Cyclic nucleotide-independent protein kinases from rabbit reticulocytes. Purification and characterization of protease-activated kinase II.

A cyclic nucleotide-independent protein kinase, protease-activated kinase II, which incorporates up to four phosphates into 40 S ribosomal protein S6, has been purified from the postribosomal supernatant of rabbit reticulocytes. Protease-activated kinase II was purified as an inactive proenzyme by chromatography on DEAE-cellulose, phosphocellulose, Sephadex G-150, and hydroxylapatite. The enzyme was activated in vitro by limited digestion with trypsin or chymotrypsin. No other mode of activation for protease-activated kinase II in vitro was identified. The proenzyme had a molecular weight of 80,000 as measured by gel filtration; following tryptic digestion, the molecular weight of the activated protein kinase was 45,000-55,000. Protease-activated kinase II required Mg2+ for activity but was inhibited by other divalent cations, monovalent cations, and fluoride ion. ATP was the phosphoryl donor in the phosphorylation reaction; GTP had no effect. In vitro, multiple phosphorylation of S6 was observed with some phosphate incorporated into S10. Phosphorylation of S6 by protease-activated kinase II has been shown to be stimulated in serum-starved 3T3-L1 cells by insulin (Perisic, O., and Traugh, J. A. (1983) J. Biol. Chem. 258, 9589-9592) and in reticulocytes by altering the pH of the incubation medium (Perisic, O., and Traugh, J. A. (1983) J. Biol. Chem. 258, 13998-14002.

Two cyclic AMP-independent protein kinases have been isolated from rabbit reticulocytes.
These enzymes have been resolved from the cyclic AMP-regulated activities by ion exchange chromatography on DEAEcellulose and phosphocellulose and assayed using casein as substrate.
For simplicity, the casein kinases were numbered in order of elution from DEAE-cellulose. Casein kinase I (CK I) bound to phosphocellulose and to sulfopropyl-Sephadex at low ionic strength at pH 6.8. Casein kinase II (CK II) did not adhere to phosphocellulose in the absence of monovalent cations, but bound when the concentration of these ions was raised to 0.25 M. This differential chromatography of CK II on phosphocellulose was used in the purification of the enzyme. Both CK I and CK II activities were purified further by hydroxylapatite chromatography. In the phosphorylation of casein, CK I preferentially utilized ATP over GTP. The K, for ATP and GTP was determined to be 13 pM and 900 pM, respectively. CK II utilized both ATP and GTP in the phosphotransferase reaction with a K,,, for ATP of 10 pM and 40 PM for GTP.

Analysis
of the highly purified CK II by polyacrylamide gel electrophoresis in sodium dodecyl sulfate showed three major bands of molecular weight 42,000, 38,000, and 24,000. The 24,000 molecular weight band was selfphosphorylated when the enzyme was incubated with magnesium and either ATP or GTP. In a similar experiment, a single protein band of 37,000 daltons was observed with CK I which was self-phosphorylated by incubation with magnesium and ATP. Velocity sedimentation experiments yielded a sedimentation coefficient of 3.2 S for CK I and 7.5 S for CK II. Preincubation of CK II with [Y-~~P]ATP followed by sucrose gradient centrifugation yielded a single, enzymatically active peak of 7.5 S which coincided with the radioactivity. A molecular weight of 144,000 + 10% was estimated for CK II by sedimentation-equilibrium which in combination with gel electrophoresis data suggests a heterogeneous subunit structure.
A number of enzyme activities which catalyze the posttranslational phosphorylation and dephosphorylation of proteins have been detected in diverse eukaryotic cells (1). These include cyclic nucleotide-regulated and cyclic nucleotide-independent protein kinases. The cyclic nucleotide-independent protein kinases are a class of enzymes whose function is not under direct control of either CAMP or cGMP and are not * This research was supported by United States Public Health Service Grant GM 21424. 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 IJSC. Section 1734 solely to indicate this fact. controlled by the same regulatory proteins as the CAMPregulated enzymes.
Cyclic nucleotide-independent protein kinases which phosphorylate casein have been partially purified from a variety of tissues, including rat liver (Z-9), human lymphocytes (lo), calf brain (ll), dogfish skeletal muscle (12), mouse plasmacytoma (I3), and rabbit reticulocytes (14) and erythrocytes (14,15). Highly purified casein kinase act.ivities have been reported recently from yeast (16), Novikoff ascites tumor cells (17), and rat liver (18). The physiological function of these enzymes remains a subject for speculation. This paper deals with the purification and properties of two cytoplasmic cyclic nucleotide-independent protein kinases from rabbit reticulocytes.

Lysate
Preparation-The preparation of the reticulocyte lyaate has been described previously (19). Protein Kinase Assay-The assay for protein kinase was carried out as previously described (20). Determination of K,,, for AZ'P and GTP-Initial velocity data were obtained at 30°C under optimal assay conditions as described above and in Table II Gel Electrophoresis-Gel electrophoresis was carried out in the presence of sodium dodecyl sulfate (24) in a slab gel apparatus (25) as previously described (20). Phosphorylase a, bovine serum albumin, creatine kinase, carbonic anhydrase, and rihonuclease were included as standards.
Their molecular weights were taken as 94,000, 68,000, 40,000, 30,000, and 13,700, respectively (26). The gel was stained, destained, and autoradiographed as described previously (20). Protein and radioactivity were quantified by scanning the gel and autoradiogram with a densitometer (EC Apparatus Corp.). Radioactivity was also quantified by excision and counting as described previously (20). and histone (Fig. 1). Two peaks of cyclic AMPindependent protein kinase activity were detected which phosphorylated casein and used ATP as the phosphate donor. matic activity was routinely detected. CK II has been previously reported in rabbit reticulocytes (formerly identified as I&) (14,29), and an activity with a similar elution pattern has also been described in rabbit erythrocytes (14, 1.5 (Fig. 2A). A minor amount of casein kinase activity did not adhere to the phosphocellulose resin. This enzyme comprised only 10% of the total casein activity applied to the phosphocellulose column. It had chromatographic properties which were similar to CK II and undoubtedly corresponded to a small contaminating amount of that enzyme. CK II behaved somewhat anomalously on phosphocellulose. In the absence of monovalent cations, about 90% of the protein kinase activity did not adhere to the resin. The fraction which did not bind to the initial phosphocellulose column was dialyzed against Buffer B containing 0.25 M NaCl and applied to a second phosphocellulose column. At this salt concentration the enzyme adhered to the resin and was eluted as a single peak of activity from 0.70 M to 0.85 M NaCl (Fig. 2B). This anomalous behavior on phosphocellulose was used to enhance the purification of the enzyme. At this point the CAMP-dependent activities were well resolved from the cyclic AMP-independent casein kinase since both type I and type II CAMP-dependent kinases were not bound to phosphocellulose under these conditions. Chromatography of CK I on Sulfopropyl-Sephadex-CK I adhered to the resin and appeared in fractions eluting between 0.3 M and 0.5 M NaCl (Fig. 3). Binding of the protein kinase to the column was very dependent on pH as the enzyme did not bind to the resin when the pH was raised to 7.1. CK I was extremely unstable after this step in the purification, and it was pooled and concentrated immediately to help stabilize the activity.   (Fig. 4). After a brief dialysis against Buffer B, the enzymes were concentrated by batch chromatography from lml hydroxylapatite columns. They were stored at 4°C in Buffer B which contained 0.4 M potassium phosphate, pH 6.8. A typical purification of CK I and CK II is summarized in Table I.

Analysis of CK I and CK II by Polyacrylamide
Gel Electrophoresis-CK I yielded a major band of molecular weight 37,000 when electrophoresed on polyacrylamide gels containing sodium dodecyl sulfate. Preincubation with [Y-~~P]ATP followed by electrophoresis and autoradiography resulted in one phosphorylated band corresponding to the 37,000 molecular weight protein (Fig. 5A). CK II was analyzed by polyacrylamide gel electrophoresis in sodium dodecyl sulfate, and three major bands with molecular weights of 42,000, 38,000, and 24,000 were observed as shown in Fig. 5B. When the enzyme was incubated with [Y-~~P]ATP and analyzed by gel I 25 50 FRACTION NO electrophoresis followed by autoradiography, radioactive phosphate was detected and associated exclusively with the 24,000-dalton protein. Only one-third as much phosphate was incorporated during the 30-min incubation period when equal concentrations of GTP were substituted for ATP. Variable amounts of CK II ranging from 6 pg to 30 pg were electrophoresed, stained with Coomassie blue R-250, and scanned with a densitometer (547 nm) to determine the relative amount of protein in each band. Integration of the traces and correction for dye binding on a weight basis (31) yielded an average molar ratio of 1.3:1.0:1.6 for the 42,000,38,000, and 24,000 molecular weight subunits, respectively.
Analytical Ultracentrifugation-CK II was centrifuged to experimental equilibrium (based on the identical distributions obtained after 16 and 21 h of centrifugation).
The long column technique of Chervenka (28)  A solution containing 5 pg of CK II was layered at low speed on a buffered 0.5 M NaCl column as described under "Methods" and the rotor accelerated to 59,780 rpm. The direction of sedimentation was from left to right. Scans were made at 4-min intervals with the monochromator set at 236 nm. A, initial absorbance trace after layering; B, after 66 min at speed. full 3 mm of the column and the logarithmic plot was linear over 70% of the distribution. A value of 144,000 rt 14,000 was calculated for the apparent weight average molecular weight of CK II. The stated error arises largely from the uncertainty in the value chosen for the apparent isopotential specific volume. In the absence of density data, we have assumed a value of 0.74 + 0.02 ml/g. Near the bottom of the column, a limiting value of about 220,000 was calculated from the slope of the logarithmic plot (32). Sedimentation velocity experiments with CK II in the optical centrifuge showed a single peak sedimenting with an~&,,~~ value of 7.5 (Fig. 6). Insufficient quantities of CK I were available to perform experiments with the optical centrifuge and, therefore, velocity experiments had to be done with sucrose density gradients. Five velocity experiments were performed which yielded a value of 3.2 + 0.15 for the sedimentation coefficient of CK I (Fig. 7A). In these experiments, the s value of CK II which had been self-phosphorylated with ATP was also determined. A value of 7.5 & 0.3 S was calculated relative to ovalbumin and catalase standards, and the enzymatic activity coincided with the radiolabel incorporated during the self-phosphorylation process (Fig.  7B). Therefore, we concluded that phosphorylation of CK II did not effect a change in the molecular weight of this enzyme. Taken together, the gel electrophoresis data and the centrifuge data suggest a structure for the CK II enzyme which would be composed of two 24,000 molecular weight subunits and one each of the 42,000 and 38,000 molecular weight subunits.
Determination of K, for ATP and GTP-Lineweaver-Burk plots were constructed for CK I and for CK II when ATP and GTP were used as phosphate donors (Fig. 8). A summary of the K, values for the enzymes is given in Table II

DISCUSSION
Previous studies on the CAMP-independent protein kinases from the postribosomal supernatant fraction from rabbit reticulocytes had shown the presence of a single peak of activity eluting from DEAE-cellulose (14). This peak was termed IIIc and was identical with CK II described here. CK I was not observed when an ammonium sulfate precipitation preceded the DEAE-cellulose chromatography step. Thus, this is the initial report on the second cytoplasmic casein kinase activity, although previous studies had shown at least two casein kinase activities were associated with the protein-synthesizing complex. Centrifugation through 0.5 M NaCl dissociated these latter activities from the complex (29). Kumar and Tao (15) have reported two CAMP-independent protein kinase activities from rabbit erythrocytes.
These enzymes were similar chromatographically to CK II and utilized both ATP and GTP in the phosphotransferase reaction. The K,,, values for ATP and GTP differed significantly from those reported here; however, this may be due to the fact that their studies were carried out at pH 9.0. We have observed that CK II aggregated when the monovalent salt concentration was less than 0.5 M. This may account for the very high molecular weight values observed by Kumar and Tao (15) and suggests that the two peaks may be different aggregation states of CK II.
The anomalous behavior of CK II on phosphocellulose has been noted. When the small amount of activity (usually less than 10%) which binds under conditions of low salt was carried through the sulfopropyl-Sephadex and hydroxylapatite steps, a high molecular weight contaminant was found to co-chromatograph.
This contaminant was not present in the phosphocellulose flow-through fraction which contained the majority of the CK II activity. Therefore, we have routinely included phosphocellulose chromatography at low salt in our procedure, even though only a small overall purification was realized.
Both CK I and CK II lose activity rapidly in the latter stages of purification which we attribute to the general decline in protein concentration.
It is important, therefore, to maintain stock solutions of these enzymes at the highest practical concentrations of protein. We have found that this was accomplished most satisfactorily by batch elution from small (1 to 2 ml) hydroxylapatite columns (95 to 100% yield). A subunit molecular weight of about 37,000 was determined for CK I by gel electrophoresis.
Assuming a globular shape for CK I, the s value of 3.2 obtained via centrifugation in sucrose translates to about 37,000 (33). This suggests that CK I is a single subunit enzyme. The molecular weight data from gel electrophoresis and the ultracentrifuge are consistent with a heterogeneous subunit structure for CK II. An enzyme with subunit molecular weights of 42,000, 38,000, and 24,000 in a ratio of 1:1:2 as suggested by gel electrophoresis would yield a native molecular weight of 128,000 + 6,500. This would be consistent with both the molecular weight of 144,000 + 14,000 obtained by equilibrium centrifugation and the 7.5 S velocity coefficient. A similar structure has been proposed for a casein kinase activity purified from Novikoff ascites tumor cells (17) and rat liver (18). Our finding that only the smaller subunit (24,000) is self-phosphorylated is also similar to that found by others (17). The radiolabeling of the native 7.5 S complex confirmed the gel electrophoresis data which suggested that it is indeed a subunit of CK II. Attempts to further purify CK II by binding the enzyme to adenosine-agarose and ATP-Sepharose or by chromatography on Sephadex G-100 showed no alteration in the subunit pattern." When the time course for the self-phosphorylation of CK II was examined, 1.7 mol