Convergent total synthesis of khafrefungin and its inhibitory activity of fungal sphingolipid syntheses

Article abstract of DOI:10.1021/jo0158128

A convergent total synthesis of khafrefungin, a novel inhibitor of fungal sphingolipid syntheses isolated from the fermentation culture MF6020, has been developed. Alkenylboronic acid 5 and alkenyliodide 6, key fragments for the total synthesis, were prepared from the corresponding achiral aldehydes using tin(II)-catalyzed and Zr(IV)-catalyzed asymmetric aldol reactions, respectively. The Suzuki coupling reaction of these two fragments was successfully performed to give 17 in good yield. Through the total synthesis, epimerization of the C4 position having a rather highly acidic proton did not occur, indicating that khafrefungin was under strict conformational constraints to prevent the epimerization process. This characteristic stability of khafrefungin has also been discussed using semiempirical calculation and synthesis. Finally, khafrefungin derivatives have also been synthesized, and their antifungal activities have been measured to obtain information on the structure-activity relationships.

Full text of DOI:10.1021/jo0158128

5580
J . Org. Chem. 2001, 66, 5580-5584
Con ver gen t Tota l Syn th esis of Kh a fr efu n gin a n d Its In h ibitor y
Activity of F u n ga l Sp h in golip id Syn th eses
Shuj Kobayashi,*,† Kouhei Mori,† Takeshi Wakabayashi,† Satoshi Yasuda,‡ and
Kentaro Hanada‡
Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku,
Tokyo 113-0033, J apan, and Department of Biochemistry and Cell Biology, National Institute of
Infectious Diseases (former National Institute of Health), Toyama, Shinjuku-ku, Tokyo 162-8640, J apan
skobayas@mol.f.u-tokyo.ac.jp
Received J une 1, 2001
A convergent total synthesis of khafrefungin, a novel inhibitor of fungal sphingolipid syntheses
isolated from the fermentation culture MF6020, has been developed. Alkenylboronic acid 5 and
alkenyliodide 6, key fragments for the total synthesis, were prepared from the corresponding achiral
aldehydes using tin(II)-catalyzed and Zr(IV)-catalyzed asymmetric aldol reactions, respectively.
The Suzuki coupling reaction of these two fragments was successfully performed to give 17 in good
yield. Through the total synthesis, epimerization of the C4 position having a rather highly acidic
proton did not occur, indicating that khafrefungin was under strict conformational constraints to
prevent the epimerization process. This characteristic stability of khafrefungin has also been
discussed using semiempirical calculation and synthesis. Finally, khafrefungin derivatives have
also been synthesized, and their antifungal activities have been measured to obtain information
on the structure-activity relationships.
In tr od u ction
Sphingolipids are members of cellular membranes and
are composed of a ceramide backbone combined with
polar ends such as monosaccharides, oligosaccharides, or
phosphocoline. Recently, these lipids have been shown
to act as a second messenger for regulating signal
transduction pathways and directly influence various
cellular events such as proliferation, differentiation, and
apoptosis; however, little is yet known about the precise
functions of sphingolipids.1
The proposed pathway for the syntheses of sphingo-
lipids is initiated by the condensation of L-serine with
palmitoyl CoA catalyzed by serine palmitoyltransferase
(SPT) to provide 3-ketosphinganine, which is converted
to sphinganine with a reductase. These steps are ob-
served in the syntheses of sphingolipids of both mammals
and fungi. In mammals, the amino group of sphinganine
is subsequently acylated by sphinganine N-acyltrans-
ferase to give dihydroceramide. After dihydroceramide
is converted to ceramide, various saccharides or phos-
phocoline is added to the resulting ceramide to provide
the corresponding glycosphingolipids or sphingomyelin,
respectively. In fungi, on the other hand, a hydroxyl
group is introduced at the C4 position of sphinganine,
probably because of its more aerobic circumstance, af-
fording phytosphingosine. N-Acylation of phytosphin-
gosine followed by condensation with inositol phosphate
provides inositol phosphorylceramide (IPC), which is
further modified by an addition of mannose to make
mannosyl inositol phosphorylceramide (MIPC) and the
addition of a second inositol phosphate group to make
F igu r e 1. Khafrefungin.
mannosyl diinositol diphosphorylceramide (M(IP)2C). It
is worth mentioning that such a difference in the bio-
synthesis of sphingolipids between mammals and fungi
is shown, suggesting that specific inhibitors of fungal
sphingolipid syntheses are considered to be attractive as
targets for antifungal therapy.
Khafrefungin is a novel antifungal agent isolated from
the fermentation culture MF6020 by a Merck group in
1997 (Figure 1).2 It has been shown to inhibit IPC
synthase, which catalyzes the previously mentioned
fungal specific step, in Sacchromyces cerevisiae and
pathogenic fungi such as Candida albicans and Crypto-
coccus neoformans. Unlike other inhibitors that inhibit
the corresponding enzyme in fungi and mammals to the
same extent, khafrefungin does not impair sphingolipid
synthesis of mammals.
We already performed the total synthesis of sphingo-
fungins B and F that have been shown to inhibit SPT,3,4
and we also investigated the antifungal activities through
the synthesis of its derivatives.4 In addition, quite
recently, we have achieved the complete stereochemical
(2) Mandala, S. M.; Thornton, R. A.; Rosenbach, M.; Milligan, J .;
Garcia-Calvo, M.; Bull, H. G.; Kurtz, M. B. J . Biol. Chem. 1997, 272,
32709.
(3) (a) VanMiddlesworth, F.; Giacobbe, R. A.; Lopez, M.; Garrity,
G.; Bland, J . A.; Bartizal, K.; Fromtling, R. A.; Polishook, J .; Zweerink,
M.; Edison, A. M.; Rozdilsky, W.; Wilson, K. E.; Monaghan, R. L. J .
Antibiot. 1992, 45, 861. (b) Zweerink, M. M.; Edison, A. M.; Well, G.
B.; Pinto, W.; Lester, R. L. J . Biol. Chem. 1992, 267, 25032.
(4) Kobayashi, S.; Furuta, T.; Hayashi, T.; Nishijima, M.; Hanada,
K. J . Am. Chem. Soc. 1998, 120, 908.
† The University of Tokyo.
‡ National Institute of Infectious Diseases.
(1) Reviews: (a) Kolter, T.; Sandhoff, K. Angew. Chem., Int. Ed.
Engl. 1999, 38, 1532. See also: (b) Dickson, R. C. Annu. Rev. Biochem.
1998, 67, 27.
10.1021/jo0158128 CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/17/2001

Convergent Total Synthesis of Khafrefungin
J . Org. Chem., Vol. 66, No. 16, 2001 5581
Sch em e 1. F ir st-Gen er a tion Tota l Syn th esis
Sch em e 2. Syn th etic P la n of Con ver gen t Tota l
Syn th esis of 1
assignment and the first total synthesis of khafrefungin.5
The strategy of our first-generation total synthesis is
summarized in Scheme 1. The polyketide acid part (2)
was prepared by successive propionate additions from
decanal, and one key step was the tin(II)-catalyzed
asymmetric aldol reaction.6 On the other hand, the
aldonic acid part (3) was synthesized from D-arabinose.
In this paper, we report our second-generation, more
convergent total synthesis of khafrefungin. The charac-
teristic stability of khafrefungin is discussed using
semiempirical calculations and synthesis. In addition,
khafrefungin derivatives have also been synthesized, and
their antifungal activities have been measured to obtain
information on the structure-activity relationships.
Sch em e 3. Syn th esis of 6a
Resu lts a n d Discu ssion
Con ver gen t Tota l Syn th esis. Our synthetic plan of
the second-generation total synthesis of khafrefungin is
shown in Scheme 2. Khafrefungin is divided into three
fragments: the aldonic acid part (3), the alkenylboronic
acid 5, and the alkenyliodide 6. Alcohol 3 was a key
intermediate in the first-generation total synthesis and
was already prepared from D-arabinose. On the other
hand, the C7-C8 bond of 4 would be constructed by the
Suzuki coupling reaction of alkenylboronic acid 5 with
alkenyliodide 6. Alkenylboronic acid 5 would be synthe-
sized by the hydroboration of an internal alkyne, which
would be prepared from aldehyde 7 according to the
Corey protocol. Aldehyde 7 was an intermediate in the
first-generation total synthesis and was prepared from
decanal via the tin(II)-catalyzed asymmetric aldol reac-
tion as a key step. On the other hand, alkenyliodide 6
would be prepared by the Wittig reaction of the corre-
sponding aldehyde that would be synthesized by the
asymmetric aldol reaction of aldehyde 8 with a propi-
onate unit.
a
Reagents and conditions: (a) 0.1 equiv of Zr(OtBu)4, 0.12 equiv
of (R)-3,3′-I2-BINOL, 0.8 equiv of PrOH, 0.2 equiv of H2O, toluene,
0 °C, 79%, syn/anti ) 20/80, 96% ee (anti); (b) TPSCl, imidazole,
DMF, rt; (c) DIBAL, CH2Cl2, -78 °C, 76% for two steps; (d)
(COCl)2, DMSO, CH2Cl2, -78 °C, followed by Et3N, rt; (e)
Ph3PdC(Me)CO2Et, THF, reflux, 76% for two steps; (f) DIBAL,
CH2Cl2, -78 °C, 95%; (g) PivCl, pyridine, catalytic amount of
DMAP, 90%.
tin(II)-catalyzed asymmetric aldol reaction could not
serve as the first step for the preparation of 6, probably
because of the instability of aldehyde 8 toward strong
Lewis acids such as metal triflates.7 On the other hand,
our laboratory has recently developed a catalytic asym-
metric aldol reaction using a chiral zirconium complex
prepared from zirconium tetra-tert-butoxide and (R)-3,3′-
diiodo-1,1′-binaphthalene-2,2′-diol ((R)-3,3′-I2-BINOL),
which affords the corresponding anti aldol adducts in
high enantioselectivities.8 The catalytic asymmetric aldol
reaction of aldehyde 8 with the silyl enol ether derived
Initially, we undertook the synthesis of alkenyliodide
6 (Scheme 3). According to our preliminary results, the
(5) Wakabayashi, T.; Mori, K.; Kobayashi, S. J . Am. Chem. Soc.
2001, 123, 1372.
(6) (a) Kobayashi, S.; Mukaiyama, T. Chem. Lett. 1989, 297. (b)
Mukaiyama, T.; Kobayashi, S.; Uchiro, H.; Shiina, I. Chem. Lett. 1991,
129. (c) Kobayashi, S.; Uchiro, H.; Fujishita, Y.; Shiina, I.; Mukaiyama,
T. J . Am. Chem. Soc. 1991, 113, 4247. (d) Kobayashi, S.; Uchiro, H.;
Shiina, I.; Mukaiyama, T. Tetrahedron 1993, 49, 1761. (e) Kobayashi,
S.; Kawasuji, T.; Mori, N. Chem. Lett. 1994, 217. (f) Kobayashi, S.;
Horibe, M. Chem.sEur. J . 1997, 3, 1472.
(7) Raymond, B.; J ose, L. C. J . Chem. Soc., Perkin Trans. 1 1990,
47.
(8) Ishitani, H.; Yamashita, Y.; Shimizu, H.; Kobayashi, S. J . Am.
Chem. Soc. 2000, 122, 5403. For the use of water in preparation of the
Zr catalyst, details will be reported in due course.

5582 J . Org. Chem., Vol. 66, No. 16, 2001
Kobayashi et al.
Ta ble 1. Su zu k i Cou p lin g of 24 w ith 26
Sch em e 4. Syn th esis of 5a
entry
solvent
MeOH
base
yield of 26/%
a
a
Reagents and conditions: (a) CBr4, PPh3, CH2Cl2, rt, 95%; (b)
1
2
3
4
5
6
7
8
9c
K2CO3
trace
trace
20
12
63
trace
trace
trace
74
EtONaa
NaOHa
TlOHa
BuLi, THF, followed by MeI, -78 °C, 93%; (c) catecholborane, 50
°C, followed by H2O.
THF-EtOH (4/1)
THF-H2O (4/1)
THF-H2O (4/1)
THF-H2O (4/1)
hexane-H2O (4/1)
benzene-H2O (4/1)
DME-H2O (4/1)
THF-EtOH (3/1)
TlOEtb
TlOEtb
TlOEtb
TlOEtb
TlOEtb
from phenyl propionate using the chiral zirconium cata-
lyst (10 mol %) proceeded smoothly in toluene at 0 °C to
give the corresponding aldol adduct 9 in 79% yield with
96% ee (anti). Adduct 9 was isolated as white crystals
and could be purified by recrystallization (>99% de,
>99% ee). The protection of the hydroxyl group of adduct
9 as its triphenylsilyl (TPS) ether followed by reduction
using DIBAL gave alcohol 11 in 76% yield for two steps,
which was oxidized to provide aldehyde 12. The Wittig
reaction of 12 with (carbethoxyethylidene)triphenyl-
phosphorane gave the ester 13 in 76% yield for two steps
with high stereoselectivity (E/Z ) >95/5). The ester 13
was then reduced using DIBAL, and the hydroxyl group
of the resulting alcohol (14) was subsequently protected
as its pivaloyl (Piv) ester to give the key fragment (6) in
86% yield for two steps.
We then synthesized another fragment 5 (Scheme 4).
Aldehyde 7, which was an intermediate in the first-
generation total synthesis,5 was treated with carbon
tetrabromide in the presence of triphenylphosphine9 to
give alkene 15. Elimination of hydrogen bromide of
alkene 15 using butyllithium and successive methylation
of the resulting lithium acetylide were carried out in the
same pot to afford the corresponding alkyne 16. The
regioselective hydroboration of alkyne 16 with cat-
echolborane followed by treatment with water gave
alkenylboronic acid 5 as a crude material, which was used
in the following transformations without further purifica-
tion.10
With both key fragments in hand, we directed our
attention to the following Suzuki coupling reaction for
the construction11 of the highly substituted diene. The
model cross-coupling reaction of alkenylboronic acid 24
with alkenyliodide 25 was first planned to be examined.
Several reaction conditions were tested, and the results
are summarized in Table 1. The use of potassium
hydroxide as a base in methanol gave no products. Poor
acceleration was also observed using sodium ethoxide and
sodium hydroxide as bases in aqueous THF. On the other
hand, thallium compounds were reported to be quite
efficient as bases used in the Suzuki coupling reaction
by Kishi and co-workers.12 Indeed, the coupling reaction
of alkenylboronic acid 24 with alkenyl iodide 25 in the
presence of a catalytic amount of tetrakis(triphenylphos-
a
b
2.0 equiv. 2.5 equiv. c 50 °C.
phine)palladium and 2.5 equiv of thallium ethoxide in
aqueous THF proceeded smoothly to afford the desired
product 26 in 63% yield. Among the tested cosolvents of
water, only THF was efficient for this thallium ethoxide-
mediated reaction. It was also noted that the thallium
ethoxide was more efficient than thallium hydroxide,
freshly prepared from thallium formate and sodium
hydroxide according to the reported procedure,13 for the
thallium source used in this reaction. In addition, thal-
lium ethoxide was found to work well in THF-ethanol
without any water, affording 26 in 74% yield.
On the basis of these model studies, the key coupling
reaction for the total synthesis was next examined
(Scheme 5). As expected, the palladium(0)-catalyzed
coupling reaction of alkenylboronic acid 5 with alkenyl-
iodide 6 using thallium ethoxide was successfully per-
formed in aqueous THF to give 17 in 63% yield (based
on 6).14 The undesired deprotection of the p-methoxy-
benzyl (PMB) group at the C12 hydroxyl group of 17
occurred in this coupling reaction (21% yield); however,
fortunately, the resulting material was easily converted
to 17 by treatment with PMB imidate. After the Piv
group of 17 was removed using DIBAL in 81% yield,
oxidation of the resulting alcohol (18) was conducted
using manganese oxide to give aldehyde 19, which was
treated with sodium chlorite to afford carboxylic acid 4
in 73% yield for two steps.15
The following steps were similar to those of the first-
generation total synthesis. The Keck esterification reac-
tion16 of carboxylic acid 4 with alcohol 3 followed by
treatment with a 1 M aqueous hydrochloric acid solution
in THF gave diol 21 in 54% yield for two steps. The
Dess-Martin periodinane-mediated oxidation17 of 21
gave aldehyde 22, which was treated with sodium chlo-
rite15 to afford carboxylic acid 23 that had already been
converted to khafrefungin (one step) in the first-genera-
tion total synthesis.
(13) Tyree, S. Y., J r. Inorg. Synth. 1967, 9, 52.
(14) The use of THF and ethanol as a mixed solvent did not work
well, giving 17 in only 21% yield, probably because of the low solubility
of the substrates used.
(15) Bal, B. S.; Childers, W. E., J r.; Pinnick, H. W. Tetrahedron 1981,
37, 2091.
(9) Corey, E. J .; Fuchs, P. L. Tetrahedron Lett. 1972, 13, 3769.
(10) Review: Lane, C. F.; Kabalka, G. W. Tetrahedron 1976, 32, 981.
(11) (a) Miyaura, N.; Suzuki, A. J . Chem. Soc., Chem. Commun.
1979, 866. Review: (b) Suzuki, A. Acc. Chem. Res. 1982, 15, 178. (c)
Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.
(12) Uenishi, J .; Beau, J . M.; Armstrong, R. W.; Kishi, Y. J . Am.
Chem. Soc. 1987, 109, 4756.
(16) Boden, E. P.; Keck, G. E. J . Org. Chem. 1985, 50, 2394.
(17) Dess, D. B.; Martin, J . C. J . Org. Chem. 1983, 48, 4155.

Convergent Total Synthesis of Khafrefungin
J . Org. Chem., Vol. 66, No. 16, 2001 5583
Sch em e 5. Syn th esis of 1a
F igu r e 2. Lowest-energy conformer of 27.
Sch em e 6. Oxid a tion of Alcoh ols 28, 30, a n d 33a
a
Reagents and conditions: (a) 0.15 equiv of Pd(PPh3)4, TlOEt,
THF, 50 °C, 63% + 21%, see text; (b) DIBAL, CH2Cl2, -78 °C,
81%; (c) MnO2, CH2Cl2, rt; (d) NaClO2, NaH2PO4, 2-methyl-2-
t
butene, BuOH/H2O (1/1), 73% for two steps; (e) 3, DCC, DMAP,
DMAP‚HCl, CH2Cl2, reflux; (f) 1 M HCl/THF (1/3.5), rt, 54% for
two steps; (g) Dess-Martin periodinane, pyridine, CH2Cl2, rt; (h)
t
NaClO2, NaH2PO4, 2-methyl-2-butene, BuOH/H2O (1/1), 55% for
two steps.
Ch a r a cter istic Sta bility. The proton at the C4 posi-
tion of khafrefungin is considered to be rather highly
acidic; however, the chirality at the C4 position of
khafrefungin was demonstrated to remain intact in the
synthesis of khafrefungin and its stereoisomer (vide
infra).5 These results seemed to suggest that khafrefun-
gin was under strict conformational constraints to pre-
vent the enolization process. To confirm this, conforma-
tion analysis of the model compound (27) was performed
using MM2 calculations.18 The results indicated that the
dihedral angle between the C5 carbonyl group and the
C4-H4 bond in the lowest energy conformer was 164°
(Figure 2) and none of the several possible conformers
could meet the dihedral demand for enolization.
To assess the influence of the steric environment of this
type of compound on enolization, the synthesis and
stability of the related compounds (29, 32, and 34) were
next investigated. As was expected, alcohol 28 was
oxidized using Dess-Martin periodinane to give ketone
29 having the same type of substitution as khafrefungin
in high yield.17 On the other hand, the oxidation of alcohol
30 lacking the methyl group at the C2 position under the
same conditions gave ketone 31 as a single product,
which presumably formed via the rapid keto-enol tau-
tomerization of the initially oxidized product (32). In
addition, the oxidation of alcohol 33 lacking the methyl
group at the C6 position was conducted under the same
conditions to give a mixture of ketones 34 and 35 (34:35
a
Reagents and conditions: (a) Dess-Martin periodinane, py-
ridine, CH2Cl2, rt, 1 h.
>14:1). These results clearly explain that the character-
istic stability of khafrefungin is due to its steric hin-
drance, especially due to the methyl group at the C2
position (Scheme 6).19
Str u ctu r e a n d Activity Rela tion sh ip s of Kh a fr e-
fu n gin . We next explored the structure and activity
relationships of khafrefungin. Synthesized khafrefungin
(1) and authentic natural khafrefungin showed antifun-
gal activity with the same MIC values (7.4 µM) (Table
2). The MIC value was similar to the previously reported
MIC value of natural khafrefungin on S. cerevisiae cells.2
Removal of the aldonic acid group (39, Figure 3) com-
pletely inactivated the antifungal activity (Table 2). Thus,
the aldonic acid group is essential for the activity of
khafrefungin. The importance of the aldonic acid group
is not simply due to the enhancement of miscibility in
water, because the replacement of the aldonic acid group
(18) (a) Hehre, W. J .; Yu, J .; Klunzinger, P. E. A Guide to Molecular
Mechanics and Molecular Orbital Calculations in SPARTAN; Wave-
function, Inc.: Irvine, CA, 1997. (b) Hehre, W. J .; Shusterman, A. J .;
Huang, W. W. A Laboratory Book of Computational Organic Chemistry;
Wavefunction, Inc.: Irvine, CA, 1996.
(19) Cf.: Evans, D. A.; Clark, J . S.; Metternich, R.; Novack, V.;
Sheppard, G. S. J . Am. Chem. Soc. 1990, 112, 866.

5584 J . Org. Chem., Vol. 66, No. 16, 2001
Kobayashi et al.
types of khafrefungin inhibited synthesis of sphingolipids
by ∼70%, whereas other synthesized isomers did not
inhibit it at all (data not shown), in agreement with the
results in terms of antifungal activity previously de-
scribed. Therefore, the configuration of the methyl group
at the C4 position in khafrefungin appears to be crucial
for inhibiting the activity of IPC synthase, the target
enzyme of khafrefungin, although it remains to be
determined whether the presence of the methyl group
at the C4 position is essential for the inhibition activity.
Con clu sion
Total synthesis of khafrefungin has been achieved
using catalytic asymmetric aldol reactions and the Suzuki
coupling reaction as key steps. The synthesis is more
convergent than our first total synthesis. For the struc-
ture of khafrefungin, strict conformational constraints to
prevent enolization were indicated, and semiempirical
calculations and synthesis have clarified the stability.
Furthermore, khafrefungin derivatives were prepared,
and their antifungal activities were measured to obtain
information on the structure-activity relationships. Be-
cause we have established two types of stereoselective
synthetic routes for khafrefungin in the previous and
present studies, these established routes would allow us
in the future to investigate the structure and activity
relationships of this drug in more detail by using various
derivatives of khafrefungin.
F igu r e 3. Khafrefungin derivatives.
Ta ble 2. An tifu n ga l Activities
compound
MIC/µM
7.4
>200
1
36
37
38
en t-1
en t-36
en t-37
en t-38
39
>200
>200
>200
>200
>200
>200
>200
Ack n ow led gm en t. The authors are grateful to Dr.
Guy H. Harris (Merck) for providing us with an au-
thentic sample of khafrefungin, spectral data, and other
valuable information. The authors also thank Mr. Hiroki
Kobayashi for his technical support. This work was
partially supported by CREST, J apan Science and
Technology Corporation (J ST), and a Grant-in-Aid for
Scientific Research from the Ministry of Education,
Science, Sport, and Culture, J apan. T.W. thanks the
J SPS fellowship for J apanese J unior Scientists.
authentic sample
7.4
by its enantiomer type (36) also resulted in suppressed
activity (Table 2). To our surprise, isomerization of the
methyl group at the C4 position (37) completely inacti-
vated khafrefungin (Table 2). Moreover, another stereo-
isomer (38) and enantiomers of the compounds 1, 36, 37,
and 38 (en t-1, en t-36, en t-37, and en t-38, respectively)
exhibited no activity (Table 2). We further examined the
effect of khafrefungin and its derivatives on the synthesis
of sphingolipids on intact yeast cells by metabolic labeling
of lipids with radioactive inositol. When the cells were
exposed to 2.5 µM drugs, both synthesized and natural
Su p p or tin g In for m a tion Ava ila ble: Experimental sec-
1
tion and copies of H and 13C NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O0158128