The protein phosphatases involved in cellular regulation

Transcription

1 Eur. J. Biochem. 145,51-56 (1984) 0 FEBS 1984 The protein phosphatases involved in cellular regulation Antibody to protein phosphatase-2a as a probe of phosphatase structure and function Susana ALEMANY, H. Y. Lim TUNG, Shirish SHENOLIKAR, Simon J. PILKIS, and Philip COHEN Department of Biochemistry, University of Dundee; Department of Pharmacology, University of South Alabama, Mobile, Alabama; and Department of Physiology, Vanderbilt University, Nashville, Tennessee (Received April 19/July 24, 1984) - EJB Antibody prepared against the catalytic subunit of protein phosphatase-2a from rabbit skeletal muscle, could completely inhibit ths enzyme, but did not significantly affect the activities of protein phosphatases-i, 2B and 2C. The antibody was used to establish the following points. 1. The three forms of protein phosphatase-2a that can be resolved by ion-exchange chromatography, termed 2Ao, 2A1, and 2A2, share the same catalytic subunit. 2. The antigenic sites on the catalytic subunit of protein phosphatase-2a remain accessible to the antibody, when the catalytic subunit is complexed with the other subunits of protein phosphatases-2ao, 2A1 and 2A2. 3. The catalytic subunits of protein phosphatase-2a from rabbit skeletal muscle and rabbit liver are very similar, as judged by immunotitration experiments. 4. Protein phosphatase-1 and protein phosphatase-2a account for virtually all the phosphorylase phosphatase activity in dilute tissue extracts prepared from skeletal muscle, liver, heart, brain and kidney, and for essentially all the glycogen synthase phosphatase activity in dilute skeletal muscle and liver extracts. 5. Protein phosphatase-2a is almost absent from the protein-glycogen complex prepared from skeletal muscle or liver extracts. 6. Protein phosphatase-2a accounts for a major proportion of the phosphatase activity in dilute liver extracts towards 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, 6-phosphofructo-1 -kinase, fructose 1,6-bisphosphatase, pyruvate kinase and phenylalanine hydroxylase, the major phosphorylated enzymes involved in the hormonal control of hepatic glycolysis and gluconeogenesis. We have recently investigated the nature of the protein phosphatases involved in cellular regulation [l -91. Based on experiments with 18 phosphoprotein substrates, we were only able to identify four protein phosphatase catalytic subunits capable of dephosphorylating the major proteins involved in the control of many metabolic processes in mammalian liver and skeletal muscle, such as glycogen metabolism, glycolysis and gluconeogenesis, aromatic amino acid breakdown, and fatty acid, cholesterol and protein synthesis. The four phosphatases have been classified into two types depending on whether they dephosphorylate the P-subunit of phosphorylase kinase and are inhibited by two thermostable proteins, termed inhibitor-1 and inhibitor-2 (type-i protein phosphatase), or whether they dephosphorylate the &-subunit of phosphorylase kinase preferentially and are insensitive to inhibitors-i and 2 Abbreviation. SDS, sodium dodecyl sulphate. Enzymes. Glycogen phosphorylase (EC l); glycogen synthase (EC ); phosphorylase kinase (EC ); cyclic- AMP-dependent protein kinase (EC ); glycogen synthase kinase-3 (EC ); protein phosphatase (EC ); 6-phosphofructo-1 -kinase (EC ); fructose-l,6-bisphosphatase (EC ); pyruvate kinase (EC ); phenylalanine hydroxylase (EC ); 6-phosphofructo 2-kinase/fructose 2,6-bisphosphatase (EC ). (type-2 protein phosphatases). Three type-2 protein phosphatases have been identified, termed 2A, 2B and 2C, that are distinct enzymes. All four protein phosphatases have broad and overlapping substrate specificities in vitro, raising the question of the relative importance of each enzyme in the dephosphorylation of different proteins in vivo. For this reason, we have attempted to measure the contribution of each phosphatase to the dephosphorylation of different substrates in tissue extracts, by using specific inhibitors. Type-I protein phosphatases can be inhibited specifically by inhibitor-2, protein phosphatase- 2B (a Ca2 +-dependent, calmodulin-stimulated enzyme) with trifluoperazine or EGTA, and protein phosphatase-2c (an Mg2+-dependent activity) with EDTA. In the presence of these substances, the residual activity has been assumed to be protein phosphatase-2a, since no other phosphatase resistant to inhibitor-2, trifluoperazine, EGTA and EDTA can be detected following fractionation of tissue extracts by anionexchange chromatography and gel filtration. However more definitive assessment of the contribution of protein phosphatase-2a requires a specific inhibitor of this enzyme. We have therefore prepared an antibody to the catalytic subunit of protein phosphatase-2a from rabbit skeletal muscle [lo]; in this paper we describe its use as a specific probe of the structure and function of the enzyme.

2 52 MATERIALS AND METHODS Proteins and 32 P-labelled phosphoprotein substrates The 33-kDa catalytic subunit of protein phosphatase-1, the 36-kDa catalytic subunit of protein phosphatase-2a, protein phosphatase-2b, glycogen phosphorylase, glycogen synthase, phosphorylase kinase, the catalytic subunit of cyclic-amp-dependent protein kinase, glycogen synthase kinase-3 and inhibitor-2 were- isolated from rabbit skeletal muscle. 6-Phosphofructo-2-kinase/fructose-2,6-bisphospha- tase, 6-phosphofructo-l-kinase, fructose-l,6-bisphosphatase, pyruvate kinase, phenylalanine hydroxylase and protein phosphatase-2c were from rat liver. The methods for preparing these proteins are referenced or detailed in two of the three accompanying papers [8, 91. Procedures for making 32Plabelled phospho-protein substrates (lo6 cpm/nmol) are detailed in [8, 9, 111. Antibody was prepared in sheep to the catalytic subunit of protein phosphatase-2a as in [lo]. The IgG fraction from the third bleed was used in all the studies described under Results and Discussion. Rabbit anti-(sheep IgG) immunoglobin was purchased from Miles Laboratories Ltd (Slough, UK). Measurement of protein concentration This was carried out by absorbance measurements at 280 nm in the case of phosphorylase and phosphorylase kinase (see [8, 91) or by the method of Bradford [12] using bovine serum albumin (A;& nm = 6.5) as a standard. Preparation of tissue extracts and protein-glycogen complex These were prepared from skeletal muscle, liver, brain and heart of normally fed rabbits as in [7], except that the proteinase inhibitors phenylmethanesulphonyl fluoride (0.1 mm), benzamidine (1.0 mm) and leupeptin (4 pg/ml) were included in the homogenisation buffer. The liver glycogen pellet was separated from the microsomal fraction as in 171. Rabbit kidneys were chopped finely with scissors and homogenised for 45 s at low speed in a Waring blendor with 3 vol. of homogenisation buffer (4.0 mm EDTA ph 7.0, 250 mm sucrose, 0.1 YO, v/v, 2-mercaptoethanol and proteinase inhibitors). The homogenate was centrifuged at x g for 10 min and the supernatant decanted. Assay of protein phosphatases Protein phosphatases were assayed as described in the preceding paper [8] and in [6]. Additions of divalent cations are specified in the legends to figures and tables presented under Results and Discussion. Protein phosphatase-1 and 2A activities were standardised using phosphorylase a (1.O mg/ ml) as substrate. 1.O unit of phosphorylase phosphatase was that amount which catalysed the release of 1.0 ymol phosphate/min in the standard assay (without divalent cations). A concentration of 1.0 mu/ml corresponds to 10% dephosphorylation of phosphorylase a/min. Immunoprecipitation of protein phosphatases Purified protein phosphatases or tissue extracts were diluted, in ice-cold 50 mm Tris/Cl ph 7.0, 0.1% (v/v) 2-mercaptoethanol, 1.O mg bovine serum albumin/ml. Ali- quots (0.02 ml) were then mixed with 0.02 ml of antibody to protein phosphatase-2a (7 mg/ml) diluted in 50 mm Tris/Cl ph 7.0, 0.1 % (v/v) 2-mercaptoethanol. After incubation for 3 h at 4 C anti-(sheep IgG) (42 mg/ml) in the same buffer (0.01 ml) was added. After a further 2 h at 4 C, solutions were centrifuged for 5 min at xg. An aliquot of the supernatant was assayed for protein phosphatase activity at a further 1 : 3 dilution in the presence of inhibitor-2 (800 U/ ml). All experiments were performed in duplicate and control incubations were carried out in which antibody to protein phosphatase-2a was replaced by dialysis buffer. The initial dilutions of the phosphatases were selected so that less than 10% release of [32P]phosphate from [32P]protein substrate, occurred in the final 30-min assay. This corresponds to a phosphorylase phosphatase concentration of 0.1 mu/ml. Skeletal muscle, liver and kidney extracts were diluted 45-fold, and brain and heart extracts 20-fold. The activities of purified protein phosphatases were stable over the 5-h incubations. However addition to rabbit anti- (sheep IgG) immunoglobulin caused a 20% decrease in the activities of protein phosphatases-i, 2A, 2B and 2C, even in control incubations where antibody to protein phosphatase- 2A had been omitted. In tissue extracts, protein phosphatase activities decreased by up to 30% over the 5-h incubation period, even in incubations where both antibody to protein phosphatase-2a and anti-(sheep IgG) immunoglobulin had been omitted. The percentage inhibition produced by antibody to protein phosphatase-2a was determined by comparison to appropriate control incubations. RESULTS AND DISCUSSION Specificity of antibody,for the catalytic subunit of protein phosphatase-2a We have previously reported that this antibody can inactivate the catalytic subunit of protein phosphatase-2a from rabbit skeletal muscle without affecting the catalytic subunit of protein phosphatase-l from the same tissue [lo]. In the present work, the ability of antibody to inhibit protein phosphatases-1, 2A, 2B and 2C was examined using a substrate (phosphorylase kinase) that is dephosphorylated efficiently by all four enzymes. These experiments demonstrated that protein phosphatase-2a could be inhibited by 95% at antibody concentrations that only inhibited protein phosphatase-2b or the catalytic subunit of protein phosphatase-l from rabbit skeletal muscle, or protein phosphatase-2c from rat liver by 10-15% (Fig. 1). Influence of antibody on the different forms of protein phosphatase-2a When liver or muscle extracts are chromatographed on DEAE-cellulose and the fractions assayed for phosphorylase phosphatase activity in the presence of inhibitor-2, two peaks of activity are observed, termed protein phosphatase-2al and 2A2 (Fig. 2). These forms have been suggested to contain the same catalytic subunit (36 kda), based on their comigration on SDS/polyacrylamide gels [5, 13, 141 and their very similar substrate specificities [2, 51. The results in Fig. 2A demonstrate that antibody to the catalytic subunit of protein phosphatase-2a can completely inhibit protein phosphatases- 2A1 and 2A2 from rabbit liver. This further supports the contention that these two phosphatases share the same catalytic subunit. These experiments also indicate that the antigen-

3 53 Fig. 1. Influence of antibody to protein phosphatase-2a on the activities ofproteinphosphatases. Assays were carried out using [32P]phosphorylase kinase (0.7 mg/ml) containing 2.1 mol phosphate/mol apyg unit (1.2 mol/mol in the cc-subunit and 0.9 mol/mol in the 1-subunit). Purified preparations of protein phosphatases-i, 2A, 2B and 2C were incubated in the presence (+) or absence (-) of antibody as described under Materials and Methods. Each enzyme was diluted to achieve 5-10% dephosphorylation of substrate during the final 30-min assay. Protein phosphatase-2b was assayed at 3 pm Ca2+ and protein phosphatase-2c at I0 mm Mg2+ as in [6]. Very similar results were obtained in three separate experiments ic sites on the catalytic subunit remain accessible to the antibody when it is complexed with the 60-kDa subunit in protein phosphatase-2a2, and the 60-kDa and 55-kDa subunits in protein phosphatase-2al [l, 5, 13, 141. Current evidence suggests that protein phosphatase-2a2 may be formed from 2A1 during purification, through the loss of the 55-kDa component [2]. When the fractions from DEAE-cellulose are frozen and thawed in the presence of 1.7% (v/v) 2-mercaptoethanol, the 36-kDa catalytic subunit dissociates from protein phosphatases-2a1 and 2Az and this produces a substantial increase in the phosphorylase phosphatase activity of 2A1 and a smaller increase in 2A2 [2, 13, 141. This freeze-thawing procedure also reveals a new peak of activity eluting before 2A1, resulting from the release of a 36-kDa catalytic subunit from an inactive species termed 2Ao. The substrate specificity of the catalytic subunit derived from 2Ao is very similar to those derived from 2A1 and 2A2 [2]. The elution profile shown in Fig. 2B confirms these results and further demonstrates that the antibody can immunoprecipitate 2Ao. The partially purified catalytic subunits produced by dissociation of protein phosphatases-2ao, 2A1 and 2A2 from rabbit liver are inhibited by similar concentrations of antibody (Fig. 3). These experiments provide further evidence for the close similarity or identity of the catalytic subunits derived from the three forms of protein phosphatase-2a. They also show that the catalytic subunits of protein phosphatase-2a from liver and skeletal muscle are immunologically very similar. When the fractions from DEAE-cellulose are assayed for phosphorylase phosphatase activity in the presence and absence of inhibitor-2, the activity in the presence of inhibitor-2 subtracted from that measured in the absence of this protein can be used to locate protein phosphatase-1 [2] (Fig. 2A). As shown in Fig. 2A, protein phosphatase-1 can also be located by using antibody to inhibit protein phosphatase-2a. This reemphasizes the specificity of the antibody, and illustrates its potential use in such screening procedures. Fig. 2. Resolution of rabbit liver protein phosphatases by chromatography in DEAE-cellulose. The column (12 x 1.9 cm) was equilibrated with 10 mm Tris/CI, ph 7.0 (25"C), 0.1 mm EGTA, 0.1% (v/v) 2- mercaptoethanol (solution A). Following application of the cytosol (35 ml), the column was washed with solution A containing 0.05 M NaCl (170ml). The column was developed with a 300-ml linear gradient of M NaCl in solution A beginning at Fraction 1 ( ). The flow rate was 40 ml/h and fractions of 6 ml were collected. A and B show elution profiles from two different chromatographies. Both chromatographies were repeated with two different liver extracts. (A) Aliquots of each fraction were diluted 10- fold and incubated in the presence or absence of antibody as described under Materials and Methods. The samples were assayed for phosphorylase phosphatase in the presence of 3.O mm MnC12, and in the presence and absence of inhibitor-2 (800 Ujml). (0) - antibody, + inhibitor-2; (+) + antibody, + inhibitor-2; (0) + antibody, - inhibitor-2; (V) activity in the presence of inhibitor-2 subtracted from activity in the absence of inhibitor-2. (B) Fractions were assayed for phosphorylase phosphatase as in A in the presence of inhibitor-2 (800 Ujml) with or without the following freeze-thawing procedure. Samples were made 2% in glycerol, 1.7% (v/v) in 2-mercaptoethanol and 1 mg/ml in bovine serum albumin, frozen for 60 min and allowed to thaw out at 4 C. (+) + antibody, + freeze-thawing; (0) - antibody, + freeze-thawing; (0) - antibody, - freezethawing Influence of antibody on protein phosphatase activities in tissue extracts Based on experiments such as those in Fig. 2A, we have suggested that protein phosphatases-1 and 2A account for virtually all the phosphorylase phosphatase activity in tissue extracts measured under our conditions [3, 71. These suggestions are now established more firmly by the results presented in Fig. 4, which show the effects of inhibitor-2, antibody, or the combination of both proteins on phosphorylase phosphatase activity in dilute tissue extracts. The combined addition of inhibitor-2 and antibody almost completely abolishes activity in skeletal muscle, liver, brain, heart and kidney extracts. The separate effects of inhibitor-2 and antibody are consistent with the relative proportions of protein phosphatases-1 and 2A in these tissues previously assessed by experiments with inhibitor-2 alone [7].

4 54 Fig. 3. Titration of the catalytic subunits derived from protein phosphatases-2ao, 2A1 and 2A2 of rabbit liver with antibody. Fractions corresponding to the activity peaks of protein phosphatases-2ao, 2A1 and 2A2, from a chromatography of liver cytosol on DEAE-cellulose similar to that shown in Fig. 2, were frozen and thawed as described in the legend to Fig. 2. The samples were diluted to 0.75 mu/ml incubated with different concentrations of antibody (plotted on a log scale) as described under Materials and Methods, and assayed for phosphorylase phosphatase activity in the presence of 1.0 mm MnC12 and inhibitor-2 (800 Ujml). (0-0) Purified catalytic subunit of protein phosphatase-2a from rabbit skeletal muscle (2Ac); (0-0) protein phosphatase-2ao; (V- V) protein phosphatase-2al ; (v-v) protein phosphatase-2a2, This experiment was repeated twice with very similar results Fig. 5. Influence of antibody and inhibitor-2 on phosphorylase phosphatase activity in extracts and glycogen pellets prepared from rabbit skeletal muscle andliver. Assays were performed in the presence of 1 mm magnesium acetate. Other conditions and abbreviations are given in the legend to Fig. 4. Two different preparations of the glycogen pellets gave very similar results Fig. 4. Influence of antibody and inhibitor-2 on phosphorylase phosphatase activity in dilute tissue extracts. Experiments were performed as described under Materials and Methods and assays were carried out in the presence of 1.0 mm free MnC12. (C) No addition of antibody or inhibitor-2; (A) + antibody, - inhibitor-2; (I) - antibody, + inhibitor-2; (AI) + antibody, + inhibitor-2. This experiment was repeated five times with essentially identical results The antibody had little effect on phosphorylase phosphatase activity in the glycogen pellet of skeletal muscle or liver, whereas inhibitor-2 could inhibit this activity almost completely (Fig. 5). This supports our previous suggestion [7] that protein phosphatase-1 is bound to this fraction relatively specifically, whereas protein phosphatase-2a is not. It should be noted that the phosphorylase phosphatase activity in liver extracts was inhibited slightly more effectively by antibody and less effectively by inhibitor-2 in Fig. 4, as compared to Fig. 5. This is due to the presence of 1.0 mm MnCI2 in the assays in Fig. 4, which stimulates protein phosphatase-2a and inhibits protein phosphatase-1 [7]. The effects of inhibitor-2 and antibody on glycogen synthase phosphatase activity in dilute skeletal muscle and liver extracts are shown in Fig. 6. These experiments were carried out using rabbit skeletal muscle glycogen synthase phosphorylated in site-2 by phosphorylase kinase or sites (3 a + 3 b + 3 c) by glycogen synthase kinase-3 (collectively termed sites-3). These appear to be the important phosphorylation sites in determining the activity of glycogen synthase in vivo [15, 161. The experiments demonstrate that protein phosphatase-1 and 2A account for virtually all the glycogen synthase phosphatase activity that can be measured in dilute skeletal muscle and liver extracts at 1.O mm + Mg2. The effects of antibody and inhibitor-2 indicated that about 25-30% of the glycogen synthase phosphatase activity in skeletal muscle extracts is accounted for by protein phosphatase-2a, and 65-70% by protein phosphatase-1 (Fig. 6). These values are similar to those reported previously [7]. Protein phosphatase- 2C, the other enzyme that is capable of dephosphorylating site-2 and sites-3 accounts for an extremely small percentage of the glycogen synthase phosphatase activity in dilute skeletal muscle and liver extracts are reported in [7](but see the Discussion in [7, 81). The antibody had little effect on glycogen synthase phosphatase activity in the glycogen pellet (Fig. 6) which veri-

5 55 Fig. 6. Influence of antibody and inhibitor-2 on glycogen synthase phosphatase activity in extracts and glycogen pellet prepared from rabbit skeletal muscle and liver. Assays were performed in the presence of 1 mm magnesium acetate using glycogen synthase phosphorylated in sites-3 (filled bars) or site-2 (open bars). These substrates contained 1.25 and 0.57 mol phosphate/mol subunit respectively, and their concentrations in the assay were 0.07 mg/ml or 0.8 pm. Other conditions and abbreviations are given in the legend to Fig. 4. This experiment was repeated with three different liver and skeletal muscle extracts and three different glycogen synthase preparations, with essentially the same result. In the control incubations the percentage radioactivity released from site-2 in the 30-min assay was 13% (muscle extract), 16% (muscle glycogen fraction), 10% (liver extract) and 15% (liver glycogen fraction). The percentage radioactivity released from sites- 3 and was 5% (muscle extract), 8% (muscle glycogen fraction), 6% (liver extract) and 8% (liver glycogen pellet) Fig. 7. Influence of antibody and inhibitor-2 on protein phosphatase activities in dilute rabbit liver extracts. Assays were performed in the presence of 1 mm magnesium acetate. The 6-phosphofructo-2-kinase/ fructose-2,6-bisphosphatase (PF2K/F2,6Pase), 6-phosphofructo-lkinase (PFIK), fructose-1,6-bisphosphatase (F1,6Pase) and pyruvate kinase (PK) substrates contained 0.6, 0.45, 0.9, and 0.8 mol phosphate/mol subunit, respectively. Other conditions and abbreviations are as in Fig. 4. The concentration of each substrate in the assays was: PF2K/F2,6Pase 0.03 mg/ml, 0.6 pm; PFIK 0.07 mg/ ml, 0.8 pm; F1,6Pase 0.14 mg/ml, 3.5 pm; PK 0.03 mg/ml, 0.5 pm. In the control incubations the percentage radioactivity released in the 30-min assay was 14% (PF2K/F2,6Pase), 9% (PK), 6% (PFIK) and 17% (F1,6Pase). Very similar results were obtained with four different liver extracts for PFIK and PK, and with two liver extracts for the other substrates fies that protein phosphatase-2a is essentially absent from this fraction (see also Fig. 4). Nearly all glycogen synthase phosphatase activity associated with the glycogen pellet of skeletal muscle and liver can be blocked by inhibitor-2 (Fig. 6), as reported previously (71. These observations suggest that the protein phosphatase-1 catalytic subunit is responsible for the dephosphorylation of both glycogen phosphorylase and glycogen synthase in the protein-glycogen complex. However, other workers have suggested that glycogen phosphorylase and glycogen synthase may be dephosphorylated by distinct protein phosphatases in the glycogen fraction [17,18]. Further studies are needed to resolve this controversy. In the preceding paper the nature of the protein phosphatases involved in the control of glycolysis and gluconeogenesis in rat liver was investigated by using 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, 6-phosphofructo-I-kinase, fructose-1, 6-bisphosphatase, pyruvate kinase and phenylalanine hydroxylase as substrates. Based on elution profiles observed after chromatography of rat liver extracts on DEAE-cellulose followed by gel filtration [S], protein phosphatase-2a appeared to account for the majority of the activity towards each of the enzymes. This view is reinforced by the results presented in Fig. 7. When assays were performed at near physiological ph (7.0) and Mg2+ concen- trations (1.0 mm) the antibody inhibited a major proportion of the phosphatase activity in dilute rabbit liver extracts acting on 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, phosphofructo-1 -kinase, fructose-l,6-bisphosphatase and pyruvate kinase. Similar results were obtained with phenylalanine hydroxylase as substrate ( sz 70% inhibition), but the measurements were not considered to be very reliable because of the low activity with this substrate in dilute liver extracts. Since protein phosphatase activities decreased up to 30% during the immunoprecipitation experiments (see Materials and Methods), it could be argued that the decrease represented the loss of one specific phosphatase, for example protein phosphatase-2c, so that the activity remaining after immunoprecipitation underestimates the contribution of this phosphatase. This possibility can be discounted because the loss of phosphorylase phosphatase activity was at least as great as for other substrates, and protein phosphatase-2c does not dephosphorylate this substrate at a significant rate. Furthermore, a similar activity loss was observed with other substrates in the absence of + Mg2, where phosphatase-2c is also inactive. In addition, the proportion of phosphorylase phosphatase activity inhibited by inhibitor-2 was similar at the start and end of the incubations. This suggests that the

6 56 loss of protein phosphatase-1 and protein phosphatase-2a activity during the experiments was similar. The present work represents the first attempt to use antibody to a protein phosphatase as a specific probe of its structure, function and substrate specificity. The results suggest that this approach can provide important information, and could be usefully employed in the study of other protein phosphatases. This work was supported by a Programme Grant and Group Support from the Medical Research Council, London, and by the British Diabetic Association (to P. C.), by the American Diabetes Association (to S.S.) and by the National Institutes of Health Grant AM18270 to (S. J. P). The glycogen phosphorylase, glycogen synthase, phosphorylase kinase, cyclic-amp-dependent protein kinase, glycogen synthase kinase-3, inhibitor-2 and protein phosphatase-2c were purified by Mrs Caroline Woodgett, Mr James Woodgett, Mr Nicholas Tonks and Dr Steven Pelech at the University of Dundee. The 6-phosphofructo-2-kinase/fructose-2,6-bisphos- phatase, 6-phospho-fructo-l-kinase, fructose-1,(i-bisphosphatase and pyruvate kinase were isolated by Dr M. Raafat El-Maghrabi and Dr Fritz Nyfeler at Vanderbilt University. Phenylalanine hydroxylase was a generous gift from Dr Michael Fisher and Professor Christopher Pogson, University of Manchester, UK. REFERENCES 1. Ingebritsen, T. S. & Cohen, P. (1983) Science (Wash. DC) 221, Ingebritsen, T. S. & Cohen, P. (1983) Eur. J. Biochem. 132, Ingebritsen, T. S., Foulkes, J. G. & Cohen, P. (1983) Eur. J. Biochem. 132, Ingebritsen, T. S., Blair, J., Guy, P., Witters, L. & Hardie, D. G. (1983) Eur. J. Biochem. 132, Pato, M., Adelstein, R. S., Crouch, D., Safer, B., Ingebritsen, T. S. & Cohen, P. (1983) Eur. J. Biochem. 132, Stewart, A. A., Ingebritsen, T. S. & Cohen, P. (1983) Eur. J. Biochem. 132, Ingebritsen, T. S., Stewart, A. A. & Cohen, P. (1983) Eur. J. Biochem. 132, Pelech, S., Cohen, P., Fisher, M. J., Pogson, C. I., El-Maghrabi, M. R. & Pilkis, S. J. (1984) Eur. J. Biochem. 145, Lim-Tung, H. Y. & Cohen, P. (1984) Eur. J. Biochem. 145, Lim-Tung, H. Y., Resink, T. J., Hemmings, B. A., Shenolikar, S. & Cohen, P. (1984) Eur. J. Biochem. 138, Cohen, P., Yellowlees, D., Aitken, A., Donella-Deana, A., Hemmings, B. A. & Parker, P. J. (1982) Eur. J. Biochem. 124, Bradford, M. M. (1976) Anal. Biochem. 72, Tamura, S., Kikuchi, H., Kikuchi, K., Hiraga, A. & Tsuiki, S. (1980) Eur. J. Biochem. 104, Tamura, S. & Tsuiki, S. (1980) Eur. J. Biochem. 114, Parker, P. J., Embi, N., Caudwell, F. B. & Cohen, P. (1982) Eur. J. Biochem. 124, Parker, P. J., Caudwell, F. B. & Cohen, P. (1983) Eur. J. Biochem. 130, Mvumbi, L., Dopere, F. & Stalmans, W. (1983) Biochem. J. 212, Gilboe, D. P. & Nuttall, F. Q. (1984) Arch. Biophys. 228, S. Alemany, H. Y. L. Tung and P. Cohen, Department of Biochemistry, University of Dundee, Dundee, Tayside, Scotland DD1 4HN S. Shenolikar, Department of Pharmacology, University of South Alabama, Medical Sciences Building, Mobile, Alabama, USA S. J. Pilkis, Department of Physiology, Vanderbilt University School of Medicine, Nashville, Tennessee, USA 37232