A scalable process for the synthesis of (E)-pterostilbene involving aqueous Wittig olefination chemistry

Article abstract of DOI:10.1016/j.tetlet.2013.09.019

A synthetic approach toward the pharmacologically active (E)-stilbene pterostilbene is described using a Wittig reaction conducted under mildly basic, aqueous conditions. A surprising, non-intuitive difference in (E)/(Z) stereoselectivity was observed comparing the two possible isomeric Wittig routes, allowing for the development of a highly efficient process to access the title stilbene derivative through a one-pot olefination deprotection sequence.

Full text of DOI:10.1016/j.tetlet.2013.09.019

Tetrahedron Letters 54 (2013) 6303–6306
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A scalable process for the synthesis of (E)-pterostilbene involving
aqueous Wittig olefination chemistry
⇑
James McNulty , David McLeod
Department of Chemistry & Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M1, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
A synthetic approach toward the pharmacologically active (E)-stilbene pterostilbene is described using a
Wittig reaction conducted under mildly basic, aqueous conditions. A surprising, non-intuitive difference
in (E)/(Z) stereoselectivity was observed comparing the two possible isomeric Wittig routes, allowing for
the development of a highly efficient process to access the title stilbene derivative through a one-pot olef-
ination deprotection sequence.
Received 22 August 2013
Accepted 05 September 2013
Available online 16 September 2013
Keywords:
Wittig reaction
Olefination
Ó 2013 Elsevier Ltd. All rights reserved.
Aqueous chemistry
Stilbene
Pterostilbene
Introduction
suffers from notoriously poor (E):(Z) stereocontrol and the prob-
lematic removal of triphenylphosphine oxide. These problems
Stilbenes and their derivatives constitute important classes of
both natural products and synthetic small molecules that are
highly valued in areas as diverse as pharmaceuticals,1 imaging
agents,2 in light emitting diodes and photovoltaic devices (as
organic dyes)3 and also in an expanding array of other applications
in the life sciences and materials chemistry.4 A selection of phar-
macologically active stilbenes and derivatives is collected in
Figure 1. These include the substituted (E)-stilbene resveratrol,1
the anticancer agent DMU2125, and pterostilbene 1,6 a compound
that is associated with a wide range of cardiovascular activities
including anti-inflammatory activity and the lowering of plasma
lipoprotein and cholesterol levels.6b Pterostilbene formulations
are currently under clinical evaluation for the treatment of high
blood pressure and in reducing oxidative stress.6c Several (Z)-stilb-
enes are endowed with potent anticancer activity including com-
bretastatin A41e,1fand the related stilstatins.7 Clinically approved
stilbenes include florbetapir F18, a compound that binds to myelin
and is used as a molecular imaging probe in positron emission
tomography (PET) in the diagnosis of disorders such as multiple
sclerosis,2 and the anti-asthmatic pharmaceutical singulair.8 While
many synthetic methods are available to access stilbenes,9 the
principal direct methods employed generally involve Wittig and
related Horner-type reactions.9a–d,7 Unfortunately, the classic Wit-
tig olefination approach to stilbenes, including pterostilbene,10
employing triphenylphosphine-derived semi-stabilized ylides,
have largely been solved through the use of aqueous Wittig meth-
ods employing ylides derived from short chain trialkylphosphines.
We have extensively investigated aqueous Wittig olefination
reactions over the last few years11 showing that stabilized and
semi-stabilized ylides can be efficiently generated and trapped un-
der increasingly milder chemical conditions delivering olefins with
high (E)-olefin stereoselectivity,11a–c including the stilbenes
DMU212 and resveratrol.11a Concerning the synthesis of stilbenes
specifically, in addition to high (E)-olefin sterocontrol,11a,c the
aqueous Wittig reaction with short-chain trialkylphosphine-de-
rived ylides provides remarkably simplified post-reaction process-
ing in comparison to classic Wittig protocols. Short chain
trialkylphosphine oxides such as triethylphosphine oxide and tri-
propylphosphine oxide are water-soluble and processing is
straightforward allowing clean separation and filtration of crystal-
line stilbene products in most cases.11a,c Where the products are
oils, the phosphine oxide side products can easily be removed
through simple aqueous/organic solvent partition.11a–d In view of
the current high interest in the development of a pterostilbene
based therapeutic6,10 and the high cost of commercial (E)-pterostil-
bene,12 we decided to investigate the possibility of a scalable aque-
ous Wittig approach to (E)-pterostilbene from readily available
precursors that would benefit from the chemical and processing
advantages described above. In this Letter we report the investiga-
tion of the two direct Wittig routes to pterostilbene, the discovery
of a non-intuitive remote meta-substituent effect on the stereose-
lectivity of the reaction and the development of a scalable, highly
optimized route to this valuable stilbene.
⇑
Corresponding author. Tel.: +1 905 525 9140x27393.
E-mail address: jmcnult@mcmaster.ca (J. McNulty).
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.09.019

6304
J. McNulty, D. McLeod / Tetrahedron Letters 54 (2013) 6303–6306
OH
N
O
O
O
18F
HO
H3C
Resveratrol
N
H
OH
O
Florbetapir F18
OH
MeO
MeO
S
Cl
N
HO
OMe
OMe
Singulair
OH
Combretastatin A4
OH
OMe
MeO
MeO
MeO
OMe
DMU212
Pterostilbene
OMe
Figure 1. A selection of pharmacologically active stilbenes and analogs.
NaBH4 gave the benzyl alcohol 13 which was directly converted to
the tripropylbenzyl phosphonium salt using tripropylphosphine
hydrochloride, affording 12b in good yield as a crystalline solid.
Once again, the Wittig reaction of aldehyde 8 using this unpro-
tected salt 12b failed to produce pterostilbene 1, however during
the course of experimentation on protected alternatives, an inter-
esting discovery was made with use of the tosyl protected phos-
phonium salt 12a that allowed for a satisfactory conclusion to
the pterostilbene synthesis. The tosyl-protected tripropylphospho-
nium salt 12a was prepared in three steps from commercial
4-hydroxybenzaldehyde 9. Protection of 9 as the tosyl derivative
gave aldehyde 10 (95%) which was reduced to the alcohol 11 using
a similar route described above for alcohol 4. The benzyl alcohol 11
was again directly converted to the benzylic phosphonium salt 12
through treatment with tripropylphosphine hydrochloride. Pheno-
lic tosylates are known to slowly saponify under aqueous basic
conditions14 and so there was initially some concern about the
aqueous Wittig reaction of salt 12a with 8. In the event, the Wittig
reaction with aldehyde 8 and phosphonium salt 12a under our
standard aqueous conditions at 100 °C was observed to be rapid
(aldehyde disappearance <3 h), but was not clean. The reaction
produced both the tosyl derivative of pterostilbene as well as fully
deprotected pterostilbene 1. Simply extending the reaction time to
Results and discussion
Two possible Wittig-disconnections can be considered as routes
to pterostilbene, A and B (see Graphical abstract). We initially
focused on the synthesis of pterostilbene following the route A
disconnection (Scheme 1). To this end, the phosphonium salt 6
was prepared from 3,5-dihydroxybenzoic acid 2 using standard
chemistry. Notably, the benzyl alcohol 4 was readily converted to
the phosphonium salt 6 on heating with tripropylphosphine
hydrochloride or via the benzyl chloride 5 allowing batches of
phosphonium salt 6 to be made reliably on a 10 g scale. Reaction
of this salt was initially attempted on the unprotected 4-hydroxy-
benzaldehyde 7a but showed no product formation using a variety
of aqueous bases. The phenol 7a was therefore protected as its THP
acetal which reacted successfully to yield the protected pterostil-
bene. Unfortunately, the (E):(Z)-stereoselectivity of this reaction
was determined to be unsatisfactory in comparison to other stilb-
enes synthesized employing trialkylbenzylphosphonium salts in
aqueous media.11a–c,i The reaction of 6 with 7a and with other pro-
tected derivatives (7b, 7c, and 7d) was investigated under a variety
of conditions however changing the phenol protecting group and
base afforded no improvement in the stereoselectivity observed
and we thus investigated the synthesis following disconnection B.
The required aldehyde 8 was prepared cleanly from the previ-
ously synthesized alcohol 4 using the Dess–Martin periodinane.
With the aldehyde in hand, work began on the para-hydroxy phos-
phonium salt 12, as outlined in Scheme 2. We initially decided to
attempt the olefination reaction using the protecting group free
phosphonium salt 12b. Reduction of 4-hydroxybenzaldehyde using
12 h resulted in
a clean, stepwise olefination/deprotection
sequence and pterostilbene was isolated in high yield (89%) and,
more importantly, with high stereoselectivity (>95:5, (E):(Z)). Of
several bases studied in this reaction, lithium hydroxide provided
the highest overall yield and the highest stereoselectivity in favor
of the (E)-isomer. As described above, as tripropylphosphine oxide
O
O
b
a
d
HO
MeO
MeO
MeO
OH
Cl
OMe
OH
3
2
4
5
OH
OMe
OMe
OMe
c
e
O
OR
MeO
MeO
H
+
-
PPr Cl
3
f
RO
+
1a-1d
6
OMe
OMe
7a R=H
7b R=Ts
7c R=Bz
(E):(Z) = 84:16 (+/- 4%)
7d R=THP
Scheme 1. Synthesis of pterostilbene 1 following Wittig disconnection A. Reagents and conditions: (a) Dimethyl sulfate (2.0 equiv), K2CO3, acetone, reflux, 95%; (b) NaBH4,
THF–MeOH, reflux, 92%; (c) tripropylphosphine hydrochloride, 100 °C, 24 h, 99%; (d) SOCl2, pyridine, 93%; (e) tripropylphosphine, PhMe, 80 °C, 99%; (f) various conditions, see
table (Supplementary data) and text.

J. McNulty, D. McLeod / Tetrahedron Letters 54 (2013) 6303–6306
6305
O
MeO
MeO
OH
a
H
4
8
OMe
O
OMe
H
O
d
b
c
PPr3+Cl-
12a
H
OH
11
HO
9
10
TsO
HO
TsO
TsO
HO
PPr3+Cl-
12b
f
OH
13
e
OH
O
PPr3+Cl-
MeO
MeO
g
H
+
RO
12
1
8
OMe
OMe
R= OTs 12a
12b (N.R.)
(91%), >95% (E)
H
Scheme 2. Synthesis of pterostilbene 1 following Wittig disconnection B. Reagents and conditions: (a) DMP, DCM, 91%; (b) TsCl, TEA, DCM, 95%; (c) NaBH4, DCM/MeOH (6:1),
SiO2, 90%; (d) tripropylphosphine hydrochloride, MeCN, 100 °C, overnight, 93%; (e) NaBH4, DCM/MeOH (6:1), SiO2, 81%; (f) tripropylphosphine hydrochloride, MeCN, 100 °C,
overnight, 91%; (g) LiOH (3 equiv), H2O (1 M), 90 °C, 12 h, 91%.
is fully water soluble, the phosphine oxide was removed through
partitioning with water/ethyl acetate and (E)-pterostilbene was
isolated in 91% yield through filtration through a silica plug and
recrystallization.
formation, and in-situ deprotection leading to 1. This process takes
advantage of many of the green features of the aqueous Wittig
reaction that have been disclosed in recent years that involve the
intermediacy of stabilized or semi-stabilized ylides.11 The use of
phosphonium salts derived of short-chain trialkylphosphines
allows for a high (E)-stereoselection as well as simple post reaction
processing allowing clean, chromatographically-free removal of
these water soluble phosphine oxide side products in many cases.
The successful use of water as solvent and weak bases ranging
from lithium hydroxide, to potassium carbonate and even weakly
Lastly, it is interesting to consider the possible origin of the
stereochemical differences observed in conducting the two regio-
isomeric Wittig approaches to pterostilbene. The Wittig reactions
of semi-stabilized ylides derived from short-chain trialkylphos-
phonium salts have been shown to provide higher than expected
(E)-olefin stereoselection in aqueous, salt-containing media.11a–c
This result may be either kinetic in origin, that is connected to
preference for the planar transition state leading to (E)-olefination
via a trans-oxaphosphetane,13c or thermodynamic through oxaph-
osphetane reversal and/or isomerization of intermediates prior to
elimination. In this regard, it is known that the presence of weakly
Lewis basic ortho-substituents (halogens or methoxy groups) in
benzylic phosphonium salts can result in increased (Z)-olefin ste-
reocontrol through enhanced kinetic control.13a,b,d It is interesting
to speculate the involvement of a minor ‘meta’-alkoxy effect con-
tributing to the increased formation of the (Z)-stilbene in reactions
with aldehyde 6, although steric factors do not appear favorable to
stabilizing the initial oxaphosphetane. The dominant factors deter-
mining stereochemistry are the employment of short chain alkyl
groups on the intermediate trialkylphosphorane resulting in
kinetic preference for the trans-oxaphosphetane as well as the po-
tential for equilibration under these aqueous, lithium salt-contain-
ing conditions. Strategic considerations toward any Wittig
disconnection advocate the use of the structurally simplest phos-
phonium salt reacting with an aldehyde, where feasible. These
results indicate the need to develop versatile synthetic approaches
to challenging cases that can allow both regioisomeric variations of
the olefination to be investigated.
basic amines such as L-proline, tosylamide etc. makes this mild
olefination chemistry attractive from a green-chemistry perspec-
tive. In addition, we wish to highlight the direct conversion of
benzylic (or allylic)11c alcohols directly (for example Scheme 2,
13–12b) to the corresponding trialkylphosphonium salts through
treatment with short-chain trialkylphosphine hydrohalide (HBr
or HCl) salts. This process eliminates the requirement for the use
of benzylic and allylic halides, which are generally cytotoxic alkyl-
ating agents/lachrymators, and removes the requirement for the
direct use of pyrophoric trialkylphosphines. Triethyl- and tripro-
pylphoshine hydrohalide and chloride salts are crystalline, air-sta-
ble materials and essentially odorless. Although hygroscopic, they
can be handled easily in the open laboratory. This direct benzyl/al-
lyl alcohol to phosphonium salt process has now been utilized
successfully in
a number of useful aqueous Wittig applica-
tions.11c–g The synthesis of pterostilbene achieved using this chem-
istry is readily scalable and economic with respect to the number
of steps required and the limited purification required throughout
the process.
Acknowledgments
We thank Cytec Canada Inc. and NSERC for financial support of
this work.
Conclusion
In conclusion, the two possible regioisomeric Wittig olefination
routes toward the synthesis of the pharmacologically active4
stilbene pterostilbene 1 were investigated under aqueous Wittig
conditions. A highly efficient, process-friendly route to pterostil-
bene was developed involving a high yielding (E)-selective stilbene
Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.tetlet.2013.
09.019.

6306
J. McNulty, D. McLeod / Tetrahedron Letters 54 (2013) 6303–6306
7. Pettit, G. R.; Thornhill, A.; Melody, N.; Knight, J. C. J. Nat. Prod. 2009, 72, 380–
References and notes
388.
8. Lipworth, B. P. The Lancet 1999, 353, 57–62.
1. (a) Aggarwal, B. B.; Bhardwaj, A.; Aggarwal, R. S.; Seeram, N. P.; Shishodia, S.;
Takada, Y. Anticancer Res. 2004, 24, 2783–2840; (b) Saiko, P.; Szakmary, A.;
Jaeger, W.; Szekres, T. Mutat. Res. 2008, 658, 68–94; (c) Sale, S.; Verschoyle, R.
D.; Boocock, D.; Jones, D. J. L.; Wilsher, N.; Ruparelia, K. C.; Potter, G. A.; Farmer,
P. B.; Steward, W. P.; Gescher, A. J. Br. J. Cancer 2004, 90, 736–744; (d) Baur, J. A.;
Sinclair, D. A. Nat. Rev. 2006, 5, 493–506; (e) Pettit, G. R.; Singh, S. B.; Niven, M.
L.; Hamel, E.; Schmidt, J. M. J. Nat. Prod. 1987, 50, 119–131; (f) Pettit, G. R.;
Rhodes, M. Anti-cancer Drug Des. 1998, 13, 183–191.
2. (a) McCoy, M. Chem. & Eng. News 2010, 11; (b) Wu, C.; Wei, J.; Tian, D.; Feng, Y.;
Miller, R. H.; Wang, Y. J. Med. Chem. 2008, 51, 6682–6688; (c) Stankoff, B.;
Wang, Y.; Bottlaender, M.; Aigrot, M. S.; Dolle, F. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 9304–9309; (d) Wu, C.; Tian, D.; Feng, Y.; Polak, P.; Wei, J. J.
Histochem. Cytochem. 2006, 54, 997–1004; (e) Lee, C.-W.; Zhuang, Z.-P.; Kung,
M.-P.; Plossl, K.; Skovronsky, D.; Gur, T.; Hou, C.; Trojanowski, J. Q.; Lee, V. M.-
Y.; Kung, H. F. J. Med. Chem. 2001, 44, 2270–2275.
3. (a) Xun, S.; Zhou, Q.; Li, H.; Ma, D.; Wang, L.; Jing, X.; Wang, F. J. Polym.Sci, Pt. A,
Polym. Chem. 2008, 46, 1566–1576; (b) Kim, C.; Choi, H.; Kim, S.; Baik, C.; Song,
K.; Kang, M. S.; Kang, S. O.; Ko, J. J. Org. Chem. 2008, 73, 7072–7079; (c) Kerr, R.
A.; Service, R. F. Science 2005, 309, 101; (d) Nazeeruddin, M. K.; De Angelis, F.;
Fantacci, S.; Selloni, A.; Viscardi, G.; Liska, P.; Ito, S.; Takeru, B.; Grätzel, M. J. Am.
Chem. Soc. 2005, 127, 16835–16847; (e) Grätzel, M. Nature 2001, 414, 338–344;
(f) McNulty, J.; Mcleod, D. Tetrahedron Lett. 2011, 52, 5467–5470.
4. Likhtenshtein, G. Stilbenes. Applications in Chemistry, Life Sciences and Materials
Science; Wiley-VCH: Weinheim, 2010.
5. Sale, S.; Tunstall, R. G.; Ruparelia, K. C.; Potter, G. A.; Steward, W. P.; Gescher, A.
J. Int. J. Cancer 2005, 115, 194–201.
6. (a) Paul, S.; Rimando, A. M.; Lee, H. J.; Ji, Y.; Reddy, B. S.; Suh, N. Cancer Prev. Res.
2009, 650–657; (b) Rimando, A. M.; Nagmani, R.; Feller, D. R.; Yokoyama, W. J.
Agric Food Chem. 2005, 53, 3403–3407; (c) See www.clinicaltrials.gov under
clinical trials, reference number NCT01267227.; (d) Rimando, A. M.; Cuendet,
M.; Desmarchelier, C.; Mehta, R. G.; Pezzuto, J. M.; Duke, S. O. J. Agric Food Chem.
2002, 50, 3453–3457.
9. (a) Orsini, F.; Verotta, L.; Lecchi, M.; Restano, R.; Curia, G.; Redaelli, E.; Wanke,
E. J. Nat. Prod. 2004, 67, 421–426; (b) Cushma, M.; Dhanapalan, N.; Gopal, D.;
Chakraborti, A. K.; Lin, C. M.; Hamel, E. J. Med. Chem. 1991, 34, 2579–2588; (c)
Simoni, D.; Roberti, M.; Invidiata, F. P.; Aiello, E.; Aiello, S.; Marchetti, P.;
Baruchello, R.; Eleopra, M.; Di Cristina, A.; Grimaudo, S.; Gebbia, N.; Crosta, L.;
Dieli, F.; Tolomeom, M. Bioorg. & Med. Chem. Lett. 2006, 16, 3245–3248; (d)
Moritani, I.; Fujiwara, Y. Tetrahedron Lett. 1967, 1119–1122; (e) Garcia-Rubia,
A.; Fernandez-Ibanez, M. A.; Arrayas, R. G.; Carretero, J. C. Chem. Eur. J. 2011, 17,
3567–3570; (f) Arvela, R. K.; Leadbeater, N. E. J. Org. Chem. 2005, 70, 1786–
1790; (g) Marti-Centelles, R.; Cejudo-Marin, R.; Falomir, E.; Murga, J.; Carda,
M.; Marco, J. A. Bioorg. & Med. Chem. 2013, 21, 3010–3015.
10. Mizuno, C. S.; Ma, G.; Khan, S.; Patny, A.; Avery, M. A.; Rimando, A. M. Bioorg. &
Med. Chem. 2008, 16, 3800–3808.
11. (a) McNulty, J.; Das, P. Eur. J. Org. Chem. 2009, 4031–4035; (b) Das, P.; McNulty,
J. Eur. J. Org. Chem. 2010, 3587–3591; (c) McNulty, J.; Das, P.; McLeod, D. Chem.
Eur. J. 2010, 16, 6756–6760; (d) McNulty, J.; Das, P. Tetrahedron Lett. 2009, 50,
5737–5740; (e) Das, P.; McNulty, J. Tetrahedron Lett. 2010, 51, 3197–3199; (f)
Das, P.; McLeod, D.; McNulty, J. Tetrahedron Lett. 2011, 52, 199–201; (g)
McLeod, D.; McNulty, J. Tetrahedron Lett. 2011, 52, 5467–5470; (h) McNulty, J.;
Calzavara, J. Tetrahedron Lett. 2011, 52, 5672–5675; (i) McNulty, J.; McLeod, D.
Chem. Eur. J. 2011, 17, 8794–8798; (j) McLeod, D.; McNulty, J. Eur. J. Org. Chem.
2012, 6127–6131.
12. Commercial samples of pterostilbene have recently been quoted at $801 for 1 g
(TCI) and $259 for 10 mg (Sigma–Aldrich).
13. (a) Byrne, P. A.; Gilheany, D. G. J. Am. Chem. Soc. 2012, 134, 9225–9239; (b)
Wang, Q.; El Khoury, M.; Schlosser, M. Chem. Eur. J. 2000, 6, 420–426; (c)
Robiette, R.; Richardson, J.; Aggarwal, V. K.; Harvey, J. N. J. Am. Chem. Soc. 2006,
128, 2394–2409; (d) McEwen, W. E.; James, A. B.; Knapczyk, J. W.; Kilingstad, V.
L.; Shiau, W.-L.; Shore, S.; Smith, J. H. J. Am. Chem. Soc. 1978, 100, 7304–7311.
14. Wolfrom, M. L.; Koos, E. W.; Bhat, H. B. J. Org. Chem. 1967, 32, 1058–1060.