Chemoenzymatic synthesis of conjugated linoleic acid

Full text of DOI:10.1021/jo981905m

9620
J . Org. Chem. 1998, 63, 9620-9621
13C NMR spectra9 of the pure isomers were consistent with
this assignment.16
Ch em oen zym a tic Syn th esis of Con ju ga ted
Lin oleic Acid
Although the 10E,12Z and 10Z,12E isomers can be
separated by silver-ion impregnated HPLC,6 no current
analytical techniques are available to separate the 9Z,11E
and 9E,11Z isomers. Therefore, it was necessary to deter-
mine more definitively the position and geometric configura-
tions of double bonds in the CLA preparation, which was
carried out as follows: (1) The double bonds of the CLA
isomers were partially reduced with hydrazine.10 (2) The
resulting E monoene fraction, isolated by silver-impregnated
silica gel column chromatography, was subjected to oxidative
ozonolysis followed by methylation.11 (3) The diesters were
isolated by silica gel chromatography, and their structures
were verified by mass spectral analyses. When this analyti-
cal procedure was used on the synthetic CLA isomers, only
dimethyl 1,11-undecanedioate and dimethyl 1,10-decanedio-
ate were obtained.12 These results not only confirm the
absence of geometric isomers but also demonstrate the
outstanding selectivity of the deprotonation-protonation
procedure.
As the CLA regioisomers are difficult to separate from
each other using chromatographic methods, we turned our
attention to the use of enzymatic methods. It is well-known
that the lipase from Geotrichum candidum13 has a unique
specificity for unsaturated fatty acids containing a 9Z double
bond. That is, esters of 9Z unsaturated fatty acids are
attacked more rapidly than other types of unsaturated fatty
acids. When the synthetic CLA methyl esters (2a and 2b)
were exposed to this lipase at pH 7.0, the 9Z,11E isomer
(2a ) was preferentially hydrolyzed with a regioselectivity
factor14 (RS) of about 9. Since selectivity can often be
enhanced by conducting the reverse reaction in organic
solvents,15 we incubated the synthetic CLA isomers with this
lipase in 1-butanol for esterification. As expected, the
enzyme preferentially esterified the 9Z,11E isomer (1a ), but
the regioselectivity (RS) was improved to only 12. This
moderate improvement led us to examine other lipases with
improved regioselectivities. After much experimentation, we
discovered that the lipase of Aspergillus niger (APF 12,
Amano) was considerably more selective. In the esterification
Chien-An Chen and Charles J . Sih*
School of Pharmacy, University of Wisconsin,
425 North Charter Street, Madison, Wisconsin 53706
Received September 21, 1998
Conjugated linoleic acid (CLA) refers to a mixture of
positional and geometric isomers of octadecadienoic acids
(18:2). It was first isolated and identified by Pariza and co-
workers from a fraction of grilled beef that possessed
antimutagenic activity.1 Since then, interest in CLA has
increased markedly because of reports that dietary CLA
reduced carcinogenesis, atherosclerosis, and body fat in
laboratory animals.2 The major isomer, 9(Z),11(E)-octadeca-
dienoic acid (1a ), is believed to be the most biologically active
CLA isomer1,2 and was shown to be the first intermediate
in the biohydrogenation of linoleic acid by bacteria in the
rumen,3 which may account for the high CLA content in
meat from ruminants and dairy products.4 However, com-
mercial CLA products5 fed to experimental animals are
complex mixtures of CLA isomers, many of which have not
been characterized. Analysis of this mixture using silver-
ion impregnated HPLC revealed the presence of at least 12
components.6 To date, there is no definitive evidence as to
which isomer within the CLA mixture is the biologically
active component or whether different isomers are respon-
sible for different bioactivities. To further study the biological
effects of CLA, it is desirable to obtain substantial amounts
of the individual pure isomers, ideally by simple methods.
Here, we describe a chemoenzymatic method for the syn-
thesis of 9(Z),11(E)-(1a ) and 10(E),12(Z)-octadecadienoic
acids (1b), the two most abundant CLA isomers found in
foods (Scheme 1).7
Since the superbasic mixture of n-butyllithium and potas-
sium tert-butoxide has been successfully used for the depro-
tonation of allylic and 1,4-pentadienyl systems with excellent
regio- and stereoselectivity,8 we selected this organometallic
reagent for the metalation of linoleic acid. In a representa-
tive experiment, 1 g of linoleic acid was treated with 3.5
equiv of this base at -78 °C in THF for 30 min. The reaction
was quenched by pouring the mixture into 50 mL of 6 N
HCl at 0 °C. After the usual workup, the crude product
mixture was treated with diazomethane and purified via
AgNO3-silica gel chromatography using hexane/ethyl ac-
etate (60:1) as the eluent to give a mixture of methyl
octadecadienoate in 77% isolated yield. Analysis of this
mixture using silver-ion impregnated HPLC6 revealed the
presence of only two peaks in a ratio of 4:6, corresponding
to the retention times of methyl 10(E),12(Z)- and 9(Z),11-
(9) 1a : 1H NMR (CDCl3, 300 MHz) δ 0.88 (3H, t, J ) 6.1 Hz), 1.20-1.50
(16H, m), 1.63 (2H, m), 2.12 (4H, m), 2.34 (2H, t, J ) 7.6 Hz), 5.28 (1H, m),
5.65 (1H, m), 5.94 (1H, t, J ) 10.8 Hz), 6.29 (1H, dd, J ) 15 Hz, 11 Hz); 13
C
NMR (CDCl3, 75 MHz) δ 180.5 (C1), 134.8 (C12), 129.8 (C9), 128.7 (C10),
125.6 (C11), 34.1 (C2), 32.9 (C13), 31.8 (C16), 29.7, 29.4, 29.1, 29.0, 28.9
(C4, C5, C6, C7, C14, C15), 27.6 (C8), 24.6 (C3), 22.6 (C17), 14.1 (C18). 1b:
1H NMR (CDCl3, 300 MHz) δ 0.89 (3H, t, J ) 6.3 Hz), 1.20-1.50 (16H, m),
1.63 (2H, m), 2.12 (4H, m), 2.34 (2H, t, J ) 7.5 Hz), 5.30 (1H, m), 5.65 (1H,
m), 5.94 (1H, t, J ) 10.9 Hz), 6.30 (1H, dd, J ) 15 Hz, 11 Hz); 13C NMR
(CDCl3, 75 MHz) δ 180.5, 134.5, 130.1, 128.6, 125.7, 34.1, 32.9, 31.5, 29.4,
29.3(8), 29.3, 29.1(8), 29.1(5), 29.0, 27.7, 24.7, 22.6, 14.1.
(10) Privett, O. S.; Nickell, E. C. Lipids 1966, 1, 98.
(11) Ackman, R. G. Lipids 1977, 12, 293.
(12) Ryhage, R.; Stenhagen, E. In Mass Spectrometry of Organic Ions;
McLafferty, F. W., Ed.; Academic Press: New York, 1963; p 399.
(13) Baillargeon, M. W.; Bistline, R. G., J r.; Sonnet, P. E. Appl. Microbiol.
Biotechnol. 1989, 30, 92. See also: J ensen, R. G. Lipids 1974, 9, 149.
(14) In systems containing two competing regioisomers, the regioselec-
tivity (RS) may be calculated using the following expression:
1
(E)-octadecadienoates (2b and 2a ), respectively. The H and
(1) Ha, Y. L.; Grimm, N. K.; Pariza, M. W. Carcinogenesis 1987, 8, 1881.
(2) (a) Ip, C.; Chin, S. F.; Scimeca, J . A.; Pariza, M. W. Cancer Res. 1991,
51, 6118. (b) Shultz, T. D.; Chew, B. P.; Seaman, W. R.; Luedecke, L. O.
Cancer Lett. 1992, 63, 125. (c) Lee, K. N.; Kritchevsky, D.; Pariza, M. W.
Atherosclerosis 1994, 108, 19. (d) Park, Y.; Albright, K. J .; Liu, W.; Storkson,
J . M.; Cook, M. E.; Pariza, M. W. Lipids 1997, 32, 853.
ln[(1 - C)(1 - SE)]
RS )
ln[(1 - C)(1 + SE)]
(3) (a) Kepler, C. R.; Tove, S. B. J . Biol. Chem. 1967, 242, 5686. (b)
Hughes, P. E.; Hunter, W. J .; Tove, S. B. Ibid. 1982, 257, 3643.
(4) Steinhart, C. J . Chem. Educ. 1996, 73, A302 and references therein.
(5) Nichols, P. L., J r.; Herb, S. F.; Riemenschneider, R. W. J . Am. Chem.
Soc. 1951, 73, 247.
(6) Sehat, N.; Yurawecz, M. P.; Roach, J . A. G.; Mossoba, M. M.; Kramer,
J . K. G.; Ku, Y. Lipids 1998, 33, 217.
(7) Lee, K. M. Conjugated Linoleic Acid and Lipids Metabolism, Ph.D.
Thesis, University of Wisconsin, 1996.
(8) (a) Schlosser, M. Pure Appl. Chem. 1988, 60, 1627. (b) Schlosser, M.;
Rauchschwalbe, G. J . Am. Chem. Soc. 1978, 100, 3258.
C denotes the extent of conversion, and SE is the excess of one
remaining substrate divided by the total remaining substrate at
conversion C. See: Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J .
J . Am. Chem. Soc. 1982, 104, 7294 and also ref 13. A 1:1 mixture of
1a and 1b was used for the determination of RS.
(15) Chen, C. S.; Sih, C. J . Angew. Chem., Int. Ed. Engl. 1989, 28, 695.
(16) (a) Berdeaux, O.; Christie, W. W.; Gunstone, F. D.; Sebedio, J . L. J .
Am. Oil Chem. Soc. 1997, 74, 1011 and references therein. (b) Body, D. R.;
Shorland, F. B. J . Am. Oil Chem. Soc. 1965, 42, 5. (c) Gunstone, F. D.; Said,
A. I. Chem. Phys. Lipids 1971, 7, 121. (d) Lie Ken J ie, M. S. F. L. K.; Pasha,
M. K.; Alam, M. S. Lipids 1997, 32, 1041.
10.1021/jo981905m CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/02/1998

Communications
J . Org. Chem., Vol. 63, No. 26, 1998 9621
Sch em e 1
mode in 1-butanol, this lipase also esterified preferentially
the 9Z,11E isomer (1a ) with an RS of 25. In a typical
experiment, 1 g of CLA consisting of a 60:40 mixture of 1a
and 1b was incubated with 100 mg of lipase APF-12 (A.
niger, Amano) in 36 mL of 1-butanol and 14 mL of water.
After being stirred at 25 °C for 18 h, the reaction was
terminated. Silicic acid column chromatography [ethyl acetate/
hexane (1:30) followed by methanol/methylene chloride (5:
95)] of the mixture afforded 510 mg of enriched (9Z,11E)-
butyl ester 3a (SE ) 0.92)14 and 610 mg of enriched 1b (SE
) 0.26). To secure 1a , the enriched 3a fraction was hydro-
lyzed with 1 M KOH back to 1a and subjected again to
regioselective esterification using the same lipase to yield
340 mg of pure 3a (SE ) 1.0), which upon base hydrolysis
afforded pure 1a (265 mg) in 44% overall yield. Further
enrichment of the 1b fraction was achieved via the selective
removal of 1a by repeating the regioselective esterification
two times. Highly enriched 1b (SE ) 0.91) was obtained in
75% overall yield.
In summary, we have shown that the superbase (n-
butyllithium/potassium tert-butoxide) reacted smoothly with
linoleic acid to give only the 9Z,11E and 10E,12Z CLA
isomers, 1a and 1b, which were conveniently separated from
each other using the lipase from A. niger via regioselective
esterification. This enzyme has a preference for the 9Z,11E
isomer, 1a , and has excellent regioselectivity. This chemoen-
zymatic method has allowed the ready preparation of the
9Z,11E and 10E,12Z CLA isomers in their highly purified
forms, which are not readily accessible by current methods.16
Further refinement of this technology for large-scale syn-
thesis is in progress, and the results of these experiments
will be reported at a later date.
Ack n ow led gm en t. This investigation was supported
in part by a grant from the National Institutes of Health.
We thank Professor Barry Trost for helpful discussions.
J O981905M