Ubiquitination of soluble and membrane-bound tyrosine hydroxylase and degradation of the soluble form Anne P. Døskeland and Torgeir Flatmark Department of Biochemistry and Molecular Biology, University of Bergen, Norway Tyrosine hydroxylase (TH) demonstrates by two-dimen- sional electrophoresis a microheterogeneity both as a soluble recombinant human TH (hTH1) and as a membrane-bound bovine TH (bTH mem ). Part of the h eterogeneity is likely due to deamidation of l abile asparagine residues. Wild-type (wt)-hTH1 was found to be a substrate for the ubiquitin ( Ub) conjugating enzyme system in a reconstituted in vitro system. When wt-hTH1 was expressed in a coupled transcription- translation TnT R -T7 reticulolysate system 35 S-labelled polypeptides of the expected molecular mass of native enzyme as well as both higher and lower molecular mass forms were observed. The amount of high-molecular-mass forms increased by time and was enhanced in the presence of Ub and clasto-lactacystin b-lactone. In pulse-chase experi- ments the amount of full-length hTH1 decreased by first- order kinetics with a half-time of 7.4 h and 2.1 h in the absence and presence of an ATP-regenerating system, respectively. The ATP-dependent degradation was inhibited by clasto-lacta cystin b-lactone. Our findings support t he conclusion that hTH1 is ubiquitinated and at least p artially degraded by the proteasomes in the reticulocyte lysate system. F inally, it is shown that the integral TH of the bovine adrenal chromaffin granule membrane (bTH mem )is ubiquitinated, most likely m onoubiquitinated. Additional Ub-conjugates of this membrane, detected by Western blot analysis, h ave not yet been identified. Keywords: tyrosine hydroxylase; ubiquitin; proteasome; chromaffin granule membrane; neuroendocrine cells. Tyrosine hydroxylase (TH, EC 1.14.16.2) catalyzes the conversion of L -tyrosine to L -dihydroxyphenylalanine ( L -DOPA), the rate-limiting step in the biosynthesis of dopamine and noradrenaline/adrenaline [ 1]. The cellular activity of TH is regulated by several alternative mech- anisms in response t o, e.g. neuronal and hormonal stimuli of neuroendocrine target cells. Both long-term transcriptional and short-term post-transcriptional m ech- anisms (notably phosphorylation) are involved in its regulation [2,3]. Beside its l ocalization in the brain [2,4], TH is present in high amount in the adrenal chromaffin cells mainly as a soluble cytosolic form (TH sol ) [5,6] and partly as a membrane-bound form (TH mem ), associated with the catecholamine secretory granules [7–9]. The molecular and cellular mechanisms involved in the degradation of t his k ey enzyme of neurotransmitter biosynthesis is, however, not yet known. The half-life of rat TH in PC-12 cells, in a subclone of PC-12 cells and in chromaffin cells has been reported to be 17 h [10], 30 h [11] and 29 ± 3 h [12], respectively, and the possibility that PEST motifs could be involved in its turnover has been suggested [13]. The possibility that the ubiquitin-proteasome pathway could play a role in its degradation is considered in the present study as the structurally closely related recombinant h uman phenylal- anine h ydroxylase (PAH, EC 1.14.16.1) [ 14,15] h as been shown to be a substrate for the ubiquitin (Ub)-conju- gating enzyme s ystem of r at liver [16]. MATERIALS AND METHODS Materials Mouse monoclonal anti-Ub Ig which recognizes free and conjugated Ub was obtained from Z ymed laboratories, Inc. (San Francisco, CA, USA). Polyclonal antibodies directed against recombinant hTH1 e xpressed in E. coli were prepared in rabbit and partially purified by ammonium sulfate precipitation. Peroxidase-conjugated antibodies [goat anti-(mouse IgG) Ig and goat anti-(rabbit IgG) Ig] were from Biorad. Rabbit anti-(mouse IgG) Ig was from Trichem Aps, Denmark. Mouse monoclonal anti-(26S proteasome) IgG (directed against p27 subunit of 20S cylinder particles) was from American Research Products (Belmont, MA, USA). Protein A–Sepharose CL-4B was from Amersham Pharmacia Biotech (Oslo, Norway). Ub C-terminal hydrolase, isopeptidase T was from Affiniti Research Products Ltd (UK), Ub aldehyde (Ubal) and clasto-lactacystin b-lactone were from Boston Biochem I nc. (Cambridge, MA, USA). Yeast hexokinase was from Roche Molecular Biochemicals (Mannheim, Germany). [ 125 I]Protein A and [ 35 S]methionine (code AG 1094) were from Amersham (Buckinghamshire, UK). The TnT R -T7 reticulocyte lysate system was from Promega (Madison, USA). Correspondence to T. Flatmark, Department of Biochemistry and Molecular Biology, University of Bergen, A ˚ rstadveien 19, N- 5 009 Bergen, Norway. Fax: + 47 55 586400, Tel.: + 47 55 586428, E-mail: torgeir.flatmark@pki.uib.no Abbreviations: TH, tyrosine hydroxylase; h TH, human TH; bTH, bovine TH; PAH, phenylalanine hydroxylase; Ub, u b iquitin; L -DOPA, L -dihydroxyphenylalanine. (Received 1 November 200 1, revised 2 January 2 002, accepted 23 January 2002) Eur. J. Biochem. 269, 1561–1569 (2002) Ó FEBS 2002 Purification of recombinant hTH1 expressed in E. coli Isoform 1 of recombinant human TH (hTH 1) expressed i n Escherich ia coli was p urified by affinity chromatography on heparin–Sepharose as described previously [17]. The con- centration of the hydroxylase was expressed in terms of enzyme subunits of 62 kDa [18]. Ubiquitination of wt-hTH1 in a reconstituted in vitro system Ubiquitination of wt-hTH1 was assayed at 3 7 °Cina reconstituted in vitro system with [ 125 I]ubiquitin and the isolated Ub-conjugating enzymes [i.e. a fraction containing the Ub-activating (E1), Ub-carrier (E2s) and Ub-protein ligase (E3)] as described f or ubiquitination of phenylalanine hydroxylase [16]. Following preincubation of the Ub-conjugating enzymes (7.6 lgproteinper55lL assay), with 1.5 l M Ubal, ubiquitination was performed with  18 l M [ 125 I]Ub by the standard assay procedure in the absence and presence of 8 l M hydroxylase. After 90 min, the reaction was quenched by t he addition of acetone, and the p rec ipitated proteins w ere analysed o n two-dimensional electrophoresis (for details, see below). After electrophor- esis, the gels were stained with Coomassie Blue R250 and dried in vacuo at 70 °C between two sheets of cellophane. To determine the distribution of 125 I-radioactivity, gels were then exposed to Hyperfilm TM-b-max for autoradi- ography. T he apparent molecular mass of the 125 I- containing bands in each lane, representing [ 125 I]Ub, free poly Ub chains and [ 125 I]Ub-conjugates, respectively, was estimated by comparison with the position of the standard proteins. Expression and degradation of hTH1 in a coupled transcription-translation reticulocyte lysate system The hTH1 w as expressed in a coupled in vitro transcription- translation system using the pET3a-hTH1 vector [18] and the TnT-T7 reticulocyte lysate system in the presence of [ 35 S]methionine essentially as described by t he supplier. 1–4 lL[ 35 S]methionine and approximately 1 lg of plasmid DNA were routinely used in the 50 lL assay. Reactions were incubated at 30 °C for the time periods indicated in the figure legen ds. From the reaction mixture 5 lL a liquots were quenched at given time points and subjected to SDS/ PAGE after heating to 56 °C for 15 min in the classical Laemmli s ample buffer as treatment of proteins at high temperature (95 °C)hasbeenshowntoresultinthe formation of aggregates especially for samples containing membrane proteins [19] and observed in the present study. The stability of hTH1 was studied in a reaction mixture containing in a final volume of 50 lL: 15 m M Hepes (pH 7 .5), 5 m M MgCl 2 ,0.25m M dithiothreitol, 1 m M methionine and 25 lL of f reshly thawed rabbit reticulocyte lysate. T he reaction was performed at 37 °C i n the pr esence of added 0.5 m M ATP, 10 m M phosphocreatine and 0.2 m gÆmL )1 creatine phosphokinase (Sigma), or in an ATP-depleted lysate obtained by a dding 2-deoxy- D -glucose (20 m M ) and hexokinase (230 UÆmL )1 ). The mixture was preincubated for 10 min and incu bation started by the addition of the l ast component, i.e. [ 35 S]methionine-labelled hTH1 (6.5% of the final volume) freshly obtained by the coupled in vitro transcription-translation system. To mon- itor hTH1 degradation, aliquots (6 lL) were, a t selected time points, added to 10 lL of reducing SDS/PAGE sample buffer containing 2 mercaptoethanol (5%), incubated 15 min at 56 °C,andappliedto10%SDS/PAGEgels. The distribution of r adioactivity in each sample lane of one- dimensional gel or in two-dimensional gel was first deter- mined in unstained gels by a b-scanner (Packard Instant Imager, Packard Inc., Canberra, Australia) and then exposed to Biomax MR (Kodak) or H yperfilm TM-b-max for autoradiography. The app arent molecular mass of t he 35 S-containing bands in each lane, representing [ 35 S]hTH1 and its derivatives, was estimated r elative t o t he position of the standard proteins. Preparation of chromaffin granule membranes Chromaffin granules from the bovine adrenal medulla were isolated by a discontinuous sucrose density-gradient, lysed (hypotonic) and centrifuged in a final discontinuous density-gradient to yield chromaffin granule ghosts essen- tially free from mitochondrial and microsomal contamin- ation [20]. Polyacrylamide gel electrophoresis Protein samples for e lectrophoresis, either from ubiquiti- nation assay or from isolated chromaffin granule ghosts, were precipitated with ice-cold acetone (sample/acet- one ¼ 1 : 3 by vol.) and kept on ice for 30 min After centrifugation (12 000 g for 15 min), the pe llets were dissolved in sample buffer and subjected to one-dimen- sional or two-dimensional gel electrophoresis. SDS/PAGE was performed according to the Laemmli p roce dure [21] in 10% (w/v) gel. One volume of the samples was routinely mixed with 1 vol. of Laemmli sample buffer a nd incubated for 15 min at 56 °C. Two-dimensional electrophoresis was performed as described previously [16]. Acetone precipita- ted proteins were dissolved in a medium containing 9.5 M urea, 2% (w/v) Chaps, 1.6% (w/v) Bio-Lyte p H 5–7, 0.4% Bio-Lyte pH 3–10 and 100 m M dithiothreitol and kept at )20 °C until used. After 1 h pre-electrophoresis at 200 V, the proteins were loaded at the basic end of the isoelec- trofocusing gel, and electrophoresis was performed at 400 v for 16 h and at 1000 v for an additional hour. The second dimension was run according to Laemmli u sing 10% (w/v) acrylamide slab gels (1 mm). The prestained protein standards (Biorad) used were phosphorylase b (101 kDa), BSA (79 kDa), ovalbumin (50.1 kDa), car- bonic anhydrase (34.7 kDa), soybean trypsin inhibitor (28.4 kDa) and lysozyme (20.8 kDa). The gels w ere stained with Coomassie Brilliant Blue, dried in vacuo at 70 °C between two sheets o f c ellophane and analysed f or radioactive proteins. Western blot analysis Proteins from chromaffin granule membranes separated by SDS/PAGE [10% (w/v) gel] were blotted electrophoretically for 3 h at 300 mA on a nitrocellulose membrane (0.45 lm pore diameter, BA 85 from Schleicher & Schuell, Dassel, Germany) in a buffer containing 48 m M Trizma base, 39 m M glycine (pH 9.2) and 20% (v/v) methanol. 1562 A. P. Døskeland and T. Flatmark (Eur. J. Biochem. 269) Ó FEBS 2002 Western blot analysis of chromaffin granule membrane proteins was performed using t he enhanced chemilumines- cence detection method with polyclonal rabbit anti-hTH1 Ig or anti-Ub Ig as primary antibody and anti-rabbit or anti-mouse horseradish peroxidase-labelled s econdary anti- body. Isotopic detection and quantitation using [ 125 I]protein A was preferentially used to ensure specificity of the TH and Ub immunoreactivity. Thus, the transferred proteins were probed with rabbit anti-TH Ig at dilution 1 : 1000 or in paralell with anti-Ub serum at the recommanded working concentration o f 2 lgÆmL )1 and w ith rabbit anti-(mouse IgG) Ig as the secondary antibody. Nitrocellulose mem- branes were then incubated with [ 125 I]protein A at the concentration o f 0.2 lCiÆmL )1 in phosphate buffered s aline containing 2.5% (w/v) dried non fat milk and 0.1% (v/v) Tween-20, and in order to vizualize 125 I-labeled proteins, they were counted in a b scanner ( Packard Instant Imager, Packard Inc., Canberra, Australia) or exposed to X-ray film for autoradiography. For Western blot analysis of chromaffin granule mem- brane proteins on one-dimensional gel, 280 lgofproteins were applied in a large well. After electrophoresis and blotting, the membrane was divided in two identical parts and probed against anti-TH Ig or anti-Ub Ig, respectively, as illustrated below. For analysis of proteins by Western blot on two-dimensional gel, the membrane blot was, after immunodetection with one antibody, for example anti-TH IgG, st ripped a nd probed with a nti-Ub IgG and vice versa. Stripping of bound antibodies was performed by incuba- ting the membrane in a buffer containing 63 m M Tris/HCl pH 6.7, 100 m M 2-mercaptoethanol and 2% (w/v) SDS at 50 °C for 30 min with occasional agitation and finally extensive washing in a large volume of Tris/NaCl/P i / Tween. Immunoisolation Chromaffin granule membrane proteins (200 lg) were solubilized in 1% (w/v) SDS and incubated at room temperature for 5 min 10 vol (920 lL) of buffer (50 m M potassium phosphate pH 7.0 containing 190 m M NaCl, 6m M EDTA, 2.5% Triton X-100 and a cocktail of protease inhibitors including 0.2 m M phenylmethanesulfonyl fluor- ide, 20 lgÆmL )1 leupeptin, 0.5 mg ÆmL )1 soybean trypsin inhibitor, 14 lgÆmL )1 pepstatin, 1 m M benzamidine) were added, followed by addition of 120 lL immunoadsorbent Protein A–Sepharose with bound IgG. The immunoad- sorbent was Protein A–Sepharose ( 10 mg of dry beads suspended and washed twice in 50 m M potassium phos- phate, pH 8.0) to which were coupled 5 lLanti-THIgGby incubating for 1 h on a rotating wheel at 4 °C. For immunoisolation the beads were mixed with samples of the membrane proteins and rocked in Eppendorf tubes f or 2hat4°C. The p rotein A–Sepharose with bound IgG–TH was pelleted b y centrifugation a t 12 000 g for 15 s, washed nine times with phosphate buffer containing 0.2% (w/v) Triton X-100 a nd finally twice with the same buffer without Triton X-100. The pellet was kept at )20 °C until used, then heated (56 °C, 10 min) in sample buffer (40 lL added). Immunoreactive material resolved by SDS/PAGE was thereafter immunoblotted with either anti-Ub Ig or anti- TH Ig, and the immunoreactivities compared. RESULTS Ubiquitination of recombinant wt-hTH1 by a reconstituted ubiquitin conjugating enzyme system As expected from our previous studies [16] on the ubiqui- tination of recombinant w ild-type human phenylalanine hydroxylase (wt-hPAH), it is seen from F ig. 1 that recom- binant wt-hTH1 i s also a substrate for the reconstituted Ub-conjugating enzyme system of rat liver. After incubation of the hydroxylase with 125 I-labelled Ub and a mixture of the purified preparations of the E 1, E2 and E 3 enzymes, the 2D-electrophoresis revealed the formation of 125 I-labelled Fig. 1. Mono- and multi/poly ubiquitination of recombinant hTH1 by a reconstituted ubiquitin conjugating enzyme system. Ubiquitination of recombinant hTH1 was performed in a reconstituted in vitro system with the U b c onju gating e nzymes E1, E2 and E3 isolated by affinity chromatography from rat liver and [ 125 I]Ub [1 6]. hPAH wit h subunit molecularmassof51kDawasusedasapositivereferenceproteinfor ubiquitination. The reaction mixture (55 lL) co ntained 7.6 lgof proteins (E1, E2 and E3 proteins), 1.5 l M Ubal, 18 l M [ 125 I]Ub in the absence of hydroxylase (A), the presence of 8 l M hPAH (B) and (CandD)of8l M hTH1. After 90 m i n, the reaction was quenched by the addition of acetone and precipitated proteins analysed on two- dimensional electrophoresis [12.5% (w/v) ( gel)] in ( A–C); 10% (w /v) gel in (D). Inset in B and C: Coomassie Brilliant Blue stained proteins from the reaction mixture containing PAH (B) and TH (C). The multiple molecular forms of hTH1 have a molecular mass for the subunit of  62 kDa; the doublet with a more acidic p I a nd a molecular mass of  100 kDa r epresents presumably the E1 enzyme [16,49]. The main forms of hTH1 are also indicated by arrows in Panel D. The [ 125 I]Ub-labelled conjugates of hTH1 are v isualized as diagonal spots of radioactivity in the autoradiogram (Panel C and D). The polyUb chains derived from 125 I-labelled Ub with 8.5 kDa a nd neutral pI are observed as vertical spots (A–C). (D) An expanded view of the area of interest (i.e. a bove  60 kDa) and co rrespon ds to the pattern of superimposed profils of stained g el and t he respective autoradio gram obtained after short exposure time. The main Coomassie Blue stained spots indicated by arrows correspond to hTH1 and E1. The auto- radiographic pattern of [ 125 I]Ub-labelled conjugates corresponds mainly to the poly/multi U b-TH conjugates vizu alized as a ladder o f at least eight distinct Ub-TH conjgates with a microheterogeneity corresponding to the enzyme as isolated. Ó FEBS 2002 Ubiquitination of tyrosine hydroxylase (Eur. J. Biochem. 269) 1563 Ub-protein conjugates derived from hTH1 in addition to poly Ub-chains. Mono-ubiquitinated and poly/multi-ub iq- uitinated species were visualized as a diagonal pattern of high M r ÔladderÕ of radioactive spots (Fig. 1C) which reflects the h eterogeneity of the Ub-protein conjugates of wt-hTH1 in terms of size and pI. Multiple molecular forms of  62 kDa and different p I values around 5.5 (Fig. 1C inset) were observed for the nonubiquitinated enzyme. A ladder of at least eight Ub-TH conjugates could be identified in the molecular mass range higher than  66 kDa (Fig. 1, panel D), which also revealed a microheterogeneity corresponding to the enzyme as i solated (arrow i n Fig. 1D). The background in the control was negligible in this relevant area as earlier reported for this reconstituted in vitro assay [16]. Thus, the diagonal spots observed on the autoradio- gram (Fig. 1D) correspond to mono- and multi/poly Ub adducts, respectively, for the wt-hTH1 with  70 kDa,  78 kDa, etc. (molecular mass increasing by multiple of  8 kDa) and increasing pI. In a ddition, a series of predominant s pots with the same neutral pI as f ree Ub and increasing M r , observed in the absence (Fig. 1A) or presence of hydroxylase (Fig. 1B,C), was distributed in a periodic pattern co rresponding to poly Ub chains [16]. Finally, some insignificant amounts of poly/multi-ubiquiti- nated proteins, representin g ubiquitination o f not yet identified liver proteins, present in the E3 preparation [16], were also o bserved ( Fig. 1A). Microheterogeneity of wt-hTH1 and bTH as observed by two-dimensional electrophoresis The recombinant wt-hTH1 expressed in E. coli revealed a microheterogeneity on two-dimensional electrophoresis (Fig. 2A) with 5–6 components of  62 kDa, differing in pI by  0.1 pH unit. A similar type o f microhetero- geneity was observed (Fig. 2 B) when the enzyme was expressed (1 h at 37 °C) in an in vitro transcription- translation system as a protein of either  62 kDa or  60 kDa subunits, where the difference in molecular mass is explained by a second initiation site in this expression system [22]. When the membrane form of bovine TH (bTH mem ), extracted from i solated adrenal chromaffin g ranule ghosts, was subjected to two-dimensional electrophoresis, the Western blot a nalysis revealed a broad distribution pattern in terms of pI (Fig. 2C) with the apparent molecular mass of  60 kDa, a value typical of the subunit of bTH in chromaffin cells [7,9]. The streaky pattern of bTH mem (Fig. 2C) is characteristic of proteins with a tendency to aggregate/precipitate around the pI [23]. In addition, two post-translational modifications of bTH may also contri- bute to this pronounced microheterogeneity. Thus, the enzyme has four possible phosphorylation sites at Ser residues in the regulatory domain, and each phosphoryla- tion lowers the pI b y  0.1 U [24,25]. Deamidation o f labile amide groups has a similar effect on pI as shown for the structurally related phenylalanine hydroxylase [26] and most likely e xp lains the microheterogeneity of the nonphos- phorylated recombinant wt-hTH1 (Fig. 2A). Thus, hTH1 contains three aspargine residues of which Asn414 (posi- tioned in a short loop between two b strands) [14] is predicted to be the most labile one on the basis of its nearest neighbour amino acids (QNG), with a half-life of  1.5 days days [27] in Tris buffer, pH 7.0. Thus, similarly to hPAH [26,28], hTH1 also occurs in multiple molecular forms which could be explained by a progressive deamidation of labile Asn residue(s). Ubiquitination and degradation of hTH1 in the reticulocyte lysate system wt-hTH1 was expressed in the coupled in vitro transcrip- tion-translation ( TnT R ) syste m and the net accumulation of [ 35 S]hTH1 was followed as a function of time (Fig. 3A,B). The typical profile of [ 35 S]hTH1 on SDS _ PAGE revealed two major bands, corresponding to subunits of  62 kDa and  60 kDa, respectively. In Fig. 2. Microheterogeneity of recombinant human tyrosine hydroxylase (hTH1) an d the me mbrane-bound form o f the bov ine enzyme (bTH mem ) as revealed by 2D-electrophoresis. (A) R ecombinant hTH1 (40 lg) expressed in E. coli and visualized by Coomassie B rilliant Blue stain- ing. ( B) [ 35 S]Methionine-labelled hTH expressed in the in vitro transcription-translation system (10 lL assay) and dete cted b y a uto- radiography. (C) bTH mem of the bovine adrenal chromaffin granula membrane. (part a) two-dimensional profil of Coomassie Brilliant Blue stained membrane proteins (500 lg). ChgA (chromogranin A) and DBH (dopamine b-hydroxylase) represent the major spots as described previously [50,51]. The position of the multiple molecular forms of bTH are indicated by bracket as confirmed by immuno- blotting using ECL detection with 20 s (part b) a nd 5 min (part c) exposures. 1564 A. P. Døskeland and T. Flatmark (Eur. J. Biochem. 269) Ó FEBS 2002 addition, on prolonged exposure (> 60 min), 35 S-labelled proteins of molecular mass around 80 kDa were observed, concomittantly to the formation of the main product. Less defined 35 S-labelled proteins were also observed in the high-molecular-mass region (Fig. 3C), in amount enh anced by the presence of U b (20 l M ) and an inhibitor o f proteasome proteolytic activity [29,30], clasto-lactacystin b-lactone (2 m M ) dissolved in dimethylsulfoxide (1%) (data not shown), and were identified as post-transcriptionally modified TH such as Ub-conjugates. Furthermore, in the hTH1-expression sytem, 35 S-labelled peptides of  34 and 28–30 kDa were observed in increasing amount concom- itantly to the formation a nd subsequent decrease of [ 35 S]TH at longer incubation time points. Thus, degrada- tion of wt-hTH1 by components in the rabbit reticulocyte lysate influencing its proteolysis were further studied (Fig. 3C,D). MgATP-dependent degradation of wt-hTH1 [ 35 S]Methionine-labelled wt-TH1 was incubated with a reticulocyte lysate as the degradation mac hinery (Fig. 4). Reticulocytes do not contain any lysosomes [31] and any MgATP-dependent degradation correlates with proteo- somal activity [ 32]. In the presence of M gATP a s ignificant decrease in the a mount of wt-hTH1 was obs erved (Fig. 4A, lower i nset) while the hTH1 degradation was relatively moderate on depletion of MgATP (Fig. 4A, upper inset). The h alf-life of hTH1 disappearance was estimated to be of 7.4 h when the lysate was depleted for MgATP vs. 2.1 h in the presence of an ATP regenerating sys tem. Based on three independent experiments the MgATP-dependent proteoly- sis gave a half-life of  4.3 h (Fig. 4B). Furthermore, when clasto-lactacystin b-lactone (2 l M ) a nd anti-(26S protea- some) IgG (2 lLper50lL assay) were added to the degradation assay in the presence of an excess of MgATP, the MgATP-dependent degradation was reduced by  60%. This finding further supports the conclusion that proteasomes in the reticulocyte lysate are involved in t he degradation of TH. Fig. 4. MgATP-dependent degradation of hTH1. Semilogarithmic plot of the degradation of [ 35 S]methionine labelled full-length ( 62 kDa ) and truncated form (  60 kDa) of hTH1 p rotei n synthesize d by the coupled in vitro transcription-translation (reticulocyte lysate) system. After synthesis for 1.5 h at 30 °C, 1 m M cold me thionine was added and incubated at 37 °C in the presence of ex cess (j, h)ordepletionof MgATP (m) (fo r details, see Materials a nd method s). Aliquots w ere removed at t imed inter vals, and labelled hTH1 (  62 plus  60 kDa forms) was quantitated by INSTANT IMAGER or subjected to autoradiography a fter SDS _ PAGE. (A) The autoradiograms shown as inset represent experiments with excess of MgATP (lower inset) vs. depletion of MgATP (upper i nset). The half-lives for TH were esti- matedto7.4hintheassaydepletedinMgATP(m),andto2.1hwith excess of MgATP (j, h). (B) The MgATP-dependent degradation (total minu s MgATP- inde pendent) gave a h alf life of 4.3 h (mean of three experiments shown by separated symbols). The curves were drawn b y linea r regression analysis [A , r ¼ 0.809 f or curve (m)and r ¼ 0.961 f or curve ( j, h); B, r ¼ 0.872 for the curve]. Fig. 3. In vitro ubiquitination and degradation of [ 35 S]methionine labelled hTH1 in the reticulocyte lysate system. hTH1 was expressed in a coupled in vitro transcription-translation (reticulocyte lysate) system as described in Materials and metho ds. F rom the 50 lLreaction mixture 5 lL aliquots were quenched at given time points subjected to SDS _ PAGE and image analysis. ( A) Ze ro control; (B) the pattern of [ 35 S]methionine-labelled proteins after 30 min incubation; (C) repre- sents the image analysis profile of [ 35 S]methionine-labelled h TH1 at 150minincubationtime,and(D)theprofileobtainedafteranaddi- tional 3 h incubation in the presence of a regenerating system for ATP (degradation assay). The main products in (B) t o (D) co rrespond to proteins  62 and  60 kDa (i.e. full-length and truncated form of hTH1) and minor proteolytic products of  34 and 28–30 kDa. (C) T he area containing [ 35 S]methionine-labelled proteins which are considered to represent U b-conjugates of hTH1 are indicated by bracket. O, origin and F, dye fr ont. Th e value 1 o n the ordinate c or- responds to about 214 cpm in (A) and (B), to 76 c.p.m. in (C) and (D). Ó FEBS 2002 Ubiquitination of tyrosine hydroxylase (Eur. J. Biochem. 269) 1565 Immunodetection of Ub-TH conjugates in bovine adrenal chromaffin granule ghosts TH is also a t arget protein for ubiquitination in vivo,as shown by SDS/PAGE and Western blot analysis of proteins extracted from highly purified bovine chromaffin granule ghosts. Immunoblots with anti-Ub IgG revealed several Ub-conjugates of molecular mass ‡ 55 kDa, with the highest intensity near the top of the gel (Fig. 5A, lane 2). One o f t he bands revealed a mobility c orresponding to that of TH immunoreactivity ( 60 kDa), with a trace amount of reactivity a t the top of the gel (Fig. 5A, la ne 1 ). O n t wo- dimensional electrophoresis, t he U b-conjugates were found to be distributed over a l arge pI interval, m ostly d etected as a smear, especially dense in t he high-molecular-mass region. Most interestingly, a strong immunoreactivity was observed as a s eries of distinct s pots of 5 .0 < pI < 5.8 in t he 63 kDa region (Fig. 5B, panel 2) w hich revealed cross-r eactivity with anti-TH Ig (Fig. 5B, panel 1). The same pattern was obtained whether the membranes were fir st probed w ith anti-Ub Ig or anti-TH Ig, a nd then r eprobed with anti-TH Ig and a nti-Ub Ig, r espectively. Due to th e similarity in molecular mass b etween the proteins cross-reacting wi th anti-TH Ig and those reacting with anti-Ub Ig, it is a ssumed that the Ub-conjugates correspond to monoubiquitinated forms of TH. The c omigration of TH- and Ub-im munoreactivities was further studied in chromaffin granule membranes solubi- lized by the detergents SDS (1%, w/v) and Triton X-100 (2.5%, w/v) [33]. bTH mem was immunoisolated, in the presence of protease inhibitors, on anti-TH IgG coupled t o protein A–Sepharose beads (see Materials and methods). After resolution on one dimensional SDS/PAGE and Western blot analysis, TH of  60 kDa and some higher molecular mass forms w ere detected with anti-TH IgG. The immunoisolates were also probed with the anti-Ub Ig, and positive immunoreactivity w as then observed at  60 kDa comigrating with TH immunoreactivity (data not shown), and thus further support the results shown in Fig. 5B. DISCUSSION The ubiquitination of many vital p roteins plays an import- ant role in regulating their functions and turnover in eukaryotic cells including neurons and n euroendocrine cells [34,35]. In the present study it is shown that a key regulatory enzyme in catecholaminergic neuroendocrine cells, i.e. tyrosine hydroxylase, is a substrate for the Ub-conjugating enzyme system, both in vitro as a soluble recombinant human enzyme and in vivo as a membrane-bound form of the enz yme i n t he c atecholamine storage/secretory gra nules of the a drenal medulla, w hich may h ave important functional implications in the central nervous system. Ubiquitination of the soluble recombinant hTH1 The finding that recombinant hTH1 is a substrate for the in vitro reconstituted Ub conjugating enzyme system of rat liver (Fig. 2) was indeed expected as the structurally homologous enzyme phenylalanine hydroxylase and its catalytic core enzyme [14] have already been found to be ubiquitinated [16]. Similarly, the in vivo turnover of the structurally homologou s enzyme tryptophan h ydroxylase [14,15] is reported to be mediated by the Ub-proteasome pathway [36]. Furthermore, ou r previous studies on two mutant forms of hTH1, associated with the clinical and metabolic phenotype of L -DOPA responsive dystonia and infantile p arkinsonism, h ave revealed a reduced cellular stability c ompared t o the wild-type form when expressed in human embryonic kidney (A293) cells [37] supporting the in vivo relevance of the observed Ub-conjugates of hTH1 formed in vitro. Thus, elimination of proteins by the Ub-proteasome pathway is co nsidered to be most active towards misfolded/misassembled and abnormal mutant proteins [38]. Energy-dependent degradation of recombinant hTH1 in the in vitro reticulolysate system Further e xp erimental e vidence in s upport of d egradation of hTH1 by the Ub-proteasome pathway was obtained in stability studies of recombinant hTH1 in the in vitro reticulolysate system (Fig. 3). Reticulocyte lysates have been used as the degradation machinery and are especially well suitable to study ubiquitin-dependent proteasomal degradation of specific proteins [32]. Indeed, reticulocytes Fig. 5. Immunodetection of bTH mem and ubiquitin-conjugates in chromaffin granule membranes. (A) C hromaffin granule ghost proteins (130 lg) were su bjected to SDS/PAGE and immunoblotted w ith antibodies directed against TH (lane 1) or Ub (lane 2) and 125 I-Protei n A as described in the Materials and methods section. O, origin; F, front. (B) C hromaffin g ranule ghost p roteins ( 800 lg) were subjected to two-dimensional electrophoresis and imm unob lotted with anti- bodies directed against hTH1 (panel 1) or Ub (panel 2). Ub-conjugates were first precisely localized on two-dimensional gel electrophoresis by immunoblotting with anti-U b Ig. The membrane was th ereafter stripped an d then reprobed with anti-TH Ig. A s shown in panel 2, Ub-conjugates were detected at the same position as bTH (panel 1). The immunoblotting procedure has be en repeated using first anti-TH and then anti-Ub Ig which also revealed colocalization of the two immunoreactivities. The time of expo sure (memb rane to film) w hich was required to detect Ub-conjugates with anti-Ub Ig was usually twofold to t hreefold longer than that required to d etect TH immuno- reactivity. 1566 A. P. Døskeland and T. Flatmark (Eur. J. Biochem. 269) Ó FEBS 2002 contain m ultiple proteolytic systems including the MgATP- ubiquitin-proteasome-dependent pathway, calpains and MgATP-independent proteases, but they contain no lyso- somes [31]. Due to the lack of lysosomal activities, the system has one limitation as compared to regular eukaryotic cells. In order to clearly distinguish proteasomal activity from other proteolytic activities, an established effective concen- tration (2 l M ) in vitro of the selective and potent protea- somal inhibitor clasto-lactacystin b-lactone was used in the present study. The inhibitor is a lactacystin analog with a potency of  10 times that of lactacystin and which beside epoxomycin [39] is the most potent and specific proteasome inhibitor [29,30,40]. In this coupled transcription-translation system a time-dependent formation of [ 35 S]methionine- labelled hTH1 w as observed, followed by i ts degradation to molecular species of  34 and 28–30 kDa, which is related to the 34-kDa core fragment of hTH1 observed on limited tryptic proteolysis [4 1]. Its degradation was found to be partly MgATP-dependent which was inhibited to about 60% by anti-(26S proteasome) IgG plus clasto-lactacystin- b-lactone. The o verall half-life of [ 35 S]methionine-labelled h TH1 was estimated to 2.1 h in the p resence of an ATP- regenerating system and to 7.4 h when the lysate was depleted of MgATP (Fig. 4). By comparison, the half-life of rat TH estimated in PC-12 cells, in a subclone of PC-12 cells and in chromaffin cells has been reported to be 17 h [10], 30 h [11] and 29 ± 3 h [12], respectively. The shorter half- lifes observed in the present reconstituted in vitro system may be explained in several ways, including a stabilization of TH in the cells as a result of i ts binding to membranes [9] and to cytosolic proteins, e.g. the 14.3.3 proteins [42], as discussed below. That TH is a substrate for the Ub-conjugating enzyme system also in vivo is further supported by our finding of mono-ubiquitinated bTH mem in highly purified bovine adrenal chromaffin granule ghosts (Fig. 5) and for the first time underlines a physio logical role of this post-translation- sal modification. Although the TH- and Ub- immunoreac- tivities revealed a comigration in the t wo selected (one- and two-dimensional) electrophoretic systems we do not know if all the TH subunits in the oligomer are ubiquitinated. A similar observation has been made for the ubiquitinated form of the m embrane-bound form o f nitric oxide s ynthase on SDS/PAGE [43]. bTH mem behaves as an i ntegral membrane protein [7,9], but the mechanism b y which it is bound is not yet resolved. Interestingly, it has been suggested that ubiquitination might promote a structural change (unfolding) of linked proteins [44], but it is not possible at this point to answer the question of whether bTH mem is inserted into the membrane before or after its ubiquitination. The finding that bTH mem is ph osphorylated by cAMP-dependent protein kinase on Ser40 in the regulatory domain [9] may support an ubiquitination of the e nzyme by the cytosolic Ub-conjugating enzyme system after its membrane insertion. In contrast to the multi/poly Ub conjugates observed for the soluble recombinant hTH1 the membrane-bound form of bTH is mono-ubiquitinated, which m ay be related to t he function of the u biquitin C-terminal h ydrolase (UCH-L1 o r PGP9.5) which is widely and often highly expressed in neuroendocrine cells [34,35], including the rat chromaffin cells [45]. From a functional point of view, the membrane localization may protect the catalytically active enzyme from degradation by t he cytosolic proteases. Thus, ubiqui- tination may play a role in the degradation of both membrane-bound and soluble TH. However, the accurate role of the ubiquitination remains t he subject of further investigation and the reason why TH is detected mainly as mono-ubiquitinated form is still unclear. Ubiquitination of proteins in the chromaffin granule membrane Previous studies on subcellular fractions of rat brain homogenates have revealed that the synaptic membrane fraction contains multiple Ub-immunoreactive bands, i.e. Ub-conjugates of 105 , 72, 60, 41 and 38 kDa, and the majority of the conjugates were found to be integral membrane proteins including some high-molecular-mass glycoproteins [46]. As the synaptic membrane fraction represents a mixture of different types of membranes, the functional significance of this finding is not clear. The specific localization of Ub-conjugates to secretory granules may suggest a function either in membrane fusion events or in the turnover of the o rganelles. Thus, it has been reported that an Ub-like conjugating enzyme system is involved in homotypic membrane fusion in Pichia pastoris [47]. Alter- natively, Ub appended to membrane proteins may represent a signal which promotes the selective sorting/sequestration and turnover of the s ecretory granules by autophagosomes as already shown for the regulated turnover of other cytoplasmic organelles (reviewed in [48]). ACKNOWLEDGEMENTS The study was supported by grants from the Research Council of Norway, from t he Novo Nordisk Foudation, The Nansen Fund, T he Blix Family Fund f or t he Advancement of Medical R esearch and the Norwegian Council on Cardiovascular Diseases. We greatly appreciate the expert technical assistance of Sidsel Riise for the preparation of hTH1 an d that o f S is sel V ik Berg e f or the p rep aration o f c hromaffin granule ghosts. REFERENCES 1. Nagatsu, T., Levitt, M. & Ud enfrie nd, S. ( 1964) Tyrosine hydro- xylase. The initial step in norepinephrine biosynthesis. J. Biol. Chem. 23 9, 2910–2917. 2. Kaufman, S. (1995) Tyrosine hydroxylase. Adv. Enzymol. Relat. Areas. Mol . Biol. 70 , 103–220. 3. Flatmark, T. (2000) Catecholamine biosynthesis and physiological regulation in neuroendocrine cells. Acta Physiol. Scand. 168, 1–17. 4. Kuc zenski, R.T. & Mandell, A.J. (1972) Regulatory properties of soluble and particulate rat brain tyrosine hydroxylase. J. Biol. Chem. 24 7, 3114–3122. 5. H aycock, J.W., George, R.J. & Wa ymire, J.C. (1985) In situ phosphorylation of tyrosine hydroxylase in chromaffin cells: localization to soluble compartments. Neurochem. Int. 7, 301–308. 6. Haavik, J., Andersson, K.K., Petersson, L. & Flatmark, T. (1988) Soluble tyrosine hydroxylase (tyrosine 3-monooxygenase) from bovine adrenal m edulla: large-scale purification and physico- chemical p roperties. Biochim. Bio phys. Acta 953, 142–156. 7.Treiman,M.,Weber,W.&Gratzl,M.J.(1983)3¢,5¢-Cyclic adenosine m onophosp hate- and Ca 2+ -calmodulin-depe ndent endogenous protein phosphorylation activity in membranes of the bovine chromaffin secretory vesicles: identification of t wo Ó FEBS 2002 Ubiquitination of tyrosine hydroxylase (Eur. J. Biochem. 269) 1567 phosphorylated components a s tyrosine hydroxylase and protein kinase regulatory subunit t ype II. J. Neurochem. 40, 6 61–664. 8. Morita, K ., Teraoka, K. & Oka, M. (1987) I nteraction of cytoplasmic tyrosine hyd roxyla se with chromaffin granule. In vitro studies on association of soluble enzyme with granule membranes and alteration in enzyme activity. J. Biol. Chem. 262, 5654–5658. 9. Kuhn, D.M., Arthr, R. Jr, Yoon, H. & Sankaran, K. (1990) Tyrosine hydroxylase in secretory granules from adrenal medulla. Evidence for an integral m embrane form. J. Biol. Chem. 26 5, 5780–5786. 10. Wu, D.K. & Cepko, C.L. (1994) The stability of endogenous tyrosine h ydroxylase protein in PC-12 cells differs from that expressed in mouse fibroblasts by gene transfer. J. Neurochem. 62, 863–872. 11. Tank, A.W., H am, L. & Curella, P. (1986) In duction of tyrosine hydroxylase by cyclic AMP and glucocorticoids in a rat pheo- chromocytoma cell line. effect o f t h e inducing ag ents alone or in combination on the enzyme levels and rate of synthesis of tyrosine hydroxylase. Mol. Pharmacol. 30 , 486–496. 12. Fernandez, E. & Craviso, G .L. (1999) Protein synthesis blockade differentially a ffects the degradation of constitu tive a nd nicotinic receptor-induced tyrosine hydroxylase protein level in isolated bovine c hromaffin c ells. J. Neurochem. 73, 1 69–178. 13. Pasinetti, G.M., Osterburg, H.H., Kelly, A.B., Kohama, S., Morgan, D.G., Reinhard, J.F. Jr, Stellwagen, R.H. & Finch, C.E. (1992) Slow changes of tyrosine hydroxylase g ene expression in dopaminergic brain neurons after neurotoxin lesio ning: a model for neuron aging. Mol. Brain . Res. 13, 63–73. 14. Flatmark, T. & Ste vens, R.C. (1999) Structural insights into the aromatic amino acid hydroxylase and their disease related mutant forms. Chem. Rev. 99 , 2137–2160. 15. Grenett, H.E., Ledley, F.D., Reed, L.L. & Woo, S.L. (1987) Full- lenght cDN A f or r abbit tryptophan hydroxylase: f unctional domains and evolution of aromatic amino acid hydroxylases. Proc. N atl Acad. Sci. USA 84 , 5530–5534. 16. Døskeland, A.P. & Flatmark, T. (2001) Conjugation of phenyl- alanine hydroxylase with polyubiquitin chains catalysed by rat liver e nzymes. Biochim. Biophys. Acta 1 547, 379–386. 17. Haavik, J., Le Bourdelles, B., Martı ´ nez, A., Flatmark, T. & Mallet, J. (1991) Recombinant human tyrosine hydroxylase iso- zymes. Reconstitution with iron and inhibitory effect of other metal ions. Eur. J. Biochem. 19 9, 371–378. 18. Le Bourdelle ` s, B., Horellou, P., Le Caer, J P., Dene ` fle, P., Latta, M., Haavik, J., G uibert, B., Mayaux, J F. & Mallet, J. (1991) Phosphorylation of human recombinant tyrosine hydroxylase isoforms 1 and 2: an additional phosphorylated residue in isoform 2, generated through alternative splicing. J. Biol. Chem. 266, 17124–17130. 19. Sagne, C., Isambert, M F. & Henry, J . (19 96) -P. & Ga snier, B. SDS-resistant aggregation of membrane proteins in application to the purification of vesicular monoamine transporter. Biochem. J. 316, 8 25–831. 20. Terland, O. & Flatmark, T. (1980) Oxidoreductase activities o f chromaffin granule ghosts isolated f rom the bovine adrenal medulla. Bio chim. Biophys. Acta 597, 318–330. 21. Laemmli, U.K. (197 0) Cleavage of structural proteins during t he assembly of the head of b acteriophage T4. Nature 227, 680–685. 22. Lu ¨ decke, B., Knappskog, P.M., Clayton, P.T., Surtees, R.A.H., Clelland, J .D., Heales, S.J., Brand, M.P., Bartholome ´ ,K.& Flatmark, T. (1996) Recessively inherited L -DOPA-responsive parkinsonism in infancy caused by a point mutation (L205P) i n the tyrosine hydroxylase g ene. Hum. Mol. Genet. 5, 102 3–1028. 23. Gørg,A.,Obermaier,C.,Boguth,G.,Harder,A.,Scheibe,B., Wildgruber, R. & Weiss, W. (2000) The current state of two- dimensional electrophoresis with immobilized pH gradients. Electrophoresis 21, 1037–1053. 24. Coˆte ´ , A., Doucet, J P. & Trifaro, J M. (1986) Phosphorylation and dephosphorylation of c hr omaffin cell p ro teins in r espo nse to stimulation. Ne uroscience 19, 629–645. 25. Pocotte, S.L., H olz, R.W. & Ue da, T. (1986) Cholinergic receptor- mediated phosphorylation a nd activation of tyrosine hydroxylase in cultured bovine adrenal chromaffin cells. J. Neurochem. 46, 610–622. 26. Solstad, T. & Flatmark, T. (2000) Microheterogeneity of recombinant human phenylalanine hydroxylase as a result o f non-enzymatic deamidations of labile amide containing amino acids. Effects on catalytic and stability properties. Eur. J. Biochem. 267, 6302–6310. 27. Robinson, N.E. & Robinson, A.B. (2001) Prediction of protein deamidation rates from p rim ary and three-dimensional structure. Proc. N atl Acad. Sci. USA 98 , 4367–4372. 28. Carvalho, R .M.N., Solstad, T ., Robinson, N .E., Robinson, A.B. & Flatmark, T. (2002) Pos sible contributions of labile a sparagine residues to differences in regulatory properties of human and rat phenylalanine hydroxylase. In Chemistry and Biology of Pteridines and Folates 2001 (Parrak, V. & Reibnegger, G., eds), Kluwer, USA. In press. 1 29. Fenteany,G.,Standaert,R.F.,Lane,W.S.,Choi,S.,Corey,E.J.& Schreiber, S.L. (1 995) Inhibition of proteasome activities and subunit-specific amino-terminal threonine modification by lacta- cystin. Science 268, 726–731. 30. Craiu, A., Gaczynska, M., Akopian, T., Gramm, C.F., Fenteany, G., G oldberg, A.L. & R ock. K.L. (1997) Lac tacystin and cl asto- lactac yst in b-lactone modify mu ltiple proteasome b-subunits and inhibit intracellular protein degradation and m ajor histocompa- tibility complex class I antigen presentation. J. Biol. Chem. 272, 13437–13445. 31. Watson, A.L., Laszlo, L. & Doherty, F.J. (1991 ) The de gradative fate of ubiquitin-protein conjugates in nucleated and enucleated cells. Acta Biol. Hung. 42, 4 9–56. 32. Hershko, A. (1988) Ub iquitin-mediated protein degradation. J. Biol. Chem. 263, 15237–15240. 33. Stoller, T.J. & Shields, D. ( 1988) Retrovirus-mediated e xpression of preprosomatostatin: posttranslational p rocessing, intracellular storage, and sec retion in GH 3 pituitary cells. J. Cell Biol. 107, 2087–2095. 34. Doran, J.F., Jackson, P., Kynoch, P.A. & Thompson, R.J. (1983) Isolation of PGP 9.5, a new human neuron-specific protein detected by high-resolution two-dimensional electrophoresis. J. Neurochem. 40, 1 542–1547. 35. Wilkinson, K.D., L ee, K ., Deshpande, S., D uerksen-Hughes, P., Boss, J.M. & Pohl, J. (1989) The neuron-specific protein PGP 9.5 is a ubiq uitin carboxyl-terminal hydrolase. Sci ence 246, 670 –673. 36. Kojima, M., Oguro, K., Sawabe, K., Iida, Y., Ikeda, R., Yamashita, A., Nakanishi, N. & Hagesawa, H. (2000) Rapid turnover of trypto phan hydro xylase i s d riven b y prote asomes in RBL2H3 cells, a serotonin p roducing mast cell line. J. Biochem. 127, 121–127. 37. Flatmark, T., Knappskog, P.M., B jørgo, E. & Martı ´ nez, A. (1997) Molecular characterization of disease related mutant forms of human phenylalanine hydroxylase and tyrosine hydroxylase. In Chemistry and Biology of Pteridines and Folates (Pfleiderer,W.& Rokos, H. , eds), pp. 503–508. Blac kwell Science, Oxford, UK. 38. Wickner, S., Maurizi, M. & Gottesman, S. (1999) Posttransla- tional quality control: fo ld ing, refolding, and degrading prote ins. Science 28 6, 1888–1893. 39. Meng,L.,Mohan,R.,Kwok,B.H.B.,Elofsson,M.,Sin,N.& Crews, C.M. (1999) Epoxomicin, a potent and selective protea- some inhibitor, exhibits in vivo anti-inflammatory activity. Proc. Natl Acad. S ci. USA 96, 1040 3–10408. 40. Dick, L.R., Cruikshank, A.A., Destree, A.T., Grenier, L., McCormack, T.A., Melandri, F.D., Nunes, S .L., Palombella, V .J., Parent, L.A., Plamondo n, L. & Stein, R.L. (1997) Mechanistic 1568 A. P. Døskeland and T. Flatmark (Eur. J. Biochem. 269) Ó FEBS 2002 studies on the inactivation of the proteasome by lactacystin in cultured cells . J. Bio l. Chem. 27 2, 182–188. 41. M artı ´ nez, A., Haavik, J., Flatmark, T., Arrondo, J.L.R. & Muga, A. (1996) Conformational properties and stability of tyrosine hydroxylase studied by infrared spectroscopy. Effect of iron/ catecholamine b inding and phosphorylatio n. J. Bi ol. C hem. 271, 19737–19742. 42. Itagaki, C., Isobe, T., Taoka, M., Natsume, T., Nomura, N., Horigome,T.,Omata,S.,Ichinose,H.,Nagatsu,T.,Greene, L.A. & Ichimura, T. (199 9) Stimulus–coupled interaction of tyrosine hydroxylase with 14–3)3 proteins. B i oche mistr y 38, 15673–15680. 43. Bender, A.T., Demady, D.R. & Osawa, Y. ( 2000) Ubiquitination of neuro nal nitric-o xide synt hase in vitro and in vivo. J. Biol. C hem. 275, 1 7407–17411. 44. Pickart, C.M. (1998) Polyubiquitin chains. In Ubiquitin and the Biology of the Cell (Peters, J. -M., H arris, J.B. & Finley, D., eds), pp. 19–63. P lenum Press, N ew York. 45. Kent, C. & Coupland, R.E. (1989) Localisation of Chromo- granin A and B, met-enkephalin-arg 6 -gyl 7 -leu 8 and PGP9.5-like immunoreactivity in the developing and adult rat adrenal medulla and extra-adrenal c hromaffin tissue. J . Anat. 166 , 213–225. 46. Beesley, P.M., Mummery, R., Tibaldi, J., Chapman, A.P., Smith, S.J. & Rider, C.C. (1995) Th e post-synap tic density: putative involvement in synapse stabilization via cadherins and covalent modification by ubiquitination. Biochem. So c. Tra ns. 23, 5 9–64. 47. Yuan, W., Stromhaug, P.E. & Dunn, W.A. (1999) G lucose- induced autophagy of peroxisomes in Pichia pastoris requires a unique E1-like protein. Mol . Biol. C ell 199, 1353 –1366. 48. Klionsky, D.J. & Emr, S.D. (2000) Autophagy as a regulated pathway of cellu lar degradation. Sc ience 29 0, 1717–1721. 49. Cook, J.C. & Chock, P.B. (1992) Isoforms of mammalian ubiquitin-activating enzyme. J. Biol. C hem. 267, 2 4315–24321. 50. Bad er, M.F., Hikita, T . & Trifaro, J .M. (1985) Calcium-dependent calmodulin binding to chromaffin granule membranes: presence of a 65-kilodalton calmodulin binding protein. J. Neurochem. 44, 526–539. 51. Winkler, H., Apps, D.K. & Fi scher-Colbrie, R. (1986) The molecular function of adrenal chromaffin granules: established facts and unresolved t o pics. Neuroscience 18 , 261–290. Ó FEBS 2002 Ubiquitination of tyrosine hydroxylase (Eur. J. Biochem. 269) 1569 . Ubiquitination of soluble and membrane-bound tyrosine hydroxylase and degradation of the soluble form Anne P. Døskeland and Torgeir Flatmark Department. characterization of disease related mutant forms of human phenylalanine hydroxylase and tyrosine hydroxylase. In Chemistry and Biology of Pteridines and Folates