Tyrosine Hydroxylase in Rat Brain Dopaminergic Nerve Terminals MULTIPLE-SITE PHOSPHORYLATION IN VIVO AND IN SYNAPTOSOMES*

Tyrosine hydroxylase, which catalyzes the initial step in catecholamine biosynthesis, is phosphorylated at serines 8, 19, 31, and 40 in intact pheochromocy- toma (PC12) cells After of rat corpus striata in vivo rat corpus striatal synapto- somes, 32P incorporation into tyrosine hydroxylase occurred predominantly at serines 19, 31, and 40. Elec- trical stimulation (30 Hz, 20 min) of the medial forebrain bundle (containing the afferent dopaminergic fibers) increased 32P incorporation into each of the three sites. Brief depolarization of the synaptosomes with elevated [K’l0 (20-60 mM, 6-30 s) or veratridine (60 NM, 2 min) produced a selective increase in 32P incorporation into Ser”. Phorbol 12,13-dibutyrate (1 PM, 6 min) increased 32P incorporation into Ser31, and CAMP-acting agents such as forskolin (10 NM, 5 min) increased 32P incorporation into Ser40. In 32P incorporation into Ser’, which was usually detectable but very low, was not regulated either in or in situ by any of the activators of signal transduction pathways. In synaptosomes, the only found to increase

the initial enzyme in the biosynthesis of catecholamines (3), and its activity appears to be rate-limiting therein. Recently, TH has been shown to be a substrate in vitro for a number of protein kinases, and the phosphorylation of TH by each of these protein kinases can lead to an increase in the catalytic activity of TH (cf. Ref. 4). TH is also phosphorylated in situ, and secretagogues increase both the phosphorylation and activity of T H (2,5,6).
Multiple-site phosphorylation of TH has been demonstrated in situ for bovine adrenal medullary chromaffin cells, rat PC12 cells, and rat superior cervical ganglia (7)(8)(9). Serine residues comprise the phosphate acceptor sites in each of the various tryptic phosphopeptides that have been isolated from rat TH, and these residues have recently been identified as serines 8, 19, 31, and 40 (10). However, in each of these catecholaminergic tissues, the proportion of TH in compartments relevant to the regulation of catecholamine biosynthesis is either small (superior cervical ganglion) or unknown (chromaffin cells and PC12 cells). Thus, it has been difficult to determine whether the observed phosphorylation and activation of T H represent biochemical events related to catecholamine biosynthesis in subcellular compartments relevant to secretion.
The analysis of TH phosphorylation and activity in neuropil or in isolated brain synaptosomes should essentially eliminate the contribution of TH from soma1 compartments. Several laboratories have described the regulation of catecholamine biosynthesis in synaptosomes from catecholaminerich brain areas such as the corpus striatum (e.g. . Depolarization of striatal nerve terminals increases dopamine biosynthesis and activates T H (e.g. Refs. [14][15][16]. And, this activation is similar to that produced by conditions which promote calcium-dependent protein phosphorylation (17). Until very recently (18)(19)(20), however, nothing was known regarding the phosphorylation of TH in striatal terminals. The present studies characterize the multiple-site phosphorylation of T H in dopaminergic nerve terminals, both in uiuo and in synaptosomes, and establish the site-specificity of stimulation-dependent increases.

MATERIALS AND METHODS~
Most of the methods and materials used in the present study have been describedpreviously in detail (10). Salient methods are described in brief immediately below, and details of additional or modified methods are presented under "Miniprint Supplement." 32P Labeling of Rat Corpus Striatum in Viuo-Carrier-free 32Pi (0.5-* Portions of this paper (including part of "Materials and Methods," part of "Results," Fig. 1 and Tables I and VI) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.
2 mCi) in 150 mM NaCl (5 pl) was delivered simultaneously into each caudate-putamen of anesthetized rats over 30 min via stereotaxically placed cannulae. The cannulae were withdrawn and a bipolar electrode was placed unilaterally into the medial forebrain bundle. Sixty min after the infusion of 32Pi, electrical stimulation (30 Hz, 200 pA, 3 ms, biphasic square wave) was applied for 20 min. At the end of the stimulation period, and with stimulation still on, rapid cryofixation of the corpus striata in situ was achieved by pressure injection of liquid nitrogen into the cranium. Heads were removed and immersed in liquid nitrogen, after which the forebrains were removed, warmed to -14 to -16 "C, and sectioned (1 or 2 mm thickness). Tissue punches from the vicinity of the cannulae tips were solubilized in hot 1% SDS and heated for 5 min in a boiling water bath.
'"P Labeling of Corplcs Striatal Synaptosomes-Pellets of crude or purified synaptosomal fractions were resuspended in incubation solution (150 mM NaC1, 15 mM Hepes, 5.5 mM D-glucose, 4.4 mM KCl, 1.2 mM MgCl,, 1.0 mM CaCI,, p H 7.4, with NaOH a t room temperature and gassed with hydrated 100% 0,) at 1 or 2 mg protein/ml and incubated for 15 min a t 37 "C. Carrier-free 32Pi was added to a final concentration of 2 or 4 mCi/ml and the incubation was continued for 45 min. Aliquots (100 pl) of the tissue were added to 100 pl of prewarmed incubation solution that contained test substances a t twice the desired final concentration. Treatment was terminated by adding 20 p1 of 10% SDS, 10 mM Tris-EDTA, p H 8, and samples were heated in a boiling water bath for 2-5 min.
Analysis of :12P Incorporation-Aliquots of the solubilized samples were taken for determinations of total protein, 32P incorporation into total protein and T H protein levels. 32P-labeled T H was isolated by SDS-PAGE after immunoprecipitation from the remainder of the sample and, in some cases, transferred electrophoretically to nitrocellulose. ' "P incorporation into T H was quantified by liquid scintillation counting of gel slices or Cerenkov counting of nitrocellulose pieces that were excised after autoradiographic localization of the '"P-labeled T H bands.
Tryptic Digestion and Phosphopeptide Separation-"P-Labeled peptides released from the gel slices or nitrocellulose pieces by limit tryptic digestion were separated by reverse-phase HPLC on a C I S column equilibrated in 0.1% trifluoroacetic acid. The 32P-labeled peptides were eluted a t 1 ml/min with an acetonitrile gradient (0.1 or O.S%/min) in 0.1% trifluoroacetic acid. An on-line detector (Radiomatic) provided "P peak integration and collection. 32P peaks were collected separately and concentrated in a Speed-Vac (Savant) prior to subsequent analysis.

Identity between TH Phosphorylation Sites
in Rat Brain and PC12 Cells When striata were prelabeled in uiuo with 0.5-2 mCi 32Pi, a single M, -62,000 32P-labeled band was immunoprecipitated from solubilized striata by affinity-purified rabbit anti-TH ( Fig. 1, left). Similar results have been presented for 32Plabeled TH from striatal synaptosomes (see Fig. 1 in Ref. 18). In both cases, 32P incorporation into T H could not be discerned with one-dimensional SDS-PAGE separations without immunoprecipitation. Immunoblots developed with affinitypurified sheep anti-TH also reacted with a single M, -62,000 band in striatal tissue processed according to either the in vivo protocol (Fig. 1, right) or the synaptosomal protocol (not presented).
When sufficient 32P incorporation was present in tryptic digests of immunoprecipitated T H after 32P-labeling in uiuo, the resulting "P-labeled peptides were separated either by reverse-phase HPLC or two-dimensional electrophoresis/ chromatography. With reverse-phase HPLC of the tryptic digest, five peaks of radioactivity were detected ana named according to their order of elution, as shown in Fig. 2.4 (left). As shown in Fig. 2B (left), the elution profile of 32P-labeled tryptic TH phosphopeptides after labeling striatal synaptosomes with "Pi was identical to that seen after labeling in uivo. Furthermore, these profiles were identical to that of T H phosphopeptides labeled in intact PC12 cells (cf. Fig. 6 in Ref. 10). Based on these separation patterns and additional data below, the phosphopeptides from TH labeled in rat corpus striatum in uiuo and in synaptosomes appeared to be identical to those in TH after labeling intact PC12 cells with 32Pi. Thus, by virtue of the sequence data from the PC12 TH phosphopeptides (lo), the following phosphorylation sites were inferred CS-1 and CS-2, Serlg; CS-3, Ser31; CS-4, Sera; and CS-5, Ser4' (Fig. 2C).
Hydrolysis (110 "C, 2 h, 6 N HCl) of each of the five peaks separated by reverse-phase HPLC revealed phosphoserine but not phosphothreonine or phosphotyrosine (not presented). In two-dimensional tryptic fingerprints (Figs. 2, A and B, right), five spots having the same pattern as those from PC12 TH were separated.
The facility with which larger amounts of 32P incorporation into TH could be achieved with the synaptosomal preparations allowed additional characterization of the peptides to support their identities with those from TH labeled in intact PC12 cells. For example, mixing experiments with the eluted peaks prior to two-dimensional electrophoresis/chromatography allowed identification of the five phosphopeptide peaks in reverse-phase HPLC in terms of their migration in the two-dimensional fingerprint system (Fig. 2B, right). Thin-layer isoelectric focusing of the striatal TH phosphopeptides in Servalyt 3-10 Precoat gels allowed the following PI values to be assigned-CS-1, -3.1; CS-2, -3.8; CS-3, 3.5; CS-4,5.0-5.5; CS-5,4.0-4.2. This compares with the following values obtained from PC12 cells: PC-1, -3.2; PC-2, -3.9; PC-The phosphopeptides were subjected to six cycles of manual Edman degradation, and subtractive analyses were performed to identify the location of phosphoserine residues relative to the NHz-terminal of the peptide and to test for the presence of multiple phosphorylation sites within a given peptide (10). After each cycle of degradation, separate aliquots of the reaction products were subjected to electrophoresis at pH 8.9 and 1.9. A single phosphorylation site was indicated for CS-1, CS-2, and CS-5 at positions 3, 4, and 3, respectively. No phosphorylation sites were revealed in either CS-3 or CS-4 for the six cycles. The phosphorylation sites in the PC peptides are at 3,4,7,7, and 3 for PC-1 through PC-5, respectively (10). In addition, the changes in mobilities of the different CS peptides that occurred in a cycle-dependent fashion (not presented) were identical to those occurring with the PC peptides (10).
A final similarity between the CS and PC peptides was demonstrated by subjecting the tryptic peptides to additional proteolysis with other endoproteinases and then rechromatography by reverse phase HPLC. These data are presented in Table 11.

32P Incorporation into TH in Vivo
The effects of electrical stimulation of the medial forebrain bundle, using stimulation parameters previously shown to activate T H and increase catecholamine biosynthesis rates (14,15), on overall 32P incorporation into TH are shown in Table 111. Electrical stimulation increased the relative 32P incorporation into TH in 16 of 17 rats. The average increase was 91% (S.E. = 18%).

( M ) P T P S A P S P Q P K G F R R A V S E Q D A K Q A E A V T S P R F I G R R Q S L I E D A R K PC-4 PC-
Separation of tryptic TH phosphopeptides after labeling in vivo or in synaptosomes with 32Pi. A (left), '"P-labeled TH was immunoprecipitated from solubilized corpus striatum after labeling with ?' Pi and subjected to SDS-PAGE. The "P-labeled TH band was digested with trypsin and applied to a C18 column equilibrated in 0.1% trifluoroacetic acid. The phosphopeptide peaks were eluted with a 0.1% acetonitrile/min gradient (dashed line). Right, lyophilized tryptic digests were spotted on thin-layer cellulose plates and electrophoresed a t pH 8.9. After electrophoresis, the plates were dried and chromatographed orthogonal to the direction of electrophoresis. The spots on the autoradiograms are labeled to indicate the corresponding peak from reversephase HPLC by analogy to the peptides isolated from "P-labeled T H in synaptosomes. 0, origin. B (left), Labeled T H was immunoprecipitated from solubilized striatal synaptosomes after labeling with ' lPi and then subjected to SDS-PAGE. Tryptic phosphopeptides were separated as in ( A ) . Data similar to these were also obtained after labeling synaptosomes that had been purified on Percoll gradients. Right, phosphopeptide peaks separated by reverse-phase HPLC (0.2% acetonitrile/min) were collected and concentrated. Aliquots of each peak containing approximately equal counts/min were combined and subjected to two-dimensional electrophoresis/ chromatography as in ( A ) . The relative migration of a given peptide was determined by inspection of autoradiograms of fingerprints from which a particular peak was individually omitted. The spots on the autoradiograms are labeled to indicate the corresponding peak from reverse-phase HPLC. The greater density of the spot labeled CS-5 resulted in part from having more CS-5 radioactivity and in part from degradation of CS-3 to CS-3'." 0, origin. C, amino acid sequence of the NH2 terminus of rat TH. Tryptic phosphopeptides from PC12 TH that were sequenced (10) are labeled and underlined.

9)
. Electrical stimulation of the medial forebrain bundle increased R2P incorporation into three of the four sites: Ser", 236 ? 50; Ser3', 176 & 36; and Ser4', 194 & 42 (mean percent of control k S.E., n = 9). '*P incorporation into SerR from both the stimulated and unstimulated samples was quantifiable in only five of the nine rats. In these five pairs of samples, "'P incorporation into Ser' in the stimulated samples was 109 & 22% of control.
During the several years over which the procedures for direct side-to-side comparisons of 32P incorporation into each As previously described for PC-3, adventitious chemical modification of CS-3 presumably involving cyclization of the NHZ-terminal Gln to pyroGlu during processing of the tryptic peptides results in a phosphopeptide ((3-3') which elutes immediately prior to CS-4 during reverse-phase HPLC and which comigrates with CS-5 in pH 8.9 two-dimensional fingerprints (10). Given the low ?lP incorporation into CS-4 (as opposed to PC-4), proper quantitation of CS-4 required that the CIR column be eluted at no greater than 0.2% acetonitrile/ min. When present, counts in CS-3' were combined with those in CS-3 to arrive at the values for "P incorporation into Ser"'. of the TH phosphopeptides were being developed, the effects of electrical stimulation were evaluated by expressing :%'P incorporation into each of the phosphopeptides (that could be quantified) relative to CS-3, the largest peak in unstimulated samples. It was noted and reported previously (21) that electrical stimulation decreased the ratio of 02P incorporation into CS-1 relative to that in CS-3. In the data from the previous paragraph, however, electrical stimulation produced a larger increase in Ser" phosphorylation than in Ser" phosphorylation. Reanalysis of all of the data in which Ser" (CS-1 plus CS-2) and SerR' (CS-3) phosphorylation could be quantified is shown in Table IV. All of the earlier animals, in which electrical stimulation decreased "P-Ser'9/"'P-Ser31, were labeled and stimulated in one of the author's laboratories, whereas all of the later animals, in which electrical stimulation increased the ratio, were labeled and stimulated in the other author's laboratory. Although the stimulation parameters were operationally matched in the two laboratories, it seems likely that the effective stimulation strength may have been different between the experimental setups (see below).

Phosphorylation of Rat
Brain Tyrosine Hydroxylase in Vivo 5653 Phosphorylation of TH in Situ in Rat Striatal Synaptosomes In a previous study (18), incubation of crude striatal synaptosomes with "Pi resulted in a time-dependent incorporation of "P into T H for at least 45 min, and multiple phosphopeptide peaks were separated by reverse-phase HPLC. Synaptosomes, while not an entirely physiological model system for investigating nerve terminal function, were used in the present study in an attempt to distinguish the participation of different signal transduction pathways in the multiple-site phosphorylation observed in uivo.
Effects of Depolarization-One obvious correlate of activation of the medial forebrain bundle would be the depolarization of striatal terminals. Thus, the effects of elevated [K'],, and veratridine were evaluated. In a previous study, elevated [K'], increased 32P incorporation into T H and, for the conditions used (40 mM, 30 s), the increase was restricted to CS-1 and CS-2 (18). In the present studies, a number of conditions were varied in an attempt to increase the relative magnitude of the K+-stimulated increase in T H phosphorylation and, thereby, the possible involvement of phosphorylation sites in addition to Ser". At relatively short treatment durations, the effects of elevated [K'],, and veratridine were both restricted to Ser" phosphorylation ( Table V). The effects of elevated [K'], on Ser" phosphorylation were rapid (<5 s) and concentration-dependent. At longer treatment durations (up to 4 min) an increase in Ser3' phosphorylation was also observed (Table V). With treatment durations approaching that used for electrical stimulation in vivo (20 min), the magnitude of the K+-dependent increases in Ser" phosphorylation became smaller, resulting in greater relative increases in Ser31 uersus Ser" phosphorylation (not presented). While a 10-20-min exposure of synaptosomes to elevated [K+I0 undoubtedly produces a number of effects in addition to depolarization, these data suggest a possible explanation of the laboratory-to-laboratory differences in the relative effects of electrical stimulation on SerIg versus Ser3' phosphorylation presented in Table IV. For example, a weaker effective stimulation strength would bias toward observing a relatively greater

TABLE IV Effects of electrical stimulation on the relative increases i n 32P
incorporation into Ser" versus Ser3' Corpus striata were labeled with 32Pi (0.5-2 mCi/side) as described under "Materials and Methods" prior to unilateral electrical stimulation of the medial forebrain bundle. Limit tryptic peptides from 32P-labeled TH were separated by reverse-phase HPLC, and 32P incorporation was quantified with an on-line radiochemical detector.
Values are ratio of 32P incorporation into SerIg divided by S e P .     (22).4 Furthermore, in perfused rat adrenal glands, stimulation-dependent increases in Ser3' phosphorylation are critically dependent upon the stimulation parameters (23). A number of additional conditions were varied in an unsuccessful attempt to reveal a K-dependent increase in Ser4' phosphorylation as well. These included lower treatment temperature (24 "C), inclusion of bicarbonate in the incubation solution, supplementation of incubation solution with constituents of cerebrospinal fluid (24), and the use of synaptosomes purified on Percoll gradients (not presented).
Effects of CAMP-acting Substances-In that Ser4' phosphorylation can be modulated by CAMP-dependent protein kinase in other catecholaminergic systems (cf. Ref. 25), the possibility that CAMP-regulating systems were either underreactive or already maximally activated in the synaptosomal preparations was tested.
In PC12 cells, adenosine that is released into the medium can act in an autocrine fashion via A? receptors to elevate cAMP levels and increase Ser4' phosphorylation (10,26); and, the effect upon Ser4' phosphorylation can mimicked by 5'-Nethylcarboxamidoadenosine and blocked by theophylline (10,26).4 However, treatment of the synaptosomes with either 5'-N-ethylcarboxamidoadenosine (10 PM, 15 min), 2-chloroadenosine (20 p~, 10 min in the presence or absence of 0.4 unit/ ml adenosine deaminase) or theophylline (100 gM, 5-15 min) did not affect "P incorporation into T H (not presented). Furthermore, inclusion of either adenosine deaminase (0.2-1.0 U/ml) or theophylline (100 g~) during the labeling period J. W. Haycock, unpublished observations. was without effect upon either basal or K-stimulated 32P incorporation into T H (not presented).
Vasoactive intestinal polypeptide and related peptides also appear to act via cAMP in a number of catecholaminergic tissues (cf. Ref. 27). In PC12 cells, superior cervical ganglia, and perfused rat adrenal glands, vasoactive intestinal polypeptide causes a selective increase in Ser4' phosphorylation (9,10,28). In the present studies, however, vasoactive intestinal polypeptide (1 gM, 15 min) was without effect on 32P incorporation into T H a t any of the sites (not presented).
In contrast, the adenylate cyclase/cAMP-dependent protein kinase system per se appeared to be both intact and capable of activation in that treatment of the synaptosomes with either forskolin, 8-bromo-CAMP, or dibutyryl cAMP produced a selective increase in Ser4' phosphorylation ( Table  V).
Effects of Other Agents on TH Phosphorylation-The gradual increase in Ser31 phosphorylation in response to elevated [K'], suggested that a signal transduction pathway more temporally dampened than the calcium influx/CAM-protein kinase I1 activation was being recruited in parallel. In PC12 cells, the phosphorylation of S e P is increased by treatment with nerve growth factor or phorbol esters (10). As shown in Table V, phorbol 12,la-dibutyrate selectively increased Ser" phosphorylation in the striatal synaptosomes. Nerve growth factor (50 ng/ml, 5-15 min) failed to produce an effect (not presented). Finally, that the phosphorylation of all four phosphorylation sites could be simultaneously increased was demonstrated by treatment of the synaptosomes with okadaic acid, an inhibitor of phosphatases 1 and 2A (Table V).

DISCUSSION
T H is phosphorylated in vitro at different sites by a number of different protein kinases (cf. Ref. 25). To the extent that the phosphorylation is associated with an increase in T H activity, each of these protein kinases has, in turn, been suggested to mediate the physiological regulation of T H (cf. Ref. 25). However, an essential criterion for establishing the physiological relevance of a protein kinase's action is the demonstration that phosphorylation of the site(s) influenced by that protein kinase in vitro occurs and is regulated in vivo. Using a rapid cryofixation method, we have demonstrated that activation of the nigrostriatal pathway increases the phosphorylation of striatal T H a t serines 19, 31, and 40.

Potential Protein KinaselSecond Messenger Systems
Serlg Phosphorylation-CAM-protein kinase I1 is the only protein kinase yet known to phosphorylate Serlg (4), and it is present and activated in nerve terminals by depolarization (29,30). Thus, a physiological role for activation of CAMprotein kinase I1 consequent to depolarization-dependent calcium influx T H seems likely. In perfused rat adrenal gland, the second messenger appears to be specifically that calcium derived from influx (as opposed to calcium mobilized from internal stores). Nicotine, which increases calcium influx without mobilizing intracellular calcium, selectively increases Ser" phosphorylation, whereas vasoactive intestinal polypeptide, which mobilizes intracellular calcium without increasing calcium influx, selectively increases Ser4' phosphorylation (23,31). In that both nicotine and vasoactive intestinal polypeptide increase catecholamine secretion via their abilities to increase intracellular calcium (31,32), the CAM-protein kinase 11-TH interaction appears to be compartmentally dis-Phosphorylation of Rat Brain Tyrosine Hydroxylase in Vivo 5655 tinct from the secretory processes. Ser4' Phosphorylation-In contrast to SerIg, Ser4' can be phosphorylated in vitro by a number of different protein kinases including CAM-protein kinase 11, CAMP-dependent protein kinase, cGMP-dependent protein kinase, protein kinase C, protein kinase N, and S6 kinase (4,33).5 Although at least several of these protein kinases are present in nerve terminals and activated by depolarization, the promiscuity of Ser4' as a phosphate acceptor makes it difficult to discern which of the possible protein kinase candidates &(are) actually responsible for the stimulation-dependent phosphorylation of Ser4'. However, the involvement of CAM-protein kinase I1 or protein kinase C in Ser4' phosphorylation seems unlikely. For example, in synaptosomes, essentially maximal increases in Ser" phosphorylation produced by elevated [K'l0 (presumably via CAM-protein kinase 11) failed to influence Ser4' phosphorylation (Table V). Similarly, treatment of synaptosomes for 5 min with a relatively high concentration (1 p~) of phorbol dibutyrate increased Ser31 but not Ser4' phosphorylation (Table V). Although longer treatment (15 min) did produce a small increase in Ser4' phosphorylation, 4-aphorbol dibutyrate did likewise (not presented), suggesting that some mechanism other than activation of protein kinase C was responsible for the Ser4' phosphorylation.
Ser31 Phosphorylation-Phorbol dibutyrate (but not 4-aphorbol dibutyrate) increased Ser31 phosphorylation; however, protein kinase C does not appear to be directly responsible for the phosphorylation of Ser3'. I n vitro, protein kinase C phosphorylates Ser4' (4,34). In fact, none of the more well characterized protein kinases (4) nor a number of more recently described protein kinases' appears to phosphorylate Ser3' in vitro. Pharmacological evidence to date suggests that increases in Ser31 phosphorylation are associated with increases in inositol phospholipid t~r n o v e r ,~ a n d efforts are being made to isolate and purify the protein kinase activity responsible for Ser31 phosphorylation.
A bimodal distribution was observed in the in vivo studies with respect to the relative increases in Ser" uersus Ser31 phosphorylation produced by electrical stimulation (Table  IV). Experiments performed in one laboratory resulted in a relatively larger increase in Ser3' phosphorylation whereas those in the other resulted in a relatively larger increase in SerIg phosphorylation. Two observations suggest that the magnitude of Ser3' phosphorylation is more directly influenced by the quantitative aspects of the stimulus than that of Serlg. For example, from Table V, the temporal course of Ser31 phosphorylation in response to elevated [K'], was slower than that of Ser". Second, in perfused rat adrenal, the magnitude of increase in Ser31 phosphorylation in response to activation of the splanchnic nerve is more dependent upon the stimulation parameters than is the increase in SerIg phosphorylation (28). Thus, it is possible that the stimulation conditions, although operationally matched between laboratories, may have produced different levels of activation of the medial forebrain bundle.
SerS Phosphorylation-Vulliet and colleagues (35) have recently isolated the protein kinase activity from PC12 cells initially shown to phosphorylate Ser'in vitro and have coined the term "proline-directed protein kinase" to reflect an apparent requirement for the sequence -Xaa-Ser/Thr-Pro-Xaato confer substrate reactivity. The levels of this protein kinase activity are very low in brain and adrenal but high in PC12 cells (36). Consistent with this observation, the major difference between the pattern of multiple-site phosphorylation of T H in PC12 cells and corpus striatum is the relative phosphorylation of Ser'. In PC12 cells, -30% of the 32P incorporation into T H was on Sera (lo), whereas in corpus striatum Ser' phosphorylation accounted for less than 10% of the total (Fig. 2, A and B ) . Although it is not known whether the proline-directed protein kinase is responsible for Ser' phosphorylation in dopaminergic nerve terminals, the physiological relevance would be moot given that Sera phosphorylation was not influenced either by electrical stimulation in vivo or by depolarization of synaptosomes. There does, however, appear to be a reasonable rate of phosphate turnover on Sera in that treatment of synaptosomes with okadaic acid increased 32P incorporation into Sera severalfold.
Relationship of Phosphorylation to Activity The stimulation parameters used in the present study were chosen on the basis of their ability to increase T H activity and DA biosynthesis (14, 15), and the activation of T H was confirmed in one rat (not presented).
Although it is not possible from the present data to determine the degree to which phosphorylation of each of the three sites would contribute to an increase in T H activity/DA biosynthesis, it is possible to adduce support for a contribution by phosphorylation at each of the different sites. (a) There is consensus that phosphorylation of Ser*' increases T H activity, and a mechanism for the activation has recently been proposed (37).

( b ) Phosphorylation of Ser" can increase T H activity in the
presence of an activator protein (38,39), although an involvement of the activator remains to be demonstrated in situ. (c) Activation of T H can be associated with what appears to be a selective increase in Ser31 phosphorylation in intact PC12 cells (40). However, given the data in Table V and  "'i-protein G wasfrom Amenham. Percoll was hom Pharmacia. Pansorbin was from Caibiochem. Aniniwpurmed sheep antibodies to catawcally &eTH holoenzyme from rat pheochromocytoma (41) were used for immunaiabeling. Affinity-pumed rabbit antibodies to SDS-denatured TH Subunits (42) were used tor immunoprecipitation (lo). Other chemicals and reagents were as Specified previously (10). IS OF "p LAW.
containinp 1% SDS, was bfough to 4CCLWl pl such that tho find mmpasiti.m WBS NOnidet P40 (51, Nonidet P40:SDS. w:w) in isotonic pH 7.4 buffer mntaining NaF. EGTA. and EDTA (10). Pansorbin (Sa-1W pl; pretreated as described (lo) was added, and the samples were 'precleared' as before (lo). The Western blots (Fig. 1, right) and v e r W rwtinely as described above. lrnmuncehemlcal delermination Of Conditions for quamilatb (>w%) immunopreciphation Wd recwery d TH w e established using TH levels was as described previously (lo) with ttw folbwinp exceptions. immunoMds w e devebped ml, 1 h). The pallets containing immunoprecipitated TH were NOT heated so that the interaction of immunoglobulin heavy chains with "%protein G di d r o t m u r in the M. 55,wO regw 01 the gsi. either from gel slices. as previously described (lo) or frm nilrocellubse after electrophoretic transfer from h; and, limit lryptic potealysis was comp!ete within 4 h. Other detsiis were as describd prwiously (10).