(Received for publication, March 27, 1997)
From the Hormone and Metabolic Research Unit,
Institute of Cellular and Molecular Pathology and the Louvain
University Medical School, Avenue Hippocrate, 75, 1200 Brussels,
Belgium and the ¶ Medical Research Council Protein Phosphorylation
Unit, Department of Biochemistry, University of Dundee,
Dundee, DD1 4HN, Scotland
To understand the insulin-induced activation of 6-phosphofructo-2-kinase (PFK-2) of the bifunctional enzyme PFK-2/fructose-2,6-bisphosphatase in heart, the effect of phosphorylation by protein kinases of the insulin signaling pathways on PFK-2 activity was studied. Purified PFK-2/fructose-2,6-bisphosphatase from bovine heart is a mixture of two isoforms (Mr 58,000 and 54,000 on SDS-polyacrylamide gels). The Mr 54,000 protein is an alternatively spliced form, lacking phosphorylation sites for protein kinases. Recombinant enzymes corresponding to the Mr 58,000 (BH1) and Mr 54,000 (BH3) forms were expressed and used as substrates for phosphorylation. The recombinant BH1 isoform was phosphorylated by p70 ribosomal S6 kinase (p70s6k), mitogen-activated protein kinase-activated protein kinase-1, and protein kinase B (PKB), whereas the recombinant BH3 isoform was a poor substrate for these protein kinases. Treatment with all protein kinases activated PFK-2 in the recombinant BH1 preparation. Phosphorylation of the recombinant BH1 isoform correlated with PFK-2 activation and was reversed by treatment with protein phosphatase 2A. All the protein kinases phosphorylated Ser-466 and Ser-483 in the BH1 isoform, but to different extents: p70s6k preferentially phosphorylated Ser-466, whereas mitogen-activated protein kinase-activated protein kinase-1 and PKB phosphorylated Ser-466 and Ser-483 to a similar extent. We propose that PKB is part of the insulin signaling cascade for PFK-2 activation in heart.
Fructose 2,6-bisphosphate (Fru-2,6-P2)1 participates in the regulation of glycolysis in liver, heart, and other mammalian tissues by controlling the activity of 6-phosphofructo-1-kinase (1). For example, in perfused rat hearts, glycolysis and Fru-2,6-P2 concentrations were increased in parallel after stimulation by adrenalin (2) and insulin (3) or by increasing the work load (4). Both adrenalin and the work load increased Fru-2,6-P2 by activating 6-phosphofructo-2-kinase (PFK-2), the enzyme responsible for Fru-2,6-P2 synthesis. This activation resulted from phosphorylation by cyclic AMP-dependent protein kinase (PKA) or by the multifunctional Ca2+/calmodulin-dependent kinase II, probably at the same sites (4). Insulin treatment also activated PFK-2 both in rat heart in vivo (3) and in isolated rat cardiomyocytes (5). To clarify the mechanism of this insulin-induced activation of PFK-2, we have studied the effects of the in vitro phosphorylation of heart PFK-2 by protein kinases of the insulin signaling cascades.
On hormone binding, the activation of the insulin receptor causes autophosphorylation at several tyrosine residues. This leads to the docking of the insulin receptor substrate-1 and -2 and their phosphorylation at multiple tyrosine residues by the receptor. Adaptor proteins containing SH2 domains are recruited, which in turn can lead to the activation of at least two signaling pathways. One of these leads to the activation of the mitogen-activated protein kinase (MAPK) cascade, via the adaptor proteins Grb2/Sos and involving the activation of Ras and Raf (6). As a consequence, the mitogen-activated protein kinase-activated protein kinase-1 (MAPKAP kinase-1) (also called p90 ribosomal S6 kinase, RSK-2, or p90rsk (7)) becomes activated. Another pathway involves activation of the lipid kinase, phosphatidylinositol 3-kinase, through recruitment of its p85 regulatory subunit and the downstream activation of p70 ribosomal S6 kinase (p70s6k) (8). Finally, a third pathway has recently been described, which also involves phosphatidylinositol 3-kinase activation (9) and which results in the activation of protein kinase B (PKB) (also known as Akt/RAC (11, 12)) (10).
Heart PFK-2/fructose-2,6-bisphosphatase (FBPase-2) differs from the other isozymes of the bifunctional enzyme (13). The various PFK-2/FBPase-2 isozymes are homodimers and are known as the liver, skeletal muscle, heart, testis, and brain isozymes, and for most, the subunit molecular weight varies between 50,000 and 60,000. The rat gene for the heart isozyme contains 16 exons, one of which (exon 15) codes for a stretch of 64 amino acids containing phosphorylation sites for PKA and protein kinase C (14). In bovine heart, two alternatively spliced forms have been described (15, 16), migrating with Mr values of 58,000 and 54,000 on SDS-polyacrylamide gels (15). Peptide sequence analysis revealed that the Mr 54,000 form lacks phosphorylation sites for PKA (Ser-466 and Ser-483) and protein kinase C (Ser-466 and Thr-475), which are encoded by exon 15 and thus present in the Mr 58,000 form (15, 17). The situation in bovine heart is further complicated by the fact that four different mRNAs have been characterized (18). One (B3) lacks exon 15 and is therefore similar to the Mr 54,000 form. The other three (B1, B2, and B4) contain exon 15 and could correspond to the Mr 58,000 form. They differ from each other by minor differences located in domains devoid of any regulatory or catalytic activities (18). Similarly, in rat heart, four mRNA species have been found (19), but information on the existence of the corresponding proteins is lacking since the heart PFK-2/FBPase-2 protein has not been purified for sequencing from this species.
In this work, we have studied the effects of phosphorylation by several protein kinases on PFK-2 activity of the heart isozyme. With the native bovine heart preparations of PFK-2/FBPase-2 as substrate, the effect of phosphorylation on PFK-2 activity could be partially masked because the preparations contain about twice as much of the short Mr 54,000 form as the complete Mr 58,000 form (15). Therefore, we mainly investigated the effects of phosphorylation on PFK-2 activity of the recombinant BH1 and BH3 forms. The former is a complete isoform and corresponds to the Mr 58,000 form, whereas the latter lacks the sequence encoded by exon 15 and thus corresponds to the short Mr 54,000 form. The protein kinases tested were p70s6k, MAPKAP kinase-1, and PKB. The effects of phosphorylation on PFK-2 activity were compared with those induced by PKA.
Radiochemicals were from Amersham Corp. Activated
MAPKAP kinase-1 (350 units/ml), from rabbit skeletal muscle (20), and activated p70s6k (14 units/ml), from the livers of
cycloheximide-treated rats (21), were purified as described. Activated
PKB (0.64 units/ml) was purified from insulin-like growth
factor-1-stimulated 293 cells overexpressing the protein (22).
Glutathione was removed from the PKB preparation by gel filtration on a
Superose 12 column (Pharmacia Biotech Inc.) in buffer containing 20 mM MOPS, pH 7, 25 mM KCl, 0.1 mM
EDTA, 5% (v/v) glycerol, 0.1% (v/v) 2-mercaptoethanol, and 0.005%
(w/v) Brij 35. The catalytic subunits of PKA (800 units/ml) (23) and
protein phosphatase 2A (PP2A; 11,000 units/ml) (24) were purified from
bovine heart as indicated. Activated MAPK (11 units/ml) was kindly
donated by Dr. J. Goris (Katholieke Universiteit Leuven, Leuven,
Belgium). Native bovine heart PFK-2/FBPase-2 was purified from
slaughterhouse tissue (15). The inhibitor peptide of PKA (PKI) (25) and
other peptides were synthesized by Drs. J. Lucchetti and V. Stroobant
(Ludwig Institute, Brussels, Belgium). Ni2+/nitrilotriacetic acid-agarose gel was obtained from
QIAGEN Inc. All other biochemicals were from Sigma or Boehringer
Mannheim.
The B1 and B3 PFK-2/FBPase-2 cDNAs were cloned in pBluescript II KS+ phagemid (called BH1 and BH3 here) and introduced into the expression vector pET3a (18). The C-terminally polyhistidine-tagged form of BH1, called BH1(His)6, was constructed from the single-stranded form of the phagemid containing the BH1 cDNA following the same procedure as described for the liver isozyme (26). The cDNA encoding BH1(His)6 was then introduced into the expression vector pET3a as described (18).
Expression and PurificationThe BH1 and BH3 isoforms of
bovine heart PFK-2/FBPase-2 were expressed and purified as described
(26). Following concentration by ultrafiltration, the preparations were
dialyzed against 200 volumes of a buffer containing 20 mM
HEPES, pH 7.5, 50 mM KCl, 5 mM
MgCl2, 0.5 mM EDTA, 0.1 mM EGTA, 1 mM potassium phosphate, 15 mM
2-mercaptoethanol, 20% (v/v) glycerol, and 0.5 µg/ml leupeptin and
stored at 80 °C.
BH1(His)6 was expressed in Escherichia coli
strain BL21(DE3) pLysE. Culture and lysis were as described above.
Purification on Ni2+/nitrilotriacetic acid-agarose was
carried out as described (26), and the preparations were stored at
80 °C.
For measurement of the changes in kinetic properties induced by phosphorylation, preparations of bovine heart PFK-2/FBPase-2 were incubated with 1 mM MgATP at 30 °C in phosphorylation buffer containing 10 mM MOPS, pH 7, 0.5 mM EDTA, 10 mM magnesium acetate, 0.1% (v/v) 2-mercaptoethanol, 5 µM PKI (except where PKA was studied), and each protein kinase as indicated in the figure and table legends. After 60 min, the reaction was stopped by 10-fold dilution in 20 mM HEPES, pH 7.5, 50 mM KCl, 0.5 mM EGTA, 5 mM EDTA, 1 mM potassium phosphate, 20% (v/v) glycerol, and 0.1% (v/v) 2-mercaptoethanol (stop buffer) and chilled in ice. Aliquots were taken for PFK-2 or FBPase-2 assay. For each protein kinase, phosphorylation had reached a maximum when the kinetic studies were carried out.
For measurements of 32P incorporation, the bovine heart
PFK-2/FBPase-2 preparations were incubated as described above with 0.1 mM [-32P]MgATP (specific radioactivity of
250 cpm/pmol) and each protein kinase as indicated in the figure
legends. The MAPKAP kinase-1 and PKA preparations were diluted prior to
use in a buffer containing 20 mM MOPS, pH 7.0, 0.1 mM EDTA, 0.01% (w/v) Brij 35, and 0.1% (v/v)
2-mercaptoethanol. The reactions were terminated by diluting aliquots
(5 µl) with 15 µl of stop buffer before boiling for 1 min with 5 µl of 5% (w/v) SDS, 20% (v/v) glycerol, 0.2% (w/v) bromphenol
blue, 100 mM dithiothreitol, and 65 mM
Tris-HCl, pH 6.8 (sample buffer), for SDS-PAGE (27) on 12.5% (w/v)
acrylamide minigels. To determine 32P incorporation, gels
were stained with Coomassie Blue and dried, and bands corresponding to
PFK-2/FBPase-2 were counted in a Hewlett-Packard Instant Imager
together with spotted dried aliquots of the diluted (1:500) stock
solution of [
-32P]MgATP used in the phosphorylation
experiments. Stoichiometries of 32P incorporation (mol/mol
of subunit) were calculated from the amount of protein loaded onto the
gel as quantified by the ninhydrin method (see below), and the
molecular weights of the PFK-2/FBPase-2 subunits which were taken as
61,520 for BH1(His)6 and 53,909 for BH3.
Synthetic peptides were incubated
in 0.1 ml of phosphorylation buffer with 0.1 mM
[-32P]MgATP (specific radioactivity of 200-1000
cpm/pmol) and each protein kinase as indicated in the legend to Table
IV. Aliquots of the reaction mixture (10 µl) were removed between 2 and 10 min for the measurement of 32P incorporation
(28).
|
Recombinant BH1(His)6
was first phosphorylated by 0.1 mM nonradioactive or
[-32P]MgATP in phosphorylation buffer, in which the
magnesium acetate and EDTA concentrations were reduced to 1 and 0.1 mM, respectively, and protein kinase as indicated in the
figure legends. After 60 min, the reaction was stopped by adding an
excess of EDTA (10 mM), followed by the indicated amount of
PP2A. Aliquots were removed at the indicated times and diluted 3- or
10-fold in stop buffer, in which the KCl was replaced by 50 mM KF, for the measurement of 32P incorporation
by autoradiography (see above) or PFK-2 activity, respectively.
PFK-2 and FBPase-2 activities were measured
(29) under the conditions described in the figure and table legends.
One unit of PFK-2 or FBPase-2 activity corresponds to the formation of 1 µmol of product/min. The protein kinases were assayed by
32P incorporation from [-32P]ATP into
peptide or protein substrates (28). These included a peptide related to
the C terminus of ribosomal protein S6 for p70s6k and
MAPKAP kinase-1 (30), the glycogen synthase kinase peptide GRPRTSSFAEG
for PKB (31), and histone IIA (1.25 mg/ml) or myelin basic protein (0.5 mg/ml) for the catalytic subunit of PKA and MAPK, respectively. PP2A
was assayed with 4-nitrophenyl phosphate as substrate (24). One unit of
protein kinase or protein phosphatase activity is the amount that
catalyzes the (de)phosphorylation of 1 nmol of substrate/min.
BH1(His)6 (50 µg) was
phosphorylated with 0.1 mM [-32P]MgATP
(specific radioactivity of 500 cpm/pmol) and 1.6 units/ml PKA, 2.8 units/ml p70s6k, 7 units/ml MAPKAP kinase-1, or 0.7 units/ml PKB as described above. The final incubation volumes were 0.15 ml for PKA, p70s6k, and MAPKAP kinase-1 and 0.3 ml for PKB.
After 2 h at 30 °C, protein was precipitated (17) and digested
in 0.2 ml of 0.1 M Tris-Cl, pH 8.6, and 2 M
urea with 1 µg of bovine trypsin overnight at 30 °C. Peptides were
separated by narrow bore HPLC (15), collected by hand in Eppendorf
tubes, and counted by Cerenkov radiation. Labeled peptides were further
purified as described (17). Aliquots of peaks containing radioactivity
(0.7 µl) were spotted with 0.7 µl of matrix, consisting of a
saturated solution of
-cyano-4-hydroxycinnamic acid in
CH3CN and 0.1% (v/v) trifluoroacetic acid (2:1) plus
substance P as an internal standard (1 pmol; M + H+ = 1348.7), and allowed to dry on the target. The mass spectrometer was a
LASERMAT 2000 (Finnigan MAT Ltd., San Jose, CA), and masses were
calculated from 20-30 cumulated spectra. Edman microsequencing was
performed with an Applied Biosystems 477A gas-phase sequencer equipped
with a phenylthiohydantoin detector. Solid-phase sequencing of peptides
was carried out as described (32).
Protein was measured by the Bradford method
(33) using -globulin as a standard or by the reaction with
ninhydrin, after total alkaline hydrolysis (34), using bovine serum
albumin as a standard. Kinetic constants were calculated by computer
fitting of the data to a hyperbola describing the Michaelis-Menten
equation by nonlinear least-squares regression.
The two recombinant isoforms of bovine heart PFK-2/FBPase-2, BH1 and BH3, were compared with the native enzyme purified from bovine heart myocardium. Analysis of the purified recombinant BH1 and BH3 preparations by SDS-PAGE revealed that they contained major bands of Mr 58,000 and 54,000, respectively, each one corresponding to the two forms of native bovine heart PFK-2/FBPase-2 (data not shown). However, truncated forms of BH1 were also observed, which have been noted previously (35). Electroblotting and N-terminal microsequencing of the intact and truncated proteins from SDS-polyacrylamide gels gave the same amino acid sequence, SGNPASSSEQ, suggesting that truncation resulted from the loss of a C-terminal fragment. This C-terminal truncation could result from a premature arrest of translation due to the presence of consecutive proline residues located between Arg-477 and Arg-496, identified by MALDI-MS peptide mapping (36) of the intact and truncated forms (data not shown).
To select and isolate the intact protein containing the C-terminal phosphorylation sites, we decided to add a 6-histidine tag at the C terminus of BH1. We previously engineered a polyhistidine tail on the liver PFK-2/FBPase-2 isozyme to facilitate its purification (26). This modification had no effect on PFK-2 or FBPase-2 activity of the liver isozyme (26). The BH1(His)6 recombinant enzyme was expressed in E. coli and purified. Analysis of the recombinant isoform by SDS-PAGE revealed a major band of Mr 61,000 ± 1000 (five preparations) and confirmed that the contamination by truncated forms was greatly reduced, if not abolished.
Kinetic Properties of PFK-2 and FBPase-2The Vmax of PFK-2 in the native bovine heart preparation was ~50 milliunits/mg of protein. However, this value was probably underestimated since the preparation was not pure. The kinetic properties of the recombinant proteins were compared. The Vmax of PFK-2 of the recombinant BH3 isoform was four to five times greater than that of recombinant BH1(His)6, and its Km for Fru-6-P was about half that of the BH1(His)6 isoform (Table I). The Vmax of FBPase-2 of BH1(His)6 was about three times that of BH3, so the PFK-2/FBPase-2 activity ratios were approximately 3 and 40, respectively. The Km of FBPase-2 for Fru-2,6-P2 of the BH1(His)6 preparation was five times higher than that of BH3 (Table I). The FBPase-2 activity of the native bovine heart preparation was too low and did not allow accurate determinations of Km and Vmax to be made.
|
PKA,
p70s6k, MAPKAP kinase-1, PKB, and, to a much lesser extent,
MAPK phosphorylated the Mr 58,000 and 54,000 bands of native bovine heart PFK-2/FBPase-2, although the extent of
phosphorylation of the Mr 58,000 band greatly
exceeded that of the Mr 54,000 band (Fig.
1). In agreement with the principal phosphorylation of
the Mr 58,000 band of the native enzyme, the
BH1(His)6 preparation was phosphorylated by PKA (0.77 ± 0.05 mol/mol of subunit; n = 3), p70s6k
(0.79 ± 0.09 mol/mol of subunit; n = 6), MAPKAP
kinase-1 (0.57 ± 0.05 mol/mol of subunit; n = 8),
and PKB (0.69 ± 0.04 mol/mol of subunit; n = 12)
(Fig. 1). BH1(His)6 was a rather poor substrate of MAPK
(0.2 mol/mol of subunit) (Fig. 1), even though exon 15 contains the
sequence PLS493P, which is in a suitable consensus for
phosphorylation by this protein kinase.
Only p70s6k and MAPKAP kinase-1 significantly phosphorylated the BH3 preparation (Fig. 1), and as expected, the stoichiometry of phosphorylation was very low (0.02 mol/mol of subunit with p70s6k). Therefore, the major phosphorylation sites reside in the sequence encoded by exon 15.
Effect of Treatment with Protein Kinases on Kinetic Properties of PFK-2 and FBPase-2Treatment of native bovine heart PFK-2/FBPase-2 with the protein kinases slightly decreased the Km of PFK-2 for Fru-6-P and increased the Vmax, except with MAPKAP kinase-1, which did not affect the latter (Tables I and II).
|
Treatment of BH1(His)6 with PKA, p70s6k, MAPKAP kinase-1, or PKB increased the Vmax of PFK-2 and decreased the Km for Fru-6-P (Tables I and II). As expected, the PFK-2 activity of BH3, which was a poor substrate for these protein kinases (Fig. 1), was little affected by treatment with the various protein kinases (Table I and data not shown for PKB). Therefore, this explains why the effects of the protein kinases on the kinetic properties of PFK-2 of the native enzyme were less pronounced than for BH1(His)6. None of the protein kinases tested modified the kinetic properties of FBPase-2 in the BH1(His)6 and BH3 preparations (Table I).
Correlation between Phosphorylation and PFK-2 ActivityWe
have studied the correlation between the extent of phosphorylation of
BH1(His)6 and PFK-2 activity for MAPKAP kinase-1, p70s6k, and PKB. Treatment with MAPKAP kinase-1 led to the
time-dependent phosphorylation of BH1(His)6,
which paralleled the increase in PFK-2 activity (Fig.
2A). Likewise, phosphorylation by
p70s6k or PKB correlated with the increase in PFK-2
activity (Fig. 3).
After phosphorylation of BH1(His)6 by MAPKAP kinase-1, the protein kinase reaction was stopped with EDTA prior to incubation with PP2A. The dephosphorylation of BH1(His)6 by PP2A correlated with the loss of PFK-2 activity (Fig. 2B). However, the activation of PFK-2 by MAPKAP kinase-1 was not totally reversed because BH1(His)6 was not completely dephosphorylated in this experiment. The small changes in kinetic properties of PFK-2 in the BH3 preparation seen after treatment with p70s6k (Table I) were not related to the phosphorylation state of the protein (data not shown).
Phosphorylation Site IdentificationBH1(His)6 was
phosphorylated with [-32P]MgATP and the protein
kinases under conditions that gave maximum phosphorylation and digested
with trypsin. Following peptide purification by narrow bore HPLC, three
labeled peaks were identified (Fig. 4). The labeled peptides were further purified and analyzed by Edman microsequencing and MALDI-MS (Table III). Solid-phase sequencing
indicated that in peaks I and II, the burst of radioactivity occurred
at Ser-466 (Fig. 4). The phosphorylation of Ser-466 by PKA has
previously been demonstrated by Edman sequencing (17, 37). In peak III, the burst of radioactivity occurred at Ser-483 (Fig. 4), and we previously demonstrated the phosphorylation of Ser-483 by PKA (17).
|
Analysis of the extent of phosphorylation of the three peaks, after treatment with the various protein kinases, indicated that PKA, MAPKAP kinase-1, and PKB labeled the same peaks (II and III), i.e. Ser-466 and Ser-483, to the same extent, whereas p70s6k mainly labeled peak I (Ser-466) and peak II (also Ser-466). Moreover, the phosphorylation of Ser-483 by p70s6k was about four times less than that of Ser-466 (Table III). Interestingly, microsequencing of the peptides in peak I, which were only phosphorylated by p70s6k at Ser-466, suggested that Thr-475 was also phosphorylated since a gap in the sequence was found at this residue and, during solid-phase sequencing, a burst of radioactivity occurred at this position (Fig. 4). However, phosphoamino acid analysis of the peptides in peak I, following total acid hydrolysis and thin-layer chromatography, indicated that the phosphorylation of Thr-475 was only ~30% of that of Ser-466 (data not shown). Thr-475 has been shown to be phosphorylated by protein kinase C in native PFK-2/FBPase-2 (8, 17, 37), but treatment with protein kinase C was without effect on PFK-2 or FBPase-2 activity (15, 38).
In this paper, we show that the native and recombinant BH1(His)6 forms of bovine heart PFK-2/FBPase-2 are new substrates of p70s6k, MAPKAP kinase-1, and PKB and that phosphorylation activates PFK-2 (Tables I and II). With recombinant BH1(His)6, all the protein kinases tested decreased the Km for Fru-6-P and increased the Vmax.
A comparison of the extent of phosphorylation of the recombinant BH1(His)6 and BH3 preparations suggested that the major phosphorylation sites reside in the sequence encoded by exon 15. Following phosphorylation of BH1(His)6 and trypsin digestion, sequence analysis and MALDI-MS suggested that PKA, MAPKAP kinase-1, and PKB phosphorylated Ser-466 and Ser-483 to a similar extent, whereas p70s6k treatment preferentially labeled Ser-466. The stoichiometry of phosphorylation by the protein kinases tended toward 1 mol of phosphate incorporated per mol of subunit, but never reached the expected value of 2 mol of incorporation on the basis of complete phosphorylation of Ser-466 and Ser-483 in each subunit. Moreover, we were not able to increase the stoichiometry of phosphorylation, even after prolonged incubation with the protein kinases. This might indicate "half-of-the-sites" phosphorylation. Alternatively, only 50% of the recombinant protein might have been in the correct conformation for phosphorylation.
The amino acid sequences surrounding Ser-466 and Ser-483 are similar to
those of Ser-9 of rabbit muscle glycogen synthase kinase-3 and
Ser-21 of rabbit glycogen synthase kinase-3
: bovine heart PFK-2,
RMRRNS466FT; bovine
heart PFK-2,
RPRNYS483VG;
muscle glycogen synthase kinase-3
,
RPRTTS9FA; and muscle
glycogen synthase kinase-3
,
RARTSS21FA.
Serine 9 in glycogen synthase kinase-3 is phosphorylated by
p70s6k and MAPKAP kinase-1 (39) and PKB (31). Indeed,
sequences of the type
RXRXXSXX,
with arginine residues at positions n
3 and n
5
of the phosphorylated serine, are recognized by p70s6k and
MAPKAP kinase-1 (21, 40). Serine 466 is in a favorable consensus for
phosphorylation by PKB (Arg-Xaa-Arg-Yaa-Zaa-(Ser/Thr)-Hyd, where Xaa is
any amino acid, Yaa and Zaa are small residues other than glycine, and
Hyd is a bulky hydrophobic residue (41)). Indeed, a synthetic peptide
containing Ser-466 was a good substrate of PKB (Table
IV). By contrast, Ser-483 is in a rather unfavorable consensus for phosphorylation by PKB, especially due to the presence of
the bulky tyrosine residue at position n
1 of Ser-483 and
the lack of a large hydrophobic residue at position n+1
(41), despite the fact that its phosphorylation was clearly
demonstrated in the intact protein (Table III). Moreover, a synthetic
peptide containing Ser-483 was a poorer substrate for PKB, with a
Km 20-fold higher than that of the peptide
containing Ser-466 (Table IV). Therefore, the primary sequence around
Ser-483 may not be the sole factor determining its phosphorylation by
PKB. An interesting possibility is that phosphorylation of Ser-483
would require prior phosphorylation of Ser-466.
While this work was in progress, we were investigating the insulin signaling pathway leading to PFK-2 activation in isolated rat cardiomyocytes (5). The activation of PFK-2 by insulin was sensitive to wortmannin, a phosphatidylinositol 3-kinase inhibitor, but was not blocked by rapamycin, which prevents the activation of p70s6k, or by PD 98059, a compound that blocks the MAPK cascade and hence MAPKAP kinase-1 activation. Therefore, the insulin signaling pathway for PFK-2 activation probably involves phosphatidylinositol 3-kinase activation, and indeed, our data show that insulin activates this enzyme in rat heart (5). The insensitivity of the insulin-induced activation of PFK-2 to rapamycin and PD 98059 in cardiomyocytes suggests that the effect is unlikely to be mediated by p70s6k or MAPKAP kinase-1.
A good candidate for mediating the insulin-induced activation of heart PFK-2 is therefore PKB, as already described for the insulin-induced inactivation of glycogen synthase kinase-3 in a skeletal muscle cell line (31). Indeed, our preliminary results indicate that PKB is activated in rat hearts in response to insulin. Furthermore, the changes in PFK-2 kinetic properties induced by phosphorylation of BH1(His)6 were similar to those observed in vivo (4), but are partially masked in the native bovine heart enzyme, which contains more of the BH3 isoform. However, rat heart may contain proportionately more of the long isoform corresponding to BH1. In conclusion, we propose that PKB is part of the insulin signaling cascade leading to PFK-2 activation in heart.
We thank Professor P. Cohen for interest and Dr. B. Caudwell (University of Dundee) for help with the solid-phase sequencing. For the gas-phase sequencing, the help of H. Degand and Dr. M. Boutry (University of Louvain) is gratefully acknowledged. Finally, we thank Dr. M.-C. Méchin for initial phosphorylation trials and M. P. Louckx for technical help.