THE JOURNALOF BIOLOGICAL

Communication CHEMISTRY
Val. 256, No. 16, Issue of August 25, pp. 8259-8262, 1981
Prrnted in U S.A.

lism of unsaturated fatty acids had been carried out with

Degradation of Unsaturated Fatty crude mitochondrial fractions. The enzymes found to be in-
Acids in Peroxisomes volved in this pathway have been assumed to be of mitochon-
drial origin.
EXISTENCE OF A 2,4-DIENOYL-CoA REDUCTASE In this report
we wish to describe results which demonstrate
PATHWAY* that peroxisomes have the enzymatic apparatus necessary for
the degradation of unsaturated fattyacids. They possess their
(Received for publication, March 6, 1981, and in revised form,
May 12, 1981) own 2,4-dienoyl-CoA reductase and metabolize 4-cis-decen-
oyl-CoA,a key metabolite in linoleic aciddegradation, accord-
Veronika D o m e s , Cornelia Baumgart, and ing to the2,4-dienoyl-CoA reductase pathway.
Wolf-H. Kunau
From the Institut furPhysiologische Chemie, Abteilung EXPERIMENTAL PROCEDURES
fur Naturwissenschaftliche Medizin, Ruhr- Universitat
Bochum, Postfach 102 148,D-4630Bochum 1, Federal Materials
Republic of Germany Clofibrate, NAD, NADPH, NADP, FAD, horseradish peroxidase,
and P-hydroxyacyl-CoA dehydrogenase were purchased from
1.Thereare two 2,4-dienoyl-CoAreductases (EC Boehringer Mannheim, GmbH, Mannheim, FRG. All other chemicals
1.3.1.-) (formerly called 4-enoyl-CoA reductase, Eur. J. were obtained from E. Merck, Darmstadt, FRG; Serva, Heidelberg,
Biochem. (1978) 91, 533-544) in rat liver, one in mito- FRG; or Sigma, Munich, FRG. (4S)-[4-3H]NADPHwas prepared as
chondria and another one in peroxisomes. described by LaBelle and Hajra (14). Acyl-CoA esters were synthe-
2.Treatmentof ratswithclofibrateincreasesthe sized as reported previously (13).
activities of the 2,4-dienoyl-CoA reductases in both cell
Methods
organelles.
3. Isolatedperoxisomesmetabolize4-cis-decenoyl- Female rats were fed either a standard diet or one enriched with
CoA, a metabolite of linoleic acid, via the 2,4-dienoyl- 3% (w/w) clofibrate for 2 weeks. The livers were homogenized in 2
CoA reductase pathway. mM 3-(N-morpholino)propanesulfonic acid, 1 mM ethylene glycol
4.From these results it is concludedthatperoxi- bis(P-aminoethyl ether)-N,N,N,N-tetraaceticacid, 0.1%ethanol, pH
somes have the enzymatic apparatus necessary for the 7.2, containing 250 mM sucrose (15).
Differential centrifugation was performed at 650 X g to produce
degradation of unsaturated fatty acids. the homogenate, at 5,000 X g to sediment heavy mitochondria, and at
27,000 X g to obtain light mitochondria (16).
Sucrose Density Gradient Centrifugation-A discontinuous su-
crose density gradient was used similar to that described by Osumi
Recently, ithas been shown for animal tissues that P- and Hashimoto (3). Itwas composed of 10 ml of 55% sucrose, 10 ml
oxidation of saturated fatty acids takes place in peroxisomes of 44.75% sucrose, and 10 ml of 33%sucrose in homogenization buffer.
as well as in mitochondria (1-4). It, has been reported that Samples of light mitochondria fractions containing 40 mg of protein
were loaded on each gradient. Centrifugation was performed for 60
three out of the four reactions of the P-oxidation cycle are min in a Sorval centrifuge RC 2B using the vertical rotor SS90 at
identical in both organelles (2,4),although they arecatalyzed 20,000 rpm in analogy tothe procedure described by Neatand
by organelle-specific enzymes(4-7). Furthermore,there is Osmundsen (15). Fractions of 1.2 ml were taken and analyzed for
considerable evidence that the fmst P-oxidation step, the in- sucrose density, protein content (A~W), and activities of marker en-
troduction of a trans-doublebond in position 2, is mediated in zymes and various P-oxidation enzymes.
both organelles by different mechanisms. In mitochondria the Enzyme Assays-In all assays 0.1% Triton X-100was added to
reaction is linked via electron-transferring proteins to the disintegrate the organelles. Catalase was measured photometrically
as described by Aebi (17).Fumarase assays were performed according
respiratory chain producing finally H20, whereas in peroxi- to Bergmeyer et al. (18). The acyl-CoA oxidase test described by
somes the hydrogens are directly transferred to oxygen, giving Hryb and Hogg (19) was used with a slight modification. The assay
rise to H202as reaction product. The enzymes catalyzing these contained 0.1 M Tris-HC1, pH 8.3,25 mM p-hydroxybenzoic acid, 1
different reactions are acyl-CoA dehydrogenases (8) and acyl- mM 4-aminoantipyrin,O.O4mg/ml of horseradish peroxidase, 0.02 mg/
CoA oxidase(s) (1, 5, 9, lo), respectively. ml of bovine serum albumin, M FAD, 0.1% Triton X-100, and
M acyl-CoA. The increase of absorbance at 500 nm was deter-
Unsaturated fatty acids just as saturated ones are oxidized
mined.
by P-oxidation, but additional enzymes are necessary to elim- The assay of acyl-CoA dehydrogenases as described previously (20)
inate the interfering cis-double bonds (11-13). Our previous was slightly modified in order to adjust it to measurements of these
findings led us to suggest that 2,4-dienoyl-CoA reductase enzymes in whole organelles. An artificial primary electron acceptor
instead of 3-hydroxyacyl-CoA epimerase is an obligatory en- could be omitted, asthe endogenous acceptor is present in abundance;
zyme for the degradation of unsaturated fatty acids (13). however, KCN had to be added to prevent losses of electrons to the
Previous studies concerning the specific catabolic metabo- respiratory chain. Acyl-CoA dehydrogenase assays contained 0.2 M
N,N-bis(2-hydroxyethyl)glycine-KOH, pH 8.0,
2.5 X M 2-p-io-
dophenyl-3-p-nitrophenyl-5-phenyl-2H-tetrazolium chloride, M
* This work was supported by the Deutsche Forschungsgemein- acyl-CoA, 0.15% Triton X-100, and M KCN. The reduction of
schaft and the Fonds der Chemischen Industrie. The costs of publi- tetrazolium salt was followed at 492 nm.
cation of this article were defrayed in part by the payment of page Consumption of 4-cis-decenoyl-CoAby either acyl-CoA oxidase or
charges. This article must therefore be hereby marked aduertise- acyl-CoA dehydrogenase I1 was also followeddirectly by the increase
ment in accordance with 18 U.S.C. Section 1734 solely to indicate of the reaction product, 2-trans, 4-cis-decadienoyl-CoA,as monitored
this fact. a t 296 nm; 19,500 M cm was taken as molar extinction coefficient
This enzyme has previously been designated as 4-enoyl-CoA (21).
reductase (13). 2,4-Dienoyl-CoAreductase was assayed as described before (13).2-

8259
8260 Peroxisomal
Degradation of Unsaturated
Acids
Fatty
Enoyl-CoA hydratase activities were measured according to thecom-
bined assay procedure described by Fong and Schulz (22).
Incubation studies were followed spectrophotometrically and
stopped by addition of 2 ml of 2 N NaOH. The reaction products were
analyzed by radio gas chromatography and mass spectrometry as
described previously (13).

RESULTS

Subcellular Fractionation of Rat Liver by Differential

Centrifugation-The results of cell fractionation by means of
differential centrifugation are summarized in Fig. 1. The data
show that clofibrate does not increase solely the specific and
absolute activities of peroxisomal enzymes but also those of
some mitochondrial ones. This is consistent with results of
recent studies from other laboratories using various hypolipi-
demic drugs (4,9, 23, 24) or different physiological conditions
(23,25). The observed differences in enzymatic activities seem
to suggest that the clofibrate effect exhibits some specificity
for the enzymes of fatty acid degradation. The specific activ-
ities of the marker enzymeswereonly slightly enhanced
(catalase, 1-5-fold) or not at all (fumarase), whereas the
specific activities of the P-oxidation enzymes were increased
4-fold(acyl-CoAoxidase, light mitochondria fraction, and
acyl-CoA dehydrogenase, heavy mitochondria fraction) and
4-fold or 10-fold (2,4-dienoyl-CoAreductase, heavy and light
mitochondria fractions).
A 4-7-fold and 2-fold increase in the activity of 2,4-dienoyl-
CoA reductase in crude mitochondrial fractions of rats fed a
clofibrate and high fat diet, respectively, has recently been i i 6 a io 12 i1 16
reported by Borrebaek et al. (26, 27). Fraction number
The distribution pattern of acyl-CoA oxidaseand acyl-CoA
FIG. 2. Distribution of enzyme activities after sucrose den-
dehydrogenase were very similar to those of the marker sity gradient centrifugation of the light mitochondrial frac-
enzymes catalase and fumarase, respectively. That different tion. Sucrose density gradient centrifugation was performed in a
vertical rotor as described under Experimental Procedures. Frac-

10 I * 10 I B tions were taken from the bottom of the centrifuge tubes. Fractiona-
tion was stopped after collecting the fractions containing mitochon-
dria.

amounts of acyl-CoA dehydrogenase(s) and fumarase are

found in the supernatant fraction is most likely due todifferent
0 50 1000 50 100 0 50 1000 50 100 leakage out of damaged mitochondria and therefore might
reflect different submitochondrial compartmentation of these
enzymes.
In contrast, the activity profiies of 2,4-dienoyl-CoAreduc-
tase were different from those of the two marker enzymes.
With control animals, the highest specific activity was found
in the heavy mitochondria fraction. However, clofibrate treat-
ment increased the specific activity in the light mitochondria
fraction by a factor of 10, and that of the heavy mitochondria
fraction only by a factor of4. These results suggested the
possibility of a duallocalization of 2,4-dienoyl-CoAreductase.
Separation of Mitochondria and Peroxisomes by Equilib-
rium Centrifugation-Mitochondria and peroxisomes of the
light mitochondria fraction were separated by density gradient
centrifugation (Fig. 2). A discontinous sucrose gradient similar
to that described by Osumi and Hashimoto (3) has been used
FIG. 1. Subcellular distribution of enzyme activities after as it gave better results than a linear gradient. Separation of
differential centrifugation. Specific activities (ordinate)are plot- the two organelles is complicated by the apparent inhomoge-
ted against relative amounts of protein in the different fractions
(abscissa). The protein content of the 650 X g supernatant neity of both organelle types. This often leads to a small
(= homogenate) was set to 100%. Specific activities of enzymes of the peroxisomal contamination of the heavy side of the mito-
homogenate are shown by dashed lines. Specific activities are given chondrial peak, or to mitochondria contaminating the light
in s/mg for catalase (17) and in nanomoles.s/mg for the other side of the peroxisomal peak. Employing the discontinous
enzymes. The burs indicate amount and activity of fractions of gradient it has been possible in most preparations to obtain
differential centrifugation: 5,000 X g sediment = heavy mitochondrial peroxisomal fractions essentially free of mitochondria. How-
fraction (EA), 27,000 X g sediment = light mitochondrial fraction (H),
and 27,000 X g supernatant (a). The enzymes assayed are: A, catalase; ever, the mitochondrial fractions were always contaminated
B, fumarase; C, acyl-CoA oxidase; 0,acyl-CoA dehydrogenase; E, 2,4- to a small extent by peroxisomes as judged by the activity
dienoyl-CoA reductase. Left, the data from the control group; rcht, profiies of catalase as well as acyl-CoA oxidase.
rats fed the clofibrate diet. Fig. 2 demonstratesfor the f i s t time the complete absence
 Peroxisomal
Degradation of Unsaturated
Acids
Fatty 8261
previous studies with mitochondrial fractions (13). When trit-
ium-labeled 4-cis-decenoyl-CoAis incubated in the absence of
A / NADPH, the radio gas chromatogram (Fig. 3A) reveals only
1500 two radioactive methyl esters in addition to the incubated
E8 acid, namely methyl 2-trans, 4-cis- and methyl 2-trans, 4-
-G 1000 trans-decadienoate. The latteris most likely an artifact arising
.
E
500
from isomerization during preparation of methyl esters for the
C gas chromatographic analysis (13).Incubations in the presence
3
s .. of NADPH led to additional metabolites (Fig. 3B): 3-decenoic
-x- acid, 2-trans-decenoic acid, 3-hydroxydecanoic acid, and oc-
I
v
lsolveni I 0)
Q tanoic acid. Only these four additional intermediates were
;
u
2000
E, radioactively labeled when radioactivity was introduced by
0
200 means of the coenzyme (NADPH) instead of the substrate (4-
{
+

1500 180
0
n 160 cis-decenoyl-CoA) (results not shown). 3-Hydroxy-4-cis-de-
1000 1LO cenoic acid and 2-cis-octenoic acid intermediates of the epi-
120 merase pathway (11) have not been found in any of these
100
500 metabolic studies.
These results strongly suggest that in peroxisomes, just as
5 10 15 20 25 30 35 LO 45
in mitochondria, 2-trans, 4-cis-decadienoyl-CoAis the sub-
Tlme lminl strate for the NADPH-dependent 2,4-dienoyl-CoAreductase
step. This conclusion is further supported by data shown in
FIG. 3. Radio gas chromatographic analyses of incubation Fig. 4. It has been possible to follow spectrophotometrically
products. A, incubation of purified peroxisomes with 4-cis-[4,5- the formation of 2-trans, 4-cis-decadienoyl-CoA in the ab-
3Hz]decenoyl-CoA without NADPH; B, incubation as in A but with
addition of NADPH. The heights of the bars represent the amounts sence of NADPH at 296 nm, the absorption maximum of 2,4-
of radioactivity detected per fractions. The values were corrected for dienoyl-CoA structures (21). Addition of NADPH led to a
a background of 50 cpm. The mass peaks represent methyl esters of decrease of this specific absorption reflecting the reduction of
the different fatty acid markers: a, octanoic acid; b, 2-cis-octenoic the diene structure.
acid; e, decanoic acid; d, 4-cis-decenoic acid; e, 3-decenoic acid; 2-
trans-decenoic acid; g, 2-trans, 4-cis-decadienoic acid; h, 2-trans, 4-
trans-decadienoic acid;i, 3-hydroxyoctanoic acid; k, 3-oxodecanoic CONCLUSION
acid; 1, 3-hydroxydecanoic acid; m, 4-cis, 3-hydroxydecenoic acid.
The experiments described in this paper strongly support
the following conclusion.
I
O
First, 4-cis-decenoyl-CoA is degraded in peroxisomes via
6
z the 2,4-dienoyl-CoAreductase pathway (Figs. 3 and 4):
4-cis-decenoyl-CoA + 2-trans, 4-cis-decadienoyl-CoA +
3-decenoyl-CoA + 2-trans-decenoyl-CoA -+
3-hydroxydecanoyl-CoA+ 3-oxodecanoyl-CoA+
octanoyl-CoA -+ + +

The stereochemistry of the double bond in 3-decenoyl-CoA

Incubation ttme Ihl
FIG. 4. Appearance and disappearance of 2-trans, 4-cis-de-
cadienoyl-CoA as measured by following the change in ab-
sorption at 296 nm. Peroxisomes purified by density gradient cen-
trifugation were incubated with 4-cis-decenoyl-CoA in the absence of
NADPH (-). Addition of NADPH resulted in a decrease of ab-
sorption (- - -).
of acyl-CoA dehydrogenase(s) from peroxisomes, thus provid-
ing additional experimental support tothe view that the
mechanisms of the first step of the P-oxidation cycle are
organelle-specific.
The activity profile of 2,4-dienoyl-CoA reductase com-
pletely paralleled that of enoyl-CoA hydratase. The latter
enzyme is known to be present in both types of organelles (5,
6).
Degradation of 4-Cis-decenoyl-CoAin Peroxisomes-Fig.
3 shows that 4-cis-decenoyl-CoAwhen incubated with purified
peroxisomes givesrise to the same metabolites as detected in
has not yet been determined for the peroxisomal reaction
sequence. The reaction product of the mitochondrial 2,4-dien-
oyl-CoA reductase possesses the trans-configuration.'
Second, the existence of a peroxisomal2,4-dienoyl-CoA
reductase, the prerequisite for this pathway in peroxisomes,
has been proven (Fig. 2). This enzyme exhibits the same
coenzyme dependence (NADPH) as the mitochondrial one.
Third, the introduction of a 2-trans-double bond by the
acyl-CoA oxidase is not prevented by a pre-existing double
bond in position 4 (conversion of 4-cis-decenoyl-CoA into 2-
trans, 4-cis-decadienoyl-CoA).Our studies with mitochondria
have shown that this is not necessarily to be expected because
only one out of the three acyl-CoA dehydrogenases from beef
liver mitochondria is capable of catalyzing this reaction.' The
fact that there may also be more than one acyl-CoA oxidase
in peroxisomes possessingdifferent substrate specificitiescan-
not yet be excluded, although there are no experimental data
so far.
Fourth, in peroxisomes, the degradation of 2-trans, 4-cis-
decadienoyl-CoA does not lead to any of the intermediates of
the epimerase pathway, thus rendering that reaction sequence
very unlikely as an alternativefor the 2,4-dienoyl-CoA reduc-
tase pathway.
Fifth, since the conversion of 3-decenoyl-CoA to 2-trans-
* Unpublished result.
8262 Degradation
Peroxisomal
decenoyl-CoA requires an enoyl-CoA isomerase, peroxisomes
evidently are equipped with all enzymes necessary to metab-
olize fatty acids containing double bonds at even numbered as
well as odd numbered carbon atoms.
Thus, if one takes all results together about peroxisomal
/?-oxidationof saturated and unsaturated fattyacids, a picture
emerges which is qualitatively very similar to that established
for mitochondria. At present, however, it remains unclear why
eukaryotic cells should possess two distinct P-oxidation sys-
tems, although some interesting hypotheses have been pro-
posed (2, 24, 28).
Acknowledgments-We gratefully acknowledge the help of Dr.
Ferdinand Wirtz-Peitz (Nattermann & Cie, Cologne) who performed
the gas chromatography-mass spectrometry analyses. We are in-
debted to U. Dorpmund for skillful technical assistance.
REFERENCES
1. Lazarow, P. B. & de Duve, C. (1976) Proc. Nutl. Acad. Sci.
U. S. A. 73,2043-2046
2. Lazarow, P. B. (1978) J. Biol. Chem. 253, 1522-1528
3. Osumi, T. & Hashimoto, T. (1978) J. Biochem. (Tokyo)83,1361-
1365
4. Osumi, T. & Hashimoto, T. (1979) J. Biochem. (Tokyo) 85, 131-
139
5. Osumi, T. & Hashimoto, T. (1979) Biochem. Biophys. Res. Com-
mun. 89,580-584
6. Osumi, T. & Hashimoto, T. (1980) Arch. Biochern. Biophys. 203,
372-383
7. Miyazawa, S., Osumi, T. & Hashimoto, T. (1980) Eur. J. Bwchem.
103,589-596
8. Beinert, H. (1963) in The Enzymes (Boyer, P. D., Lardy, H. A.,
and Myrback, K., eds) 2nd Ed, Vol. 7, pp. 447-466, Academic
Press, New York
of Unsaturated Fatty Acids
9. Inestrosa, N., Bronfman, M. & Leighton, F. (1979) Life Sei. 25,
1127-1136
10. Osumi, T., Hashimoto, T. & Ui, N. (1980) J. Biochem. (Tokyo)
87,1735-1746
11. Stoffel, W. (1966) Nuturwissenschaften 53,621-630
12. Kunau, W. H. & Bartnik, F.(1974) Eur. J. Biochem. 48,311-318
13. Kunau, W.H. & Dommes, P. (1978) Eur. J. Biochem. 91, 533-
544
14. LaBelle, E. F., Jr. & Hajra, A. K. (1972) J. Biol. Chem. 247,5825-
5834
15. Neat, C. E. & Osmundsen, H. (1979) Biochem. J. 180,445-448
16. van Tol, A. & Hiilsmann, W. C. (1969) Biochim. Biophys. Acta
189,342-353
17. Aebi, H. (1974) in Methoden der Enzymatischen Analyse (Berg-
meyer, H. U., ed) 3rd Ed, pp. 713-718, Verlag Chemie, Wein-
heim, FRG
18. Bergmeyer, H. U., Gawehn, K. & GraB1, M. (1974) in Methoden
der Enzymatischen Analyse(Bergmeyer, H. U., ed) 3rd Ed, pp.
479-480, Verlag Chemie, Weinheim, FRG
19. Hryb, D. J. & Hogg, J. F. (1979) Biochem. Biophys. Res. Commun.
87,1200-1206
20. Dommes, V. & Kunau, W. H. (1976) Anal. Biochem. 71,571-578
21. Wakil, S. J. & Hiibscher, G. (1960) J. Biol. Chem. 235, 1554-1558
22. Fong, J. C. & Schulz, H. (1976) J. Biol. Chem. 252,542-547
23. Mannaerts, G. P., Debeer, L. J., Thomas, J. & De Schepper, P. J.
(1979) J. Biol. Chem. 254,4589-4595
24. Miyazawa, S., Furuta, S., Osumi, T. & Hashimoto, T. (1980)
Biochim. Biophys. Acta 630,367-374
25. Ishii, H., Horie, S. & Suga, T. (1980) J. Biochem. (Tokyo) 87,
1855-1858
26. Borrebaek, B., Osmundsen, H. & Bremer, J. (1980) Biochem.
Biophys. Res. Commun. 93, 1173-1180
27. Borrebaek, B., Osmundsen, H., Christiansen, E. N. & Bremer, J.
(1980) FEBS Lett. 121.23-24
28. Osmundsen, H., Neat, C. E. & N o m , K. R. (1979) FEBS Lett.
99,292-296