Purification and Properties of a Distinct Protamine Kinase from the Cytosol of Bovine Kidney Cortex*

A protamine kinase has been purified to apparent homogeneity from extracts of the cytosol of bovine kidney cortex. This protamine kinase exhibited an ap- parent M, = 43,000 as estimated by gel permeation chromatography on Sephacryl S-200 and an apparent M, = 45,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified protamine kinase exhibited about 5% activity with casein, 8% with histone H2B, and <0.1% with histone H1, histone H4, glycogen synthase a from rabbit skeletal muscle, ovalbumin, bovine serum albumin, and phos-vitin. The activity of the highly purified protamine kinase was unaffected by cyclic AMP (up to 0.1 mM), cyclic GMP (up to 0.1 mM), the heat-stable protein inhibitor of cyclic AMP-dependent protein kinase (up to 100 pg/ml), heparin (up to 100 pg/ml), EGTA (up to 1 mM), ca2+ (up to 1 mM), calmodulin (up to 0.5 pM) in the absence or presence of Ca2+ (0.05 mM), and phos- phatidylserine (up to 40 pg/ml) and/or diolein (up to 1 pg/ml) in the absence or presence of Ca2+ (up to 0.5 mM). Experiments in which extracts of kidney cytosol were incubated with [-pS2P]ATP and MgClz revealed that the phosphorylation of numerous polypeptides was

The regulation of enzyme activity by phosphorylation/ dephosphorylation is an important mechanism for the coordinated control of cellular activity in response to extracellular stimuli (1, 2). Over the last few years, it has become evident that cells contain numerous protein kinases. The substrate specificity and subunit composition of many of these enzymes have been determined (reviewed in Ref. 3). The availability of this information has provided a firm molecular basis for further characterization of the structure, physiological role, and regulation of these kinases.
Recently, two forms of a protamine kinase have been identified from the soluble fraction of bovine kidney mitochondria (4). Both forms of this protamine kinase were present in an inactive form in mitochondrial extracts and were only detected following an initial chromatography on poly(L-lysine)agarose. The two forms of the mitochondrial protamine kinase were separated by chromatography on protamine-agarose, and each form was purified about 80,000-fold to apparent homo-* This work was supported in part by National Institutes of Health Biomedical Research Support Grant Award SO7 RR07160. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
geneity. Both forms of the protamine kinase exhibited an apparent M, = 45,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and gel filtration chromatography on Sephacryl S-200. Both forms of the protamine kinase underwent autophosphorylation and exhibited identical substrate specificities. The properties of the two forms of the protamine kinase of bovine kidney mitochondria distinguished these enzymes from previously described protein kinases.
In studies designed to determine the subcellular distribution of the mitochondrial protamine kinase in bovine kidney, we detected 40-100-fold higher protamine kinase activity in extracts of cytosol. In order to determine the relationship, if any, of cytosolic protamine kinase to the two forms of the mitochondrial protamine kinase, the protamine kinase of kidney cytosol was examined. In this paper, we report the purification of the cytosolic protamine kinase to apparent homogeneity and describe studies on the substrate specificity of this protamine kinase. The results indicate that the cytosolic protamine kinase is distinct from the two forms of the mitochondrial protamine kinase and also from previously described protein kinases. The results also indicate that the cytosolic protamine kinase exhibits activity toward a large number of physiological polypeptides. phosphoryl groups was limited to <3 nmol. This is equivalent to <20% conversion of the limiting substrate.
O t h r Protein Kinase Assays-Determination of mitochondrial protamine kinase activity was performed as described for the cytosolic protamine kinase except that 0.25 mg of ovalbumin and 1.5 mM MgC12 were used instead of bovine serum albumin and 10 mM MgC12. One unit of mitochondrial protamine kinase activity is defined as the amount of enzyme that incorporated 1 pmol of 32P into protamine/ min. The assay for cyclic AMP-dependent protein kinase activity was as described for the cytosolic protamine kinase except that 1.0 mg of histone H2B was used instead of protamine, and reactions were carried out in the absence and presence of 10 p~ cyclic AMP. One unit of cyclic AMP-dependent protein kinase is defined as the amount of enzyme that incorporated 1 pmol of 32P into histone H2B/min. Protein kinase C activity was determined as described for the protamine kinase except that 1.0 mg of histone H1 was used instead of protamine, and 0.5 mM CaCI2 and 40 pg/ml phosphatidylserine were included in the incubations. One unit of protein kinase C activity is defined as the amount of enzyme that incorporated 1 pmol of 32P into histone Hl/min. Protein was determined as described (6). Polyacrylamide gel electrophoresis was performed in slab gels (12% acrylamide) with 0.1% sodium dodecyl sulfate and Tris/glycine buffer, pH 8.3 (7). Protein bands were detected by staining with Coomassie Blue. Radioactive bands were located with Kodak X-Omat AR-5 film.

RESULTS
Purificatwn of the Cytosolic Protamine Kinuse-All operations were carried out at 2-5 "C. Fresh bovine kidney was obtained from a local abattoir and transferred to the laboratory on ice. The kidney cortex was removed and homogenized in a Waring blender for 1 min in 2 volumes of buffer A (10 mM KP04, pH 7.3, containing 250 mM sucrose, 1 mM EDTA, 1 mM benzamidine, 0.1 mM phenylmethylsulfonyl fluoride, and 14 mM j3-mercaptoethanol). The homogenate was centrifuged at 14,000 x g for 30 min in a Beckman JA-10 rotor. The pellets were discarded and to the extract (step 1) was added, with stirring, 0.08 volumes of 50% (w/v) aqueous poly(ethy1ene glycol) 8000. After 20 min, this solution was centrifuged and the precipitate discarded. The supernatant was then applied onto a column (14 X 10 cm) of DEAEcellulose equilibrated in buffer B (50 mM imidazole, pH 7.3, 10% glycerol, 1 mM EDTA, 1 mM benzamidine, 0.1 mM phenylmethylsulfonyl fluoride, and 14 mM B-mercaptoethanol). The column was then washed with 4 liters of buffer B containing 0.1 M NaCI. About 80% of cyclic AMP-dependent protein kinase and protein kinase C-like activities were eluted in this wash (not shown). The major protamine kinase activity (about 70% of the total recovered from DEAE-cellulose) was then eluted with buffer B containing 0.25 M NaCl. To this solution (step 2) was added, with stirring, solid ammonium sulfate (390 g/liter). The solution was stirred for 30 min and centrifuged at 30,000 x g for 30 min in a Beckman JA-14 rotor. The supernatant was discarded and the pellets resuspended in buffer B and dialyzed against 20 volumes of the same buffer with three changes in 16 h. This solution (step 3) was applied to a column (2. were concentrated to 2 ml on Q-Sepharose as described in the text. This solution was then applied to a second column (2.5 X 95 cm) of Sephacryl S-200 equilibrated and developed with buffer B containing 0.5 M NaC1. The flow rate was 18 ml/h, and 2.0-ml fractions were collected. Molecular weights are shown as M, X The protein standards were bovine serum albumin, ovalbumin, and ribonuclease. Vo was determined with ferritin and V, with FMN. Protein was determined with 0.1-ml aliquots by the Bradford procedure (6). Inset, sodium dodecyl sulfate-polyacrylamide gel electrophoresis pattern of the purified protamine kinase (10 pg). The gel was stained with Coomassie Blue. The position of marker proteins, M, X from top to bottom is phosphorylase, bovine serum albumin, ovalbumin, carbonic anhydrase, trypsin inhibitor, and a-lactalbumin.   Table 11.
of M e and [y3*P]ATP, even after overnight incubation at 30 "C (not shown).
Comparison of the Catalytic Properties of the Cytosolic and Mitochondrial Protamine Kinases-The optimal concentration of M$+ was 10 mM for the cytosolic protamine kinase and 1.5 mM for the two forms of the mitochondrial enzyme (Fig. 2). The apparent K, for M e was estimated to be about 1.8 mM for the cytosolic kinase and about 0.04 mM for the mitochondrial kinases. At 1 mM, Ca2+ had little or no effect on the activity of the cytosolic protamine kinase but inhibited the two forms of the mitochondrial protamine kinase by 80%.
Half-maximal inhibition of the mitochondrial protamine kinases occurred at about 0.5 mM Ca2+. EGTA' (up to 1 mM) was without effect on the cytosolic or mitochondrial enzymes. Similarly NaCl (up to 0.5 M) or KC1 (up to 0.5 M) was without effect.
Theabbreviationusedis: EGTA, [ethylenebis(oxyethylenenitrilo)] tetraacetic acid. Protamine was preferentially phosphorylated by the purified cytosolic and mitochondrial protamine kinases. The two forms of the mitochondrial protamine kinase exhibited activity toward histone H1 (-go%), bovine serum albumin (-48%), and glycogen synthase from rabbit skeletal muscle (-15%). With these substrates the cytosolic protamine kinase was essentially inactive. The cytosolic protamine kinase exhibited some activity with histone H2B (-8%) and casein (-5%). With these substrates the mitochondrial protamine kinases were inactive. The cytosolic and mitochondrial protamine kinases exhibited little activity ((0.1%) with histone H4, ovalbumin, phosvitin, and the synthetic polypeptide poly (Glu,Tyr) (4:l). With protamine as substrate, each of the purified enzymes exhibited a broad pH optimum of 5.7-9.0. Differences in the substrate specificities of the purified cytosolic and mitochondrial protamine kinases are summarized in Table 11.
The activity of the cytosolic protamine kinase, like the activities of the mitochondrial protamine kinases (4), was unaffected by cyclic AMP (up to 0.1 mM) or cyclic GMP (up to 0.1 mM). Similarly, the heat-stable inhibitor of cyclic AMPdependent protein kinase at concentrations (up to 100 pglml) that completely inhibited the catalytic subunit of cyclic AMPdependent protein kinase with either histone H2B or protamine as substrate was without effect on the activity of the cytosolic protamine kinase with protamine, histone H2B, or casein. Heparin up to 100 pg/ml had no effect on the cytosolic enzyme with protamine or casein as a substrate. The activity of the cytosolic protamine kinase, like the activities of the two forms of the mitochondrial protamine kinase (this study and Ref. 4), was also unaffected by calmodulin (up to 0.5 p~) with or without Ca2+ (0.05 mM). In addition, phosphatidylserine (up to 40 pg/ml) and/or diolein (up to 1 pg/ml), in the presence or presence of Ca2+ (up to 0.5 mM), were without effect on the activities of the cytosolic protamine kinase and the two forms of the mitochondrial kinase. At 0.2 mM, GTP did not replace ATP in the cytosolic or mitochondrial protamine kinase reactions.
Physiological Role of the Cytosolic Protamine Kinase-As a first step in our studies on the physiological role of the cytosolic protamine kinase, extracts of kidney cytosol were incubated with [y3*P]ATP and MgClz in the absence and The final concentration of all the proteins tested was 10 mg/ml except for glycogen synthase a (1 mg/ml) and casein (6 mg/ml). The activity of the kinases with bovine serum albumin was determined as described (4). All other reactions were performed as described under "Experimental Procedures." *Activity is expressed as percent of the activity observed with protamine. e Mitochondrial protamine kinase I and mitochondrial protamine kinase I1 refer to the forms of this enzyme which elute at 0.6 and 0.8 M NaCl from protamine-agarose, respectively (4).
The relative activity of the cytosolic protamine kinase with the proteins shown was not altered when the activity of this enzyme was determined at 1.5 mM M$+ (not shown). presence of highly purified preparations of the protamine kinase (Fig. 3). A marked increase in the phosphorylation of numerous polypeptides was observed in the incubations which contained the purified protamine kinase (Fig. 3B).

DISCUSSION
The protamine kinase described in this paper was detected in studies designed to determine the subcellular distribution of two forms of a protamine kinase that were identified and purified to apparent homogeneity from extracts of bovine kidney mitochondria (4). We estimate that extracts of bovine kidney cytosol contained between 40-and 100-fold higher protamine kinase activity than was present in mitochondrial extracts. Based on the total recoveries following fractionation of extracts on DEAE-cellulose, we estimate that about 30% of the total cytosolic protamine kinase activity was due to cyclic AMP-dependent protein kinase, protein kinase C, and an unidentified protamine kinase which did not bind to DEAE-cellulose (not shown). The remaining protamine kinase activity (-70%) was due to the protamine kinase described in this paper.
In order to provide a firm molecular framework to examine the properties of cytosolic protamine kinase, a procedure was developed to purify this enzyme about 30,000-fold to apparent homogeneity from extracts of bovine kidney cytosol ( Table  I). The purified cytosolic protamine kinase exhibited an apparent M , = 43,000 as determined by gel permeation chromatography on Sephacryl S-200 (Fig. 1) and an apparent MI = 45,000 as estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 1, inset). Although the apparent MI of the cytosolic protamine kinase is similar to the apparent MI of the mitochondrial protamine kinases (4), in contrast to these enzymes (4), the cytosolic enzyme did not undergo autophosphorylation. In addition, although the steps employed to purify the major protamine kinase of kidney cytosol (Table I) were similar to the steps used in the purification of the mitochondrial protamine kinases (4), the cytosolic enzyme exhibited several distinct chromatographic prop- 32P]ATP (180 cpm/pmol) and 10 mM MgCL After 5 min of incubation at 30 "C, reactions were terminated by the addition of an equal volume of denaturing sample buffer, heated for 5 min at 100 "C, and then subjected to polyacrylamide gel electrophoresis (7). The gel was stained with Coomassie Blue, washed extensively, dried, and exposed to x-ray film. The position of marker proteins, M, X from top to bottom is phosphorylase, bovine serum albumin, ovalbumin, carbonic anhydrase, trypsin inhibitor, and dactalbumin. The protamine kinase used in these experiments was from step 7, Table I. erties. Thus, by contrast to the two forms of the mitochondrial protamine kinase which coeluted from Q-Sepharose at about 0.2 M NaCl (4), the cytosolic protamine kinase was eluted from this column at about 0.35 M NaCl. In addition, by contrast to the two forms of the mitochondrial protamine kinase recovered at about 0.6 and 0.8 M NaCl from protamineagarose (4), only one form of the cytosolic kinase was recovered from protamine-agarose at about 0.7 M NaCl.
The substrate specificity of the highly purified preparations of the cytosolic protamine kinase indicates that this enzyme is distinct from the two forms of the mitochondrial protamine kinase (Table 11). Thus, by contrast to the two forms of the mitochondrial protamine kinase, the cytosolic enzyme was inactive with histone H1, bovine serum albumin, and glycogen synthase a from rabbit skeletal muscle. In addition, by contrast to the two forms of the mitochondrial protamine kinase, the cytosolic enzyme exhibited some activity with histone H2B (-8%) and casein (-5%). Moreover, the optimal concentration of M e for the cytosolic protamine kinase was 10 mM as compared to 1.5 mM for the mitochondrial protamine kinases (Fig. 2). At 10 mM Mg+, the activity of the mitochondrial protamine kinase was inhibited by 40% (Fig. 2). Furthermore, the activities of the two forms of the mitochondrial protamine kinase were inhibited 80% by 1 mM Ca2+. In contrast, the activity of the cytosolic protamine kinase was unaffected by this divalent cation.
The catalytic properties of the highly purified preparations of the cytosolic protamine kinase also differentiate this enzyme from previously described cytosolic protein kinases. Thus, by contrast to cyclic AMP-dependent protein kinase, phosphorylase kinase, Ca2+/calmodulin-dependent multiprotein kinase, glycogen synthase kinase -3 and -4, type I and type I1 casein kinases, protein kinase C, and a new glycogen synthase kinase (Refs. 2,4,8-11, and references therein), the cytosolic protamine kinase was essentially inactive with glycogen synthase a from rabbit skeletal muscle. In addition, the purified protamine kinase was unaffected by 0.1 mM cyclic AMP, 0.1 mM cyclic GMP, or by the heat-stable protein inhibitor of the cyclic AMP-dependent protein kinase (up to 100 pglml), and the enzyme exhibited only about 8% activity with histone H2B, a good substrate for cyclic AMP-dependent protein kinase (this study and Ref. 12). These properties also indicate that this protamine kinase is distinct from cyclic AMP-dependent protein kinase (13) and cyclic GMP-dependent protein kinase (14). Similarly, the activity of the protamine kinase was unaffected by EGTA (up to 1 mM), Ca2+ (up to 1 mM), or calmodulin (up to 0.5 p~) with or without Ca2+ (0.05 mM), and only 0.15 mol of phosphoryl groups was incorporated per mol of phosphorylase 6 monomer following 1 h of incubation with the kinase. These properties differentiate the protamine kinase from phosphorylase kinase (15) and other Ca2'/calmodulin-dependent protein kinases (e.g. Ref. 16). The cytosolic protamine kinase also exhibited <0.1% activity with histone H1 in the absence or presence of phosphatidylserine (up to 40 pg/ml) and/or diolein (up to 1 pg/ ml), with or without Ca2+ (up to 0.5 mM). These properties distinguish the protamine kinase from the catalytic domain and native forms of protein kinase C (17), as well as other histone H1 kinases (18,19). The purified cytosolic protamine kinase was also inactive with histone H4, a substrate of S6 kinase (20), and protease-activated kinase I (21) and I1 (22,23). Similarly, by contrast to the type I and type I1 casein kinases (8), the cytosolic protamine kinase exhibited only about 8% activity with casein and was essentially inactive with phosvitin. Heparin (up to 100 pglml), a potent inhibitor of type I1 casein kinases (8), also had no effect on protamine kinase activity with either protamine or casein as substrates.
The physiological role and regulation of the protamine kinase of kidney cytosol remain to be elucidated. The experiments presented in Fig. 3, however, indicate that this protamine kinase exhibits activity toward a large number of physiological proteins. These results raise the possibility that this protamine kinase may occupy a central position in the coordinated regulation of a number of metabolic processes and that by analogy to other protein kinases (1-3), this protamine kinase may be a key target for regulation by hormones and other extracellular stimuli.