Phosphorylation of Protein Kinase Cδ (PKCδ) at Threonine 505 Is Not a Prerequisite for Enzymatic Activity EXPRESSION OF RAT PKCδ AND AN ALANINE 505 MUTANT IN BACTERIA IN A FUNCTIONAL FORM

A structural feature shared by many protein kinases is the requirement for phosphorylation of threonine or tyrosine in the so-called activation loop for full enzyme activity. Previous studies by several groups have indicated that the isotypes α, βI, and βII of protein kinase C (PKC) are synthesized as inactive precursors and require phosphorylation by a putative “PKC kinase” for permissive activation. Expression of PKCα in bacteria resulted in a nonfunctional enzyme, apparently due to lack of this kinase. The phosphorylation sites for the PKC kinase in the activation loop of PKCα and PKCβII could be identified as Thr497 and Thr500, respectively. We report here that PKCδ, contrary to PKCα, can be expressed in bacteria in a functional form. The activity of the recombinant enzyme regarding substrate phosphorylation, autophosphorylation, and dependence on activation by 12-O-tetradecanoylphorbol-13-acetate as well as the Km values for two substrates are comparable to those of recombinant PKCδ expressed in baculovirus-infected insect cells. By site-directed mutagenesis we were able to show that Thr505, corresponding to Thr497 and Thr500 of PKCα and PKCβII, respectively, is not essential for obtaining a catalytically competent conformation of PKCδ. The mutant Ala505 can be activated and does not differ from the wild type regarding activity and several other features. Ser504 can not take over the role of Thr505 and is not prerequisite for the kinase to become activated, as proven by the unaffected enzyme activity of respective mutants (Ala504 and Ala504/Ala505). These results indicate that phosphorylation of Thr505 is not required for the formation of functional PKCδ and that at least this PKC isoenzyme differs from the isotypes α, βI, and βII regarding the permissive activation by a PKC kinase.

PKC␦ is a ubiquitously expressed PKC isoenzyme (12) and exhibits some unique properties. The tyrosine kinase c-Src selectively phosphorylates the type ␦ PKC isoenzyme in vitro. The tyrosine phosphorylation induces a modification of PKC␦ activity exhibiting some substrate selectivity (13). Tyrosine phosphorylation of PKC␦ could be demonstrated also in vivo (14,15). In addition to the substrate specificity acquired by tyrosine phosphorylation, PKC␦ appears to possess an intrinsic substrate specificity, for example, toward the elongation factor 1␣ and an elongation factor 1␣ peptide (16). PKC␦ is autophosphorylated to a much higher degree than the other isoenzymes, and in contrast to other kinases, including PKC isoenzymes, it is able to accept GTP as a phosphate donor for autophosphorylation (17). Based on results obtained with cells overexpressing PKC␦, some role of this PKC isoenzyme in growth suppression and induction of differentiation has been suggested (18 -20). The finding that PKC␦ is lost from immortalized human keratinocytes after stable transfection with a c-Ha-ras oncogene also points to this possible function (21).
Here we report on another, possibly unique, feature of PKC␦. Contrary to PKC␣ (22)(23)(24), PKC␦ could be expressed in bacteria in a functional form. Moreover, we were able to demonstrate by site-directed mutagenesis that phosphorylation by a putative "PKC kinase" of Thr 505 , unlike that of the corresponding Thr 497 and Thr 500 in PKC␣ (24) and ␤ II (25), respectively, is not essential for a permissive activation of PKC␦.
Polymerase Chain Reaction Amplification and Cloning of Wild Type and Mutant PKC␦ cDNA-For the construction of a PKC␦ full-length cDNA with an NdeI restriction site at the initiation signal ATG and an EcoRI restriction site behind the stop codon TGA, the following oligonucleotide primers were used: 5Ј-AAA GGA TCC CAT ATG GCA CCG TTC CTG CGC-3Ј as 5Ј-primer and 5Ј-TCT GGG AAT TCA CTA CTA TTC CAG GAA TTG CTC-3Ј as 3Ј-primer (synthesized by W. Weinig, German Cancer Research Center). For polymerase chain reaction amplification (cycle profile: 94°C/5 min; 10 ϫ 94°C/15 s, 56°C/30 s, 72°C/2 min; 15 ϫ 94°C/15 s, 56°C/30 s, 72°C/2 min, plus cycle elongation of 20 s for each cycle; 72°C/7 min), a rat PKC␦ full-length cDNA clone of 3000 base pairs was used as template. The resulting cDNA of 2048 base pairs was cut with NdeI and EcoRI and cloned into the NdeI-EcoRI-cut expression vector pET28. The resulting plasmid was termed pET28␦wt and used for transformation of BL21(DE3)pLysS cells.
Polymerase chain reaction-mediated site-directed mutagenesis was performed using the "overlap extension" method according to Ho et al. (26). A rat PKC␦ full-length clone of 3000 base pairs served as a template. The following pairs of mutagenic oligonucleotides were used (only the oligonucleotides that are sense with respect to the PKC␦ cDNA are given, and changed bases are underlined): (T/A)505, 5Ј-CGG GCC AGC GCT TTC TGC GGC-3Ј; (S/A)504, 5Ј-GAG AAC CGG GCC GCC ACA TTC TGC GGC ACT-3Ј; and (S/A)504(T/A)505, 5Ј-GAG AAC CGG GCC GCC GCA TTC TGC GGC ACT-3Ј. Successful mutation was confirmed by sequencing. The PKC␦ mutant cDNAs were cloned into pET28 as described above. The resulting plasmids were termed pET28␦Ala 505 , pET28␦Ala 504 , and pET28␦Ala 504 /Ala 505 .
Bacterial Expression of PKC␦-E. coli BL21(DE3)pLysS cells were transformed with the various plasmids indicated above, grown at 37°C for 12 h, diluted 1:100 in 1 liter of fresh LB medium supplemented with 50 g/ml kanamycin and 30 g/ml chloramphenicol, and incubated at 24°C until the absorbance (600 nm) reached 0.5-0.7. Induction was performed with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside for 16 h at 24°C. Cells were sedimented and resuspended with ice-cold buffer I (30 mM MES, pH 6.5, 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol) or for further purification on nickel-nitrilo-triacetic acid resin with buffer II (50 mM sodium phosphate, pH 8.0, 500 mM NaCl, 10 mM imidazole). Resuspension in buffer resulted in an absorbance (600 nm) of approximately 40. After sonication with a Branson sonifier and centrifugation at 80,000 ϫ g for 45 min at 4°C, the supernatant (termed bacterial extract) was used as a source of bacterial recombinant His-tagged PKC␦. The insoluble fraction was resuspended in SDS sample buffer and boiled for 5 min before application to PAGE. Partial Purification of Recombinant His-tagged PKC␦-Partial purification of soluble PKC␦ was achieved by metal chelate affinity chromatography of the bacterial extract under native conditions using nickel-nitrilo-triacetic acid resin and following the manufacturer's recommendation. Bound proteins were eluted with imidazole (100, 150, 200, 250, and 500 mM). PKC␦ was detected in the 100 mM imidazole fraction by immunoblotting and assaying PS-and TPA-stimulated kinase activity.
Recombinant PKC␦ from Baculovirus-infected Insect Cells-Sf9 cells were infected with the recombinant baculovirus, and cells were extracted as described previously (27).
Protein Kinase Assay-Phosphorylation reactions were carried out in a total volume of 100 l containing buffer C (50 mM Tris-HCl, pH 7.5, 10 mM ␤-mercaptoethanol), 4 mM MgCl 2 , 10 g of PS, 100 nM TPA, 2 g of pseudosubstrate-related peptide ␦ or 30 g of histone III-S as substrate, 10 l of bacterial extracts diluted as indicated in figure legends, and 37 M ATP containing 1 Ci [␥-32 P]ATP. In some experiments, PS and TPA were omitted, and in some others, Gö 6976 or Gö 6983 at concentrations indicated in figure legends were added. After incubation for 7 min at 30°C, the reaction was terminated by transferring 50 l of the assay mixture onto a 20-mm square piece of phosphocellulose paper (Whatman p81). After washing the paper three times in deionized water and twice in acetone, radioactivity was determined by liquid scintillation counting. 1 unit of kinase activity equals 1 nmol of phosphate incorporated into substrate/min.

FIG. 1.
A, expression of PKC␦ in E. coli BL21(DE3)pLysS cells. Bacteria were transformed with the plasmid pET28 alone or the plasmid pET28␦wt (wt) containing the full-length cDNA of rat PKC␦ (PKC␦ wild type) and were grown as described under "Experimental Procedures." Bacteria were extracted with 1 ml of buffer I and buffer I-insoluble proteins dissolved in 1 ml of sample buffer containing 1% SDS. Soluble (a) and insoluble (b) proteins (3 l each) were separated by SDS-PAGE. PKC␦ (arrow) was identified by immunoblotting with a PKC␦-specific antibody and an alkaline phosphataseconjugated goat anti-rabbit IgG as secondary antibody. Molecular masses were determined from the standard proteins myosin (205 kDa), ␤-galactosidase (116 kDa), phosphorylase b (97 kDa), bovine serum albumin (66 kDa), and ovalbumin (45 kDa). B, comparison of the concentration of recombinant PKC␦ in extracts from bacteria and baculovirus-infected insect cells. Extracts from bacteria that were transformed with the plasmids pET28␦wt or pET28␦Ala 505 and produced the PKC␦ wild type (wt) or the PKC␦Ala 505 mutant (Ala505) and extracts from baculovirus-infected Sf9 insect cells expressing PKC␦ wild type (Baculo) were diluted as indicated (g protein). On separation of the extracted proteins by SDS-PAGE the amount of PKC␦ was estimated by immunoblotting (see A).
Autophosphorylation and Tyrosine Phosphorylation of PKC␦-Autophosphorylation was carried out essentially as described for the protein kinase assay, but 37 M ATP containing 8 Ci of [␥-32 P]ATP was added and substrate was omitted. The reaction was terminated by addition of 10% trichloroacetic acid. Precipitated proteins were redissolved in sample buffer, separated by SDS-PAGE (7.5%), and visualized by immunoblotting and autoradiography with an exposure time of 16 h. Tyrosine phosphorylation was performed essentially as autophosphorylation, but 37 M ATP (without [ 32 P]ATP), 250 M MnSO 4 , and 6 units c-Src were added, and incubation at 30°C was performed for 15 min.
Immunoblotting-Immunoblotting was performed as described previously (28). As the primary antibody a polyclonal PKC␦-specific antibody (7) or a monoclonal anti-phosphotyrosine antibody and as the secondary antibody alkaline phosphatase-or horseradish peroxidaseconjugated goat anti-rabbit IgG (see figure legends) were used. Immunoblots were stripped by incubation with 100 mM ␤-mercaptoethanol, 2% SDS, and 63 mM Tris-HCl, pH 6.7, at 60°C for 1 h.
Determination of Protein Concentration-Protein concentration was determined with the protein dye reagent concentrate (according to the method of Bradford; Ref. 53) from Bio-Rad, using bovine serum albumin as standard.

RESULTS
At transformation with the expression vector pET28␦wt containing the full-length cDNA of rat PKC␦, although not with the pET28 vector alone, the E. coli cells BL21(DE3)pLysS produced PKC␦, as demonstrated by immunoblotting of bacterially expressed proteins with a PKC␦-specific antibody (Fig. 1A). A portion of the ectopically expressed rat PKC␦ was soluble in a buffer without detergent (herein termed bacterial extract). The insoluble fraction, present in the bacteria probably in the form of inclusion bodies, was dissolved in sample buffer containing 1% SDS. Slow growth of the bacteria at 24°C was expected to result in an increased portion of active recombinant enzyme.
PKC␦ could not be detected on staining of bacterial proteins with Coomassie Blue but only by immunoblotting. The concentration of PKC␦ in the bacterial extract was approximately one-third of that in an extract of insect cells that produced PKC␦ on infection with a recombinant baculovirus (Fig. 1B). The bacterial recombinant PKC␦ was found to be enzymatically active. As shown in Fig. 2, histone III-S was phosphorylated by the recombinant enzyme in the bacterial extract in a TPA-dependent manner. Only weak incorporation of phosphate occurred in the absence of PS and TPA or in the presence of PS alone. An extract of bacteria transformed with the pET28 vector alone served as a control and did not show any TPAinducible kinase activity. To be able to compare the kinase activity of recombinant PKC␦ from bacteria with that from baculovirus-infected insect cells, nearly equal amounts of PKC␦ had to be applied to the kinase assay. This was achieved by diluting the insect cell extract 1:11 (final concentration of protein in the assay, 0.08 mg/ml) and the bacterial extract 1:4 (final concentration of protein in the assay, 0.2 mg/ml), according to the data on the concentration of PKC␦ in both extracts (see Fig. 1B). The activity of PKC␦ produced by the bacteria proved to be comparable to that of PKC␦ expressed in insect cells (Fig. 2). Enzyme activity was measured with the substrates histone III-S (Fig. 2) and pseudosubstrate-related peptide ␦ (data not shown, but see Fig. 6). Moreover, no significant difference between both enzymes could be detected regarding either activation by TPA (Fig. 2) or specific activities and K m values for the pseudosubstrate-related peptide ␦ and histone III-S (Table I).
The expression of functional PKC␦ in bacteria indicated that PKC␦, unlike PKC␣ (24) and ␤ II (25), does not require phosphorylation of Thr 505 (corresponding to Thr 497 and Thr 500 in PKC␣ and ␤ II , respectively) as a prerequisite of a catalytically competent form of the enzyme. This could be assumed, since bacteria are thought to lack the as yet unidentified protein kinase, which is responsible for phosphorylation of this site (22)(23)(24). To prove that Thr 505 is not essential for the activation of PKC␦, we exchanged threonine 505 for alanine by sitedirected mutagenesis and transformed bacteria with the expression vector pET28␦Ala 505 containing the PKC␦-mutant cDNA. The PKC␦ mutant (PKC␦Ala 505 ) was expressed by the bacteria as effectively as the PKC␦ wild type (Fig. 1B) and did not significantly differ from the wild type regarding its TPAstimulated kinase activity (Fig. 2). Equal amounts of wild type and mutant were applied to the kinase assay, as determined by immunoblotting (see above and  peptide ␦ as substrate, and 1.1 and 1.5 units/mg, respectively, with histone III-S as substrate) as well as the K m values (5 and 6 M, respectively, for peptide ␦, and 125 and 140 M, respectively, for histone III-S) were essentially the same (Table I).
For a more detailed characterization of bacterial recombinant PKC␦ wild type and the mutant PKC␦Ala 505 we compared both enzymes with respect to several features, such as autophosphorylation, tyrosine phosphorylation by Src, and inhibition by staurosporine-related inhibitors.
The wild type as well as mutant in the respective bacterial extracts could be autophosphorylated in the presence of PS and TPA, as shown in Fig. 3. In the absence of PS and TPA or in the presence of PS alone, no or just weak autophosphorylation was observed. Regarding intensity and TPA dependence of autophosphorylation, the mutant PKC␦Ala 505 behaved like the wild type. No autophosphorylated PKC␦ could be detected in extracts of bacteria transformed with the pET28 vector alone.
Phosphorylation of PKC␦ at tyrosine residue(s) in vitro by the tyrosine kinase Src, a characteristic property of this PKC isoenzyme, results in decreased mobility of the protein in SDS-PAGE (13). Both recombinant enzymes, wild type and PKC␦Ala 505 , were phosphorylated by Src and exhibited decreased mobility on tyrosine phosphorylation, as shown by immunoblotting with an anti-PKC␦ antibody (Fig. 4A). Immunoblotting with an anti-phosphotyrosine antibody proved that phosphorylation did indeed occur at tyrosine residue(s) (Fig.  4B). Again, no significant difference between wild type and mutant could be seen.  Fig. 2) were phosphorylated with 6 units of c-Src as described under "Experimental Procedures." PKC␦ was identified by immunoblotting with an anti-PKC␦ antibody (A). After stripping of the blot, tyrosine-phosphorylated proteins, including PKC␦, were identified using an anti-phosphotyrosine antibody (B), as described under "Experimental Procedures." As secondary antibodies goat anti-rabbit and goat anti-mouse peroxidaseconjugated antibodies were used for the first and second blots, respectively. The location of PKC␦ is indicated by arrows (note the shift on tyrosine phosphorylation).

FIG. 5. Suppression of the kinase activity of bacterial recombinant PKC␦ wild type (wt) and PKC␦Ala 505 mutant (Ala505) by the staurosporine-derived inhibitors Gö 6976 (A) and Gö 6983 (B).
The enzyme activity of PKC␦ in the diluted extracts (10 l, see Fig.  2) was determined in the absence or presence of the inhibitors (concentrations as indicated) by the kinase assay, as described under "Experimental Procedures." Kinase activity was determined in the presence of PS and TPA and is given as percentage of control (activity in the absence of inhibitor).
Finally, we studied the inhibition of both types of recombinant PKC␦ by the staurosporine-related protein kinase inhibitors Gö 6976 and Gö 6983, which are specific for PKC. The two inhibitors are known to differ significantly in their capacity to suppress PKC␦, with IC 50 values of 1 M and larger for Gö 6976 and 10 -100 nM for Gö 6983 (29 -31). A similar difference in inhibitory potency was observed with the bacterial recombinant PKC␦ wild type and mutant (Fig. 5). Suppression by Gö 6976 occurred, with IC 50 values in the range of 10 -30 M and by Gö 6983 in the range of 25 nM.
We could not entirely exclude the possibility that in PKC␦ Ser 504 might be able to take over the role of Thr 505 and might serve as an essential phosphorylation site, such as Thr 497 and Thr 500 of PKC␣ and ␤ II , respectively. Therefore, we expressed two other mutants of PKC␦, PKC␦Ala 504 and PKC␦Ala 504 / Ala 505 , in the bacteria, in which serine 504 alone or serine 504 and threonine 505 were exchanged for alanine. As shown in Fig. 6, both mutants actively phosphorylated the pseudosubstrate-related peptide ␦ in a TPA-dependent manner and thus did not differ from the wild type. The same could be demonstrated with autophosphorylation of the PKC␦ mutants (data not shown).
In preliminary experiments bacterial recombinant PKC␦ (wild type and mutants) containing a short His tag could be partially purified by affinity chromatography on a nickel-nitrilo-triacetic acid resin. On elution from the column with 100 mM imidazole the purity of the enzymes was around 20%, as estimated from Coomassie Blue-stained SDS-polyacrylamide gels. The specific activities (the mean of two experiments), as determined with histone III-S as substrate, were 46

DISCUSSION
Fabbro and co-workers (32) provided the first evidence of PKC␣ being synthesized as an inactive nonphosphorylated precursor, which is at first converted to a transient and finally to a "mature" phospho form. Further studies by this and another group showed that bacterial expression of PKC␣ results in a recombinant protein devoid of kinase activity (22)(23)(24). It was suspected that phosphorylation by another protein kinase is necessary for PKC␣ to gain the ability of being activated, and that bacteria lack this putative PKC kinase. According to our results, PKC␦ can be expressed in bacteria in a functional form. Agreeing with previous reports (22)(23)(24), we were unable to express enzymatically active PKC␣ in bacteria using the same expression vector and the same conditions as for the expression of PKC␦ (data not shown). The activity of bacterially expressed PKC␦ regarding substrate phosphorylation and TPA depend-ence as well as its K m values for two substrates are comparable to those of recombinant PKC␦ expressed in baculovirus-infected insect cells. Moreover, partial purification of recombinant PKC␦ from bacterial extracts by one-step affinity chromatography yields an enzyme with a specific activity (46.2 units/mg protein) comparable to that of partially purified native PKC␦ from porcine spleen, which previously was found to be 15 units/mg protein on the second purification step (phenyl-Sepharose) and 54 units/mg protein on the third purification step (protamine-agarose; see Ref. 7). As the partially purified bacterial enzyme is around 20% pure, its specific activity is also in good agreement with the specific activity of the native enzyme purified to homogeneity from porcine spleen (304.2 units/mg protein; Ref. 7). Provided that the bacteria strain used in our studies indeed lacks the PKC kinase, these results indicate that PKC␦ does not require phosphorylation by this kinase to become a functional enzyme. Some protein kinases are able to autoactivate during bacterial expression by a presumably cotranslational intermolecular phosphorylation (33).
The sites phosphorylated by another kinase were identified in bovine PKC␣ (24) and rat PKC␤ II (25) as Thr 497 and Thr 500 , respectively. These threonines are located in an activation loop that is also crucial for the regulation of other protein kinases (34 -39). Replacement of these critical residues in PKC␣ (24) or PKC␤ II (25) with a neutral, nonphosphorylatable residue results in kinases that cannot be activated. The corresponding site in rat PKC␦ is Thr 505 . Bacterial expression of a PKC␦ mutant containing alanine in position 505 yields a fully functional kinase. The mutant does not differ from the wild type in many respects, such as effectiveness of expression in bacteria, TPA-stimulated kinase activity, K m values for substrates, autophosphorylation, tyrosine phosphorylation by Src, and inhibition by staurosporine-related inhibitors. This clearly demonstrates that functional PKC␦ can be expressed in bacteria, and that Thr 505 , contrary to Thr 497 and Thr 500 in PKC␣ and ␤ II , respectively, is not a critical site for permissive activation of PKC␦. As Ser 504 is another amino acid that can be phosphorylated in this region of PKC␦, we wished to exclude the possibility that this residue might be able to replace Thr 505 as a phosphorylation site for the putative PKC kinase and thus as the critical site for permissive activation. Exchange of Ser 504 alone or both Ser 504 and Thr 505 for alanine does not result in any loss of kinase activity. Partial purification of the bacterial PKC␦ mutants Ala 505 and Ala 504 /Ala 505 by affinity chromatography yields enzymes with specific activities of 59.5 and 54.2 units/mg protein, respectively, which are comparable to those of partially purified bacterial wild type PKC␦ and native PKC␦ from porcine spleen (see above). This proves that neither Thr 505 nor Ser 504 is essential for gaining a catalytically competent conformation of PKC␦. As each PKC isoenzyme known so far contains threonine in a position corresponding to positions 497 and 500 of PKC␣ and ␤ II , respectively, the findings regarding the critical role of this threonine residue for the activation process of PKC␣ and ␤ II have been thought to be valid for all PKC isoenzymes (24,40). According to our results, however, at least the isotype ␦ is an exception to this apparent rule. PKC␦ can be activated even when containing a neutral, nonphosphorylatable amino acid instead of this threonine residue.
In accordance with reports on a stepwise phosphorylation of PKC␣ (32,41,42) and on the in vivo phosphorylation sites of PKC␤ II (40,43), Newton and co-workers (40,44) have shown that phosphorylation of threonine 500, putatively by the PKC kinase, enables PKC␤ II to autophosphorylate at Thr 641 and Ser 660 . Phosphorylation at Thr 641 replaces the requirement for phosphate on Thr 500 and stabilizes the functional form of the enzyme. A report by Zhang et al. (45) indicates that in PKC␤ I phosphorylation of Thr 642 (corresponding to Thr 641 in PKC␤ II ) is an early event in the processing of the newly synthesized enzyme and is required for enzymatic functioning. Very recently, Bornancin and Parker (46) reported that phosphorylation of Thr 638 of PKC␣ (corresponding to Thr 642 of PKC␤ I and Thr 641 of PKC␤ II ) is not required for the catalytic function of the enzyme per se, but serves to control the duration of activation by regulating the rate of dephosphorylation and inactivation of the protein. This is achieved through the cooperative interaction between Thr 638 and the catalytic core site, Thr 497 . It is conceivable that PKC␦ also requires autophosphorylation of the corresponding site, i.e. Ser 643 , for stabilization of a catalytically competent conformation. As indicated by our results, however, PKC␦ is able to autophosphorylate without having been prephosphorylated by another kinase. As a consequence, a putative regulation of this PKC kinase would not affect PKC␦, and thus PKC kinase might be a target for a differential regulation of PKC isoenzymes. Identification of the in vivo phosphorylation sites of PKC␦ should allow an answer to the question of whether any of the three in vivo phosphorylation sites found in PKC␤ II (40,43) play a role also in PKC␦.
In the crystal structure of cAPK (47,48) the dianionic phosphoryl group of Thr 197 -P in the activation loop neutralizes a cluster of positively charged residues from several regions of the protein. Thr 197 of cAPK aligns with Thr 505 of PKC␦. To get a better idea of the structural situation around Thr 505 in PKC␦, we exchanged the side chains in the crystal structure of the catalytic core of cAPK for those of PKC␦ according to the alignment of Hanks and Quinn (49) while keeping the backbone unchanged (Fig. 7). Two of the side chain interactions with Thr 197 -P in cAPK are not possible in PKC␦; His 87 from the small lobe is replaced by Cys 391 in PKC␦, and Thr 195 is replaced by Ala 503 . A cysteine in the position equivalent to His 87 of cAPK is also conserved in PKC␣ and PKC␤ II , which both require phosphorylation at Thr 497 and Thr 500 , respectively, for activity. The basic residues Arg 470 and Lys 494 , corresponding to Arg 165 and Lys 189 of cAPK, are both conserved in PKC␦. It is likely that Arg 165 , which precedes the catalytic base, and perhaps Lys 189 are conserved in protein kinases and thus also in PKC␦ for similar functions. Both residues form salt bridges to Thr 197 -P in cAPK. The contact to Arg 165 may directly promote the correct assembly of the active site by controlling the orientation of the catalytic base Asp 166 via its peptide backbone, whereas the contact to Lys 189 may help in correctly positioning the metal binding loop, i.e. essentially Asp 184 . However, from this model it is not apparent how in PKC␦ the dispensable Thr 505 -P may be functionally substituted. Several possibilities can be discussed to explain the observed Thr 505 -independent activity of PKC␦. If an ionic interaction at this site is also needed for PKC␦, it could be provided by a bound ion, similarly as in casein kinase 1 (50,51). On the other hand, a negatively charged residue from outside the catalytic core might reach into the activation site. Another possibility is that the PKC␦ activation loop, which contains the three-residue insert Gly 499 -Glu 500 -Asn 501 , folds back to orient the Glu 500 carboxylate in a position where it can interact with Arg 470 and Lys 494 . Finally, nonionic interactions with Arg 470 , similarly to mammalian casein kinase 1 (51), are conceivable. The question remains why residue 505 is conserved as a threonine in PKC␦. The fact that Thr 505 is dispensable for the permissive activation of PKC␦ does not exclude its phosphorylation for other purposes, such as protein-protein interaction, as indicated in the interaction of cAPK catalytic and regulatory subunits (52), or enzyme inactivation.