Journal ofNeurochemistryRaven Press, Ltd., New York© 1994 International Society for Neurochemistry

Activation of Protein Kinase C by Purified Bovine Brain14-3-3 : Comparison with Tyrosine Hydroxylase Activation

M. Tanji, R. Horwitz, *G. Rosenfeld, and J . C. Waymire

Department ofNeurobiology and Anatomy and *Department of Pharmacology,University of Texas Medical School, Houston, Texas, U.S.A .

Abstract : In the course of the purification of 14-3-3 pro-tein (14-3-3) we found that 14-3-3 isolated from bovineforebrain activates protein kinase C (PKC), rather thanthe previously reported protein kinase C inhibitory activity(KCIP) . We have characterized the 14-3-3 activation ofPKC. The physical properties of purified PKC activatorare the same as those previously reported for 14-3-3 andKCIP ; i .e ., (1) it is composed of subunits of molecularweight 32,000, 30,000, and 29,000 ; (2) it is homogeneouswith respect to molecular weight, as judged by nativegradient-gel electrophoresis, with a molecular weight of53,000 ; and (3) it is composed of at least six isoformswhen analyzed by reverse-phase HPLC . The concentra-tion dependence of PKC activation by 14-3-3 is in thesame range as that shown previously for KCIP inhibitionof PKC, and as that required for 14-3-3 activation of tyro-sine hydroxylase ; a maximal stimulation of two- to three-fold occurs at 40-100 lug/ml . 14-3-3's activation of PKCis sensitive to a-chymotrypsin digestion but is not heatlabile . Activation is specific to PKC; at least two otherprotein kinases, cyclic AMP- and calcium/calmodulin-dependent protein kinases, are not activated . The activa-tion of PKC by 14-3-3 is independent of phosphatidylser-ine and calcium and, as such, is an alternative mechanismfor the activation of PKC that obviates its translocationto membranes . Key Words: Protein kinase C-14-3-3protein-Bovine forebrain-Inhibition-Tyrosine hy-droxylase-Kinase C inhibitory protein .J. Neurochem. 63, 1908-1916 (1994) .

14-3-3 is an acidic protein that is highly enrichedin brain tissue but also exists in a wide variety oftissues and organisms including keratinocytes (Lefferset al., 1993), Xenopus (Martens et al ., 1992), Dro-sophila melanogaster (Swanson and Ganguly, 1992),plants (Hirsch et al ., 1992; Wu et al ., 1992), and yeast(van Heusden et al ., 1992) . It is estimated to be atleast 1% of brain protein (Boston et al ., 1982) . In1987 another acidic brain protein, termed "activatorprotein," was shown to be highly related to 14-3-3(Ichimura et al ., 1987) . The activator protein was orig-inally described by Yamauchi et al . (1981) as an activa-tor of tyrosine hydroxylase and tryptophan hydroxy-

1908

lase, the two rate-limiting enzymes in catecholamineand serotonin synthesis, respectively . These hydroxy-lases require the activator protein for increased activityinduced by their phosphorylation by calcium/calmodu-lin-dependent protein kinase 11 . The similarities be-tween 14-3-3 and the activator protein include subunitcomposition, molecular weight, amino acid content,immunological cross-reactivity, and most important,the capacity to activate these two hydroxylases (Ichi-mura et al ., 1987) . Further analysis of 14-3-3 hasshown that it is not a single protein, but instead afamily of seven to nine isoforms (Ichimura et al .,1988) . Amino acid sequencing of cyanogen bromidecleaved peptides of the isoforms and cloning of threeisoforms has shown each to have unique amino acidsequences but that each is highly homologous to theothers (Ichimura et al ., 1988) . Thus far, three novelbrain mRNAs for 14-3-3s have been cloned and se-quenced. The amino acid sequence of each isoformindicates the proteins have very acid C-terminal andneutral N-terminal regions . Acidic C-termini are hy-pothesized to activate phosphorylated hydroxylases byinteracting with the phosphorylated regions of theseenzymes to relieve constrained enzymatic activity(Ichimura et al ., 1988) .

Within the last 3 years three additional biologicalactivities of 14-3-3 have been noted . Exol, a proteinreported by Morgan and Burgoyne (1992a,b) to benecessary for catecholamine exocytosis in permeabil-ized bovine adrenal chromaffin cells, is homologousto 14-3-3 . Consistent with this observation, Wu et al .(1992) have shown that anti-14-3-3 inhibits catechola-

Received December 13, 1993 ; revised manuscript received Febru-ary 1, 1994; accepted March 21, 1994.

Address correspondence and reprint requests to Dr. J. C. Waymireat Department of Neurobiology and Anatomy, University of TexasMedical School, 6431 Fannin Street, Houston, TX 77225, U.S .A.

Abbreviations used: CaM kinase II, Ca21 /calmodulin-dependentprotein kinase II ; CAMP, cyclic AMP; KCIP, kinase C inhibitorprotein ; 6-MPH,, DL-6-methyl-5,6,7,8-tetrahydropterin ; PKC, pro-tein kinase C; SDS-PAGE, sodium dodecyl sulfate- polyacrylamidegel electrophoresis; TBS, Tris-buffered saline ; TCA, trichloroaceticacid .

14-3-3 PROTEIN ACTIVATES TYROSINE HYDROXYLASE AND PKC

mine secretion stimulated by adding back chromaffincell cytosolic proteins . Zupan et al . (1992) have re-ported the cloning and isolation of a protein with Ca2+ -activated phospholipase A activity from brain, whichhas a marked sequence similarity to 14-3-3 . Toker etal . (1992) have reported the isolation of a protein witha high degree of homology to 14-3-3s that is a novelinhibitor of protein kinase C (PKC) (Aitken et al .,1990 ; Toker et al ., 1992) and have termed this activity"PKC inhibitor protein" (KCIP) .

In the course of purifying 14-3-3 by the procedureof Toker et al . (1992) for use in studies of tyrosinehydroxylase regulation, we discovered another biologi-cal activity of this family of proteins . Instead of thePKC inhibitory activity reported previously joker etal ., 1992), our preparation activated PKC. Here wedescribe the purification of 14-3-3, the properties ofthe activation of PKC by 14-3-3, and compare theactivation of PKC with 14-3-3's activation of tyrosinehydroxylase. A preliminary report of this work hasappeared (Horwitz et al ., 1992) . While this work wasin progress, Isobe et al . (1992) reported that a rena-tured isoform of 14-3-3 also activates PKC.

MATERIALS AND METHODS

[y-32P]ATP (4,500 Ci/mmol) and '21I-protein A (30 Ci/g) were purchased from ICNBiomedicals, Inc. (Irvine, CA,U.S.A .) . L-[3,5-3H]Tyrosine (56 Ci/mmol) was from Am-ersham Co . (Arlington Heights, IL, U.S.A .) . Nitrocellulosemembrane (0.45 ym) was from Schleicher & Schuell, Inc.(Keene, NH, U.S.A .) . DEAE-Sephacel, phenyl-Sepharose4B, MonoQHR5/5, Superose-6 and 12, and molecular masscalibration kits were from Pharmacia LKB Biotechnology(Piscataway, NJ, U.S.A .) . Benzamidine hydrochloride hy-drate was from Aldrich (Milwaukee, WI, U.S .A .) and leu-peptin was from Chemicon (Temecula, CA, U.S .A .) . DL-6-Methyl-5,6,7,8-tetrahydropterin (6-MPH 4 ) was from Calbio-chem Co. (La Jolla, CA, U.S.A.) . L-Phosphatidylserine wasfrom Avanti Polar-Lipids, Inc. (Alabaster, AL, U.S.A.) . a-Chymotrypsin was purchased from AMRESCO (Solon, OH,U.S.A .) . Cellulose phosphate paper (P81) was from What-man (Maidstone, England). Other materials (analytical re-agent grade) were from Sigma (St. Louis, MO, U.S.A .) orBio-Rad (Richmond, CA, U.S .A .) .A synthetic peptide (Met-His-Arg-Gln-Glu-Thr-Val-Asp)

used to assay Cat+ /calmodulin-dependent protein kinase II(CaM kinase II) was generously provided by Dr . P. T. Kelly(Department of Neurobiology and Anatomy, University ofTexas Medical School, Houston, TX, U.S.A .) . Calmodulinand synapsin I were purified from bovine forebrain as de-scribed by Gopalakrishna and Anderson (1982) and Bahlerand Greengard (1987), respectively . The catalytic subunit ofcyclic AMP (cAMP)-dependent protein kinase was purifiedfrom bovine heart as described (Okuno and Fujisawa, 1990) .PKC was purified from bovine or rat forebrain by a modifi-cation of the procedure of Walsh et al . (1984) . DEAE-Sephacel anion-exchange chromatography and Superose-6HR10/30 size-exclusion column chromatography steps wereinserted before phenyl-Sepharose hydrophobic interactionchromatography and the last step used MonoQcolumn chro-matography instead of DEAE-Sephacel chromatography .

1909

PKC was not separated into separate isozymes . All datapresented are from bovine forebrain PKC unless otherwisestated . CaM kinase II was prepared from rat or bovine fore-brain as described by McGuinness et al. (1983) . Tyrosinehydroxylase was purified from bovine adrenal medullae bya procedure to be published elsewhere (M . Tanji and J. C.Waymire) . The antibody to 14-3-3 was generated by Dr .Gary Rosenfeld. Its specificity was confirmed by using theantibody to screen a rat brain cDNA expression library . Puri-fication of the antibody-positive plaques and a 1 .9-kb cDNAinsert yielded the deduced amino acid sequence of 14-3-3~(Aitken et al ., 1990) .

Immunoreactive assay of 14-3-3One microliter of sample was spotted on nitrocellulose

membrane . After spotting, membranes were washed withTris-buffered saline (TBS) buffer (8 mM Tris-Cl, pH 7 .5,containing 0.15 M NaCl), dried on a filter paper, and sub-merged in TBST buffer (TBS buffer containing 0.05%Tween 20). The membrane was blocked with 1 % bovineserum albumin in TBST buffer for 30 min and incubatedwith a 1 :2,000 dilution ofprimary rabbit polyclonal antibodyagainst 14-3-3 in TBST buffer . After washing the membranewith four changes of TBST buffer for 20 min, the membranewas incubated with a 1 :7,500 dilution of second antibodyand 121I-protein A (0.2 yCi/ml) in TBST for 30 min. Themembrane was washed with five changes of TBST bufferfor 25 min and dried. The radioactivity of each spot wascounted in an LKB Model 1270 y-counter. All procedureswere performed at room temperature. The observed radioac-tivity was proportional to the amount of spotted protein,between 0 and 130 leg of protein per spot.

Purification of 14-3-3Bovine forebrain was obtained fresh from the local slaugh-

terhouse, frozen immediately in liquid nitrogen, and kept at-80°C until used . All procedures were performed at 4°C oron ice if not described otherwise. Frozen brains (50 g) werepulverized on dry ice and -10 g was homogenized in 200ml of buffer A [20 mM Tris-Cl, pH 7.5, containing 2 mMEDTA, 10 mM EGTA, 1 mM dithiothreitol, and 8 .55%sucrose (wt/vol) to which protease inhibitors had beenadded immediately before use (10 yg/ml soybean trypsininhibitor, 75 l-tg/ml phenylmethylsulfonyl fluoride, 10 mMbenzamidine, 10 Fig/ml leupeptin, and 1 ug/ml pepstatin,final concentration)] . Homogenates were combined and cen-trifuged at 35,000 g for 25 min. The supernatant was loadedonto a 2.5 X 27-cm column of DEAE-Sephacel equilibratedin buffer B (20 mM Tris-Cl, pH 7.5, 2 mM EDTA, 5 mMEGTA, and 2 mM dithiothreitol) . After washing the columnwith 400 ml of buffer C (20 mM Tris-Cl, pH 7.5, containing1 mM EDTA, 1 mM EGTA, and 1 mM dithiothreitol),proteins were eluted with a linear NaCl gradient (0-0.5 M)in buffer C at a flow rate of 2 ml/min . Five-milliliter fractionswere collected; two peaks of 14-3-3 immunoreactivity werepooled and combined .NaCl was added to immunoreactive fractions (2 .5 M final

concentration) and they were applied to a 2.5 X 8-cm phe-nyl-Sepharose 4B column (Pharmacia, Inc.) equilibrated in2.5 M NaCl in buffer C. After washing with 120 ml of 2.5MNaCl in buffer C, protein was eluted from the column bya linear gradient of 2.5-0 M NaCl in 100 ml of buffer C,followed by 80 ml of buffer C. The flow rate was 1 ml/minand 3-ml fractions were collected. Immunoreactive fractionswere pooled and dialyzed against buffer C overnight .

J. Neurochem ., Vol . 63, No . S, 1994

191 0

The desalted phenyl-Sepharose immunoreactive pooledfractions were divided into 11 fractions and each chromato-graphed separately on a MonoQHR5/5 column (Pharmacia,Inc.) equilibrated in buffer C. Protein was eluted with agradient of 0-0.6 M NaCl in 10 ml of buffer C, followedby 0.6-1.0 M NaCI in 5 ml of buffer C. The flow ratewas 0.5 ml/min and 0.5-ml fractions were collected. Eachfraction was analyzed by immunoreactive assay, westernblotting, and tyrosine hydroxylase activation assay (see be-low) . Although three immunoreactive peaks were eluted,labeled P1, P2, and P3, only the last to elute, P3, activatedtyrosine hydroxylase and migrated on sodium dodecyl sul-fate-polyacrylamide gel electrophoresis (SDS-PAGE) withthe molecular weight subunits characteristic of 14-3-3 (mo-lecular weight of 29,000, 30,000, and 31,000 detected byeither Coomassie Blue staining or western blotting with theantibody to 14-3-3) . Peak P3 was pooled and dialyzed over-night against a saturated ammonium sulfate solution(pH 7.0).

Precipitated protein was collected by centrifugation andredissolved in 500 ul of buffer C followed by gel filtrationon a Superose-12 HR10/30 column (Pharmacia, Inc.) equili-brated with buffer C. The flow rate was 0.1 ml/min and 0.5-ml fractions were collected. A single major immunoreactiveand protein peak eluted immediately after the void volume .Two-thirds of the purified 14-3-3 was stored at -80°C andthe remainder stored at 4°C.

Determination of molecular massThe molecular mass of purified 14-3-3 was determined by

polyacrylamide-gel electrophoresis and gel-filtration chro-matography . Before polyacrylamide-gel electrophoresis, theprotein was dissolved in 5-10 /.cl of gel buffer (90 mM Tris-borate, pH 8.3, containing 3 mM EDTA) containing 12%glycerol and 0.8 mg/ml bromophenol blue and applied to4-30% polyacrylamide gel. Commercially available highmolecular mass markers (Pharmacia, Inc.) were used as stan-dards. Electrophoresis was performed at 8.8 V/cm for 16 hat room temperature . After electrophoresis, the gel wasstained for protein with Coomassie Brilliant Blue R-250.The molecular mass determination by gel-filtration col-

umn chromatography was performed using a Superose-6HR10/30 column (Pharmacia, Inc.) equilibrated with bufferC containing 0.1 M NaCl . Thyroglobulin, ferritin, catalase,and bovine serum albumin (1-3 Nag) were used as molecularweight standards. The flow rate was 0.05 ml/min .

Reverse-phase HPLCPurified 14-3-3 was analyzed by reverse-phase HPLC ac-

cording to the procedure of Toker et al . (1992) . Forty micro-grams of protein from the Superose-12 HR10/30 columnwas applied to a Brownlee RP 300, 100 X 2.1-mm column(Applied Biosystems, Inc.) equilibrated in 0.1% heptofluor-obutyric acid and eluted with a two-step gradient of H2O/acetonitrile containing 0.1% heptofluorobutyric acid as anion-suppressing agent. Acetonitrile was increased to 45%over 10 min, then further increased to 50% over 50 min ata flow rate of 0.2 ml/min . 14-3-3 isoforms eluted during the45-50% acetonitrile gradient . Protein was monitored at A215-The 14-3-3 isoform peaks were collected and analyzed bySDS-PAGE .

Assay of tyrosine hydroxylase and PKC activationActivation of tyrosine hydroxylase by 14-3-3 was exam-

ined by a modification of the method of Yamauchi et al .

J. Neurochem., Vol . 63, No . 5, 1994

M. TANJI ET AL.

(1981) . 14-3-3 was incubated for 5 min at 30°C in 50 mMHEPES/OH, pH 7.0, containing 10 MM M902, 0.4 mMCaC1 2 , 0.1 mM ATP, 2 pM calmodulin, 10 nM Ca2+/cal-modulin-dependent protein kinase 11, and 10 nM tyrosinehydroxylase in a volume of 30 pl . Tyrosine hydroxylaseactivity was measured by a modification (Waymire et al .,1991) of the procedure of Reinhard et al . (1986) and initi-ated by adding 70 p1 of a solution to bring the final concen-tration to 150 mM Tris-maleate, pH 6.4, 50 pM L-tyrosine,5 mM ascorbate, 0.45 mg/ml catalase, and 0.5 mM 6-MPH,After 20 min at 30°C the reaction was stopped on ice and 1ml of 0.1 M hydrochloric acid containing 7.5% charcoal wasadded to each assay. After centrifugation at 3,000 g for 2min, 400 pl of supernatant was taken for determination ofradioactivity. Background activity was performed in thesame way except that tyrosine hydroxylase was omitted fromthe assay. The tyrosine hydroxylase assay was linear overreaction time for at least 30 min.PKC activity was examined by incubating 20 units (trans-

fer of 20 pmol inorganic PO4 from ATP to histone Sigmatype HIS per minute) of purified PKC in the presence andabsence of 14-3-3 at 30°C for 5 min in 30 pl of assay mixturecontaining 20 mM HEPES/OH, pH 7.5, 10 MM MgC12, 0.5mM CaC12 , 5 yg/ml diolein, 0.5 mg/ml histone (type HIS),50 IZg/ml phosphatidylserine, and 50 pM [y- 32P]ATP (500pCi/pmol). Diolein and phosphatidylserine were sonicatedtogether before addition . The reaction was started by theaddition of PKC and stopped on ice, a 25-Ml aliquot spottedonto cellulose phosphate paper (2 X 2 cm), and the paperwas washed seven times with 75 mM acetic acid . The boundradioactivity was measured by assessing Cherencov radia-tion by liquid scintillation spectroscopy . Alternatively, thereaction mixture was treated with s volume of 50% trichloro-acetic acid (TCA) and centrifuged at 3,000 g for 10 min.The precipitated proteins were separated on a 5-20% gradi-ent SDS-PAGE. The gel was dried, submitted to autoradiog-raphy, and the radioactive bands excised for quantitation byassessing Cherencov radiation .Assay of CaM kinase II and cAMP-dependentprotein kinaseThe activity of CaM kinase It was measured by a modifi-

cation of the method of Kennedy et al . (1983) . Sampleswere incubated 2 min at 30°C in 35 MI of 50 mM Tris-Cl,pH 7.5, containing 10 MM MgC12 , 50 AM [y- 32p]ATP(500cpm/pmol), 1.7 pM calmodulin, 0.4 mM EGTA, 5 mM 2-mercaptoethanol, 1 .4 MM CaC12, and 15 .8 /cM substratepeptide (Met-His-Arg-Gln-Glu-Thr-Val-Asp) . The reac-tions were stopped on ice and 20-p1 aliquots spotted ontocellulose phosphate P81 paper (2 X 2 cm) . After washingthe paper seven times with 75 mM acetic acid, radioactivitywas assessed as described above.cAMP-dependent protein kinase activity was measured by

a modification method of Witt and Roskoski (1975) . Thekinase was incubated at 30°C, 5 min in 100 y1 of 50 mMMES/Tris, pH 6.3, containing 10 MM MgC12, 0.3 mMEGTA, 10 PM [y-32P]ATP (600 cpm/pmol), 3 pM cAMP,and 75 lug/ml Kemptide (Leu-Arg-Arg-AIa-Ser-Leu-Gly) .Reactions were stopped by spotting on cellulose phosphatepaper P81 (2 X 2 cm) and after rinsing radioactivity wasanalyzed as described above.a-Chymotryptic proteolysis of 14-3-3

14-3-3 (80 /ig) was digested with 2 wg of a-chymotrypsinat 37°C, 2 h, in 100 /sl of 20 mM NH4HCO3 . The reactionwas stopped by heating at 90°C, 10 min.

14-3-3 PROTEIN ACTIVATES TYROSINE HYDROXYLASE AND PKC

RESULTS

Purification of 14-3-314-3-3 was purified using DEAE-Sephacel-, phe-

nyl-Sepharose 4B-, MonoQ-, and Superose-12-col-umn chromatography . The presence of 14-3-3 in elu-tion fractions was monitored using a polyclonal anti-body. Although we confirmed that this antibodystrongly cross-reacts with 14-3-3 (it detected proteinwith the appropriate subunit composition and biologi-cal activity ; see below), it also weakly cross-reactedwith other proteins .

After each chromatography step, immunoreactivefractions were pooled for the next purification step .After MonoQ column chromatography, three immuno-reactive peaks were apparent. The first two peaks, P1and P2, did not activate tyrosine hydroxylase andshowed no evidence of 14-3-3 subunits when analyzedon SDS-PAGE using either western blotting or directstaining of the gels with Coomassie Brilliant Blue (notshown) . In contrast, peak P3 exhibited protein bandsof molecular weight 29,000-32,000 consistent with14-3-3 and activated tyrosine hydroxylase . The thirdmeasure we used to assess the presence of 14-3-3 ineach fraction was KCIP activity . However, instead ofthe predicted inhibition of PKC, no KCIP activity wasobserved and a single peak of PKC activator, whichcorresponded to P3, was found; P3 produced a nearlytwofold activation of the kinase (see below) . Becauseof the potential importance of 14-3-3's activation ofPKC, we further characterized this activity . In addition,we compared 14-3-3's capacity to activate PKC withits activation of tyrosine hydroxylase . For these charac-terization studies, 14-3-3 purified through the addi-tional step of size-exclusion chromatography on Su-perose-12 was used . The final product was purified 34-fold with a yield of 16%. Because nonspecific immuno-reactivity existed in the early, crude fractions, this esti-mate of fold purification is only an approximation .

Properties of purified 14-3-3After purification, 14-3-3 eluted as a single protein

peak when applied either to a Superose-6 column ora second MonoQ column (data not shown) and wasresolved as three bands of molecular weight 32,000,30,000, and 29,000 on SDS-PAGE (Fig . IA) . Theestimated molecular weight of native purified 14-3-3determined by both native gel electrophoresis (Fig .113) and gel-filtration column chromatography on Su-perose-6 (not shown) was 53,000 . That the proteinmigrated as a single band on native gradient-gel elec-trophoresis indicated that 14-3-3 was homogeneous .

Three cycles of freeze/thaw did not effect either themigration pattern on SDS-PAGE or the capacity toactivate either tyrosine hydroxylase or PKC (data notshown) . No difference in these same measures wasobserved after 6 months' storage at either 0 or -80°C .The purified protein was stored at -80°C . The sameproperties of activation of tyrosine hydroxylase and

1911

FIG . 1 . Polyacrylamide gel electrophoresis of purified 14-3-3.Approximately 16 kg of purified 14-3-3 was applied to each gel.Proteins were visualized with Coomassie Brilliant Blue R-250 . A:Electrophoresis was performed with 13.5% gel in the presenceof 0 .1 % SIDS . B: Native gel electrophoresis was performed on4-30% gradient gel (pH 8 .3) .

PKC were obtained whether 14-3-3 was purified fromfrozen or fresh bovine forebrain .As demonstrated previously for both 14-3-3 (Ichi-

mura et al ., 1988) and KCIP Joker et al ., 1992), ourpreparation of bovine 14-3-3 separated into multipleisoforms when analyzed by reverse-phase HPLC (Fig .2, numbered 1-6) . The protein separated into five tosix UV peaks when chromatographed by the proceduredescribed by Toker et al . (1992) . Except for peaks 1,4, and 5, the purified isoform peaks showed singlebands with a relative molecular weight of 29,000 onSDS-PAGE (data not shown) . As reported by Tokeret al . (1992) , peak 1 showed little or no protein stain-ing . As peak 4 contained two protein bands (29,000and 30,000), this peak may contain more than oneisoform ; probably Y2 and ~ . As peak 5 also containedtwo protein bands (29,000 and 32,000), this peak ispredicted to also be composed of more than one iso-form (Ichimura et al ., 1988 ; Toker et al ., 1992) . Be-cause past studies have shown the e isoform to be a32,000 polypeptide, peak 5 likely contains the c iso-form, plus one other isoform, probably ~ . Verificationof peak identity will require amino acid sequence de-termination of the polypeptide in each peak .

Characteristics of 14-3-3 activation of PKCPurified 14-3-3's activation of PKC was compared

with several other proteins' capacities to activate the pro-

J. Neurochem., Vo[. 63, No. 5, 1994

1912

FIG . 2. Reverse-phase HPLC of 14-3-3 . Forty micrograms of14-3-3 from the Superose-12 HR10/30 column step was appliedto a Brownlee RP 300 (100 x 2 .1 mm) column equilibrated in0 .1 % heptofluorobutyric acid and eluted with a two-step gradi-ent of H2O/acetonitrile containing 0 .1 % heptofluorobutyric acidas described in Materials and Methods .

tein kinase. Figure 3 shows that the activation was nota nonspecific protein effect ; no activation of PKC wasproduced by three other proteins when added to the PKCassay at a similar concentration as 14-3-3 . Likewise, PKCwas the only protein kinase of three tested that was acti-vated by 14-3-3 . Bovine brain CaM kinase 11 and bovineheart CAMP-dependent protein kinase catalytic subunitwere not influenced by 14-3-3 (Fig . 4) . Analysis of theinteraction of 14-3-3 with cofactors involved in the acti-vation of PKC showed that the activation by 14-3-3 was

FIG . 3. Comparison of 14-3-3 with several proteins for the ca-pacity to stimulate the activity of PKC . 14-3-3 (" - " ), synap-sin I (" - " ), calmodulin, (0 - V), and bovine serum albumin( " - " ) were added to the assay mixture in the concentrationshown and the activity of PKC measured as described in Materi-als and Methods . 14-3-3 alone (O - O) had no PKC activity .

J. Neurochem., Vol. 63, No. 5, 1994

M. TANJI ET AL.

FIG . 4. Comparison of the effect of 14-3-3 on the activity ofthree protein kinases . Each protein kinase was assayed with(closed bars) and without (open bars) 120 jig/ml 14-3-3 as de-scribed in Materials and Methods .

totally independent of phosphatidylserine (Figs . 5 and6), calcium, or diacylglycerol (Fig . 5) . The activationproduced by 14-3-3 was not due to a nonproteinaceousmolecule that copurifies with 14-3-3 . Digestion of 14-3-3 with a-chymotrypsin reduced its capacity to activatethe PKC (data not shown) . Further, 14-3-3 did not in-crease PKC activity by artifactually increasing the bind-ing of the phosphorylated histone to the phosphocellulosepaper ; a twofold increase in 3 '-PO4-labeled histone wasobserved in 14-3-3-containing assays if SDS-PAGE wasused to quantitate histone phosphorylation directly (datanot shown) .Comparison of the activation of tyrosinehydroxylase and PKCThe existence of a relationship between 14-3-3's

capacity to activate phosphorylated tyrosine hydroxy-

FIG . 5 . Effects of phosphatidylserine (PS, 80 Mg/ml) and cal-cium (Ca2+' 0 .5 mM) on PKC activity without (open bars) orwith (filled bars) 14-3-3 (120 1Lg/ml) . A: PKC assayed in theabsence of diacylglycerol . B : PKC assayed in the presence ofdiacylglycerol (5 leg/ml) .

14-3-3 PROTEIN ACTIVATES TYROSINE HYDROXYLASE AND PKC

FIG. 6 . Effect of phosphatidylserine concentration on the activa-tion of PKC by 14-3-3 . PKC activity was examined as describedin Materials and Methods in the absence (0 - V) and presenceof 120 pg/ml 14-3-3 (" - " ) at indicated concentrations ofphosphatidylserine . All assays contained 0.5 mM Ca21 .

lase and PKC was evaluated by comparing the concen-tration dependence and heat stability of the two pro-cesses . Figure 7 shows that the concentration of 14-3-3required for the activation of the tyrosine hydroxylaseand PKC are within the same concentration range (40-120 ttg/ml) . This optimal concentration is the sameas that reported previously by Yamauchi and Fujisawa(1981) to be optimal for the activation of tyrosinehydroxylase by 14-3-3 and the same concentration of14-3-3 reported to inhibit PKC Joker et al ., 1992) .In contrast to the similar concentration dependenceof the tyrosine hydroxylase and PKC activation, thecapacity of 14-3-3 to activate the two enzymes haddifferent temperature sensitivities . Figure 8 shows thatheating 14-3-3 at 50-100 °C for 10 min progressivelydecreased its capacity to activate tyrosine hydroxylase,with no activation occurring if the protein was incu-bated at 70°C or above. No decrease in PKC activationby 14-3-3 was produced by heating . 14-3-3 retainedits activation potency even after incubation at 100°Cfor 10 min.

DISCUSSION

14-3-3 was isolated from bovine forebrain by a com-bination of conventional and high performance chro-matography Joker et al ., 1992) . The molecularweight of 53,000 in native condition and three subunitsof molecular weights 32,000, 30,000, and 29,000 aresimilar to those reported previously and confirm thatthe protein is a dimer (Grasso et al ., 1977 ; Boston etal ., 1982) . Also consistent with previous analysis of

1913

14-3-3 (Ichimura et al ., 1987), reverse-phase HPLCshowed 14-3-3 to be composed of multiple isoforms .It is not possible, as yet, to say whether each isoformis a homo- or heterodimer . Based on the coelution of14-3-3 immunoreactivity and UV absorption at 280nm and nondenaturing PAGE, the purified 14-3-3 ap-pears to be homogeneous .

In contrast to the report that 14-3-3 inhibits PKCJoker et al ., 1992), we have shown that 14-3-3 puri-fied from bovine brain activates PKC two- to threefold .The activation is not an artifact of our PKC assaysystem (which relies on the binding of phophorylatedhistone to ion-exchange paper) as TCA-precipitatedhistone is likewise more phosphorylated in the pres-ence of 14-3-3 . The observed activation of PKC by14-3-3 agrees with the report by Isobe et al . (1992) that14-3-3 dimers composed of either ~, /3, or E subunits ofbovine brain 14-3-3 (S2, 62 , and 0, isolated by re-verse-phase chromatography of the purified 14-3-3 andthen reconstructed, activate PKC. The amount of acti-vation of PKC reported for these three homodimers isthe same as the activation of PKC by native 14-3-3 reported here . The observation that native 14-3-3activates PKC demonstrates that the activation re-ported by Isobe et al. (1992) is not the product of 14-3-3 denaturation and reassembly of the subunits intohomodimers .We characterized several aspects of the activation

FIG . 7 . Comparison of the concentrations of 14-3-3 required toactivate tyrosine hydroxylase and PKC . A: The effect of 14-3-3on tyrosine hydroxylase activity was measured as described inMaterials and Methods in the absence (" - " ) and the pres-ence (A - A) of prior phosphorylation of tyrosine hydroxylaseby CaM kinase II . B : The effect of 14-3-3 concentration on theactivity of purified PKC measured as described in Materials andMethods (A - A) . 14-3-3 alone had no PKC activity(* - " ) .

J. Neurochem., Vol. 63, No. 5, 1994

1914

FIG. 8. Comparison of the effect of elevated temperature on 14-3-3 activation of tyrosine hydroxylase (bottom) and PKC (top).The 14-3-3 (0.7 mg/ml) or (0.15 mg/ml) was incubated 10 minat the temperatures shown before addition to the activation mix-ture of tyrosine hydroxylase or to the assay mixture of PKC.Enzyme assays were conducted as described in Materials andMethods.

of PKC by 14-3-3 . First, that several other proteinswould not substitute for the 14-3-3 protein to activatePKC indicates that the activation by 14-3-3 is not dueto the nonspecific stabilization effects that proteinsmay have on enzymes . Second, that several other pro-tein kinases were not influenced by 14-3-3 shows theactivation to be specific to PKC. That a-chymotrypsindestroyed 14-3-3's activity indicates that the stimula-tion is due to the protein component of the purified14-3-3 and not a trace contaminant that copurified with14-3-3 . The observation that 14-3-3's activation ofPKC was fully stable to heat treatment at 100°C for10 min contrasts with the heat lability of 14-3-3's acti-vation of tyrosine hydroxylase and suggests that sec-ondary and/or tertiary structure is important for tyro-sine hydroxylase activation by the bovine 14-3-3 butnot for the activation of PKC. This also implies thatthe basis for the interaction of 14-3-3 with tyrosinehydroxylase and PKC is probably different .

Although the mechanism of the activation of PKCby 14-3-3 is not yet known, the observation that 14-3-3 activated PKC in the presence of phosphatidylserine,calcium, or diacylglycerol suggests a mechanism thatcomplements the activation of PKC by these cofactors .Because PKC is predicted to be a very flexible protein(Huang, 1989), a hypothetical mechanism of the acti-vation is that 14-3-3 binds to the PKC to produce

J. Neurochem., Vol. 63, No . 5, 1994

M. TANJI ET AL.

conformational changes that activate the enzyme orenhance activation induced by other cofactors (for ex-ample, phosphatidylserine, or the combination of cal-cium and phosphatidylserine) . An intriguing possibil-ity is that 14-3-3 binding to PKC induces a conforma-tional change in the protein kinase that is similar tothe change in conformation that is associated with PKCtranslocation to the cell membranes . As such, 14-3-3binding to PKC could be a means of activating nonpar-ticulate PKC, obviating translocation .The observed activation of PKC by 14-3-3 in the

present studies, in contrast to the previously reportedinhibition of PKC Joker et al ., 1992), is perplexing .However, it appears that our unpublished results fromrecent studies resolve this discrepancy . We have iso-lated a protein that copurifies with 14-3-3, unless stepsare taken to eliminate it, which causes 14-3-3 to be aninhibitor of PKC . The purification and properties ofthis "switching" protein will be published elsewhere(manuscript in preparation) .

Although the role of 14-3-3 in cells remains un-known, some evidence of the regulation of cellularfunction exists . The best-documented influence of 14-3-3 is its capacity to activate monoamine hydroxylases(Yamauchi et al ., 1981 ; lchimura et al ., 1987) . Be-cause 14-3-3 is required for the activation of tyrosineand tryptophan hydroxylases via phosphorylation byCaM kinase 11, it may be involved in monoamine bio-synthesis regulation . That the CaM kinase II phosphor-ylation site on tyrosine hydroxylase is phosphorylatedin response to stimulation of intact cells (Waymire etal ., 1988) and rat brain synaptosomes (Haycock andHaycock, 1991) strongly supports this conclusion . AsPKC also phosphorylates and activates tyrosine hy-droxylase, but on a different serine residue, 14-3-3 mayhave an additional positive influence on catecholaminesynthesis through enhancement of PKC-mediatedphosphorylation of tyrosine hydroxylase .

14-3-3 appears to be important for the regulation ofcatecholamine secretion as well . Morgan and Bur-goyne (1992a,b) report that a 14-3-3-like protein,Exol, restores catecholamine exocytosis to permeabil-ized bovine chromaffin cells that have lost their secre-tory response because of the apparent loss of criticalcytoplasmic proteins through leakage out of permeabil-ized cells . Exol restores secretion and can act synergis-tically with PKC in doing so . One explanation for theseobservations is that Exol, like 14-3-3, activates PKC.However, in Morgan and Burgoyne's studies, Exoldid not influence PKC activity (Morgan and Burgoyne,1992a) . The observation that 14-3-3 is highly concen-trated in adrenal medulla and autonomic ganglia isconsistent with an important role in these tissues (Isobeet al ., 1989) .Beyond a specific role for 14-3-3 in monoaminergic

cells, that 14-3-3 isoforms are ubiquitous in eukaryotes(Aitken et al ., 1992) implies a very basic function .One possibility is that these proteins are involved inthe expression of second messenger-mediated phos-

14-3-3 PROTEIN ACTIVATES TYROSINE HYDROXYLASE AND PKC

phorylation cascades . The observation that 14-3-3shares sequence homology and subunit size with sev-eral membrane proteins (RACKS) that are reported toact as intracellular receptors for activated PKC

(Mochly-Rosen et al ., 1991a,b) has led to the proposi-tion that the different isoforms of 14-3-3 may be bind-ing proteins for corresponding isoforms of PKC (Lef-fers et al ., 1993) . As such, a possible role of the 14-3-3 isoforms could be to direct the subcellular localiza-tion of isoforms of PKC. In this regard, the inhibitionor activation of PKC by 14-3-3 could be very im-portant.

Whether 14-3-3, itself, is regulated directly by phos-phorylation is unknown. Because 14-3-3 isoforms haveconsensus sequences for phosphorylation by severalprotein kinases, the possibility exists that a modulationof their function in these systems could occur. Studiesso far, however, have shown only little phosphoryla-tion of 14-3-3s (Morgan and Burgoyne, 1992a ; Tokeret al ., 1992 ; M. Tanji and J . C . Waymire, unpublishedobservations) .

In conclusion, we have shown that purified bovinebrain 14-3-3 activates PKC. Characterization of 14-3-3's activation of PKC has demonstrated that (1) only14-3-3, of several proteins tested, is able to activatePKC, (2) other protein kinases are not activated by14-3-3, (3) the activation of PKC is independent ofphosphatidylserine and calcium, and (4) the activationof PKC by 14-3-3 has a different sensitivity to temper-ature than 14-3-3's activation of tyrosine hydroxylase,indicating that the mechanisms of 14-3-3's activationof these two enzymes are probably disparate.

Acknowledgment : This work was supported by UnitedStates Public Health Service grant NS 11061 to J.C.W .

REFERENCESAitken A., Ellis C. A., Harris A., Sellers L. A., and TokerA. (1990)

Kinase and neurotransmitters . Nature 344, 594.Aitken A., Collinge D. B., Van Heusden B. P. H., Isobe T., Rose-

boom P. H., Rosenfeld G., and Soll J . (1992) 14-3-3 proteins :a highly conserved, widespread family of eukaryotic proteins.Trends Biochem. Sci. 17, 498-501 .

Bahler M. and Greengard P. (1987) Synapsin I bundles F-actin ina phosphorylation-dependent manner. Nature 326, 704-707.

Boston P. F., Jackson P., Kynoch P. A. M., and Thompson R. J.(1982) Purification, properties, and immunohistochemical lo-calisation of human brain 14-3-3 protein. J. Neurochem. 38,1466-1474.

Gopalakrishna R. and Anderson W. B. (1982) Ca"-induced hy-drophobic site on calmodulin : application for purification ofcalmodulin by phenyl-Sepharose affinity chromatography . Bio-chem. Biophys. Res. Commun . 104, 830-836.

Grasso A., Roda G., Hogue-Angeletti R. A., MooreB. W., and PerezV. J. (1977) Preparation and properties of the brain specificprotein 14-3-2. Brain Res. 124, 497-507.

Haycock J. W. and Haycock D. A. (1991) Tyrosine hydroxylase inrat brain dopaminergic nerve terminals . Multiple-site phosphor-ylation in vivo and in synaptosomes . J. Biol. Chem . 266, 5650-5657 .

Hirsch S., Aitken A., Bertsch U., and Soll J. (1992) A plant homo-logue to mammalian brain 14-3-3 protein and protein kinase Cinhibitor . FEBS Lett. 296, 222-224.

1915

Horwitz R., Tanji M., and Waymire J. C. (1992) Purified 14-3-3protein from bovine brain activates both tyrosine hydroxylaseand protein kinase C. Soc. Neurosci. Abstr. 18, 284.

Huang K.-P. (1989) The mechanism of protein kinase C activation.Trends Neurosci . 12, 425-432.

Ichimura T., Isobe T., Okuyama T., Yamauchi T., and Fujisawa H.(1987) Brain 14-3-3 protein is an activatorprotein that activatestryptophan 5-monooxygenase and tyrosine 3-monooxygenasein the presence of Ca 21, calmodulin-dependent protein kinaseII . FEBS Lett. 219, 79-82.

Ichimura T., Isobe T., Okuyama T., Takahashi N., Araki K., KuwanoR., and Takahashi Y. (1988) Molecular cloning of cDNA cod-ing for brain-specific 14-3-3 protein, a protein kinase-dependentactivator of tyrosine and tryptophan hydroxylases . Proc . Natl.Acad. Sci. USA 85, 7084-7088.

Isobe T., Ichimura T., and Okuyama T. (1989) Chemistry and cellbiology of neuron- and glia-specific proteins . Dev. Neurosci.14, 238-246.

Isobe T., Hiyane Y., IchimuraT., Okuyama T., Takahashi N., NakajoS., and Nakaya K. (1992) Activation of protein kinase C by the14-3-3 proteins homologous with Exol protein that stimulatescalcium-dependent exocytosis . FEBS Lett. 308, 121-124.

Kennedy M. B., McGuinness T. L., and Greengard P. (1983) Acalcium/ calmodulin-dependent protein kinase activity frommammalian brain that phosphorylates synapsin I : partial purifi-cation and characterization . J. Neurosci . 3, 818-831.

Leffers H., Madsen P., Rasmussen H. H., Honore B ., AndersenA. H., Walbum E., Vanderkerekhove J., and Celis J. E. (1993)Molecular cloning and expression of the transformation sensitive epithelial marker "stratifin" : a member of a protein familythat has been involved in the protein kinase C signaling path-way. J. Mol. Biol. 231, 982-998.

Martens G. J., Piosik P. A., and Danen E. H. (1992) Evolutionaryconservation of 14-3-3 proteins . Biochem. Biophys. Res. Com-mun. 184, 1456-1459.

McGuinness T. L., LaiY., Greengard P., Woodgett J. R., and CohenP. (1983) A multifunctional calmodulin-dependent protein ki-nase : similarities between skeletal muscle glycogen synthasekinase and a brain synapsin I kinase. FEBS Lett. 163, 329-334.

Mochly-Rosen D., Khaner H., and Lopez J. (1991a) Identificationof intracellular receptor proteins for activated protein kinase C.Proc . Natl. Acad. Sci. USA 88, 3997-4000.

Mochly-Rosen D., Khaner H., Lopez J., and Smith B. L. (1991b)Intracellular receptors for activated protein kinase C. J. Biol.Chem . 266, 14866-14868.

Morgan A. and Burgoyne R. D. (1992a) Interaction between proteinkinase C and Exol (14-3-3 protein) and its relevance to exo-cytosis in permeabilized adrenal chromaffin cells . Biochem. J.286, 807-811.

Morgan A. and Burgoyne R. D. (1992b) Exol and Exo2 proteinsstimulate calcium-dependent exocytosis in permeabilized adre-nal chromaffin cells . Nature 355, 833-836.

Okuno S . and Fujisawa H. (1990) Stabilization, purification andcrystallization of catalytic subunit of cAMP-dependent proteinkinase from bovine heart. Biochim. Biophys. Acta 1038, 204-208.

Reinhard J. F. Jr., Smith G. K., and Nichol C. A. (1986) A rapidand simple sensitive assay for tyrosine-3 monooxygenase basedupon the release of 3H20 and adsorption of ['H]-tyrosine bycharcoal . Life Sci. 39, 2185-2189.

Swanson K. D. and Ganguly R. (1992) An activator protein oftyrosine and tryptophan hydroxylase of the mammalian nervoussystem . Gene 113, 183-190.

Toker A., Sellers L. A., Amess B., Patel Y., Harris A., and AitkenA. (1992) Multiple isoforms of a protein kinase C inhibitor(KCIP-1/ 14-3-3) from sheep brain-amino acid sequence ofphosphorylated forms. Eur. J. Biochem. 206, 453-461 .

van Heusden G. P. H., Wenzel T. J., Lagendijk E. L., de SteenamH. Y., and van den Berg J. A. (1992) Characterization of theyeast BMH1 gene encoding a putative protein homologous tomammalian protein kinase II activators and protein kinase Cinhibitors . FEBS Lett, 302, 145-150.

J. Neurochem., Vol . 63, No . 5, 1994

1916

Walsh M . P ., Valentine K . A ., Ngai P . K., Carruthers C . A ., andHollenberg M . D . (1984) Ca"-dependent hydrophobic-interac-tion chromatography : isolation of a novel Cat+ -binding proteinand protein kinase C from bovine brain . Biochem. 1224, 117-127 .

Waymire J . C ., Johnston J. P ., Hummer-Lichteig K., LloydA ., VignyA ., and Craviso G . L . (1988) Phosphorylation ofbovine adrenalchromaffin cell tyrosine hydroxylase . Temporal correlation ofacetylcholine's effect on site phosphorylation, enzyme activa-tion, and catecholamine synthesis. J. Biol . Chem . 263, 12439-12447 .

Waymire J. C ., Craviso G . L ., Lichteig K ., Johnston J . P ., BaldwinC ., and Zigmond R . E . (1991) Vasoactive intestinal peptidestimulates catecholamine biosynthesis in isolated adrenal chromaffin cells : evidence for a cyclic AMP-dependent phosphory-lation and activation of tyrosine hydroxylase . J. Neurochem .57, 1313-1324.

Witt J . J . and Roskoski R . Jr . (1975) Rapid protein kinase assay

J. Neurochem., Vol. 63, No . S, 1994

M. TANK ET AL.

using phosphocellulose-paper absorption . Anal. Biochem. 66,253-258 .

Wu Y . N., Vu N.-D ., and WagnerP . D . (1992) Anti-(14-3-3 protein)antibody inhibits stimulation of noradrenaline (norepinephrine)secretion by chromaffin-cell cytosolic proteins . Biochem. J. 285,697-700.

Yamauchi T . and Fujisawa H . (1981) Tyrosine 3-monooxygenase isphosphorylated by Ca t+-calmodulin-dependent protein kinase,followed by activation by activator protein . Biochem . Biophys .Res. Commun. 100, 807-813 .

Yamauchi T ., Nakata H ., and Fujisawa H . (1981) A new activatorprotein that activates tryptophan 5-monooxygenase and tyrosine3-monooxygenase in the presence of Ca"-, calmodulin-depen-dent protein kinase . J. Biol. Chem. 256, 5404-5409.

Zupan L . A ., Steffens D. L ., Berry C . A ., Landt M., and GrossR . W . (1992) Cloning and expression of a human 14-3-3 proteinmediating phospholipolysis . J. Biol. Chem. 267, 8707-8710 .