A general synthesis of sphinganines through multicomponent catalytic asymmetric aziridination

Article abstract of DOI:10.1002/ejoc.201301766

A catalytic asymmetric synthesis of all four stereoisomers of sphinganine is described starting from hexadecanal. Utilizing either the (R) or (S) enantiomer of a BOROX catalyst, a multicomponent reaction of this aldehyde with an amine and ethyl diazoacetate gives rise to enantiomeric aziridine-2- carboxylates. Access to all diastereomers of sphinganine is realized upon ring opening of the enantiopure aziridine-2-carboxylate at the C-3 position by direct S_N2 attack of an oxygen nucleophile, which occurs with inversion of configuration and by ring expansion of an N-acyl aziridine to an oxazolidinone and then hydrolysis. Overall, this process results in the formal ring opening of the aziridine with an oxygen nucleophile with retention of configuration. The synthesis of all four stereoisomers of sphinganine was achieved by multi-component asymmetric aziridination of hexadecanal. Complete stereocontrol is realized with the proper choice of the chirality of the BOROX catalyst and the introduction of an oxygen substituent at the 3-position of the aziridine with either retention or inversion. MEDAM = tetramethyldianisylmethyl. Copyright

Full text of DOI:10.1002/ejoc.201301766

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201301766
A General Synthesis of Sphinganines through Multicomponent Catalytic
Asymmetric Aziridination
Munmun Mukherjee,[a] Yubai Zhou,[a] Anil K. Gupta,[a] Yong Guan,[a] and
William D. Wulff*[a]
Keywords: Asymmetric synthesis / Aziridination / Multicomponent reactions / Sphingolipids
A catalytic asymmetric synthesis of all four stereoisomers of
sphinganine is described starting from hexadecanal. Utiliz-
ing either the (R) or (S) enantiomer of a BOROX catalyst, a
multicomponent reaction of this aldehyde with an amine and
ethyl diazoacetate gives rise to enantiomeric aziridine-2-
carboxylates. Access to all diastereomers of sphinganine is
realized upon ring opening of the enantiopure aziridine-2-
carboxylate at the C-3 position by direct SN2 attack of an
oxygen nucleophile, which occurs with inversion of configu-
ration and by ring expansion of an N-acyl aziridine to an
oxazolidinone and then hydrolysis. Overall, this process re-
sults in the formal ring opening of the aziridine with an oxy-
gen nucleophile with retention of configuration.
Introduction
Sphingolipids consist of several subclasses of compounds
that are involved in cell structure and regulation, and they
are based on a structurally related family of hundreds of
compounds known as sphingoid bases, often termed “long-
chain bases”.[1–3] The three major core units in sphingo-
lipids are sphinganine, sphingosine, and phytosphingosine
(Scheme 1). N-Acylated derivatives of the sphingoid bases
are members of the ceramide family. Glycosphingolipids are
ceramides with one or more sugar units attached to the hy-
droxy group at the 1-position. Members of this group in-
clude cerebrosides, which have a single glucose or galactose.
Gangliosides have three sugars, one of which must be a
sialic acid. The number of different head groups in the
glycosphingolipids is enormous, and a variety of “omics”
web sites are devoted to this class of compounds.[3] Phos-
phosphingolipids are ceramides in which a phosphate group
is bound as a phosphate ester to the hydroxy group at the
1-position. For sphingomyelins, this phosphate group is
either a phosphorylcholine or a phosphoroethanolamine.
Sphingolipids were recognized as far back as 1884, and
much has since been learned about their biochemistry.[4] Er-
rors in their metabolism has led to several inherited human
diseases including diabetes,[5] cancers,[6] infection by micro-
organisms,[7] Alzheimer’s disease,[8] heart disease, and an
Scheme 1. Sphingoid bases and the four stereoisomers of sphingan-
ine.
array of neurological syndromes.[9] They are involved in
nearly all aspects of cell regulation including proliferation,
differentiation, adhesion, neuronal repair, and signal trans-
duction.[10] The natural configuration of both sphinganine
and sphingosine is the d-erythro configuration; however, it
has been found that for both classes of compounds (and
their derivatives) the stereochemistry can play a large role
in their bioactivity. For example, the l-threo diastereomer
of sphinganine (safingol 4) is an antineoplastic and antipso-
riatic drug[11] and has been investigated for its ability to
inhibit protein kinase C.[12] As a consequence of this and
other differences in bioactivities, all four of the isomers of
sphinganines and sphingosines have been prepared and
their biological properties investigated.[1–3]
[a] Department of Chemistry, Michigan State University,
East Lansing, MI 48824, USA
E-mail: wulff@chemistry.msu.edu
http://www2.chemistry.msu.edu/faculty/wulff/myweb26
/index.htm
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201301766
1386
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 1386–1390

A General Synthesis of Sphinganines
involve a three-component asymmetric aziridination of
hexadecanal with a BOROX catalyst for which the (S)
enantiomer would give the (2R,3R) enantiomer of aziridine
8g and the (R) enantiomer would give the (2S,3S) enantio-
mer of 8g. Access to all four stereoisomers of sphinganine
would result from stereocontrolled opening of each enantio-
mer of aziridine 8g with both retention and inversion of
configuration.
Background
The history of the syntheses of the sphinganines has been
reviewed[13] and covers the period from the first synthesis in
1951 up to 2004, and many other syntheses have appeared
subsequent to this review.[14] The earlier syntheses of sphin-
ganines tended to be nonselective, which gave mixtures of
diastereomers that needed to be separated and enantiomers
that needed to be resolved. The most successful applications
with asymmetric catalysis involved the use of Sharpless
asymmetric dihydroxylation,[15] Sharpless asymmetric epox-
idation,[16] and Sharpless kinetic resolution of allylic
alcohols,[14n] although it has not been demonstrated if these
Results and Discussion
Three years ago, we introduced the first three-component
methods can be used for all four of the stereoisomers of the catalytic asymmetric aziridination reaction (MCAZ) by
sphinganines. Other catalytic asymmetric methods utilized bringing together an aldehyde, an amine, and a diazo com-
in the synthesis of a specific sphinganine stereoisomer are pound (Scheme 3).[18a] The catalyst for this reaction was a
the asymmetric hydrogenation of β-oxo esters[14m] and a chiral polyborate anion (BOROX catalyst) that was as-
proline-based Mannich reaction.[14p] As might be expected, sembled in situ from a vaulted biaryl ligand and B(OPh)3
the shortest synthesis (two steps from hexadecanal) involves by the amine substrate.[18b,19] This reaction was only re-
an asymmetric catalytic reaction in which the chiral center ported with VAPOL–BOROX catalyst 11 for the multicom-
at the nitrogen-substituted carbon atom is created in the ponent aziridination with MEDAM amine, and the results
stereogenic step. Shibasaki and co-workers reported that a from a screening of catalysts 12 and 13 derived from
1,1Ј-bi-2-naphthol (BINOL)–lanthanum catalyst will effect VANOL and tBu2VANOL with several aldehydes are pre-
the nitroaldol reaction (Henry reaction) between 2-nitroeth- sented in Table 1. The reactions in Table 1 were allowed to
anol and hexadecanal.[17a] Although this reaction gave good proceed for 24 h with the catalyst (5 mol-%) to ensure that
diastereoselectivity (91:9) and good asymmetric induction all of the substrates were fully converted, but many of the
(97%ee), the reaction was limited in that it could produce substrates were undoubtedly converted in far less time. For
only one diastereomer and prolonged reaction times were example, the reaction of the VAPOL catalyst (5 mol-%) in
required: 10 turnovers required 6.8 d. Testimony to the im- Entry d of Table 1 was complete in 1 h.[18a] In previous
practicality of this method is found in a recent study that studies on the aziridination of preformed imines, the
concluded that for the large-scale synthesis (100 g) of the VANOL catalyst gave essentially the same results as the
same sphinganine, it was best to run the reaction without a VAPOL catalyst, but the catalyst derived from tBu2VANOL
chiral catalyst and separate the diastereomers and enantio- 16 was found to give improved inductions with preformed
mers.[14b] This reaction seems to have come full circle, as benzhydryl imines derived from aliphatic and heteroarene-
this approach with this reaction was first reported in carbaldehydes.[20a] The results in Table 1 for the VAPOL
1951.[17b]
catalyst are taken from ref.[18a] The data in the table reveal
Our goal was to develop an asymmetric catalytic azirid- that all three ligands gave very high asymmetric inductions
ination that would provide direct access to all four of the for arenecarbaldehydes and that tBu2VANOL ligand 16
stereoisomers of sphinganine (Scheme 2). The plan was to gave higher yields than unsubstituted VANOL ligand 15.
Scheme 2. Proposed syntheses of all four sphinganines.
Eur. J. Org. Chem. 2014, 1386–1390
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1387

M. Mukherjee, Y. Zhou, A. K. Gupta, Y. Guan, W. D. Wulff
SHORT COMMUNICATION
The tBu2VANOL ligand gave higher asymmetric induction Table 1. Comparison of VAPOL, VANOL, and tBu2VANOL
BOROX catalysts.[a]
for aliphatic aldehydes than either VANOL or VAPOL.
Note that the reactions of primary aliphatic aldehyde 7g
were performed at –10 °C rather than at 25 °C (Table 1, En-
try g). It was previously observed for primary aliphatic al-
dehydes that the yields are much higher at lower tempera-
tures.[18a] The reaction of aldehyde 7g was also optimized
with regard to temperature, concentration, and catalyst
loading (see the Supporting Information). The higher asym-
metric inductions with the tBu2VANOL ligand could be
useful for nonsolid aziridines, the optical purity of which
cannot be enhanced by crystallization.[20b]
[a] Unless otherwise specified, all reactions were run at 0.5 m in
amine in toluene on a 0.5 mmol scale with 10 (1.2 equiv.) and 7
(1.05 equiv.) at 25 °C for 24 h and went to 100% completion with
the catalyst (5 mol-%). The catalyst was prepared by stirring a mix-
ture of the ligand (5 mol-%), commercial B(OPh)3 (15 mol-%), and
9 (100 mol-%) in toluene at 80 °C for 0.5 h. This was followed by
the addition of 4 Å molecular sieves and 7 and then 10. (S)-
VANOL and (R)-tBu2VANOL were used. [b] Data taken from
ref.[18a] [c] Yield of isolated cis-aziridine 8 after silica gel chromatog-
raphy. [d] Determined by HPLC on a chiral column. [e] Reaction
was performed at –10 °C at 0.2 m and with 10 (2.0 equiv.).
resulting N-H-aziridine was treated with Boc anhydride to
give N-Boc-aziridine 17 in 80% yield. We found that cis-
aziridine 17 could be ring-expanded to give only cis-oxazol-
idinone 18 in 90% yield, and the optimal Lewis acid was
Sc(OTf)3 (Table 2).
Table 2. Ring expansion of N-Boc-aziridine 17 with retention of
configuration.
Scheme 3. Three-component catalytic asymmetric aziridination.
Thus, with ready access to either enantiomer of aziridine
8g (Table 1), we viewed the biggest challenge to the realiza-
tion of the proposed synthesis outlined in Scheme 2 as the
opening of the aziridine ring with an oxygen nucleophile
with retention of configuration.[21] The ring expansion of
N-tert-butoxycarbonyl- (Boc-)aziridines to oxazolidinones
[a] Unless otherwise specified, all reactions were performed with 17
has been reported to occur with retention of configuration
and would represent a tidy solution to this problem.[22] We
approached this with some trepidation, because in our ex-
perience the ring expansion of an N-Boc-cis-aziridine with
a phenyl group in the 3-position is not selective and gives
mixtures of cis- and trans-oxazolidinones.[23] The ring ex-
pansion of trans-aziridines has been reported to proceed
(0.2 mmol) that was 0.2 m in CH2Cl2 with the Lewis acid (10 mol-
%). [b] Determined by analysis of the crude reaction mixture by 1H
NMR spectroscopy with Ph3CH as an internal standard. n.d. =
not determined. Yield of isolated product after silica gel
chromatography is given in parentheses. [c] Catalyst 50 mol-%.
[d] A complex mixture of unidentified products was observed.
The synthesis of all four stereoisomers of sphinganine is
with high stereoselectivity with Cu(OTf)2 (Tf = trifluoro- summarized in Scheme 4. Aziridine 8g is crystalline, and
methylsulfonyl) as the optimum Lewis acid.[22] Under the the optical purity could be enhanced to 98% by crystalli-
same conditions, the ring expansion of cis-aziridines was zation (80% recovery). Alternatively, the aziridination could
also reported to be stereoselective, although no experimen- be performed with (R)-tBu2VANOL–BOROX catalyst 13 to
tal procedures were given and no yields or data were pre- give ent-8g in a higher yield (97 vs. 80%) with 98%ee. A
sented for any cis product. In any event, the stage was set comparison of the aziridination with preformed imine from
to test the ring expansion by removing the MEDAM group aldehyde 7g was also performed (see the Supporting Infor-
in aziridine 8g with triflic acid in acetonitrile,[24] and the mation, Section C). The MCAZ protocol was also success-
1388
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 1386–1390

A General Synthesis of Sphinganines
Scheme 4. Synthesis of all four stereoisomers of sphinganine.
fully performed at the gram scale (5 mmol, 1.5 g) with no
effect on the yield (85% on 2.6 g of 8g) or the ee (96%) by
using the VAPOL–BOROX catalyst. The ring opening of
aziridines with oxygen nucleophiles with inversion of con-
figuration is known and typically requires an electron-with-
drawing group on the nitrogen atom.[25] Thus, Boc-aziridine
17 was treated with neat formic acid at room temperature,
which resulted in ring opening with formate and subsequent
O- to N-formyl migration. The N-formyl group was re-
moved with hydrochloric acid, and finally the ester was re-
duced to give l-threo-sphinganine isomer 4 (safingol) in
70% yield over the three steps. Instead of ring opening with
inversion of configuration, N-Boc-aziridine 17 can be ring-
expanded with retention of configuration with the aid of
scandium triflate; this procedure resulted in oxazolidione
18 in 90% yield. From this oxazolidinone, the d-erythro iso-
mer of sphinganine (1) was prepared in 70% yield by hy-
drolysis and reduction with lithium aluminum hydride
(LAH). Enantiomeric sphinganines d-threo-sphinganine 6
and l-erythro-sphinganine 5 were prepared in an analogous
fashion beginning with aldehyde 7g and the (R)-BOROX
catalyst.
Scheme 5. Alternative route by direct ring opening of unactivated
MEDAM aziridines.
Alternatively, the d-threo and l-threo isomers could be
accessed by direct opening of unactivated aziridine 8g and
ent-8g (Scheme 5). Treatment of 8g with trifluoroacetic acid
(TFA) in CH2Cl2 at room temperature for 48 h resulted in
ring opening with clean inversion of configuration. Hydrol-
ysis of the trifluoroacetyl ester gave amino alcohol 19 in
83% yield. Reduction gave MEDAM-protected amino diol
20 in 96% yield. The MEDAM group in 20 was removed
by hydrogenolysis in the presence of Boc anhydride to give
Conclusions
The syntheses described in this work represent the first
use of a single catalytic asymmetric reaction to prepare all
four stereoisomers of sphinganine. The key step is a multi-
component catalytic asymmetric aziridination of hexadeca-
nal with an amine and ethyl diazoacetate. Aziridine 8g and
its enantiomer can be obtained with the proper enantiomer
of the BOROX catalyst in 96–98%ee depending on the li-
gand used in the catalyst. The four stereoisomers of sphin-
ganine follow from the ring opening of the each aziridine
with an oxygen nucleophile under conditions that proceed
selectively with inversion or retention of configuration.
N-Boc-protected l-threo-sphinganine 21 in 87% yield.[26]
A
similar set of transformations allowed the conversion of az-
iridine ent-8g into N-Boc-protected d-threo-sphinganine
ent-21 in four steps in 59% overall yield. The advantage of
this route is that the large MEDAM group on the nitrogen
atom greatly reduces the polarity of these intermediates,
which facilitates purification.
Supporting Information (see footnote on the first page of this arti-
cle): Procedures and spectroscopic data for all new compounds.
Eur. J. Org. Chem. 2014, 1386–1390
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1389

M. Mukherjee, Y. Zhou, A. K. Gupta, Y. Guan, W. D. Wulff
SHORT COMMUNICATION
1655–1664; i) H. P. Kokatla, R. Sagar, Y. D. Vankar, Tetrahe-
dron Lett. 2008, 49, 4728–4730; j) N. Kim, S. H. Lee, S. K.
Namgoong, Bull. Korean Chem. Soc. 2009, 30, 695–699; k) C.
Seguin, F. Ferreira, C. Botuha, F. Chemla, A. Perez-Luna, J.
Org. Chem. 2009, 74, 6986–6992; l) A. C. Allepuz, R. Badorrey,
M. D. Diaz-de-Villegas, J. A. Galvez, Eur. J. Org. Chem. 2009,
6172–6178; m) R. Ait-Youcef, X. Moreau, C. Greck, J. Org.
Chem. 2010, 75, 5312–5315; n) P. Kumar, A. Dubey, V. G. Pur-
anik, Org. Biomol. Chem. 2010, 8, 5074–5086; o) M. E. Jung,
S. W. Yi, Tetrahedron Lett. 2012, 53, 4216–4220; p) M. V. Rao,
K. K. S. Reddy, B. V. Rao, Tetrahedron Lett. 2012, 53, 5993–
5995; q) Q. Li, H. Zhang, C. Li, P. Xu, Chin. J. Chem. 2013,
31, 149–153; r) E. D. D. Calder, A. M. Zaed, A. Sutherland, J.
Org. Chem. 2013, 78, 7223–7233.
Acknowledgments
This work was supported by the National Institute of General
Medical Sciences (GM 094478).
[1] J. Liao, J. Tao, G. Lin, D. Liu, Tetrahedron 2005, 61, 4715–
4733.
[2] H.-S. Byun, X. Lu, R. Bittman, Synthesis 2006, 2447–2474.
[3] S. T. Pruett, A. Bushnev, K. Hagedorn, M. Adiga, C. A.
Haynes, M. C. Sullards, D. C. Liotta, A. H. Merrill Jr., J. Lipid
Res. 2008, 49, 1621–1639.
[4] J. L. W. Thudichum, A Treatise on the Chemical Constitution of
Brain, Bailliere, Tindall, and Cox, London, 1884.
[5] S. A. Summers, D. H. Nelson, Diabetes 2005, 54, 591–602.
[6] D. E. Modrak, D. V. Gold, D. M. Goldenberg, Mol. Cancer
Ther. 2006, 5, 200–208.
[7] L. J. Heung, C. Luberto, M. Del Poeta, Infect. Immun. 2006,
74, 28.
[15] a) R. A. Fernandes, P. Kumar, Tetrahedron: Asymmetry 1999,
10, 4797–4802; b) R. A. Fernandes, P. Kumar, Eur. J. Org.
Chem. 2000, 3447–3449; c) L. He, H.-S. Byun, R. Bittman, J.
Org. Chem. 2000, 65, 7618–7626.
[16] W. E. Roush, M. A. Adam, J. Org. Chem. 1985, 50, 3752–3757.
[17] a) H. Sasai, T. Tokunaga, S. Watnabe, T. Suzuki, N. Itoh, M.
Shibasaki, J. Org. Chem. 1995, 60, 7388–7389; b) C. A. Grob,
E. F. Jenny, H. Utzinger, Helv. Chim. Acta 1951, 34, 2249–2254.
[18] a) A. K. Gupta, M. Mukherjee, W. D. Wulff, Org. Lett. 2011,
13, 5866; b) A. K. Gupta, M. Mukherjee, G. Hu, W. D. Wulff,
J. Org. Chem. 2012, 77, 7932–7944.
[8] Z. Zhou, H. Zhou, P. J. Walian, B. K. Jap, Biochemistry 2007,
46, 2553–2563.
[9] T. Kolter, K. Sandhoff, Biochim. Biophys. Acta Biomembr.
2006, 1758, 2057–2079.
[10] For an overview of sphingolipid functions, see: a) A. H. Mer-
rill Jr, C. C. Sweeley, Biochemistry of Lipids, Lipoproteins and
Membranes (Eds.: D. E. Vance, J. E. Vance), Elsevier, Amster-
dam, 1996, chapter 12, p. 309; b) Y. D. Vankar, R. R. Schmidt,
Chem. Soc. Rev. 2000, 29, 201.
[19] G. Hu, A. K. Gupta, R. H. Huang, M. Mukherjee, W. D.
Wulff, J. Am. Chem. Soc. 2010, 132, 14669–14675.
[20]
a) Y. Guan, Z. Ding, W. D. Wulff, Chem. Eur. J. 2013, 19,
15565–15571; b) Y. Zhang, A. Desai, Z. Lu, G. Hu, Z. Ding,
W. D. Wulff, Chem. Eur. J. 2008, 14, 3785–3803.
[11] USP Dictionary of USAN and International Drug Names, US
Pharmacopeia, Rockville, Maryland, 2000, p. 636.
[12] G. K. Schwartz, J. Jiang, D. Kelsen, A. P. Albino, JNCI J. Natl.
Cancer Inst. 1993, 85, 402.
[21] For reviews and aziridine ring-opening reactions, see: a) E. Hu,
Tetrahedron 2004, 60, 2701–2743; b) P. Lu, Tetrahedron 2010,
66, 2549–2560; c) S. Stankovic, M. D’hooghe, S. Catak, H.
Eum, M. Waroquier, V. Van Speybroeck, N. De Kimpe, H.-J.
Ha, Chem. Soc. Rev. 2012, 41, 643–665.
[22] C. Tomasini, A. Vecchione, Org. Lett. 1999, 1, 2153.
[23] Z. Lu, Y. Zhang, W. D. Wulff, J. Am. Chem. Soc. 2007, 129,
7185.
[24] M. Mukherjee, A. K. Gupta, Z. Lu, Y. Zhang, W. D. Wulff, J.
Org. Chem. 2010, 75, 5643–5660.
[25] J. Legters, J. G. H. Willems, L. Thijs, B. Zwanenburg, Recl.
Trav. Chim. Pays Bas 1992, 111, 59.
[13] A. R. Howell, R. C. So, S. K. Richardson, Tetrahedron 2004,
60, 11327–11347.
[14] a) D. Enders, A. Müller-Hüwen, Eur. J. Org. Chem. 2004,
1732–1739; b) L. H. Zhang, D. C. Oniciu, R. Mueller, B. H.
McCosar, E. Pop, ARKIVOC 2005, x, 285–291; c) R. M.
Ndonye, D. P. Izmirian, M. F. Dunn, K. O. A. Yu, S. A. Por-
celli, A. Khurana, M. Kronenberg, S. K. Richardson, A. R.
Howell, J. Org. Chem. 2005, 70, 10260–10270; d) Y. Cai, C.-C.
Ling, D. R. Bundle, Org. Biomol. Chem. 2006, 4, 1140–1146; e)
Y.-S. Tian, J.-E. Joo, V.-T. Pham, K.-Y. Lee, W.-H. Ham, Arch.
Pharmacal Res. 2007, 30, 167–171; f) J.-Y. Mun, A. Onorato,
F. C. Nichols, M. D. Morton, A. I. Saleh, M. Welzel, M. B.
Smith, Org. Biomol. Chem. 2007, 5, 3826–3833; g) A. Sharma,
S. Gamre, S. Chattopadhyay, Tetrahedron Lett. 2007, 48, 633–
634; h) E. Abraham, S. G. Davies, N. L. Millican, R. L. Nichol-
son, P. M. Roberts, A. D. Smith, Org. Biomol. Chem. 2008, 6,
[26] The removal of the Boc group from N-Boc-d-threo-sphingan-
ine has been reported in 85% yield: R. C. So, R. Ndonye, D. P.
Izmirian, S. K. Richardson, R. L. Guerrera, A. R. Howell, J.
Org. Chem. 2004, 69, 3233.
Received: November 25, 2013
Published Online: January 27, 2014
1390
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 1386–1390