Dehydrogenase Phosphatase JIANGONG YAN, JANET E

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 4953-4956, May 1996 Biochemistry

Role of the regulatory subunit of bovine pyruvate dehydrogenase phosphatase JIANGONG YAN, JANET E. LAWSON, AND LESTER J. REED Biochemical Institute and Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, TX 78712 Contributed by Lester J. Reed, January 16, 1996

ABSTRACT Bovine pyruvate dehydrogenase phospha- was expressed in Escherichia coli and purified to near homo- tase (PDP) is a Mg2+-dependent and Ca2+-stimulated het- geneity (3). EB and E2 were obtained by resolution of the PDC erodimer that is a member of the protein phosphatase 2C (11). The E2 contained small amounts of tightly bound E3- family and is localized to mitochondria. Insight into the binding protein (protein X) and PDH kinase (E2-X-K com- function of the regulatory subunit of PDP (PDPr) has been plex). The peptides RRASVA and RRATVA were prepared gained. It decreases the sensitivity of the catalytic subunit of manually by solid-phase synthesis using Boc-Ala-Pam resin PDP (PDPc) to Mg2+. The apparent Km of PDPc for Mg2+ is and Boc-protected amino acids (Peptides International). The increased about 5-fold, from about 0.35 mM to 1.6 mM. The peptides were characterized by amino acid composition anal- polyamine spermine increases the sensitivity of PDP but not ysis and by amino acid sequence determination. Protein kinase PDPc to Mg2+, apparently by interacting with PDPr. PDPc but A catalytic subunit was obtained from Sigma and [y-32P]ATP not PDP can use the phosphopeptide RRAT(P)VA as a sub- was from New England Nuclear. strate. These observations are interpreted to indicate that Phosphorylated Substrates. 32P-labeled inactive PDC (P- PDPr blocks or distorts the active site of PDPc and that PDC) was prepared by incubating for 30 min at 30°C a solution spermine produces a conformational change in PDPr that containing 10-20 mg of PDC in buffer A [25 mM 3-(N- reverses its inhibitory effect. These findings suggest that PDPr mM may be involved in the insulin-induced activation of the morpholino)propanesulfonate (Mops) buffer, pH 7.3/2.5 mitochondrial PDP in adipose tissue, which is characterized MgCl2/0.1% 2-mercaptoethanol/10% glycerol] and 0.1 mM by a decrease in its apparent Km for ['y-32P]ATP (106 cpm/nmol) in a final volume of 2 ml. The pH Mg2+. was adjusted to -5.3 to precipitate the PDC and to remove a trace amount of PDP. The precipitate was dissolved in a small Pyruvate dehydrogenase phosphatase (PDP) is a mitochon- volume of buffer A and the solution was passed through a drial protein serine/threonine phosphatase that catalyzes the Sephadex G-50 (superfine) column (1.5 x 10 cm) equilibrated dephosphorylation and concomitant reactivation of the pyru- with buffer A to remove the vate dehydrogenase (PDH or El) component of the PDH radioactive ATP. The phosphor- multienzyme complex (PDC) (1, 2). A catalytic subunit of PDP ylated complex contained 3-8 nmol of phosphoryl groups per (PDPc) is a member of the protein phosphatase 2C family (3). mg of protein. PDP is a heterodimer consisting of a Mg2+-dependent and 32P-labeled P-El was prepared by incubating a solution Ca2+-stimulated PDPc of Mr 52,600 and a flavoprotein (that containing 10 mg of E1, 50 ,ug of E2-X-K complex, and 0.1 mM we will call PDPr) ofMr 95,800 with hitherto unknown function [,y-32P]ATP in 2 ml of buffer A for 1 h at 30°C. The solution (2, 4, 5). The apparent Km of PDP for Mg2> varies with the was adjusted to pH 5.8 to precipitate P-El, which was dissolved assay buffer from about 1.5 to 5 mM. The apparent Km for Ca>2 in a small volume of buffer A, the solution was dialyzed against is - 1 ,uM (5, 6). three changes of buffer A, and then centrifuged at 35,000 rpm Both phosphorylated El (P-E1) and PDP must be bound to and 4°C for 1-5 h in a Beckman model Optima TLX ultra- the 60-mer icosahedral dihydrolipoamide acetyltransferase centrifuge. The P-El contained 9-12 nmol of phosphate (E2) component of PDC to obtain a maximum rate of dephos- groups per mg of protein. P-El was also prepared by resolution phorylation. Ca>2 apparently mediates the specific binding of of P-PDC (11). PDP to E2 in juxtaposition to P-El (7). This orientation The peptides RRASVA and RRATVA were phosphory- decreases the apparent Km of PDP for P-El (7) and for Mg2> lated by incubating for 8-10 h at 30°C a solution containing 10 (6). There appears to be two Ca2+-binding sites, one intrinsic mg peptide, 25 mM Tris HCl (pH 7.3), 2.5 mM MgCl2, 0.1% to PDPc and a second produced by association of PDP with E2 2-mercaptoethanol, 10% glycerol, 0.1 mM [,y-32P]ATP, and 40 (2). units of protein kinase A catalytic subunit in a final volume of At subsaturating concentrations of Mg2>, PDP activity is 2 ml (12). The reaction was terminated by adding 2 ml of 60% stimulated by polyamines, particularly spermine (8). Like acetic acid, and the solution was passed through an anion- Ca2+, spermine decreases the apparent Km of PDP for Mg2+ exchange column (1.0 x 4 cm) of Dowex 2X8-100 equilibrated but by a different but hitherto unknown mechanism (6, 8, 9). with 30% acetic acid. The flowthrough, which contained the This paper reports that PDPr decreases the sensitivity of PDPc phosphopeptide, was lyophilized. The peptide was dissolved in to Mg2+ and that this effect is reversed by spermine, which 1 ml of 25 mM Tris-HCl (pH 7.3), 0.1% 2-mercaptoethanol, apparently interacts with PDPr. These observations are po- and the pH was adjusted to 7.3 with 1 M Tris (pH 10). tentially significant in understanding the molecular basis of the PDP Assay. PDP or PDPc was diluted in buffer A containing insulin-induced activation of the mitochondrial PDP. 1 mg/ml of BSA and an aliquot was preincubated at 30°C for 2 min with 10 mM MgCl2 and 0.1 mM CaCl2 (except where MATERIALS AND METHODS noted). The reaction was initiated by adding an amount of P-PDC or P-El containing -0.1 nmol of 32P-labeled phos- Materials. Highly purified PDC and PDP were prepared phate groups in a final volume of 40 ,ul. The reaction was from bovine kidney mitochondria (10, 11). Recombinant PDPc Abbreviations: PDH or E1, pyruvate dehydrogenase; PDC, PDH The publication costs of this article were defrayed in part by page charge complex; P-E1, phosphorylated E1; P-PDC, phosphorylated PDC; E2, payment. This article must therefore be hereby marked "advertisement" in dihydrolipoamide acetyltransferase; PDP, pyruvate dehydrogenase accordance with 18 U.S.C. §1734 solely to indicate this fact. phosphatase; PDPc, catalytic subunit; PDPr, regulatory subunit. 4953 Downloaded by guest on October 1, 2021 4954 Biochemistry: Yan et al. Proc. Natl. Acad. Sci. USA 93 (1996)

terminated after 90 s by adding 200 ,ul of 20% trichloroacetic 60 - acid. The mixture was centrifuged at 12,000 rpm for 2 min in an Eppendorf microcentrifuge, and a 200-,l aliquot of the supernatant fluid was transferred into 1 ml of scintillation fluid 50 - (Bio-Safe II, Research Products International) and counted. One unit is defined as the amount of phosphatase that releases 1 nmol of 32p; per min. Protein was determined as described by -40 Bradford (13). With the phosphopeptide substrates, assay conditions were as described above, except that 25 mM Tris HCI buffer was used instead of Mops buffer and 32p; was determined as the phosphomolybdic acid complex (14). 30 Ca2+-Dependent Binding of PDP and PDPc to E2. Solutions of PDP and PDPc were centrifuged at 35,000 rpm for 30 min at 4°C before use. Solutions containing 20 jig of PDP or PDPc, 20- 200 jig of E2, or 550 ,ug of PDC and 0.2 mM Ca2+ or 0.2 mM EGTA in 1 ml of buffer A were incubated on ice for 30 min. The mixtures were centrifuged at 35,000 rpm and 4°C for 2.5 10 - h in the TLS55 rotor of a Beckman model Optima TLX ultracentrifuge to separate unbound PDP and PDPc from E complexed PDP and PDPc (bound to E2). The supernatant 0 fluids were removed, the pellets were washed once with buffer E A, covered with 50 Al of buffer A, and allowed to stand for 4-6 140- h with occasional gentle shaking to dissolve the pellets. Ali- quots of the supernatant fluids and the redissolved pellets were 120- assayed for PDP activity and were subjected to SDS/PAGE. 80- /

RESULTS 100 Sensitivity of PDP and PDPc to Mg2+. With P-PDC as substrate and in the presence of a saturating concentration of 80 Ca2+, PDPc was more sensitive than PDP to Mg2+ (Fig. 1A). PDPr appareptly decreases the sensitivity of PDPc to Mg2+. 60- The Km values of PDPc and PDP for Mg2+ were about 0.35 mM and 1.6 mM, respectively. With uncomplexed P-E1 and uncomplexed PDP or PDPc (E2 absent), the Km values of PDPc 40- and PDP for Mg2+ were both -1.8 mM. Effect of Ca2+ on Sensitivity of PDP and PDPc to Mg2+. With P-El complexed with E2 (i.e., P-PDC) as substrate, but not with uncomplexed P-E1, and in the presence of a saturating 20 concentration of Mg2+, Ca2+ stimulates the activity of PDP about 10-fold. Ca2+ decreases the Km of PDP for complexed P-E1 (7) and for Mg2+ (6). As shown in Fig. 1, Ca2+ increased 0 2 4 6 8 10 12 the sensitivity of both PDP and PDPc to Mg2+. In the absence j/[Mg2+] (mMMyl of Ca2+, the Km of PDP and PDPc for Mg2+ increased to about 3.5 and 2.6 mM, respectively. FIG. 1. Effect of Ca2+ on the sensitivity of PDP (-) and PDPc (m) Ca2+-Dependent Binding of PDP and PDPc to E2. Mixtures to Mg2+. (A) The assay mixtures contained 0.5 ,ug of PDPc or an of PDP or PDPc and an excess of E2 with Ca2+ (200 ,uM) or equivalent amount (1.5 j.g) of PDP, 0.1 mM CaCl2, the specified without Ca2+ (200 ,uM EGTA) were centrifuged at 35,000 rpm concentrations of MgCl2, and P-PDC containing 0.14 nmol of 32p- for 2.5 h in a Beckman model TLS55 swinging-bucket rotor to labeled protein-bound phosphate groups in 40 ,ul of buffer A. Assay conditions were as described. (B) The assay mixtures contained 0.1 separate the large PDP-E2 and PDPc-E2 complexes from mM EGTA instead of Ca2+. Other components and conditions were unbound PDP and PDPc. The pellet and supernatant fractions as inA. Each point represents the mean of values obtained from three were assayed for PDH phosphatase activity and were subjected separate experiments. to SDS/PAGE (data not shown). The results (Table 1) show that both PDP and PDPc bound to E2 in the presence, but not in the absence, of Ca2+. Similar results were obtained with Comparison ofSubstrate Specificities ofPDP and PDPc. Km mixtures of the PDC and PDP or PDPc (data not shown). and kcat values of PDP and PDPc for several substrates are Effect of Spermine on Sensitivity of PDP and PDPc to Mg2+. presented in Table 2. At saturating Mg2+ and Ca2+ concen- Polyamines, of which spermine is the most effective, stimulate trations, P-PDC is a better substrate for PDP than for PDPc. the activity of PDP at subsaturating Mg2+ concentrations by The apparent Km values were 0.75 and 2.3 ,uM, respectively. decreasing its Km for Mg2+ (6, 8, 9). With P-PDC as substrate Although the Km values of PDP and PDPc for uncomplexed and in the presence of a saturating concentration of Ca2 , 0.5 P-E1 (E2 absent) were appreciably higher than for complexed mM spermine increased the sensitivity of PDP to Mg2+, but P-E1, the kcat values were essentially the same. Although the had little, if any, effect, on the sensitivity of PDPc to Mg2+ (Fig. phosphopeptide RRAT(P)VA was comparable with uncom- 2). The apparent Km of PDP for Mg2+ was decreased about plexed P-EB as a substrate for PDPc, it was ineffective with 5-fold by spermine. With uncomplexed P-E1 as substrate (E2 PDP. This observation suggests that PDPr blocks access of the absent), spermine had little, if any, effect on the sensitivity of phosphopeptide to the active site of PDPc. Attempts to elicit PDP to Mg2+ (Fig. 2). These findings indicate that spermine PDP activity toward the phosphopeptide with E2 and spermine interacts with PDPr and that both PDP and P-E1 must be were unsuccessful. The phosphothreonyl hexapeptide was a bound to E2 to observe spermine stimulation of PDP activity. better substrate than the phosphoseryl hexapeptide, as shown Downloaded by guest on October 1, 2021 Biochemistry: Yan et al. Proc. Natl. Acad. Sci. USA 93 (1996) 4955 Table 1. Ca2+-dependent binding of PDP and PDPc to E2 Table 2. Kinetic parameters for PDP and PDPc Phosphatase PDP PDPc activity Substrate Kmi, ,M kcat, min- Kmi, ,M kcat, min- PDP PDPc P-PDC 0.75 65 2.3 65 With Ca2+ P-E1 60 65 23 65 Supernatant D.A.* D.A. RRAT(P)VA N.D.* 18 7 Pellet 4.0 3.5 RRAS(P)VA N.D.* 53 7 Without Ca2+ The Km values are based on the concentrations of 32P-labeled Supernatant 3.9 3.7 phosphoryl groups in the substrates. Pellet D.A. D.A. *No detectable activity. D.A., detectable activity. *D.A. (-I%). molecule between Ca2+-mediated binding sites on E2 so that it can dephosphorylate more than one P-El molecule. In previously with rabbit skeletal muscle protein phosphatase 2C2 contrast to Ca>, spermine apparently interacts with the PDPr (12). subunit of PDP, possibly producing a conformational change that improves access of Mg2+ to its binding site (i.e., the active site) on PDPc. Consistent with these proposals are the obser- DISCUSSION vations that (i) the Ca>2 stimulation is not observed with Comparison of the properties of the native PDP heterodimer uncomplexed P-El (E2 absent) (7), or with other phospho- and recombinant PDPc has provided insight into the function protein (C. H. MacGowan, P. Cohen, Z. Damuni, and L.J.R., of PDPr. The major regulators of PDP activity are Mg2, Ca>, unpublished data) or phosphopeptide (15) substrates, whether and, apparently, an allosteric effector that is mimicked by or not E2 is present; and (ii) stimulation of PDP activity by spermine. PDP is inactive in the absence of Mg2+ and its spermine at subsaturating Mg2> concentrations is not ob- activity toward P-E, complexed with E2 is stimulated about served with PDPc or with uncomplexed P-El. 10-fold by Ca2+ in the presence of Mg2+ (5, 7). Both Ca2> and It appears that PDPr may participate in the Ca2+-mediated spermine increase the sensitivity of PDP to Mg2> (i.e., de- binding and/or orientation of PDPc on E2. At saturating Mg2+ crease the Km of PDP for Mg2+) but by different mechanisms and Ca2+ concentrations, the apparent Km of PDPc for P-El (6, 8). Ca2+ also decreases about 20-fold the apparent Km of complexed with E2 is about 3-fold larger than the apparent Km PDP for P-El complexed with E2 (7). Ca2+ apparently medi- of PDP for complexed P-EB. Consistent with this finding is the ates the specific binding of PDP (or PDPc) to the 60-mer E2 observation that PDPc binds less tightly than PDP to an in juxtaposition to the bound P-El, resulting in an intramo- E2-agarose affinity column in the presence of Ca2+ (data not lecular dephosphorylation (Fig. 3). In the absence of Ca>, shown). This procedure is used in the purification of PDP (2, PDP does not bind to E2 so that the dephosphorylation is 4). intermolecular and less efficient. Although not indicated in These findings suggest that PDPr may be involved in an Fig. 3, a mechanism must exist to permit movement of a PDP insulin-induced signaling pathway leading to stimulation of the activity of the mitochondrial PDC, particularly in adipose 100 tissue. A 2- to 3-fold increase in activity occurs within a few

C 60 a - E C: 40

MP DP 20 2* D -- .1I...Q l... + 0 0 5 10 15 20 25 FIG. 3. Diagrammatic representation of the role of Ca2+ in the regulation of PDP activity. In the presence of Ca>, the PDP het- 1/ [Mg2+] (mM)- erodimer (PDPc-PDPr) or PDPc binds to E2 in juxtaposition to the bound P-El so that the dephosphorylation is intramolecular. In the FIG. 2. Effect of spermine on sensitivity of PDP and PDPc to absence of Ca>, PDP is not bound to E2, therefore, the dephosphor- Mg2+. Assay mixtures containing 0.5 jig of PDPc (o) or 1.5 jig of PDP ylation is intermolecular; the rate is limited by diffusion-collision and (0), 0.1 mM CaCl2, and variable concentrations of MgCl2 in 30 ,ud of decreases to a value that is comparable with that observed with buffer A were incubated, respectively, in the absence (o, 0) or uncomplexed P-EB. It is not yet clear whether Ca2+ binding to PDP presence (-, *) of 0.5 mM spermine for 3 min at 30°C. A 10-jil aliquot (or PDPc) produces conformational changes that promote specific of P-PDC containing 0.1 nmol of 32P-labeled phosphate groups was attachment of PDP (or PDPc) to E2 or whether Ca2+ participates added to start the reaction. Assay conditions were as described. In a directly in the attachment. PDPr decreases the sensitivity of PDPc to parallel experiment, the assay mixtures contained 1.5 jig of PDP and Mg>, i.e., raises the Km for Mg2>, apparently by distoring or blocking 0.1 nmol of P-EB instead of P-PDC without (A) and with (A) 0.5 mM the Mg2+-binding site (i.e., the active site). El is a heterotetramer spermine. Each point represents the mean of values from four separate (a2032) that undergoes phosphorylation on its a subunits. ,B subunits experiments. anchor the a subunits to the E2 core. Downloaded by guest on October 1, 2021 4956 Biochemistry: Yan et al. Proc. Natl. Acad. Sci. USA 93 (1996) minutes of exposure of rat adipose tissue to insulin, is accom- 4. Pratt, M. L., Maher, J. F. & Roche, T. E. (1982) Eur. J. Biochem. panied by a net dephosphorylation of complexed P-E1, and 125, 349-355. results in stimulation of PDP activity (16). This latter stimu- 5. Denton, R. M., Randle, P. J. & Martin, B. R. (1972) Biochem. J. lation is characterized by a decrease in the Km of the phos- 128, 161-163. phatase for Mg2+, but apparently by a Ca2+-independent 6. Thomas, A. P., Diggle, T. A. & Denton, R. M. (1986) Biochem. mechanism (6, 8, 9). That spermine apparently mimics the J. 238, 83-91. action of insulin has been noted (6, 8). The recent observation 7. Pettit, F. H., Roche, T. E. & Reed, L. J. (1972) Biochem. Biophys. that the insulin-induced stimulation of PDC activity in rat fat Res. Commun. 49, 563-571. cells is insensitive to wortmannin suggests the existence of a 8. Damuni, Z., Humphreys, J. S. & Reed, L. J. (1984) Biochem. distinct signaling pathway leading to the activation of PDP Biophys. Res. Commun. 24, 95-99. (17). Several investigators have proposed that the insulin- 9. Rahmatullah, M. & Roche, T. E. (1988) J. Bio. Chem. 263, induced activation of the mitochondrial PDP is mediated by a 8106-8110. putative inositol phosphate glycan produced from a glycolipid 10. Pettit, F. H. & Reed, L. J. (1982) Methods Enzymol. 89,376-386. precursor(s) in the plasma membrane by a phosphatidylinos- 11. Linn, T. C., Pelley, J. W., Pettit, F. H., Hucho, F., Randall, D. D. itol-specific phospholipase C (18, 19). Our findings suggest & Reed, L. J. (1972) Arch. Biochem. Biophys. 148, 327-342. that spermine may mimic an insulin "mediator" (allosteric 12. Donella-Deanna, A., MacGowan, C. H., Cohen, P., Marchiori, F., Meyer, H. E. & Pinna, L. A. (1990) Biochim. Biophys. Acta 1051, effector) that by interacting with PDPr increases the sensitivity 199-202. of PDP to Mg2+. At a mitochondrial free Mg2+ concentration 13. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. of -0.4 mM (20), a decrease in the Km for Mg2+ would cause 14. Wollny, E., Watkins, K., Kramer, G. & Hardesty, B. (1984) J. a marked stimulation of PDP activity. The recent cloning of Biol. Chem. 259, 2484-2492. cDNA encoding PDPr (unpublished data) should facilitate 15. Davis, P. F., Pettit, F. H. & Reed, L. J. (1977) Biochem. Biophys. further studies on its function. Res. Commun. 75, 541-549. 16. Denton, R. M., Midgley, P. J. W., Rutter, G. A., Thomas, A. P. We thank Drs. Thomas Roche, Richard Denton, and Marvin & McCormack, J. G. (1989) Ann. N.Y Acad. Sci. 573, 285-296. Hackert for helpful discussions. This work was supported by U.S. Public Health Service Grant GM-06590 and by a grant from the 17. Moule, S. K., Edgell, N. J., Welsh, G. I., Diggle, T. A., Foulstone, Foundation for Research. E. J., Heesom, K. J., Proud, C. .G. & Denton, R. M. (1995) Biochem. J. 311, 595-601. 1. Linn, T. C., Pettit, F. H. & Reed, L. J. (1969) Proc. Natl. Acad. 18. Saltiel, A. R. (1987) Endocrinology 120, 967-972. Sci. USA 62, 234-241. 19. Larner, J., Huang, L. C., Suzuki, S., Tang, G., Zhang, C., 2. Teague, W. M., Pettit, F. H., Wu, T.-L., Silberman, S. R. & Reed, Schwartz, C. F. W., Romero, G., Luttrell, L. & Kennington, A. S. L. J. (1982) Biochemistry 21, 5585-5592. (1989) Ann. N.Y Acad. Sci. 573, 297-305. 3. Lawson, J. E., Niu, X.-D., Browning, K. S., Le Trong, H., Yan, 20. Corkey, B. E., Duszynski, J., Rich, T. L., Matschinsky, B. & J. & Reed, L. J. (1993) Biochemistry 32, 8987-8993. Williamson, J. R. (1986) J. Biol. Chem. 261, 2567-2574. Downloaded by guest on October 1, 2021