Eur. J. Biochem. 18Y, 539-546 (1990) 0FEBS 1990
Purification and properties of carnitine acetyltransferase from human liver Wilma BLOISI Irma COLOMBO ’, Barbara GARAVAGLIA Roberto GIARDINI ’, Gaetano FINOCCHIARO I,
I,
and Stefano DIDONATO Istituto Nazionale Neurologico C. Besta, Dipartimento di Biochimica e Genetica, Milano, Italy
’ Dipartimento di Fisiologia Generale e Biochimica, Facolti di Farmacia, Universita di Milano, Italy Istituto Nazionale per lo Studio e la Cura dei Tumori, Dipartimento di Patologia, Milano, Italy (Received October 12, 1989/January 4,1990) - EJB 89 1235
Carnitine acetyltransferase was purified from the supernatant obtained after centrifugation of human liver homogenate to a final specific activity of 78.75 unit . mg-’ with acetyl-CoA as a substrate. Human carnitine acetyltransferase is a monomer of 60.5 kDa with maximum activity in the presence of propionyl-CoA and a pH optimum of 8.7. Apparent K , values for acetyl-CoA are three times lower than for decanoyl-CoA. K,,, values for L-carnitine in the presence of acetyl-CoA are six times lower than in the presence of decdnoyl-CoA. K, values for acetylcarnitine are three times lower than for octanoylcarnitine. The polyclonal antibodies against human carnitine acetyltransferase recognize a 60.5-kDa peptide in the purified preparation of human liver and brain homogenates and in immunoblots of mitochondrial and peroxisomal fractions from human liver. Immunoprecipitation and SDSjPAGE analysis of 35S-labelledproteins produced by human fibroblasts indicate that mitochondrial carnitine acetyltransferase is synthesized as a precursor of 65 kDa. We also purified carnitine acetyltransferase from the pellet obtained after centrifugation of liver homogenate. The pellet was extracted by sonication in the presence of 0.5% Tween 20. The chromatographic procedures for the purification and the kinetic, physical and immunological properties of pellet-extracted carnitine acetyltransferase are similar to those of carnitine acetyltransferase purified from the supernatant of human liver homogenate. The enzyme carnitine acetyltransferase catalyzes the reversible transfer of short-chain acyl groups between CoA and carnitine so that these groups can cross intracellular membranes 11, 21; for a review see [3]. Carnitine acetyltransferase was purified from different species [4 - 111. This enzyme shows high activity in rat heart, brown adipose and testicular tissues [12] and has also been detected in the nervous system 113,141,where it was suggested to be a carrier of acetyl-CoA to the extra-mitochondria1 space for choline acetylation [15]. On the other hand, the uniform distribution of carnitine acetyltransferase in the nervous system [14] rather indicates a general role for this enzyme and recent evidence suggests that 2-oxaloacetate is the carrier of acetyl groups for acetylcholine biosynthesis [16]. Carnitine acetyltransferase activity has been found to be primarily associated with the inner mitochondrial membrane, but also in the soluble matrix of peroxisomes and in microsomes from the endoplasmic reticulum, where the enzyme is tightly membrane bound 117, 181. Deficiency in carnitine acetyltransferase activity was present in the brain and other tissues of a patient suffering from fatal ataxic encephalopathy 1191 and also, in the bydin, heart and kidney of a baby with poor respiration, failure to thrive and low levels of esterified carnitine in urine 1201. Valproic Correspondence to G. Finocchiaro, Istituto Nazionale Neurologico, Dipartimento di Biochimica e Genetica, via Celoria 11, I20133 Milano, Italy Abbreviations. Nbs’, 5,S’-dithiobis(2-nitrobenzoicacid). Enzymes. Carnitine acctyltransferase (EC 2.3.1.7); catalase (EC 1.11.1.6); glutamate dehydrogenase (EC 1.4.1.3).
acid, an anti-epileptic drug which may trigger fatal hepatic encephalopathy in children 1211 induces an eightfold increase in carnitine acetyltransferase activity [22], suggesting that this activity should be investigated in patients with Reye-like metabolic crises. In order to study the possible role of carnitine acetyltransferase in these disorders it is necessary to know the characteristics of human carnitine acetyltransferase. In this paper we report the purification of carnitine acetyltransferase from human liver, the preparation of polyclonal antibodies against this enzyme, its kinetic, physical and immunological properties and experiments regarding its subcellular localization. EXPERIMENTAL PROCEDURES
Materials Acyl-CoA esters, acylcarnitines, Ponceau S, phenylmethylsulphonyl fluoride, Percoll and molecular mass markers for gel filtration were obtained from Sigma (Saint Louis, USA). 5,5’-Dithiobis(2-nitrobenzoicacid) (Nbs,) was from BDH. L-Carnitine was a gift from Sigma Tau (Rome, Ita1y ). Proteins used as molecular mass markers in SDSjPAGE were from Pharmacia AB (Uppsala, Sweden) or from BioRad Laboratories (Richmond, California). DEAE-Sepharose CL-6B, Agarose-hexane-CoA and Sephacryl S-300 were supplied by Pharmacia. Matrex gel blue A was obtained from Amicon Corp. (Danvers, USA). Hydroxylapatite, SDS/ PAGE standards, goat anti-(rabbit IgG) antibodies conjugated with alkaline phosphatase, 5-bromo-4-chloro-3-indolyl
540 phosphate p-toluidine salt, p-nitroblue tetrazolium chloride and nitrocellulose were obtained from Bio-Rad Laboratories. Bicinchoninic acid protein assay reagent was purchased from Pierce (Oud-Bejerland, the Netherlands). Ampholine polyacrylamide gel plate, pH 3.5-9.5 was supplied by LKB (Bromma, Sweden). Nycodenz was supplied by Accurate Chemicals and Scientific Corp. (Westbury, USA). L-[~%]Methionine was obtained from Amersham (Buckinghamshire, England), Rhodamine 6G from Kodak Co. (Rochester, USA) and staphylococcal A cells from BRL (Rockville, USA). Enzyme assays Carnitine acetyltransferase was assayed in the forward reaction (formation of acetylcarnitine) by following the production of CoASH, which forms a yellow covalent adduct with Nbs,. The reaction was monitored spectrophotometrically at 412 nm at 37°C. The reaction volume contained 0.1 M Tris/ HC1, pH 8.0, 0.1 mM EDTA, 0.1 % Tween 20, 0.2 mM Nbsz, 0.1 mM acetyl-CoA and 1 mM L-carnitine in 0.5 ml. The reverse reaction (formation of acetyl-CoA) was assayed by spectrophotometer at 233 nm [23] at 25 "C employing 500 pM acetyicarnitine and 190 pM CoA. K,,, profiles with different substrates were assayed by varying the concentration of acyl-CoAs (forward reaction) and acylcarnitines (reverse reaction). One unit of enzyme activity corresponds to the formation of 1 pmol acylcarnitine or acyl-CoA . min- . mg protein- '. For the determination of pH optima, a 0.1 M KH,PO,/ Tris/glycine buffer, adjusted to the desired pH, was used in 0.5 ml 0.1 mM EDTA, 0.1% Tween/20, 0.2 mM Nbs2, 0.1 mM acetyl-CoA. Glutamate dehydrogenase and catalase were marker enzymes for mitochondria and peroxisomes, respectively. Their activity was determined following the procedures described in [24, 251.
'
Protein assays Protein concentration was determined by using bicinchoninic acid protein assay reagent according to the manufacturer's instructions. Preparation of homogenates from human liver and brain Human liver was obtained from autopsy 24 h after death and stored at -80°C. After thawing, the liver was homogenized with Ultra-Turrax in 4 vol. (massjvol.) 20 mM Tris/ HC1, pH 9.4, 5 mM EDTA, 0.5 mM phenylmethylsulphonyl fluoride (Buffer A) at 4°C. A further homogenization was obtained by four or five strokes of a Teflon pestle in a glass potter (Glenco). Human brain biopsy was obtained for diagnostic purposes from the front temporal cortex of a patient with a neoplastic lesion. The sample was homogenized by glass/glass potter in 0.1 M Tris/HCl pH 7.5, 0.1% Tween 20. The homogenate was spun for 5 min at 14000xg and the supernatant (5.5 mg . ml-') was used for immunoblotting. Purijication of human carnitine acetyltransferase The homogenate obtained from 69 g human liver from an autopsy was centrifuged for 1 h at 40000xg at 4°C (J. A. 17 rotor, Beckman). Ammonium sulphate was added to the supernatant to a final concentration of 40%, the mixture was
stirred for 30 min and then centrifuged at 10000 x g for 30 min at 4 "C. The pellet was discarded and ammonium sulphate was added to the supernatant to 60% final concentration. After centrifugation at 10000 x g for 30 min at 4"C, the pellet was resuspended in 220-ml20 mM Tris/HCl pH 9.4,5 mM EDTA and dialyzed against the same buffer until the Nessler reaction became negative. The sample was then applied to a DEAE-Sepharose column (2.6 cm x 35 cm) and eluted with 500 ml of a 0 - 0.4 M lincar gradient of NaCl in 10 mM potassium phosphate pH 8.0,0.5 mM EDTA (buffer B) at a flow rate of 30 ml/h. Fractions 46 - 60 (7 ml each) were collected, dialyzed against 8 1 buffer B, diluted to 500 ml and applied to a Matrex gel blue A column (1.5 cm x 38 cm). A 0 - 1-M linear gradient of NaCl in 350 ml buffer B was used for the elution at a flow rate of 20 ml/h. Fractions 55 -74 (4 ml each) were pooled and dialyzed against 10 mM potassium phosphate, pH 8.0, diluted to 95 ml and applied to a hydroxylapatite column (1 cm x 18 cm). Fractions 55 - 72 (2 ml each), eluted with a 0.01 - 0.4 M linear gradient of 160 ml potassium phosphate pH 8.0 at a flow rate of 10 ml/h, were pooled, diluted to 400 ml and applied to an agarose-hexane-CoA column (1 cm x 8 cm). A 0-0.8-M linear gradient of NaCl in 100 ml buffer B was carried out at a flow rate of 20 ml/h. Fractions 23 - 30 (2 ml each) contained purified carnitine acetyltransferase and were pooled, desalted and concentrated to a final volume of 1 ml. Determination of native and subunit molecular mass The native molecular mass was determined by gel filtration using a Sephacryl S-300 column (0.5 cm x 100 cm) equilibrated with 10 mM potassium phosphate pH 8.0, 0.1 M NaCl with molecular mass markers as references. The subunit molecular mass was determined by SDSjPAGE on a 10% slab gel [26], by comparison with standard proteins of high and low molecular mass. Gels (16 cm x 20 cm regular gels, 8 cm x 10 cm minigels, 0.75 mm thick) werc stained with 0.25% (mass/vol.) Coomassie blue and destained with 10% (by vol.) acetic acid/ 40% (by vol.) methanol solution. Determination of the isoelectric point The isoelectric point of carnitine acetyltransferase was determined using Ampholine polyacrylamide gel plates 3, pH 3.5 - 9.5, according to the manufacturer's instructions (LKB). Nbs, inactivation 0.22 pg purified soluble carnitine acetyltransferase were incubated at room temperature for different intervals (15 s, 30 s, 60 s and 75 s) with 0.2 mM Nbsz in 10 mM Tris/HCl, pH 8.0. The forward reaction was assayed by adding the substrates as described above. Inactivation by Ca2+ Different concentrations (mM) of CaCl, (2.5, 5, 10, 20, 30, 60, 100) were added to the reaction mixture containing 15 pg purified carnitine acetyltransferase. Enzyme activity was assayed as described above, with acetyl-CoA as a substrate. Production of antibodies Antibodies against human carnitine acetyltransferase were raised in the rabbit (New Zealand white) by intradermal injec-
54 1 Table 1. Summary of the purification steps of carnitine ace1yltransfera.wf r o m human liver Human liver from autopsy (69 g) was homogenized as described under Experimental Procedures and centrifuged at 40000 x g for 1 h. Supernatant was used Tor purification. Carnitine acetyltransferase activity was measured using 0.1 m M acetyl-CoA as a substrate. It was not possible to assay carnitine acetyltransferase activity in the total homogenate because it was too concentrated. Therefore purification and yield were computed taking supernatant as 1.0 and 100.0, respectively. Units are defined as pmol acetylcarnitine produced . min-' protein- ' ~~
Step
Total activity
Total protein
Specific activity
Purification
Yield
Supernatant Ammonium sulphate (40- 60%) DEAE-Sepharose CL-6B Matrex gel blue A Hydroxylapatite Agarose-Hexane-CoA
unit 267 225 210 190 117 63
mg 9200.0 4300.0 579.0 37.0 10.0 0.8
unit. mg-' 0.03 0.05 0.36 5.14 11.70 78.75
-fold 1 2 12 171 390 2623
Yo 100.0 84.0 78.7 71 . I 43.8 23.6
~
tion of 0.5 mg purified enzyme, diluted with an equal volume of Freund's complete adjuvant. A booster was given 10 days later with 0.5 mg carnitine acetyltransferase in Freund's incomplete adjuvant. A second booster was injected 10 days later. Immunotitration
Different concentrations of anti-(carnitine acetyltransferase) antibodies were incubated for 1 h at room temperature with 18 pg purified carnitine acetyltransferase. Samples were spun for 10 min at 14000 x g and the supernatants were used for enzyme assay, as described above. Iinmunoblo t ting
After electrophoresis, the gel was blotted overnight onto nitrocellulose filters in 25 inM Tris/70 mM glycine at 150 mA. Antiserum was diluted 100-fold in 50 mM Tris/HCl pH 8.0, 150 mM NaC1, 5% (mass/vol.) non-fat dry milk. Goat anti(rabbit IgG) antibodies conjugated with alkaline phosphatase were used for detection, according to the manufacturer's instructions (Bio-Rad). Labelling of fibroblasts in culture and immunoprecipitation
Fibroblasts were grown in Eagle's minimum essential medium containing 10% foetal calf serum and non-essential amino acids, at 37°C in 75-cm2 flasks. The confluent monolayers were preincubated with 10 ml labelling medium (60% Puck's saline F, 15% dialyzed foetal calf serum and 10% 0.5 M glucose) for 1 h at 37°C. Fibroblasts were then incubated with 7 nil labelling medium containing 100 pCi [35S]methioninefor 1 h a t 37"C, washed twice withphosphatebuffered saline and harvested with 3 ml NaCl/EDTA/Triton/ HCl/Met. Rhodamine 6G 6.3 pM, when used, was added to the labelling medium 30 min after preincubation and incubated for an additional 30 min. The cell lysate was subjected to immunoprecipitation using 7.5 pl rabbit anti-(carnitine acetyltransferase) antiserum according to lkeda et al. [27]. Purijication of carnitine ucetyltransferase from the membrane-extracted fraction of human liver
The homogenate obtained from liver taken from autopsy was centrifuged for 1 h at 40000 x g (J. A. 17 rotor, Beckman). The pellet was sonicated twice for 30 s in 80 ml buffer A
Fig. 1. Polyacrylamide gel electrophoresis of purified carnitine acetyltransferase. Carnitine acetyltransferase purified from the supernatant obtained after 40000 x g centrifugation of human liver or from the pellet extracted with 0.5% Tween 20 and sonication (approximately 1.5 pg and 2.5 pg each) were analyzed by SDSjPAGE on a 10% slab gel. Protein were subsequently stained by Coomassie blue. The molecular mass of the standards (kDa) (Bio-Rad) is indicated on the right. (A) Pellet-extracted carnitine acetyltransferase. (B) Supernatant carnitine acetyltransferase
containing0.5% Tween 20. After dilution to 200 ml the sample was centrifuged at 40000 x g for 30 min at 4°C. The supernatant was subjected to the chromatographic procedures described above, with 0.2% Tween 20 present in all the buffers.
RESULTS AND DISCUSSION Purification and physical properties of carnitine acetyltransferase
Carnitine acetyltransferase was purified to homogeneity from the supernatant of human liver homogenate as described under Experimental Procedures. The procedure is highly reproducible and the purification steps are summarized in Table 1. The final carnitine acetyltransferase, with a specific activity of 78.75 unit. mg-', was purified 2623-fold over the activity detected in the 40000 x g supernatant of human liver homogenate. The yield was 23.6% of the original amount. The enzyme was quite stable during purification and when
542 Table 2. Amino acid composition of human liver carnitine acetyltransferase loo[
Amino acid
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Valine Cysteine Methionine Isoleucine Leucine Tyrosine Phenylalanine Lysine Histidine Arginine Tryptophan
Amino acid/enzyme mol/mol 48 28 36 63 33 38 42 31 -
16 25 55 20 22 36 18 25 -
stored in 50% glycerol at -20°C maintained its activity unaltered for several months. The native molecular mass of carnitine acetyltransferase was estimated by Sephacryl S-300 gel filtration using pamylase (200 kDa), alcohol dehydrogenase (1 50 kDa), albumin (67 kDa) and carbonic anhydrase (29 kDa) as standards. Carnitine acetyltransferase peak activity corresponded to a molecular mass of 60 kDa (data not shown). Determination of subunit molecular mass was obtained by SDSjPAGE and Coomassie staining of the final enzyme preparation. Fig. 1 shows a single protein band corresponding to carnitine acetyltransferase purified from human liver homogenate supernatant (lane 2). The calculated apparent molecular mass of the subunit was 60.5 kDa. These results indicate that carnitine acetyltransferase is a monomer, similar to mouse liver carnitine acetyltransferase [9]. Carnitine acetyltransferase was previously proposed to have a heterodimeric structure [7] but Miyazawa et al. demonstrated that the dimer is the consequence of proteolytic degradation during purification [8]. We used the proteolytic inhibitor, phenylmethylsulphony1 fluoride, during human liver homogenization and ammonium sulphate fractionation which significantly limited the degradation of carnitine acetyltransferase. The molecular mass of carnitine acetyltransferase from other species was similar: the enzyme from pigeon breast muscle was about 55 kDa [4], the beef heart enzyme was 62 kDa [lo], while mitochondria1 and peroxisomal carnitine acetyltransferase from Cundidu tropicalis were 60 kDa and 79 kDa, respectively [I I]. The PI value for human carnitine acetyltransferase is 6.3. Mouse carnitine acetyltransferase has a PI of 6.8 [9] while alkaline PI values were obtained in the other species tested (see Table 2 in [3]). The amino acid composition of human carnitine acetyltransferase is shown in Table 2. Compared with rat liver carnitine acetyltransferase [8], human carnitine acetyltransferase has a higher number of glycine residues (38 versus 28) and a lower number of alanine residues (42 versus 73). Human carnitine acetyltransferase also has a more aspartic acid residues (48 versus 45) and a fewer residues of glutamic acid,
1
C2
C3
I
I
I
I
Cb C 6 C8 C10 C12 Acyl- CoA chain length
-
m u
C1L
C16
Fig. 2. Substrate specificity of human carnitine acetyltransferase. AcylCoAs with carbon-chain length ranging over C2-16 (0.1 mM) were used as substrate. Enzyme activity was determined as described under Experimental Procedures
Table 3. Apparent Michuelis constants of human carnitine acetyltransferase K,,, values for acyl-CoAs were determined using a final concentration of 1 mM L-carnitine and 10-200 pM acyl-CoA. When determining K,,, values for L-carnitine a final concentration of 100 pM acyl-CoA and 0.1 -2 mM L-carnitine were used. Assays were performed as described under Experimental Procedures
acyl-CoA
L-carnitine
PM Acetyl-CoA Propionyl-CoA Butyryl-CoA Hexanoyl-CoA Octanoyl-CoA Decanoyl-CoA
21.3 28.0 43.5 54.9 50.3 64.0
91 86 152 120 148 580
lysine, histidine and arginine (63 versus 73, 36 versus 49, 18 versus 19 and 25 versus 27, respectively). These differences may account for the more acidic isoelectric point of human carnitine acetyltransferase compared to that of rat carnitine acetyltransferase [28]. Catalytic properties The substrate specificity of carnitine acetyltransferase (shown in Fig. 2) demonstrates that the enzyme is very active with acetyl-CoA, propionyl-CoA and butyryl-CoA. The maximum activity is obtained with propionyl-CoA, 30% of the peak activity is achieved with octanoyl-CoA while palmitoyl-CoA is not a substrate. Maximum carnitine acetyltransferase activity with propionyl-CoA or butyryl-CoA was detected in pig heart [2] and rat and mouse liver [8,9], respectively. In the other species tested, the activity is strongly reduced with octanoyl-CoA and absent with long-chain acylCoAs [3]. The apparent Michaelis constants of carnitine acetyltransferase for different substrates are listed in Table 3. Values for both acyl-CoAs and carnitine were determined with saturating levels of the other substrates. K, values are low for short-
543
Antibody (PIof antiserum)
Fig. 3. Carnitine acetyltransferase titration by antibodies. Different concentration of anti-(carnitine acetyltransferase) antibodies were incubated with 18 pg purified carnitine acetyltransferase for 1 h at room temperature. Samples were spun for 10min at 14000xg and the supernatant was assayed for carnitine acetyltransferase activity as described under Experimental Procedures
chain acyl-CoAs and increase with increasing carbon-chain length. These values follow the same profile as those reported by Farrell et al. [9]. K, for L-carnitine is six times higher in the presence of decanoyl-CoA than in the presence of acetylCoA. This implies that short-chain acyl-CoAs are preferential substrates when the concentration of L-carnitine is low. We also determined K, of human carnitine acetyltransferase for acetylcarnitine, propionylcarnitine and octanoylcarnitine. The lowest K,,, (0.42 mM) was obtained for acetyltransferase; a value of 0.65 inM was obtained for propionylcarnitine while the highest K, (1.39 mM) was obtained for octanoylcarnitine. V,,, for acetyl-CoA and propionyl-CoA (forward reaction) and for acetylcarnitine and propionylcarnitine (reverse reaction) were also determined in a sample of partially purified carnitine acetyltransferase. The values for acetyl-CoA and propionyl-CoA were 48.2 pmol . mg-I and 79 pmol . mg-l, respectively, while with acetylcarnitine and propionylcarnitine the activity was 27.1 pmol . mg-I and 23.6 pmol . mg-'. Both K , and V,,, values suggest that the forward reactionpredominates in vivo and that carnitine acetyltransferase acts preferentially on the formation of acetylcarnitine from intramitochondrial acetyl-CoA. The pH optimum for carnitine acetyltransferase activity was determined by assaying acetylcarnitine formation (forward reaction) with a pH range over 6.0 - 10.0. The pH optima for carnitine acetyltransferase are in the alkaline range with a peak at pH 8.7. A similar value was found in pig heart carnitine acetyltransferase while pH optima in other species were lower (see Table 2 in [3]). In contrast with mouse peroxisomal carnitine acetyltransferase activity, which is not affected by 10 mM Ca2+ [9], human carnitine acetyltransferase was inhibited to 85.5%, a difference which could be partly related to the different pH values used in the assay. The activity decreased by increasing the concentration of Ca2+to 12.8% in the presence of 100 mM Ca2 . +
Inzmunochemicalstudies
Polyclonal antibodies were obtained by immunizing a New Zealand white rabbit with purified carnitine acetyltransferase,
Fig. 4. lmmunohlotting of' purified carnitine acetyltransferase. After blotting onto nitrocellulose filter standard proteins were stained with amido black (lane 1) while purified carnitine acetyltransferase (8.7 pg) was matched with anti-(carnitine acetyltransferase) antiserum 100fold diluted (lane 2)
Fig. 5. Immunoblotting of purified curnitine acetyltranjferase and human brain homogenate. The human brain homogenate was prepared as described under Experimental Procedures. The supernatant obtained after centrifugation at 16000xg was run together with purified carnitine acetyltransferase on a 9% polyacrylamide minigel. Reaction with anti-(carnitine acetyltransferase) antibodies was performed as described previously. (1) Purified carnitine acetyltransferase (0.9 pg); (2) human brain homogenate (1 10 pg). The arrowhead indicates the 60.5-kDa band in human brain homogenate
as described under Experimental Procedures. The titer of the antiserum was first detected by immunoprecipitation experiments. Fig. 3 shows that the activity of 18 pg carnitine acetyltransferase decreases with increasing antiserum concentration and is abolished after incubation with 20 p1 antiserum. This antiserum has been used for immunoblotting and immunoprecipitation studies on different samples. As shown in Fig. 4, lane 2 , anti-(carnitine acetyltransferase) antiserum, 100-fold diluted, identifies a 60.5-kDa band when purified carnitine acetyltransferase is loaded onto the gel. The supernatant obtained after centrifuging human brain homogenate was also subjected to Western blot analysis. Anti-(carnitine acetyltransferase) antibodies recognized one band in the supernatant (Fig. 5, lane 2) with identical molecular mass to
544
1
2
Table 4. Enzyme specific activities in total homogenate and subcellular fractions of human liver 10 g human liver were used for subcellular fractionation following methods described by Ramsay et al. (1987) and Ghosh and Haira (1 986). Total homogenate, mitochondria and peroxisomes were treated with 0.5% Tween 20 before activities were assayed
3
97kDa
Fraction
glutamate de- catalase hydrogenase
68 kDa
43 kDa
Fig. 6. Polyacrylamide gel electrophoresis of13 5SJmethionine-lahelled curnitine ucetyltransferase synthesized in normal cultured human fibroblasts. Confluent monolayers of fibroblast were labelled with [35S]mcthionine in the presence and absence of Rhodamine 6G and cell extracts were subjected to immunoprecipitation with anti-(carnitine acetyltransferase) antibodies (7.5 PI). The immunoprecipitates were analyied on 10% SDSjPAGE as described under Experimental Procedures. The precursor and the mature form of carnitine acetyltransferase are shown in lanes 1 and 2. respectively. Standard proteins are in lane 3
purified carnitine acetyltransferase (lane 1). Since carnitine acetyltransferase activity is present in the nervous system [13, 14, 19, 201 this 60.5-kDa band very likely represents human brain carnitine acetyltransferase. We also labelled human fibroblasts with [35S]methionine in the presence or absence of Rhodamine 6G, an inhibitor of mitochondrial enzyme processing [29]. The labelled products were immunoprecipitated with anti-(carnitine acetyltransferase) antibodies and protein A from Staphylococcus aureus. After SDSjPAGE on a 10% slab gel, one radioactive band corresponding to 65 kDa was visible in the sample treated with Rhodamine 6G (Fig. 6, lane 1). This precursor is 4.5kDa larger than mature carnitine acetyltransferase (60.5-kDa) which is clearly detectable in the sample without Rhodamine 6G (Fig. 6, lane 2). This evidence indicates that antibodies against purified carnitine acetyltransferase recognize the mitochondrial form of this enzyme. Indeed, most of the mitochondrial proteins encoded in the nucleus are synthesized as cytoplasmic precursor of larger molecular mass and converted in the mitochondria into the mature form [30].
Carnitine acetyltransjerase activity in human mitochondrial and peroxisomes Since a peroxisomal carnitine acetyltransferase with properties similar to human carnitine acetyltransferase has already been described [9], we investigated the existence of peroxisomal carnitine acetyltransferase in human livcr. Starting from a sample of human liver obtained about 2 h
Specific activity of
Homogenate Mitochondria Peroxisomes
1.10 0.065 2.96 f 0.125 0.52 0.030
*
*
0.10 0.009 0.01 k 0.000 2.31 & 0.511
carnitine acetyltransferase
0.03 k 0.003 0.06 0.002 0.14 f 0.012
Fig. 7. Immunohlotting of total homogenate and peroxisomal und mifochondrial fractions f r o m human liver. Peroxisomes and mitochondria were obtained from fresh human liver as described under Experimental Procedures. Total homogenate, mitochondria, peroxisomes and pellet-extracted carnitine acetyltransferase were electrophoresed on a 10% polyacrylamide gel, blotted onto a nitrocellulose filter and incubated with anti-(carnitine acetyltransferase) antibodies diluted 100 times. ( 1 ) Total homogenate (456 pg proteins); (2) mitochondrial fraction (199 pg proteins); (3) peroxisomal fraction (88 pg proteins). Purified carnitine acetyltransferase (lane 4)is indicated by the arrowhead
after surgery, we prepared enriched fractions of mitochondria and peroxisomes by the method described above. Table 4 shows the specific activity of marker enzymes and carnitine acetyltransferase in liver homogenate, mitochondria and peroxisomes. The specific activity of catalase increased 23-fold in peroxisomes when compared with that in the homogenate. The specific activity of glutamate dehydrogenase increased 2.7-fold in mitochondria compared to that in the homogenate. Carnitine acetyltransferase activity was observed both in mitochondria and in peroxisomes. The specific activities of mitochondria and peroxisomes are 2-fold and 4.7fold higher, respectively, than that in the homogenate. Total homogenate, mitochondria and peroxisomes were subjected to immunoblot analysis using anti-(carnitine acetyltransferase) antibodies (Fig. 7). A 60.5-kDa band ofmolecular mass identical to carnitine acetyltransferase is visible in the three preparations and a band 2-kDa larger than carnitine acetyltransferase, barely visible in the homogenate, is enriched in peroxisomes. Miyazawa et al. co-purified a peptide of molecular mass slightly higher than rat liver carnitine acetyltransferase which is suggested to be the precursor of mitochondrial carnitine acetyltransferase [XI. We incubated 90 pg of the peroxisomal preparation for 30 min at 25 C with 200 pg of frozen-thawed mitochondria
545
Fig. 8. Immunohlot of pellet-extracted carnitine acetyltransferase. Pellet-extracted carnitine acetyltransferase (2 pg) was electrophoresed on a 10% slab gel, blotted onto a nitrocellulose filter and incubated with anti-(carnitine acetyltransferase) antibodies diluted 100 times as described under Experimental Procedures
in sucrose buffer. When the sample was analysed by 10% SDSjPAGE the 62.5-kDa peptide was not modified (not shown). This suggests that the peptide we observed is not the precursor of human mitochondrial carnitine acetyltransferase, in agreement with the observation that such a precursor in fibroblasts is 2 kDa larger than 62.5 kDa. Interestingly, in the peroxisomal fraction we found identical octanoyl-CoA and acetyl-CoA activities, both inhibited by anti-(carnitine acetyltransferase) antibodies (data not shown). Since the activity of purified carnitine acetyltransferase with octanoyl-CoA is only 50% of the activity with acetyl-CoA (see Fig. 2), the residual octanoyl-CoA activity in peroxisomes could be attributed to human carnitine octanoyltransferase, a peroxisomal enzyme characterized in rat and mouse liver [9, 311. Therefore, it is possible that the 62.5-kDa peptide recognized by anti-(carnitine acetyltransferase) antibodies in peroxisomes is carnitine octanoyltransferase. A large 67-kDa band, visible in the human liver homogenate and faintly in mitochondria and peroxisomes, disappeared after incubation of antibodies with albumin (not shown). This indicates that albumin, although not visible (see Figs 1 and 4, Fig. 5, lane 1 and Fig. 7, lane 4) is present in the preparation of purified carnitine acetyltransferase and detectable by anti-(carnitine acetyltransferase) antibodies in human liver homogenate, but not in fibroblast homogenate (see Fig. 6, lanes 1 and 2). Another faint band of molecular mass slightly larger than albumin is visible in immunoblots of peroxisomes (Fig. 7, lane 3); the identity of this band is unknown. Other bands of smaller molecular mass, especially in the hornogenate of human liver, are likely to be degradation products of carnitine acetyltransferase due to the action of proteases and to repeated freezing and thawing of the sample. Purification and properties of membrane-extracted carnitine acetyltransferase ,from human liver
Markwell et al. showed that in rat Liver approximately 52% of carnitine acetyltransferase activity is associated with
the mitochondrial fraction, 14% with the peroxisomal fraction, and 34% is bound to a lipid-rich membranous fraction, partially derived from microsomes [17]. Mitochondria1 carnitine acetyltransferase was originally located in the interniembrane space and in the matrix, according to subfractionation experiments performed with digitonin in rat liver mitochondria 132) but Edwards et al. suggested that carnitine acetyltransferase only bound to the inner mitochondrial membrane [ 181. We observed that the 40000 x g pellet from human liver homogenate, extracted by sonication in the presence of 0.5% Tween 20, contained carnitine acetyltransferase activity. The extract was subjected to the same chromatographic procedures as used for carnitine acetyltransferase purified from the supernatant of human liver hornogenate (see Experimental Procedures). The final carnitine acetyltransferase activity with acetyl-CoA as a substrate, was 72 unit . mg-' with a purification of 2889-fold over the pellet extract (data not shown). The final preparation electrophoresed in a polyacrylamide gel, and stained with Coomassie blue is shown in Fig. 1, lane A. Carnitine acetyltransferase from the pellet and from the supernatant have identical molecular masses. The substrate specificity of membrane-extracted carnitine acetyltransferase is identical to the substrate specificity of carnitine acetyltransferase from the supernatant, with maximum activity for propionyl-CoA (data not shown). Michaelis constants determined either for acyl-CoAs or L-carnitine, as well as pH optima and pl values (data not shown) are all simiiar to those obtained for carnitine acetyltransferase purified from the human liver supernatant. The final preparation of membrane-extracted carnitine acetyltransferase is identified by immunoblot with carnitine acetyltransferase antiserum (Fig. 8) and its activity is also inhibited after incubation with 25 p1 antiserum (data not shown). Therefore, carnitine acetyltransferase purified from the supernatant and the pellet from human liver homogenates share a considerable degree of similarity and further studies will be needed to understand the molecular mechanisms of their subcellular localization. The amino acid composition of carnitine acetyltransferasc was determined by K. Williams and K. Stone (Department of Molecular Biophysics and Biochemistry, Yale University Medical School, New Haven CT, USA). This work has been partially supported by a grant from the Muscular Dystrophy Association for the project: Molecular cloning of human carnitine acyltransferases. I. C. was supported by a grant from the Department of General Physiology and Biological Chemistry, Milan University, School of Pharmacology.
REFERENCES 1. Friedman, S. & Fraenkel, G. (1955) Arch. Biochem. Biophjs. 59,
491 - 501. 2. Fritz, I. B., Schultz, S. K. & Srere, P. A. (1963) J . Biol. Chem. 238,2509-2517. 3. Colucci, W. I. & Gandour, R. D. (1988) Bioorg. Chem 16, 307334. 4. Chase, J . F. A., Pearson, D. 3 . & Tubbs, P. K. (1965) Biochim. Bioph-vs. Actu 96, 162 - 165. 5. Chase, J . F. A. & Tubbs, P. K. (1969) Biochem. J . 111,225-235. 6. Markwell, M. A. K. & Bieber, L. L. (1976) Arch. Biochem. Biophys. 172, 502- 509. 7. Mittal, B. & Kurup, C. K. R. (1980) Biochim. Biophys. Actu 619, 90 - 97. 8. Miyazawa, S., Ozasa, H., Furuta, S., Osumi, T. & Hashimoto, T. (1982) J . Biochem. (Tokyo) 93,439-451. 9. Farrell, S. O., Fiol, C . J., Reddy, 1. K. & Bieber, L. L. (1984) J . Bid. Chem. 259,13089-13095.
546 10. Huckle, W. R. & Tamblyn, T. M. (1983) Arch. Biochem. Biophys. 226,94-110. 11. Ueda, M., Tanaka, A. & Fukui, S. (1982) Eur. J . Biochem. 124, 205 -210. 12. Marquis, N. R. & Fritz, I. B. (1965) J. Biol. Chem. 240, 21932196. 13. Fritz, I. B. (1963) Adv. Lipid. Res. I, 285-334. 14. McCaman, R. E., McCaman, M. W. & Stafford, M. L. (1966) J . Biol. Chem. 241,930-934. 15. Dolezal, V. & Tuceck, S. (1981) 1. Neurochem. 36,1323-1330. 16. Willoughby, J., Craig, F. E., Harvey, S. A. K. & Clark, J. €3. (1989) J . Neurochem. 52,896-901. 17. Markwell, M. A. K., McGroarty, E. J., Bieber, L. L. & Tolbert, N. E. (1973) J . Biol. Chem. 248, 3426-3432. 18. Edwards, Y. H., Chase, J. F. A., Edwards, M. R. &Tubbs, P. K . (1974) Eur. J . Biochem. 46, 209-215. 19. DiDonato, S., Rimoldi, M., Moise, A., Bertagnolio, B. & Uziel, G. (1979) Neurology 29, 1578-1583. 20. Przyrembel, H. (1987) J . ZnheritedMetab. Dis. 10, 129-146. 21. Gerber, N., Dickinson, R. G., Harland, R. C., Lynn, R. K., Houghton, D., Antonias, J. I. & Schimschock, J. C. (1979) J . Pediatr. 95, 142- 144. 22. Singh, Y., Liu, G. A. & Krishna, G. (1987) J. Toxicol. Environ. Health 22. 459 - 469.
23. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Meth0d.s ofenzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol 1, p. 439, Verlag Chemie Weinheim, Academic Press, Inc. 24. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Methods ofenzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol. 1, pp. 461 -462, Verlag Chemie Weinheim, Academic Press Inc. 25. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Methods of enzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol 1, pp. 438 -439, Verlag Chemie Weinheim, Academic Press, Inc. 26. Laemmli, U. K . (1 970) Nature 227,680 - 685. 27. Ikeda, Y., Keese, S. M. & Tanaka, K. (1985) Proc. Natl Acad. Sci. U S A 82,7081 -7085. 28. Markwell, M. A. K., Tolbert, N. E. & Bieber, L. L. (1976) Arch. Biochem. Biophys. 176,479-488. 29. Kuzela, S., Joste, V. & Nelson, B. D. (1986) Eur. J . Biochem. (Tokyo) 154, 553 - 557. 30. Hay, R., Boehni, P. & Gasser, S. (1984) Biochim. Biophys. Actu 779, 65 - 87. 31. Miyazawa, S., Ozasa, H., Osumi, T. & Hashimoto, T. (1983) J . Biochem. (Tokyo) 94, 529 - 542. 32. Brdiczka, D., Gerbitz, K. & Pette, D. (1969) Eur. J . Biochem. 11, 234 - 240.
Purification and properties of carnitine acetyltransferase from human liver Wilma BLOISI Irma COLOMBO ’, Barbara GARAVAGLIA Roberto GIARDINI ’, Gaetano FINOCCHIARO I,
I,
and Stefano DIDONATO Istituto Nazionale Neurologico C. Besta, Dipartimento di Biochimica e Genetica, Milano, Italy
’ Dipartimento di Fisiologia Generale e Biochimica, Facolti di Farmacia, Universita di Milano, Italy Istituto Nazionale per lo Studio e la Cura dei Tumori, Dipartimento di Patologia, Milano, Italy (Received October 12, 1989/January 4,1990) - EJB 89 1235
Carnitine acetyltransferase was purified from the supernatant obtained after centrifugation of human liver homogenate to a final specific activity of 78.75 unit . mg-’ with acetyl-CoA as a substrate. Human carnitine acetyltransferase is a monomer of 60.5 kDa with maximum activity in the presence of propionyl-CoA and a pH optimum of 8.7. Apparent K , values for acetyl-CoA are three times lower than for decanoyl-CoA. K,,, values for L-carnitine in the presence of acetyl-CoA are six times lower than in the presence of decdnoyl-CoA. K, values for acetylcarnitine are three times lower than for octanoylcarnitine. The polyclonal antibodies against human carnitine acetyltransferase recognize a 60.5-kDa peptide in the purified preparation of human liver and brain homogenates and in immunoblots of mitochondrial and peroxisomal fractions from human liver. Immunoprecipitation and SDSjPAGE analysis of 35S-labelledproteins produced by human fibroblasts indicate that mitochondrial carnitine acetyltransferase is synthesized as a precursor of 65 kDa. We also purified carnitine acetyltransferase from the pellet obtained after centrifugation of liver homogenate. The pellet was extracted by sonication in the presence of 0.5% Tween 20. The chromatographic procedures for the purification and the kinetic, physical and immunological properties of pellet-extracted carnitine acetyltransferase are similar to those of carnitine acetyltransferase purified from the supernatant of human liver homogenate. The enzyme carnitine acetyltransferase catalyzes the reversible transfer of short-chain acyl groups between CoA and carnitine so that these groups can cross intracellular membranes 11, 21; for a review see [3]. Carnitine acetyltransferase was purified from different species [4 - 111. This enzyme shows high activity in rat heart, brown adipose and testicular tissues [12] and has also been detected in the nervous system 113,141,where it was suggested to be a carrier of acetyl-CoA to the extra-mitochondria1 space for choline acetylation [15]. On the other hand, the uniform distribution of carnitine acetyltransferase in the nervous system [14] rather indicates a general role for this enzyme and recent evidence suggests that 2-oxaloacetate is the carrier of acetyl groups for acetylcholine biosynthesis [16]. Carnitine acetyltransferase activity has been found to be primarily associated with the inner mitochondrial membrane, but also in the soluble matrix of peroxisomes and in microsomes from the endoplasmic reticulum, where the enzyme is tightly membrane bound 117, 181. Deficiency in carnitine acetyltransferase activity was present in the brain and other tissues of a patient suffering from fatal ataxic encephalopathy 1191 and also, in the bydin, heart and kidney of a baby with poor respiration, failure to thrive and low levels of esterified carnitine in urine 1201. Valproic Correspondence to G. Finocchiaro, Istituto Nazionale Neurologico, Dipartimento di Biochimica e Genetica, via Celoria 11, I20133 Milano, Italy Abbreviations. Nbs’, 5,S’-dithiobis(2-nitrobenzoicacid). Enzymes. Carnitine acctyltransferase (EC 2.3.1.7); catalase (EC 1.11.1.6); glutamate dehydrogenase (EC 1.4.1.3).
acid, an anti-epileptic drug which may trigger fatal hepatic encephalopathy in children 1211 induces an eightfold increase in carnitine acetyltransferase activity [22], suggesting that this activity should be investigated in patients with Reye-like metabolic crises. In order to study the possible role of carnitine acetyltransferase in these disorders it is necessary to know the characteristics of human carnitine acetyltransferase. In this paper we report the purification of carnitine acetyltransferase from human liver, the preparation of polyclonal antibodies against this enzyme, its kinetic, physical and immunological properties and experiments regarding its subcellular localization. EXPERIMENTAL PROCEDURES
Materials Acyl-CoA esters, acylcarnitines, Ponceau S, phenylmethylsulphonyl fluoride, Percoll and molecular mass markers for gel filtration were obtained from Sigma (Saint Louis, USA). 5,5’-Dithiobis(2-nitrobenzoicacid) (Nbs,) was from BDH. L-Carnitine was a gift from Sigma Tau (Rome, Ita1y ). Proteins used as molecular mass markers in SDSjPAGE were from Pharmacia AB (Uppsala, Sweden) or from BioRad Laboratories (Richmond, California). DEAE-Sepharose CL-6B, Agarose-hexane-CoA and Sephacryl S-300 were supplied by Pharmacia. Matrex gel blue A was obtained from Amicon Corp. (Danvers, USA). Hydroxylapatite, SDS/ PAGE standards, goat anti-(rabbit IgG) antibodies conjugated with alkaline phosphatase, 5-bromo-4-chloro-3-indolyl
540 phosphate p-toluidine salt, p-nitroblue tetrazolium chloride and nitrocellulose were obtained from Bio-Rad Laboratories. Bicinchoninic acid protein assay reagent was purchased from Pierce (Oud-Bejerland, the Netherlands). Ampholine polyacrylamide gel plate, pH 3.5-9.5 was supplied by LKB (Bromma, Sweden). Nycodenz was supplied by Accurate Chemicals and Scientific Corp. (Westbury, USA). L-[~%]Methionine was obtained from Amersham (Buckinghamshire, England), Rhodamine 6G from Kodak Co. (Rochester, USA) and staphylococcal A cells from BRL (Rockville, USA). Enzyme assays Carnitine acetyltransferase was assayed in the forward reaction (formation of acetylcarnitine) by following the production of CoASH, which forms a yellow covalent adduct with Nbs,. The reaction was monitored spectrophotometrically at 412 nm at 37°C. The reaction volume contained 0.1 M Tris/ HC1, pH 8.0, 0.1 mM EDTA, 0.1 % Tween 20, 0.2 mM Nbsz, 0.1 mM acetyl-CoA and 1 mM L-carnitine in 0.5 ml. The reverse reaction (formation of acetyl-CoA) was assayed by spectrophotometer at 233 nm [23] at 25 "C employing 500 pM acetyicarnitine and 190 pM CoA. K,,, profiles with different substrates were assayed by varying the concentration of acyl-CoAs (forward reaction) and acylcarnitines (reverse reaction). One unit of enzyme activity corresponds to the formation of 1 pmol acylcarnitine or acyl-CoA . min- . mg protein- '. For the determination of pH optima, a 0.1 M KH,PO,/ Tris/glycine buffer, adjusted to the desired pH, was used in 0.5 ml 0.1 mM EDTA, 0.1% Tween/20, 0.2 mM Nbs2, 0.1 mM acetyl-CoA. Glutamate dehydrogenase and catalase were marker enzymes for mitochondria and peroxisomes, respectively. Their activity was determined following the procedures described in [24, 251.
'
Protein assays Protein concentration was determined by using bicinchoninic acid protein assay reagent according to the manufacturer's instructions. Preparation of homogenates from human liver and brain Human liver was obtained from autopsy 24 h after death and stored at -80°C. After thawing, the liver was homogenized with Ultra-Turrax in 4 vol. (massjvol.) 20 mM Tris/ HC1, pH 9.4, 5 mM EDTA, 0.5 mM phenylmethylsulphonyl fluoride (Buffer A) at 4°C. A further homogenization was obtained by four or five strokes of a Teflon pestle in a glass potter (Glenco). Human brain biopsy was obtained for diagnostic purposes from the front temporal cortex of a patient with a neoplastic lesion. The sample was homogenized by glass/glass potter in 0.1 M Tris/HCl pH 7.5, 0.1% Tween 20. The homogenate was spun for 5 min at 14000xg and the supernatant (5.5 mg . ml-') was used for immunoblotting. Purijication of human carnitine acetyltransferase The homogenate obtained from 69 g human liver from an autopsy was centrifuged for 1 h at 40000xg at 4°C (J. A. 17 rotor, Beckman). Ammonium sulphate was added to the supernatant to a final concentration of 40%, the mixture was
stirred for 30 min and then centrifuged at 10000 x g for 30 min at 4 "C. The pellet was discarded and ammonium sulphate was added to the supernatant to 60% final concentration. After centrifugation at 10000 x g for 30 min at 4"C, the pellet was resuspended in 220-ml20 mM Tris/HCl pH 9.4,5 mM EDTA and dialyzed against the same buffer until the Nessler reaction became negative. The sample was then applied to a DEAE-Sepharose column (2.6 cm x 35 cm) and eluted with 500 ml of a 0 - 0.4 M lincar gradient of NaCl in 10 mM potassium phosphate pH 8.0,0.5 mM EDTA (buffer B) at a flow rate of 30 ml/h. Fractions 46 - 60 (7 ml each) were collected, dialyzed against 8 1 buffer B, diluted to 500 ml and applied to a Matrex gel blue A column (1.5 cm x 38 cm). A 0 - 1-M linear gradient of NaCl in 350 ml buffer B was used for the elution at a flow rate of 20 ml/h. Fractions 55 -74 (4 ml each) were pooled and dialyzed against 10 mM potassium phosphate, pH 8.0, diluted to 95 ml and applied to a hydroxylapatite column (1 cm x 18 cm). Fractions 55 - 72 (2 ml each), eluted with a 0.01 - 0.4 M linear gradient of 160 ml potassium phosphate pH 8.0 at a flow rate of 10 ml/h, were pooled, diluted to 400 ml and applied to an agarose-hexane-CoA column (1 cm x 8 cm). A 0-0.8-M linear gradient of NaCl in 100 ml buffer B was carried out at a flow rate of 20 ml/h. Fractions 23 - 30 (2 ml each) contained purified carnitine acetyltransferase and were pooled, desalted and concentrated to a final volume of 1 ml. Determination of native and subunit molecular mass The native molecular mass was determined by gel filtration using a Sephacryl S-300 column (0.5 cm x 100 cm) equilibrated with 10 mM potassium phosphate pH 8.0, 0.1 M NaCl with molecular mass markers as references. The subunit molecular mass was determined by SDSjPAGE on a 10% slab gel [26], by comparison with standard proteins of high and low molecular mass. Gels (16 cm x 20 cm regular gels, 8 cm x 10 cm minigels, 0.75 mm thick) werc stained with 0.25% (mass/vol.) Coomassie blue and destained with 10% (by vol.) acetic acid/ 40% (by vol.) methanol solution. Determination of the isoelectric point The isoelectric point of carnitine acetyltransferase was determined using Ampholine polyacrylamide gel plates 3, pH 3.5 - 9.5, according to the manufacturer's instructions (LKB). Nbs, inactivation 0.22 pg purified soluble carnitine acetyltransferase were incubated at room temperature for different intervals (15 s, 30 s, 60 s and 75 s) with 0.2 mM Nbsz in 10 mM Tris/HCl, pH 8.0. The forward reaction was assayed by adding the substrates as described above. Inactivation by Ca2+ Different concentrations (mM) of CaCl, (2.5, 5, 10, 20, 30, 60, 100) were added to the reaction mixture containing 15 pg purified carnitine acetyltransferase. Enzyme activity was assayed as described above, with acetyl-CoA as a substrate. Production of antibodies Antibodies against human carnitine acetyltransferase were raised in the rabbit (New Zealand white) by intradermal injec-
54 1 Table 1. Summary of the purification steps of carnitine ace1yltransfera.wf r o m human liver Human liver from autopsy (69 g) was homogenized as described under Experimental Procedures and centrifuged at 40000 x g for 1 h. Supernatant was used Tor purification. Carnitine acetyltransferase activity was measured using 0.1 m M acetyl-CoA as a substrate. It was not possible to assay carnitine acetyltransferase activity in the total homogenate because it was too concentrated. Therefore purification and yield were computed taking supernatant as 1.0 and 100.0, respectively. Units are defined as pmol acetylcarnitine produced . min-' protein- ' ~~
Step
Total activity
Total protein
Specific activity
Purification
Yield
Supernatant Ammonium sulphate (40- 60%) DEAE-Sepharose CL-6B Matrex gel blue A Hydroxylapatite Agarose-Hexane-CoA
unit 267 225 210 190 117 63
mg 9200.0 4300.0 579.0 37.0 10.0 0.8
unit. mg-' 0.03 0.05 0.36 5.14 11.70 78.75
-fold 1 2 12 171 390 2623
Yo 100.0 84.0 78.7 71 . I 43.8 23.6
~
tion of 0.5 mg purified enzyme, diluted with an equal volume of Freund's complete adjuvant. A booster was given 10 days later with 0.5 mg carnitine acetyltransferase in Freund's incomplete adjuvant. A second booster was injected 10 days later. Immunotitration
Different concentrations of anti-(carnitine acetyltransferase) antibodies were incubated for 1 h at room temperature with 18 pg purified carnitine acetyltransferase. Samples were spun for 10 min at 14000 x g and the supernatants were used for enzyme assay, as described above. Iinmunoblo t ting
After electrophoresis, the gel was blotted overnight onto nitrocellulose filters in 25 inM Tris/70 mM glycine at 150 mA. Antiserum was diluted 100-fold in 50 mM Tris/HCl pH 8.0, 150 mM NaC1, 5% (mass/vol.) non-fat dry milk. Goat anti(rabbit IgG) antibodies conjugated with alkaline phosphatase were used for detection, according to the manufacturer's instructions (Bio-Rad). Labelling of fibroblasts in culture and immunoprecipitation
Fibroblasts were grown in Eagle's minimum essential medium containing 10% foetal calf serum and non-essential amino acids, at 37°C in 75-cm2 flasks. The confluent monolayers were preincubated with 10 ml labelling medium (60% Puck's saline F, 15% dialyzed foetal calf serum and 10% 0.5 M glucose) for 1 h at 37°C. Fibroblasts were then incubated with 7 nil labelling medium containing 100 pCi [35S]methioninefor 1 h a t 37"C, washed twice withphosphatebuffered saline and harvested with 3 ml NaCl/EDTA/Triton/ HCl/Met. Rhodamine 6G 6.3 pM, when used, was added to the labelling medium 30 min after preincubation and incubated for an additional 30 min. The cell lysate was subjected to immunoprecipitation using 7.5 pl rabbit anti-(carnitine acetyltransferase) antiserum according to lkeda et al. [27]. Purijication of carnitine ucetyltransferase from the membrane-extracted fraction of human liver
The homogenate obtained from liver taken from autopsy was centrifuged for 1 h at 40000 x g (J. A. 17 rotor, Beckman). The pellet was sonicated twice for 30 s in 80 ml buffer A
Fig. 1. Polyacrylamide gel electrophoresis of purified carnitine acetyltransferase. Carnitine acetyltransferase purified from the supernatant obtained after 40000 x g centrifugation of human liver or from the pellet extracted with 0.5% Tween 20 and sonication (approximately 1.5 pg and 2.5 pg each) were analyzed by SDSjPAGE on a 10% slab gel. Protein were subsequently stained by Coomassie blue. The molecular mass of the standards (kDa) (Bio-Rad) is indicated on the right. (A) Pellet-extracted carnitine acetyltransferase. (B) Supernatant carnitine acetyltransferase
containing0.5% Tween 20. After dilution to 200 ml the sample was centrifuged at 40000 x g for 30 min at 4°C. The supernatant was subjected to the chromatographic procedures described above, with 0.2% Tween 20 present in all the buffers.
RESULTS AND DISCUSSION Purification and physical properties of carnitine acetyltransferase
Carnitine acetyltransferase was purified to homogeneity from the supernatant of human liver homogenate as described under Experimental Procedures. The procedure is highly reproducible and the purification steps are summarized in Table 1. The final carnitine acetyltransferase, with a specific activity of 78.75 unit. mg-', was purified 2623-fold over the activity detected in the 40000 x g supernatant of human liver homogenate. The yield was 23.6% of the original amount. The enzyme was quite stable during purification and when
542 Table 2. Amino acid composition of human liver carnitine acetyltransferase loo[
Amino acid
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Valine Cysteine Methionine Isoleucine Leucine Tyrosine Phenylalanine Lysine Histidine Arginine Tryptophan
Amino acid/enzyme mol/mol 48 28 36 63 33 38 42 31 -
16 25 55 20 22 36 18 25 -
stored in 50% glycerol at -20°C maintained its activity unaltered for several months. The native molecular mass of carnitine acetyltransferase was estimated by Sephacryl S-300 gel filtration using pamylase (200 kDa), alcohol dehydrogenase (1 50 kDa), albumin (67 kDa) and carbonic anhydrase (29 kDa) as standards. Carnitine acetyltransferase peak activity corresponded to a molecular mass of 60 kDa (data not shown). Determination of subunit molecular mass was obtained by SDSjPAGE and Coomassie staining of the final enzyme preparation. Fig. 1 shows a single protein band corresponding to carnitine acetyltransferase purified from human liver homogenate supernatant (lane 2). The calculated apparent molecular mass of the subunit was 60.5 kDa. These results indicate that carnitine acetyltransferase is a monomer, similar to mouse liver carnitine acetyltransferase [9]. Carnitine acetyltransferase was previously proposed to have a heterodimeric structure [7] but Miyazawa et al. demonstrated that the dimer is the consequence of proteolytic degradation during purification [8]. We used the proteolytic inhibitor, phenylmethylsulphony1 fluoride, during human liver homogenization and ammonium sulphate fractionation which significantly limited the degradation of carnitine acetyltransferase. The molecular mass of carnitine acetyltransferase from other species was similar: the enzyme from pigeon breast muscle was about 55 kDa [4], the beef heart enzyme was 62 kDa [lo], while mitochondria1 and peroxisomal carnitine acetyltransferase from Cundidu tropicalis were 60 kDa and 79 kDa, respectively [I I]. The PI value for human carnitine acetyltransferase is 6.3. Mouse carnitine acetyltransferase has a PI of 6.8 [9] while alkaline PI values were obtained in the other species tested (see Table 2 in [3]). The amino acid composition of human carnitine acetyltransferase is shown in Table 2. Compared with rat liver carnitine acetyltransferase [8], human carnitine acetyltransferase has a higher number of glycine residues (38 versus 28) and a lower number of alanine residues (42 versus 73). Human carnitine acetyltransferase also has a more aspartic acid residues (48 versus 45) and a fewer residues of glutamic acid,
1
C2
C3
I
I
I
I
Cb C 6 C8 C10 C12 Acyl- CoA chain length
-
m u
C1L
C16
Fig. 2. Substrate specificity of human carnitine acetyltransferase. AcylCoAs with carbon-chain length ranging over C2-16 (0.1 mM) were used as substrate. Enzyme activity was determined as described under Experimental Procedures
Table 3. Apparent Michuelis constants of human carnitine acetyltransferase K,,, values for acyl-CoAs were determined using a final concentration of 1 mM L-carnitine and 10-200 pM acyl-CoA. When determining K,,, values for L-carnitine a final concentration of 100 pM acyl-CoA and 0.1 -2 mM L-carnitine were used. Assays were performed as described under Experimental Procedures
acyl-CoA
L-carnitine
PM Acetyl-CoA Propionyl-CoA Butyryl-CoA Hexanoyl-CoA Octanoyl-CoA Decanoyl-CoA
21.3 28.0 43.5 54.9 50.3 64.0
91 86 152 120 148 580
lysine, histidine and arginine (63 versus 73, 36 versus 49, 18 versus 19 and 25 versus 27, respectively). These differences may account for the more acidic isoelectric point of human carnitine acetyltransferase compared to that of rat carnitine acetyltransferase [28]. Catalytic properties The substrate specificity of carnitine acetyltransferase (shown in Fig. 2) demonstrates that the enzyme is very active with acetyl-CoA, propionyl-CoA and butyryl-CoA. The maximum activity is obtained with propionyl-CoA, 30% of the peak activity is achieved with octanoyl-CoA while palmitoyl-CoA is not a substrate. Maximum carnitine acetyltransferase activity with propionyl-CoA or butyryl-CoA was detected in pig heart [2] and rat and mouse liver [8,9], respectively. In the other species tested, the activity is strongly reduced with octanoyl-CoA and absent with long-chain acylCoAs [3]. The apparent Michaelis constants of carnitine acetyltransferase for different substrates are listed in Table 3. Values for both acyl-CoAs and carnitine were determined with saturating levels of the other substrates. K, values are low for short-
543
Antibody (PIof antiserum)
Fig. 3. Carnitine acetyltransferase titration by antibodies. Different concentration of anti-(carnitine acetyltransferase) antibodies were incubated with 18 pg purified carnitine acetyltransferase for 1 h at room temperature. Samples were spun for 10min at 14000xg and the supernatant was assayed for carnitine acetyltransferase activity as described under Experimental Procedures
chain acyl-CoAs and increase with increasing carbon-chain length. These values follow the same profile as those reported by Farrell et al. [9]. K, for L-carnitine is six times higher in the presence of decanoyl-CoA than in the presence of acetylCoA. This implies that short-chain acyl-CoAs are preferential substrates when the concentration of L-carnitine is low. We also determined K, of human carnitine acetyltransferase for acetylcarnitine, propionylcarnitine and octanoylcarnitine. The lowest K,,, (0.42 mM) was obtained for acetyltransferase; a value of 0.65 inM was obtained for propionylcarnitine while the highest K, (1.39 mM) was obtained for octanoylcarnitine. V,,, for acetyl-CoA and propionyl-CoA (forward reaction) and for acetylcarnitine and propionylcarnitine (reverse reaction) were also determined in a sample of partially purified carnitine acetyltransferase. The values for acetyl-CoA and propionyl-CoA were 48.2 pmol . mg-I and 79 pmol . mg-l, respectively, while with acetylcarnitine and propionylcarnitine the activity was 27.1 pmol . mg-I and 23.6 pmol . mg-'. Both K , and V,,, values suggest that the forward reactionpredominates in vivo and that carnitine acetyltransferase acts preferentially on the formation of acetylcarnitine from intramitochondrial acetyl-CoA. The pH optimum for carnitine acetyltransferase activity was determined by assaying acetylcarnitine formation (forward reaction) with a pH range over 6.0 - 10.0. The pH optima for carnitine acetyltransferase are in the alkaline range with a peak at pH 8.7. A similar value was found in pig heart carnitine acetyltransferase while pH optima in other species were lower (see Table 2 in [3]). In contrast with mouse peroxisomal carnitine acetyltransferase activity, which is not affected by 10 mM Ca2+ [9], human carnitine acetyltransferase was inhibited to 85.5%, a difference which could be partly related to the different pH values used in the assay. The activity decreased by increasing the concentration of Ca2+to 12.8% in the presence of 100 mM Ca2 . +
Inzmunochemicalstudies
Polyclonal antibodies were obtained by immunizing a New Zealand white rabbit with purified carnitine acetyltransferase,
Fig. 4. lmmunohlotting of' purified carnitine acetyltransferase. After blotting onto nitrocellulose filter standard proteins were stained with amido black (lane 1) while purified carnitine acetyltransferase (8.7 pg) was matched with anti-(carnitine acetyltransferase) antiserum 100fold diluted (lane 2)
Fig. 5. Immunoblotting of purified curnitine acetyltranjferase and human brain homogenate. The human brain homogenate was prepared as described under Experimental Procedures. The supernatant obtained after centrifugation at 16000xg was run together with purified carnitine acetyltransferase on a 9% polyacrylamide minigel. Reaction with anti-(carnitine acetyltransferase) antibodies was performed as described previously. (1) Purified carnitine acetyltransferase (0.9 pg); (2) human brain homogenate (1 10 pg). The arrowhead indicates the 60.5-kDa band in human brain homogenate
as described under Experimental Procedures. The titer of the antiserum was first detected by immunoprecipitation experiments. Fig. 3 shows that the activity of 18 pg carnitine acetyltransferase decreases with increasing antiserum concentration and is abolished after incubation with 20 p1 antiserum. This antiserum has been used for immunoblotting and immunoprecipitation studies on different samples. As shown in Fig. 4, lane 2 , anti-(carnitine acetyltransferase) antiserum, 100-fold diluted, identifies a 60.5-kDa band when purified carnitine acetyltransferase is loaded onto the gel. The supernatant obtained after centrifuging human brain homogenate was also subjected to Western blot analysis. Anti-(carnitine acetyltransferase) antibodies recognized one band in the supernatant (Fig. 5, lane 2) with identical molecular mass to
544
1
2
Table 4. Enzyme specific activities in total homogenate and subcellular fractions of human liver 10 g human liver were used for subcellular fractionation following methods described by Ramsay et al. (1987) and Ghosh and Haira (1 986). Total homogenate, mitochondria and peroxisomes were treated with 0.5% Tween 20 before activities were assayed
3
97kDa
Fraction
glutamate de- catalase hydrogenase
68 kDa
43 kDa
Fig. 6. Polyacrylamide gel electrophoresis of13 5SJmethionine-lahelled curnitine ucetyltransferase synthesized in normal cultured human fibroblasts. Confluent monolayers of fibroblast were labelled with [35S]mcthionine in the presence and absence of Rhodamine 6G and cell extracts were subjected to immunoprecipitation with anti-(carnitine acetyltransferase) antibodies (7.5 PI). The immunoprecipitates were analyied on 10% SDSjPAGE as described under Experimental Procedures. The precursor and the mature form of carnitine acetyltransferase are shown in lanes 1 and 2. respectively. Standard proteins are in lane 3
purified carnitine acetyltransferase (lane 1). Since carnitine acetyltransferase activity is present in the nervous system [13, 14, 19, 201 this 60.5-kDa band very likely represents human brain carnitine acetyltransferase. We also labelled human fibroblasts with [35S]methionine in the presence or absence of Rhodamine 6G, an inhibitor of mitochondrial enzyme processing [29]. The labelled products were immunoprecipitated with anti-(carnitine acetyltransferase) antibodies and protein A from Staphylococcus aureus. After SDSjPAGE on a 10% slab gel, one radioactive band corresponding to 65 kDa was visible in the sample treated with Rhodamine 6G (Fig. 6, lane 1). This precursor is 4.5kDa larger than mature carnitine acetyltransferase (60.5-kDa) which is clearly detectable in the sample without Rhodamine 6G (Fig. 6, lane 2). This evidence indicates that antibodies against purified carnitine acetyltransferase recognize the mitochondrial form of this enzyme. Indeed, most of the mitochondrial proteins encoded in the nucleus are synthesized as cytoplasmic precursor of larger molecular mass and converted in the mitochondria into the mature form [30].
Carnitine acetyltransjerase activity in human mitochondrial and peroxisomes Since a peroxisomal carnitine acetyltransferase with properties similar to human carnitine acetyltransferase has already been described [9], we investigated the existence of peroxisomal carnitine acetyltransferase in human livcr. Starting from a sample of human liver obtained about 2 h
Specific activity of
Homogenate Mitochondria Peroxisomes
1.10 0.065 2.96 f 0.125 0.52 0.030
*
*
0.10 0.009 0.01 k 0.000 2.31 & 0.511
carnitine acetyltransferase
0.03 k 0.003 0.06 0.002 0.14 f 0.012
Fig. 7. Immunohlotting of total homogenate and peroxisomal und mifochondrial fractions f r o m human liver. Peroxisomes and mitochondria were obtained from fresh human liver as described under Experimental Procedures. Total homogenate, mitochondria, peroxisomes and pellet-extracted carnitine acetyltransferase were electrophoresed on a 10% polyacrylamide gel, blotted onto a nitrocellulose filter and incubated with anti-(carnitine acetyltransferase) antibodies diluted 100 times. ( 1 ) Total homogenate (456 pg proteins); (2) mitochondrial fraction (199 pg proteins); (3) peroxisomal fraction (88 pg proteins). Purified carnitine acetyltransferase (lane 4)is indicated by the arrowhead
after surgery, we prepared enriched fractions of mitochondria and peroxisomes by the method described above. Table 4 shows the specific activity of marker enzymes and carnitine acetyltransferase in liver homogenate, mitochondria and peroxisomes. The specific activity of catalase increased 23-fold in peroxisomes when compared with that in the homogenate. The specific activity of glutamate dehydrogenase increased 2.7-fold in mitochondria compared to that in the homogenate. Carnitine acetyltransferase activity was observed both in mitochondria and in peroxisomes. The specific activities of mitochondria and peroxisomes are 2-fold and 4.7fold higher, respectively, than that in the homogenate. Total homogenate, mitochondria and peroxisomes were subjected to immunoblot analysis using anti-(carnitine acetyltransferase) antibodies (Fig. 7). A 60.5-kDa band ofmolecular mass identical to carnitine acetyltransferase is visible in the three preparations and a band 2-kDa larger than carnitine acetyltransferase, barely visible in the homogenate, is enriched in peroxisomes. Miyazawa et al. co-purified a peptide of molecular mass slightly higher than rat liver carnitine acetyltransferase which is suggested to be the precursor of mitochondrial carnitine acetyltransferase [XI. We incubated 90 pg of the peroxisomal preparation for 30 min at 25 C with 200 pg of frozen-thawed mitochondria
545
Fig. 8. Immunohlot of pellet-extracted carnitine acetyltransferase. Pellet-extracted carnitine acetyltransferase (2 pg) was electrophoresed on a 10% slab gel, blotted onto a nitrocellulose filter and incubated with anti-(carnitine acetyltransferase) antibodies diluted 100 times as described under Experimental Procedures
in sucrose buffer. When the sample was analysed by 10% SDSjPAGE the 62.5-kDa peptide was not modified (not shown). This suggests that the peptide we observed is not the precursor of human mitochondrial carnitine acetyltransferase, in agreement with the observation that such a precursor in fibroblasts is 2 kDa larger than 62.5 kDa. Interestingly, in the peroxisomal fraction we found identical octanoyl-CoA and acetyl-CoA activities, both inhibited by anti-(carnitine acetyltransferase) antibodies (data not shown). Since the activity of purified carnitine acetyltransferase with octanoyl-CoA is only 50% of the activity with acetyl-CoA (see Fig. 2), the residual octanoyl-CoA activity in peroxisomes could be attributed to human carnitine octanoyltransferase, a peroxisomal enzyme characterized in rat and mouse liver [9, 311. Therefore, it is possible that the 62.5-kDa peptide recognized by anti-(carnitine acetyltransferase) antibodies in peroxisomes is carnitine octanoyltransferase. A large 67-kDa band, visible in the human liver homogenate and faintly in mitochondria and peroxisomes, disappeared after incubation of antibodies with albumin (not shown). This indicates that albumin, although not visible (see Figs 1 and 4, Fig. 5, lane 1 and Fig. 7, lane 4) is present in the preparation of purified carnitine acetyltransferase and detectable by anti-(carnitine acetyltransferase) antibodies in human liver homogenate, but not in fibroblast homogenate (see Fig. 6, lanes 1 and 2). Another faint band of molecular mass slightly larger than albumin is visible in immunoblots of peroxisomes (Fig. 7, lane 3); the identity of this band is unknown. Other bands of smaller molecular mass, especially in the hornogenate of human liver, are likely to be degradation products of carnitine acetyltransferase due to the action of proteases and to repeated freezing and thawing of the sample. Purification and properties of membrane-extracted carnitine acetyltransferase ,from human liver
Markwell et al. showed that in rat Liver approximately 52% of carnitine acetyltransferase activity is associated with
the mitochondrial fraction, 14% with the peroxisomal fraction, and 34% is bound to a lipid-rich membranous fraction, partially derived from microsomes [17]. Mitochondria1 carnitine acetyltransferase was originally located in the interniembrane space and in the matrix, according to subfractionation experiments performed with digitonin in rat liver mitochondria 132) but Edwards et al. suggested that carnitine acetyltransferase only bound to the inner mitochondrial membrane [ 181. We observed that the 40000 x g pellet from human liver homogenate, extracted by sonication in the presence of 0.5% Tween 20, contained carnitine acetyltransferase activity. The extract was subjected to the same chromatographic procedures as used for carnitine acetyltransferase purified from the supernatant of human liver hornogenate (see Experimental Procedures). The final carnitine acetyltransferase activity with acetyl-CoA as a substrate, was 72 unit . mg-' with a purification of 2889-fold over the pellet extract (data not shown). The final preparation electrophoresed in a polyacrylamide gel, and stained with Coomassie blue is shown in Fig. 1, lane A. Carnitine acetyltransferase from the pellet and from the supernatant have identical molecular masses. The substrate specificity of membrane-extracted carnitine acetyltransferase is identical to the substrate specificity of carnitine acetyltransferase from the supernatant, with maximum activity for propionyl-CoA (data not shown). Michaelis constants determined either for acyl-CoAs or L-carnitine, as well as pH optima and pl values (data not shown) are all simiiar to those obtained for carnitine acetyltransferase purified from the human liver supernatant. The final preparation of membrane-extracted carnitine acetyltransferase is identified by immunoblot with carnitine acetyltransferase antiserum (Fig. 8) and its activity is also inhibited after incubation with 25 p1 antiserum (data not shown). Therefore, carnitine acetyltransferase purified from the supernatant and the pellet from human liver homogenates share a considerable degree of similarity and further studies will be needed to understand the molecular mechanisms of their subcellular localization. The amino acid composition of carnitine acetyltransferasc was determined by K. Williams and K. Stone (Department of Molecular Biophysics and Biochemistry, Yale University Medical School, New Haven CT, USA). This work has been partially supported by a grant from the Muscular Dystrophy Association for the project: Molecular cloning of human carnitine acyltransferases. I. C. was supported by a grant from the Department of General Physiology and Biological Chemistry, Milan University, School of Pharmacology.
REFERENCES 1. Friedman, S. & Fraenkel, G. (1955) Arch. Biochem. Biophjs. 59,
491 - 501. 2. Fritz, I. B., Schultz, S. K. & Srere, P. A. (1963) J . Biol. Chem. 238,2509-2517. 3. Colucci, W. I. & Gandour, R. D. (1988) Bioorg. Chem 16, 307334. 4. Chase, J . F. A., Pearson, D. 3 . & Tubbs, P. K. (1965) Biochim. Bioph-vs. Actu 96, 162 - 165. 5. Chase, J . F. A. & Tubbs, P. K. (1969) Biochem. J . 111,225-235. 6. Markwell, M. A. K. & Bieber, L. L. (1976) Arch. Biochem. Biophys. 172, 502- 509. 7. Mittal, B. & Kurup, C. K. R. (1980) Biochim. Biophys. Actu 619, 90 - 97. 8. Miyazawa, S., Ozasa, H., Furuta, S., Osumi, T. & Hashimoto, T. (1982) J . Biochem. (Tokyo) 93,439-451. 9. Farrell, S. O., Fiol, C . J., Reddy, 1. K. & Bieber, L. L. (1984) J . Bid. Chem. 259,13089-13095.
546 10. Huckle, W. R. & Tamblyn, T. M. (1983) Arch. Biochem. Biophys. 226,94-110. 11. Ueda, M., Tanaka, A. & Fukui, S. (1982) Eur. J . Biochem. 124, 205 -210. 12. Marquis, N. R. & Fritz, I. B. (1965) J. Biol. Chem. 240, 21932196. 13. Fritz, I. B. (1963) Adv. Lipid. Res. I, 285-334. 14. McCaman, R. E., McCaman, M. W. & Stafford, M. L. (1966) J . Biol. Chem. 241,930-934. 15. Dolezal, V. & Tuceck, S. (1981) 1. Neurochem. 36,1323-1330. 16. Willoughby, J., Craig, F. E., Harvey, S. A. K. & Clark, J. €3. (1989) J . Neurochem. 52,896-901. 17. Markwell, M. A. K., McGroarty, E. J., Bieber, L. L. & Tolbert, N. E. (1973) J . Biol. Chem. 248, 3426-3432. 18. Edwards, Y. H., Chase, J. F. A., Edwards, M. R. &Tubbs, P. K . (1974) Eur. J . Biochem. 46, 209-215. 19. DiDonato, S., Rimoldi, M., Moise, A., Bertagnolio, B. & Uziel, G. (1979) Neurology 29, 1578-1583. 20. Przyrembel, H. (1987) J . ZnheritedMetab. Dis. 10, 129-146. 21. Gerber, N., Dickinson, R. G., Harland, R. C., Lynn, R. K., Houghton, D., Antonias, J. I. & Schimschock, J. C. (1979) J . Pediatr. 95, 142- 144. 22. Singh, Y., Liu, G. A. & Krishna, G. (1987) J. Toxicol. Environ. Health 22. 459 - 469.
23. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Meth0d.s ofenzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol 1, p. 439, Verlag Chemie Weinheim, Academic Press, Inc. 24. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Methods ofenzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol. 1, pp. 461 -462, Verlag Chemie Weinheim, Academic Press Inc. 25. Bergmeyer, U. H., Gawehn, K. & Grassl, M. (1974) in Methods of enzymatic analysis, 2nd edn (Bergmayer, U. H., ed.) vol 1, pp. 438 -439, Verlag Chemie Weinheim, Academic Press, Inc. 26. Laemmli, U. K . (1 970) Nature 227,680 - 685. 27. Ikeda, Y., Keese, S. M. & Tanaka, K. (1985) Proc. Natl Acad. Sci. U S A 82,7081 -7085. 28. Markwell, M. A. K., Tolbert, N. E. & Bieber, L. L. (1976) Arch. Biochem. Biophys. 176,479-488. 29. Kuzela, S., Joste, V. & Nelson, B. D. (1986) Eur. J . Biochem. (Tokyo) 154, 553 - 557. 30. Hay, R., Boehni, P. & Gasser, S. (1984) Biochim. Biophys. Actu 779, 65 - 87. 31. Miyazawa, S., Ozasa, H., Osumi, T. & Hashimoto, T. (1983) J . Biochem. (Tokyo) 94, 529 - 542. 32. Brdiczka, D., Gerbitz, K. & Pette, D. (1969) Eur. J . Biochem. 11, 234 - 240.
