Cdk5: a novel role in learning and memory

Learning and memory are processes by which organisms acquire, retain and retrieve information. They result in modifications of behavior in response to new or previously encountered stimuli thereby enabling adaptation to a permanently changing environment. Protein phosphorylation has long been known to play a key role in triggering synaptic changes underlying learning and memory. Although intracellular phosphorylation and dephosphorylation is orchestrated by a complex network of interactions between a number of protein kinases and phosphatases, significant advances in the understanding of neuronal mechanisms underlying learning and memory have been achieved by investigating the actions of individual molecules under defined experimental conditions, brain areas, neuronal cells and their subcellular compartments. On the basis of these approaches, the cyclic AMP protein kinase (PKA), protein kinase C (PKC) and extracellularly regulated protein kinases 1 and 2 (Erk-1/2) have been identified as the core signaling pathways in memory consolidation. Here we review recent findings demonstrating an important novel role for Cdk5 in learning and memory. We suggest that some of the well-characterized roles of Cdk5 during neurodevelopmental processes, such as interactions with distinct cytoplasmic and synaptic target molecules, may be also involved in synaptic plasticity underlying memory consolidation within the adult central nervous system.     

Adenosine 3′:5′-cyclic phosphate-dependent and -independent protein kinases of renal brush border membranes: Solubilization,separation, and characterization of multiple forms

Multiple protein kinase activities were found in the luminal segment of the renal proximal tubule cell plasma membrane (brush border membrane). Membranes were extracted with Lubrol, with no loss in activity, and the extract was chromatographed on diethylaminoethyl cellulose with a salt gradient. With protamine as substrate, activity eluted in two peaks, designated I and IIb, and was cyclic AMP independent. With histone VII-S, one peak, designated IIa, appeared, which eluted slightly ahead of IIb and was cyclic AMP dependent. The three activities eluted in their original patterns following rechromatography. Histone kinase activity in the combined IIa+b fraction was stimulated threefold by cyclic nucleotides (K_a = 0.013 and 0.94 μM for cyclic AMP and cyclic GMP, respectively) by increasing V. Cyclic AMP binding activity eluted with histone kinase activity. Rechromatography of IIa+b on diethylaminoethyl cellulose containing 1 μm cyclic AMP resulted in passage through the column of most of the histone kinase activity (IIa) prior to the salt gradient, but retention of kinase IIb, which again eluted in its original position. Characterization of the separated enzymes revealed that kinase I was highly specific for protamine and totally insensitive to cyclic AMP and a specific protein inhibitor of cyclic AMP-dependent kinases. Kinase IIa was relatively specific for histones and was completely inhibited by the protein inhibitor. Kinase IIb was nonspecific, catalyzing phosphorylation of protamine, casein, histones, and phosvitin in decreasing order of activity, and was insensitive to cyclic AMP and the protein inhibitor. Exposure of intact brush border membranes to elevated temperatures revealed that phosphorylation of intrinsic membrane proteins and protamine was thermolabile, whereas cyclic AMP-dependent histone kinase activity was relatively thermostable. These findings implicate cyclic AMP-independent protamine kinases in the cyclic AMP-independent autophosphorylation of the brush border membrane.     

Inactivation of cyclic GMP-dependent protein kinase by N alpha-tosyl-L-lysine chloromethylketone

Cyclic GMP-dependent protein kinase from bovine lung and cyclic AMP-dependent protein kinase from bovine heart are inactivated by N^α-tosyl-L-lysine chloromethylketone (TLCK) in the presence of cyclic GMP and cyclic AMP, respectively. The inactivation of both protein kinases is pseudo-first order, suggesting the rate limiting step is beyond the binding of TLCK. Cyclic GMP-dependent protein kinase is inactivated less than $\text{1}\text{4}$ as rapidly as cyclic AMP-dependent protein kinase, although it shows a higher apparent affinity for TLCK. Cyclic AMP stimulated the rate of inactivation of cyclic AMP-dependent protein kinase 10-fold but cyclic GMP stimulated the rate of inactivation of cyclic GMP-dependent protein kinase only 1.5-fold. The rate of inactivation of cyclic GMP-dependent protein kinase by TLCK is sufficiently rapid (half-time of about 30 min at 37°C with 2 mM TLCK) to account for the effects of TLCK on cell growth observed by others.     

Characterization of two adenosine 3':5'-phosphate-dependent protein kinase species from Chinese hamster ovary cells.

Chinese hamster ovary cells exhibit several characteristic morphological and physiological responses upon treatment with agents which increase the intracellular level of adenosine 3':5'-phosphate (cyclic AMP). To better understand the mechanism of these cyclic AMP-mediated responses, we separated two cyclic AMP-dependent protein kinases (ATP:protein phosphotransferase, EC 2.7.1.37) (protein kinase I and protein kinase II) from the cytosol of Chinese hamster ovary cells by DEAE-cellulose chromatography and studied their properties. Protein kinase I is eluted at a lower salt concentration than protein kinase II and is stimulable to 10 times its basal catalytic activity, while protein kinase II is stimulable only 2-fold. Both kinases are completely dissociated by cyclic AMP and inhibited by specific cyclic AMP-dependent protein kinase inhibitor. They have similar Km values for magnesium (approximately 1 mM), cyclic AMP (approximately 60 nM), and ATP (approximately 0.1 mM), and the dissociation constant (Kdis) for cyclic AMP (approximately 13 nM) is the same for both enzymes. However, they appear to have different substrate preferences and cyclic AMP-binding properties in that cyclic AMP bound to protein kinase II exchanges readily with free cyclic AMP, while that bound to protein kinase I is not exchangeable. The native enzymes have different sedimentation coefficients (6.4 S for protein kinase I and 4.8 S for protein kinase II), whereas those of the activated enzymes are the same (2.9--3.0 S). It appears that the two cyclic AMP-dependent protein kinases which differ from each other in their regulatory subunits may play different roles in the mediation of cyclic AMP action in Chinese hamster ovary cells.     

A cyclic AMP dependent protein kinase in Dictyostelium discoideum

A cyclic AMP-dependent protein kinase was found to appear during the time course of development of $\text{Dictyostelium}\text{discoideum}$. No cyclic AMP dependency was observed at any stage of development in crude 110,000 X G soluble extracts. After partial purification, however, extracts from post-aggregation stages contained enzyme that was activated up to 6-fold by cyclic AMP, whereas protein kinase from earlier stages was not affected by cyclic AMP. Likewise, cyclic AMP binding activity increased from the aggregation to the slug stage of development. Approximately one-half of the total cyclic AMP binding activity co-purified with the cyclic AMP dependent protein kinase. The enzyme from $\text{Dictyostelium}$ showed similarities to mammalian protein kinases with respect to its kinetic properties but differed in its behavior on ion-exchange chromatography.     

Further studies on cyclic adenosine 3':5'-monophosphate protein kinase from dimorphic fungus Mucor rouxii

In this paper, cyclic adenosine-3′:5′-monophosphate-dependent protein kinase from yeast-like cells of Mucor rouxii is characterized. A scheme of partial purification is described together with K_m for ATP (15 μm), histone (0.2 mg/ml), half-maximal activation constant for cyclic AMP (30 nm), and dissociation constant for the binding of cyclic AMP (40 nm). This enzyme is similar to type II protein kinases in two main aspects: the elution position in DEAE-cellulose chromatography and the readiness of its reassociation. But it has a singular characteristic: it does not dissociate completely with cyclic AMP alone (even at concentrations as high as 0.3 mm) unless histone or NaCl is present. NaCl displays several roles: helps dissociation, prevents inactivation of the catalytic subunit, inhibits enzyme activity, and does not prevent reassociation as occurs with type II protein kinases. Once the holoenzyme is dissociated, cyclic AMP is essential to maintain the enzyme in the dissociated state.     

Effects of cyclic adenosine 3': 5'-monophosphate on phosphoprotein kinase and phosphatase fractions prepared from rat liver nuclei.

A soluble rat liver nuclear extract containing total RNA polymerase activities also exhibits appreciable amounts of protein kinase activity. This unfractionated protein kinase catalyzes the phosphorylation of both endogenous proteins and exogenous lysine-rich histone in the presence of [γ-³²P]ATP and Mg²⁺. The optimal concentration of Mg²⁺ is 5 mm for histone phosphorylation and 25 mm for the phosphorylation of endogenous proteins. Cyclic AMP has no effect on the phosphorylation of lysine-rich histone by this unfractionated nuclear protein kinase. However, addition of cyclic AMP causes a reduction in the ³²P-labeling of an endogenous protein (CAI) which can be characterized by its mobility during SDS-acrylamide gel electrophoresis and elution in the unbound fraction of a DEAESephadex column. If CAI is first labeled with ³²P and then incubated with 10^?6m cyclic AMP under conditions where protein kinase activity is inhibited, the presence of the cyclic nucleotide causes a loss of the ³²P-labeling of this protein, implying the activation of a substrate-specific protein phosphatase. When rat liver RNA polymerases are purified by DEAE-Sephadex chromatography, protein kinase activity is found in the unbound fraction and in those column fractions containing RNA polymerase I and II. The fractionated protein kinases exhibit different responses to cyclic AMP, the unbound protein kinase being stimulated and the RNA polymerase-associated protein kinases being dramatically inhibited. A second protein (CAII) whose phosphorylated state is modified by cyclic AMP is found within the DEAE-Sephadex column fractions containing RNA polymerase II. The cyclic nucleotide in this case appears to reduce labeling of CAII by inhibition of the protein kinase activity which co-chromatographs with both CAII and RNA polymerase II. Based on molecular weight estimates, neither CAI nor CAII appears to be an RNA polymerase subunit. The identity of CAI as a protein factor whose phosphorylated state influences nuclear RNA synthesis is suggested by the fact that addition of fractions containing CAI to purified RNA polymerase II inhibits the activity of this enzyme, but only if CAI has been previously incubated in the presence of cyclic AMP.     

Action of cyclic nucleotide analgoues in Chinese hamster ovary cells

Cyclic nucleotide analogues have been tested for their ability to cause the morphological conversion of Chinese hamster ovary cells in culture, as well as for effects on cyclic AMP-related enzymes. The ability of the analogues to inhibit the cyclic AMP phosphodiesterase activity and to activate the cyclic AMP-dependent protein kinase activity in cell extracts has been measured. Cell cultures were incubated with the analogues and the effects on morphology, intracellular level of cyclic AMP, and in vivo protein kinase activation were determined. All analogues which induced the morphological conversion also caused in vivo activation of the cyclic AMP-dependent protein kinase. Only N⁶,O^2′-dibutryl and N⁶-monobutyryl cyclic AMp caused caused on increase in intracellular cyclic AMP, presumably through inhibition of the intracellular cyclic AMP phosphodiesterase activity. The increase in cyclic AMP appears to cause the protein kinase activation. However, analogues such as 8-bromo and 8-benzylthio cyclic AMP do not cause any change in intracellular cyclic AMP level and appear to activate the intracellular cyclic AMP-dependent protein kinase directly.     

Characterization of protein kinases from adrenal medulla

Since phosphorylation of chromosomal proteins by cyclic AMP-dependent protein kinases (EC 2.7.1.37) enhances template activity of adrenal medulla chromatin (9), we have studied the properties and regulation of protein kinases isolated from chromaffin cell cytosol and nuclei. DEAE-cellulose chromatography revealed three peaks of kinase activity in the nucleus (nPKI, nPKII, nPKIII) and two in the cytosol (cPKI, cPKII). The three nuclear enzymes, as well as cPKII, did not require cyclic AMP to express their catalytic activity, nPKI and nPKIII preferred acidic substrates as PO ₄ ^3– acceptors, while nPKII and the cytosol enzymes preferred basic PO ₄ ^3– acceptors. Enzyme recombination experiments using protein kinase regulatory subunits from cytosol suggested that cPKII was the catalytic subunit of cPKI. In contrast, the nuclear enzymes were not catalytic subunits of the cyclic AMP-dependent protein kinase in the cytosol (cPKI). Only the cytosol protein kinases could be inhibited by endogenous heat-stable protein kinase inhibitors. The nuclear and cytosol cyclic AMP-independent protein kinases were distinguishable on the basis of their sedimentation constants as well as Mg²⁺ and Mn²⁺ requirements.     

The effect of growth state on the activity of the protein kinase isoenzymes an increase in type II cyclic AMP-dependent protein kinase is correlated with growth arrest in G1 phase

Native polyacrylamide gels have been used to resolve protein kinase isoenzymes from cultured cells and the protein kinases have been identified by carrying out phosphorylation reactions in the gel. Following electrophoresis, the gels were incubated with histome and [γ-³²P]ATP. The gels were then thoroughly washed and dried down, and the protein kinases were located by autoradiography. Protein kinase activity as measured in the gel system was a linear function of cytosol protein concentration up to about 100 μg per channel and incorporation of ³²P into histone was time dependent. Three bands of protein kinase activity were resolved in cytosol samples from baby hamster kidney (BHK) fibroblasts. The band with the lowest relative mobility utilized histone IIA or casein equally well as substrate protein whereas bands 2 and 3 demonstrated a clear preference for histone. Bands 2 and 3 displayed a relative mobility in electrophoresis that was identical to that observed for cyclic AMP-dependent protein kinases I and II from rat liver. Treatment of cyctosol samples with cyclic AMP prior to electrophoresis resulted in the disappearance of cyclic AMP-dependent protein kinases from the gel profile. This method was employed to identify bands 2 and 3 as cyclic AMP-dependent protein kinases. The protein kinases in growth-arrested cells were compared with proliferating cells. We have observed a 3.5-fold increase in the activity of Type II protein kinase as the cells arrest growth in G₁ phase of the cell cycle. This increase in Type II is correlated with the increase in cells blocked in G₁ and a decrease in II Type activity appears to be an early event in permitting cells to leave G₁ and resume growth.     

Effects of adenosine 3′ : 5′-monophosphate and guanosine 3′ : 5′-monophosphate on calcium uptake and phosphorylation in membrane fractions of vascular smooth muscle

The effects of adenosine 3′ : 5′-monophosphate (cyclic AMP), guanosine 3′ : 5′-monophosphate (cyclic GMP) and exogenous protein kinase on Ca uptake and membrane phosphorylation were studied in subcellular fractions of vascular smooth muscle from rabbit aorta. Two functionally distinct fractions were separated on a continuous sucrose gradient: a light fraction enriched in endoplasmic reticulum (fraction E) and a heavier fraction containing mainly plasma membranes (fraction P).While cyclic AMP and cyclic GMP had no effect on Ca uptake in the absence of oxalate, both cyclic nucleotides inhibited the rate of oxalate-activated Ca uptake when used at concentrations higher than 10^?5 M. The addition of bovine heart protein kinase to either fraction produced an increase in the rate of oxalate-activated Ca uptake which was further augmented by cyclic AMP. Cyclic GMP caused smaller stimulations of protein kinase-catalyzed Ca uptake than cyclic AMP.Mg-dependent phosphorylation, attributable to endogenous protein kinase(s), was inhibited in fraction E by low concentrations (10^?8 M) of both cyclic AMP and cyclic GMP. In fraction P, an inhibition by cyclic AMP occurred also at a concentration of 10^?8 M, while with cyclic AMP a concentration of 10^?5 M was required for a similar inhibition. Bovine heart protein kinase stimulated the phosphorylation of the membrane fractions much more than Ca uptake. In fraction E, in the presence of bovine protein kinase, both cyclic AMP and cyclic GMP stimulated phosphorylation up to 200%. Under these conditions, no stimulation was observed in fraction P.These results are compatible with the hypothesis that in vascular smooth muscle soluble rather than particulate protein kinases are involved in the regulation of intracellular Ca concentration.     

Activation of protein kinase(s) by glucagon and cyclic AMP in the rat liver: Relationship to metabolic effects

Although it is known that protein kinases are activated by cyclic AMP, the role of the activated kinase in the gluconeogenic response to cyclic AMP is not known. Therefore, we examined whether the inhibition of the gluconeogenic resposne in the liver is due to an interference with the activation of protein kinase in the following situations: (1) adrenalectomy, (2) Na⁺-free perfustae, (3) administration of local anesthetic. We measured protein kinase activity indirectly by measureing incorporation of ³²P into proteins of the perfused liver, and directly by measuring the enzyme activity. We found no significant inhibition of activation of protein kinase in teh above experimental conditions. It seems that in the intact liver, activation of protein kinase by itself is not sufficient to evoke metabolic responses.In order to clarify whether teh requirement for ion redistribution is specific for the gluconeogenic response or not, the lipolytic and antilipogenic effects of glucagon and cyclic AMP were examined. Na⁺-free persurfate, local anesta high K⁺ did interfere with the lipolytic and antilipogenic responses to these agents just as it interfered with the fluconeogenic response. It is likely that ion redistribution evoked by glucagon and cyclic AMP is essential to the expression of most, if not all, metabolic effects.     

Effects of Ca2+, calmodulin and cyclic AMP on the phosphorylation of endogenous proteins by homogenates of rat islets of Langerhans

Activation of Ca²⁺-calmodulin- and cyclic AMP-dependent protein kinases has been suggested to be involved in stimulus-secretion coupling in the pancreatic β-cell. To study the properties of such kinases and their endogenous protein substrates homogenates of rat islets of Langerhans were incubated with [γ-³²P]ATP. Phosphorylated proteins were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis and detected by autoradiography. The phosphorylation of certain proteins could be enhanced by Ca²⁺ plus calmodulin or by cyclic AMP. The major effect of Ca²⁺ and calmodulin was to stimulate the phosphorylation of a protein (P53) of molecular weight 53 100±500 (n = 15). Maximum phosphorylation of protein P53 occurred within 2 min with 2 μM free Ca²⁺ and 0.7 μM calmodulin. Incorporation of label into protein P53 was inhibited by trifluoperazine or W7 but not by cyclic AMP-dependent protein kinase inhibitor. Phosphorylation of a protein of similar molecular weight could be enhanced to a lesser extent in the absence of Ca²⁺ but in the presence of cyclic AMP and 3-isobutylmethylxanthine: this phosphorylation was blocked by cyclic AMP-dependent protein kinase inhibitor. Cyclic AMP also stimulated incorporation of label into polypeptides of molecular weights 55 000 and 70–80 000. The results are consistent with the hypothesis that protein phosphorylation mechanisms may play a role in the regulation of insulin secretion.     

Binding proteins for adenosine 3′:5′-cyclic monophosphate in bovine adrenal cortex

Five peaks of cyclic AMP-binding activity could be resolved by DEAE-cellulose chromatography of bovine adrenal-cortex cytosol. Two of the binding peaks co-chromatographed with the catalytic activities of cyclic AMP-dependent protein kinases (ATP-protein phosphotransferase, EC 2.7.1.37) of type I or type II respectively. A third binding protein was eluted between the two kinases, and appeared to be the free regulatory moiety of protein kinase I. Two of the binding proteins for cyclic AMP, sedimenting at 9S in sucrose gradients, could also bind adenosine. They bound cyclic AMP with an apparent equilibrium dissociation constant (K(d)) of about 0.1mum, and showed an increased binding capacity for cyclic AMP after preincubation in the presence of K(+), Mg(2+) and ATP. The two binding proteins differed in their apparent affinities for adenosine. The isolated regulatory moiety of protein kinase I had a very high affinity for cyclic AMP (K(d)<0.1nm). At low ionic strength or in the presence of MgATP, the high-affinity binding of cyclic AMP to the regulatory subunit of protein kinase I was decreased by the catalytic subunit. At high ionic strength and in the absence of MgATP the high-affinity binding to the regulatory subunit was not affected by the presence of catalytic subunit. Under all experimental conditions tested, dissociation of protein kinase I was accompanied by an increased affinity for cyclic AMP. To gain some insight into the mechanism by which cyclic AMP activates protein kinase, the interaction between basic proteins, salt and the cyclic nucleotide in activating the kinase was studied.     

Immunofluorescent localization of cyclic nucleotide-dependent protein kinases on the mitotic apparatus of cultured cells

Cyclic nucleotides and cyclic nucleotide-dependent protein kinases have been implicated in the regulation of cell motility and division, processes that depend on the cell cytoskeleton. To determine whether cyclic nucleotides or their kinases are physically associated with the cytoskeleton during cell division, fluorescently labeled antibodies directed against cyclic AMP, cyclic GMP, and the cyclic nucleotide- dpendent protein kinases were used to localize these molecules in mitotic PtK1 cells. Both the cyclic GMP-dependent protein kinase and the type II regulatory subunit of the cyclic AMP-dependent protein kinase were localized on the mitotic spindle. Throughout mitosis, their distribution closely resembled that of tubulin. Antibodies to cyclic AMP, cyclic GMP, and the type I regulatory and catalytic subunits of the cyclic AMP-dependent protein kinase did not label the mitotic apparatus. The association between specific components of the cyclic neucleotide system and the mitotic spindle suggests that cyclic nucleotide-dependent phosphorylation of spindle proteins, such as those of microtubules, may play a fundamental role in the regulation of spindle assembly and chromosome motion.     

Cyclic AMP-dependent protein kinase from Ustilago maydis

Summary Ustilago maydis was surveyed for cyclic AMP-dependent protein kinase activity. Using a combination of ion-exchange and molecular filtration techniques, we demonstrate that there is only one form of cyclic AMP-dependent protein kinase in the cytosolic fraction of the fungus. The kinase activity is specifically activated by cyclic AMP and utilizes protamine and kemptide as substrates. Most, if not all, of the cyclic AMP binding detected in the soluble fraction is associated with the protein kinase activity. Cyclic AMP-dependent protein kinase is completely dissociated by cyclic AMP into catalytic and regulatory subunits having an apparent molecular weight of 35 000 daltons as judged by sucrose gradient centrifugation.Post graduate fellow from the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, Argentina).Career investigator from CONICET.     

Cyclic AMP-dependent memory mutants are defective in the food choice behavior of <Emphasis Type="Italic">Drosophila</Emphasis>

Acute choice behavior in ingesting two different concentrations of sucrose in Drosophila is presumed to include learning and memory. Effects on this behavior were examined for four mutations that block associative learning (dunce, rutabaga, amnesiac, and radish). Three of these mutations cause cyclic AMP signaling defects and significantly reduced taste discrimination. The exception was radish, which affects neither. Electrophysiological recordings confirmed that the sensitivity of taste receptors is almost indistinguishable in all flies, whether wild type or mutant. These results suggest that food choice behavior in Drosophila involves central nervous learning and memory operating via cyclic AMP signaling pathways.     

Altered cyclic AMP-dependent protein kinase activity in a mutant adrenocortical tumor cell line

A somatic cell genetic approach has been used to evaluate the role of cyclic AMP-dependent protein kinase in ACTH action on adrenal steroidogenesis. A mutant clone, 8BrcAMP^r-1, previously was isolated from an ACTH-sensitive adrenocortical tumor cell line (clone Y1) following mutagenesis and selective growth in 8-bromoadenosine 3′, 5′-monophosphate. This study demonstrates that the 8BrcAMP⁴-1 cells have an altered cyclic AMP-dependent protein kinase. The protein kinase in the cytosol of the mutant characteristically requires, for half-maximal activity, concentrations of cyclic AMP 7-fold higher than those required by the enzyme in preparations from the parent. The cytosolic cyclic AMP-dependent protein kinases of Y1 and 8BrcAMP^r-1 cells chromatograph similarly on columns of DEAE-cellulose. From each cell line, a major peak of activity (≥ 70% of recovered activity), designated as Peak I, elutes with 0.04–0.06 M NaCl; a second peak of activity, designated as Peak II, elutes with 0.12–0.14 M NaCl. Protein kinase activity in the Peak I fraction of mutant cells has a decreased apparent affinity (4-fold) for cyclic AMP relative to the corresponding fraction of parental Y1 cells. The protein kinase activities present in Peak II fractions from Y1 and mutant cells are indistinguishable. The protein kinase mutant exhibits poor steroidogenic responses to added ACTH and cyclic AMP; and as shown previously does not display the growth arrest and morphological changes produced in Y1 by these agents. These results suggest that cyclic AMP-dependent protein kinase is important in the regulation of adrenal steroidogenesis, morphology and growth by ACTH.     

Purification and characterization fo a phosvitin kinase from the thyroid gland

1. Two cyclic AMP independent protein kinases phosphorylating preferentially acidic substrates have been identified in soluble extract from human, rat and pig thyroid glands/ Both enzymes were retained on DEAE-cellulose. The first enzyme activity eluted between 60 and 100 mM phosphate (depending on the species), phosphorylated both casein and phosvitin and was retained on phosphocellulose; this enzyme likely corresponds to a casein kinase already described in many tissues. The second enzyme activity eluted from DEAE-cellulose at phosphate concentrations higher than 3000 mM, phosphorylated only phosvitin and was not retained on phophocellulose. These enzymes were neither stimulated by cyclic AMP, cyclic GMP and calcium, nor inhbiited by the inhibitor of the cyclic AMP dependent protein kinases. 2. The second enzyme activity was purified from pig thyroid gland by the association of affinity chromatography on insolubilized phosvitin and DEAE-cellulose chromatography. Its specific activity was increased by 8400. 3. The purified enzyme (phosvitin kinase) was analyzed for biochemical and enzymatic properties. Phosvitin kinase phosphorylated phosvitin with an apparent K_m of 100 μg/ml; casein, histone, protamine and bovine serum albumin were not phosphorylated. The enzyme utilized ATP as well as GTP as phosphate donor with an apparent K_m of 25 and 28 μM, respectively. It had an absolute requirement for Mg²⁺ with a maximal activity at 4 mM and exhibited an optimal activity at pH 7.0. The molecular weight of the native enzyme was 110 000 as determined by Sephacryl S300 gel filtration. The analysis by SDS-polyacrylamide gel electrophoresis revealed a major band with a molecualr weight of 35 000 suggesting a polymeric structure of the enzyme.