Regulation of protein phosphatase activity by the deinhibitor protein

REGULATION OF PROTEIN PHOSPHATASE ACTIVITY BY T H E DEINHIBITOR PROTEIN J. GORIS, E. WAELKENS, T. CAMPS and W. MERLEVEDE Afdeling Biochemie, Departement Humane Biologie, Faculteit Geneeskunde, Katholieke Universiteit Leuven, B-3000 Belgium

INTRODUCTION

Mammalian tissues have been shown to contain active and inactive forms of a multisubstrate protein phosphatase (protein phosphatase: EC 3.1.3.16) which is responsible for the dephosphorylation amongst others of glycogen synthase b (glycogen synthase: EC 2.4.1.11), phosphorylase a (phosphorylase: EC 2.4.1.1) and the/3-subunit of phosphorylase b kinase (phosphorylase kinase: EC 2.7.1.38). This ATP,Mg-dependent protein phosphatase can be activated by an activating factor (kinase FA) which has been characterized as a bifunctional protein displaying also synthase kinase activity (glycogen synthase kinase: EC 2.7.1.37) independent of cyclic AMP and Ca ions (1-3). The ATP,Mg-dependent protein phosphatase system has been characterized in more detail in skeletal muscle. Besides the inactive phosphatase subunit (Fc), a heat stable modulator protein (M) is involved, so that the [Fc-M]complex is the correct unit for the FA-ATP,Mg-mediated activation. The Fcsubunit by itself can be partially activated by metal ions, such as Mn z+ (4). The major active protein phosphatase present in the rabbit skeletal muscle extract is associated with the glycogen particle and also contains modulator activity. Addition of extra modulator protein causes a time- and concentrationdependent conversion of the phosphatase to the inactive ATP,Mg-dependent enzyme form (5). The exact molecular mechanism of activation is still under investigation: a transient phosphorylation of the modulator protein seems to induce an intramolecular conformational change in the catalytic entity (Fc) of the phosphatase, which is reflected in the reversible activation of the enzyme (1). In mammalian liver glycogen is of minor use for local energy supply, and glycogenolysis will provide glucose to maintain blood levels for the benefit of other tissues in conditions of glucose shortage. Glycogen will accumulate when glucose is abundant, and thus the glycogen content of the normal liver can vary from less than 0.1% to more than 10%. Glycogen accumulation occurs with both phosphorylase and synthase as dephosphorylated enzymes, which requires an operative dephosphorylation during glycogen synthesis. The lack of sensitivity of the glycogen bound protein phosphatase towards AER 2z-i"

467

468

J. GORIS, el al.

inhibitor- 1 and the modulator protein, then called inhibitor-2 according to the nomenclature proposed by Huang and Glinsmann (6), led to the discovery of the deinhibitor protein neutralizing the effect of both inhibitors on phosphorylase phosphatase (7). Besides, it was shown that the dephosphorylation and inactivation of inhibitor-1 by a protein phosphatase isolated from the liver glycogen pellet is controlled by the deinhibitor protein (8). We now report the extensive purification and characterization of the deinhibitor protein isolated from dog liver. In addition to neutralizing the inhibitory effect of inhibitor-1 and modulator protein and bringing about the dephosphorylation of inhibitor-l, the deinhibitor protein also affects the activation-inactivation process of the ATP,Mg-dependent phosphatase.

MATERIALS

AND

METHODS

The experimental materials and procedures, including the preparation of enzymes, modulator protein and 32p-labeled phosphorylase a, enzyme assays and analytical methods have been described previously (4, 9-12). Skeletal muscle inhibitor-1 was prepared according to (13). 32P-labeled inhibitor-1 (5 to 6 X 106 cpm per nmol), with an incorporation of phosphate close to 1 mol per mol was prepared by using ['y-32p]ATP and the catalytic unit of protein kinase (protein kinase: EC 2.7.2.37) prepared according to (14); after boiling the phosphorylated form of inhibitor-1 was reisolated by Sephadex G-50 gel filtration. Protein determinations were carried out by the method of Bradford (15). One unit ofphosphatase activity is defined as the amount of enzyme which releases 1 nmol of [32P]phosphate per min at 30°C in a 30 ALlassay containing 1 mg per ml of [ 32p] phosphorylase a. The deinhibitor was routinely assayed by its ability to relieve the inhibition of the ATP,Mg-dependent phosphatase, produced either by inhibitor-1 or by the modulator in the following assay: 10 #1 containing 10 mU of ATP,Mg-dependent phosphatase (16) and the appropriate amount of FA (10) for full activation, 0.4 mM ATP and 2 mM MgCI2, an amount of phosphatase modulator or inhibitor-1 which inhibits the 10 mU by 50%, 20 mM Tris-HC1 (pH 7.4), 0.5 mM dithiothreitol and 10 t~g bovine serum albumin, preincubated for 10 min with l0 ul deinhibitor sample. Labeled phosphorylase a (10 #1, 3 mg per ml) was added and after 10 min the reaction was stopped by adding trichloroacetic acid and the soluble 32p counted as in (9). Fractions to be assayed for deinhibitor but containing phosphatase activity were boiled prior to assay. One unit ofdeinhibitor activity increases the activity of the phosphatase by one unit under these conditions. The effect of the deinhibitor protein was linear with concentrations up to 4 mU in this assay. When the effects of the deinhibitor protein were studied on purified

PROTEIN P H O S P H A T A S E D E I N H I B I T O R

469

ATP,Mg-dependent phosphatase (9), free Fc (4) or spontaneously active phosphorylase phosphatase isolated from the liver glycogen pellet (see further) some basic or latent phosphatase activity, stimulated by the deinhibitor protein but insensitive to both inhibitor-1 and modulator protein, was observed. Since this variable enzyme activity is apparently due to a contaminating phosphatase, it was subtracted from the control values. The first steps in the purification of the deinhibitor-free protein phosphatase and deinhibitor protein from the glycogen pellet were essentially the same as published previously (8). The cocktail of protease inhibitors, however, was not mandatory. The DEAE-cellulose was equilibrated with 10 mM Tris-HCl, pH 7.8, 0.5 mM dithiothreitol, 1 m u E D T A and 1 mM E G T A (Buffer A). The protein phosphatase activity was measured with phosphorylase a as substrate. The pooled fractions of the deinhibitor-free phosphatase (pool 2) were lyophilized, redissolved in a small vol of buffer B (20 mM Tris-HC1, pH 7.4, 0.5 mM dithiothreitol) and dialyzed against the same buffer. The sample was then applied to an Ultrogel ACA-34 column (2.5 × 100 cm) previously equilibrated with 0.1 M NaC1 in buffer B. The fractions containing the phosphatase eluting in a symmetrical peak with a molecular weight of about 35,000 were pooled and passed directly over a Blue Sepharose CL-6B column (1.5 × 9 cm) equilibrated with 0.1 M NaCI in buffer B. The resin was washed with the same buffer until no more proteins were eluted and the phosphatase was then flushed out with 0.3 M NaCI in buffer B. The active fractions were pooled, diluted 3-fold with buffer B and absorbed on a polylysine-Sepharose 4B column (0.9 x 5 cm) in an identical way. The resin was then eluted with a 40 ml linear salt gradient going from 0.1 M to 0.6 M NaC1 in buffer B. The phosphatase, which eluted in a sharp symmetrical peak, was pooled, dialyzed and concentrated by dialysis against 10% polyethylene glycol and then against 60% glycerol in buffer B. The enzyme (specific activity + 17,000 U per mg) was stored as such at -20°C.

RESULTS

Purification of the Deinhibitor Protein After the DEAE-cellulose chromatography of the resuspended glycogen pellet (8), the deinhibitor containing fractions were pooled (+ 300 ml) and Triton X- 100 was added to obtain a final concentration of 0.2%. This solution was heat-treated by dropping small portions of it at a time with continuous stirring into 100 ml boiling buffer A. The mixture was kept for another 5 min in the boiling water bath. The solution was then cooled on ice and filtered through Whatman nr 1 filter paper. The filtrate was lyophilized, redissolved in a minimal vol of buffer A and dialyzed against 500 ml buffer A for 1 hr with one change in dialysis fluid. In large scale preparations the lyophilized

470

J. GORIS, et al. TABLE 1. P U R I F I C A T I O N OF DOG LIVER D E I N H I B I T O R PROTEIN Deinhibitor activity

Purification step 1. DEAE-cellulose eluate after boiling 2. Lyophilized-dialyzed 3. Sephadex G-75 eluate 4. DEAE-cellulose breakthrough 5. Blue Sepharose eluate

Total protein (mg)

Total activity (units)

Specific activity (units/mg)

Yield (%)

95

2300

24,2

100

87 20.2 12.6 0.5

2275 1320 1050 382

26.1 65.3 83.3 765,6

99 57 46 17

Starting material was 3.5 kg of fresh dog liver. The deinhibitor activity was measured as indicated under "Methods". The activity of the deinhibitor protein in the pooled DEAE-cellulose eluate after boiling was taken as 100%, total protein being less than 0.1% of the protein in the crude extract.

preparations were kept at -20°C, pooled after solubilization and lyophilized again in order to reduce the vol before the next purification step. In a typical preparation (Table 1) the clear dialyzed sample (now +_ 15 ml) was then applied to a column of Sephadex G-75 Superfine (2.5 x 95 cm) previously equilibrated with buffer A. The deinhibitor eluted in one symmetrical peak (Fig. 1) with a Kd value of 0.36. The active fractions were pooled, brought to 20 mM with NaC1 and rechromatographed on a DEAEcellulose column (0.9 × 2 cm) equilibrated with buffer B. The breakthrough fraction, containing more than 95% of the deinhibitor protein recovered in this step, was directly applied to a Blue Sepharose CL-6B column (1.5 × 13 cm) equilibrated with buffer A. After extensive washing with the same buffer the deinhibitor protein was flushed off the column with 0.4 M NaCI in the same buffer. The active fractions were pooled, lyophilized, dissolved in a minimum vol of buffer B and dialyzed by the microdialysis method on a Millipore filter (VSWP, pore size 0.025 taM) during 45 min against the same buffer. In sodium dodecyl sulfate polyacrylamide gel electrophoresis a pattern of protein staining as shown in Figure 2 was observed. When the unstained gels were sliced and assayed for deinhibitor after elution with buffer B, the deinhibitor peak was found at the position of the large diffuse band. Table 1 shows the extent of purification and yields at the different stages in the preparation of the deinhibitor protein. The amount ofdeinhibitor protein after boiling of the fractions obtained by DEAE-cellulose chromatography was taken as 100% activity. At this stage the amount of protein represents less than 0.1% of the protein measured in the crude extract. Detection of the deinhibitor protein in the crude extract or high speed supernatant is impossible, since too many factors interfere with the phosphorylase phosphatase activity and thus with the deinhibitor assay. Furthermore, it is difficult to evaluate the recovery in the boiling step, since even if the heat

471

PROTEIN PHOSPHATASE DEINHIBITOR 2

1

3

4

t

t

!/

15

\

I ¢"i

t~

o

!

10

"7-

;\

J

'~-o- ~x.,_.~,//J / \ \ / 3O 20°'o ~o doNr ¢o ~o ~o ~bo Fraction

FIG. 1. Sephadex G-75 gel filtration of the deinhibitor protein. 15 ml containing 43.5 mg of protein obtained in step 2 (see Table 1) were applied onto a Sephadex G-75 column (2.5 × 95 cm) equilibrated in 20 mM Tris, 0.5 mM DTT, 1 mM EDTA, 1 mM E G T A pH 7.4. Fractions of 4.4 ml were collected and assayed after a 10-fold dilution for deinhibitor activity (O O). Arrows indicate the elution position of Btue dextran (1), myoglobin (2), cytochrome C (3) and vitamin B~2 (4). Optical density was measured at 280 nM (-~ -~).

stability of the deinhibitor protein is beyond question, coprecipitation during the heat denaturation of the phosphatase and other proteins could occur. The recovery of the deinhibitor protein was clearly improved by including 0.2% Triton X-100 in the boiling step; most of the detergent was removed during the subsequent gel filtration step.

Physical Characteristics of the Deinhibitor Protein In sodium dodecyl sulfate polyacrylamide gel electrophoresis the mobility of the deinhibitor corresponds to that of a protein with a molecular weight of 8,300, using as marker proteins: insulin /3-chain (2,300), insulin c~-chain

472

3. GORIS, et aL

(3,500), bovine pancreatic trypsin inhibitor (6,500), cytochrome C (12,500), myoglobin (17,500), chymotrypsinogen (25,000), ovalbumin (45,000) and bovine serum albumin (68,000). It should be noticed, however, that the standard curve of log. mol. wt. against Rf value is not linear in the low molecular weight range (below 12,000). In 5-20% sucrose gradient centrifugation, the deinhibitor migrates as a protein with a sedimentation coefficient of $20,~ = 1.1, using C 14 labeled cytochrome C as an internal marker ($20,~ = 1.9). Assuming a partial specific vol of 0.725, this would correspond to a molecular weight of 5,500. In Sephadex G-75 gel filtration, the deinhibitor runs at the same position as myoglobin (17,500). The Stokes radius of the deinhibitor can therefore be estimated as 19.75 /~. Using the Stokes radius and the sedimentation coefficient, the molecular weight of the deinhibitor protein can be calculated by the method of Siegel and Monty (17) as 8,900 and the frictional ratio (f/fo) as 1.4. This molecular weight corresponds quite well with the molecular weight found in sodium dodecyl sulfate polyacrylamide gel electrophoresis. The deinhibitor appears to be stable in extreme conditions such as 20% trichloroacetic acid, 80% ethanol or boiling in sodium dodecyl sulfate, but is trypsin or pronase labile.

Effect of the Deinhibitor Protein on the ATP, Mg-dependent Protein Phosphatase The ATP,Mg-dependent protein phosphatase can be represented as [FcM], representing a two subunit structure in which Fc is the free catalytic subunit and M the modulator protein. This inactive [Fc-M]-complex is proposed to be the correct substrate for the FA-ATP,Mg-dependent activation, but this notation does not necessarily indicate a 1:1 stoichiometric complex between Fc and M (1). Preparations of the phosphatase which are deficient in M are not activated by the kinase FA and are referred to as "free F c''. Addition of modulator protein to these preparations makes them better substrates for the FA-ATP,Mg-mediated activation (4). The extent as well as the rate of activation by kinase FA in the presence of ATP,Mg largely depends upon the concentration of modulator present (18). This would indicate that the [Fc-M]-complex may bind extra "inhibitory" modulator protein. Clearly the modulator protein is necessary for the activation of the ATP,Mgdependent protein phosphatase, but if present in higher concentrations this effector also becomes inhibitory. Since the "deinhibition" was the first observed property of the deinhibitor protein, it was of interest to investigate the effect of the deinhibitor on the stimulation as well as the inhibition brought about by the modulator protein. This is shown in Figure 3. When the activation of Fc is allowed to proceed in the presence of increasing concentrations of modulator protein a gradual transition from stimulation to

0.5

c-

0./4

O o

D

(J

0.3 O 0

m

c~ °

0.2

m

J~ c0

m

Q;

O

0.1

t

t

DYE FRONT

START SLICE N U M B E R

FIG. 2. Polyacrylamide gel electrophoresis of the deinhibitor protein in the presence of sodium dodecyl sulfate. 10 #g of purified deinhibitor protein were electrophoresed using 15% polyacrylamide gels in 0.1% sodium dodecyl sulfate. The gels were sliced in 2 mm portions and the soluble proteins were extracted overnight in buffer B containing 1 mg per ml bovine serum albumin and assayed for deinhibitor protein. Identical gels were stained for proteins using Coomassie brilliant blue R-250 and destained by diffusion.

473

PROTEIN P H O S P H A T A S E D E I N H I B I T O R

09

08

/

0.7

06

/

A--

o.s

o ~ ° ~

o~ o~ g_

0.3

o

0.2

0.1

i

i

025 05

i

i

i

1

2

4

nM [ M o d u t a t o r ]

FIG. 3. Effect of the deinhibitor protein on the activation of free Fc in the presence of increasing concentrations of modulator protein. Free Fc (0.9 nM) was preincubated with kinase FA (2 U/ml) 0.2 mM ATP, 1 mM MgCI2 and increasing concentrations of modulator protein in the presence (A A) or absence (O O) of 300 nM deinhibitor protein during 10 min at 30°C in 30 ~1. After this preincubation l0 ~l phosphorylase a (4 m g / m l ) were added and phosphatase activity measured after a 5-min incubation at 30°C.

inhibition is observed. This confirms earlier observations (18). However, in the presence of deinhibitor protein the phosphatase activity reaches a higher level when using suboptimal levels of modulator and at higher concentrations of modulator the "inhibitory" effect of the modulator is counteracted. Increasing the concentration of the deinhibitor protein with suboptimal concentrations of modulator protein (Fig. 4A) shows that the activity, which is proportional to the amount of modulator, is further increased by the addition of deinhibitor protein. This increase, which is concentration dependent, reaches a plateau at higher concentrations of deinhibitor protein; the level of this plateau is determined by the concentration of modulator protein. When free Fc is incubated with increasing and inhibitory

474

J. GORIS, et al. B 0.~

o-'----o~_~ 0@ x

x fx

/

0..~


x~ ' / 0.6

~O--O

m

o

-

/

oO.~

-

.cQ . 0/. .c Q.

A

A 0.1

0.; 3-0--0

I

o

-

260

I

I

'x

o t

I

25 50 100

-

/

x

400

10 20 40

d

0

i

160

I

320

nM [Oeinhi bitor]

FIG. 4. Effect of the deinhibitor protein on the activation of free Fc in the presence of sub- and supra-optimal concentrations of modulator protein. (A) Free Fc (3 riM)was preincubated with kinase FA (3 U/ml), 0.2 mM ATP, 1 mMMgCI2,0.04 nM(O O),0.08nm (~ A),0.16nM (D D) and 0.32 nM(X X) modulator protein in the presenceof increasingconcentrations of deinhibitor protein during 10 rain at 30°C in 30 #1. After this preincubation 10 #1 of phosphorylase a (4 mg/ml) were added and phosphatase activity measured after a 10-min incubation at 30°C. (B) Free F¢ (0.5 nM) was preincubated with kinase FA (1.5 U/ml), 0.2 mM ATP, 1 mM MgC12, 0.16 nM (O ©), 0.64 nM (A A), 1.28 nM (O O) and 10.24 nM (× ×) modulator protein in the presence of increasingconcentrations of deinhibitor protein and assayed as in Fig. 4A. c o n c e n t r a t i o n s of m o d u l a t o r (Fig. 4B), the d e i n h i b i t o r protein can restore the m a x i m a l activity; however, the A 5° increases. F r o m these results it appears that the i n h i b i t i o n of the A T P , M g o d e p e n d e n t p h o s p h a t a s e by the m o d u l a t o r can be neutralized by the deinhibitor, b u t that o n the contrary the effect of the m o d u l a t o r in the activation process is stimulated. This observation again emphasizes the dual role of the m o d u l a t o r p r o t e i n a n d suggests that the m e c h a n i s m s involved in these effects are not necessarily related. It is also clear that the d e i n h i b i t o r protein c a n n o t replace the m o d u l a t o r in the FA-ATP,Mgd e p e n d e n t activation process. T o investigate further the stimulatory effect of the d e i n h i b i t o r p r o t e i n on the activation process, the F A - A T P , M g - d e p e n d e n t activation of the [Fc-M Icomplex was e x a m i n e d in the presence of s u b o p t i m a l c o n c e n t r a t i o n s of kinase FA (Fig. 5). A p p a r e n t l y the rate of activation, which is d e p e n d e n t u p o n the FA c o n c e n t r a t i o n , is not influenced by the presence of the d e i n h i b i t o r protein. H o w e v e r instead of leveling off as in the absence of deinhibitor, the activation proceeds further in its presence. The further increase in activity is n o t due to a

PROTEIN PHOSPHATASE DEINHIBITOR

475

/v

iI

2

•

fl

iii

In

///

i

/¢///

o

IL

! ill / rll / /

I~o ,

/

o ~ "

Igin.

o--

o

preincubati0n

FIG. 5. Effect of the deinhibitor protein on the activation of the ATP,Mg-dependent protein phosphatase by rate limiting concentrations of kinase FA. [FcM]-enzyme (6.6 nM) was preincubated with 0.8 U (e,O), 1.7 U (A,/x,),3.4 U (=,D)and 6.8 U (V,V) protein kinase FA per ml, 0.2 mM ATP and 1 mM MgCI2 at 30°C in the presence (. . . . ) and absence ( ) of 200 nM deinhibitor protein. For the phosphatase assay 10 /~1 samples were incubated with 10/~1 of phosphorylase a (2 mg/ml) during 2 min.

stabilizing effect of the deinhibitor on one of the factors involved in the activation, since when the deinhibitor is added after the plateau is reached a further t i m e - d e p e n d e n t activation is observed. F r o m these results it appears that due to the action of the d e i n h i b i t o r protein more inactive phosphatase is made available for the activation process, or more likely that in an e q u i l i b r i u m situation between activation a n d inactivation, the phosphatase is stabilized in its active c o n f o r m a t i o n . This effect becomes very clear when the rate of activation is limited by the a m o u n t of kinase FA. U n d e r these c o n d i t i o n s the extent of activation is increased by the deinhibitor protein while the rate of activation of the A T P , M g - d e p e n d e n t phosphatase remains u n c h a n g e d .

Effect of the Deinhibitor Protein on the Spontaneously Active Phosphorylase Phosphatase Inhibition by inhibitor-1. While the effect o f i n h i b i t o r - I o n the p h o s p h a t a s e is truly a n o n - t i m e d e p e n d e n t inhibition, the effect o f the d e i n h i b i t o r is time ~R

22-P*

476

J. GORIS, et al.

dependent and a 5 to 7 min preincubation is required to see its full effect (not shown). As shown in Figure 6 the deinhibitor protein can completely neutralize the effect of inhibitor-l, without adding kinase F A and ATP,Mg. This result is quite different from the observation with modulator protein and shows quite clearly that no inactivation of the phosphatase is observed with inhibitor-1 and that in antagonizing the effect of inhibitor-I the deinhibitor protein functionally behaves as a true deinhibitor. Furthermore, both factors behave antagonistically since the 15° for inhibitor-1 and the A 5° for the deinhibitor change in concert. Inactivation by the modulator protein. Before the role of the modulator protein in the activation process became clear (4), evidence was provided that the protein had the capacity to reverse the FA-ATP,Mg-mediated activation, and it was observed that the modulator protein, then still called inhibitor-2, was responsible for the time-dependent inactivation of a partially purified

8C

~

•

L0

~ o

t~s 21s

;

;o

2'0

nM [Inhibitor I] FIG. 6. Effect of the deinhibitor protein on the inhibition of the spontaneously active liver protein phosphatase by inhibitor-]. After a 10-rain preincubation at 30°C without (© ©) or with 15 riM(× ×), 30nM (D D),60 nM(e -') or ]20nM (V 7) deinhibitor protein, the phosphatase (0.65 U/m]) was added to an equal vol (20 #1) containing 2 mg/m] phosphorylase a and increasing concentrations of inhibitor-l, shown as final concentrations, for a 10-min incubation period. Phosphatase activity is expressed as percent of the enzyme activity in the absence of added inhibitor.

PROTEIN P H O S P H A T A S E D E I N H I B I T O R

477

preparation of glycogen bound phosphorylase phosphatase isolated from dog liver (19). This was recently confirmed for the active protein phosphatase isolated from the glycogen pellet of skeletal muscle (5) and now for the deinhibitor-free phosphatase isolated from the glycogen pellet of dog liver. Figure 7 shows the modulator concentration dependent inactivation of the protein phosphatase. When the enzyme was preincubated with the deinhibitor protein, the phosphatase was clearly stabilized and much less sensitive to the effect of modulator protein at concentrations that otherwise completely

100

80

°, I

\\

60

I\

\\o

/.0

0's,

i

i

2

i

,

aM [Moclutator] FIG. 7. Effect of the deinhibitor protein on the interconversion of the spontaneously active liver protein phosphatase to the ATP,Mg-dependent enzyme form. The phosphatase (1.25 U / m l ) was incubated with 625 nM deinhibitor protein (closed symbols) or with increasing concentrations of modulator protein (open symbols) for 10 min at 30°C. Subsequently, 20 #1 of the phosphatase preincubated with the deinhibitor protein were incubated for 10 rain at 30°C with 20 ~1 containing increasing concentrations of modulator protein in the presence (e e) or absence (A A) of protein kinase FA (2.5 U/ml), 0.05 mM A T P and 0.25 mM MgCI2. 20 /~1 samples of the phosphatase preincubated with the modulator protein were incubated for 10 min at 30°C with 20 ~1 buffer B containing I mg bovine serum albumin (A A), 625 nM deinhibitor protein (x x), protein kinase FA (5 U/ml), 0.1 mM ATP and 0.5 mM MgCI2 in the absence (O O) or presence of 310 nM deinhibitor protein (t~ D). Phosphatase activity was measured after addition of 10 #1 of phosphorylase a (5 m g / m l ) in a 10-min incubation. The concentrations of modulator protein are those in the final 50 #1 assay. Phosphatase activity is expressed as percent of the enzyme activity in the absence of added modulator protein.

478

J. GORIS, et

al.

inactivated the enzyme. This is in sharp contrast to the absence of any effect of the deinhibitor when added after inactivation of the phosphatase by the modulator. After inactivation by the modulator protein, activity could be restored only partially by an incubation with kinase FA in the presence of ATP,Mg. Complete reactivation, however, can be obtained when the inactivated phosphatase is incubated with FA and ATP,Mg in the presence of the deinhibitor protein.

DISCUSSION The deinhibitor protein, which was discovered during studies on the lack of sensitivity of crude fractions of glycogen bound phosphorylase phosphatase to inhibitor-1 and modulator protein (7) has now been purified extensively from dog liver. The deinhibitor protein has furthermore been detected in liver glycogen particles from rats and rabbits (unpublished observations). Its presence in the cytosol cannot be excluded because of the difficulties encountered with the assay procedure. However, since glycogen synthesis requires both glycogen synthase and phosphorylase as dephosphorylated enzymes and therefore an active multisubstrate protein phosphatase, insensitive to inhibitor-1 and the modulator protein, the deinhibitor protein could preferentially be associated with the protein glycogen complex (20). This phenomenon could be more evident in liver which evidently can build up great glycogen stores, largely to the benefit of other tissues. The protein phosphatase present in the resuspended liver glycogen pellet dephosphorylates and inactivates inhibitor-1 in the absence o f M n 2+. This activity of the phosphatase, which is lost during purification of the enzyme, can be restored upon addition of the deinhibitor protein (8). This requirement makes the presence of unphysiological concentrations of Mn 2+ unnecessary and thus completes the physiological phosphorylation-dephosphorylation cycle for inhibitor-1. This could explain to a certain extent the lack of sensitivity of the glycogen bound phosphatase to inhibitor-1 but this phenomenon should be distinguished from the "deinhibition" effect, since deinhibition is observed using the deinhibitor in the nM range. The concentration of added deinhibitor is about a 100-fold higher for dephosphorylation of inhibitor-l, which explains why deinhibition is observed under conditions where no dephosphorylation of inhibitor-1 occurs (not shown). Since the discovery of the deinhibitor protein (7) much progress has been made in the understanding of the activation-inactivation of the multisubstrate protein phosphatase. Investigating in more detail the ATP,Mg-dependent protein phosphatase system isolated from skeletal muscle, it has been shown that at least three protein factors are involved: an inactive phosphatase subunit (Fc), a heat stable modulator protein (M) and an activating protein factor (kinase FA) which turned out to be a very efficient

PROTEIN PHOSPHATASE D E I N H I B I T O R

479

synthase kinase. The [Fc-M]-complex appears to be the correct unit for the FA-ATP,Mg-dependent activation. These observations have been summarized in a recent review (1). In this proposal the modulator protein is necessary for the FA-ATP,Mg-dependent activation of the inactive phosphatase and cannot only inhibit the spontaneously active phosphatase (which explains the original appelation of inhibitor-2) but also causes a time- and concentrationdependent inactivation of the enzyme. Therefore, it should come as no surprise that the effects of the deinhibitor protein, as related to the modulator protein, are more complex than what was first described as a simple "deinhibition" (7). In the activation of the ATP,Mg-dependent phosphatase (Fc-M) as well as in the reactivation of the spontaneously active phosphatase inactivated by preincubation with the modulator protein, full activity can only be observed when the activation by FA-ATP,Mg proceeds in the presence of the deinhibitor protein. This is clear when the activation of Fc is allowed to proceed with suboptimal concentrations of modulator protein or when the activation is brought about by rate limiting concentrations of protein kinase FA. In the first case the final activity reached, which is primarily determined by the amount of modulator available, is higher with the deinhibitor. In the second case, the rate limited activation, defined by the amount of kinase FA present, proceeds further in time in the presence of the deinhibitor. From these observations it is obvious that certainly no antagonism exists between the modulator and the deinhibitor protein in the activation process. As to the inactivation of the spontaneously active phosphatase which is brought about by the modulator protein, two new probably related observations are of importance. First, the inactivation requires higher concentrations of modulator when the phosphatase is preincubated with the deinhibitor protein. Secondly, the FA-ATP,Mg-dependent reactivation is only complete in the presence of the deinhibitor protein. The mechanism of activation-inactivation of the ATP,Mg-dependent protein phosphatase has been studied in more detail (1, 21). The activation of the inactive [FcM]-enzyme by kinase FA is initiated by a phosphorylation of the modulator moiety [M] which produces the transition of the inactive catalytic subunit [Fc] into an active conformation [Fc*]. The phosphorylated state of the modulator is only transitory and a rapid dephosphorylation, presumably autocatalytic, results in the formation of a fully active [Fc*M]enzyme, which then slowly reverts back to the FA-ATP,Mg-dependent enzyme form. In this activation-inactivation process the modulator remains associated in the active as well as the inactive phosphatase complex as a regulatory subunit. This representation is at variance with the proposal of Cohen and coworkers (22) of a dissociation of the catalytic subunit and the phosphorylated modulator protein during activation. However, the experiments leading to this hypothesis were carried out with a catalytic

480

J. GORIS, et aL

subunit (molecular weight: 35,000) which is probably the result of a partial proteolysis (23) and free modulator protein, which probably does not exist as such under physiological conditions (24). The effect of the deinhibitor protein on the inactivation of the spontaneously active protein phosphatase as well as on the FA-ATP,Mg-dependent activation of its inactive enzyme-form could be explained by a stabilizing effect on the slow conversion of the active protein phosphatase to the inactive [FcM]-enzyme. The deinhibitor protein is a relatively small heat- and acid-stable peptide with a molecular weight of about 9,000. Its behavior in the different procedures used to determine its molecular weight is consistent with the assumption that the deinhibitor is not a globular protein and that it may have little ordered structure. In this it resembles its "antagonists", since both inhibitor-I (13) and the modulator protein (25) are asymmetrical proteins with little ordered structure. As with inhibitor-1 and the modulator protein the problem of a possible artifact, as reviewed in (1), will arise, especially since the deinhibitor protein can only be isolated or assayed after a boiling step. The effects of the deinhibitor protein are observed at very low concentrations and explain in several ways the lack of sensitivity to inhibitor-1 and modulator protein of the glycogen bound phosphatase from liver. The deinhibitor protein is clearly different from the heat stable activator isolated from liver and skeletal muscle by Khandelwal and Zinman (26) which was specific for the dephosphorylation of histone and was proposed to be of no influence on the interaction with what was described as inhibitor-2. Grankowski et al. (27) have isolated two heat stable protein activators from rabbit reticulocytes, one specific for the dephosphorylation of 40S ribosomal subunits or phosphohistone (molecular weight: 12,500) and the other one specific for eukaryotic initiation factor-2 (molecular weight: 17,400). Both activators are dependent upon Mn 2÷ and seem to interact with the phosphatase and not with the substrate. The effect was furthermore dependent on an ethanol precipitation of the phosphatase preparation. Because of the difference in criteria a comparison is difficult to assess. In kidney Wilson et al. (28) identified a heat stable activator protein of phosphorylase phosphatase. Based on the large discrepancy in molecular weight the deinhibitor protein and this activator appear to be different. The protein kinase inhibitor purified from skeletal muscle (29) and the deinhibitor protein have similar physical characteristics. Although from a physiological point of view it would make sense if both activities reside in the same protein, it was shown in crosswise assays that this is not the case (30). Since the insulin generated factor described by Larner et al. (31) stimulates glycogen synthase phosphatase activity but inhibits cyclic AMP-dependent protein kinase activity as well, this factor also appears to be unrelated to the deinhibitor protein, which does not inhibit the protein kinase (30). Thus the deinhibitor protein has pronounced effects on the interaction

PROTEIN PHOSPHATASEDEINHIBITOR

481

between the active protein phosphatase and inhibitor-1 and on the activationinactivation process of the ATP,Mg-dependent protein phosphatase. In this way this factor could determine the relative sensitivity of the multisubstrate protein phosphatase to inhibitor-1 and the modulator protein. In the classification of protein phosphatases involved in cellular regulation (32) a fundamental and convenient criterium is usually the sensitivity to inhibitor-1 and the modulator protein. The role of the deinhibitor in the interaction between the multisubstrate ATP,Mg-dependent phosphatase and inhibitor-1 or the modulator should be a warning in these attempts of classification. Preliminary results have shown that addition of deinhibitor to a crude gel filtered extract enhances the phosphorylase phosphatase activity severalfold. Because it appears that the ratio deinhibitor protein: inhibitor-1 or modulator can determine the activity of the major multisubstrate protein phosphatase, it is tempting to speculate that the deinhibitor could be a factor subject to metabolic and hormonal regulation. Since the deinhibitor protein can neutralize the effect of inhibitor-I even without the dephosphorylation as a prerequisite, the deinhibitor could protect and increase the phosphatase activity and thus antagonize the effect of cyclic AMP without a necessary decrease in the concentration of the cyclic nucleotide.

SUMMARY AND CONCLUSIONS In liver and muscle the major active phosphorylase and synthase phosphatase activity is associated with the glycogen particle. When we examined the effect of the inhibitor-I and modulator protein on the enzyme present in crude glycogen fractions from dog liver, the phosphorylase phosphatase was not or only slightly affected. Since the enzyme isolated from the glycogen complex by DEAE-cellulose chromatography could be inhibited by inhibitor-I as well as the modulator protein, it was assumed that an unknown mechanism or factor present in the glycogen fraction was responsible for this reduced sensitivity of the protein phosphatase. This led to the discovery (7) of the deinhibitor protein which has now been extensively purified from dog liver. The deinhibitor protein was shown to be thermostable, ethanol- and trichloroacetic acid-resistant, but non-dialyzable and it was destroyed by pronase or trypsin. The apparent molecular weight was estimated at about 17,500 in gel filtration, 8,300 in sodium dodecyl sulfate polyacrylamide gel electrophoresis and 5,500 in sucrose density gradient centrifugation, behavior which is consistent with the assumption that the deinhibitor protein may have little ordered structure. Glycogen synthesis requires both phosphorylase and glycogen synthase as dephosphorylated enzymes. The interaction of the deinhibitor protein with the protein phosphatase brings about several effects which, when considered together, could all facilitate the dephosphorylation of glycogen synthase and

482

J. GORIS, et al.

p h o s p h o r y l a s e . T h e protein p h o s p h a t a s e present in a resuspended glycogen pellet d e p h o s p h o r y l a t e s inhibitor-1 in the absence o f M n 2+. This ability o f the p h o s p h a t a s e , which is lost during purification o f the enzyme, can be restored u p o n a d d i t i o n o f the d e i n h i b i t o r protein. O w i n g to the association o f the d e i n h i b i t o r p r o t e i n with the active p h o s p h a t a s e the e n z y m e b eco m es insensitive to inhibition by inhibitor-1 a n d the m o d u l a t o r protein, an d m o r e resistant to the c o n v e r s i o n into the F g - A T P , M g - d e p e n d e n t form, b r o u g h t a b o u t by the m o d u l a t o r protein. D u r i n g the a c t i v a t i o n o f the A T P , M g d e p e n d e n t p h o s p h a t a s e u n d e r conditions where kinase FA is rate limiting, the d e i n h i b i t o r p r o te in increases the level w i t h o u t affecting the rate o f activation.

ACKNOWLEDGEMENTS T h e au t h o rs are grateful to Ms. Rita Bollen for expert technical assistance. This w o r k was s u p p o r t e d by the " O n d e r z o e k s f o n d s K. U . L e u v e n " an d by the " F o n d s voor G e n e e s k u n d i g W e t e n s c h a p p e l i j k O n d e r z o e k " .

REFERENCES I.

W. M E R L E V E D E , J. R. V A N D E N H E E D E , J. GORIS and S.-D. YANG, Regulation of the ATP,Mg-dependent protein phosphatase, Curr. Top. Cell. Reg. 23, 177-215 (1984).

2. W. MERLEVEDE, T. CAMPS, J. GORIS, J. R. VANDENHEEDE and S.-D. YANG, ATP,Mg-dependent protein phosphatase in mammalian tissues, Wiss. Tag. Bet. Univ. Halle 3, 135-146 (1983). 3. W. MERLEVEDE, The role of dephosphorylation by the ATP,Mg-dependent protein phosphatase in mammalian glycogen metabolism, Verhandelingen van de Koninklijke Academie voor Geneeskunde van Belgie, 45, 275-302 (1983). 4. S.-D.YANG, J. R. VANDENHEEDE and W. MERLEVEDE, Identification of inhibitor-2 as the ATP,Mg-dependent protein phosphatase modulator, J. Biol. Chem. 256, 10213-10234 (1981). 5. J. R. VANDENHEEDE, S.-D. YANG and W. MERLEVEDE, Role of the modulator protein in the interconversion of rabbit skeletal muscle protein phosphatase, Bioehem. Biophys. Res. Commun., 115, 871-877 (1983). 6. F. L. HUANG and W. H. GLINSMANN, Separation and characterisation of two phosphorylase phosphatase inhibitors from rabbit skeletal muscle, Europ. J. Biochem. 70, 419-426 (1976). 7. G. DEFREYN, J. GORIS and W. MERLEVEDE, A deinhibitor protein neutralizing the effect of the protein inhibitors on dog liver phosphorylase phosphatase, FEBS Letters 79, 125-128 (1977). 8. J. GORIS, T. CAMPS, G. DEFREYN and W. MERLEVEDE, The dephosphorylation of protein phosphatase inhibitor-1 is controlled by the deinhibitor protein, FEBS Letters 134, 189-193 (1981). 9. S.-D. YANG, J. R. VANDENHEEDE, J. GORIS and W. MERLEVEDE, ATP,Mgdependent protein phosphatase from rabbit skeletal muscle. I. Purification of the enzyme and its regulation by the interaction with an activating protein factor, J. Biol. Chem. 255, 11759-11767 (1980). 10. J. R. VANDENHEEDE, S.-D. YANG, J. GORIS and W. MERLEVEDE, ATP,Mgdependent protein phosphatase from rabbit skeletal muscle. II. Purification of the activating factor and its characterisation as a bifunctional protein also displaying synthase kinase activity, J. Biol. Chem. 255, 11768-11774 (1980).

PROTEIN PHOSPHATASE DEINHIBITOR 11.

12. 13. 14.

15. 16. 17. 18. 19. 20.

21. 22.

23.

24. 25. 26. 27.

28.

483

J . R . VANDENHEEDE, S.-D. YANG and W. MERLEVEDE, Rabbit skeletal muscle

protein phosphatase(s): identity of phosphorylase- and synthase-phosphatase and interconversion to the ATP,Mg-dependent enzyme form, J. Biol. Chem. 256, 5894-5900 (1981). S.-D. YANG, J. R. VANDENHEEDE and W. MERLEVEDE, A simplified procedure for the purification of the protein phosphatase modulator (inhibitor-2) from rabbit skeletal muscle, FEBS Letters 132, 293-295 (1981). G.A. NIMMO and P. COHEN, The regulation of glycogen metabolism. Purification and characterisation of protein phospbatase inhibitor-1 from rabbit skeletal muscle, Europ. J. Biochem. 87, 341-351 (1978). V. KINZEL and D. KUBLER, Single step purification of the catalytic subunit(s) of cyclic 3,5-adenosine monophosphate-dependent protein kinase(s) from rat muscle, Biochem. Biophys. Res. Commun. 71, 257-264 (1976). M . M . BRADFORD, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72, 248-254 (1976). J. GORIS, F. DOPERE, J. R. VANDENHEEDE and W. MERLEVEDE, Regulation of liver phosphorylase phosphatase. ATP,Mg-mediated activation of the partially purified dog liver enzyme, FEBS Letters 117, 117-121 (1980). L.M. SIEGEL and K. J. MONTY, Determination of molecular weights and frictional ratios of macromolecules in impure systems: aggregation of urease, Biochem. Biophys. Res. Commun. 19, 494-499 (1965). S. SIVARAMAKRISHNAN, J. R. VANDENHEEDE and W. MERLEVEDE, Characterization of different forms of kinase FA from rabbit skeletal muscle, Advances in Enzyme Regulation 23, 321-330 (1983). J . R . VANDENHEEDE, J. GORIS, S.-D. YANG, T. CAMPS and W. MERLEVEDE, Conversion of active protein phosphatase to the ATP,Mg-dependent enzyme form by inhibitor-2, FEBS Letters 127, 1-3,(1981). W. MERLEVEDE, Properties and regulation of liver glycogen phosphorylase phosphatase, Verhandelingen van de Koninklijke Academie voor Geneeskunde van Belgie 42, 135-155 (1980). S. JURGENSEN, E. SHACTER-NOIMAN, C. Y. HUANG, P. B. CHOCK, S.-D. YANG, J. R. VANDENHEEDE and W. MERLEVEDE, On the mechanism of activation of the ATP,Mg-dependent protein phosphatase by kinase FA, J. Biol. Chem., in press. T.J. RESINK, B. A. HEMMINGS, H. Y. LIM TUNG and P. COHEN, Characterisation of a reconstituted Mg-ATP-dependent protein phosphatase, Europ. J. Biochem. 133, 455-461 (1983) J . R . VANDENHEEDE, S.-D. YANG, J. GORIS and W. MERLEVEDE, Regulation of rabbit muscle MgATP-dependent protein phosphatase by interaction with cyclic AMP- and Ca2+-independent synthase kinases, pp. 497-513 in Cold Spring Harbor Symposia on Cell Proliferation: vol. 8 Protein Phosphorylation, Cold Spring Harbor, L.I., New York (1981). S.-D. YANG, J. R. VANDENHEEDE and W. MERLEVEDE, The heat labile phosphatase modulator (inhibitor-2) complex from rabbit skeletal muscle, Biochem. Biophys. Res. Commun. 113, 439-445 (1983). J.G. FOULKES and P. COHEN, The regulation of glycogen metabolism. Purification and properties of protein phosphatase inhibitor-2 from rabbit skeletal muscle, Europ. J. Biochem. 105, 195-203 (1980). R.L. KHANDELWAL and S. M. ZINMAN, The presence of a heat-stable activator of phosphoprotein phosphatase for the dephosphorylation of phosphorylated histone in rabbit liver, Biochem. Biophys. Res. Commun. 82, 1340-1345 (1978). N. GRANKOWSKI, D. LEMUSVIRTA, G. B. STEARNS, G. KRAMER and B. HARDESTY, The isolation and partial characterisation of two-substrate-specificprotein activators of the reticulocyte phosphoprotein phosphatase, J. Biol. Chem. 255, 5755-5762 (1980). S.E. WILSON, R. L. MELLGREN and K. K. SCHLENDER, Isolation of a heat stable protein activator of phosphorylase phosphatase, FEBS Letters 146, 331-334 (1982).

484 29. 30. 31. 32.

3. GORIS, et al. S. WHITEHOUSE and D. A. WALSH, Purification of a physiological form of the inhibitor protein of the cAMP-dependent protein kinase, J. Biol. Chem. 257, 6028-6032 (1982). E. WAELKENS, J. GORIS, D. A. WALSH and W. MERLEVEDE, About the non identity of protein phosphatase deinhibitor and protein kinase inhibitor, Biochem. Soc. Trans., in press. J. LARNER, G. GALASKO, K. CHENG, A. A. DE PAOLI-ROACH, L. HUANG, P. D A G G Y and J. KELLOG, Generation by insulin of a chemical mediator that controls protein phosphorylation and dephosphorylation, Science 206, 1408-1410 (1979). T. S. INGEBRITSEN and P. COHEN, The protein phosphatases involved in cellular regulation. 1. Classification and substrate specificities, Europ. J. Biochem. 132, 255-261 (1983).