Diastereoselective concise syntheses of the polyhydroxylated alkaloids DMDP and DAB

Article abstract of DOI:10.1016/j.tetlet.2013.11.068

A diastereoselective concise synthesis of the iminosugars DMDP and DAB is presented starting from l-xylose and affording the two alkaloids in good yields of 35% and 22% over seven and eight steps, respectively. The Petasis borono-Mannich reaction of 3,5-di-O-benzyl-l-xylofuranose with benzylamine and (E)-styrylboronic acid served as the nitrogen-introducing key step furnishing the new C-N bond in an entirely diastereoselective manner. A chemo- and regioselective O-mesylation followed by an intramolecular S_N2- cyclisation allowed the formation of the pyrrolidine ring. Ozonolysis of the styryl double bond and subsequent reduction to form the C-5 hydroxymethyl substituent followed by hydrogenolysis of the benzyl protecting groups concluded the DMDP synthesis. Furthermore, an unexpected fragmentation process during the ozonolysis reaction also gave access to the C-5 decarbinolated DMDP derivative DAB.

Full text of DOI:10.1016/j.tetlet.2013.11.068

Tetrahedron Letters 55 (2014) 475–478
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Diastereoselective concise syntheses of the polyhydroxylated
alkaloids DMDP and DAB
⇑
Marc E. Bouillon , Stephen G. Pyne
School of Chemistry, University of Wollongong, Wollongong, New South Wales 2522, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
A diastereoselective concise synthesis of the iminosugars DMDP and DAB is presented starting from
Received 16 September 2013
Revised 30 October 2013
Accepted 14 November 2013
Available online 23 November 2013
L
-xylose and affording the two alkaloids in good yields of 35% and 22% over seven and eight steps, respec-
tively. The Petasis borono–Mannich reaction of 3,5-di-O-benzyl- -xylofuranose with benzylamine and
L
(E)-styrylboronic acid served as the nitrogen-introducing key step furnishing the new C–N bond in an
entirely diastereoselective manner. A chemo- and regioselective O-mesylation followed by an intramo-
lecular SN2-cyclisation allowed the formation of the pyrrolidine ring. Ozonolysis of the styryl double bond
and subsequent reduction to form the C-5 hydroxymethyl substituent followed by hydrogenolysis of the
benzyl protecting groups concluded the DMDP synthesis. Furthermore, an unexpected fragmentation
process during the ozonolysis reaction also gave access to the C-5 decarbinolated DMDP derivative DAB.
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Polyhydroxylated alkaloids
Iminosugars
Petasis reaction
2,5-Dideoxy-2,5-imino-
D
-mannitol
-arabinitol
1,4-Dideoxy-1,4-imino-
D
The polyhydroxylated alkaloid 2R,5R-di(hydroxymethyl)-3R,
4R-dihydroxypyrrolidine (DMDP) (1), also known as 2,5-dideoxy-
It is well-known and a topic of numerous publications, that imi-
nosugars such as DMDP and DAB and their derivatives mimic car-
bohydrates and therefore act as inhibitors or moderators of various
glycosidases,8 enzymes that are responsible for catalysing the
hydrolysis of the glycosidic bonds in complex carbohydrates and
glycoconjugates.9 Due to their biological activity, it is not surpris-
ing that these alkaloids are considered to be powerful agents
against various diseases including diabetes, cancer, and HIV, by
means of altering the metabolism of carbohydrates, the properties
of glycoproteins and carbohydrate-mediated interactions of host
cells with infective viruses.10 These interesting pharmaceutical
properties make DMDP, DAB and related alkaloids interesting syn-
thetic targets, with the challenge to develop new general, practical
and concise routes to synthesise these compounds.
To date, the majority of the synthetic approaches to DAB, DMDP
and their stereoisomers use carbohydrates or carbohydrate deriva-
tives as starting materials for the straightforward installation of
the three or four neighbouring stereocentres, respectively. Intro-
duction of the amino function in the carbohydrate scaffold is then
accomplished by reductive amination or by nucleophilic substitu-
tion of one hydroxyl group with a nitrogen nucleophile such as
the azide anion or benzylamine. These syntheses often involve pro-
longed protection/deprotection steps resulting in low overall
yields.11
2,5-imino-D-mannitol, was first isolated by Welter and co-workers
from the leaves of the tropical legume Derris elliptica in 1976,1 and
has been found since then in many different natural sources,2
mostly plants, but also some microorganisms. The isolation of its
smaller congener, the decarbinolated DMDP derivative 1,4-dide-
oxy-1,4-imino-D-arabinitol (DAB or D-AB1) (2), was reported in
1985, almost simultaneously, from both the fruits of the African
legume Angylocalyx boutiqueanus by Nash et al.3 and the fronds of
the fern Arachniodes standishii by Hatanaka and co-workers.4 Like
DMDP, DAB was subsequently found in a diverse range of plant
species, occasionally accompanied by its 2-O- and 5-O-b-D-
glycosides, as well as its N-methyl and N-hydroxyethyl deriva-
tives.2 Although the first syntheses of both alkaloids date back to
the mid-1980s,5 new approaches are still regularly published so
that to date, more than 30 syntheses of DAB and DMDP,6 as well
as their enantiomers, have been reported. Furthermore, DAB and
DMDP can be considered as subunits of more complex alkaloids
including australine and casuarine, as well as members of the
broussonetine and hyacinthacine alkaloid families (Fig. 1), and
are therefore often synthetic by-products, intermediates or even
starting materials en route to the synthesis of these higher
alkaloids.7
In our continuing effort to synthesise polyhydroxylated alka-
loids,12 we now report our syntheses of DMDP and DAB starting
⇑
Corresponding author.
from the commercially available carbohydrate,
L-xylose. The key
E-mail
addresses:
meb984@uowmail.edu.au,
Marc.Bouillon@gmx.de
(M.E. Bouillon).
step in our synthetic approach was the three-component,
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.11.068

476
M. E. Bouillon, S. G. Pyne / Tetrahedron Letters 55 (2014) 475–478
appropriate starting material bearing the essential free OH-group
at the -position relative to the carbonyl functionality as the coor-
dination site for the boronic acid (Fig. 2).
This bis-benzylated -xylofuranose derivative is a literature
known compound14 that can easily be synthesised on a large scale
in three steps from -xylose (Scheme 1). Adopting and improving
the synthesis of Matsuda and Terashima,15
-xylose (4) was con-
a
L
L
L
verted into the desired Petasis precursor 3 in 80% yield over three
steps.
With sufficient amounts of 3 in hand we then turned to the key
step of our synthetic route, the Petasis reaction of 3 with benzyl-
amine and (E)-styrylboronic acid to establish the key intermediate,
amino alcohol 7. Initially, we applied the conditions originally pro-
posed by Petasis et al.16 for the synthesis of b-amino alcohols, how-
ever, furnishing compound 7 only in a moderate yield of 53%
accompanied by various unidentified by-products that could not
be separated by column chromatography. We therefore investi-
gated the effect of different solvents and solvent mixtures on the
above Petasis reaction with the goal to improve the yield and
purity of our b-amino alcohol. Our initial experiments employed
common solvents such as dichloromethane, methanol and etha-
nol–water mixtures. These modifications, however, did not result
in any significant improvements. In fact, when dichloromethane
was applied as the solvent the yield and purity of compound 7 be-
came worse and dropped to less than 25%. Then again, following a
recent development in borono–Mannich chemistry to apply fluori-
nated alcohols such as 1,1,1,3,3,3-hexafluoro-iso-propanol (HFIP)
and 2,2,2-trifluoroethanol (TFE) as solvent or co-solvent,17 we
found that TFE as the solvent in our Petasis reaction gave the best
result improving the yield to 76%. Moreover, the purity of the crude
product was considerably better than the purities of the products
obtained when MeOH or EtOH were used as solvents. The exact
reason why these fluorinated alcohols in some cases have such
an outstanding effect on the yield and/or the rate of Petasis reac-
tions is as yet unknown. It is generally assumed that the effect
originates in their enhanced ability to stabilise polarised or ionic
intermediates and transition states.18 The increased acidity of
these solvents may also play a role in the efficiency of the process,
possibly due to a catalytic effect on the formation of the interme-
diate iminium ion. The newly formed stereocentre of the amino
group at C-5 in compound 7 was expected to have the desired
(R)-configuration based on reports that the Petasis reaction usually
provides 1,2-anti-amino alcohols via an iminium boronate inter-
mediate (A) as shown in Scheme 1.16,19 This assumption was later
confirmed in the eventual synthesis of our primary target com-
pound, DMDP (1).
Following the Petasis reaction was the SN2/5-exo-tet-cyclisation
of the amino alcohol 7 to form the pyrrolidine derivative 8a. This
was achieved by treating a solution of 7 in CH2Cl2 with 1.075 equiv
of MsCl in the presence of an excess amount of Et3N at À10 °C to
0 °C, followed by gradually warming of the reaction mixture to re-
flux temperature. Gratifyingly, the secondary OH-group at C-6 was
chemo- and regioselectively mesylated furnishing the desired pyr-
rolidine 8a after internal SN2-cyclisation in 78% yield. Nevertheless,
a small amount of the mesylate 8b (6%) was also formed. Due to
the fact that the pyrrolidine alcohol 8a and its mesylate 8b have al-
most identical Rf values it was impossible to separate these com-
pounds by column chromatography at this stage of the synthesis.
Therefore, the mixture of 8a and 8b had to be used for the follow-
ing ozonolysis step. To avoid possible N-oxidation during the ozon-
olysis reaction, the free amine was treated initially with 2 M HCl
(2 M solution in Et2O) forming the hydrochloride salt, which was
then treated with ozone (ꢀ33% O3 in O2) at À78 °C.20 The yield of
10 after reduction of the intermediate ozonide with NaBH4, how-
ever, was only a moderate 56%. Nevertheless, when we applied
an alternative procedure, as reported by Behr and co-workers21
Figure 1. DMDP and DAB as well as their appearance as a subunit in more complex
alkaloids.
one-pot synthesis of the amino alcohol 7 via the Petasis borono–
Mannich reaction of the xylose derivative 3 with benzylamine
and (E)-styrylboronic acid. A distinct advantage of this Petasis ap-
proach is the highly diastereoselective outcome of the reaction; the
amino function is added to the carbohydrate scaffold with excel-
lent diastereomeric excess (>99%) in a single step. Mild reaction
conditions and the easy accessibility of all the components also
make this method especially convenient.
Synthesis of DMDP
In contrast to previous syntheses of polyhydroxylated alkaloids
in our group,13 we decided to use an already partly protected sugar
derivative as the aldehyde component in the Petasis borono–Man-
nich reaction, rather than an unprotected sugar, to avoid lengthy
protecting group manipulations further on in the synthesis. Retro-
synthetic analysis led to 3,5-di-O-benzyl-L-xylofuranose (3) as an
Figure 2. 3,5-Di-O-benzyl-L-xylofuranose (3) as a precursor for the Petasis reaction.

M. E. Bouillon, S. G. Pyne / Tetrahedron Letters 55 (2014) 475–478
477
Scheme 1. Synthesis of DMDP and DAB. Reagents and conditions: (a) 1. acetone, cat. H2SO4, anhydrous CuSO4, 25 °C, 28 h, then concd NH4OH, 2. 0.2% HCl, 25 °C, 3.5 h, then
NaHCO3; (b) NaH, BnCl, TBAI, THF, 0 °C to reflux; (c) 0.5 M H2SO4, 1,4-dioxane, reflux, 4 h; (d) benzylamine, (E)-styrylboronic acid, TFE, 25 °C, 24 h; (e) MsCl, Et3N, CH2Cl2,
À10 °C to 40 °C, 4 h; (f) NaH, BnBr, TBAI, THF, 0 °C to rt, 48 h; (g) 1. H2SO4 (1 M in Et2O), Et2O, 0 °C, 2. O3 (ꢀ33% in O2), CH2Cl2, À78 °C, then EtOH, NaBH4, À78 °C to rt; (h)
1. H2SO4 (1 M in Et2O), Et2O, 0 °C, 2. O3 (ꢀ33% in O2), CH2Cl2, À78 °C, 3. Et3N, À78 °C to À5 °C, 2 h, then EtOH, NaBH4, À5 °C to rt; (i) Pd black/Pearlman’s catalyst (1:1), H2
(1 atm), AcOH, rt, 18 h.
Surprisingly, besides the desired alcohol 12, we also isolated a sec-
ond compound by column chromatography from the crude mix-
ture of the ozonolysis reaction in 28% yield. This compound was
identified as the pyrrolidine 13 in which the C-5 hydroxymethyl
substituent had been completely cleaved.
A possible mechanism for the loss of this substituent is shown
in Scheme 2 and involves the fragmentation of the intermediate
secondary ozonide 14 and formation of an intermediate pyrrolidi-
nium ion 15 that is subsequently reduced by sodium borohydride
to the observed pyrrolidine derivative 13.
The repetition of the ozonolysis reaction of pyrrolidinol 8a
showed that the decarbinolation process also occurred, however,
affording pyrrolidine 9 only as a minor by-product in 8% yield.
The free hydroxy group apparently has a stabilizing effect during
the ozonolysis that is diminished with its benzylation. We con-
cluded that if our proposed mechanism is correct, it would be pos-
sible to promote this fragmentation process by the addition of a
base to the secondary ozonide 14 before its reductive work-up. If
Scheme 2. Proposed mechanism for the loss of the C-5 substituent via fragmen-
tation of the secondary ozonide 14.
successful then this small modification in the ozonolysis reaction
using the pyrrolidinium bisulfate instead of the hydrochloride salt
of compound 8a, we were able to improve the yield of pure diol 10
to 77%. The concluding hydrogenolytic cleavage of the benzyl pro-
tecting groups over Pd-enriched Pearlman’s catalyst finally affor-
ded DMDP (1) in 97% yield (Scheme 1).
could be used to divert the synthetic route to also prepare the imi-
nosugar DAB (2). As it turned out treatment of 14 with Et3N at
À78 °C prior to the reduction with NaBH4 indeed facilitated the
formation of pyrrolidine 13. The yield of this DAB precursor 13
could be increased from 28% to nearly 50%, whereas the yield of
pyrrolidinol 12 dropped down from 59% to less than half (25%).
The concluding hydrogenolysis of all the benzyl groups under the
same conditions employed in the DMDP synthesis eventually affor-
ded DAB (2) in 100% yield.
Synthesis of DAB
A small modification of our DMDP synthesis also gave access to
the iminosugar DAB (2). In an early effort to separate the surplus C-
4 mesylate 8b from alcohol 8a to obtain a pure precursor for the
ozonolysis, we benzylated compound 8a under standard condi-
tions obtaining the fully benzyl-protected pyrrolidine 11 in 93%
yield. The less polar benzyl ether 11 could now easily be separated
from the mesylate 8b by column chromatography and was subse-
quently submitted to the above ozonolysis-reduction sequence.
In conclusion, the synthesis of DMDP (1) has been accomplished
in seven consecutive steps from the commercially available carbo-
hydrate
L-xylose (4). The first steps were the transformation of
L-xylose into its known derivative, 3,5-di-O-benzyl-L-xylofuranose
(3), in 80% overall yield. The key step of our synthesis was the
one-pot, three-component Petasis borono–Mannich reaction of 3
with benzylamine and (E)-styrylboronic acid which provided the

478
M. E. Bouillon, S. G. Pyne / Tetrahedron Letters 55 (2014) 475–478
6. For a review on the syntheses and biology of DMDP, see: Wrodnigg, T. M.
Monatsh. Chem. 2002, 133, 393–426.
amino diol 7 as a single diastereomer in a good yield of 76%. The
subsequent N-alkylative SN2/5-exo-tet cyclisation with inversion
of the configuration at C-2 furnished the desired pyrrolidine ring
with the required all-(R)-stereochemistry in 78% yield. The pyrro-
lidinol 8a was then subjected to an ozonolysis reaction of the styryl
double bond which provided the pyrrolidine diol 10 in 77% yield.
The concluding hydrogenolytic cleavage of the benzyl protecting
groups finally afforded DMDP (1) in an overall yield of 35% from
7. (a) Trost, B. M.; Horne, D. B.; Woltering, M. J. Angew. Chem. 2003, 115, 6169–
6172; (b) Trost, B. M.; Horne, D. B.; Woltering, M. J. Chem. Eur. J. 2006, 12,
6607–6620; (c) Izquierdo, I.; Plaza, M. T.; Tamayo, J. A. Tetrahedron 2005, 61,
6527–6533; (d) Durugkar, K. A. Dissertation; National Chemical Laboratory
(NCL): Pune, India, 2009.
8. (a) Nash, R. J.; Watson, A. A.; Asano, N. Polyhydroxylated alkaloids that inhibit
glycosidases. In Alkaloids: Chemical and Biological Perspectives; Pelletier, S. W.,
Ed.; Elsevier Science Ltd.: Oxford, U.K, 1996; 11, 345–376; (b) Asano, N.; Nash,
R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetrahedron: Asymmetry 2000, 11, 1645–
1680.
L-xylose (4), or in 44% yield over four steps from the known 3,5-
di-O-benzyl derivative 3, respectively.
9. Elbein, A. D. FASEB J. 1991, 5, 3055–3063.
10. For biological and therapeutical evaluation on glycosidase inhibitory activities,
see: (a) Stütz, A. E. Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond; Wiley-VCH: Weinheim, 1999; (b) Watson, A. A.; Nash, R. J. In
Biodiversity: New Leads for the Pharmaceutical and Agrochemical Industries;
Wrigley, S. K., Hayes, M. A., Thomas, R., Chrystal, E. J. T., Nicholson, N., Eds.; The
Royal Society of Chemistry, 2000; pp 172–176; (c)Iminosugars. From Synthesis
to Therapeutic Applications; Compain, P., Martin, O. R., Eds.; Wiley & Sons:
Chichester, 2007; (d) Winchester, B.; Fleet, G. W. J. Glycobiology 1992, 2, 199–
210; (e) Asano, N.; Kato, A.; Watson, A. A. Mini-Rev. Med. Chem. 2001, 1, 145–
154; (f) Asano, N. Curr. Top. Med. Chem. 2003, 3, 471–484; (g) Asano, N.
Glycobiology 2003, 13, 93R–104R; (h) Asano, N. Glycosidase-Inhibiting
Alkaloids: Isolation, Structure, and Application. In Modern Alkaloids—
Structure, Isolation, Synthesis and Biology; Fattorusso, E., Taglialatela-Scafati,
O., Eds.; Wiley-VCH, 2008; pp 111–138; (i) Winchester, B. G. Tetrahedron:
Asymmetry 2009, 20, 645–651; (j) Nash, R. J.; Kato, A.; Yu, C.-Y.; Fleet, G. W. J.
Fut. Med. Chem. 2011, 3, 1513–1521.
Serendipitously, a small modification of the synthesis also gave
access to the known iminosugar DAB (2). Treatment of the inter-
mediate secondary ozonide 14 with Et3N during the ozonolysis
work-up of 11 promoted a fragmentation process which provided
the DAB precursor 13 in almost 50% yield. The subsequent final
hydrogenolysis of the remaining benzyl groups then afforded
DAB (2) in 22% yield over eight steps from
L-xylose (4), or 27% over
four steps from 3,5-di-O-benzyl- -xylofuranose 3, respectively.
L
Spectroscopic data for both synthesised iminosugars were in good
agreement with the published data of the natural products.
Acknowledgements
11. For a review on syntheses of DMDP and DAB starting from carbohydrates, see:
Synthesis of Naturally Occurring Nitrogen Heterocycles from Carbohydrates; El
Ashry, E. S. H., El Nemr, A., Eds.; Blackwell Publishing, 2005. pp 1–29.
12. (a) Lindsay, K. B.; Pyne, S. G. J. Org. Chem. 2002, 67, 7774–7780; (b) Lindsay, K.
B.; Pyne, S. G. Aust. J. Chem. 2004, 57, 669–672; (c) Tang, M.; Pyne, S. G. J. Org.
Chem. 2003, 68, 7818–7824; (d) Tang, M.; Pyne, S. G. Tetrahedron 2004, 60,
5759–5767; (e) Au, C. W. G.; Nash, R. J.; Pyne, S. G. Chem. Commun. 2010, 713–
715; (f) Jiangseubchatveera, N.; Bouillon, M. E.; Liawruangrath, B.;
Liawruangrath, S.; Nash, R. J.; Pyne, S. G. Org. Biomol. Chem. 2013, 11, 3826–
3833.
We are thankful to Dr. Wilford Lie for his assistance with NMR
measurements, and Dr. John Korth and Karin Maxwell for their
help with high-resolution mass spectrometry. Furthermore, we
thank the University of Wollongong and the Australian Research
Council for funding this project.
Supplementary data
13. (a) Davis, A. S.; Pyne, S. G.; Skelton, B. W.; White, A. H. J. Org. Chem. 2004, 69,
3139–3143; (b) Machan, T.; Davis, A. S.; Liawruangrath, B.; Pyne, S. G.
Tetrahedron 2008, 64, 2725–2732; (c) Davis, A. S.; Ritthiwigrom, T.; Pyne, S. G.
Tetrahedron 2008, 64, 4868–4879; (d) Ritthiwigrom, T.; Willis, A. C.; Pyne, S. G.
J. Org. Chem. 2010, 75, 815–824; (e) Moosophon, P.; Baird, M. C.;
Kanokmedhakul, S.; Pyne, S. G. Eur. J. Org. Chem. 2010, 3337–3344.
14. Toyao, A.; Tamura, O.; Takagi, H.; Ishibashi, H. Synlett 2003, 35–38.
15. (a) Matsuda, F.; Kawasaki, M.; Terashima, S. Tetrahedron Lett. 1985, 26, 4639–
4642; (b) Matsuda, F.; Terashima, S. Tetrahedron 1988, 44, 4721–4736.
16. Petasis, N. A.; Zavialov, I. A. J. Am. Chem. Soc. 1998, 120, 11798–11799.
17. (a) Nanda, K. K.; Trotter, B. W. Tetrahedron Lett. 2005, 46, 2025–2028; (b)
Kumagai, N.; Muncipinto, G.; Schreiber, S. L. Angew. Chem., Int. Ed. 2006, 45,
3635–3638; (c) Kaiser, P. F.; Churches, Q. I.; Hutton, C. A. Aust. J. Chem. 2007, 60,
799–810.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.tetlet.
2013.11.068. These data include MOL files and InChiKeys of the
most important compounds described in this article.
References and notes
1. Welter, A.; Jadot, J.; Dardenne, G.; Marlier, M.; Casimir, J. Phytochemistry 1976,
15, 747–749.
2. For a list of natural sources of DMDP, DAB and their derivatives, see Watson, A.
A.; Fleet, G. W. J.; Asano, N.; Molyneux, R. J.; Nash, R. J. Phytochemistry 2001, 56,
265–295.
18. Jourdan, H.; Gouhier, G.; van Hijfte, L.; Angibaud, P.; Piettre, S. R. Tetrahedron
Lett. 2005, 46, 8027–8031.
3. Nash, R. J.; Bell, E. A.; Williams, J. M. Phytochemistry 1985, 24, 1620–1622.
4. (a) Furukawa, J.; Okuda, S.; Saito, K.; Hatanaka, S.-I. Phytochemistry 1985, 24,
593–594; (b) Jones, D. W. C.; Nash, R. J.; Bell, E. A.; Williams, J. M. Tetrahedron
Lett. 1985, 26, 3125–3126.
19. Pyne, S. G.; Au, C. W. G.; Davis, A. S.; Morgan, I. R.; Ritthiwigrom, T.; Yazici, A.
Pure Appl. Chem. 2008, 80, 751–762.
20. Faugeroux, V.; Génisson, Y.; Andrieu-Abadie, N.; Colié, S.; Levade, T.; Baltas, M.
Org. Biomol. Chem. 2006, 4, 4437–4439.
5. (a) Fleet, G. W. J.; Smith, P. W. Tetrahedron Lett. 1985, 26, 1469–1472; (b) Fleet,
G. W. J.; Smith, P. W. Tetrahedron 1987, 43, 971–978; (c) Card, P. J.; Hitz, W. D. J.
Org. Chem. 1985, 50, 891–893; (d) Fleet, G. W. J.; Smith, P. W. Tetrahedron 1986,
42, 5685–5692.
21. Gautier-Lefebvre, I.; Behr, J.-B.; Guillerm, G.; Muzard, M. Eur. J. Med. Chem.
2005, 40, 1255–1261.