Pure & Appl. Chem., Vol. 63, No. 1, pp. 13-22,1991. Printed in Great Britain. @ 1991 IUPAC
Recent progress in the synthesis of butenolide carotenoids and retinoids Masayoshi Ito Kobe Women's College of Pharmacy Motoyamakita-machi, Higashinada-ku, Kobe 658, Japan Abstract- Two kinds of method (the Wittig and sulphone methods) to prepare 4-alkylidenebutenolides are described. An empirical rule regarding the NMR chemical shifts of H-3 in 4-alkylidenebutenolides displaying an extended conjugaed system at the C-2 position is shown. Total synthesis of racemic peridinin and pyrrhoxanthin has been accomplished by means of the sulphone method and optically active peridinin has been synthesised. Recent work on the synthesis of butenolide retinoids is also described.
INTRODUCTION
Of ca. 600 naturally occurring carotenoids of known structure (ref. I), butenolide , peridininol (2) , anhydroperidinin (2), pyrrhoxanthin (4), carotenoids, peridinin (1) and pyrrhoxanthinol (5) are unusual C37-skeletal nor-carotenoids because of the presence of a 4-alkylidenebutenolide system (6) carrying an allene or an acetylene function in the main polyene chain. They are found in the photosynthetic dinoflagellates. The principal carotenoid of the planktonic algae causing "red tide", peridinin, was first isolated in 1890 (ref. 2) and its unique C tricyclic formulation was published in 1971 by the Jensen group (refs. 3 , 4 ) . The stereo%&cure was determined in 1977 by the same group (refs.
5,6).
OH
40-
R': alkyl or polyene chain
6
R2: polyene chain
13
-7
M. IT0
14
On the other hand, from the point of carotenoid function, peridinin is known as an auxiliary light-harvesting pigment for photosynthesis. Because of the importance of peridinin in photosynthesis in the sea, total synthesis of peridinin has been an important goal. The synthesis of peridinin is difficult because of its complicated structure and instability. Therefore, total synthesis of peridinin presents an interesting challenge to synthetic chemists in the carotenoid field. In addition, a new C40-butenolide carotenoid, uriolide (1) , was recently found in a prasinophyte (ref. 7). S Y N T H E S I S OF 4-ALKYLIDENEBUTENOLIDES
Two procedures, a Wittig method (refs. 8,9,10) and a sulphone method (ref. I I ) , were developed to prepare the 4-alkylidenebutenolides (6) which display extended conjugation at the C-2 position. The numbering system used throughout this section is that for the butenolides and illustrated in 6.
i) Wittig method (Schemes 1 and 2) Treatment of the retinoidal acid anhydride (8) with acetyl phosphorane (2) gave the ylidenebutenolide (10) which was converted into the retinoidal ylidenebutenolide In this Wittig reaction, regioselectivity at the carbonyl groups of the acid anhydride (8) was not high. Attempts to extend the conjugation by use of compounds (10)or (11) were frustrated by the instability of the intermediates. In addition, the ketone group was unreactive towards the conjugated polyene ylide. in (10)
(u) .
Scheme 1
R
R R=H, OMe
Ph3P=CHC02Me P
R The lactone-Wittig salt, derived from the corresponding hydroxy-butenolide (12)and triphenylphosphonium bromide, was condensed with conjugated aldehydes in the presence of NaH to give carotenoidal ylidenebutenolides (2)in low yield together with many byproducts. This reaction is presumed to proceed &y the lactone-ylide intermediate
(2t) * Scheme 2
1 ) PPh3'HBr 2 ) NaH/CH2C12
R'
12
3) O H C Y R2
R'
R L H , OH R 2=
Both Wittig methods were found to be inappropriate for the preparation of alkylidenebutenolides with longer conjugated chains because of the drastic reaction conditions. However, by combination of two Wittig methods, E and isomers of 4-alkylidenebutenolides having the common structure (2)were prepared. Consequently, we found an empirical rule that in compounds of this type, the NMR signal for H-3 in the 4z isomer was observed at isomer was found downfield below 67.00 to 7.20, whereas the corresponding signal for the These values can be used to determine the stereochemistry of the ylidene 67.40 (Table 1). part of the conjugated double bond chromophore.
z
Synthesis of butenolide carotenoids and retinoids
15
TABLE 1. Chemical shifts ( 6 p p m ) of H-3 in 4-alkylidenebutenolides (ref. 12)
R' O
3q R 2 H
R1=a, R2=d R'=b, R2=d R1=a, R2=I R ~ = c ,R2=l R'=a, R2=m R'=g, R 1= g , R'=g, R'=g, R'=g, R'=g, R 1 =g, R1=a, R'=b, R'=a, R'=b, R 1 =g,
R2=l R2=d R2=h R2=i R2=f R2=j R2=k R2=e R2=e RZ=f RZ=f R2=e
42-series
4E-series
7.08
7.93
7.08 7.02 7.07 7.04
7.95 7.42 7.44 7.42
7.05 7.16 7.0 7.08 7.13 7.05 7.11 7.00 7.03 7.09 7.11
7.44 7.96 8.01 7.51 7.46 7.42 7.49
a:
e0e0p0 b: Me0
.A0
d:
e:
C:
HO
0k0C02Me
C02Me g: Me
h: C 0 2 M e
i:
*-
C02Me
j: 0-
7.50
0$
f:
m:
0
+y+Q
ii)Sulphone method (Scheme 3) The reaction of the conjugated formyl ester (16)with various allylic sulphones (12)in the presence of lithium di-isopropylamide (LDA) at -78"C gave conjugated alkylidenebutenolides (18)in moderate yields as a mixture (ca. 1 : I ) of and isomers about the ylidene double bond (Table 2). In this reaction, addition of sulphone-anion to the aldehyde, cyclisation of the resulting hydroxy ester, and elimination of the sulphone group took place successively in one pot to give the expected products (18).
z
Scheme 3
COzMe
-
RCHZSOzPh
+R
0
1
LDA/THF:Hex. =1:1 -78oC
u
t
R=conjugated polyene chain
M. IT0
16
Entry
Sulphone (Lp
Product
(U)
I
Total yield of E and Z
56%
+SO,Ph L
2
46 %
3
46 %
@Lso2ph
4
33 %
* 5 I
37 %
SOlPh I
*
I
I
This was improved to 49% yield by use of n-BuLi in THF containing HMPA
By use of the sulphone method, the first synthesis of peridinin and pyrrhoxanthin was with the conjugated C22-allenic achieved by the reaction of the C1 epoxy formyl ester (19) , respectively, (Scheme 4 ) . sulphone (20)o r C22-acetylenic sulphone (21)
HoxToAc wcHo Jza
Scheme 4
C02Me
H
P
h
O
2
S
w
H
f C02Me
H
XCHO -
Synthesis of butenolide carotenoids and retinoids
17
S Y N T H E S I S O F CIS-EPOXY FORMYL ESTER
The 4-butyl-dimethylsilyl(TBDMS) ether (3) of 4-hydroxy-2,2,6-trimethylcyclohexone (2) (ref. 13) was converted into the enol triflate in 89% yield by the reaction with N-phenyltrifluoromethanesulphonimide( Tf2NPh) (ref. 14) in the presence of LDA (Scheme 5). A coupling reaction (ref. 15) of the triflate (3) with methyl acrylate, in the presence of a palladium (11) catalyst, afforded in 93% yield the diene-ester (25) which was reduced with lithium aluminium hydride (LAH) and acetylated to give the allylic acetate (26, 8 0 % ) . The allylic sulphone (7) was prepared in high yield by the reaction with sodium sulphinate (ref. 16). Functionalisation (introduction of carbomethoxyl and alcatalysed by Pd(PPh lyl groups) of the Zdphone (7) and deprotection gave the compound (28)which was oxidised regioselectively at the terminal vinyl group and the sulphone group eliminated to afford a 21%) and (30, 17%), which were each obtained pure mixture of the formyl ester isomers (2, by preparative h.p.1.c.. Iodine-catalysed isomerisation of the isomer (3)provided a mixture ( 3 : 4 ) of 2 and 30. Epoxidation of compound with pchloroperbenzoic acid (MCPBA) gave a mixture of the Q(B)-epoxide (3,56%] and m(a)-epoxide (32, 19%). The stereostructures of both isomers were confirmed by H NMFi data (ref. 11).
(a)
(z)
Scheme 5
0 HO
93%
2,
TBDMSO
22
M(
1) LDA
2
OTf C H , = C H C 0 2 M p
TBDMSO
TfZNPh
Pd(PPh,),CI, Et,N/DMF 93 %
24
89 %
BYco2"' 1) LAH
TBDMSO
2, Ac20/Py 80 %
25
TBDMSO
Pd (PPh,), 89 %
26
-
1) n-BuLi/CICO,Me 2) NaH/
HovcHo CHZSO2Ph
TBDMSO
Br
1) Os04,NaI04
/ v /
3) ( ~ - B u ) ~ N F
HO
2)
A1203
28
61%
22
C0,Me
C0,Me
MCPBA
1,
Z : E =4:3
- WCHO C02Me
19%
C02Me
HO
'
HO M c d 7 2 2 , H O
S Y N T H E S I S OF C22-ALLENIC A N D C ~ Z - A C E T Y L E N I C SULPHONES
The known C1 acetylenic diacetate (2,ref. 17) was transformed through the epoxide (2, ref. 13) to -?he allenic aldehyde (35,ref. 18) which possesses the required three chiral centres in the C -allenic component (Scheme 6). Wittig condensation of the C15-allenic aldehyde (35)wig$ the C7-phosphonium salt (6) and deprotection of the product gave a mixture of two C 2-allenic aldehydes (37)and (%) which were cleanly separated by preparative or (2) was independently transformed to the (allh.p.1.c. in tEe dark. Each isomer E)-allenic sulphone (20) in three steps (Scheme 6). The formation of the (all-E)-sulphone
m)
M. IT0
18
Scheme 6
a@
+ I,,
CH,OA~
5 steps
to
_L
-* MCPBA
Y
AcO
epoxide
CH,OA~
+
AcO
28%
66 %
112
1) NaBH4 2) AczOlPy 3) PhSOzNa / i-PrOH-H~O reflux
*
63% from all-E 51% from 112
AcO
33
Scheme 6
(20) from the (IlZ)-aldehyde (2) is presumed to result from isomerisation in the refluxing conditions. The C22-acetylenic component was prepared similarly as shown in Scheme 7. The C15-acetylenic diacetate (2)was converted into the hydroxyaldehyde (2) which was condensed with the C7-Wittig salt (%). After deprotection and purification by pre arative h.p.1.c. this gave the (all-@-C 2-aldehyde (40,41%), the (IlZ)-isomer (& 23%p, l, and the (9z)-isomer (42, 9%), respectivefy. The (all-E)-acetylenic sulphone (21) was synthesised independently from the two aldehydes (L& and LJ), but the (92)-isomer TQ)was always produced in ca. equal or greater amount than the (all-E)-isomer
(a).
I NaOMe 2) HC I
9%
23 9%
4 1 5%
112
HO
41
92
\
CHO 3) PhSOzNa/i-PrOH-HzO/reflux
21 AcO
31% from all-E 20% from 112
AcO 31% from all-E 45% from 112
CHzSOzPh
42
Synthesis of butenolide carotenoids and retinoids
19
TOTAL S Y N T H E S I S OF PERIDININ A N D P Y R R H O X A N T H I N
The m - e p o x i d e (2)was condensed with the (all-E)-allenic sulphone (20) in the presence of LDA at -78% to afford, in 9% yield, the expected products (Scheme 8)rpreparative and its (Il'E)-isomer in pure form in ca. equal h.p.1.c. in the dark gave peridinin (1) yield. The spectral properties and h.p.1.c. behaviour of the synthetic peridinin were in good agreement with those of the natural specimen (ref. 19). Condensation between the trans-epoxide (2)and the (all-E)-acetylenic sulphone (21) in the presence of LDA at -78°C produced a mixture (1:l) of pyrrhoxanthin (4)and its (11 lE)-isom r in 13% yield; these were cleanly separated by preparative h.p.1.c. in the dark. The 7H NMR data of synthetic pyrrhoxanthin were in accordance with those of natural pigment (ref. 20). Scheme 8
Ho4ToAc
Ph0$3HzC 2p
II
II
1) L D A / T H F : H e x . ( l : l ) / - 7 8 ~ 2) C0,Me
21
I
+
+
1 1'E - i so me r
-
11' E is0 m e r
S Y N T H E S I S OF OPTICALLY ACTIVE PERIDININ
The synthetic peridinin was a mixture of diastereoisomers from which optically active isomers could not be isolated by h.p.1.c.. Thus, with the readily available (4&6@-4hydroxy-2,2,6-trimethylcyclohexanone (44, ref. 21 ) as starting material, the same pathway as described in the synthesis of the racemic peridinin was used for the synthesis of the optically active form. The optically active C1 epoxy formyl ester was prepared in 12 steps and the C allenic sulphone in an optically active form was prepared in 13 steps. Condensation of2$he two components produced optically active peridinin (Scheme 9) whose spectral data (W-VIS, IR, NMR, and MS) were identical with those of the natural specimen. In addition, its circular dichroism (CD) spectrum (Fig. 1) was nearly superimposable on that reported by the Jensen group (ref. 6 ) . This is the first total synthesis of optically active peridinin. Scheme 9
49
12 step/
steps
HO CI5-component
Cz2-component
t
peridinin
M. IT0
20
A€ nm 0
-2
-4
natural peridinin synthetic peridinin
-----
-6
Fig. 1. CD spectra of natural peridinin and optically active synthetic peridinin (in EPA solution)
S Y N T H E S I S OF BUTENOLIDE RETINOIDS
The most attractive compound among the recently described retinoidal butenolides is manoalide (45, ref. 22) which inhibits phospholipase A2 and possesses topical antiinflammatory activity. Manoalide contains a B-substituted y-hydroxybutenolide moiety whose efficient synthesis was described recently by Isoe et al.(ref. 2 3 ) . This provides a general synthetic method for y-hydroxybutenolideswith various substituents, and employs photosensitized oxygenation of substituted a-trimethylsilylfuran, and chemoselective oxidation of a furan ring possessing tri- and tetra-substituted olefins in the side chain. This method was applied to the total synthesis of manoalide (G)which was achieved by two groups (refs. 24,25), by the approaches shown in Schemes 10 and 11.
-
Scheme 10
n-BuLi / THF -78°C
0
CHO
o,, hv CH,CI,-MeOH
-78°C Rose Bengal
@u:R=H manoalide OH
R = M e manoalide24-methylether
21
Synthesis of butenolide carotenoids and retinoids
-
Scheme 11
LDA, DMPU
THF I -78°C
CHO
CHO
TMS 1) 300nm / PhH
TMS
BF3OEt2 I CHZCI, -55°C
2) '02 / MeOH
-45 OH
The precursor (&) of the retinoidal acid anhydride (8) possesses a y-hydroxybutenolide moiety with a conjugated substituent at the B-position. Cytotoxic activity of the retinoidal butenolide (&) on mouse neuro cancer cells has been examined. Compound (&) was found to be about 90 times more potent than retinoic acid (refs. 26,27). In addition, & and related compounds (Q) showed anti-ulcer activity against ulcer models induced by HC1-ethanol, though compound (12)exhibited no activity (ref. 28).
Acknowledgements I am grateful to Emeritus Professor K. Tsukida for valuable discussions and to the devoted collaborations of the members (Dr. T. Iwata, Mrs. Y. Yamano, and Miss Y. Shibata) in my laboratory. I am indebted to Dr. Y. Tanaka, Kagoshima University and Professor S. LiaaenJensen, The Norwegian Institute of Technology, University of Trondheim, for their invaluable gift of natural specimens. Finally, I also appreciate the financial and chemical support of Kuraray Co., Ltd. Japan.
REFERENCES
1. H. Pfander, Key to Carotenoids, 2nd Fa., Birkhber, Basel, 1987. 2. F. Schctt, Ber. Deut. Bot. Ges., 8, 9-32 (1890).
3. H.H.oStrain, W.A. Svec, K. Aitzetmiiller, M.C. Grandolfo, J.J. Katz, H. Kj+sen, S. Norgard, S. Liaaen-Jensen, F.T. Haxo, P. Wegfahrt and H. Rapoport, J. Am. Chem. Soc.,
93,
1823-1 825 (1971 )
.
4. H.H. Strain, W.A. Svec, P. Wegfahrt, H. Rapoport, F.T. Haxo, S. Norg%rd, H. Kj@sen and S. Liaaen-Jensen, Acta Chem. Scand. B, 3, 109-120 (1976). 5. J.E. Johansen, S. Liaaen-Jensen and G. Borch, in Marine Natural Products Chemistry, (D.J. F a d h e r and W.H. Fenical eds.), Plenum Press, pp. 225-2376. J.E. Johansen, G. Borch and S. Liaaen-Jensen, Phytochemistry, 19,441-4.44 (1980). 7. P. FOSS, R.R.L. Guillard and S. Liaaen-Jensen, Ph tochemistr 25 119-124 (1986). ;8?). 8. M. Ito, T. Iwata and K. Tsukida, Heterocycle, 1-9 9. M. Ito, T. Iwata and K. Tsukida, Chem. Pharm. Bull., 2, 1709-1716 (1984). 10. M. Ito, Y. Hirata, Y. Shibata, A. Sat0 and K. Tsukida, J. Nutr. Sci. Vitaminol., 2,
.
31 3-31 8 ( 1987) 11. M. Ito, Y. Hirata, Y. Shibata and K. Tsukida,.J. Chem. Soc., Perkin Trans. I, 197-199 (1990). 12. M. Ito, in Chemistry and Biolo of Synthetic Retinoids, (M.I. Dawson and W.H. Okamura
eds.) , CRC Press, p p . T - d O r
13. E. Widmer, Pure Appl. Chem., $7, 741 -752 (1985). 24, 979-982 (1983). 14. J.E. McMurry and W.J. Scott, Tetrahedron
m.,
M. IT0
22
15. W.J. Scott, M.R. PeEa, K. Swgrd, S.J. Stoessel and J.K. Stille, J. Org. Chem., 50, 2302-2108 (1985). ._. . . _ ~ ,._,. 16. K. Inomata, T. Yamamoto and H. Kotake, Chemistry Lett., 1981, 1357-1360. 17. A.J. Davies, A. Khare, A.K. Mallams, R.A. Massy-Westropp, G.P. Moss and B.C.L. Weedon, J. Chem. SOC., Perkin Trans. I, 2147-2157 (1984). 18. M. Ito, Y. Hirata, K. Tsukida, N. Tanaka, K. Hamada, R. Hino and T. Fujiwara, -Chem. --') Pharm. Bull 36, 3328-3340 (1988). 19. !fhis was kindly supplied by Dr. Y. Tanaka, Kagoshima University, Japan. 20. H NMR data of natural pyrrhoxanthin were kindly provided by Prof. S. Liaaen-Jensen, ~
--
-
The Norwegian Institute of Technology, University of Trondheim, Norway. 21. K;G.W. Leuenberger, W. Boguth, E. Widmer and R. Zell, Helv. Chim. Acta, f 1976). 22. 23 * . ..~ ... 24 * S. Katsumura; S. Fujiwara ani S. Isoe, Tetraheiron Lett., 2t 25.
9,1832-1849
26. 27. M. Ito, T. Iwata and K. Tsukida, Chem. Pharm. Bull., 28. M. Ito, to be published.
&,
4346-4351 (1986).
