Casein kinases and their protein substrates in rat liver cytosol: Evidence for their participation in multimolecular systems

248

Biochimica etBioph~ica Acta 846 (1985) 248 256 Elsevier

BBA 11540

C a s e i n k i n a s e s a n d t h e i r p r o t e i n s u b s t r a t e s in r a t l i v e r c y t o s o l : e v i d e n c e f o r t h e i r p a r t i c i p a t i o n in m u l t i m o l e c u l a r s y s t e m s F l a v i o M e g g i o , P a t r i z i a A g o s t i n i s a n d L o r e n z o A. P i n n a Istituto di Chirnica Biologica, Universitiz di Padova, Via F. Marzolo 3, 35131 Padova (Italy)

(Received February 21st, 1985)

Key words: Casein kinase: Protein kinase; Phosphoprotein; (Rat liver cytosol)

We have shown by gel filtration on Sepharose 4B at low ionic strength that casein kinases S (type 1), heparin-insensitive, and TS (type 2), heparin-inhibited, of rat liver cytosol participate in two distinct multimolecular systems, V e / V o = 1.25 and V e / V , , = 1.90, respectively, both less retarded than the peak of cAMP-dependent protein kinase activity ( V ~ / V o = 2.04). Both casein kinase I and casein kinase II complexes are unstable in 0.5 M NaCI, giving rise by gel filtration under these conditions to the free forms of casein kinase S ( V e / V o = 2.37, M r 34000) and casein kinase T S ( V ~ / Vo = 2.10, M r 130000), respectively. in contrast, the elution volume of cAMP-dependent protein kinase activity is always the same irrespective of the ionic strength of the medium. Casein kinase I, accounting for the whole casein kinase S activity of cytosol, also contains a phosphorylatable 31-kDa protein (p31) which is a substrate of casein kinase S, since its phosphorylation is insensitive to heparin, the heat-stable inhibitor and trifluoperazine, but it is prevented by beryllium. Casein kinase II, on the other hand, apparently results from the association of the whole casein kinase TS (type 2) of rat liver cytosol with a 90-kDa protein substrate (p90) which is distinct from glycogen synthase according to their different peptide mappings. The radiolabelling of pg0 is inhibited by heparin, unlabeled G T P and polyglutamates, while it is dramatically and specifically enhanced by polylysine. At least three more protein bands of M r 58000, 52000 and 37000 are phosphorylated by casein kinase T S in the casein kinase II fraction: their co-elution with casein kinase TS, however, seems to be accidental and their radiolabeling in the presence of polylysine is almost neglegible compared to that of p90. It is concluded that p31 and p90 may represent specific targets of casein kinase S and casein kinase TS, respectively, whose intimate association with the enzymes could be functionally significant.

Introduction

Casein kinases are ubiquitous protein kinases insensitive to either cyclic nucleotides or Ca 2+ and phospholipids, and which are operationally defined by their preference for casein and phosvitin, over histones, as artificial substrates (reviewed in Refs. 1 and 2). According to their properties casein kinases can be grouped into two main families: type 1 casein kinases (termed also NI, when isoAbbreviation: PMSF, pheny|methylsulfonyl fluoride.

lated from nuclei, or S or A) phosphorylate almost exclusively serine residues of casein and use only ATP as phosphate donor, whereas type 2 casein kinases (termed also NII, TS or G) predominantly affect threonine residues of casein and use both G T P and ATP as phosphate donor. These two classes of casein kinases also sharply differ in their molecular weight and quaternary structure: while type 1 enzymes exhibit relatively low M r (between 22000 and 50000, but most frequently around 35 000) and have been described as monomeric [2] or in very few cases dimeric [3]

0167-4889/85/$03.30 ~;' 1985 Elsevier Science Publishers B.V. (Biomedical Division)

249 proteins, the type 2 casein kinases exhibit higher M r (120000-200000) and an oligomeric structure often resulting from the association of a and /3 subunits to give a2/32 tetramers [2,4]. This information has been derived entirely from the comparison of the subunit(s) Mf calculated by SDS-polyacrylamide gel electrophoresis with the molecular weights of the holoenzymes determined by gel filtration a n d / o r ultracentrifugation under non-denaturing conditions. It should be noted, however, that the media employed for non-denaturating gel filtrations and ultracentrifugations always included high concentrations of either NaC1 or KC1 (0.4-1.0 M), conferring upon them remarkable dissociating capacity if compared with the intracellular conditions. In the absence of salts it has actually been reported that some casein kinases tend to form large aggregates exhibiting very high molecular weights [5,6]. On the other hand, it has been shown that sometimes protein kinases become intimately associated with their substrates to form multienzymatic complexes [7,8]. It is possible therefore that the large aggregates with casein kinase activity are not mere artifacts but may represent functional complexes whose composition could shed some light on the as yet unclear physiological role of these enzymes. In view of this it seemed interesting to re-evaluate under non-dissociating conditions the apparent M r of the type 1 and type 2 casein kinases which are present in rat liver cytosol (termed casein kinase S and casein kinase TS, respectively) and to start a study of the proteins associated with their catalytic moieties, as they might represent potential substrates a n d / o r modulators of these enzymes. Materials and Methods Preparation and fractionation of rat liver cytosol. Rat livers were chopped with scissors and homogenized with about 3 vol. of ice-cold, low-ionicstrength medium containing 0.5 M sucrose, 0.1 M Tris-HC1 (pH 7.0), 5 mM EDTA, 50 /~M PMSF and 60 #g/1 leupeptin, using three strokes of a Potter-Elvehjem homogeniser. Nuclei and unbroken cells were discarded by 15 min centrifugation at 700 x g. All the remaining particulate materials were removed by 90 min centrifugation at 105 000 x g. The clear supernatant was dialysed

for 5 h against a large excess of 25 mM NaH2PO 4 (pH 7.0) containing 15 mM 2-mercaptoethanol, 2 mM EDTA, 0.01% Brij 35 and 0.05 mM PMSF (buffer A) and the proteins were concentrated about 5-fold either by ultrafiltration or by precipitation with 70% saturated ammonium sulphate. In this latter case the precipitate was dissolved in a small volume of buffer A and dialysed overnight against the same buffer. The concentrated cytosol was fractionated by gel chromatography at either low or high ionic strength. Low-ionic-strength gel filtration was accomplished through a 106 × 1.95 cm Sepharose 4B column equilibrated and operated with buffer A. The flow rate was about 15 m l / h and 3.5-ml fractions were collected. The single fractions were monitored for absorbance at 280 and 260 nm and for protein kinase activity tested toward either casein or histones, in the presence and absence of specific effectors such as heparin and cAMP. High-ionic-strength gel filtration was performed through the same column equilibrated and operated with buffer A including also 0.5 M NaC1 (buffer B). Prior to application on the column the concentrated cytosol was also dialysed overnight against buffer B. In some experiments ionic strength was first increased by dialysis against buffer B and than decreased by a second dialysis against buffer A, in order to remove NaC1; the following gel filtration was also performed in buffer A. The pattern in this case was identical to the one obtained without the dialysis against buffer B. Protein kinase assays. Casein kinase activity was determined essentially as previously described [9], using whole casein as protein substrate, cAMP-dependent protein kinase was tested toward histones type II AS (Sigma) in the presence of 1/~M cAMP [10]. The values of 32p incorporation into histones observed in the absence of cAMP were subtracted to give the activity dependent on cAMP. Purification of casein kinases. Highly purified preparations of casein kinase S (type 1) and casein kinase TS (type 2) were obtained as previously described [9]. Their M r values determined by gel filtration either in the presence or absence of 0.5 M NaC1 were 34000 and 130000, respectively. M r determinations. The apparent M r values of purified casein kinases and of casein kinase-containing fractions were determined by gel filtration

250

through either Sepharose 4B or Ultrogel AcA34 equilibrated with either buffer A (low ionic strength) or buffer B (high ionic strength). The columns were calibrated with the following proteins: ferritin ( M r 410000), glucose-6-phosphate dehydrogenase (104000), lactate dehydrogenase (140000), bovine serum albumin (67000), pepsin (36 000).

4B at low ionic strength the casein kinase activity is resolved into two peaks, both less retarded than marker lactate dehydrogenase (Fig. 1A). The first peak (casein kinase I) is insensitive to heparin and overlaps quite symmetrically a peak of ultraviolet absorbance whose A260 is greater than its A=8(,.

.DH k

Detection of phosphorylatable protein substrates. Small aliquots (25-50 /.tl) of the fractions eluted from Sepharose 4B were incubated for 10 min at 37°C in 0.1 ml of medium containing 50 mM Tris-HC1 (pH 7.5), 12 mM MgC12, 10 /~M [y32p]ATP with a specific radioactivity of about 3000 cpm/pmol and 50 mM NaCI. In some experiments other effectors were also added as detailed in the figures. The reaction was stopped by addition of 2% SDS, 1% 2-mercaptoethanol, 2 mM EDTA (final concentrations) and the samples were subjected to gel electrophoresis on vertical slabs of 11% polyacrylamide in the presence of 0.1% SDS, essentially according to Laemmli [11]. The gels were stained with Coomassie brilliant blue, dried and autoradiographed as previously described [9]. The apparent M r values of the prominent radiolabeled bands were calculated by calibration with the following marker proteins: glycogen phosphorylase (subunit M r 93000); bovine serum albumin (67000); ovalbumin (45 000); glyceraldehyde phosphate dehydrogenase (36000); trypsinogen (25000) and ribonuclease

LK-II

c

T L g • lO

:

Results

Gel chromatographic behauiour of casein kinases: effects of ionic strength and purification When the clear rat liver post-microsomai supernatant is subjected to gel filtration on Sepharose

['~I

/

~ /_. i . ,L

//~", ',. -

.:!

05

:

o

:,t)~1

\k ". I

,,

i

|

',~

/

,),

',

.,~

"

/

'i>, E LDH v

L '~

6

(I 3 700). Peptide mapping. The radiolabeled bands were excised from the dried gels, eluted overnight with 50 mM Tris-HC1 (pH 7.5) containing 1% SDS and digested with CNBr in 70% formic acid prior to 18% polyacrylamide gel electrophoresis [12]. Uhracentrifugation. Sucrose gradient ultracentrifugation of the casein kinase I fraction from Sepharose 4B in either the presence or the absence of 0.5 M NaC1 was performed according to Hosey and Tao [6].

CK-I

CK-t

tl/ ,/ )

Vo

IlL

,o

I

al

B

II

t

I o , \ I I v/ ',

L 05

I

/ V I~

~2

~

I

20

40

60

80 fraction number

lO0

Fig. l. Resolution of rat liver cytosol protein kinase activities by Sepharose 4B gel filtration at low (uppere figure) and high (lower figure) ionic strength. Casein kinase activity in the absence (e) and presence ( O ) of heparin (1 /~g/ml). The histogram denotes cAMP-dependent protein kinase activity tested on histones. The dotted and broken lines indicate absorbance at 260 and 280 nm, respectively. L D H denotes the elution volume of lactate dehydrogenase ( M r 140000).

251

The more retarded peak (casein kinase II), in contrast, is almost completely inhibited by heparin and is eluted ahead of endogenous cAMP-dependent histone kinase. A precise determination of M r of casein kinases I and II was hindered by the failure to obtain linear calibration curves on Sepharose 4B. Their a p p r o x i m a t e values are (1500-2500).103 and (400-800)-103 for casein kinase I and casein kinase II, respectively. Essentially the same profile is obtained whether or not rat liver cytosol is precipitated with 70% ammonium sulphate or is exposed to high salt concentrations prior to dialysis and gel filtration at low ionic strength (see experimental section.) If the gel filtration takes place at ionic strength close to the physiological (0.025 M sodium phosphate plus 0.1 M NaC1) the elution volume of heparin-inhibited casein kinase II is unchanged; heparin-insensitive casein kinase I, though somewhat more retarded than it was in the absence of salts (Table I), is still eluted before casein kinase II, and it is again symmetrically overlapped by an ultraviolet peak whose absorbance at 260 nm largely exceeds that at 280 nm (not shown). If, however, the gel filtration is performed in the presence of 0.5 M NaC1 the resulting profile of casein kinase activity is deeply modified (Fig. 1B): both peaks are now eluted after both lactate dehydrogenase and cAMP-dependent kinase; moreover

the heparin-insensitive peak is more retarded than the heparin sensitive one. Their M r values evaluated by gel filtration through Ultrogel AcA-34 were 34000 and 130000, respectively, corresponding to those of purified casein kinase S and casein kinase TS. The same result was obtained if casein kinase I and casein kinase II from Sepharose 4B at low ionic strength were directly submitted to Ultrogel AcA-34 gel filtration in the presence of 0.5 M NaC1. It can be concluded that both the casein kinase fractions detectable at low ionic strength are dissociated by 0.5 M NaCI, originating smaller components whose sizes correspond to the M r of purified casein kinases S (type 1) and TS (type 2). Such a behaviour, as well as the different sensitivity to heparin, strongly suggested that the large fraction I from Sepharose 4B contains the type 1 casein kinase S, while the more retarded fraction II is accounted for by type 2 casein kinase TS. Such a conclusion is fully confirmed by the additional findings that peak casein kinase I phosphorylates only serine residues of B-casein and can use only ATP as phosphate donor, whereas casein kinase II phosphorylates threonine residues of fl-casein and can use also GTP as phosphate donor (not shown). The different gel chromatographic behaviours at low and high ionic strength would indicate that both cytosolic casein kinases S and TS participate

TABLE I E L U T I O N V O L U M E S OF PROTEIN K I N A S E ACTIVITIES A N D OF T H E P H O S P H O R Y L A T A B L E PROTEINp90 UPON S E P H A R O S E 4B G E L F I L T R A T I O N A T LOW A N D H I G H IONIC S T R E N G T H The crude fractions casein kinases I and It and cAMP-dependent histone kinase from Sepharose 4B (Fig. 1A) were resubmitted to gel filtration through the same column equilibrated with either buffer A (low ionic strength) or buffer B (including also 0.5 M NaC1) or buffer A including 0.1 M NaCI ('physiological' ionic strength). Casein kinases S and TS purified to near homogeneity were also subjected to gel filtration at low and high ionic strength after having been dialyzed overnight against buffer A or buffer B, respectively. The elution volumes of p90 have been calculated from an experiment like that described in Fig. 4. CK, casein kinase; HK, histone kinase. V~/V o

CK-I (heparin-insensitive) CK-II (heparin-inhibited) H K (A-kinase) Purified Ck-S (heparin-insensitive) Purified Ck-TS (heparin-inhibited) p90

values

low ionic strength

high ionic strength

physiological ionic strength

1.25 1.90 2.00 2.35 2.11 1.90

2.37 2.10 2.04 2.37 2.10 2.01

1.45 1.90 -

252

in large complexes which dissociate in 0.5 M NaC1. A priori this might occur either by simple polymerization of enzyme molecules or through their interaction with other proteins. The former hypothesis, however, seems very unlikely considering that once purified to near-homogeneity both casein kinases TS and S exhibit the same V J Vo ratios in either the presence or the absence of NaC1 (Table I). Crude cytosol, in contrast, after treatment with high salt concentrations still gives rise to the high molecular weight complexes provided that salts are removed by dialysis. Moreover, if purified casein kinase TS radiolabeled by autophosphorylation with [y-32p]ATP is added to crude cytosol and subjected to gel filtration in the absence of NaC1 a consistent amount of radiolabeling is eluted together with the large casein kinase II complex rather than in the position of pure casein kinase TS (not shown). It is very likely therefore that the large forms of casein kinases S and TS evidenced by Sepharose 4B gel chromatography in the absence of salt are due to the interaction of these enzymes with other protein molecules.

Phosphorylatable proteins co-eluting with casein kinases The idea that both types of casein kinases in the crude cytosol are associated with phosphorylatable substrates is supported by phosphorylation experiments, showing that the large Sepharose 4B fractions exhibiting protein kinase activity also contain proteins which can be phosphorylated by the co-eluting protein kinases. In particular a phosphorylatable polypeptide of M r 31 000 (p31) can be visualized when the casein kinase I fraction is incubated with [7-32p]ATP • Mg 2+ (Fig. 2). Such a phosphorylation is neither inhibited by trifluoperazine (not shown), heparin, polyglutamate or unlabeled GTP, nor stimulated by cAMP, while it is abolished by 0.1 mM BeSO4, which has been reported [13] to act as a specific inhibitor of type 1 casein kinases. Consequently the phosphorylation of the 31-kDa protein can be promoted also by purified casein kinase S after the endogenous kinase(s) have been inactivated by heating the casein kinase I fraction at 55 ° for 5 min. It is very likely, therefore, that p31 represents a physiological target of casein kinase S, associ-

E

.-=

3 ÷

68K

4-

4-

+

+

--

36K gmO I I m ~ ( I I I D mNw I~,,m

~p31

~

~.

25 K

1

2

3

4

5

6

7

8 v

CK- I

CK-I 5rain, at 55tC

Fig. 2. Detection of phosphorylatable proteins in the heparin insensitive casein kinase I fraction casein from Sepharose 4B. Comparable aliquots of the casein kinase I fraction (CK-I in Fig. IA) were analyzed for the presence of phosphorylatable proteins by incubation with [y-32 P]ATP and different effectors, as indicated, followed by polyacrylamide gel electrophoresis in the presence of SDS. Samples 7 and 8 were pre-heated at 55°C for 5 min before being incubated either without or with purified casein kinase S. The autoradiographies are shown. The 34 kDa band (denotes Ck-S) in lane 8 is due to the autophosphorylation of casein kinase S.

ated with the enzyme and probably with other proteins into a very large cytosolic complex. The casein kinase II fraction from Sepharose 4B, accounting for type 2 casein kinase activity, on the other hand, undergoes phosphorylation at several protein bands (Fig. 3). Most of these radiolabeled bands are abolished by either heparin or polyglutamates or unlabeled GTP, indicating that their phosphorylation is actually catalyzed by endogenous type 2 casein kinase TS. In contrast, some bands are enhanced by the addition of polylysine, which stimulates casein kinase TS [4]. This polycation exhibits an especially remarkable effect on a 90 kDa band (p90), which becomes predominant over the other ones. Spermine, in contrast, has no evident effect under our conditions employing 12 mM Mg :+ (not shown). An incontrovertible demonstration that the heparin-sensitive and polylysine-stimulated radiolabeled bands are actually phosphorytated by endogenous casein kinase TS was provided by

253 62

=

.~

'

65

67

69

71

73

75

77

79

83

~'

g

=

-NaCI

8 o

+

+

+

+

+

g

,5

"g

®

4-

4 c

+NaCI

~ p g O

93 K - -

~1--I

68 K - -

L

45K - 36K - -

p90~

- NaCI.

..... m

25K - -

® p90~ 1

2

3

4

CK- II

5

6

7

+NaCt

8

OR-If

P-e,tt tr,ated Fig. 3. Detection of phosphorylatable protein in the heparin-inhibited casein kinase II fraction casein from Sepharose 4B. Comparable aliquots of the fraction casein kinase II (CK-II in Fig. 1A) were analyzed for the presence of phosphorylatable proteins by incubation with [y-32P]ATP and different effectors as indicated, followed by polyacrylamide gel electrophoresis in the presence of SDS. Samples 7 and 8 were obtained by treating casein kinase II with phosphocellulose in order to remove the endogenous casein kinase prior to incubation with [•-32P]ATP in either the absence or the presence of purified casein kinase TS. The autoradiographs are shown, fl Denotes the autophosphorylatable fl-subunit of casein kinase TS. The lines on the left indicate the position of marker proteins.

their d i s a p p e a r a n c e u p o n t r e a t m e n t of casein k i n a s e II with p h o s p h o c e l l u l o s e , which b i n d s casein k i n a s e TS very efficiently, a n d b y their s u b s e q u e n t r e - a p p e a r a n c e after a d d i t i o n of purified casein k i n a s e TS (Fig. 3, lanes 7 a n d 8).

Specific association of casein kinases S and TS with the 31 kDa and 90 kDa substrate proteins T h e S e p h a r o s e 4B casein kinase activity profile at low ionic strength is n o t altered b y p r e i n c u b a tion with either 0.5% d e o x y c h o l a t e or a - a m y l a s e ( n o t shown), thus ruling out the suggestion that m e m b r a n e f r a g m e n t s a n d / o r glycogen particles c o u l d a c c o u n t for the h i g h - M r c o m p l e x e s exhibiting casein k i n a s e activity. O n the other hand, the co-elution of casein kinases S a n d TS with p31 a n d p90, respectively, seems to be n o t a m e r e coincidence, b u t r a t h e r the c o n s e q u e n c e of their inclu-

p90~

"~ ~ l ~

-NaCL

© p90 ~

+NaCL

62

65

67

69

71

73

75

fraction

77

79

83

number

Fig. 4. Elution behaviours of casein kinase TS and its protein substrate p90 at low and high ionic strength. Two aliquots of the same cytosolic preparation were subjected to Sepharose 4B gel filtration in either the absence or the presence of 0.5 M NaCI (see methods and Fig. 1). The fractions indicated by numbers were analyzed for: (A) casein kinase activity; (B) presence of the protein band with M r =90000 (by polyacrylamide gel electrophoresis in SDS and Coomassie staining); (C) intrinsic phosphorylation of p90 (incubation with [~,32p]ATP, followed by polyacrylamide gel electrophoresis in SDS and autoradiography). Only the segments of the gels including the 90 kDa protein (donated by arrows) are shown.

sion into definite s u p r a m o l e c u l a r aggregates. A c t u a l l y casein kinase S a n d p31 c o - p u r i f y also u p o n Sepharose 2B gel c h r o m a t o g r a p h y a n d sucrose g r a d i e n t ultracentrifugation, p r o v i d e d that the ionic strength is k e p t low (not shown). O n the o t h e r hand, the 90 k D a s u b s t r a t e of casein kinase TS, which is also evident as a C o o m a s s i e - s t a i n a b l e b a n d , exactly co-elutes with casein kinase TS from S e p h a r o s e 4B o p e r a t e d at low ionic strength (Fig. 4), whereas the o t h e r phosp h o r y l a t a b l e p r o t e i n s p r e d o m i n a t e along the de-

254

G S. p~90

unit M r, is not consistent with the size of the radiolabeled CNBr fragment originating from it (Fig. 5). Its M r in fact is much higher than that of the CNBr fragment (CB-2) including the sites phosphorylated by casein kinase 2 in both skeletal muscle and liver glycogen synthases [14,15]. Discussion

~K ~K,..

,,. /

Fig. 5. 32p-labelled peptide mappings of p90 and glycogen synthase phosphorylated by casein kinase TS. The endogenous protein p90, radiolabeled by casein kinase TS as indicated in Fig. 3, lane 8, and excised from the gel, and skeletal muscle glycogen synthase subjected to the same procedure after phosphorylation by casein kinase TS were digested with CNBr and the radiolabeled fragments were resolved by 18% polyacrylamide gel electrophoresis in SDS and revealed by autoradiography. The apparent M r values of the radiolabeled peptides were evaluated by calibration as in Ref. 12.

scending slope of the casein kinase TS peak of activity. If, however, the elution is performed in the presence of 0.5 M NaC1 the V~ values of both casein kinase TS and the 90 kDa protein band are significantly increased (Table I). A comparison between the pictures of Fig. 4 obtained at low and high ionic strength also shows that the phosphorylation of p90 by [7-32p]ATP occurs only if the gel chromatography is performed in the absence of NaC1, in order to allow the coelution of p90 and casein kinase TS. In the presence of NaCI p90 is no longer detectable as a radiolabeled band in any Sepharose 4B fraction, indicatiang that casein kinase TS has been completely separated from it. The identification of the 90 kDa protein with glycogen synthase, suggested by the similar sub-

The experiments described in this paper show that, at low ionic strength, the whole casein kinase activity of rat liver cytosol is associated with large complexes whose apparent M r values are much higher than the actual molecular weights of either casein kinase S (type 1) and casein kinase TS (type 2). The larger complex, casein kinase I, eluted from Sepharose 4B with a V J V ratio of 1.2, accounts for the whole heparin-insensitive casein kinase S (type 1), while heparin-inhibited and polylysine-stimulated casein kinase TS (type 2) is fully associated with a more retarded fraction (casein kinase II) having a V J V o ratio of 1.8. Both complexes are unstable in 0.5 M NaC1, giving rise to the 'free' forms of casein kinases exhibiting the expected M r of 34000 and 130000 for casein kinases S (type 1) and TS (type 2), respectively. Such a dissociation is reversible if salt is removed by dialysis from the crude cytosolic fraction; however, once purified, both casein kinase S and casein kinase TS always display their regular molecular weights (34000 and 130000, respectively), irrespective of the ionic strength of the medium. This would rule out any polymerization of the free forms of casein kinases and support the concept that the large aggregates isolated at low ionic strength arise rather from the interaction of casein kinase S and TS with different proteins. Interactions between casein kinases and their protein substrates have already been described [7,8,16]. Also in our case the large complexes, casein kinase I and casein kinase II, have been shown to contain phosphorylatable proteins which are specifically affected by the respective casein kinases. In particular the less retarded fraction casein kinase I, accounting for the whole casein kinase S activity, includes a 31-kDa protein (p31) which is phosphorylated by casein kinase S itself. The very large size of the casein kinase I complex, if c o r n -

255 pared with the low M r of both casein kinase S and p31, and its unusually high 260 n m / 2 8 0 nm absorbance ratio suggest that this fraction is composed of other constituents, besides casein kinase S and p31, including also polynucleotides. It should also be noted that the peaks of casein kinase S activity and ultraviolet absorbances at 260 and 280 nm co-elute quite symmetrically (see Fig. 1A) and they also comigrate upon sucrose gradient ultracentrifugation (not shown), suggesting that casein kinase S is actually part of a homogeneous multimolecular system where it may play a specific role. The apparent M r of casein kinase I (ranging between 1500.103 and 2500.103 ) is of the same order of that of the high molecular mass aminoacyl-tRNA synthetase complexes of eukaryotes [17], which are known .to include also ribonucleic acids [18] and to undergo regulation by phosphorylation-dephosphorylation [19]. The possibility that casein kinase S might participate in such a complex should be casefully considered. On the other hand, the more retarded casein kinase II fraction, accounting for the whole casein kinase TS (type 2) activity of rat liver cytosol seemingly results from the association of casein kinase TS with a co-eluting phosphorylatable protein whose subunit M r is 90000 (p90). It should be noted on this matter that other proteins which can be phosphorylated by casein kinase TS are also present in the casein kinase II fraction, with M r 58000, 52000, 37000 and 24000. Their elution profiles however, evidenced by SDS-polyacrylamide gel electrophoresis of single fractions from Sepharose 4B, only partially overlap the casein kinase II peak, suggesting that their co-elution with casein kinase TS might be not as physiologically meaningful as that of p90. Although the subunit M r of this protein is similar to that of glycogen synthase, a substrate of type 2 casein kinase [14,15], such an identification was definitely ruled out by their quite different 32p-labeled peptide mappings. It should be concluded, therefore, that at least two forms of bound casein kinase TS are present in rat liver cytoplasm: one associated with glycogen particles [12], where it phosphorylates glycogen synthase, while the other one, described here, is a soluble complex with p90 and maybe with other substrates. The identification of p90 would be very helpful for the under-

standing of the physiological role of cytosolic rat liver casein kinase TS. Among the known targets of type 2 casein kinases, including, besides glycogen synthase, also troponin T, acetyl-CoA carboxylase, RII, calsequestrin, ornithine decarboxylase and inhibitor 2, only the androgen receptor exhibits a subunit M r (87000) similar to that of p90, but its presence in liver would be unexpected and, moreover, its phosphorylation seems to be mediated only by nuclear and not by cytosolic casein kinase 2 [20]. A possible identification of p90 with M 1, a 91-kDa microsomal protein readily phosphorylated by casein kinase TS [12], is ruled out by the very different sizes of their radiolabeled CNBr digestion products (Fig. 5 and Ref. 12). Irrespective of the actual nature of the proteins associated with rat liver casein kinases S and TS, there is anyway compelling evidence that the cytosolic forms of both these enzymes are not free but participate in two distinct multimolecular system. In this respect they behave differently from the cAMP-dependent protein kinase which exhibits in both high- and low-ionic-strength media the same elution volume, consistent with the free form of the holoenzyme. Such a difference may reflect distinct commitments: while cAMP-dependent protein kinase could be simultaneously involved in several processes, the two "independent" casein kinases are sequestered into functional compartments which may restrict their competence toward just one or a few of their many potential targets.

Acknowledgements This work was supported by grants from Italian Consiglio Nazionale delle Ricerche (83.00454.04) and Ministero della Pubblica Istruzione. The skilled technical assistance of Mr. G. Tasinato and the excellent secretarial aid of Miss. M. Vettore are gratefully acknowledged.

References 1 Pinna, L.A., Meggio, F. and Donella-Deana, A. (1980) in Protein Phosphorylation and Bioregulation (Thomas, G., Podest~, E.J. and Gordon, J., eds.), pp. 8-16, Karger, Basel 2 Hathaway, G.M. and Traugh, J.A. (1982) Curr. Top. Cell. Regul. 21, 101-127

256 3 Thornburg, W., O'Malley, A.F. and Lindell, T.J. (1978) J. Biol. Chem. 253, 4638-4641 4 Meggio, F., Brunati, A.M. and Pinna, L.A. (1983) FEBS Lett. 160, 203-208 5 Baggio, B. and Moret, V. (1971) Biochem. Biophys. Acta 250, 346-350 6 Hosey, M.M. and Tao, M. (1977) Biochim. Biophys. Acta 482, 348-357 7 Rose, K.M., Stetler, D.A. and Jacob, S.T. (1981) Proc. Natl. Acad. Sci USA 78, 2833-2837 8 Atmar, K.J. and Kuehn, G.D. (1981) Proc. Natl. Acad. Sci. USA 78, 5518-5522 9 Meggio, F., Donella-Deana, A. and Pinna, L.A. (1981) J. Biol. Chem. 256, 11958-11961 10 Meggio, F., Chessa, G., Borin, G., Pinna, L.A. and Marchiori, F. (1981) Biochim. Biophys. Acta 662, 94-101 11 Laemmli, U.K. (1970) Nature 227, 680-685 12 Meggio, F., Brunati, A.M., Donella-Deana, A. and Pinna, L.A. (1984) Eur. J. Biochem. 138, 379-385

13 Cummings, B., Kaser, M.R., Wiggins, G., Ord, M.G. and Stocken, L.A. (1982) Biochem. J. 208, 141-146 14 Picton, C., Aitken, A., Bilham, T. and Cohen, P. (1982) Eur. J. Biochem. 124, 37-45 15 Imazu, M., Strickland, W.G., Chrisman, T.D. and Exton, J.H. (1984) J. Biol. Chem. 259, 1813-1821 16 Tao, M., Conway, R. and Cheta, S. (1980) J. Biol. Chem. 255, 2563-2568 17 Dang, C.V., Johnson, D.L. and Yang, D.C.H. (1982) FEBS Lett. 142, 1-6 18 Saxholm, H.J.K. and Pitot, H.C. (1979) Biochim. Biophys. Acta 562, 386 399 19 Damuni, F., Caudwell, F.B. and Cohen, P. (1982) Eur. J. Biochem. 129, 57-65 20 Goueli, S.A., Holtzman, J.L. and Ahmad, K. (1984) Biochem. Biophys. Res. Commun. 123, 778-784