First synthesis of Xerulin, an inhibitor of the biosynthesis of cholesterol

Article abstract of DOI:10.1055/s-1999-3162

Starting from L-ascorbic acid, crotonaldehyde, (trimethylsilyl)acetylene and the stannylated alcohol 15, the title compound 4 was synthesized for the first time. L-Ascorbic acid was elaborated into phosphonium bromide 19 with a high degree of Z-stereoselectivity while the other starting materials were combined for obtaining the unsaturated aldehyde 5. A Wittig reaction between this aldehyde and the ylide derived from bromide 19 provided xerulin (4) along with a mixture of isomers which was readily separable.

Full text of DOI:10.1055/s-1999-3162

LETTER
1227
First Synthesis of Xerulin, an Inhibitor of the Biosynthesis of Cholesterol
Konrad Siegel, Reinhard Brückner*
Institut für Organische Chemie und Biochemie, Albert-Ludwigs-Universität, Albertstr. 21, D-79104 Freiburg, Germany
Fax +49-761-2036100; E-mail: reinhard.brueckner@organik.chemie.uni-freiburg.de
Received 15 April 1999
Dedicated to Professor Armin de Meijere (Georg-August-Universität Göttingen) at the occasion of his 60th birthday.
X
Y
Abstract: Starting from L-ascorbic acid, crotonaldehyde,
(trimethylsilyl)acetylene and the stannylated alcohol 15, the title
compound 4 was synthesized for the first time. L-Ascorbic acid was
elaborated into phosphonium bromide 19 with a high degree of Z-
stereoselectivity while the other starting materials were combined
for obtaining the unsaturated aldehyde 5. A Wittig reaction between
this aldehyde and the ylide derived from bromide 19 provided xer-
ulin (4) along with a mixture of isomers which was readily separa-
ble.
Me3Si
Y = H: 9
Y = I: 10
X = O:
7
b) (not isolated)
a) 81%
X = CBr2: 8
c) 67%
d) 53% over 2 steps.
I
Key words: butenolides, butyrolactones, C-C coupling, b-elimina-
tion, Wittig reaction
OH
R
R = Me3Si: 12
R = H: 13
11
e) 82%
Steglich, Anke, et al. isolated and elucidated structurally
three unique g-alkylidenebutenolides from Xerula melan-
otricha Dörfelt (Scheme 1): xerulin (4), dihydroxerulin
(3), and xerulinic acid (which is xerulin with a COOH in-
stead of the CH3 group).1 Xerulin (4) and dihydroxerulin
(3) arose as inseparable 90:10 - 65:35 mixtures. They
inhibited the biosynthesis of cholesterol in HeLa S3 cells
(ID50 = 1 mg/g) without being cytotoxic.
f) 55%
OH
14
7
g) 81%
O
PPh3
ref. 2
+
6
O
O
O
1
2
5
7
6
Bu3Sn
OH
O
17
O
C16 C17: 3
C16 C17: 4
15
16
Scheme 2 a) CBr4 (2.0 equiv), PPh3 (4.0 equiv.), CH2Cl2 °CÆroom
temp., 4 h - b) BuLi (1.0 equiv), THF, -78 °C, 5 min, 0 °C, 10 min, -
78 °C, I2 (1.0 equiv), room temp., 10 min; utilized without work-up
immediately thereafter. - c) MeLi (2.2 equiv), THF, -78 °C, 20 min,
r.t., 10 min, I2 (1.0 equiv), -78 °C Æ r.t., 1 h - d) Me3Si-C∫C-I (1.1
equiv), Pd(dba)2 (2 mol-%), AsPh3 (8 mol-%), 15 (1.0 equiv), THF,
r.t., 3 h - e) K2CO3 (1.0 equiv), MeOH, r.t., 14 h. - f) Pd(dba)2 (5
mol-%), CuI (15 mol-%), THF/iPr2NH (5:8, v:v), r.t., 15 min. - g)
Dess-Martin periodinane (1.1 equiv), CH2Cl2 (not dry), r.t., 5 min.
this work
7
Ph3P
O
+
6
O
O
5
6
Scheme 1
Recently, we described the first synthesis of dihydroxeru-
lin (3).2 It began with stereoselective preparations of pho- it delivered up to 30% of the desired trans,Z isomer 3
phorus ylide 1 and lactone aldehyde 2 and ended with a along with up to 25% of a mixture of at least two stereoi-
Wittig reaction between these entities. Irrespective of somers. This lack of stereocontrol implies that phophorus
whether it was effected in the presence or absence of LiBr, ylide 1 is neither ”stable” (creating preferentially a trans-
Synlett 1999, No. 8, 1227–1230 ISSN 0936-5214 © Thieme Stuttgart · New York

1228
K. Siegel, R. Brückner
LETTER
C=C bond) nor ”unstable” (creating preferentially a cis-
C=C bond) but ”semistable” which quality almost ex-
cludes that a high degree of stereocontrol can be exerted.3
As a consequence, when we tried to synthesize the next
representative of Steglich´s and Anke´s Xerula metabo-
lites, namely xerulin (4), and chose to rely again upon a
Wittig reaction for establishing the C6=C7 bond, we decid-
ed to switch roles. We introduced the acyclic portion of
the target molecule as an aldehyde 5 and the heterocyclic
portion as a phosphorus ylide 6 and describe in the follow-
ing how this was accomplished.
19
base; 5 (9-24 h)
7
11
9
5
3
2
10
8
6
17
O
O
O
4
+
9
18
16
7
11
6
The acyclic portion of xerulin was assembled from cro-
tonaldehyde (7), (trimethylsilyl)acetylene (9), and the
stannylated pentadienol 154 (Scheme 2).5 Crotonaldehyde
(7) was C1-elongated to the dibromodiene 8 (81%) which
was subjected to a Fritsch-Buttenberg-Wiechell rear-
rangement under the Corey-Fuchs conditions6. The result-
ing lithioalkyne was quenched with iodine7 furnishing the
iodopentenyne 11 in 67% yield. Commercial (trimethyls-
ilyl)acetylene (9) was oxidized by successive treatment
with BuLi and iodine7 giving a THF solution of the corre-
sponding iodoalkyne 10. It was used as such for a
Pd(dba)2 catalyzed coupling with the stannylalcohol 154
providing the unsaturated alcohol 12 in 53% overall
yield.8 Desilylation in basic methanol9 furnished the
terminal alkyne 13 (82%). It was combined with io-
doalkyne 11 in a Pd(dba)2/CuI-catalyzed Cadiot-Chodk-
iewicz reaction10 providing the desired diyne 14 and
almost no self-coupling products.11 The sequence of
Scheme 2 was terminated by a Dess-Martin oxidation12
leading to the desired aldehyde 513 in 81% yield.
10
8
17
5
cis-4
18
16
3
O
+
2
<3% other isomer(s)
Condi- Base
tions
19:Base:5 Solvent
T
Yield 4 Yield cis-4 +
[ºC]
[%] isomer(s) [%]
19
0
1
2
NEt3
DBN
2.0 : 2.0 : 1 CH2Cl2
2.0 : 2.0 : 1 CH2Cl2
-78 - 20
-78 - 20
-78 - 20
-78 - 20
20
20
0
6
3
KOtBu
2.0 : 2.0 : 1
THF
THF
5
12
0
4
KHMDS 2.0 : 2.0 : 1
12
0
5
dimsyl-Na 2.0 : 2.0 : 1 DMSO
12
26
19
16
28
7
6
BuLi
BuLi
1.1 : 1.1 : 1
2.0 : 2.0 : 1
THF
THF
-78 - 20
0 - 20
20
12
27
20
16
27
7
7
8
K2CO3
K2CO3
K2CO3
K2CO3
1.5 : 1.4 : 1 CH2Cl2
2.0 : 2.0 : 1 CH2Cl2
5.0 : 5.5 : 1 CH2Cl2
5.0 : 5.0 : 1 CH2Cl2
9
50
10
11
50
90
Scheme 4
HO
OH
H
O
5
3
Heating this compound in trimethylphosphite led to the
butenolide-containing phosphonate 18 (95% Z) in 95%
yield while an SN reaction with triphenylphosphine pro-
vided quantitatively the corresponding phosphonium salt
19 (96% Z) at room temperature. Configurational as-
signments of these species are based on the high-field
shifted 300 MHz 1H-NMR resonances of protons 3-H (al-
ways) and 5-H (except in 19) in Z- compared with E-iso-
mers (Table 1);15 we had observed similar shift
differences in g-alkylidenebutenolides earlier.2,15
HO
OH
O
X
5 steps
(ref. 2)
O
O
L-ascorbic acid
X = OH (Z:E = 96:4): 16
X = Br (Z:E = 97:3): 17
a) 72%
b) 95%
c) 100%
5
3
5
3
(MeO)2P(=O)
Br Ph3P
O
O
O
Table 1 Stereochemically relevant 300 MHz 1H-NMR shifts of
g-alkylidenebutenolides 16-19
O
18 (Z:E = 95:5)
19 (Z:E = 96:4)
Z isomer
δ(5-H)
E isomer
δ(3-H)
Scheme 3. a) PPh3 (2.5 equiv), CBr4 (2.5 equiv), CH2Cl2, 0 °C, 30
min. - b) P(OMe)3 (excess), 80 °C, 2 h - c) PPh3 (1.0 equiv), MeCN,
r.t., 14 h.
δ(3-H)
δ(5-H)
7.39 2 5.48 2
7.83 2 5.91 2
16
17
18
19
7.41
7.36
7.56
5.55
5.28
5.86
7.73
7.62
8.80
6.00
5.67
5.57
Potential precursors of the heterocyclic moiety of xerulin
were synthesized from the g-(hydroxyethylidene)buteno-
lide 16 (96% Z; Scheme 3) which we had made accessible
from L-isoascorbic acid for synthesizing dihydroxerulin.2
The OH group of compound 16 was replaced by Br14 giv-
ing the brominated butenolide 17 in 72% yield (97% Z).
Synlett 1999, No. 8, 1227–1230 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
First Synthesis of Xerulin, an Inhibitor of the Biosynthesis of Cholesterol
1229
1
Table 2 H NMR shifts and vicinal coupling constants of natural xerulin (4), synthetic xerulin,
and the synthetic butenolide cis-4 (500 MHz, CDCl3)
2-H
3-H
5-H
6-H 7-H
8-H
9-H 10-H 11-H 16-H 17-H 18-H3
natural 4 1
synthetic 4
cis-4
6.17 7.35 5.88 6.81
6.37 - 6.54
6.44
6.76 5.75 5.60 6.33
6.40 6.78 5.77 5.62 6.35
6.45 6.77 5.78 5.60 6.33
1.83
1.84
1.83
6.19 7.38 5.90 6.83 6.49
6.20 7.42 6.27 6.62 6.34
6.78
J2,3
J5,6
J6,7 J7,8
J8,9
J9,10 J10,11
J16,17 J17,18
-
-
natural 4 1
synthetic 4
cis-4
5.5
5.3
5.3
11.8 14.5
10.5 15.5
10.5 15.3
11.4 15.5
15.8
15.8
15.7
6.8
7.0
6.9
-
-
-
-
-
-
-
-
11.7 14.2 10.7
12.0 11.6 11.9
14.2
14.0
Exploratory Horner-Wadsworth-Emmons reactions be- compounds 4 which allow to construct the C6=C7 bond
tween lithiated phosphonate 18 and aldehyde 5 failed to with stereocontrol.
provide even traces of xerulin. Gratifyingly, Wittig reac-
tions between aldehyde 5 and the ylide derived from
butenolide 19 gave the desired olefin under various condi-
Acknowledgement
We are indebted to the Fonds der Chemischen Industrie (stipend for
K. S.) and the Deutsche Forschungsgemeinschaft ("Sonderfor-
schungsbereich 416" at the Georg-August-Universität Göttingen)
for financing this project, to the Degussa AG for donating PdCl2,
and to the Schering AG for a gift of triflic anhydride.
tions (Scheme 4). A suspension of K2CO3 in CH2Cl2 gave
up to 55% olefin (entry 10), BuLi in THF 53% (entry 7),
and NEt3 in CH2Cl2 39% (entry 1). KHMDS as a base was
mediocre (Æ24% yield; entry 4), KOtBu poor (Æ11%
yield; entry 3) and DBN as well as dimsyl-Na did not lead
to products (entries 2, 5).
References and Notes
Unfortunately, irrespective of the reaction conditions the
Wittig reactions of Scheme 4 were non-stereoselective.
Each of them furnished almost equal amounts of the de-
sired trans-olefin 4 (∫ xerulin; up to 28% yield)16 and its
isomer cis-4.17 However, the latter was easily separable by
chromatography on silicagel (petroleum ether / tert-bu-
tylmethylether 7:1 Æ petroleum ether / tert-butylmethyl-
ether / dichloromethane 4:1:0.2 Æ 2:1:19; 4 eluted first,
cis-4 second).
(1) D. Kuhnt, T. Anke, H. Besl, M. Bross, R. Herrmann, U.
Mocek, B. Steffan, W. Steglich, J. Antibiotics 1990, 43, 1413-
1420.
(2) K. Siegel, R. Brückner, Chem. Eur. J. 1998, 3, 1116-1122.
(3) B. E. Maryanoff, A. B. Reitz, Chem. Rev. 1989, 89, 863-927.
(4) Ref. 2 and footnote 22 therein.
(5) All new compounds gave satisfactory 1H-NMR spectra (300
MHz, CDCl3), IR spectra (film), correct combustion analyses
(except phosphonate 18 and xerulin, both of which, however,
provided correct high resolution mass spectra).
(6) Method: E. J. Corey, P. L. Fuchs, Tetrahedron Lett. 1972,
3769-3972.
(7) Procedure: L. Brandsma, Preparative Acetylenic Chemistry,
Elsevier, Amsterdam - Oxford - New York - Tokyo, 1988, p.
152.
(8) Method: G. J. Hollingworth, J. B. Sweeney, Synlett 1993, 463-
465.
(9) Method: W. B. Austin, N. Bilow, W. S. Kellegham, K. S. Y.
Lau, J. Org. Chem. 1981, 46, 2280-2286.
1
The 500 MHz H-NMR spectrum of synthetic xerulin in
CDCl3 was inaccordance with the data of the natural prod-
uct (Table 2).1 In addition, it allowed a computer analysis
of the previously unresolved multiplet of 7-H, 8-H and 9-
H.16 This analysis revealed a coupling constant J8,9 = 14.2
Hz and thereby confirmed the suspected1 trans-configura-
tion of the C8=C9 bond. In compound cis-4, the cis-config-
uration of the newly formed C6=C7 bond was deduced
from J6,7 = 11.6 Hz and the Z-configuration of the C4=C5
bond from d3-H = 7.42; the latter value is smaller than d ª
7.9 expected2,15 for an E isomer. The Z-configuration of
the C4=C5 bond in compound cis-4 is underlined by the
(10) Method: J. Withyak, J. B. Chan, Synth. Commun. 1991, 21,
977-979.
(11) Cf., for instance, C. Cai, A. Vasella, Helv. Chim. Acta 1995,
78, 2053-2064.
(12) Method: a) D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48,
4155-4156; b) S. D. Meyer, S. L. Schreiber, J. Org. Chem.
1994, 59, 7549-7552.
close resemblance of d3-H,cis-4 = 7.42 with d3-H, natural xerulin
7.35 and d3-H, synthetic xerulin = 7.38.
=
(13) all-trans-2,4,10-Dodecatriene-6,8-diyn-1-al (5): 1H NMR
(300 MHz, CDCl3): d = 1.86 (dd, J12,11 = 6.8, 4J12,10 = 1.9, 12-
H3), 5.62 (dmc, J10,11 ª 15.9, 10-H), 6.15 (d, J5,4 = 15.4, 5-H),
6.23 (dd, J2,3 = 15.5, J2,1 = 8.0, 2-H), 6.40 (dq, J11,10 = 15.9,
In summary, xerulin (4) has been synthesized by a conver-
gent route and was obtained in pure form for the first time.
The only reason why this synthesis cannot be called ´ste-
reoselective´ is that the terminating Wittig reaction did not
show more stereocontrol than the Wittig reaction - of op-
posite polarity - used for synthesizing dihydroxerulin (3).
We are currently looking for modified syntheses of these
J
11,12 = 7.0, 11-H), 6.87 (dd, J4,5 = 15.3, J4,3 = 11.5, 4-H), 7.10
(dd, J3,2 = 15.5, J3,4 = 11.3, 3-H), 9.59 (d, J1,2 = 7.5, 1-H).- IR
(CDCl3): n = 3035, 2915, 2825, 2745, 2195, 1680, 1645, 1605,
1445, 1290, 1215, 1160, 1115, 985, 950, 915, 750, 735, 715
Synlett 1999, No. 8, 1227–1230 ISSN 0936-5214 © Thieme Stuttgart · New York

1230
K. Siegel, R. Brückner
LETTER
cm-1.- Analysis: calc for C12H10O (170.2): C 84.68, H 5.92;
(17) Z-5-[(2Z,4E,6E,12E)-Tetradecatetraene-8,10-
found: C 84.50, H 6.10.
diynylidene)-5H-2-furanone (cis-4): 1H NMR [500 MHz,
CDCl3, contaminated with <10% isomer(s)]: d = 1.83 (dd,
J18,17 = 6.9, 4J18,16 = 1.9, 18-H3), 5.60 (dqd, J16,17 = 15.8, 4J16,18
= 1.6, 7J16,11 = 1.2, 16-H), 5.78 (d, J11,10 = 15.6, 11-H), 6.20 (d,
(14) Procedure: E. H. Axelrod, G. M. Milne, E. E. van Tamelen, J.
Am. Chem. Soc. 1970, 92, 2139-2141.
(15) F. C. Görth, A. Umland, R. Brückner, Eur. J. Org. Chem.
1998, 1055-1062.
J2,3 = 5.3, 2-H), 6.27 (d, J5,6 = 12.4, 5-H), 6.33 (dq, J17,16
=
(16) Xerulin (4): 1H NMR (500 MHz, CDCl3): d = 1.84 (dd, J18,17
15.7, J17,18 = 6.9, 17-H), superimposed by 6.34 (dd, J7,6 ª J7,8
ª 11.5, 7-H), 6.45 (dd, J9,8 = 14.1, J9,10 = 11.4, 9-H), 6.62 (dd,
J6,7 ª J6,5 ª 11.6, 6-H), 6.77 (dd, J10,11 = 15.3, J10,9 = 11.4, 10-
H), superimposed by 6.78 (dd, J8,9 = 13.9, J8,7 = 12.3, 8-H),
7.42 (d, J3,2 = 5.2, 3-H).- 13C NMR [125.7 MHz, CDCl3,
contaminated with <10% isomer(s)]: d = 19.01 (C-18), 72.42,
79.35, 80.53, and 83.37 (C-12, C-13, C-14, C-15), 109.67 (C-
5), 109.85 (C-16), 112.56 (C-11), 119.33 (C-2), 123.77 (C-6),
130.00 (C-8), 134.33 (C-7), 135.77 (C-9), 142.84 (C-3),
143.60 (C-10), 144.07 (C-17), 149.69 (C-4), 169.29 (C-1).-
IR (CDCl3): n = 2925, 2855, 2255, 1775, 1750, 1675, 1600,
1530, 1465, 1380, 1335, 1160, 1105, 1065, 995, 925, 890,
755, 720, 705, 650 cm-1 - UV (MeOH): lmax (lg e) = 405
(6.12), 423 (6.12) nm.-.m/z = 262.0993 (M≈) 2 mDa
confirmed by HRMS (EI, 70 eV).
= 6.9, 4J18,16 = 1.9, 18-H3), 5.62 (dqd, J16,17 = 15.8, 4J16,18 = 1.8,
7J16,11 = 1.0, 16-H), 5.77 (d, J11,10 = 15.3, 11-H), 5.90 (d, J5,6
=
11.7, 5-H), 6.19 (d, J2,3 = 5.3, 2-H), 6.35 (dq, J17,16 = 15.8, J17,18
= 7.0, 17-H), 6.40-6.56 (m, 7-H, 8-H, 9-H)*, 6.78 (dd, J10,11
=
15.3, J10,9 = 10.5, 10-H), in part superimposed by 6.83 (dd, J6,7
= 14.2, J6,5 = 11.9, 6-H), 7.38 (d, J3,2 = 5.2, 3-H). *This
spectrum was simulated by the program DAVINX“ under
WINDOWS“ as a 7-spin system comprising 5-H - 11-H.
Varying the non-observable d and J values systematically
until the calculated spectrum showed the best fit with the
experimental spectrum, we found d7-H = 6.49, d8-H = 6.44, d9-H
= 6.40, J7,8 = 10.7, and J8,9 = 14.2.- 13C NMR (125.7 MHz,
CDCl3): d = 18.92 (C-18), 72.54, 79.40, 80.66, and 83.37 (C-
12, C-13, C-14, C-15), 109.94 (C-16), 112.18 (C-11), 114.72
(C-5), 118.92 (C-2), 127.78 (C-6), 135.13, 135.60, and 137.87
(C-7, C-8, C-9), 142.59 (C-3), 143.85 (C-10), 144.02 (C-17),
149.53 (C-4), 169.40 (C-1).- IR (CDCl3): n = 3030, 2960,
2925, 2855, 2250, 2190, 1775, 1750, 1530, 1335, 1215, 1105,
1065, 995, 920, 895, 805, 765, 715 cm-1.- UV (MeOH): lmax
(lg e) = 404 (6.02), 421 (6.02) nm. m/z = 262.0993 (M≈) 2
mDa confirmed by HRMS (EI, 70 eV).
Article Identifier:
1437-2096,E;1999,0,08,1227,1230,ftx,en;G11799ST.pdf
Synlett 1999, No. 8, 1227–1230 ISSN 0936-5214 © Thieme Stuttgart · New York