Y. J. Chin. Pharm. Sci. 2000, 9, 116 (12 Kg of Dichroa febrifuga leaves were extracted
with EtOH to generate 40 mg of neodichroine).
10. Patnam, R.; Chang, F. –R.; Chen, C. –Y.; Kuo, R. –Y.; Lee, Y. –H.; Wu, Y. –C. J. Nat. Prod. 2001, 64, 948 (2 kg
of Hydrangea chinensis roots were extracted with MeOH to ultimately generate 10 mg of hydrachine A).
11. For a recent review concerning indolizidine and quinolizidine natural products, see: Michael, J. P. Alkaloids
2016, 75, 1.
12. Chang, F. –R.; Lee, Y. –H.; Yang, Y. –L.; Hsieh, P. –W.; Khalil, A. T.; Chen, C. –Y.; Wu, Y. –C. J. Nat. Prod. 2003,
66, 1245.
13. (a) McLaughlin, N. P.; Evans, P. J. Org. Chem. 2010, 75, 518; (b) Smullen, S.; Evans. P. Tetrahedron 2017, 73,
5493.
14. Trauner, D.; Danishefsky, S. J. Tetrahedron Lett. 1999, 40, 6513.
15. For recent febrifugine syntheses, see: (a) Sukemoto, S.; Oshige, M.; Sato, M.; Mimura, K. –I.; Nishioka, H.;
Abe, H.; Harayama, T.; Takeuchi, Y. Synthesis 2008, 3081; (b) Sieng, B.; Ventura, O. L.; Bellosta, V.; Cossy, J.
Synlett 2008, 1216; (c) Sudhakar, N.; Srinivasulu, G.; Rao, G. S.; Rao, B. V. Tetrahedron: Asymm. 2008, 19, 2153;
(d) Liu, R. –C.; Huang, W.; Ma, J. –Y.; Wei, B. –G.; Lin, G. –Q. Tetrahedron Lett. 2009, 50, 4046; (e) Wijdeven, M.
A.; van den Berg, R. J. F.; Wijtmans, R.; Botman, P. N. M.; Blaauw, R. H.; Schoemaker, H. E.; van Delft, F. L.;
Rutjes, F. P. J. T. Org. Biomol. Chem. 2009, 7, 2976; (f) Ashoorzadeh, A.; Archibald, G.; Caprio, V. Tetrahedron
2009, 65, 4671; (g) Emmanuvel, L.; Kamble, D. A.; Sudalai, A. Tetrahedron: Asymm. 2009, 20, 84; (h) Wang, R.;
Fang, K.; Sun, B. –F.; Xu, M. –H.; Lin, G. –Q. Synlett 2009, 2301; (i) Pansare, S. V.; Paul, E. K. Synthesis 2013, 45,
1863.
16. For examples of -aminooxylation-HWE sequences, see: (a) Zhong, G.; Yu, Y. Org. Lett. 2004, 6, 1637; (b)
Kondekar, N. B.; Kumar, P. Org. Lett. 2009, 11, 2611; (c) Gotoh, H.; Hayashi, Y. Chem Commun. 2009, 3083; (d)
Chouthaiwale, P. V.; Kotkar, S. P.; Sudalai, A. Arkivoc 2009, 88; (e) Cui, M.; Song, H.; Feng, A.; Wang, Z.; Wang
Q. J. Org. Chem. 2010, 75, 7018; (f) Vishwajeet, J.; Kondekar, N. B.; Kumar, P. Org. Lett. 2010, 12, 2762; (g)
Sabitha, G.; Yadagiri, G. C. K.; Yadav, J. S. Tetrahedron Lett. 2010, 51, 3824; (h) Kumaraswamy, G.; Sadaiah, K.;
Raghu, N. Tetrahedron: Asymmetry 2012, 23, 587; (i) Devalankar, D. A.; Chouthaiwale, P. V.; Sudalai, A.
Tetrahedron: Asymmetry 2012, 23, 240; (j) Reddy, S.; Venkateswarlu, Y. Tetrahedron Lett. 2013, 54, 4617; (k)
Doherty, W.; Evans, P. J. Org. Chem. 2016, 81, 1416.
17. Gérard, S.; Nollet, G.; Vande Put, J.; Marchand-Brynaert, J. Bioorg. Med. Chem. 2002, 10, 3955.
18. For a short review on Eschenmoser’s salt, see: Winterfeldt, E. J. Prakt. Chem. 1994, 336, 91.

Highlights
• New organocatalysis-based method for the synthesis of optically active febrifugine (both
enantiomers)
• First total synthesis of reported quinolizidine natural product
• Clarification of the structure of the reported natural product

Asymmetric synthesis of (–)- and (+)-neodichroine/hydrachine A from (+)- and (–)-febrifugine
OH
H
Shaun Smullen and Paul Evans
N
OH
OH
N
O
NMe
3 steps
lit.
2
O
O
4 steps
O
I
N
O
BocN
Cbz
O
N
N
Cbz
N
H
MeOH, ,
8%
98% ee
40%
N
(+)-Febrifugine
-aminooxylation-
Horner-Wadsworth-Emmons
Intramolecular
Mannich
()-Neodichroine/
hydrachine A