(Received for publication, December 11, 1996)
From the Friedrich Miescher Institute, P. O. Box 2543, CH-4002 Basel, Switzerland and the § Department of Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, Texas 75235
The influence of inositol phosphates and
phosphoinositides on the isoform of the RAC-protein kinase B
(RAC/PKB) was studied using purified wild type and mutant kinase
preparations and a recombinant pleckstrin homology (PH) domain. Binding
of inositol phosphates and phosphoinositides to the PH domain was
measured as the quenching of intrinsic tryptophan fluorescence.
Inositol phosphates and D3-phosphorylated phosphoinositides bound with affinities of 1-10 µM and 0.5 µM,
respectively. Similar values were obtained using RAC/PKB expressed and
purified from baculovirus-infected Sf9 cells in the fluorescence assay.
The influence of synthetic dioctanoyl derivatives of
phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol
3,4,5-trisphosphate on the activity of RAC/PKB purified from
transfected COS-1 cells was studied. Phosphatidylinositol 3,4,5-trisphosphate was found to inhibit the RAC/PKB kinase activity with half-maximal inhibition at 2.5 µM. In contrast,
phosphatidylinositol 3,4-bisphosphate stimulated kinase activity
(half-maximal stimulation at 2.5 µM). A mutant RAC/PKB
protein lacking the PH domain was not affected by D3-phosphorylated
phosphoinositides. These results demonstrate that the PH domain of
RAC/PKB binds inositol phosphates and phosphoinositides with high
affinity, and suggest that the products of the phosphatidylinositide
3-kinase can act as both a membrane anchor and modulator of RAC/PKB
activity. The data also provide further evidence for a link between
phosphatidylinositide 3-kinase and RAC/PKB regulation.
The stimulation of receptor tyrosine kinases (RTK)1 by agonists leads to immediate activation of intracellular signal transduction pathways. The assembly of multiprotein complexes at the plasma membrane is one important feature of RTK signal transduction mechanisms (reviewed in Refs. 1 and 2). Numerous studies suggest that the activation of phosphatidylinositide 3-kinase (PI3-K) by growth factors is also involved (3, 4), leading to the accumulation of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3) and phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P2); these metabolites are assumed to act as second messengers (5). Recently RAC-protein kinase B (RAC/PKB) has emerged as a key player in the PI3-K-stimulated signaling pathway, based on the inhibition of its activation by wortmannin (6-11).
RAC/PKB is a subfamily of the second messenger-regulated
serine/threonine kinases (12). Three isoforms (,
,
) have been identified, each consisting of an amino-terminal pleckstrin homology (PH) domain, a central kinase domain, and a serine/threonine-rich carboxyl-terminal region (13-17). Various stimuli, such as insulin, PDGF, epidermal growth factor, basic fibroblast growth factor, serum
(6-10), and protein phosphatase inhibitors (9), lead to activation of
RAC/PKB kinase activity. The activation is promoted by signals
emanating from RTK-regulated PI3-K and is accompanied by an increase in
serine/threonine phosphorylation of RAC/PKB (6, 9). The phosphorylation
of two sites has been shown to be critical for activation by insulin
and insulin growth factor-1 (18): Thr-308 in the T-loop and Ser-473 in
the Ser/Thr-rich carboxyl-terminal region. Interestingly, the Ser-473
site is conserved in many of the second messenger subfamily of protein
kinases. The inactivation of kinase activity by in vitro
phosphatase treatment suggest that protein phosphatase 2A is also
involved in the regulation of RAC/PKB (9).
The inhibition of growth factor-mediated RAC/PKB activation by
wortmannin and LY 294002 suggests that RAC/PKB plays a crucial role in
perceiving and transmitting signals from RTK-activated PI3-K (6-11).
In particular, the amino-terminal PH domain of RAC/PKB appears to be
important in kinase regulation because its deletion abrogates
activation by PDGF and serum (7, 9). The PH domain is a module of
approximately 100 amino acids identified in over 90 different proteins
(19-21). The primary structure of the PH domain varies in different
proteins. The three-dimensional structures have been determined for
pleckstrin (22), dynamin (23), -spectrin (24), and phospholipase
C
-1 (25) and all are similarly folded. It was noted that the
three-dimensional structure of the pleckstrin PH domain had
similarities to lipid-binding proteins (22). Subsequently, inositol
phosphates and phosphoinositides were found to bind to PH domains from
pleckstrin (26),
-spectrin (27), and phospholipase C
-1 (28,
29).
Structural studies have revealed the critical amino acid residues
involved in the binding of inositol phosphates. The structures of the
-spectrin and the phospholipase C
-1 (PLC
-1) PH domains complexed with inositol (1,4,5)-trisphosphate
(Ins(1,4,5)P3), indicated that Trp-23 of
-spectrin and
Trp-36 of PLC
-1 are hydrogen bonded to a phosphate group of the
ligand (27, 25). The existence of an homologous Trp-22 in RAC/PKB
prompted us to use the intrinsic tryptophan fluorescence of the RAC/PKB
PH domain to monitor interactions with phosphoinositides and inositol
phosphates. Here we demonstrate that this PH domain binds with high
affinity to certain inositol phosphates and to
PtdIns(3,4,5)P3 and PtdIns(3,4)P2. Conflicting results on the influence of D3-phosphorylated phosphoinositides on
RAC/PKB activity (7, 30) also led us to examine the effect of these
ligands on RAC/PKB kinase activity. The data obtained show that these
compounds can both inhibit and activate the kinase in a PH
domain-specific manner.
The dioctanoyl derivatives of PtdIns(3,4)P2 and PtdIns(3,4,5)P3 were synthesized according to the method of Reddy et al. (31). PtdIns(4,5)P2 and inositol phosphates were purchased from Fluka, Switzerland.
Construction of Expression VectorsThe RAC/PKB PH domain
consisting of amino acids 1-131 was subcloned as a
5-BamHI/3
-EcoRI PCR fragment into pGEX-2T
(Pharmacia) vector. The hemagglutinin (HA) epitope-tagged constructs
(HA-RAC/PKB
and HA-
NRAC/PKB
) were previously described (9,
18).
The 1.6-kilobase pair EcoRI fragment of
human RAC/PKB (13) was subcloned into the EcoRI site of
the baculovirus transfer vector pVL1392 (Invitrogen) in the correct
orientation. The plasmid was cotransfected into Sf9 cells together with
the purified wild-type baculovirus (AcMNPV) DNA. The culture
supernatant was collected and the recombinant virus was purified by
limiting dilution and screened by dot-blot hybridization. The pure
recombinant virus was amplified and then used for recombinant protein
production.
Escherichia coli JM109 transformed with the
pGEX2T PH131 expression construct was diluted in 2 × YT media
containing 100 µg/ml ampicillin, grown at 37 °C to a density
OD600 of 0.8 to 0.9, and then induced with 0.1 mM isopropyl-1-thio--D-galactopyranoside for
3 h at 25 °C. Bacterial pellets were lysed in 5 volumes of 50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride
using lysozyme (1 mg/ml). After lysis, the extracts were treated with
0.1% (w/v) sodium deoxycholate and 20 µg/ml DNase I for 30 min at
4 °C and then centrifuged at 20,000 × g for 45 min
at 4 °C. Glutathione-agarose (Sigma) was added to the supernatant and incubated for 1 h at 4 °C with gentle agitation. The
glutathione-agarose beads were collected by centrifugation, washed with
phosphate-buffered saline, and the fusion protein eluted with 10 mM glutathione in 50 mM Tris-HCl, pH 8.0, 1 mM EDTA for 2 h at room temperature. Following
cleavage with 2 units of thrombin (Sigma) per mg of fusion protein, the
PH domain preparation was purified to homogeneity by Mono Q FPLC
(Pharmacia). The PH domain was eluted from the column with a 20-ml
linear gradient of 0-600 mM NaCl in 50 mM Tris-HCl, pH 8.0, 1 mM EDTA with a flow rate of 1 ml/min.
The PH domain eluted at 450 mM NaCl and peak fractions were
examined by SDS-PAGE.
The Sf9 cells were harvested 72 h after infection with recombinant baculovirus by centrifugation at 1000 × g for 10 min. The cells were washed twice with TBS (Tris-buffered saline: 20 mM Tris-HCl, pH 7.5, 120 mM NaCl) and then lysed with extraction buffer (50 mM Tris-HCl, pH 7.5, 1 mM DTT, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 20 µM leupeptin, 20 µM aprotinin) at 4 °C. The lysate was centrifuged at 14,000 × g for 10 min, the supernatant was diluted 5-fold with the appropriate column buffer and applied directly to the Mono Q column (HR 5/5, Pharmacia). The column was equilibrated in 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM DTT. The sample was applied at a flow rate of 0.5 ml/min and the column was washed with 2 bed volumes before a linear ascending gradient from 0 to 0.4 M NaCl (20 ml) and then from 0.4 to 1.0 M NaCl (10 ml) in column buffer was applied at a flow rate of 0.5 ml/min and fractions of 0.5 ml were collected. Fractions were analyzed for protein content and by SDS-PAGE and immunoblotting for RAC/PKB. The fractions containing RAC/PKB were pooled and concentrated using a Centricon-30 to 200 µl and subjected to further purification by gel filtration chromatography on a Superose 12 (HR 10/30) column (Pharmacia). The column was equilibrated in 50 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1 mM EDTA, 0.02% NaN3. The protein was eluted in the above buffer at a flow rate of 0.2 ml/min. Fractions of 0.5 ml were collected and analyzed for protein content by SDS-PAGE.
Purification of HA-tagged RAC/PKBEpitope-tagged HA-RAC/PKB
was purified from 40 × 10-cm dishes of transfected COS-1 cells
without treatment or following stimulation with 0.1 mM
pervanadate as described previously (9). The cells were lysed in a
buffer containing 50 mM Tris-HCl, pH 7.5, 1% Nonidet P-40,
120 mM NaCl, 1 mM EDTA, 50 mM
sodium fluoride, 40 mM -glycerophosphate, 0.1 mM pervanadate, 100 nM okadaic acid, 2 µg/ml
leupeptin, 2 µg/ml aprotinin, 1 mM benzamidine, and 0.5 mM phenylmethylsulfonyl fluoride. The extract was
centrifuged for 15 min at 12,000 × g and 4 °C, and the
supernatant applied to a 1-ml affinity column of the HA-specific
monoclonal antibody 12CA5 covalently attached to protein A-Sepharose
according to the supplier's protocol (Pierce). The bound protein was
eluted stepwise with 1 mg/ml HA-epitope peptide (AAYPYDVPDYA) for
1 h at 4 °C. The pooled fractions were dialyzed for 3 h at
4 °C against 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM EDTA, 1 mM DTT, 5 nM
okadaic acid. The dialyzed protein preparation was loaded onto a HR5/5
Mono Q column (FPLC Pharmacia) at a flow rate of 1 ml/min and protein
eluted with a 20-ml gradient of 50-550 mM NaCl in 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM DTT, 5 nM okadaic acid. The kinase eluted at
about 350 mM NaCl and the fractions were analyzed by
Western blot and assayed for kinase activity (as described below).
Fractions were pooled and spin concentrated with a Centricon-30
(Amicon). HA-
NRAC/PKB was purified in the same way from
pervanadate-stimulated COS-1 cells, but the kinase eluted from Mono Q
at 300 mM NaCl. Protein concentrations were determined by
the Bradford assay using the Bio-Rad kit.
Kinase assays were performed at
30 °C in a total reaction volume of 50 or 100 µl consisting of 50 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 1 mM DTT, 1 µM cAMP-dependent
protein kinase inhibitor, 30 µM Crosstide (GRPRTSSFAEG)
(10), 50 µM ATP with [-32P]ATP (specific
activity 4000 cpm/pmol). Where indicated, PtdIns(3,4)P2 or
PtdIns(3,4,5)P3 or other inositol phosphates were added.
The reactions were equilibrated at 30 °C for 5 min and initiated by addition of purified protein. At appropriate time points, 10-µl aliquots of the reaction solution were spotted onto P81 filters (Whatman) and washed 4 times with 1% phosphoric acid. Filter-bound radioactivity was determined by scintillation counting. Under the
conditions used the assay is linear for 60 min (see Fig.
4A).
Fluorescence Spectroscopy
The fluorescence measurements were recorded using a Perkin Elmer Luminescence Spectrometer LS 50 B at 25 °C in 20 mM Hepes-NaOH, pH 7.0, 10 mM MgCl2 at a protein concentration of 1.6-3.2 µg/ml (0.1-0.2 µM) in a volume of 1 ml. The purified protein was excited at a wavelength of 290 nm and fluorescence measured at 345 nm. Titration data were fitted using simple weighting to a rectangular hyperbolic function by nonlinear regression analysis (Grafit 3.0 Erithacus software).
To study the function
of the PH domain of RAC/PKB we expressed the PH domain (amino acids
1-131 of the RAC/PKB sequence) as a soluble glutathione
S-transferase fusion protein. The fusion protein was
isolated by affinity chromatography, cleaved by thrombin, and purified
to homogeneity by Mono Q ion-exchange chromatography (Fig.
1A). Full-length wild type RAC/PKB was
expressed in baculovirus-infected Sf9 cells and purified as described
under "Materials and Methods." The final purification step involved
gel filtration on Superose 12; the purity of the protein obtained by
this procedure is shown in Fig. 1B. The RAC/PKB obtained
from Sf9 cells had a low specific activity and therefore was probably
only phosphorylated on Ser-124 and Thr-450 (18). Phosphorylation of
Thr-308 and Ser-473 is required for activation of RAC/PKB (18). In
order to produce activated RAC/PKB we therefore chose to express the
kinase in COS-1 cells. The hemagglutinin-tagged RAC/PKB protein was
overexpressed in COS-1 cells and purified as outlined under
"Materials and Methods." The different purified enzyme
preparations were used for the characterization of their interaction
with inositol phosphates and phosphoinositides.
Binding of Inositol Phosphates to RAC/PKB
We examined the
binding of inositol phosphates to the bacterially expressed PH domain
and with RAC/PKB purified from baculovirus-infected Sf9 cells.
Following excitation at 290 nm, the PH domain exhibited an emission
spectra typical for tryptophan with a maximum fluorescence at 345 nm;
the fluorescence was quenched by approximately 30% upon addition of
ligands (Fig. 2A). Similar changes in the
fluorescence were detected for the full-length native protein isolated
from Sf9 cells, but the fluorescence was only quenched maximally 15% upon addition of ligands (Fig. 2B). No shift of the maximum
of the emitted fluorescence was observed with both purified
preparations in the presence of ligands. Changes in the fluorescence in
the presence of inositol phosphates were used to determine the
dissociation constants (Kd) for the complexes.
Inositol phosphates bound to the PH domain with micromolar affinities (Table I). For Ins(1)P and Ins(2)P the Kd values were 7.0 and 8.4 µM, respectively (Table I). With Ins(1,4)P2 the affinity was about 3-fold higher with a Kd of 2.4 µM. The highest affinities for inositol phosphates were found with Ins(1,4,5)P3 and Ins(1,3,4,5)P4, with Kd values of 1.2 µM and 1.5 µM, respectively.
|
In contrast the affinities for other inositol tetrakisphosphates were
lower, and Kd values for Ins(3,4,5,6)P4
and Ins(1,2,3,5)P4 were 6.4 and 5.0 µM,
respectively. The affinity of the PH domain for
Ins(1,2,3,4,5,6)P6 was slightly higher with a
Kd of 8.7 µM. The observed affinity
for Ins(1,4,5)P3 is about 30-fold higher than values
previously reported for -spectrin (27), and 3500-fold higher than
the affinity of 1-(
-glycerophosphoryl)-inositol(4,5)-bisphosphate for dynamin (32). The PLC
-1 PH domain bound Ins(1,4,5)P3
6-fold better than the RAC/PKB PH domain, showing a clear preference for Ins(1,4,5)P3 and PtdIns(4,5)P2 (28). In
contrast to the PH domain of PLC
-1, we found specific binding of
D3-phosphorylated inositol phosphates by the RAC/PKB PH domain;
however, the RAC/PKB PH domain distinguished poorly between
Ins(1,3,4,5)P4 and Ins(1,4,5)P3. Furthermore, the phosphate groups at positions 3, 4, and 5 of the
inositol ring were important for the interaction with the RAC/PKB PH
domain, as additional phosphate groups interfered with binding,
as observed with Ins(3,4,5,6)P4,
Ins(1,2,3,5)P4, and Ins(1,2,3,4,5,6)P6.
To examine whether there are dramatic differences in the binding characteristics between the isolated PH domain and the full-length protein, we further determined the affinities for several inositol phosphates to the purified RAC/PKB expressed in Sf9 cells (Table I). Here, we measured a dissociation constant of 11.8 µM for Ins(1)P and 4.1 µM for Ins(1,4)P2, for Ins(1,2,3,4,5,6)P6 the dissociation constant was 7.5 µM. As for the isolated PH domain the inositol phosphates with the highest affinities were found with Ins(1,4,5)P3 and Ins(1,3,4,5)P4 with Kd values of 3.2 and 2.5 µM, respectively. The overall picture of the affinities for the full-length protein was therefore similar to that of the isolated PH domain.
Binding of Phosphoinositides to RAC/PKBTo study the
interaction of the RAC/PKB PH domain with phosphoinositides we used
PtdIns(4,5)P2 and the dioctanoyl derivatives (31) of
PtdIns(3,4)P2 and PtdIns(3,4,5)P3 (Fig.
3A). The Kd determined for
PtdIns(4,5)P2 was 2.5 µM, which is in the
same range as values reported for dynamin (32) and PLC-1 (28). The
values determined for PtdIns(3,4)P2 and
PtdIns(3,4,5)P3, 570 and 400 nM, respectively,
reflect the preference of the RAC/PKB PH domain for D3-phosphorylated
phosphoinositides. The stoichiometry of binding examined by Scatchard
analysis indicated a 1:1 complex for PtdIns(3,4,5)P3 (Fig.
3B).
With the dioctanoyl derivatives of PtdIns(3,4)P2 and PtdIns(3,4,5)P3 we were able to determine their affinities for the full-length RAC/PKB; the Kd values for PtdIns(3,4,5)P3 and PtdIns(3,4)P2 were 0.7 and 1.0 µM, respectively, and both formed 1:1 complexes as determined by Scatchard analysis.
Influence of Phosphoinositides on RAC/PKB ActivityTo further
characterize the effects of inositol phosphates and phosphoinositides
on RAC/PKB, we examined the influence of these compounds on the
activities of purified wild type HA-RAC/PKB protein and of a construct
lacking the PH domain. HA-RAC/PKB and HA-NRAC/PKB were overexpressed
in COS-1. The wild type protein was purified from both untreated and
pervanadate stimulated cells in order to obtain enzyme preparations
which were dephosphorylated or phosphorylated on the activating sites
Thr-308 and Ser-473 (9, 18). Kinase preparations were purified by
immunoaffinity chromatography using the 12CA5 monoclonal antibody
immobilized on protein A-Sepharose (see "Materials and Methods")
followed by anion exchange chromatography over Mono Q. The in
vitro kinase activity of RAC/PKB was assayed using Crosstide (Fig.
4A), a peptide derived from GSK-3 (see
"Materials and Methods"), which is efficiently phosphorylated by
RAC/PKB (10). The specific activity of HA-RAC/PKB from
pervanadate-stimulated cells was 350 pmol/min/mg compared to 18 pmol/min/mg of the enzyme obtained from unstimulated cells. The
specific activity of HA-
NRAC/PKB was 14 pmol/min/mg indicating that
this protein was not activated by pervanadate treatment, as reported
previously (9).
The initial rates of peptide phosphorylation by RAC/PKB were measured
in the presence of increasing concentrations of the dioctanoyl
derivatives of PtdIns(3,-4)P2 and
PtdIns(3,4,5)P3. Surprisingly, we found that
PtdIns(3,4,5)P3 inhibited HA-RAC/PKB activity in a
concentration-dependent manner, with half-maximal inhibition at 2.5 µM and 95% inhibition at 10 µM (Fig. 4B). In contrast,
PtdIns(3,4)P2 stimulated HA-RAC/PKB activity up to 2-fold (Fig. 4C). PtdIns(4,5)P2 did not influence the activity of
HA-RAC/PKB. However, PtdIns(3,4)P2 at 6 µM
appeared to prevent inhibition of HA-RAC/PKB activity by 2.5 µM PtdIns(3,4,5)P3 and was even able to
revert the effect by causing a slight stimulation of the activity (Fig.
5). Addition of 15 µM
Ins(1,3,4,5)P4 reduced the inhibition caused by 2.5 µM PtdIns(3,4,5)P3, probably by competition for the same binding site on the PH domain. Significantly, at concentrations up to 40 µM, Ins(1,3,4,5)P4
did not induce the same effect with HA-RAC/PKB as
PtdIns(3,4,5)P3. This result indicates a different mode of
binding to the PH domain compared with PtdIns(3,4,5)P3.
To further characterize the influence of PtdIns(3,4,5)P3 or PtdIns(3,4)P2 on RAC/PKB, we examined their effects on enzyme purified from unstimulated cells, which is not phosphorylated on the activating residues. The effects of the two compounds were comparable to the activated RAC/PKB. PtdIns(3,4)P2 increased the activity 2.5-fold and PtdIns(3,4,5)P3 inhibited kinase activity with a half-maximal concentration of 2.2 µM (data not shown). Thus the phosphorylation state of the kinase had little influence on phosphoinositide interactions under the conditions used.
To determine whether the PH domain was involved in mediating these
effects on kinase activity, we purified a mutant RAC/PKB lacking the
NH2-terminal 116 amino acids (HA-NRAC/PKB). As expected, the kinase activity of HA-
NRAC/PKB was not affected by
phosphoinositides (Fig. 6). Neither the addition of 10 µM PtdIns(3,4,5)P3 nor 10 µM
PtdIns(3,4)P2 had any significant influence on the
HA-
NRAC/PKB activity in vitro. Furthermore, 10 µM PtdIns(4,5)P2 had no effect on the
activity. Thus, the PH domain apparently influenced kinase activity and
the overall effect was dependent on the ligand bound.
Taken together the results so far available indicate that the PtdIns(3,4,5)P3 and PtdIns(3,4)P2 derivatives influence kinase activity via an intramolecular interaction. Alternatively they could induce dimerization and this ultimately affects kinase activity. These possibilities are currently being investigated.
Recently we demonstrated that RAC/PKB activation by insulin and insulin growth factor-1 is mediated by phosphorylation of Thr-308 in the T-loop, and Ser-473 in the carboxyl-terminal regulatory domain (18). The implication of these data is that RAC/PKB is regulated by an upstream kinase. However, the fact that the activation is potently inhibited by wortmannin suggests that PI3-K plays a pivotal role in this process. We suggested previously that a putative RAC/PKB upstream kinase could be directly regulated by phosphoinositides (18). Moreover, it is also possible that phosphoinositides directly interact with RAC/PKB during the activation process.
In this article we have investigated the interaction of inositol
phosphates and phosphoinositides with RAC/PKB. Our results show that
the RAC/PKB PH domain binds inositol phosphates and phosphoinositides,
and that the activity of RAC/PKB is influenced in vitro by
PtdIns(3,4,5)P3 and PtdIns(3,4)P2. The
affinities for these compounds were higher than those recently
determined for pleckstrin, -spectrin, and dynamin (26, 27, 32). The highest affinities were found for Ins(1,4,5)P3 and
Ins(1,3,4,5)P4 and for the D3-phosphorylated
phosphoinositides PtdIns(3,4)P2 and
PtdIns(3,4,5)P3. Phosphate groups at positions 3, 4, and 5 of the inositol ring were found to be important for the interaction of
the phosphoinositides and inositol phosphates with the RAC/PKB PH
domain. Studies with Ins(3,4,5,6)P4,
Ins(1,2,3,5)P4, and Ins(1,2,3,4,5,6)P6 showed
that additional phosphates interfered with binding to the PH domain. In
contrast to the PH domain of PLC
-1, we found specific binding of
D3-phosphorylated inositol phosphates and phosphoinositides by the
RAC/PKB PH domain. However, the discrimination between the tested
ligands was less pronounced than for phospholipase C
-1 (28).
Obviously the mechanism and/or the site for binding inositol phosphates
are different compared to PLC-1 where structural restraints hinder
the binding of D3-phosphorylated inositol phosphate (25).
Ins(1,4,5)P3 seems to be buried in a binding pocket of PLC
-1, which is formed by the
1/
2,
3/
4, and
5/
6
loops, where the ligand makes 12 hydrogen bonds with nine amino acids contributed by the loops. The interactions of ligands with the loop
3/
4 seems to be unique for this PH domain and could explain the
high affinity of PLC
-1 for Ins(1,4,5)P3 and
PtdIns(4,5)P2 (25). In the crystal structure of the complex
of Ins(1,4,5)P3 and
-spectrin, the
Ins(1,4,5)P3 seems to bind to a different location compared
to PLC
-1 and is more exposed to solvent (25, 27). Here the binding
pocket is formed by the
1/
2 and
5/
6 loops (27). The number
of hydrogen bonds is decreased, such that five amino acids form seven
hydrogen bonds to bind the ligand, which may account for the decreased
affinity. The phosphates at the D-4 and D-5 positions are important for
the interaction. Inspection of the
-spectrin PH domain structure
does not reveal candidate residues that would potentially interact with
a phosphate group at the D-3 position, or occlude ligand binding (27).
Indeed a change in the circular dichronic spectrum was observed with Ins(1,3,4,5)P4 but not for compounds lacking the phosphates
at D-4 or D-5 positions (27). For other PH domains it is not yet known
in detail whether they bind to D3-phosphorylated inositol phosphates.
Presumably the diverse primary structures lead to variations in the
binding site resulting in different affinities and specificities for
inositol phosphates.
The primary sequence of the potential inositol phosphate binding site
of RAC/PKB is likely to be more closely related to PLC-1 than
-spectrin, as shown in Fig. 7. Analysis of the
affinities of inositol phosphates for the PH domain of RAC/PKB predicts
a binding site with high affinities for D3-phosphorylated compounds and
poor discrimination against other inositol phosphates. RAC/PKB shows
amino acid residues homologous to PLC
-1 in the presumed variable
loop
3/
4. The loop
3/
4 in PLC
-1 is involved in the binding of Ins(1,4,5)P3 (25) and it is possible that the
homologous region of RAC/PKB is also involved in binding of the ligand,
as well as the
1/
2 and
4/
5 loops. This could explain the
higher affinities compared to dynamin and
-spectrin. Furthermore,
our results suggest that there should be specific residues in RAC/PKB that interact with the D3 phosphate group of the ligand. We are currently in the process of determining the structure of the kinase PH
domain complexed with different inositol phosphates.
Several studies have shown that growth factor receptors stimulate PI3-K activity, leading to the activation of RAC/PKB (6-10). Activation of PI3-K leads to an elevation in the levels of D3-phosphorylated phosphoinositides in cells (reviewed in Refs. 3-5). The specific binding of these phosphoinositides to the RAC/PKB PH domain could provide the link between PI3-K and RAC/PKB. In accordance with this suggestion previous studies revealed that activation of RAC/PKB by serum and PDGF was ablated with constructs lacking a functional PH domain (7, 9). However, it should be noted that insulin is able to stimulate RAC/PKB mutants which lack the PH domain (34).2 The mechanistic reasons for these differences remain to be established.
Franke et al. (7) showed that D3-phosphorylated PtdIns produced by immunoprecipated PI3-K activated RAC/PKB in vitro. However, in vivo it is not clear whether PtdIns is a substrate of PI3-K; PtdIns(4)P and PtdIns(4,5)P2 are more likely physiological substrates (3, 4). Here, we demonstrate that the D3-phosphorylated phosphoinositides PtdIns(3,4,5)P3 and PtdIns(3,4)P2 bind to the PH domain of RAC/PKB and have opposing effects on RAC/PKB activity. PtdIns(3,4,5)P3 completely inhibits the activity of RAC/PKB in vitro and PtdIns(3,4)P2 stimulates the activity of RAC/PKB about 2-fold. Depending on the turnover of PtdIns(3,4,5)P3 after cellular stimulation it is possible that the interaction of RAC/PKB with PtdIns(3,4,5)P3 has a minor function in the cellular environment. In thrombin- or fMet-Leu-Phe-stimulated platelets the peak response of PtdIns(3,4,5)P3 already declined after 1 min whereas the level of PtdIns(3,4)P2 increased and was maintained over several minutes (35, 36). In PDGF-stimulated smooth muscle cells, or insulin-stimulated Chinese hamster ovary cells the levels of both PtdIns(3,4,5)P3 and PtdIns(3,4)P2 are increased over several minutes (37, 38). The conversion from PtdIns(3,4,5)P3 to PtdIns(3,4)P2 could possibly be mediated by the mitogen-stimulated inositol polyphosphate 5-phosphatase SHIP (39). It is conceivable that in some cell types, RAC/PKB interacts predominantly with PtdIns(3,4)P2 and that the ratio of PtdIns(3,4,5)P3 to PtdIns(3,4)P2 may modulate the activation of RAC/PKB.
The high affinities of D3-phosphorylated phosphoinositides for the
RAC/PKB PH domain is consistent with this domain being important for
membrane localization. Given the in vivo concentrations of
PtdIns(4,5)P2 and D3- phosphorylated phosphoinositides
reported (40), the affinities of the RAC/PKB PH domain for
phosphoinositides are sufficient for membrane attachment.
Discrimination by the PH domain between PtdIns(3,4)P2,
PtdIns(3,4,5)P3, and PtdIns(4,5)P2 would depend
on their concentrations in the cell membrane. The influence of
PtdIns(3,4)P2 and PtdIns(3,4,5)P3 on RAC/PKB
activity is reminiscent of the allosteric behavior of phospholipase
C-1, which is activated by binding of PtdIns(4,5)P2 to
its PH domain (33, 41). It is not known whether the PH domain of
RAC/PKB undergoes a conformational change upon ligand binding. The fact that we see no shift in the emission maximum of the isolated PH domain
and the full-length protein suggests no conformational changes in the
environment of the tryptophan involved in binding. James and colleagues
(30) showed binding of RAC/PKB to phosphoinositide containing vesicles
with similar affinities as we found with our fluorescence assay, but
they did not observe any influence on the activity of RAC/PKB in the
presence of PtdIns(3,4)P2 or PtdIns(3,4,5)P3 under their experimental conditions. They did not investigate the
binding mechanism or possible conformational changes involved upon
ligand binding. Clearly, the kinetics of the association and
dissociation and possible accompanying conformational changes has to be
studied in more detail. The fact that RAC/PKB isolated from
unstimulated cells showed the same behavior, with similar half-maximal
concentrations of inhibition and activation by
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 argues that
for this interaction the phosphorylation state of the protein has a
minor influence. The maximal activation induced by
PtdIns(3,4)P2 in vitro is lower than the
activation of RAC/PKB achieved by phosphorylation in vivo.
Therefore, a conformational change induced by PtdIns(3,4)P2
alone would not be sufficient to cause maximal activation as proposed
by Franke et al. (7).
In order to more fully understand the role of the PH domain in kinase activation we have recently prepared a series of membrane targeted RAC/PKB constructs.2 The rationale for these experiments was that if the PH domain were important for kinase activation by localizing the protein at the membrane, artificially targeting the RAC/PKB should promote activation. In summary, our results show that membrane targeting was sufficient to promote kinase activation by phosphorylation on Thr-308 and Ser-473. These results therefore suggest that an upstream RAC/PKB kinase or kinases are localized on the membrane. The fact that a "kinase-dead" RAC/PKB construct was also phosphorylated on the same residues supports this suggestion. During the preparation of this manuscript Kohn et al. (34) reported that membrane targeting of RAC/PKB in TRMP canine kidney epithelial cells overexpressing the PDGF receptor and CHO.IR cells overexpressing the insulin receptor lead to a substantial increase of kinase activity. With both studies (34)2 deletion of the PH domain from the membrane targeted construct did not block or inhibit activation. These results illustrate that phosphorylation of RAC/PKB is the principal mechanism for kinase activation following stimulation by growth factors or insulin (9, 18).
The interaction of RAC/PKB with PtdIns(3,4)P2 could therefore promote both membrane localization and cause a conformational change which allows efficient phosphorylation by an as yet unidentified kinase. It is also possible that PtdIns(3,4)P2 and PtdIns(3,4,5)P3 generated after cell stimulation activates this upstream kinase which phosphorylates Thr-308 and Ser-473, working in concert with the PtdIns(3,4)P2 induced conformational change to fully activate RAC/PKB. Isolation and characterization of the upstream kinase will be required to fully delineate the mechanism of RAC/PKB activation, and further study of RAC/PKB regulation by RTK-activated PI3-K should provide significant insights into lipid signaling mechanisms and the involvement of PH domains.
We thank Drs. J. Hofsteenge, T. Schäfer, W. Gläsner, and T. Millward for helpful discussions and comments on the manuscript.