Biochemical characterization of rat brain protein kinase C isozymes.

7 downloads 192 Views 829KB Size Report
Biochemical characteristics of three rat brain pro- tein kinase C isozymes, types I, 11, and 111, were com- pared with respect to their protein kinase and phorbol.
Vol. 263, No. 29, Issue of October 15, pp. 14839-14845,1988 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Biochemical Characterization of Rat Brain ProteinKinase C Isozymes* (Received for publication, February 18,1988)

Kuo-Ping Huang, FreesiaL. Huang, Hiroki Nakabayashi, andYasuyoshi Yoshida From the Section onMetabolic Regulation, Endocrinology and Reproduction Research Branch, National Institute of Child Health and Human Deuelopment, National Institutes of Health, Bethesda, Maryland 20892

Protein kinase C, a Ca2+-activated and phospholipid-deBiochemical characteristics of three rat brain protein kinase C isozymes, types I, 11, and 111, were com- pendent serine/threonine proteinkinase, has been implicated pared withrespect to their protein kinase and phorbol in the regulation of many cellular functions (1, 2). A t least ester-binding activities. All three isozymes appeared three protein kinase C isozymes, designated types I, 11, and to be alike in theirphorbol ester-binding activitiesas 111, have been isolated from rat (3-5), monkey (6), and rabbit evidenced by their similar K d for phorbol 12,13-dibu- brains (7). Theseenzymes all have the same molecular weight tyrate and requirements for Ca” and phospholipids. of 82,000 and exhibit protein kinase activity dependent on However, differences with respect to the effector-me- Ca2+,phospholipid, and diacylglycerol or phorbol ester (3-7). diated stimulation of protein kinase activity were de- However, they are distinguishable by their sites of autophostectableamongthese isozymes. Thetype I enzyme phorylation and immunoreactivities toward the specific anticould be stimulated by cardiolipin toa greater extent bodies against each of these isozymes (3, 6). Recently, it has than those of the type I1 and I11 enzymes. In thepres- been shown that the type I, 11, and I11 protein kinase C ence of cardiolipin, the concentrationsof dioleoylglyisozymes are products of y, 0, and CY genes of this kinase cero1 or phorbol 12,13-dibutyrate required for halfmaximal activation (Alh)of the type I enzyme were family, respectively (4, 8). Analysis of the deduced amino acid sequences of the cloned nearly an orderof magnitude lower than those for the cDNAs coding for the three protein kinase C isozymes retype I1 and I11 enzymes. In the presence of phosphatidylserine, differences in the AHof dioleoylglycerol and vealed an overall sequence homology in two major domains phorbol 12,13-dibutyrate for the three isozymes of which confer the phospholipid/diacylglycerol or phorbol ester protein kinase C were less significant thanthose meas- binding and the kinase activity (9-16). These two domains ured in the presence of cardiolipin. Nevertheless, the have been separated chromatographically after limited proAH of these two activators for the typeI enzyme were teolysis (17-20) and identified as phospholipid-dependent and lower than those for the type I1 and I11 enzymes. At Ca2+-independentphorbol ester-bindingproteinandCa2+/ high levels of phosphatidylserine (greater than15 mol phospholipid-independent protein kinase. All these protein %), binding of phorbol 12,13-dibutyrate to the type I kinase C isozymes exhibit four conserved (Cl-C4) and five enzyme evoked a corresponding stimulationof the ki- variable (Vl-V5) sequence regions in their primary structure. nase activity, whereas bindingof this phorbol ester to The two major domains, each consists of two conserved rethe typeI1 and I11 enzymes produced a lesser degree of gions, are separated by one large variable region (V3) where kinase stimulation. For all three isozymes, the concen- the protease-sensitive sites are located. Overall, the type I trations of phosphatidylserine required for half-max- protein kinase C, encoded by y gene, displays the highest imum [3H]phorbol 12,13-dibutyrate binding were al- degree of divergence within five variable regions. Comparison most an order of magnitude less than those for kinase of the amino acid sequences of protein kinase C family reveals stimulation. Consequently, neither isozyme exhibited a significant kinase activity at lower levels of phos- extensive conservation within each type of the isozyme from phatidylserine (less than5 mol %) and phorbol 12,13- different sources (10). Thesefindings suggest that thevariable dibutyrate (50 nM), a condition sufficient to promote regions of different protein kinase C isozymes may play an near maximal phorbol ester binding. In addition to important role in defining distinct functional specificities of their different responses to the various activators, the these enzymes. Previously, we have characterized these enzymes by imthree protein kinase C isozymes also have differentK , values for protein substrates. The type I enzyme ap- munological methods and found that each protein kinase C peared to have lower K , values for histone IIIS, myelin isozyme was distinctively present in different brain regions, basic protein, poly(lysine, serine) (3:l) polymer, and cell types, and subcellular localizations, and they were differprotamine than those for the type I1 and I11 enzymes. entially expressed during development’ (6, 8). These results These resultsdocumented that the three protein kinase indicate that each protein kinase C isozymemay regulate C isozymes were distinguishable in their biochemical specific cellular responses. The present study was undertaken properties. In particular, the type I enzyme, which isa to determine the biochemical characteristics of the threetypes brain-specific isozyme, is distinct from the typeI1 and of rat brainprotein kinase C. These enzymes have been 111 enzymes, bothhaveawidespreaddistribution separated by hydroxylapatite column chromatography (3) and among differenttissues. identified byisozyme-specific antibodies (6). By using the phospholipid/detergent mixed micellesassay method (21), we observed that the kinase activities of these isozymeswere distinguishable with respect to their phospholipid require* The costs of publication of this article were defrayed in part by ment, sensitivities to activation by diacylglycerol and phorbol the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Yoshida, Y., Huang, F. L., Nakabayashi, H., and Huang, K.-P. (1988) J . Biol. Chem. 263,9868-9873.

14839

14840 Protein

Brain

Rat

ester, andK,,, of different protein substrates.Overall, the type I enzyme is strikingly different from the type I1 and 111 enzymes, both of which are quitesimilar, albeit having different cellular localizations in rat brain (8). EXPERIMENTALPROCEDURES

Materials-The following materials were obtained from the indicated sources: histone IIIS, poly(lysine, serine) (3:l) (PLS),2 protamine-free base, EGTA, and poly(arginine, serine) (3:l) (PAS) from Sigma; [-y-32P]ATP,and [3H]phorbol 12,ls-dibutyrate (PDBu) from Du Pont-New England Nuclear; phosphatidylserine (PS), lysophosphatidylserine (lyso-PS),phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidic acid (PA), cardiolipin (CL), and dioleoylglycerol (DG) from Avanti Polar Lipids, Birmingham, AL; phorbol, phorbol 13-acetate, phorbol 12-myristate, phorbol 12-myristate 13-acetate, phorbol 12, 13-dibutyrate, and 4aphorbol 12, 13-dibutyrate from LC Services Corp., Woburn, Mass,; Nonidet P-40 from Pharmacia LKB Biotechnology Inc.; GF/C glass fiber filters from Whatman; and rabbit brain myelin basic protein from Calbiochem. Methods-Protein kinase C activity was measured in 25 pl of Trisc1 buffer (30 mM), pH 7.5, containing 6 mM magnesium acetate, 0.12 mM [Y-~'P]ATP, 0.25 mM EGTA, 0.4 mM CaC12, 0.04% Nonidet P40, 100 pg/ml PS, 20 pg/ml DG, 1 mg/ml histone IIIS, and protein kinase. The mol % of 100 pg/ml PS (0.12 mM) and 20 pg/ml DG (0.03 mM) in 0.04% Nonidet P-40 (0.66 mM)is approximately 15 and 4, respectively. The Ca2+-and PS-independent activity was measured under the same condition without Ca'+ and phospholipid, but containing 2 mM EGTA. Measurements of 32Pincorporation into protein was performed as described previously (22). One unit of protein kinase activity is defined as the amount of enzyme catalyzing the incorporation of 1 nmol of phosphate from ATP into protein substrate per min at 30 "C under the assay condition. Phospholipids in chloroform were evaporated to dryness under Nz and resuspended in 20 mM TrisC1 buffer, pH 7.5, with sonication and vortexing. DG and phorbol esters were dissolved in ethanol. The lipid/detergent mixed micelles were prepared by adding Nonidet P-40 to glass test tubes containing DG or phorbol ester previously evaporated off ethanol and followed by the addition of phospholipid. The concentration of Nonidet P-40 (0.04%)used in the assay was chosen empirically so that DG or PDBu could support a maximal stimulation of the kinase activity in the presence of 100 pg/ml PS. The M, of the Nonidet P-40 micelles determined by molecular sieve chromatography on Bio-Si1 TSK-400 column equilibrated with 20 mM Tris-C1 buffer, pH 7.5, containing 6 mM magnesium acetate and 0.02% Nonidet P-40 was 81,000to 85,000. The estimated aggregation number of Nonidet P-40 per micelle was 140, a value comparable to those of the Triton X-100 micelles. The M, of 0.04% Nonidet P-40 and 100 pg/ml PS mixed micelles in the same buffer was approximately 105,000 and that of protein kinase C and Nonidet P-40/PS/DG mixed micelles was approximately 200,000. It was estimated that one protein kinase C monomer bound to one Nonidet P-IO/PS/DG mixed micelle.One mol % DG in theseNonidet P-4O/PS mixed micelles wasassumed to contain 7.8 pM. Binding of [3H]PDBu was measured as previously described (19). The reaction mixture (0.2 ml) containing 30 mM Tris-C1 buffer, pH 7.5, 6 mM magnesium acetate, 0.25 mM EDTA, 0.4 mMCaC12, 0.04% Nonidet P-40, 100 pg/ml PS, 0.5 mg/ml bovine serum albumin, 50 nM [3H]PDBu, and protein kinase C. The reaction mixture was incubated at room temperature for 30 min and followed by incubation at 4 "C for 30 min after adding 0.5 ml of 30% DEAE-cellulose in 20 mM Tris-C1 buffer, pH 7.5. Bound [3H]PDBuwas separated from free ligand by filtering through Whatman GF/C glass fiber filter and washed five times each with 1 ml of ice-cold 20 mM Tris-C1 buffer, pH 7.5. The nonspecific binding was determined under the same condition with the addition of 100 pM nonradioactive PDBu. Rat brain protein kinase C isozymes were purified to near homogeneity by the previously described procedure (3). The purity of each isozyme preparation was monitored by the isozyme-specific antibod-

Kinase C Isozymes ies (6) to avoid cross-contamination. The purified enzymes were diluted in 20 mM Tris-C1 buffer, pH 7.5, containing 1mM dithiothreitol, 0.5 mM EDTA, 0.5 mM EGTA, 10% glycerol, and 0.5 mg/ml bovine serum albumin before assay. Proteinconcentrations were determined by the dye-binding method (23) using bovine plasma albumin as standard. Calculations of enzyme kinetics and ligand binding data were carried out by a computer program designed to make statistical analysis of a group of dose-response curves (24). The concentration of free Ca2+was estimated by using established computer program (25). RESULTS

Requirement of Calcium for Different Protein Kinase C Isozymes-Members of the protein kinase C gene family exhibit sequence homology in the PS/phorbol ester-binding domain and kinase catalytic domain. However, there is no consensus "E-F" hand (26) Ca2+-bindingsequence for all three protein kinase C isozymes (9, 10). Previously, we demonstrated that both the phosphorylation of histone 111s and binding of [3H]PDBu by all these enzymes were dependent on Ca2+using the sonicated lipid vesicles assay method. Since Ca2+is known to cause fusion of phospholipid vesicles containing phosphatidylserine, we employed the detergent/phospholipid mixed micelles assay method (21) to evaluate the requirement of Ca2+for these enzymes (Fig. 1). Under this assay condition, the kinase activities for all these isozymes were negligible at all Ca2' concentrations with PS alone. In the presence of 15 mol % PS and 4 mol % DG, the kinase activities for all three isozymeswere stimulated by Ca2+, Maximal stimulation was observed at approximately 10 PM free Ca2+.The A , of Ca2+for all three enzymes was approximately 0.2-0.4 PM. Similar results were also obtained for the Ca2+requirement ( K d = 0.3-0.6 I . ~ M )for [ 3 H ] P D B ~ binding (data notshown). These resultsindicate that all three protein kinase C isozymes have similar affinities for Ca2+ in the presence of PS and DG or PDBu. Specificity of the Various Phospholipids in the Stimulation of Protein Kinase C Isozymes-The phospholipid/detergent mixed micelles assay method provides an ideal condition for measuring the kinase activity under a defined lipid environment. In the presence of 4 mol % DG or 1.6 PM PDBu, all three kinases were maximally stimulated by PS or CL (Fig. 2). Other phospholipids, such as PA, PG, lyso-PS, and PI were less effective, and PC and PE were ineffective. The three protein kinase C isozymeswere similar in their orders of

Log I&+ I

FIG. 1. Ca2+requirement for protein kinase C activity. ProThe abbreviations used are: PLS, poly(lysine, serine) (3:l);PAS, tein kinase activity was measured under the standard assay condition poly(arginine, serine) (3:l); EGTA [ethylenebis(oxyethylenenitrilo)] containing 15 mol % P S (100 pg/ml) alone (open symbols) or 15 tetraacetic acid; PDBu, phorbol 12,13-dibutyrate; PS, phosphatidyl- mol % P S and 4 mol % DG (100 pg/ml PS and 20 pg/ml DG) (filled serine; lyso-PS, lysophosphatidylserine; PC, phosphatidylcholine; symbols) and increasing concentrations of Ca". The enzyme activities PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PA, phos- in the presence of 2 mM EDTA without PS were subtracted from the I1 , (A-A, phatidic acid PI,phosphatidylinositol; CL, cardiolipin; DG, dioleoyl- measured activities for type I ( o " 0 , U) A-A), and 111 (M W , ) protein kinase C. glycerol.

C Isozymes

Kinase Protein Brain Rat

"0

3 - 6

14841

9

12

150

CL CONCENTRATION (mol % I n

PS

LvsoPS

PC

PI

PE

CL

PA

3

6

9

12

15

PS CONCENTRATION (mol % I

FIG.3. Effects of CL and PS on protein kinase C activity. Protein kinase activities of type I (W, u), I1 (A-A, A-A), and I11 (M M , ) isozymes were measured with increasing concentrations of CL (panel A ) or P S (panel B ) in the absence (open symbols)or presence (filled symbols) of 1.6 pM PDBu.

PG

FIG.2. Effect of different phospholipids on protein kinase C activity. Protein kinase activities for type I, 11, and I11 isozymes (panels I , ZI, and I l l ) were measured under standard assay conditions wlth 15 mol % phospholipid (open bar), 15 mol % phospholipid and 1.6 PM PDBu (hatched bar),and 15mol % phospholipid and 4 mol % DG (solid bar). The basal activity with 2 mM EGTA was subtracted from all measurements. The kinase activity in the presence of P S and DG was taken as 100%. TABLEI Requirement of phospholipids for [3H]PDBu binding Binding of [3H]PDBu was carried out under standard assay conditionscontaining 0.15 fig each of the purified protein kinase C isozyme, 15 mol % of each phospholipid, and 50 nM PDBu. Binding was expressed as percent of the maximal value usinn PS. Percent of maximum binding Phospholipid

PS Lyso-PS PI PC PE CL PA PG

Twe I

T w e I1

T w e 111

100 10.5 2.5

100 12.9 0.6

0

0

2.4 6.6 70 23

3.3 7.4 61 17.8

100 11.9 2.5 0 1.2 7.6 80 17.6

preference for these phospholipids. With theexception of CL, a similar order of preference for these phospholipids was also observed for t3H]PDBu binding with each of the three isozymes (Table I). In the presence of 15 mol % CL, binding of t3H]PDBu was less than 10% of that with PS for all three enzymes. A similar phenomenon was also observed by Hannunand Bell (27) using aprotein kinase Cpreparation containing all these isozymes. The kinase activities of these enzymes could be distinguished when assayed in the presence of CL without DG or PDBu. In the presence of 15 mol % CL alone, the type I kinase activity was approximately 60% of that with CL and DG or PDBu(Fig. 2). In comparison, the activities of the type I1 and I11 protein kinase C isozymes were less than 10% of their respective activities with CL and DG or PDBu. These enzymes responded differently to increasing CL concentrations without DG or PDBu (Fig. 3A). Significant stimulation by CL was observed only for the type I but notfor the type I1 or I11 enzymes. In the presence of 1.6 PM PDBu, the concentrations of CL required for half-maximal kinase activity, Alh, ranged from 2.7 f 0.32, 5.3 f 0.8, and 4.9 f 0.8 mol % for the type I, 11, and I11 enzymes, respectively. In comparison, the

Y

No DG Addition

PB P-12"

P-18A PMA aPDBu PDBu

FIG.4. Effects of different phorbol esters on protein kinase C activity. Protein kinase activities of type I, 11,and I11 isozymes (panels Z, ZI, and ZZZ) were measured in the presence of 15 mol % PS (open bar) or CL (solid bar) with 1.6 p M phorbol ( P B ) ,phorbol 12myristate (P-12-M),phorbol 13-acetate (P-13-A),phorbol 12-myristate 13-acetate ( P M A ) ,a-PDBu, andPDBu. The kinase activity with 15 mol % PS and4 mol % DG was taken as 100%. The basal activity with 2 mM EGTA without lipid was subtracted from each measurement. kinase activity with PS without activator was negligible for all three isozymes. The A, of PS in the presence of 1.6 p~ PDBu for the type I, 11, and I11 protein kinase C isozymes were 4.9 f 0.25, 5.3 k 0.34, and 3.5 k 0.24 mol %, respectively (Fig. 3B). These results indicate that the type I enzyme can be distinguished from the type I1 and I11 enzymes when assayed with CL without other activator. Specificity of Phorbol Esters in the Stimulation of Protein Kinase C Isozymes-Protein kinase C activity was determined by the mixed micelles of Nonidet P-4O/PS or CL (15 mol %) in the presence of a 1.6 pM concentration of the various phorbol ester analogs (Fig. 4). In the presence of PS, both PMA and PDBu activated all three isozymes to a level comparable to that by DG. Phorbol, phorbol 13-acetate, and aPDBu were ineffective, and phorbol12-myristatewas approximately 20% as effective as PMA. In the presence of CL, similar results were obtained for the type I1 and I11 enzymes;

Protein BrainRat

14842

Kinase C Isozymes

however, stimulation of the typeI enzyme bythe active tumorpromoting phorbol esters wasless significant because of high basal activity. These results indicate that all three protein kinase C isozymes appear to have similar specificitiesfor phorbol esters. Stimulation of ProteinKinase C Isozymes by DG and PDBu-Even though all three protein kinase C isozymeswere stimulated by saturated concentrations of DG (4 mol %) or PDBu (1.6 ~ L Mto ) a similar extent, the A, of these activators for these isozymes was different. The An of PDBu in the presence of 15 mol % PS for the type I, 11, and I11 protein kinase C isozymes was 14.3 k 1.1, 43.7 f 4.2, and 52.4 f 3.8 nM, respectively (Fig. 5). TheAu of DG for the typeI, 11, and 111 protein kinase Cisozymes was 0.24 f 0.01,0.59 f 0.08,

-2 E

I

I

I

90-

I

I

1

,

I

I

I

Type I

Type 10.4

* 0.16 -1

%

Type 1 I 8.3 i 0.6 mol % Tvpe 111 3.2 f 0.5 mol % I

0

I

4.8 2.4

I

7.2

I

I

9.6 14.4 12

I

I

I

I

16.8 19.2 21.6

DIOLEOYLGLYCEROL (mol % )

FIG.8. Stimulation of Protein kinase c isozymes by DG in the presence of CL. Protein kinase activities were measured under standard assay conditions with 15 mol % CL and increasing concentrations of DG using types I (W), I1 (A-A), and I11 (U protein ) kinase C.

PDBu CONCENTRATION InM)

FIG.5. Stimulation of protein kinase C isozymes by PDBu in the presence of PS. Protein kinase activities were measured under standard assay conditions with 15 mol % PS and increasing concentrations of PDBu using type I (U I1 (A-A), ), and 111 (c".) protein kinase C.

t

fu

a

In

K d = 17 f 2.4nM

PDBu CONCENTRATION (nMI

i

FIG.9. Correlation between PDBu binding and kinase activity for protein kinase C isozymes. Binding of [3H]PDBu (open symbols) and measurement of protein kinase C activity (filled symbols) were carried out under standard assay conditions containing 15 mol % PS. Binding of PDBu in the presence of 100 nM [3H]PDBu and the kinase activity in the presence of 800 nM PDBu was taken as 100%.

E

W

Y

45

a

zY z c

Type 10.24 i 0.01 m o l %

30

and 0.41 f 0.05 mol %, respectively (Fig. 6). When the kinase activity was measured in the presence of Nonidet P-4O/CL 15 mixed micelles, the AI,*of PDBu was13.6 k 4.8,228 k 21, and a 96 f 14 nM (Fig. 7), and thatof DG was 0.4 f 0.16, 8.3 f 0.6, and 3.2 f 0.5 mol % (Fig. 8 ) for types I, 11, and I11 protein DIOLEOYLGLYCEROL (mol % I A , value kinase C, respectively. These data indicate that the FIG.6. Stimulation of protein kinase C isozymes by DG in of PDBu or DG for the type I enzyme was not significantly the presence of PS. Protein kinase activities were measured under altered when assayed with either PS or CL. In comparison, standard assay conditions with 15 mol % PS and increasing concenthe AH of PDBu or DG for the type I1 and I11 enzymes was trations of DG using types I (c".), I1 (A-A), and 111 significantly higher in the presence of CL than PS. Hence, (U protein ) kinase C. the A, of DG or PDBu for the type I enzyme differs most significantly with the type I1 and I11 enzymes in the presence of Nonidet P-4O/CL mixed micelles. Overall, the type I enzyme appears to be most sensitive to activation by DG or PDBu. Binding of pH]PDBu by ThreeProtein Kinase C Isozymes-Protein kinase C binds [3H]PDBu in the presenceof PS and Ca2+. Under the standard assay condition containing 15 mol % PS, all three isozymes had similar Kd values for Type II 228 * 21 nM PDBu (Fig. 9). Forthetype I enzyme, there was a close correspondence between binding of [3H]PDBu and the PDBumediated activation of protein kinase activity (Fig. 9A). In PDBu CONCENTRATION InMI contrast, binding of PDBu to the type I1 and I11 enzymes did not evoke a corresponding activation of the kinase activities FIG.7. Stimulation of protein kinase C isozymes by PDBu (Fig.9, B and c),indicating that binding of additional P D B ~ in the presence of CL. Protein kinase activities were measured molecules is required for maximal kinase activation. Binding under standard assay conditions with 15 mol % cL and increasing concentrations of P D B using ~ typesI (u), 11 (A-A), and 111 of PDBu and the resulting stimulation of the type I protein kinase C was evident only a t saturated concentrations of PS (c".) protein kinase C. Type 1 I 0.59 ? 0.08 mol % Type 111 0.41

~~

~~

~~

?

0.05 mol %

C Isozymes

Kinase Protein BrainRat

14843

(greater than 15 mol %). At a low concentration of PDBu (50 Among the four protein substrates tested, both type I and 11 nM) used in the binding assay, binding of this phorbol ester enzymes have highest VmB.with myelin basis protein,whereas to the kinase reached a near maximal level of 6 mol % Ps, the Vmalwith histone HIS and myelin basic protein were however, without evoking a significant activation of any of almost equivalent for the type I11 enzyme. The V- of PLS these kinases (Fig. 10).Under this assay condition (containing and protamine for all three kinases ranged between 20 and 6 mol % PS and 50 nM PDBu), autophosphorylation of any 40% of those with histone 111s. In general, the Km values of of these isozymes was almost negligible after 20 min of incu- various protein substrates for the type I enzyme were less bation in the presence of [y3'P]ATP (data not shown). The than those for type I1 and I11 enzymes, indicating that the concentrations of PS required for half-maximum t3H]PDBu type I enzyme has thehighest affinity for all these substrates. binding were 0.72 f 0.08, 0.35 -+ 0.02, and 0.51 f 0.01 mol % In spite of the observed differences in the K,,, and vmax of for the type I, 11, and I11 protein kinase C isozymes, respec- different protein substrates for all three isozymes, we could tively. These values were at least an order of magnitude lower not detect any difference in the sites of phosphorylation in than those required for kinase activation. histone H1 and myelin basic protein by two-dimensional Comparison of the Protein SubstrateSpecificities of Protein peptide mapping analysis (data not shown). Kinnse C Isozymes-All threeprotein kinase C isozymes DISCUSSION phosphorylated histone 111sand myelin basic protein dependent on CaZ+,PS, and DG. Phosphorylations of PLS and PAS The three rat brain protein kinase C isozymes purified by were independent of all the effectors. None of these kinases hydroxylapatite column chromatography have been shown to phosphorylated casein. The kinetic parameters of various be the products of three distinct genes (4, 8). Based on the protein substrates for the three isozymes are listed in Table deduced amino acid sequences of the cloned cDNA, it has 11. Phosphorylations of histone IIIS, myelin basic protein, been suggested that homology in thekinase catalytic domain and protamine were measured in the presence of Ca", PS, as well as in the PS/phorbol ester-binding domain exists for and DG, and phosphorylation of PLS was in the absence of all three proteinkinase C isozymes. Indeed, all theseenzymes these effectors. The K , values of histone TIIS for the type I bind [3H]PDBu dependent on Ca2+and PS and phosphorylate and I1 enzymes were approximately one-half of that for the histone IIIS andmyelin basic protein in the presence of Ca2+, type I11 enzyme. The K,,, of myelin basic protein for the type PS, and DG or PDBu. In addition to the conserved regions, I1 enzyme was almost twice as much as those for the type I which confer the kinase and phorbol ester-binding activities, and I11 enzymes, and the K,,, of PLS for the type I enzyme at least five variable sequence regions (Vl-V5) have been was one-half of those for the type I1 and I11 enzymes. The K , identified for all theseprotein kinase C isozymes. It was values of protamine for all three isozymes were comparable. suggested that these variable regions may confer the unique characteristics of these enzymes. In thisstudy, we have carried out detailed analysis to characterize the kinase catalytic function and phorbol ester-binding property of these enzymes. Our results indicate that all three enzymes bind [3H]PDBu 40 with equivalent Kd and have similar requirements of Ca2+ and phospholipids. These enzymes are, therefore, not readily dis30 tinguishable by their binding of [3H]PDBu. These findings suggest that the binding of phorbol ester to the high affinity 20 binding sites is not affected by the surrounding variable 10 sequence regions of each isozyme. The expression of protein kinase C activity in response to '0 3 6 9 12 15 the various effectors is morecomplex than phorbol ester PS CONCENTRATION (mol O/o) binding. The complexity arises in part from the possible FIG. 10. Effect of PS concentrations of the binding of [3H] interactions among the various components, such as kinase, PDBu and the kinase activity. Binding of [3H]PDBu (o"--o, protein substrate, and effectors (28-30). The detergent/phosA-A, G " U ) and measurement of protein kinase activity pholipid mixed micelles(21) used in the present study at least (M A-A, , U) were carried out under the standard assay provide a uniform lipid environment for the kinase reaction. conditions containing 50 nM I3H]PDBu (for binding) or nonradioactive PDBu (for kinase activity measurement) and increasing concen- Although the concentration of Nonidet P-40 (0.04%) used was close to the critical micelle concentration, the physical U) I1 , (A-A, trations of PS using types I ( C b - 0 , A-A), and 111 ( e " 0 , c".) protein kinase C . states of Nonidet P-40 micelles and Nonidet P-4O/PS mixed TABLEI1 Kinetic parameters of protein kinase C isozymes with different protein substrates Protein kinase C activity was measured under standard assay conditions containing 30 mM Tris-C1 buffer, pH 7.5, 6 mM magnesium acetate, 0.12 mM [T-~'P]ATP,0.25 mM EGTA, 0.4 mM CaC12,0.04% Nonidet P-40, 15 mol % PS, 4 mol % DG, 0.01-1.0 mg/ml histone IIIS, myelin basic protein, or protamine, and 18 ng, 12 ng, and 8 ng of type I, 11, and 111isozymes, respectively. Protein kinase activity with PLS assubstrate was measured under the same condition containing 2 mM EGTA without CaC12, PS, and DG. The kinetic parameters were expressed as mean f S.E. MBP, myelin basic protein. Type

Protein substrate

I Vmax

unitsjmg

Histone IIIS

MBP PLS Protamine

2,301 k 126 4,229 f 149 65024.4 f 22 464 + 22

Type I1 K, d m l 139.9 f 20

Vm*= unitsjmg 5,238 f 330

75.1 f 6.6 2.2 31.6 f 2.4

9,908 f 880 2,321 +- 75 1,931 f 54.5

+

Type 111

K, _______ r*g/ ml

176 f 27 167 5 31 62.5 -+ 8.5 48.3 f 2.5

VmSX

unitslmg

10,852 f 1,247 10,239 f 329 2,720 f 235 1.919 f 51

K, dm1 346 2 42

58.9 k 5.2 61.1 -+ 9.2 38.5 k 2.2

14844

Kinase Protein Brain Rat

micelles determined by molecular sieve chromatography are similar to their Triton X-100 counterparts (21, 27, 31). We estimated that themole ratio of mixed micelles to the kinase would be 500-1000 under the assay conditions. These conditions seem to provide a good sensitivity to distinguish the three protein kinase C isozymes. Our results, in general confirm those of Hannun et al. (21, 27, 31),who assayed protein kinase C activity at a higher concentration of detergent (0.3% Triton X-100) using preparations containing a mixture of isozymes. Under the present assay conditions, the Alh of Ca2+ for all three kinases are identical, indicating that the Ca2+binding sites for all three kinases are also highly conserved. These binding sites, however, have not yet been clearly identified from their deduced amino acid sequences. The A , of Ca2+ for the activation of protein kinase C activity with histone 111sas substrateis comparable to that for the binding of phorbol ester for all three isozymes, indicating that Ca2+ interacts directly with the kinase rather than with the substrate. Phosphorylations of the synthetic polymers PLS and PAS by all three kinases were independent of Ca2+.These protein substrates likely interact with all three isozymes in a similar fashion such that theCa2+-requirementis not evident. In spiteof the similarities among the various protein kinase C isozymes with respect to their Kd for PDBu binding and Au for Ca", subtle differences are detectable among them. In particular, the type 1enzyme exhibits ahigher kinase activity with CL in the absence of activator than those of the type I1 and I11 enzymes. In thepresence of both CL and DG or PDBu, all three kinases display near maximal activity. Hence, the kinase activity ratio of CL/(CL + DG or PDBu) is highest (0.6-0.8) for the type I enzyme and can be used as an activity marker for this enzyme. Furthermore, in the presence of CL, the A , ofDG or PDBu for the type I enzyme is almost an order of magnitude lower than those corresponding values for the type I1 and I11 enzymes. In the presence of PS, the same kinetic parameters for the typeI enzyme are also lower, although not as significant, than those for the type I1 and I11 enzymes. In general, the type I enzyme appears to be more readily activated by either DG or PDBu. This conclusion is most obvious from the results shown in Fig. 9, where we demonstrated that thebinding of PDBu to thetype I enzyme in the presence of saturated levels of PS results in a corresponding activation of the kinase. This phenomenon was, however, not observed either for the type I1 or I11 enzymes. Maximal stimulation of these latter two enzymes required more than a stoichiometric amount of PDBu, implying that additional PDBu molecules must interact with the kinases. Under the binding assay condition containing 50 nM PDBu, the concentrations of PS required for maximal PDBu binding for all three kinases are much less than those for the kinase activation (Fig. 10). We have observed a similar phenomenon when the assays were carried out with the sonicated PS vesicles using the assay method of Uchida and Filburn (32). It should be noted that, in this study, for the assay of PDBu binding, we have included DEAE-cellulose to facilitate the separation of bound from free ligand. The effect of DEAEcellulose in changing the equilibrium of binding has not been taken into consideration. Since the addition of DEAE-cellulose to the assay mixture could only dilute the concentration of PS, the estimated PS requirement for PDBu binding can be considered as an upper limit. At 6 mol % PS, binding of PDBu reaches near maximal without evoking a significant stimulation of the kinase activity with histone 111s as substrate. Thisfinding confirms the previous results of Bazzi and Nelsestuen (28) who demonstrated that a complex formation between protein kinase C and lipid vesicles containing PS/

C Isozymes PC/PE/PDBu/Ca2+ was not sufficient to generate maximal kinase activity. In addition, we have also observed that in the presence of 50 nM PDBu, 6 mol % PS, and [-p32P]ATP, autophosphorylation of protein kinase C was negligible; extensive autophosphorylation of all three isozymes became evident only after supplementation of PS up to 15 mol %. These results further support the contention that thebinding of PDBu at a low concentration of PS is not sufficient to evoke a maximal kinase activity. Since 50 nM PDBu or PMA is sufficient to promote most cellular responses and protein phosphorylation in the cultured cells, we predict that thelocal concentration of PS may be greater than thepredicted values of 8-10 mol % in the plasma membrane. It is also likely that the phorbol ester orDG-mediated activation of protein kinase C may involve the clustering of PS on the plasma membrane upon translocation of the kinase. Throughout this study, we have used histone 111s asa protein substrate toanalyze the kinetic parameters for different protein kinase C isozymes. This protein has been shown to be an excellent substrate for determining the Ca2+/PS/DGdependent kinase activity. Phosphorylation of histone 111sby all three protein kinase C isozymes displayed substrate saturation kinetics. Similar kinetic behaviors were also observed with myelin basic protein, PLS, and protamine,even though these protein substrates interact differently with phospholipids (29). In general, the K, values of these protein substrates for the type I enzyme were lower than those for the type I1 and I11 enzymes regardless of the fact that the assays were carried out in the presence or absence of Ca2+and lipids. These findings indicate that these kinases have different affinity for different protein substrates. While this study was in progress, several groups reported the isolation of multiple forms of protein kinase C isozymes by hydroxylapatite column chromatography (4,5, 7). These enzymes were found to respond differently to the activation by free fatty acids (32) and to have different kinetic parameters with respect to PS, Ca2+ (7), DG, PDBu, and protein substrates ( 5 ) .Jaken and Kiley (7) demonstrated that, when assayed with sonicated lipid vesicles, type I enzyme was relatively less Ca2+-dependentand appeared to have a lower K, for PS in the absence of DG or PDBu than the type I1 and I11 enzymes. Pelosin et al. ( 5 ) ,using the PC andPS mixed micelles assay method, found that the type I enzyme was stimulated to a lesser extent by DG or PMA and had a lower K, for protamine than thetype I1 and I11 enzymes. Although the assay conditions employed bythese groups were different from our study, all the results generally point to the same conclusion that the various forms of protein kinase C are distinguishable by their catalytic properties. Therefore, these enzymes may respond differently to a variety of stimuli at different cellular and subcellular locales to produce specific responses. REFERENCES 1. Nishizuka, Y. (1986) Science 233,305-312

2. Berridge, M. J. (1987) Annu. Reu. Biochem. 56,159-193 3. Huang, K.-P., Nakabayashi, H., and Huang, F. L. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,8535-8539 4. Kikkawa, U., Ono, Y., Ogita, K., Fujii, T., Asaoka, Y.,Sekiguichi, K., Kosaka, Y., Igarashi, K., and Nishizuka, Y. (1987) FEBS Lett. 217,227-231 5. Pelosin, J.-M., Vilgrain, I., and Chambaz, E. M. (1987) Biochem. Biophys. Res. Commun. 31,382-391 6. Huang, F. L.,Yoshida, Y . , Nakabayashi, H., and Huang, K.-P. (1987) J. Biol. Chem. 262,15714-15720 7. Jaken, S., and Kiley, S. C. (1987) Proe. Natl. A d . Sci. U. S. A. 84,4418-4422 8. Huang, F. L., Yoshida, Y., Nakabayashi, H., Knopf, J. L., Young,

Protein Brain Rat

Kinase C Isozymes

S. W. 111, and Huang, K.-P. (1987) Biochem. Biophys. Res. Commun. 149,946-952 9. Parker, P. J., Coussens, L., Totty, N., Rhee, L., Young, S., Chen, E., Stabel, S., and Ullrich, A. (1986) Science 233,853-859 10. Coussens, L., Parker, P. J., Rhee, L., Yang-Feng, T. L., Chen, E., Waterfield, M. D., Francke, U., and Ullrich, A. (1986) Science 233,859-866 11. Ono, Y., Kurokawa, T., Kawahara, U., Nishimura, O., Marumoto,

R., Igarashi, K., Sugino, y., Kikkawa, u., Ogita, K., and Nishizuka, Y. (1986) FEBS Lett. 203,111-115 12. Ono, Y., Kurokawa, T., Fujii, T., Kawahara, K., Kikkawa, U., Ogita, K., Sugino, y., and Nishizuka, y. (1986) FEBS Lett. 206,347-352 13. Knopf, J. L., Lee, M.-H., Sultzman, L. A., Kriz, R. W., Loomis, c. R., Hewick, R. M', and R' M' (1986) 469 491-502 14. Housey, G . M., O'Brian, C. A., Johnson, M. D., Kirschmeier, P., and Weinstein, I. B. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 1065-1069 28. 126 15. Ohno, S., Kawasaki, H., Imajoh, T., Suzuki, K., Inagaki, M., Yokokura, H., Sakoh, T,, and Hidaka, H. (1987) Nature 3 2 5 , 161-166 16. Carpenter, D., Jackson, T., and Hanley, M. R. (1987) Nature 325,107-108 17. Kishimoto,A., Kajikawa, N., Shiota, M., and Nishizuka, Y. (1983) J. Biol. Chem. 258, 1156-1164 18. Huang, K.-P., and Huang, F. L. (1986) J. Biol. Chem. 261, 14781-14786

14845

19. Huang, K.-P., and Huang, F.L. (1986) Biochem. Biophys. Res. Commun. 139, 320-326 20. Lee, M.-H., and Bell, R. M. (1986) J. Bwl. Chem. 261, 1486714870 21. Hannun, Y., Loomis, C. R., and Bell, R. M. (1985) J. Biol. Chem. 260,10039-10043 22. Huang, K.-P., and Robinson, J. C. (1976) Anal. Biochem. 72, 593-599 23. Bradford, M. (1976) ~ n a lBiochem. . 72, 248-254 24. DeLean, A., Munson, P. J., and Rodbard, J. (1978) Am. J. Physwl. 236, E97-E102 25. Perrin, D. D., and Sayee, 1. G . (1967) Talanta 14,833-842 26. Manalan, A. S., and Klee, C. B. (1986) Adu. Cyclic. Nucleotide Protein Phosphorylation Res. 18, 227-278 27. Hannun, y. A., and Bell, R. M. (1986) J. BioL C h m . 261,93419347 Bazzi, D., M. and Nelsestuen, G . L.(1987) Biochemistry 26,11529. Bazzi, M. D.9 and Nelsestuen, G . L.(1987) Biochemistry 26, 1974-1982 30. Bazzi, M.D.9 and Nelsestuen, G . L- (1987) BWChem. BbPhYS. Res. Commun. 147,248-253 31. Hannun, Y. A., Loomis, C. R., and Bell, R.M. (1986) J. Biol. Chem. 261,7184-7190 32. Uchida, T., and Filburn, C.R. (1984) J. Biol. Chem. 269,1231112314