Conformational studies on the Δ8(E, Z)-sphingolipid desaturase from helianthus annuus with chiral fluoropalmitic acids as mechanistic probes

Article abstract of DOI:10.1021/jo100542q

(Figure presented) The Δ⁸-sphingolipid desaturase from sunflower (Helianthus annuus) converts phytosphinganine into a mixture of Δ⁸-(E)- and -(Z)-phytosphingenines by removal of two syn-hydrogen atoms from anti-, and gauche-conformations of the substrate. With chiral (R)-6-, (S)-6-, (R)-7-, and (S)-7-fluoropalmitic acids the importance of conformations for the formation of (E)- and (Z)-isomers was investigated by using growing yeast cells expressing the desaturase from H. annuus. The fluoropalmitic acids were readily incorporated into a series of fluorinated phytosphinganines. The desaturation products of the major C₁₈- fluorophytosphinganine demonstrate that different conformations of the relevant aliphatic segment of the sphingolipids can be exposed to the active center of the enzyme resulting in (E)- or (Z)-fluoroalkenes. The presence of a fluorine atom at the position of the initial hydrogen removal C8-H_R led to a complete suppression of the desaturation reaction, while replacement of C8-H_S with fluorine generated a mixture of mainly (Z)- and trace amounts of (E)-fluoroolefine. Fluorine at C9 of the phytosphinganine precursors did not interfere with the initial C-H activation step and produced (E)- and (Z)-fluoroalkenes in the same ratio as observed for the nonfluorinated precursors. Hydroxylated byproducts of the desaturation process were not observed. These results strongly support the importance of conformations of the transition states during desaturation as the relevant criterion for the relative ratio of (E)- and (Z)-alkenes.

Full text of DOI:10.1021/jo100542q

pubs.acs.org/joc
Conformational Studies on the Δ8(E,Z)-Sphingolipid Desaturase
from Helianthus annuus with Chiral Fluoropalmitic Acids As
Mechanistic Probes
Andreas Habel,† Petra Sperling,‡,§ Stefan Bartram,† Ernst Heinz,‡ and Wilhelm Boland*,†
†
€
Department of Bioorganic Chemistry, Max Planck Institute for Chemical Ecology, Hans-Knoll-Strasse 8,
D-07745 Jena, Germany, and ‡Biocenter Klein Flottbek and Botanical Garden, University of Hamburg,
Ohnhorststrasse 18, D-22609 Hamburg, Germany
boland@ice.mpg.de
Received March 22, 2010
The Δ8-sphingolipid desaturase from sunflower (Helianthus annuus) converts phytosphinganine into
a mixture of Δ8-(E)- and -(Z)-phytosphingenines by removal of two syn-hydrogen atoms from anti-,
and gauche-conformations of the substrate. With chiral (R)-6-, (S)-6-, (R)-7-, and (S)-7-fluoropal-
mitic acids the importance of conformations for the formation of (E)- and (Z)-isomers was
investigated by using growing yeast cells expressing the desaturase from H. annuus. The fluoropal-
mitic acids were readily incorporated into a series of fluorinated phytosphinganines. The desatura-
tion products of the major C18-fluorophytosphinganine demonstrate that different conformations of
the relevant aliphatic segment of the sphingolipids can be exposed to the active center of the enzyme
resulting in (E)- or (Z)-fluoroalkenes. The presence of a fluorine atom at the position of the initial
hydrogen removal C8-HR led to a complete suppression of the desaturation reaction, while
replacement of C8-HS with fluorine generated a mixture of mainly (Z)- and trace amounts of (E)-
fluoroolefine. Fluorine at C9 of the phytosphinganine precursors did not interfere with the initial
C-H activation step and produced (E)- and (Z)-fluoroalkenes in the same ratio as observed for the
nonfluorinated precursors. Hydroxylated byproducts of the desaturation process were not observed.
These results strongly support the importance of conformations of the transition states during
desaturation as the relevant criterion for the relative ratio of (E)- and (Z)-alkenes.
Introduction
(LCB), an amide-linked fatty acyl chain of varying chain length,
and a polar head.4 The predominant LCBs are C18 amino
alcohols such as sphinganine (3) and 4-hydroxysphinganine
Sphingolipids are characteristic lipid components of all euka-
ryotic cells in which they have structural roles as membrane com-
ponents and signaling functions in many regulatory circuits.1-3
The sphingolipids are generally composed of a long-chain base
(2) Zheng, W.; Kollmeyer, J.; Symolon, H.; Momin, A.; Munter, E.;
Wang, E.; Kelly, S.; Allegood, J. C.; Liu, Y.; Peng, Q.; Ramaraju, H.;
Sullards, M. C.; Cabot, M.; Merrill, J. A. H. Biochim. Biophys. Acta,
Biomembr. 2006, 1758, 1864–1884.
(3) Spiegel, S.; Milstien, S. Nat. Rev. Mol. Cell Biol. 2003, 4, 397–407.
(4) Pruett, S. T.; Bushnev, A.; Hagedorn, K.; Adiga, M.; Haynes, C. A.;
Sullards, M. C.; Liotta, D. C.; Merrill, A. H. J. Lipid Res. 2008, 49, 1621–
1639.
*To whom correspondence should be addressed. Phone: (þ49) 3641
571200. Fax: (þ49) 3641 571202.
§ Deceased 2008.
(1) Sphingolipid Biology, 1st ed.; Hirabayashi, Y., Igarashi, Y., Merrill,
A. H. J., Eds.; Springer: Tokyo, Japan, 2006.
DOI: 10.1021/jo100542q
r
Published on Web 06/24/2010
J. Org. Chem. 2010, 75, 4975–4982 4975
2010 American Chemical Society

Habel et al.
JOCArticle
SCHEME 1. Biosynthesis of Unsaturated (8E,Z)-Phytoceramides
found in many organisms. Many of them, and in particular
plants and fungi, can introduce additional double bonds into
their LCB resulting in additional variation, with the unsatu-
rated 4-hydroxy-8-sphingenine prevailing in plants.5 The
first and committed step in the de novo sphingolipid synthe-
sis is catalyzed by a serine palmitoyltransferase (EC
2.3.1.50), which condenses serine (1) and palmitoyl-CoA
(2) to 3-ketosphinganine. The 3-ketosphinganine is then
reduced to the LCB sphinganine (3). Following acyl-transfer
(ceramide synthase) and hydroxylation at C4, a phytocer-
amide (N-acyl 4-hydroxysphinganine) 4 is obtained.
SCHEME 2. Proposed Mechanism for a Nonstereospecific
Sphingolipid Δ8-Desaturase from Plants
In plants, the LCBs can be additionally desaturated
between C8 and C9 yielding a mixture of (E)- and (Z)-
isomers of Δ8-unsaturated LCBs 5 and 6, namely (8E/Z)-4-
hydroxy-8-sphingenines (phytosphingenines).6 Heterolo-
gous expression of plant-derived cDNAs encoding the Δ8-
desaturases in yeast cells (S. cerevisiae) resulted in significant
proportions of both (E)- and (Z)-isomers of 4-hydroxy-8-
sphingenines (Scheme 1). Since these compounds were not
produced by the wild-type yeasts, their formation in the
transgenic cells could be clearly attributed to the activity of
the expressed nonstereospecific sphingolipid Δ8-desaturase.
Moreover, depending on the origin of the Δ8-desaturase,
different but characteristic (E/Z)-ratios ranging from pure
(E)-isomers (from fungi7) via 7:1 and 1:1 to predominating
(Z)-isomers (from plants8-14) were observed with the same
expression system.5 The deduced amino acid sequences of the
proteins show high similarities to membrane-bound acyl
lipid desaturases bearing a highly conserved multihistidine
di-iron coordination site.
The desaturases are assumed to convert the saturated sub-
strate into an olefin by a two-step radical mechanism: Initial
C-H activation by a bis(μ-oxo)-diiron complex (Scheme 2)
removes the first hydrogen atom from the carbonyl-proximal
carbon. This step is rate limiting and displays a high intrinsic
primary kinetic isotope effect (KIE, kH/D ≈ 3-8), whereas the
second hydrogen atom is lost without a significant KIE.
Desaturation proceeds suprafacially and removes two syn-
oriented vicinal hydrogen atoms from the substrate.15-17 The
simultaneous formation of (E,Z)-olefins by a single enzyme was
rationalized by removal of two vicinal syn-oriented hydrogen
atoms from the substrate in different conformations. In sun-
flower Δ8-sphingolipid desaturases a low but distinct kinetic iso-
topic effect suggested a preferential attack of the reactive di-iron
center onto the C8-HR (kH/D=1.91) of the substrate to yield
(5) Sperling, P.; Heinz, E. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids
2003, 1632, 1–15.
(6) Lynch, D. V.; Dunn, T. M. New Phytol. 2004, 161, 677–702.
(7) Takakuwa, N.; Kinoshita, M.; Oda, Y.; Ohnishi, M. Curr. Microbiol.
2002, 45, 459–461.
(8) Sun, D.; Froman, B. E.; Orth, R. G.; MacIsaac, S. A.; Larosa, T.;
Dong, F.; Valentin, H. E. J. Chromatogr. Sci. 2009, 47, 895–901.
(9) Ryan, P. R.; Liu, Q.; Sperling, P.; Dong, B.; Franke, S.; Delhaize, E.
Plant Physiol. 2007, 144, 1968–1977.
(10) Garcia-Maroto, F.; Garrido-Cardenas, J. A.; Michaelson, L. V.;
Napier, J. A.; Alonso, D. L. Plant Mol. Biol. 2007, 64, 241–250.
(11) Sperling, P.; Libisch, B.; Zahringer, U.; Napier, J. A.; Heinz, E. Arch.
Biochem. Biophys. 2001, 388, 293–298.
(12) Michaelson, L. V.; Longman, A. J.; Sayanova, O.; Stobart, A. K.;
Napier, J. A. Biochem. Soc. Trans. 2002, 30, 1073–1075.
(13) Sperling, P.; Zahringer, U.; Heinz, E. J. Biol. Chem. 1998, 273,
28590–28596.
(14) Sperling, P.; Blume, A.; Zahringer, U.; Heinz, E. Biochem. Soc.
Trans. 2000, 28, 638–641.
the (E)-isomer after loss of the neighboring C9-HS (kH/D
=
1.16) via an energetically favored anti-orientation. (Z)-Isomers
are produced via two different gauche-orientations after initial
attack onto either the C9-HR (kH/D =3.97) or the C8-HR
(kH/D=2.07) followed by loss of the vicinal hydrogen atom18
(Scheme 2).
(15) Shanklin, J.; Cahoon, E. B. Annu. Rev. Plant Physiol. Plant Mol.
Biol. 1998, 49, 611–641.
(16) Behrouzian, B.; Buist, P. H. Prostaglandins, Leukotrienes Essent.
Fatty Acids 2003, 68, 107–112.
(17) Buist, P. H. Nat. Prod. Rep. 2004, 21, 249–262.
(18) Beckmann, C.; Rattke, J.; Oldham, N. J.; Sperling, P.; Heinz, E.;
Boland, W. Angew. Chem., Int. Ed. 2002, 41, 2298–2300.
4976 J. Org. Chem. Vol. 75, No. 15, 2010

Habel et al.
JOCArticle
SCHEME 3. Mechanistic Alternatives for the Desaturation of Chiral 8- or 9-Fluorophytosphinganines by the Δ8-Desaturase from
Helianthus annuusa
aOnly one gauche-conformation is shown.
According to this model, the final ratio of the (E/Z)-olefins
will be determined by the ratio of the differently populated
conformations of the substrate with either anti- or gauche-
orientation around the C8-C9 axis. Since the enzyme
accepts both substrate conformations, the controlled mani-
pulation of steric or electronic effects within the substrate can
be expected to alter the products from various conforma-
tions in a predictable fashion. In particular, fluorinated
substrates are promising tools for this type of modification
since the exchange of a hydrogen atom by fluorine does not
greatly affect the size or shape of the substrate but strongly
modifies the electronic properties of the adjacent bonds.
Certain conformations of the chiral fluorosubstrates will
expose the inert fluorine atom to the active site of the
desaturase, which will not yield a product. Only after a
conformational change can desaturation products be ex-
pected and, hence, the fluorinated substrates can be used as
valuable mechanistic probes. With respect to the established
major reaction pathway to (E)-olefins implying an initial
attack onto the C8-HR of the anti-oriented substrate fol-
lowed by syn-elimination of the C9-HS (Scheme 2), different
stereochemical outcomes can be envisaged for the transfor-
mation of chiral 8- or 9-fluorophytosphinganines by the
sunflower desaturase (Scheme 3).
1. (R)-8-Fluorophytosphinganine: The substrate confor-
mation en route to the major (E)-isomer exposes the fluorine
atom to the active center of the desaturase. The initial attack
onto the C8-FR is, thus, prohibited and no olefin should be
formed. Instead, the formation of a fluorohydrin19,20 should be
possible from a gauche-conformation with initial attack at C9.
2. (S)-8-Fluorophytosphinganine: The substrate can be
directly converted into a (Z)-fluoroalkene via the preferred
anti-orientation of the substituents R1 and R2. The formation
of an (E)-fluoroalkene would require a gauche-conformation.
Note that due to priorities in nomenclature assignment the (Z)-
fluoroalkene has the same architecture of the carbon skeleton
as the nonfluorinated (E)-alkene.
3. (R)-9-Fluorophytosphinganine: Expected products are
the (Z)-fluoroalkene from the anti-conformer and a fluoro-
hydrin from the two gauche-conformers.
4. (S)-9-Fluorophytosphinganine: The anti-conformer
is expected to generate a fluorohydrin. The two gauche-
conformations should produce an (E)-fluoroalkene as
illustrated in Scheme 3.
(19) Buist, P. H.; Behrouzian, B.; Alexopoulos, K. A.; Dawson, B.; Black,
B. Chem. Commun. 1996, 2671–2672.
(20) Buist, P. H.; Alexopoulos, K. A.; Behrouzian, B.; Dawson, B.; Black,
B. J. Chem. Soc., Perkin Trans. 1 1997, 2617–2624.
J. Org. Chem. Vol. 75, No. 15, 2010 4977

Habel et al.
JOCArticle
SCHEME 4. Synthesis of Chiral Fluoropalmitic Acids
Helianthus annuus (pHaDES8) were grown at 30 °C for 48 h
in complete minimal-dropout-uracil medium.14 For desaturase
expression the cultures were induced with galactose and incu-
bated in the presence of cerulenin27 and a chiral fluoropalmitic
acid for 5 d at 28 °C. Cerulenin was added to avoid channeling
of the fluoropalmitic acid into the general fatty acid biosynthesis
of the yeast. After optimization of the growth conditions with
respect to temperature, nutrient composition, inhibitor and sub-
strate concentrations, as well as time (5 d at 28 °C), fluorinated
4-hydroxysphinganines and 4-hydroxysphing-8-enines (>95%
labeled) could be isolated from the cultures. The optimization
was essential, since the fluoropalmitic acids, unlike palmitic
acid, reduced the growth of the yeast cultures which ceased
completely at substrate concentrations >90 μg mL-1
.
Analysis of Fluorinated Sphingolipids. Established workup
procedures of complex sphingolipids from plant tissue make use
of a strong base such as Ba(OH)2 to liberate the LCBs from the
ceramides.28,29 Since these procedures gave only low yields for
the fluorinated LCBs, an alternative approach without cleavage
of the amide bond of the phytoceramide moiety was developed.
The presence of a vicinal diol in the phytoceramide moieties
allows a selective and highly efficient cleavage of the phytocer-
amide 17 in simple as well as in more complex sphingolipids
between the hydroxy groups at C3 and C4 by using Pb(OAc)4.30
This procedure gave high yields of fluoroaldehydes which after
prepurification by passage over Florisil were converted into the
corresponding N,N-dimethylhydrazones (DMZ). Unlike the
free fluoroaldehydes the N,N-dimethylhydrazones show char-
acteristic ion fragments that allow a rapid identification and a
detailed analysis of the metabolites by GLC-MS (Figure 1). All
hydrazones of the saturated fluoroaldehydes display an intense
molecular ion (Figure 1A) and a major fragment at m/z 85
resulting from allylic cleavage next to the NdC double bond and
a fragment at m/z 86 corresponding to the McLafferty rearran-
gement of the hydrazone moiety.31,32 The other minor fragment
ions are generated mainly by repeated alkyl cleavage. The typical
EI-induced elimination of HF was not observed with the N,N-
dimethylhydrazones representing an additional major advan-
tage of this procedure. The molecular ions of the derivatized
fluoroolefins (Figure 1B) are less abundant, but still allow
identification of the compounds.
The stereochemistry of the unsaturated 5- and 6-fluoro-
aldehydes was determined by GLC with synthetic references.
The stereo- and regioisomers display different retention
times and allow the unequivocal determination of the con-
figuration of the enzymatically introduced double bonds
(see the Supporting Information).
Administration of Chiral Fluoropalmitic Acids and Analysis
of Fluorinated Metabolites. Despite their effects on culture
growth the fluorinated palmitic acids were taken up by the
yeasts, and under optimized conditions acceptable yields of
fluorinated sphingolipids were obtained. The chiral fluoro-
palmitic acids 15 and 16 were clearly incorporated into the
LCB backbone as indicated by analysis of the corresponding
N,N-dimethylhydrazones obtained after treatment of the
In summary, depending on the configuration of the 8- or
9-fluorosphinganine either (E)- or (Z)-fluoroolefins should be
generated in a predictable fashion proving the concept that
different substrate conformations lead to the nonstereospecific
generation of (E,Z)-olefins by a single enzyme. Here we report
on the stereochemical analysis of the Δ8-sphingolipid desaturase
from sunflower (Helianthus annuus) using chiral 8- and 9-fluoro-
sphinganines as substrates. The enantiospecific synthesis of
chirally fluorinated palmitic acids as precursors for the fluori-
nated phytoceramides and their conversion to fluorinated
sphingolipids as well as the mechanistic consequences of the
F/H-exchange on the desaturation are described.
Results
Synthesis of Chiral Fluoropalmitic Acids. Fluorinated pal-
mitic acids can be used as distinguished mechanistic probes
for the desaturases since the transgenic yeast generates the re-
quired 8- and 9-fluorophytosphinganines from these substrates
(Scheme 1). The synthesis of the chiral fluoropalmitic acids was
achieved along the protocol of Habel et al.21 as outlined in
Scheme 4. The readily available chiral terminal epoxides 9 (10)
(>99% ee)22,23 were alkylated at C1 with terminally unsatu-
rated Grignard reagents in the presence of Cu(I). The resulting
secondary alcohols 11 (12) were obtained in high yield and high
ee (>99%)21 as confirmed by Mosher ester derivatization (see
the Supporting Information).
After recrystallization from pentane at -20 °C, the alcohols
were converted with diethylaminosulfur trifluoride (DAST) at
-78 °C into the corresponding fluoroalkenes 13 (14) with
inversion of configuration.24,25 The olefins were converted
to the chiral fluoropalmitic acids (R)-15, (S)-15, (R)-16, and
(S)-16 by oxidative degradation with RuCl3 and NaIO4.26
Administration of Fluoropalmitic Acids to Transgenic Yeast
Cells. Cultures of the yeast strain S. cerevisiae INVSc1
(Invitrogen) containing the Δ8-sphingolipid desaturase from
(21) Habel, A.; Boland, W. Org. Biomol. Chem. 2008, 6, 1601–1604.
(22) Schaus, S. E.; Brandes, B. D.; Larrow, J. F.; Tokunaga, M.; Hansen,
K. B.; Gould, A. E.; Furrow, M. E.; Jacobsen, E. N. J. Am. Chem. Soc. 2002,
124, 1307–1315.
(23) Nielsen, L. P. C.; Stevenson, C. P.; Blackmond, D. G.; Jacobsen,
E. N. J. Am. Chem. Soc. 2004, 126, 1360–1362.
(24) Singh, R. P.; Shreeve, J. n. M. Synthesis 2002, 2002, 2561–2578.
(25) Leroy, J.; Hebert, E.; Wakselman, C. J. Org. Chem. 1979, 44, 3406–
3408.
(26) Carlsen, P. H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org.
Chem. 1981, 46, 3936–3938.
(27) Nomura, S.; Omura, S.; Hata, T.; Horiuchi, T. J. Biochem. 1972, 71,
783–796.
(28) Christie, W. W. In Lipid Analysis, 3rd ed.; The Oily Press:
Bridgwater, UK, 2003; pp 181-202.
(29) Morrison, W. R.; Hay, J. D. Biochim. Biophys. Acta 1970, 202, 460–467.
(30) Karlsson, K. A.;Martenss., EBiochim. Biophys. Acta 1968, 152, 230–233.
(31) Spiteller, G.; Kern, W.; Spiteller, P. J. Chromatogr. A 1999, 843, 29–98.
(32) Goldsmith, D.; Djerassi, C. J. Org. Chem. 1966, 31, 3661–3666.
4978 J. Org. Chem. Vol. 75, No. 15, 2010

Habel et al.
JOCArticle
crude lipid extracts with Pb(OAc)4 and N,N-dimethylhydrazine
(see Experimental Section). The main product was the expected
hydrazone 18 of 5-fluoropentadecanal resulting from diol-
cleavage of 8-fluorophytosphinganine 17 (Figure 1). In addition
we found minor amounts of the N,N-dimethylhydrazones of
saturated fluorinated C16-toC18-aldehydes derived from C19-to
C21-fluorophytosphinganines, probably resulting from a com-
bination of chain elongation by acetate units and R-oxidation of
the fatty acid precursor. Moreover, fluorinated N,N-dimethyl-
hydrazones from long-chain C25- and C26-aldehydes (derived
from C28- and C29-fluorophytosphinganines) were detected.
Nonfluorinated hydrazones derived from C18- and C20-phyto-
sphinganines were only found in trace amounts due to the
inhibition of fatty acid elongation by cerulenin. All N,N-dime-
thylhydrazones were identified by references and their mole-
cular ions with high-resolution mass spectroscopy (HRMS). The
stereochemistry of the unsaturated fluoroaldehydes, resulting
from desaturation of the fluorinated phytosphinganines by the
H. annuus enzyme, was determined by GLC comparison with
synthetic references (see the Supporting Information). Thus,
administration of (R)-6-fluoropalmitic acid ((R)-15) generated
(R)-8-fluorophytosphinganine and administration of (S)-6-
fluoropalmitic acid ((S)-15) resulted in (S)-8-fluorophytosphin-
ganine (27), which was predominantly transformed to (Z)-8-
fluoro-8-phytosphingenine (30) along with only trace amounts
of the (E)-8-fluoro-8-phytosphingenine (31). (R)-8-Fluoro-
phytosphinganine (26) did not yield any desaturation product.
According to Figure 2 and Table 1 the administration of (S)-7-
fluoropalmitic acid ((S)-16) resulted predominantly in the for-
mation of (E)-9-fluoro-8-phytosphingenine (33) as the major
unsaturated product deduced from the signal of the N,N-di-
methylhydrazone-derivative 23 of (E)-6-fluoro-5-pentadecenal
(Figure 2). A summary of the fluorinated cleavage products and
their fluorophytosphinganine precursors resulting from the in-
corporation of the chiral fluoropalmitic acids is given in Table 1.
According to Scheme 3 the formation of erythro- or threo-
fluorohydrins is in principle a feasible reaction pathway. To
detect these compounds, the cleavage products were addition-
ally derivatized with MSTFA and analyzed by GLC-MS.
However, no evidence for the generation of fluorohydrins was
found. Also the direct analysis of the crude reaction mixture of
the fluorophytosphinganines by LC-MS analysis13 gave no hint
for the formation of fluorohydrins.
FIGURE 1. Mass spectra of (un)saturated N,N-dimethylhydra-
zones 18 and 20. The fluorinated metabolites were analyzed by
GLC-MS after oxidative cleavage of the phytosphinganine to
fluoroaldehydes and derivatization with N,N-dimethylhydrazine.
Discussion
Expression of the gene, coding for the phytosphinganine
Δ8-desaturase from sunflower in the yeast S. cerevisiae,
FIGURE 2. Chromatographic separation of N,N-dimethylhydrazones of fluoroaldehydes. Separation of a mixture of synthetic references
N,N-dimethylhydrazones: (A) (S)-5-fluoropentadecanal ((S)-18), (Z)-5-fluoropentadec-5-enal (20), and (E)-5-fluoropentadec-5-enal (21). The 5/6
fluoro regioisomers of the N,N-dimethylhydrazones of E- and Z-fluoropentadecanals respectively show identical retention times. (B) Representative
GLC-MS chromatogram of fluorinated metabolites obtained after administration and metabolism of (S)-7-fluoropalmitic acid ((S)-16). (Ion traces of
m/z 86 are shown, complete analysis for all incubations is provided in the Supporting Information, Figure S1).
J. Org. Chem. Vol. 75, No. 15, 2010 4979

Habel et al.
JOCArticle
TABLE 1. Unsaturated Fluoroaldehydes from Fluorosphingenine Metabolites
fluorophyto-
sphinganine
(E)-5-fluoro-5-
pentadecenal (%)
(Z)-5-fluoro-5-
pentadecenal (%)
(E)-6-fluoro-5-
pentadecenal (%)
(Z)-6-fluoro-5-
pentadecenal (%)
substrate
(R)-15
(S)-15
(R)-16
(S)-16
26
27
28
29
5
95
100
100
TABLE 2. Product Spectrum of the Nonstereospecific Δ8(E,Z)-Desaturase from H. annuusa
aProduct spectrum observed after application of chirally fluorinated fatty acids. Gray background highlights the observed products (R = acyl).
resulted in the parallel formation of (E)- and (Z)-8-phyto-
sphingenine in a ratio of 88:12. Different population of anti- and
gauche-conformations at the active site of the enzyme in combi-
nation with a syn-elimination of the relevant hydrogen atoms
were assumed to control this stereochemical outcome.18 To
evaluate the importance of conformations in the transiion state
of the desaturation, we administered chiral fluoropalmitic acids
to the transgenic yeast. As expected, the palmitic acids were con-
verted to chiral fluorophytosphinganines after suppression of
the yeast fatty acid elongation by cerulenin. These chiral fluoro-
phytosphinganines served as precursors for the Δ8-desaturase
and yielded (E,Z)-fluorophytosphingenines with predictable
configurations as outlined in Scheme 3. The stereochemistry
of the desaturated products was analyzed after cleavage of
the vicinal diol of the phytoceramide moieties by Pb(OAc)4
followed by derivatization of the resulting aliphatic fragments
with N,N-dimethylhydrazine to the corresponding N,N-di-
methylhydrazones. This procedure allowed both sensitive detec-
tion of the relevant fragments and determination of their
configuration by GLC-MS with authentic references.
Scheme 3. Although fluorohydrins have been reported pre-
viously as trace products of a desaturase reaction,33 the
current study gave no evidence for their formation.
If a fluorine atom replaces the C8-HR in 26 no product is
observed since none of the known desaturases is able to
remove a fluorine radical. Also the initial attack onto the
C9-HR of 26 in combination with a gauche-conformation
does not lead to an alkene since fluorine is a bad leaving
group. The alternative formation of a fluorohydrin was also
not observed. Without the presence of the fluorine atom, the
initial attack onto the C9-HR in combination with a gauche-
conformation of the substrate would generate a (Z)-8-
phytosphingenine.18 The desaturation of the diastereomeric
(S)-8-fluorophytosphinganine (27) delivers both (Z)- and
(E)-8-fluoro-8-phytosphingenines (30 and 31) with the (Z)-
isomer prevailing (E:Z=5:95). Accordingly, the fluorinated
and the nonfluorinated substrates are converted via the same
geometry of the transition state into the unsaturated pro-
ducts. Only the population of the respective anti- and gauche-
conformations is slightly different. While the desaturation of
According to the product distribution compiled in Table 2
the stereochemistry of all fluoroalkene products was in
agreement with the considerations and predictions based in
(33) Carvalho, F.; Gauthier, L. T.; Hodgson, D. J.; Dawson, B.; Buist,
P. H. Org. Biomol. Chem. 2005, 3, 3979–3983.
4980 J. Org. Chem. Vol. 75, No. 15, 2010

Habel et al.
JOCArticle
the natural phytosphinganine by the sunflower desaturase
generates a 88:12 ratio of the (E)- and (Z)-olefins,14 the (S)-8-
fluorophytosphinganine (27) is converted to the (Z)- and (E)-
8-fluoro-8-phytosphingenines (30 and 31) in a ratio of 95:5.
This altered ratio indicates that the fluorinated substrate
strongly favors the anti-configuration at the active center of
the desaturase (Scheme 3). An exclusive production of the
(Z)-9-fluoro-8-phytosphingenine (32) was observed after
administration of the (R)-9-fluorophytosphinganine (28).
This result is easily explained by an initial attack onto the
C8-HR and syn-elimination of two vicinal hydrogen atoms
via an anti-conformation of the substrate. An initial attack
on the C9-F is excluded and generates no product. As
expected, administration of the (S)-7-fluoropalmitic acid
((S)-16) results in the formation of (S)-9-fluorophytosphin-
ganine (29), which produces the (E)-9-fluoro-8-phytosphin-
genine (33) as the sole reaction product. In this case only the
initial attack on the C9-HR in combination with the gauche-
conformation of the substrate is productive.
(72 mg, 0.38 mmol), and (R)-1,2-epoxyundecane ((R)-10)
(0.5 g, 2.9 mmol). Yield: 0.52 g (70%). ee:>97%, according to
1H NMR of the corresponding Mosher ester (see Supporting
Information). IR (cm-1): 3319, 3243, 3080, 2955, 2924, 2873,
2851, 1643, 1466, 1378, 1352, 1133, 1101, 1068, 1046, 1020, 992,
912, 868, 821, 722, 642. 1H NMR (400 MHz, CDCl3): δ 0.88 (t,
J=6.8 Hz, 3H), 1.16-1.50 (m, 24H), 1.96-2.11 (m, 2H), 3.51-
3.63 (m, 1H), 4.84-5.05 (m, 2H), 5.81 (ddt, J=16.9, 10.2, 6.7 Hz,
1H). 13C NMR (100 MHz, CDCl3): δ 14.1, 22.7, 25.5, 25.7, 28.9,
29.2, 29.3, 29.57, 29.63, 29.7, 31.9, 33.7, 37.4, 37.5, 72.0, 114.2,
139.1. EIMS (70 eV) m/z: 253 (0.3, Mþ - H), 236 (2), 208 (6), 194
(3), 183 (3), 166 (2), 157 (16), 155 (8), 138 (5), 127 (13), 109
(100), 97 (46), 95 (23), 83 (89), 81 (36), 69 (50), 67 (72), 57 (47),
55 (77). HRMS (EI) m/z: calcd for C17H34O 254.260966, found
254.260286.
(S)-Heptadec-1-en-8-ol ((S)-12): Prepared from Mg turnings
(0.18 g, 7.4 mmol), 6-bromo-1-hexene (8) (1 g, 6.4 mmol), CuI
(0.11 g, 0.56 mmol), and (S)-1,2-epoxyundecane ((S)-10) (0.73 g,
1
4.3 mmol). Yield: 0.86 g (79%). ee: > 97%, according to H
NMR of the corresponding Mosher ester (see the Supporting
Information). Spectroscopic data are in agreement with (R)-12.
Fluoro-1-alkenes, Representative Procedure. (S)-7-Fluorohep-
tadec-1-ene ((S)-13): A cold solution (-78 °C) of (diethyl-
amino)sulfur trifluoride (DAST, 0.56 g, 3.5 mmol) in CH2Cl2
(20 mL) was treated dropwise with a solution of (R)-11 (0.6 g, 2.3
mmol) in the same solvent (2 mL). Stirring was continued over-
night at room temperature. After hydrolysis with water (15 mL)
the product was extracted with CH2Cl2, washed with sat. NaHCO3
solution, and dried over anhydrous Na2SO4. Purification was
achieved by column chromatography (SiO2) with hexane for
elution. Yield: 0.53 g (87%). IR (cm-1): 3078, 2927, 2856, 1641,
1464, 1378, 1353, 1130, 993, 910, 723, 639. 1H NMR (400 MHz,
CDCl3): δ 0.88 (t, J=7.0 Hz, 3H), 1.68-1.22 (m, 24H), 2.09-
2.01 (m, 2H), 4.45 (dtt, J = 49.12, 7.89, 3.99 Hz, 1H), 4.94 (ddt,
J=10.18, 2.31, 1.23 Hz, 1H), 5.00 (ddt, J=17.12, 2.10, 1.58 Hz,
1H), 5.81 (ddt, J=16.94, 10.19, 6.68 Hz, 1H). 13C NMR (100
MHz, CDCl3): δ 139.0, 114.3, 95.2, 93.9, 35.3, 35.2, 35.1, 35.04,
33.7, 31.9, 29.54, 29.53, 29.52, 29.3, 29.0, 28.8, 25.15, 25.12, 25.01,
24.97, 22.7, 14.1. EIMS (70 eV) m/z: 256 (0.1, Mþ), 236 (3),
208 (23), 194 (3), 180 (4), 165 (5), 152 (9), 151 (11), 137 (12),
124 (19), 123 (22), 110 (26), 109 (38), 95 (96), 82 (84), 67 (100),
54 (70). HRMS (EI) m/z: calcd for C17H33F 256.256630, found
256.255409.
Conclusion
Taken together, our results clearly demonstrate that the
nonstereospecific desaturase from sunflower simultaneously
generates (E)- and (Z)-alkenes via different conformations of
the substrate at the active center of the enzyme. Replacement
of one of the involved hydrogen atoms by fluorine either
prevents the removal of a hydrogen atom, and/or directs the
conformation of the substrate to anti or gauche, which
generated one isomer predominantly (entries 27 and 29 of
Table 2). These findings clearly demonstrate that not the
intrinsic mechanism of the desaturation at the di-iron center,
but instead the conformation of the substrate determines the
stereochemistry of the resulting alkenes. Most likely the
shape of the active site also determines the stereochemistry
of specific (E)- or (Z)-desaturases by allowing only one
defined conformation in the substrate protein interactions.
Experimental Section
Synthesis of Substrates. (R)-Heptadec-1-en-7-ol ((R)-11): The
synthesis was performed according to a published procedure21 from
Mg turnings (0.17 g, 7 mmol), 5-bromo-1-pentene (7) (0.9 g, 6
mmol), CuI (0.1 g, 0.53 mmol), and (R)-1,2-epoxydodecane ((R)-9)
(0.65 g, 3.5 mmol). Yield: 0.76 g (85%). ee: > 97% according to 1H
NMR of the corresponding Mosher ester (see the Supporting
Information). IR (cm-1): 3329, 3239, 3081, 2958, 2921, 2873,
2848, 1643, 1467, 1379, 1345, 1132, 1102, 1066, 1044, 1021, 990,
931, 913, 865, 839, 721, 640. 1H NMR (400 MHz, CDCl3): δ 0.88 (t,
J = 6.8 Hz, 3H), 1.52-1.16 (m, 24H), 2.00-2.11 (m, 2H), 3.52-
3.63 (m, 1H), 4.89-5.05 (m, 2H), 5.81 (ddt, J = 16.9, 10.2, 6.7 Hz,
1H). 13C NMR (100 MHz, CDCl3): δ 14.1, 22.7, 25.1, 25.7, 29.0,
29.3, 29.6, 29.63, 29.64, 29.7, 31.9, 33.7, 37.3, 37.5, 71.9, 114.3, 138.9.
EIMS (70 eV) m/z: 253 (0.5, Mþ - H), 236 (2), 208 (11), 180 (2), 171
(12), 152 (3), 138 (3), 124 (6), 113 (10), 111 (14), 97 (41), 95 (100), 83
(32), 69 (40), 57 (25), 55 (37). HRMS (EI) m/z: calcd for C17H34O
254.260966, found 254.262222.
(R)-7-Fluoroheptadec-1-ene ((R)-13): Prepared from (S)-11
(1.4 g, 5.5 mmol) and DAST (1.32 g, 8.2 mmol). Yield: 1.17 g
(83%). Spectroscopic data are in agreement with (S)-13.
(S)-8-Fluoroheptadec-1-ene ((S)-14): Prepared from (R)-12
(0.30 g, 1.2 mmol) and DAST (0.29 g, 1.8 mmol). Yield: 0.25 g
(83%). IR (cm-1): 3078, 2927, 2856, 1642, 1464, 1379, 1354,
1131, 994, 911, 723. 1H NMR (500 MHz, CDCl3): δ 0.88 (t, J=
7.0 Hz, 3H), 1.69-1.21 (m, 24H), 2.09-2.02 (m, 2H), 4.45 (ddt,
J=23.23, 7.93, 4.02 Hz, 1H), 5.03-4.91 (m, 2H), 5.81 (ddt, J=
16.90, 10.16, 6.67 Hz, 1H). 13C NMR (125 MHz, CDCl3): δ
139.0, 114.3, 95.2, 93.9, 35.3, 35.2, 35.1, 35.0, 33.7, 31.9, 29.54,
29.53, 29.52, 29.3, 29.0, 28.8, 25.15, 25.11, 25.01, 24.97, 22.7,
14.1. EIMS (70 eV) m/z: 256 (0.1, Mþ), 236 (7), 208 (22), 195 (3),
180 (5), 166 (6), 152 (14), 137 (17), 123 (26), 111 (41), 110 (42), 96
(95), 82 (100), 80 (69), 69 (45), 67 (22), 56 (61), 55 (55). HRMS
(EI) m/z: calcd for C17H33F 256.256629, found 256.256510.
(R)-8-Fluoroheptadec-1-ene ((R)-14): Prepared from (S)-12
(0.8 g, 3.1 mmol) and DAST (0.75 g, 4.7 mmol). Yield: 0.62 g
(77%). Spectroscopic data are in agreement with (S)-14.
Preparation of Fluoropalmitic Acids, Representative Proce-
dure. (S)-6-Fluorohexadecanoic Acid ((S)-15): Ruthenium tri-
chloride hydrate (9 mg, 2.2 mol %) was added to a solution
of sodium metaperiodate (1.73 g, 8.1 mmol) and (S)-13 (0.5 g,
2 mmol) in a solvent mixture of CCl4 (4 mL), CH3CN (4 mL),
and H2O (6 mL). The mixture was stirred vigorously for 2 h at
(S)-Heptadec-1-en-7-ol ((S)-11): Prepared from Mg turnings
(0.33 g, 13.4 mmol), 5-bromo-1-pentene (7) (1.8 g, 12.2 mmol),
CuI (0.19 g, 1 mmol), and (S)-1,2-epoxydodecane ((S)-9) (1.5 g,
1
8.1 mmol). Yield: 1.73 g (84%). ee: > 97%, according to H
NMR of the corresponding Mosher ester (see the Supporting
Information). Spectroscopic data are in agreement with (R)-11.
(R)-Heptadec-1-en-8-ol ((R)-12): Prepared from Mg turnings
(0.12 g, 4.9 mmol), 6-bromo-1-hexene (8) (0.72 g, 4.4 mmol), CuI
J. Org. Chem. Vol. 75, No. 15, 2010 4981

Habel et al.
JOCArticle
room temperature. Then, CH2Cl2 (10 mL) was added and the
phases were separated. The aqueous phase was extracted with
CH2Cl2. The combined organic extracts were dried over anhy-
drous Na2SO4 and concentrated. The residue was purified by
flash chromatography on silica gel with Et2O/hexane/formic
acid 1:1:0.01 for elution. Yield: 0.43 g (80%). IR (cm-1): 3060,
2953, 2944, 2915, 2870, 2847, 1695, 1469, 1443, 1407, 1316, 1274,
1258, 1234, 1209, 1132, 1098, 1065, 1021, 990, 971, 930, 856, 828,
786, 733, 719. 1H NMR (400 MHz, CDCl3): δ 0.88 (t, J=6.8 Hz,
3H), 1.17-1.75 (m, 24H), 2.37 (t, J=7.43 Hz, 2H), 4.35-4.57
(m, 1H). 13C NMR (100 MHz, CDCl3): δ 14.1, 22.7, 24.5,
24.6, 24.7, 25.1, 25.1, 29.3, 29.49, 29.53, 29.57, 29.59, 31.9, 33.8,
34.7, 34.9, 35.1, 35.3, 93.4, 95.0, 179.2. EIMS (70 eV) m/z: 254
(8, Mþ - HF), 236 (58), 192 (21), 165 (7), 151 (10), 138 (12), 123
(15), 111 (28), 97 (58), 83 (72), 69 (71), 57 (63), 55 (100). HRMS
(EI) m/z: calcd for C16H30O2 (Mþ - HF) 254.224580, found
254.224823.
directly induced with 10 mL of galactose (20% solution, w/v) to
a final concentration of 2%. Cerulenin (925 μL of 3 mM EtOH
solution) and the fluoropalmitic acid (280 μL of 100 mM EtOH
solution) were added to the culture flask and the inoculated
cultures were incubated for 5 d at 28 °C in a rotary incubator/
shaker set at 220 rpm. The yeast cells were harvested by
centrifugation, resuspended in 4 mL of CHCl3/MeOH (1:1;
v:v), and mechanically disrupted with glass beads by vortexing
for 3 min, followed by stirring for 2 h at 4 °C. Yeast cells were
separated from solvent and the extraction was repeated with
CHCl3/MeOH (2:1; v:v). After separation of the solvent, yeast
cells were resuspended in 5 mL of CHCl3/MeOH (2:1; v:v) and
stirred overnight at 4 °C. The combined organic extracts were
filtered and evaporated.
Analysis of Phytosphingolipids. The residue was dissolved in
2 mL of CH2Cl2 and cooled to 4 °C. Pb(OAc)4 (5-10 mg) was
added followed by an incubation for 2 h at room temperature.
The solution was then passed through a short plug of silica gel
(Pasteur pipet) for removal of residual Pb(OAc)4 and the plug
was washed with Et2O (5 mL). The solvent was evaporated
under a stream of argon.
Derivatization of Released Aldehydes. The residue was dis-
solved in hexane (1 mL) and N,N-dimethylhydrazine (5 μL) was
added followed by incubation for 1 h. The solvent was evapo-
rated under a stream of argon, the residue was redissolved in
CH2Cl2, and an aliquot was applied to GC/MS.
Instrumental Analysis of the Biological Extracts. GC/
MS analysis of N,N-dimethylhydrazones was performed on a
GC-TOF mass spectrometer, a 30 m DB-1 column (ZB-1MS,
30 m ꢀ 0.25 mm, film thickness 0.25 μm; Phenomenex, Torance,
CA, USA) installed. Helium was used as carrier gas (1 mL
min-1); injector temperature: 220 °C; split less mode: GC oven
temperature program: 50 °C (maintained for 2 min) at 10 deg
min-1 to 180 °C followed by 5 deg min-1 to 250 °C followed by
30 deg min-1 to 280 °C (maintained for 2 min). Mass spectra
were recorded in positive EI mode (70 eV).
(R)-6-Fluorohexadecanoic Acid ((R)-15): Prepared as above
from (R)-13 (0.5 g, 2 mmol), NaIO4 (1.73 g, 8.1 mmol), and
RuCl3 H2Ο (9 mg, 2.2 mol %). Yield: 0.4 g (75%). Spectro-
3
scopic data are in agreement with (S)-15.
(S)-7-Fluorohexadecanoic Acid ((S)-16): Prepared as above
from (S)-14 (0.2 g, 0.8 mmol), NaIO4 (0.68 g, 3.2 mmol), and
RuCl3 H2Ο (5 mg, 2.2 mol %). Yield: 0.15 g (71%). IR (cm-1):
3
3060, 2953, 2915, 2848, 1695, 1470, 1444, 1407, 1315, 1296, 1274,
1240, 1210, 1129, 1100, 1066, 1044, 1019, 967, 893, 858, 807, 719.
1H NMR (400 MHz, CDCl3) δ 0.88 (t, J=6.85 Hz, 3H), 1.20-
1.71 (m, 24H), 2.33 (t, J=7.47 Hz, 2H), 4.45 (dtt, J=48.70, 7.70,
3.68 Hz, 1H). 13C NMR (100 MHz, CDCl3): δ 14.1, 22.7, 24.6,
24.77, 24.82, 25.1, 25.2, 28.9, 29.3, 29.50, 29.53, 31.9, 33.6, 34.8,
35.03, 35.07, 35.3, 93.6, 95.2, 178.2. EIMS (70 eV) m/z: 254
(6, Mþ - HF), 236 (52), 192 (15), 165 (7), 152 (12), 138 (12), 125
(17), 111 (24), 97 (49), 83 (62), 69 (70), 57 (45), 55 (100). HRMS
(EI) m/z: calcd for C16H30O2 (Mþ - HF) 254.224580, found
254.224739.
(R)-7-Fluorohexadecanoic Acid ((R)-16): Prepared as above
from (R)-14 (0.4 g, 1.6 mmol), NaIO4 (1.37 g, 6.4 mmol), and
RuCl3 H2Ο (10.5 mg, 2.2 mol %). Yield: 0.31 g (73%). Spectro-
scopic data are in agreement with (S)-16.
Acknowledgment. We thank S. Lorenz for EI-high-resolu-
tion mass spectra. Financial support of this work by the
Deutsche Forschungsgemeinschaft, SFB 436 “Metal Mediated
Reactions Modeled after Nature” and the Max Planck Society
is gratefully acknowledged.
3
Administration of Fluoropalmitic Acids to Growing Cultures of
S. cerevisiae. The galactose-inducible strain of S. cerevisiae
INVSc1 (Invitrogen) containing the Δ8-sphingolipid desaturase
from Helianthus annuus (pHaDES8) has been described
earlier.14 A starter culture (5 mL) of transformed S. cerevisiae
cells was grown at 30 °C for 48 h in complete minimal-dropout-
uracil medium containing 2% (w/v) raffinose (ForMedium).
This culture was used to inoculate 100 mL of medium, contain-
ing 1% Tergitol in a 500 mL Erlenmeyer flask. The cultures were
Supporting Information Available: Experimental details and
1H NMR, 13C NMR, infrared, and mass spectra of substrates
and reference compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
4982 J. Org. Chem. Vol. 75, No. 15, 2010