Coenzyme a in purified peroxisomes is not freely soluble in the matrix but firmly bound to a matrix protein

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Vol. 139, No. 3, 1986

Pages 1195-1201

September 30, 1986

COENZYME A IN PURIFIED PEROXISOMES IS NOT FREELY SOLUBLE IN THE MATRIX BUT FIRMLY BOUND TO A MATRIX PROTEIN 1

Paul P. Van Veldhoven and @uy P. Mannaerts Afdeling Farmakologie, Katholieke Unive~iteit Leuven, B-3000 Leuven, Belgium Received July 30, 1986

SUMMARY: On subfractionation of purified rat liver peroxisomes in matrical, peripheral membrane, integral membrane and core protein fractions, the endogenous peroxisomal CoA was released together with the matrix proteins. The released CoA could not be measured by an enzymatic cycling assay unless the matrix proteins were denatured by acid treatment or by heating at alkaline pH. The cofactor could not be removed by dialysis of the matrix proteins unless salt was added. It was not displaced by exogenous CoA. It migrated into sucrose density gradients together with a protein of approximately 80 kDa. The results indicate that peroxisomal CoA is firmly bound to a matrix protein and that the presence of CoA inside purified peroxisomes does not necessarily imply that the peroxisomal membrane is impermeable to this cofactor. ® 1986AcademicPress, Inc.

In

the

course

characterization that

their

measurements on

Further

phospholipid

purified rat

the sucrose

demonstrated that of other

proteins

studies

permeable to

permeate through

laboratory, confirmed

variety (4).

membrane is

substrates easily

addition,

initial

on

the

biochemical

sucrose and

that the three

oxidases, known at that time, do not show latency, implying that

permeability our

their

of peroxisomes, de Duve and collaborators (1,2) established

the peroxisomal

peroxisomal

of

vesicles

indicated that

membrane. Direct

liver peroxisomes,

permeability of

conducted in

peroxisomes and, in

isolated peroxisomes are readily permeable to a

small molecules

permeability

the peroxisomal

such as

studies

reconstituted

on with

NAD* (3), CoA, ATP and carnitine

isolated

peroxisomes

peroxisomal

integral

and

on

membrane

the peroxisomal permeability to the above mentioned

This work was supported by 8rants from the Belgian 'Fonds voor Geneeskundig Wetenschappelijk Onderzoek' and the 'Onderzoeksfonds van de Katholieke Universiteit Leuven'. 0006-291X/86 $1.50 1195

Copyright © 1986 by Academic Press, lne. AII rights of reproduction in any Jorm reserved.

Vol. 139, No. 3, 1986

cofactors but

is not

by the

forming membrane

our

protein 2 (4). to

The

water

to small

nonselective soluble

membrane of

a nonseleetive

permeability

molecules

is

in

of

the

agreement

pore-

peroxisomal with

our

peroxisomes do not contain pyridine nucleotides,

carnitine (5), but it is difficult to reconcile with

the purified

CoA (6).

been considered

decided to

the peroxisomal

that purified

findings that

membrane we

small

nucleotides o r

unesterified has

mediated by the presence of specific membrane transloeases

presence in

observations adenine

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

In fact, as evidence

organelles contain

significant amounts

of

the presence of CoA in purified peroxisomes for the

charged compounds

impermeability of the peroxisomal

(7). In order to solve the discrepancy

investigate the nature of the peroxisomal CoA pool in greater

detail. METHODS

Peroxisomes were purified from livers of male clofibrate-treated rats by a combination of differential centrifugation and isopycnic centrifugation in iso-osmotic self-generating Percoll gradients as described earlier (3). The purified peroxisomes were subfractionated by two procedures (4), which will be described in detail elsewhere 2. Procedure I consisted of the following successive treatments: two sonications in 10 mM 3-(N-morpholino)propanesulfonate buffer pH 7.2, containing I mM EDTA pH 7.2, I mM dithiothreitol and 0.5 mM phenylmethylsulfonylfluoride (buffer A), which released the soluble matrix proteins; treatment with I M NaCI, which released the peripheral membrane proteins; treatment with 1 % (w/v) Triton X-tO0 and with 1 % (w/v) Triton X-tO0 plus I M NaCl, which solubilized the integral membrane proteins. In procedure II the peroxisomes were first sonicated twice in 10 mM pyrophosphate buffer pH 9, containing I mM EDTA pH 9, I mM dithiothreitol and 0.5 mM phenylmethylsulfonylfluoride (buffer B), which released the soluble matrix proteins and the peripheral membrane proteins. The remaining membranes and cores were then treated with detergent and with detergent plus NaCl as described above, which solubilized the integral membrane proteins. Portions of the peroxisomal matrix proteins, released after sonication in buffer A, were placed in dialysis bags and dialyzed at room temperature against 100 volumes of buffer A, containin$ no additional salt or various amounts of NaCI. Aliquots were removed from the bags at different time points and analyzed for CoA. Peroxisomal proteins, released after sonication in buffer B, were concentrated by dialysis against solid polyethylene glycol 20 000 to a final concentration of 4 ms per ml. 2 ml were layered on top of a linear sucrose gradient [5-35 % (w/v) sucrose in buffer B; 38 ml] and centrifuged in a Beckman VTi 50 rotor for 12 hours at 167,000 g. Fractions of equal volume were collected and analyzed for their polypeptide pattern by sodium dodecyl sulfate gel eleetrophoresis and for their CoA content. Electrophoresis was performed in 10 to 20 % (w/v) acrylamide gradient slab gels with a 3 % (w/v) acrylamide stacking layer, as described by Laemmli (8). The gels were stained with Coomassie blue R (~).

2p.p. Van Veldhoven, W.W. Just and G.P. Mannaerts: manuscript submitted. 1196

Vol. 139, No. 3, 1986

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

CoA present in purified peroxisomes and in subperoxisomal fractions was measured by a fluorimetric adaptation of the cycling assay of Veloso and Veech (10). This method measures the sum of unesterified CoA and acetyl-CoA. Since we have previously found that in purified peroxisomal fractions acetyl-CoA does not exceed 10-15 % of this sum (6), CoA and acetyl-CoA were not determined separately in the present experiments. CoA was measured on neutralized samples that had been deproteinized first with perchloric acid or heated at 70 ° for 20 min at pH 9.5 in the presence of 5 mM dithiothreitol. No difference in CoA content was found between acid- and base-treated samples, indicating the virtual absence of long chain acyl-CoA esters, which precipitate under acidic conditions and which are hydrolyzed under alkaline conditions. Catalase, urate oxidase, acyl-CoA oxidase, protein and phospholipids were determined as described previously (11,12). 3-Ketoacyl-CoA thiolase and carnitine octanoyltransferase were measured according to the method of Miyazawa et al. (13,14). 3-Hydroxyacyl-CoA dehydrogenase was measured as described by Purata et al. (15). RESULTS AND DISCUSSION When different matrix

purified rat

liver peroxisomes

constituents, most

enzymes (Table

of the

were subfractionated

CoA was

released

into

together

their

with

the

I), indicating a predominantly matrical localization.

Table I: Matrix localization of peroxisomal CoA. Purified peroxisomes were separated into their constituents by various successive treatments as described in Methods. Enzyme and CoA release were measured after each treatment. Results are expressed as percentage of total recovered activity or amount, and are means z S.E.M. for the number of experiments indicated in parentheses. The CoA content of purified whole peroxisomes was 0.70 ± 0.11 nmol/mg protein (n = 7). Acyl-CoA oxidase and 3-ketoacyl-CoA thiolase, two soluble matrix enzymes, followed the same pattern of release as catalase; the bulk of 3-hydroxyacyl-CoA dehydrogenase, which behaves as a peripheral membrane protein, was released after NaCI treatment (procedure I) or after sonication in pyrophosphate buffer (procedure I I ) ; the bulk of phospholipids (membranes) was solubilized after detergent treatment; urate oxidase (cores) was not released (data not shown in the Table). T~EAT~NT

CATALAS"E

CoA

Release, ~ of Total Pl-oeedure I (n = 3) Sonication NaCl Triton X-100 Triton X-tOO + NaCl

85.7 3.0 9.4 0.6

-+ 1.7 t 0.7 z 1.5 -+ 0.1

69.4 11.7 13.3 1.7

+- 0.7 -+ 3.5 -+ 2.8 z 0.6

% Remainir~ % Recovery

0.4 +- 0.I 90.4 -+ 5.9

3.4 -+ 1.1 93.2 -+ 8.5

96.4 ± 2.7 3.3 ± 2.7 0.2 -+ 0.1

97.4 +- 0.8 0.7 -+ 0.4 0.3 -+ 0.1

0.2 + 0.I 104.9 ± 7.5

I .6 -+ 0.6 92.2 -+ 9. I

Procedure II (n = 4) Sonication, pyrophosphate Triton X-tO0 Triton X-tO0 + NaCI % Remaining % Recovery

1197

Vol. 139, No. 3, 1986

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

100 - - ~ = ~ . -

~s 5o 5

200

~00

"

I000

Dialysis lime (minufes) Fig, I : Behavior of peroxisomal m a t r i x OoA during d i a l y s i s . Portions of peroxisomal matrix proteins released after sonication of purified peroxisomes in buffer A (see Methods), were placed in dialysis bags and dialyzed against 100 volumes of buffer A, containing no salt (II) or NaCI: 50 mM (A); 100 mM (@); 200 mM (~); 500 mM (W). At different time points aliquots were removed from the dialysis baEs and analyzed for CoA. The initial CoA content was I .22 nmol/mg matrix protein. Results are expressed as percentaEe of the initial amount.

The

denaturation treatment could in

n o t freely

cofactor is

of the

measured with

the absence

were

released matrix

the matrix,

proteins

however. Without prior

either

by

perchloric

acid

or by heating in the presence of dithiothreitol at alkaline pH, it

not be

bound,

soluble in

the employed enzymatic cycling assay. Moreover,

of salt it could not dialyzed, indicating that it was firmly

presumably to dialyzed in

a matrix

protein (Fig.

I). When

the matrix proteins

the presence

of increasing

salt concentrations,

CoA was

gradually

released, evoking the possibility that an electrostatic binding is

involved.

On separation

in

sucrose density gradients, CoA could not be found in the top fractions of

the

gradient, but

proteins

which

it migrated

assumes a

before diffused

8radient together

with the matrix

according to the method of McEwen (16), was 5.31 ± 0.44 S (n = 3),

corresponds to

proteins

into the

(Fig. 2). The sedimentation coefficient of the CoA binding protein,

estimated

one

of the matrix proteins by rate zonal centrifugation

a relative

spherical form

were first they were into the

molecular mass of approximately 80 kDa, if

for the

incubated with

centrifuged in

binding

exogenous

protein.

When

radioactively

the

matrix

labelled

CoA

the sucrose gradients, the radioactive CoA

gradients durin8

centrifugation, but 1198

it did not migrate

Vol. 139, No. 3, 1986

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ab 1

5

10

15

96 67 '7, ~3 ×

x

30 20."

14..~

B

10

"-.... i 5

10

15 Fraction

Fig. 2: Separation of peroxisomal proteins on sucrose density gradlents. Peroxisomal proteins, released after sonication of purified peroxisomes in buffer B, were concentrated as described in Methods. Before they were separated on sucrose density gradients, the concentrated proteins were divided into two portions, one of which was incubated for 30 min. at 4°C with 2 nmol [~H(G)] Coenzyme A (specific radioactivity: 4.5 Ci/mol) per mg protein. After centrifugation, the gradient fractions were analyzed for their polypeptide pattern (A) and for their CoA content (g). A: lane a: molecular weight markers; lane b: 150 ~g of total proteins released; lanes 1-16: proteins contained in 0.1 ml of The gradient fractions. Fractions I and 16 are the bottom and top fractions respectively. B, solid line: CoA content of the gradient, centrifuged in the absence of exogenous radioactive CoA. The CoA content of the fractions is expressed as percentage of the total CoA content of the gradient, which was 0.97 nmol/mg protein. Recovery of 0oA after centrifugation: 98 %. B, broken line: content of exogenously added radioactive CoA of the gradient, centrifuged in the presence of radioactive CoA. The content of exogenous radioactive CoA of the fractions is expressed as percentage of the total added radioactive CoA. Recovery: 101%.

together

with

exogenously did

the

added CoA

not exchange

pattern

of

and Since

of two

proteins

(Fig.

2).

This

indicates

that

the

did not bind to possibly unoccupied sites and that it

with the endogenous bound CoA. Analysis of the polypeptide

the

eleetrophoresis presence

matrix

gradient revealed

fractions

that

the

polypeptides with

one polypeptide 3-ketoacyl-CoA

by

sodium

presence

of

dodeeyl CoA

sulfate

coincided

gel

with

molecular masses of approximately

the

40 kDa

with a molecular mass of approximately 30 kDa (Fig. 2). thiolase

requires 1199

CoA

and

since

the

peroxisomal

Vol. 139, No. 3, 1986

thiolase

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

is composed

of two subunits of 40 kDa (17), it is conceivable that

CoA

is associated

not

be detected after centrifugation, probably because of its lability (17).

CoA

did not

acyl-CoA

with this

migrate in

esters

gradients:

as

enzyme. Unfortunately, thiolase activity could

the gradients with the following enzymes, which use

substrate

acyl-CoA

and

oxidase,

whose

activity

3-hydroxyacyl-CoA

was

measured

dehydrogenase,

in

the

carnitine

octanoyltransferase (data not shown). Our

findings satisfactorily explain the presence of CoA inside purified

organelles,

whose membranes

significance,

if any, of the CoA binding is not clear, however. Since CoA is

released

from its

possible

that little

The

putative binding

noncovalent

and

surprisingly

low, since

most

of CoA likely

the presence of salt, it is

suggests that the nature of the binding is electrostatic.

The

dissociation

rate

is

the binding survived the long-lastin8 procedures of

purification and

centrifugation

protein in

or no CoA is bound to this protein in the intact cell.

salt-induced release

peroxisome

are readily permeable to CoA. The physiological

or dialysis.

subfractionation, followed The possibility

that CoA

by gradient density is covalently bound,

e.g.

as a glutamyl-CoA thiol ester enzyme intermediate as has been described

for

succinyl-CoA: 3-ketoacid CoA transferase (18), seems unlikely in view of

the

longevity of

the binding

at neutral

pH, its lability at acidic pH and

its lability in the presence of salt.

REFERENCES

I. de Duve, C., and Baudhuin, P. (1966) Physiol. Rev. 46, 323-357. 2. Baudhuin, P. (1969) Ann. N.Y. Acad. Sci. 168, 2141228. 3. Van Veldhoven, P., Debeer, L.J., and Mannaerts, G.P. (1983) Biochem. J. 210, 685-693. 4. Van Veldhoven, P.P. (1986) Ph. Thesis, Katholieke Universiteit Leuven, Belgium. 5. Mannaerts, 6.P., Van Veldhoven, P., Van Broekhoven A., Vandebroek, G., and Debeer, L.J. (1982) Biochem. J. 204, 17-23. 6. Van Broekhoven, A., Peeters, M.C., Debeer, L.J., and Mannaerts, G.P. (1981) Biochem. B i o p h y s . Res. Commun. 100, 305-312. 7. B o r s t , P. (1986) B i o c h i m . B i o p h y s . A e t a 866, 179-203. 8. Laemmli, V.K. (1970) N a t u r e 227, 680-685. 9. Bonner, W.M., and Laskey, R.A. (1974) Eur. J. Biochem. 46, 83-88. 10. Veloso, D., and Veech, R.L. (1975) in : Methods in E r ~ l o s y (Colowick, B.P., and Kaplan, N.O., eds.) vol. 35, pp. 273-278, Academic Press, New York. 1200

Vol. 139, No, 3, 1986

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

11. Declereq, P.E., Haagsman, H.P., Van Veldhoven, P., Debeer, L.J., Van Golde, L.M.G., and Mannaerts, G.P. (1984) J. Biol. Chem. 259, 9064-9075. 12. Van Veldhoven, P., and Mannaerts, G.P. (1985) Bioehem. J. 227, 737-741. 13. Miyazawa, S., Osumi, T., and Hashimoto, T. (1980) Eur. J. Biochem. 103, 589-596. 14. Miyazawa, S., Ozasa, H., Osumi, T., and Hashimoto, T. (1983) J. Bioehem. 94, 529-542. 15. Furata, S., Miyazawa, S., Osumi, T., Hashimoto, T. and Li, N. (1980) J. Bioehem. BB, 1059-1070. 16. McEwen, C.R. (1967) Anal. Biochem. 20, 114-119. 17. Miyazawa, S., Furata, S., Osumi, T., Hashimoto, T., and Li, N. (I781) J. Bioehem. 90, 511-517. 18. Solomon, F., and Jencks, W.P. (1969) J. Biol. ~hem. 244, 1079-1081.

1201