Application of iron carbonyl complexation to the selective total synthesis of sanshools

Article abstract of DOI:10.1016/j.tetlet.2012.08.135

We focused on the iron-complexes of sanshools (hydroxyl-α-sanshool, hydroxyl-β-sanshool, and γ-sanshool) as the key intermediates for the selective synthesis of these structurally unstable compounds. Consequently, we developed a concise and selective method for the total synthesis of sanshools in 5-6 steps (26-45% overall yield), including complexation of dienes with iron tricarbonyl group.

Full text of DOI:10.1016/j.tetlet.2012.08.135

Tetrahedron Letters 53 (2012) 6000–6003
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Application of iron carbonyl complexation to the selective total synthesis
of sanshools
⇑
Katsuyuki Aoki, Yasushi Igarashi , Hiroaki Nishimura, Isao Morishita, Kimitoshi Usui
Kampo Formulations Development Center, Tsumura & Co., 3586 Yoshiwara, Ami-mach, Inashiki-gun, Ibaraki 300-1192, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
We focused on the iron-complexes of sanshools (hydroxyl-a-sanshool, hydroxyl-b-sanshool, and c-san-
Received 19 July 2012
Revised 10 August 2012
Accepted 22 August 2012
Available online 5 September 2012
shool) as the key intermediates for the selective synthesis of these structurally unstable compounds. Con-
sequently, we developed a concise and selective method for the total synthesis of sanshools in 5–6 steps
(26–45% overall yield), including complexation of dienes with iron tricarbonyl group.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Hydroxyl-
Hydroxyl-b-sanshool
-Sanshool
a-sanshool
c
Iron complex
Zanthoxylum fruit
Daikenchuto
Introduction
Although this was the first report describing the total synthesis
of 1 and 2 with high geometrical selectivity, there were rooms
Sanshools, such as hydroxy-
(2), and c-sanshool (3), were originally isolated as unstable unsat-
a
-sanshool (1), hydroxy-b-sanshool
for further improvement with respect to the stability of the reac-
tion intermediates, which were relatively unstable, and the cost
effectiveness of the reagents used in the synthesis.
urated aliphatic amides from the dried fruits of the Japanese pep-
per (Zanthoxylum piperitum) (Fig. 1).1 Each one of these compounds
elicits unique tingling and sometimes numbing sensations on the
tongue, presumably through the activation of transient receptor
potential channels2 and inhibition of potassium channels.3 As
these compounds were also detected in the human plasma and ur-
ine following the oral administration of Daikenchuto (DKT), a
known traditional Japanese medicine,4 they are, therefore, consid-
ered as key compounds for delineating the mechanism of action of
DKT.
Incidentally, the first metal complex of a conjugated diene was
reported in 1930 by Reihlen et al.9 Since then, many researchers
have shown interest in using the diene-metal complexes as the
starting material in organic synthesis, as a result of which conju-
gated diene-metal complexes of nearly all transition metals had
been prepared and/or characterized. Therefore, numerous reviews
devoted to the application of conjugated metal-diene complexes in
organic synthesis are now available. In this regard, it is noteworthy
Despite their interesting pharmacological profiles, instability of
sanshools (such as isomerization, oxidation, polymerization, and/
or photo-degradation), due to their conjugated triene structures,5
made it difficult to prepare them synthetically since isolation.
Thus, even though low-selective Wittig reaction was successfully
O
O
N
N
H
OH
H
OH
used for the synthesis of a-, b-, and c-sanshools, all of which lacked
hydroxy-α-sanshool (1)
hydroxy-β-sanshool (2)
a hydroxyl group,6,7 geometrically selective total synthesis of 1, 2,
and 3 remained elusive so far.
Recently, we have developed a concise method for achieving the
O
total synthesis of two sanshools (hydroxyl-a- and hydroxy-b-san-
shool) by utilizing the Suzuki–Miyaura Coupling reaction.8
N
H
γ-sanshool (3)
⇑
Corresponding author. Tel.: +81 29 889 3822; fax: +81 29 889 2158.
E-mail address: igarashi_yasushi@mail.tsumura.co.jp (Y. Igarashi).
Figure 1. Structures of hydroxy-a-sanshool (1), hydroxy-b-sanshool (2), and
c-sanshool (3).
0040-4039/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2012.08.135

K. Aoki et al. / Tetrahedron Letters 53 (2012) 6000–6003
6001
that the (diene)iron tricarbonyl complexes10 have especially be-
come useful reagents in organic synthesis due to the ease of prep-
Z-selectivity, which could have resulted from the steric effect of
Fe(CO)3 complex, was observed under this condition, and 7 was ob-
tained as the single geometrical isomer. It was reported earlier that
not only the E-isomer but also the Z-isomer was low-selectively
formed (E/Z = 1/4.9) in a similar reaction where the diene did not
have the protection of metal complexation.7 In our Wittig reaction
between 2,4-hexadienal and compound 11, the corresponding Z-
isomer was also low-selectively formed (E/Z = 1/6) in 72% yield.
In general, the Wittig reaction yields preferentially Z-alkenes for
nonstabilized triphenylphosphonium ylides under salt-free condi-
tions. The stereochemical outcome of the Wittig reaction is be-
lieved to be the result of steric effects16 (1,2 and 1,3 interactions)
that develop as the bulky triphenyl group on phosphonium ylide
and the functional group of aldehyde approach one another. There-
fore, the high Z-selectivity was observed at more sterically-bulky 9
compared to unprotected 2,4-hexadienal. Consequently, as shown
here, we achieved creating the E, E, Z-triene structure. Subsequent
reduction of 7 with DIBAL-H, followed by HWE reaction using
Douglas’s condition17 gave 10 relatively easily in 75% yield. Amida-
tion of 10 with 1-amino-2-methylpropan-2-ol,8 then gave com-
pound 4 in good yield. Iron carbonyl complex 4 was obtained as
more stable solid than unprotected 11,5 which was known to exist
as unsuitable material at room temperature. Finally, de-protection
of 4 using trimethylamine-oxide in acetonitrile resulted in 82%
yield of compound 1.
aration, stability toward
a variety of reaction conditions,
availability of multiple methods for the removal of the iron tricar-
bonyl (Fe(CO)3) group, and low costs of iron carbonyls.
When considering 2,4-hexadienal as the initial substrate for the
preparation of the triene structure, we thought that the protection
imparted by the Fe(CO)3 moiety on the diene structure would be
stereo-electronically effective (e.g., provide protection to the labile
triene structure, improve reaction selectivity) in controlling the
constricted course of reaction required for achieving the total syn-
thesis of sanshools. We also thought that if these iron-complexes
could exist as relatively stable solids, then each sanshool could
be prepared timely on demand. With these objectives in mind,
we focused on the iron-complex of each sanshool as the potentially
important intermediate.
Results and discussion
The retrosynthetic strategy for obtaining 1, 2, and 3 is shown in
Scheme 1. The similar concept using iron carbonyl complex for
olefination reaction has been already reported.11 Briefly, the first
complex formation of 2, 4-hexadienal with Fe3(CO)12 would pro-
vide the corresponding Fe(CO)3-aldehyde 9. Next, Wittig12 or Julia
olefination13 of 9 would then result in Fe(CO)3-nitriles 7 and 8 as
the geometrical isomers. Then, reduction of 7 and 8 with diisobu-
tylaluminium hydride (DIBAL-H), followed by Horner–Wads-
worth–Emmons (HWE) reaction and/or amidation would give
Fe(CO)3-carboxilic acid derivatives as the important intermediates
4, 5, and 6. Finally, synthesis of the desired sanshools 1, 2, and 3
would be achieved by de-protection of the intermediates 4, 5,
and 6. In this Letter, we report the application of iron carbonyl
complexation to the selective total synthesis of sanshools 1, 2,
and 3.
Synthesis of hydroxy-b-sanshool (Scheme 3)
To obtain hydroxy-b-sanshool, we next explored avenues to
selectively synthesize compound 8. The results of Julia-Kocienski
olefination18 of 9 with the benzothiazol-2-yl sulfone 13 under var-
ious reaction conditions are summarized in Table 2. A typical sol-
vent effect on geometrical selectivity was observed in Method A.
Thus, the reaction carried out in tetrahydrofuran (THF) was shown
to provide higher E-selectivity than that in N,N-dimethylformam-
ide (DMF) (entry 1 vs entry 2). The amount of 13 and/or the
amount of base used in the reaction did not influence the yield
in Method A (entry 2 vs entry 3). On the other hand, the yield
was significantly improved and the selectivity was maintained in
reaction Method B, and compound 8 was obtained in 74% yield
(60% isolated yield) (entry 3 vs entry 4). Interestingly, the reaction
Method C did not work at all (entry 5). Although the reason of the
case has not been clarified yet, we expect a stable and regenerable
complex involving 9 and base to generate in reaction Method C.
This arises from the quantitative recovery of 9 after work-up. Thus,
as shown here, we achieved creating the E, E, E-triene structure.
Iron carbonyl complex 8 was obtained as more stable solid than
unprotected 21 which was known to exist as unsuitable material
at room temperature. Subsequently, compound 2 was prepared
Synthesis of hydroxy-a-sanshool (Scheme 2)
Compound 9 was prepared by complexation of 2,4-hexadienal
with dodecarbonyl triiron (Fe3(CO)12) following a method de-
scribed earlier.14 Next, we explored for the optimal condition for
the Wittig reaction between compound 9 and compound 11,15
which was prepared separately, and the results are summarized
in Table 1. In this reaction, compound 7 was synthesized in low
yield when either n-butyl lithium (n-BuLi) or potassium tert-
butoxide (t-BuOK) was used as a base (entries 1 and 2), and in
low to moderate yield when sodium hydride was used as a base
(entries 3 and 4). However, the result obtained using sodium
bis(trimethylsilyl)amide (NaHMDS) as a base was very promis-
ing—it yielded compound 7 in high yield (entry 5). Indeed, high
Fe(CO)3
O
N
O
N
R
R
n
n
H
H
1, (Z), n = 1, R = OH
2, (E), n = 1, R = OH
3, (Z), n = 2, R = H
4, (Z), n = 1, R = OH
5, (E), n = 1, R = OH
6, (Z), n = 2, R = H
CHO
CHO
Fe(CO)3
Fe(CO)3
CN
7, (Z)
8, (E)
9
2,4-hexadienal
Scheme 1. The retrosynthetic strategy for the sanshools (1, 2 and 3).

6002
K. Aoki et al. / Tetrahedron Letters 53 (2012) 6000–6003
Fe(CO)3
a)
b)
CHO
2,4-hexadienal
CHO
Fe(CO)3
80%
CN
O
9
7
4
Fe(CO)3
Fe(CO)3
c)
d)
90%
O
75%
N
H
OH
OH
10
O
e)
N
H
93%
OH
1
Ph3P+
CN Br-
11
Scheme 2. Reagents and reaction conditions: (a) Fe3(CO)12, toluene, reflux; (b) See Table 1; (c) (1) DIBAL-H, toluene, À20 oC, (2) Zn(OTf)2, TMEDA, (EtO)2POCH2CO2H, THF,
room temperature (rt); (d) EDC, HO-Bt, NH2CH2CMe2OH,8 CH2Cl2, rt; (e) Me3NO, CH3CN, rt.
O
Fe(CO)3
Fe(CO)3
b)
a)
CN
9
OH
70%
8
12
O
O
Fe(CO)3
c)
d)
N
H
N
H
OH
OH
94%
83%
5
2
N
N
13
N
SO2
CN
N
Ph
Scheme 3. Reagents and reaction conditions: (a) See Table 2; (b) (1) DIBAL-H, toluene, À20 oC, (2) Zn(OTf)2, TMEDA, (EtO)2POCH2CO2H, THF, room temperature (rt); (c) EDC,
HO-Bt, NH2CH2CMe2OH,8 CH2Cl2, rt; (d) Me3NO, CH3CN, rt.
Fe(CO)3
Fe(CO)3
O
a)
7
6
N
H
+
6
14
b)
O
O
N
H
N
78%
H
3
H
N
(EtO)2
15
P
O
O
Scheme 4. Reagents and reaction conditions: (a) (1) DIBAL-H, toluene, À20oC, (2) See Table 3; (b) Me3NO, CH3CN, room temperature.
Table 1
Synthesis of
c-sanshool (Scheme 4)
Synthesis of 7 via Wittig olefination
Entry
1
Condition
Yield (%)
28
To explore the possibility of synthesizing
c
-sanshool, we next
(1) 11, n-BuLi, HMPA, THF, 0 °C, 0.5 h
(2) 9, THF, À78?0 oC, 4 h
carried out Horner–Wadsworth–Emmons olefination of 7 with
compound 15,19 which was prepared previously. The results are
summarized in Table 3. When NaHMDS was used as the base under
low temperature condition, the isomer 14, but not the desired
compound 6, was obtained (entries 1 and 2). The addition of 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU) moderately improved the
geometrical selectivity (entry 2). On the other hand, the use of lith-
ium hydroxide20 as the base made it possible to selectively synthe-
size compound 6 (entry 3). Use of increased amounts of 15 and
lithium hydroxide in the reaction mixture further improved the
yield of 6 (entry 3), even when the reaction time was significantly
reduced (compare reaction times of entry 3 vs entry 4). However,
the geometrical selectivity did not improve further when DBU
2
3
4
5
(1) 11, t-BuOK, DMSO-THF, rta, 0.25 h
(2) 9, THF, À20oC, 1 h
17
55
21
90
(1) 11, NaH, DMSO-THF, 0 °C, 1 h
(2) 9, THF, 0 oC, 2.5 h
(1) 11, NaH, DMSO, rt, 1 h
(2) 9, DMSO, rta, 1 h
(1) 11, NaHMDS THF, 0 °C?rt, 1.5 h
(2) 9, THF, À50 °C?rta, 2.5 h
a
rt: Room temperature.
from compound 8 in high yield following the de-protection proce-
dure described above.

K. Aoki et al. / Tetrahedron Letters 53 (2012) 6000–6003
6003
Table 2
Synthesis of 8 via Julia-Kocienski olefination
Entry
Methoda
13 (eq)
KHMDS (eq)
Condition
Solvent
Yieldb (%)
7/8b
1
2
3
4
5
A
A
A
B
C
1.1
1.1
1.5
1.5
1.5
1.1
1.1
2.0
2.0
2.0
À55oC ? À30 °C 1 h
DMF
THF
THF
THF
THF
47
36/64
2/98
4/96
4/96
À78 °C, 2 h ? rtc, over nightd
À78 °C, 2 h ? rtc, over nightd
À78 °C, 2 h ? rtc, over nightd
À78 °C, 1 h ? rtc, over nightd
46
41e
74 (60)e
No reaction
Not tested
a
A: base was added to the sulfone and then carbonyl was added. B: base was added to a mixture of sulfone and carbonyl. C: base was added to the carbonyl and then
sulfone was added.
b
Yield and isomer ratio were calculated by integration of the 1H NMR (400 MHz) spectra of the mixture of 7 and 8.
rt: Room temperature.
The reaction mixture was stirred for approximately 20 h (reactions times were not optimized).
Isolated yield.
c
d
e
Table 3
Supplementary data
Synthesis of 6 via HWE olefination
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2012.08.
135.
Entry
1
Condition
Isolated yield
15 (1.3 equiv), NaHMDS (1.4 equiv),
THF, À20 °C ꢀrta, 4 h
6 (19%) 14 (36%)
2
15 (1.4 equiv), NaHMDS (1.4 equiv),
DBU (5 equiv), THF, À20 °C ꢀrt, 1.5 h
15 (1.3 equiv), LiOH (1.5 equiv), MS4Ac,
THF, refld, 12 h
6 (28%) 14 (30%)
6 (40%)
References and notes
3b
4b
5b
1. (a) Yasuda, I.; Takeya, K.; Itokawa, H. Phytochemistry 1982, 21, 1295–1298; (b)
Yasuda, I.; Takeya, K.; Itokawa, H. Chem. Pharm. Bull. 1981, 29, 1791–1793.
2. Koo, J. Y.; Jang, Y.; Cho, H.; Lee, C-H.; Jang, K. H.; Chang, Y. H.; Shin, J.; Oh, U. Eur.
J. Neurosci. 2007, 26, 1139–1147.
3. Bautista, D. M.; Sigal, Y. M.; Milstein, A. D.; Garrison, J. L.; Zorn, J. A.; Tsuruda, P.
R.; Nicoll, R. A.; Julius, D. Nat. Neurosci. 2008, 11, 772–779.
4. (a) Iwabu, J.; Watanabe, J.; Hirakura, K.; Ozaki, Y.; Hanazaki, K. Drug Metab.
Dispos. 2010, 38, 2040–2048; (b) Munekage, M.; Kitagawa, H.; Ichikawa, K.;
Watanabe, J.; Aoki, K.; Kono, T.; Hanazaki, K. Drug Metab. Dispos. 2011, 39,
1784–1788.
15 (2.0 equiv), LiOH (4.0 equiv), MS4Ac,
THF, refld, 2 h
6 (56%)
15 (2.0 equiv), LiOH (4.0 equiv), MS4Ac,
DBU (1.0 equiv), THF, refld, 2 h
6 (54%)
a
b
c
rt: room temperature.
Trace amount of 14 was observed by TLC.
Aldrich 2–3
refl: reflux.
l activated powder of MS4A was used (500 mg/100 mg of 7).
d
5. Yang, X. J. Agric. Food Chem. 2008, 56, 1689–1696.
6. Sonnet, P. E. J. Org. Chem. 1969, 34, 1147–1149.
7. Crombie, L.; Fisher, D. Tetrahedron Lett. 1985, 26, 2481–2484.
8. Igarashi, Y.; Aoki, K.; Nishimura, H.; Morishita, I.; Usui, K. Chem. Pharm. Bull.
2012, 60, 1088–1091.
9. Reihlen, H.; Grüghl, A.; von Hessling, G.; Pfrengle, O. Leibig Ann. Chem. 1930,
482, 161.
10. (a) William, A.; Donaldson Curr. Org. Chem. 2000, 4, 837–868; (b) William, A.
D.; Subhabrata, C. Eur. J. Org. Chem. 2009, 3831–3843.
11. (a) Lellouche, J. P.; Gigou-Barbedette, A.; Grée, R. Bull. Soc. Chim. Fr. 1993, 129,
605–624; (b) Wada, A.; Hiraishi, S.; Takamura, N.; Date, T.; Aoe, K.; Ito, M. J.
Org. Chem. 1997, 62, 4343–4348; c) Wada, A.; Fujioka, N.; Imai, H.; Shichida, Y.;
Ito, M. Bioorg. Med. Chem. Lett. 1998, 8, 423–426.
was added to the reaction mixture to promote the E-selectivity
(entry 5). Iron carbonyl complex 6 was obtained as more stable so-
lid than unprotected 31 which was known to exist as unsuitable
material at room temperature. Next, following the procedure de-
scribed above, de-protection of compound 6 provided compound
3 in 78% yield.
In summary
12. (a) Maercker, A. Org. React. 1965, 14, 270–490; (b) Carruthers, W. In Some
Modern Methods of Organic Synthesis; Cambridge University Press: Cambridge,
UK, 1971; pp 81–90; (c) Hoffmann, R. W. Angewandte Chemie International
Edition 2001, 40, 1411–1416.
A concise and selective method for the total synthesis of san-
shools (hydroxyl-
achieved by complexation of the labile triene structure with
Fe(CO)3. Hydroxy- -sanshool (1) was synthesized from the com-
mercially available 2,4-hexadienal in 45% overall yield using six
reaction steps. Hydroxy-b-sanshool (2) was also synthesized in
a-, hydroxyl-b-, and c-sanshool) has been
13. (a) Julia, M.; Paris, J. M. Tetrahedron Lett. 1973, 4833–4836; (b) Kocienski, P. J.;
Lythgoe, B. J. Chem. Soc., Perkin Trans 1 1980, 1045–1050.
14. Iwan, W.; Benson, M. K.; Keith, R.; Liam, R. C. Org. Lett. 2006, 8(20), 4389–4392.
15. Gary, A. M.; Ruth, F. J. Org. Chem. 2006, 71, 6135–6140.
16. (a) Schlosser, M.; Schaub, B. J. Am. Chem. Soc. 1982, 104, 5821–5823; (b) Vedejs,
E.; Fleck, T.; Hara, S. J. Org. Chem. 1987, 52, 4637–4639; (c) Vedejs, E.; Marth, C.
F. J. Am. Chem. Soc. 1988, 110, 3948–3958; (d) Vedejs, E.; Fleck, T. J. J. Am. Chem.
Soc. 1989, 111, 5861–5871; (e) Mari, F.; Lahti, P. M.; McEwen, W. E. Heteroat.
Chem. 1991, 2, 265–276; f) Mari, F.; Lahti, P. M.; McEwen, W. E. J. Am. Chem. Soc.
1992, 114, 813–821; (g) Gu, Y.; Tian, S.-K. Top. Curr. Chem. 2012, 1–42.
17. Douglas, J. S.; Paul, H. Synthesis 2006, 21, 3654–3660.
a
26% overall yield using six reaction steps, whereas c-sanshool (3)
was synthesized in 31% overall yield using five reaction steps. For-
tunately, iron carbonyl complex 4, 5, and 6 were obtained as stable
solids at room temperature. Therefore, our method is optimal as far
as the process control of labile sanshools is concerned, as each san-
shool could be prepared by de-protection of the corresponding
intermediate whenever needed. In addition, the method described
here is economic and affordable, as none of the reagents used here
is expensive.
Thus, a simple method for the selective synthesis of sanshools
has become available by complexation of iron carbonyls to the la-
bile conjugated polyene structure of the sanshools. As the present
method can accommodate a wide range of polyene compounds, it
could also be used for the synthesis of other natural products, such
as many fatty acids.
18. Smith, A. B., III; Zehong, W. J. Org. Chem. 2000, 65, 3738–3753.
19. Same procedure as in the preparation of 4 gave phosphine reagent 15. (See also
Reference 17).
EDC, HO-Bt,
NH2CH2CHMe2
H
(EtO)2
OH
in CH2Cl2 r.t.
88%
(EtO)2
N
P
P
O
O
O
O
16
15
20. Takacs, J. M.; Jaber, M. R.; Clement, F. C.; Walters, C. J. Org. Chem. 1998, 63,
6757–6760.