163

Molecular and Cellular Biochemistry 178: 163–168, 1998.© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

Purification and characterization of the rat livergamma-butyrobetaine hydroxylase

Stéphane Galland, Françoise Le Borgne, Denis Guyonnet,Pierre Clouet and Jean DemarquoyUniversité de Bourgogne, Unité de Recherche en Nutrition Cellulaire et Métabolique, BP400, F-21011 Dijon Cedex,France

Received 17 February 1997; accepted 11 June 1997

Abstract

The biosynthesis of carnitine from lysine and methionine involves five enzymatic reactions. γ-butyrobetaine hydroxylase (BBH;EC 1.14.11.1) is the last enzyme of this pathway. It catalyzes the reaction of hydroxylation of γ-butyrobetaine to carnitine.This enzyme had never been purified to homogeneity from rat tissue. This paper describes the purification and characterizationof the rat liver BBH. This protein has been purified some 413 fold by ion exchange, affinity and gel-filtration chromatographiesand appears as a dimere of 43,000 Daltons subunits by PAGE. The affinity chromatography column used in the purificationprocess utilizes 3-(2,2,2-trimethylhydrazinium)propionate (THP), a BBH inhibitor, as the ligand. Polyclonal antibodies wereraised against the liver enzyme. They were able to precipitate BBH activity in either a crude liver extract or a purified fractionof the enzyme. Furthermore, it crossreacts with a 43 kDa protein in the liver. No evidence for extra hepatic enzyme was found.(Mol Cell Biochem 178: 163–168, 1998)

Key words: carnitine biosynthesis, γ-butyrobetaine hydroxylase, rat liver, fatty acid oxidation

Address for offprints: J. Demarquoy, Université de Bourgogne, Unité de Recherche en Nutrition Cellulaire et Métabolique, BP400, F-21011 DijonCedex, France

Introduction

Carnitine plays a major role in the oxidation of fatty acidsallowing the transfer of fatty acids through the inner mito-chondrial membrane to the mitochondrial matrix where β-oxidation occurs. Carnitine biosynthesis involves fiveenzymatic reactions. The ultimate precursors of this synthesisare two amino acids: lysine and methionine [1]. Five enzymesare required for carnitine synthesis. The first reaction iscatalyzed by nuclear enzymes (protein-lysine methyltrans-ferase; EC 2.1.1.X, X = 43, 59, 60). During this reaction, S-adenosylmethionine provides three methyl groups for themethylation of peptide-linked lysyl residues to form peptide-linked 6-N-trimethyllysine (TML) [2]. TML is released fromprotein by protein turnover and then reaches the mitochondriato be hydroxylated to 3-hydroxy 6-N-trimethyllysine, thisreaction, catalyzed by 6-N-trimethyllysine dioxygenase (EC1.14.11.8) [3], takes place in the mitochondrial matrix. The

next step occurs in the cytoplasm and during this reaction,catalyzed by the serine transhydroxy methylase, the 3-hydroxy 6-N-trimethyllysine is clived to glycine and γ-butyrobetaine aldehyde [4], that is further oxidized toγ-butyrobetaine. This NAD-dependent reaction is catalyzedby the cytosolic γ-butyrobetaine aldehyde dehydrogenase(EC 1.2.1.47) [5]. Finally, γ-butyrobetaine is hydroxylatedto carnitine by γ-butyrobetaine hydroxylase (BBH; EC1.14.11.1), a cytosolic enzyme [6].

The first four enzymes of the carnitine biosynthesispathway are expressed in virtually any tissues. On theopposite, in rat, BBH activity was only detected in the liverand to some extent in the testis [7–9]. Our goal was tocharacterize the rat liver BBH. We first purify the enzyme anddetermine kinetic parameters of the reaction on purifiedenzyme. Antibodies directed against the liver enzyme wereraised and used to precise BBH tissular location and to relateBBH activity to the presence of immunoreactive material.

164

This allowed to determine in which tissue the enzyme ispresent and therefore where carnitine is biosynthesized.

Materials and methods

Materials

AnimalsAll studies were performed on adult male Wistar rats (Dépré,Saint Doulchard, France) weighting between 200–250 g.They were given food and water ad libitum and were fasted18 h prior to use in experiments.

Chemicals[3H]acetyl-CoA was obtained from Amersham Radiochemi-cal Centre (Amersham, Bucks, UK). 3-(2,2,2-trimethylhydra-zinium)propionate (THP) was a generous gift of Dr. Bundulis(Grindex, Riga, Latvia). EAH Sepharose 4B, Sephacryl 200HR, Sephadex G25 and MonoQ HR were purchased fromPharmacia (Pharmacia Biotech, St Quentin en Yvelines,France). Other chemicals or drugs were purchased fromSigma (Sigma, France) and were of analytical grade.

Methods

Cytosolic fractionOrgans were minced, washed 3 times with ice-cold homo-genization buffer (Sucrose 300 mM, EGTA 1 mM in Tris-HCl50 mM, pH 7.5), briefly dried and weighted. They werehomogenized in the same buffer with 6 strokes of a loose-fitting Teflon pestle at 200 rpm in an ice-cooled ElvehjemPotter. The homogenate was centrifuged at 13,000 × g for 30min at 4°C. The resulting supernatant was centrifuged at100,000 × g for 60 min at 4°C. The final supernatant repre-sented the cytosolic fraction and was used for proteindetermination and enzyme purification.

BBH activityTwo different assays were used. For most experiments, BBHactivity was determined on cytosolic fractions as describedby [10] with minor modifications. A thermal denaturation ofthe enzyme (65°C, 2 min) was used to stop the reactioninstead of an organic extraction and the sample was subse-quently spun 2 min at 13,000 × g (assay #1). Carnitine formedwas determined as in [11]. For determining BBH activity inother organs than the liver, another assay described in [12]was used (assay #2). Enzyme activity was expressed in nmoleof carnitine formed per min and per µg protein.

BBH purificationBBH purification was performed from rat liver cytosol. Allthe subsequent steps were carried out at 4°C. Cytosol (160–200 ml) was taken to 50% saturation with solid (NH

4)

2SO

4

(v/v), stirred for 60–90 min and then centrifuged at 13,000 ×g for 30 min. Supernatant was taken to 60% saturation andcentrifuged under the same conditions. The resulting pelletwas dissolved in 10 ml of a solution of Tris-HCl 20 mM, pH7.5 (Tris buffer) and desalted using a 2.6 × 40 cm G25 columnequilibrated with the same buffer. The protein fractions werepooled and applied to a 2.6 × 40 cm MonoQ HR columnequilibrated with the Tris buffer. The column was washedwith a stepwise gradient of 100 mM of KCl in Tris buffer.BBH was eluted with a linear gradient (from 100–300 mMof KCl in the Tris buffer). The active fractions were pooledand applied to a 1.6 × 20 cm column of EAH Sepharose 4Bon which THP had been covalently linked. The column waswashed with the Tris buffer. Bound proteins were eluted witha stepwise gradient of 0.1, 0.5 and 1 M NaCl in the same Trisbuffer. Active fractions were pooled and placed onto a 2.6 ×100 cm Sephacryl 200HR column equilibrated and run withthe Tris buffer containing 0.2 M KCl. Depending of the purityof fractions, a second ion-exchange chromatography mayhave been used. The nature of the purified fractions wasestimated by SDS-PAGE.

Polyacrylamide gel electrophoresisSodium dodecyl sulfate PAGE were done according to theprocedure of Weber and Osborn [13]. The final concentrationof acrylamide in the gel was 12.5%. They were run at aconstant current of 30 mA. After the electrophoresis iscompleted, proteins were stained using Coomassie blue.Molecular weight markers were used in all electrophoreses.

Protein concentration determinationDuring chromatographic steps, protein concentration wasestimated according to its absorbance at 280 nm. Moreprecise determination was performed using the BCA proce-dure (Biorad) with serum albumine as a standard.

Kinetic studiesThese were done on purified enzyme. Incubations wereperformed as above (assay #1 ) except the concentration ofeach reagent was different for each experiment. The concen-trations and incubation times are indicated in the legend ofthe corresponding figure or table.

Polyclonal antibodiesAntibodies were raised against the denatured purified BBH.Virtually pure fractions of BBH (purity > 95% as estimatedafter Coomassie blue revelation) were submitted to a SDS-PAGE, then stained. The bands were cut out and used to makeantibodies into rabbits. A total of 200 µg of protein were used

165

for the 4 injections. IgG fraction of the serum was isolatedafter (NH

4)

2SO

4 precipitation.

Immunoprecipitation of rat liver BBHTo assess the specificity of the antibodies, an immuno-neutralization of BBH was done. Pure BBH or a crudecytosolic extract was incubated in the presence of variousamount of pre-immune or anti-BBH sera. A suspension ofStaphilococus Aureus (Cowan strain, SAC) was added to themixture (10 µl of a 10% suspension for 1 µl of serum). Themixture was incubated for 2 h at 4°C and the antibody-antigen-protein A(SAC) pellet was eliminated by a 12,000× g centrifugation (1 min) and the BBH activity estimated inthe supernatant.

Western blotWestern analysis was used for determining tissular BBHlocation. After SDS-PAGE of various samples, the gel wastransferred to a PVDF membrane (Polyscreen, Dupont NEN)in a 192 mM glycine-25 mM Tris pH 8.3 transfer buffercontaining 20% methanol at 30 volts overnight. After transferis completed, the membrane was first saturated with TBS-Tween 0.05%-BSA 1% and then incubated in the presenceof anti-BBH antibodies (1/200 dilution in TBS-Tween0.05%). After 3 washes with TBS-Tween 0.05%, the mem-brane was incubated with diluted horseradish peroxidase(HRP) conjugated goat anti-rabbit IgG (1/10000 in TBS-Tween 0.05%). The HRP activity was detected using achemiluminescence procedure (Renaissance Kit, DupontNEN).

Results

Purification and characterization of rat liver BBHPurification of rat liver BBH was accomplished using severalchromatographic steps. After ammonium sulfate precipitation(50–60%), proteins were desalted and applied to a mono Qcolumn. Active fractions from the mono Q column wereapplied to an affinity chromatography where the ligand wasthe THP (also called mildronate) a BBH inhibitor [14]. Theproteins were eluted with a NaCl gradient, BBH activityappeared at a concentration of 0.5 M NaCl. However, the

capacity of this column appeared to be limited and only smallamount of sample can be applied to the column, therefore agel filtration on a Sephacryl 200 HR was added to theprocedure. The active fractions appeared as a single peak justafter the void volume at a Kav of 0.225. A summary of theentire purification is given in Table 1.

The molecular weight of the purified enzyme was esti-mated on gel filtration using Sephacryl 200 HR and by SDS-PAGE. By comparison with standards, a molecular weightof 68,000 and 43,000 Da respectively was found. Thesefindings confirm that BBH is a dimer of subunits of identicalsize. No difference in the size of both subunits has been foundusing this approach.

Catalytical propertiesCatalytical properties of rat liver BBH were studied withpurified enzyme. Using the assay described in the Materialsand methods section (i.e. in the presence of catalase), kineticparameters were determined 3–4 times and the average wasused for K

m determination (All were perfectly matching).

Constants for γ-butyrobetaine and α-ketoglutarateThe K

m values for γ-butyrobetaine and α-ketoglutarate were

respectively 0.08 and 0.125 mM. Concentrations of γ-butyrobetaine above 0.2 mM were found to inhibit thereaction as well as concentration of α-ketoglutarate higherthan 1 mM (Fig. 1, A and B).

Effect of Fe2+ and ascorbateNo activity for the rat liver BBH was detected when eitherFe2+ ions or ascorbate were omitted from the reaction mixture.The K

m value for ascorbate was 6.6 mM (Fig. 1C).

Effect of catalaseWhen catalase was added to the assay medium, a 10 foldincrease in the BBH activity was found. The maximumactivity was detected at a concentration of 2 mg catalase. TheK

m value for catalase was 1.1 µM (this was calculated using

a molecular weight of 240,000 for catalase) (Fig. 1D).

Effect of THPTHP is a non competitive inhibitor of the mammalian BBH.This inhibition is complete at a concentration of 0.05 mM (Fig.2). The IC

50 was found for 13.3 µM.

Table 1. Purification of γ-butyrobetaine hydroxylase from 100 g of rat liver

Step Protein BBH activity Specific activity Yield Purification(nmol/min) (nmol/min/mg) (%) (n-fold)

Cytosol 3072 1506 0.49 100 –MonoQ™ 318 1076 3.38 7 1 7THP-EAH Sepharose 21 538 25.62 36 52Gel filtration 1.6 324 202.51 22 413

Protein and activity determination was performed as described in the Materials and methods section.

166

Immunological studiesIgG fraction of the antiserum or crude antiserum from rabbitinjected with purified rat liver BBH reacts with the purifiedenzyme and the enzyme present in a crude cytosolic fraction.The antiserum also inhibits the BBH activity while a pre-immune serum does not (Fig. 3).

Tissular location of BBH in rat organsWestern blot analysis of BBH protein in various tissuesshowed that the antibodies raised against the rat liver BBHwere only able to recognize the liver enzyme. A strong signalof immunoreactive material was found at a molecular weightof 43 kDa in the liver lane (Fig. 4). Under the same conditionsor under less stringent conditions, a very feint band of cross-reacting material was detected in the testis and the brain at a45 kDa molecular weight. The anti-BBH antibodies alsocross-reacted with 2 proteins in the kidney whose size wasrespectively around 40 and 44 kDa. No cross-reactingmaterial was found in the other examined tissue (i.e. heart,epididyme . . .).

BBH activity was measured in crude cytosolic and ammo-nium sulfate precipitated fractions extracted from the same

tissues. The highest level of BBH activity was found in theliver (1.1 nmol/min/µg protein). Little to no activity was foundin the testis (0.03 nmol/min/µg protein), kidney, heart, brainand epididyme. We found a good correlation between BBHactivity and immunologically-reacting material.

Fig. 1. BBH activity at various concentration of substrates and cofactors. The purified enzyme (1 µg) was incubated for 15 min at 37°C in the presence ofdifferent concentration of A: γ-butyrobetaine; B: α-ketoglutarate; C: ascorbate; D: catalase and the regular amount of the other components of the reactionsystem. In the Lineweaver-Burk plot, V is the amount of carnitine formed during the incubation and [S] the concentration of the substrate or cofactor addedto the mixture.

Fig. 2. Effect of increasing concentration of THP on BBH activity. BBHactivity was assayed using assay#1 except that THP was added to thereaction system.

167

Discussion

γ-butyrobetaine hydroxylase has been previously purified tohomogeneity from Pseudomonas sp. AK1, an induciblebacterial strain [15], from calf liver [16] and human kidney[17] but never from rat liver. It was interesting to purify andget antibodies against this rat enzyme since most nutritionalstudies concerning carnitine are done with rats. We developeda purification process that is based on an affinity chroma-tography, using THP as the ligand. However, this affinitychromatography was not efficient enough to be used alone.The yield was relatively low (around 50%) as well as thepurification (8–15 times). It seems that the THP, because ofits own ionic properties may act like an ion-exchangechromatography, decreasing the specificity and the yield.

Using all these chromatographic steps, we were able topurify to homogeneity the rat liver BBH. As in other speciesand previously reported [15, 16, 18, 19] the enzyme appearedas a dimer of identical size subunits. No evidence for differencein the structure of the subunits was found. They bothappeared as a single band in SDS-PAGE and no real attempt

of separating both subunit was made. This means that bothsubunit might be of a slight different size as it has beenreported in the past [19].

Rat liver BBH showed a strong affinity for its substrates(γ-butyrobetaine and α-ketoglutarate). Km values found forthese components were in the same range that those foundin other species. The rat enzyme, also, showed similarrequirements that the bovine or human BBH, i.e. the presenceof ferrous iron, ascorbate and catalase.

Tissular expression of BBH activity has been well docu-mented. The rat enzyme activity has been mainly found inthe liver [20] and the testis [9]. The kinetic properties of bothenzymes are different and there was no evidence that thesetwo enzymes were the same. Using usual BBH activitymeasurement, we were unable to detect any significantactivity in the testis. This is not surprising since the authorswho described high level of BBH activity in this organ havehad to modify the conditions of the assay. Using this secondassay we did not find any significant activity either. Further-more, if no significant activity has been detected in the testis,we did not find any significant amount of immunoreactiveprotein in this organ. Under low stringency conditions, asingle feint band with a size that has been estimated around45 kDa may be detected. The respective intensity of the liverand testis immunoreactive bands was estimated at a ratio of25+ to 1. Since the BBH found in the testis was reported tohave the same specific activity as the liver enzyme, we thinkthat if there is BBH activity in the testis, this enzyme isdifferent of the liver since it is not recognized by polyclonalantibodies directed against the liver enzyme. Furthermore wedid not find any BBH activity in this tissue even if carnitineconcentration is particularly high.

Immunoreactive material was also found in very lowamount in the kidney and in the brain of adult rat. For thekidney, 2 bands were detected whose sizes were different ofthe liver BBH. In the brain a single band was detected, its sizewas estimated at 45 kDa. These findings suggest that as inthe testis it may exist other forms of BBH in these organs.These proteins that slightly cross-react with antibodies raisedagainst rat liver BBH are not of the same molecular weightand their ability to convert γ-butyrobetaine to carnitine is notdetectable using a classical assay. In Human, BBH activityis not restricted to liver but is also present in kidney [17]where it has been reported to have the highest specific activity[21] and also in the brain. In rat, our results suggest that theseorgans may synthesize one or two proteins that are slightlyrelated to the liver BBH.

Our ultimate goal will be to clone the rat liver cDNAencoding for BBH and to check for the presence of transcriptsin other organs than the liver and to assess for similarities thatmay exist between these different proteins.

Fig. 3. Inhibition of rat liver BBH by an IgG fraction prepared from rabbitantiserum raised against the rat liver BBH.

Fig. 4. Immunodetection of BBH in various tissues by western blot. Tenµg of protein were subjected to a SDS-PAGE, the gel was transferred to aPVDF membrane that was incubated in the presence of anti-BBH antibodies.Ab-Ac interactions were visualized by chemiluminescence (RenaissanceKit, Dupont NEN).

168

Acknowledgements

The authors wish to thank Dr Bundulis for providing THP.This work was supported by the Région Bourgogne, theFondation pour la Recherche Médicale and the Ministère dela Recherche et de l’Enseignement Supérieur.

References

1. Tanphaichitr V, Broquist HP: Site of carnitine biosynthesis in the rat. JNutr 104: 1669–1673, 1974

2. Labadie J, Dunn WA, Aronson NN: Hepatic synthesis of carnitine fromprotein-bound trimethyl-lysine. Lysosomal digestion of methyl-lysine-labelled asialo-fetuin. Biochem J 160: 85–95, 1976

3. Hulse JD, Ellis SR, Henderson LM: Carnitine biosynthesis. β-hydroxylation of trimethyllysine by an α-ketoglutarate-dependentmitochondrial dioxygenase. J Biol Chem 253: 1654–1659, 1978

4. Bradley Hochalter J, Henderson LM: Carnitine biosynthesis: Theformation of glycine from carbons 1 and 2 of 6-N-trimethyl-L-lysine.Biochem Biophys Res Commun 70: 364–366, 1976

5. Hulse JD, Henderson LM: Carnitine biosynthesis. Purification of 4-N-trimethylaminobutyraldehyde dehydrogenase from beef liver. J BiolChem 255: 1146–1151, 1980

6. Lindstedt G, Lindstedt S: Cofactor requirements of γ-butyrobetainehydroxylase from rat liver. J Biol Chem 245: 4178–4186, 1970

7. Haigler HT, Broquist HP: Carnitine synthesis in rat tissue slices.Biochem Biophys Res Commun 56: 676-681,1974

8. Rebouche CJ: Sites and regulation of carnitine biosynthesis in mammals.Federation Proc 41: 2848–2852, 1982

9. Carter AL, Abney TO, Braver H, Chuang AH: Localization of γ-

butyrobetaine hydroxylase in the rat testis. Biol Reprod 37: 68–72, 198710. Paul HS, Sekas G, Adibi SA: Carnitine biosynthesis in hepatic

peroxisomes. Demonstration of γ-butyrobetaine hydroxylaseactivity. Eur J Biochem 203: 599–605, 1992

11. McGarry JD, Foster DW: An improved and simplified radio-isotopic assay for the determination of free and esterifiedcarnitine. J Lipid Res 17: 277–281, 1976

12. Carter AL, Stratman FW: A rapid and sensitive assay of γ-butyro-betaine hydroxylase. Febs Lett 111: 112–114, 1980

13. Weber K, Osborn M: The reliability of molecular weight determi-nations by dodecyl sulfate-polyacrylamide gel electrophoresis. J BiolChem 244: 4406–4412, 1969

14. Simkhovich BZ, Shutenko ZV, Meirena DV, Khagi KB, MezapukeRJ, Molodchina TN, Kalvins IJ, Lukevics E: 3-(2,2,2-trimethyl-hydrazinium)propionate (THP) – a novel γ-butyrobetaine hydroxylaseinhibitor with cardioprotective properties. Biochem Pharmacol 37:195–202, 1988

15. Lindstedt G, Lindstedt S, Nordin l: Purification and properties of γ-butyrobetaine hydroxylase from Pseudomonas sp AK 1. Biochemistry16: 2181 –2188, 1977

16. Kondo A, Blanchard JS, Englard S: Purification and properties of calfliver γ-butyrobetaine hydroxylase. Arch Biochem Biophys 212: 338–346, 1981

17. Lindstedt G, Lindstedt S, Nordin l: γ-butyrobetaine hydroxylase inhuman kidney. Scand J Clin Lab Invest 42: 477–485, 1982

18. Lindstedt S, Nordin I: Multiple forms of γ-butyrobetaine hydroxylase(EC 1.14.11.1). Biochem J 223: 119–127, 1984

19. Ruetschi U, Nordin I, Odelhog B, Jornvall H, Lindstedt S: γ-butyrobetaine hydroxylase. Structural characterization of the Pseudo-monas enzyme. Eur J Biochem 213: 1075–1080, 1993

20. Lindstedt G: Hydroxylation of γ-butyrobetaine to carnitine in rat liver.Biochemistry 6: 1271 –1282, 1967

21. Rebouche CJ, Engel AG: Tissue distribution of carnitine biosyntheticenzymes in man. Biochim Biophys Acta 630: 22–29, 1980