(Received for publication, May 16, 1996, and in revised form, October 4, 1996)
From the Division of Medicinal Chemistry and Pharmaceutics, College of Pharmacy, University of Kentucky, Lexington, Kentucky 40536-0082
The present study takes a novel
approach to explore the mode of action of phosphoinositide 3-kinase
lipid products by identifying a synthetic peptide
W-NG28-43 (WAAKIQASFRGHMARKK) that displays discriminative
affinity with phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3). This PtdIns(3,4,5)P3-binding
peptide was discovered by a gel filtration-based binding assay and
exhibits a high degree of stereochemical selectivity in
phosphoinositide recognition. It forms a 1:1 complex with
PtdIns(3,4,5)P3 with Kd of 2 µM, but binds phosphatidylinositol 4,5-bisphosphate
(PtdIns(4,5)P2) and phosphatidylinositol 3,4-bisphosphate
(PtdIns(3,4)P2) with substantially lower affinity (5- and
40-fold, respectively) despite the largely shared structural motifs
with PtdIns(3,4,5)P3. Other phospholipids examined,
including phosphatidylserine, phosphatidylcholine, phosphatidylinositol, and phosphatidylethanolamine, show low or negligible affinity with the peptide. Several lines of evidence indicate that this phosphoinositide-peptide interaction is not due to
nonspecific electrostatic interactions or phospholipid aggregation, and
requires a cooperative action among the hydrophobic and basic residues
to exert the selective recognition. CD data suggest that the peptide
acquires an ordered structure upon binding to
PtdIns(3,4,5)P3. Further, we demonstrate that
PtdIns(3,4,5)P3 enhances the phosphorylation rate of this
binding peptide by protein kinase C (PKC)- in a
dose-dependent manner. In view of the findings that this
stimulatory effect is not noted with other PKC peptide substrates
lacking affinity with PtdIns(3,4,5)P3 and that PKC-
is
not susceptible to PtdIns(3,4,5)P3 activation, the activity enhancement is thought to result from the substrate-concentrating effect of the D-3 phosphoinositide, i.e. the
presence of PtdIns(3,4,5)P3 allows the peptide to bind to
the same vesicles/micelles to which PKC is bound. Moreover, it is
noteworthy that neurogranin, the full-length protein of
W-NG28-43 and a relevant PKC substrate in the forebrain,
binds PtdIns(3,4,5)P3 with high affinity. Taken together, it is plausible that, in addition to PKC activation, PtdIns(3,4,5)P3 provides an alternative mechanism to
regulate PKC activity in vivo by recruiting and
concentrating its target proteins at the interface to facilitate the
subsequent PKC phosphorylation.
Among various phospholipids in the plasma membrane,
phosphatidylinositol phosphates have received much attention because of their pivotal role in transmembrane signal transduction (1, 2). Two
major phosphoinositide-mediated signaling cascades have been
characterized, both of which originate from phosphatidylinositol 4,5-bisphosphate
(PtdIns(4,5)P2).1 In the
canonical pathway, phospholipase C activation leads to a rapid
production of D-myo-inositol 1,4,5-trisphosphate
(Ins(1,4,5)P3) and diacylglycerol, which elicit
Ca2+ release and protein kinase C (PKC) activation,
respectively (3). The second pathway entails phosphoinositide 3-kinase
(PI 3-kinase) of which the activation produces transient accumulations
of two novel phosphoinositides, phosphatidylinositol
3,4,5-trisphosphate (PtdIns(3,4,5)P3) and
phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P2) (4,
5). This PI 3-kinase pathway has been implicated in diverse
cellular responses to growth factors, including mitogenesis (6, 7),
chemotaxis (8, 9), membrane trafficking (2), actin reorganization (10),
receptor down-regulation (11), cell survival (12), and so forth. These
D-3 phosphoinositides are not substrates for phospholipase C,
indicating that they do not serve as precursors to phosphoinositol
second messengers (13). Although evidence is accumulating that the PI
3-kinase lipid products are important cellular regulators, their
molecular targets remain unclear. Recent evidence indicates that
PtdIns(3,4,5)P3 and PtdIns(3,4)P2 activate
Ca2+-independent PKC family members ,
, and
(14,
15). This stimulatory effect in conjunction with the action of
diacylglycerol is thought to exert a sustained PKC activation (16).
Another PI 3-kinase product, phosphatidylinositol 3-phosphate, was
implicated in the activation of a protein kinase encoded by the
Akt proto-oncogene via a Pleckstrin homology domain (17). In
addition, PtdIns(3,4,5)P3 was shown to disrupt the
association of PI 3-kinase with tyrosine-phosphorylated proteins by
binding to the Src homology 2 (SH2) domains of the p85 subunit (18).
Moreover, PtdIns(3,4,5)P3 and PtdIns(3,4)P2 exhibited high affinity with profilin, which provides a putative link
between PI 3-kinase and actin rearrangement (19).
Conceivably, these D-3 inositol phospholipids initiate downstream signaling by rapidly recruiting and/or activating target proteins. Evidence suggests that the different lipid products of PI 3-kinase activate different molecular targets responsible for diverse physiological consequences (16, 18). Considering the presence of various phosphatidylinositol phosphates and other phospholipids in the plasma membrane, selective recognition of the D-3 phosphoinositides at the interface represents a crucial issue in addressing their mode of action. Thus, to better understand the molecular basis by which the PI 3-kinase lipid products regulate cellular functions, our effort has focused on delineating the mode of recognition of these phosphoinositides. In this study, by using a gel filtration-based binding analysis, we have identified a synthetic 17-mer peptide W-NG28-43 that displays discriminative affinity with PtdIns(3,4,5)P3. It is worthy to note that this PtdIns(3,4,5)P3-binding peptide corresponds to the PKC phosphorylation and calmodulin-binding domain of a brain-specific protein neurogranin (20). Our observation that neurogranin also binds PtdIns(3,4,5)P3 with high affinity leads to a hypothesis in which PtdIns(3,4,5)P3 serves as a targeting site to facilitate the phosphorylation of specific PKC substrates. Moreover, this PtdIns(3,4,5)3-binding peptide provides a useful model to delineate the PtdIns(3,4,5)P3-recognition site on putative target proteins.
1-O-(1,2-Di-O-palmitoyl-sn-glycero-3-phosphoryl)-D-myoinositol
3,4,5-trisphosphate (PtdIns(3,4,5)P3) and
1-O-(1,
2-di-O-palmitoyl-sn-glycero-3-phosphoryl)-D-myo-inositol 3,4,-bisphosphate (PtdIns(3,4)P2) were prepared as
previously reported (21). Both synthetic phosphoinositides were
characterized by 1H and 31P NMR and fast atom
bombardment mass spectrometry, in which no appreciable impurity was
detected. PtdIns(4,5)P2, phosphatidylcholine (PtdCho),
phosphatidylethanolamine (PtdEA), phosphatidylinositol (PtdIns), and
phosphatidylserine (PtdSer) were products from Sigma. 1,2-Dioctanoyl-sn-glycerol was obtained from Calbiochem.
Inositol phosphates used in this study were synthesized according to
the reported procedures (22). Colorimetric PKC assay and lissamine rhodamine-labeled PKC and PKA substrates (including glycogen synthase peptide, myelin basic protein peptide4-14, PKC
pseudosubstrate, epidermal growth factor receptor peptide, neurogranin
peptide28-43), epsilon peptide, and Kemptide) were
obtained from Pierce. PKC- was kindly provided by Drs. Alex Toker
and Kiyotaka Nishikawa. Other peptides used in this investigation were
either purchased from commercial sources or synthesized by an Applied
Biosystems Peptide Synthesizer in the Macromolecular Structure Analysis
Facility at University of Kentucky. All custom-synthesized peptides
were characterized by fast atom bombardment mass spectrometry in the Mass Spectrometry Facility at the University of Kentucky. Micelles containing individual phospholipids were prepared by sonicating in
distilled water in a model 1210 Bransonic ultrasonic cleaner for 5 min.
The screening method is based on the principle that binding of a peptide to PtdIns(3,4,5)P3-containing micelles will change its elution profile on gel filtration chromatography. This technique has been applied to the study of protein-phospholipid interactions (23, 24). Individual peptides (125 µM) were incubated with micelles consisting of pure PtdIns(3,4,5)P3 (theoretical concentration, 125 µM), in a final volume of 100 µl, for 30 min at room temperature. The mixture was chromatographed on a Sephacryl S-200 column (1 × 10 cm) equilibrated with 10 mM Tris/HCl, pH 7.5, containing 75 mM KCl. The column was eluted with the same buffer at 0.5 ml/min, and fractions of 0.45 ml were collected. Protein assays were performed by a Coomassie Blue dye-binding method or by UV absorbance at 230 nm. The micelle-bound peptide was eluted in the void volume, and was well separated from the free peptide. The amount of micelle-bound peptide was calculated as the difference between the total amount applied to the column and the amount of the free peptide. The dissociation constant (Kd) was estimated according to Equation 1.
![]() |
(Eq. 1) |
Fluorescence spectra were recorded at 30 °C with a Hitachi F-2000 spectrophotometer. Interactions between W-NG28-43 and various phospholipids were assessed by monitoring the tryptophan fluorescence with excitation wavelength at 292 nm. The buffer used for the fluorescence experiments consisted of 25 mM Tris/HCl and 100 mM KCl, pH 7.5. Individual phosphoinositides, in the form of micelles in 15 µl of distilled water, was gradually introduced into 785 µl of the buffer containing 7 µM W-NG28-43. Within the concentration range of phosphoinositides used, the bulk solution remained clear.
Circular Dichroism SpectroscopyCD spectra were recorded with a JASCO J720 spectropolarimeter at room temperature in a 20 mm path length cell. The solution contained 25 µM W-NG28-43 and 50 µM individual phosphoinositides in 25 mM Tris/HCl and 100 mM KCl, pH 7.5. The following settings were used: wavelength range, 200-250 nm; bandwidth, 1 nm; step resolution, 0.5 nm; scan speed, 10 millidegree/min. Each spectrum represented an average of 10 scans with base-line subtraction.
31P NMR SpectroscopyDecoupled 31P NMR spectra were recorded at 25 °C in 10-mm tubes on a Varian VRX400 spectrometer at 161.9 MHz. Aliquots of a W-NG28-43 solution were introduced into a 1-ml solution containing 0.8 mM PtdIns(3,4,5)P3 and 10% (v/v) deuterium oxide. The spectrum width was 20 kHz with 10-µs pulse width and 0.8-s acquisition time. For each spectrum, 5,000 acquisitions were obtained over a 4,000-s period. External H3PO4 and deuterium oxide were employed as a chemical shift reference and a locking signal, respectively.
Protein Kinase C AssayEffect of
PtdIns(3,4,5)P3 on PKC-mediated phosphorylation of
NG28-43 and other PKC substrates was examined using a
colorimetric method developed by Pierce (25). The reaction mixture (25 µl) contained 30 mM Tris/HCl, pH 7.4, 50 mM
NaCl, 2 mM ATP, 10 mM MgCl2, 0.1 mM CaCl2, 0.002% Triton X-100, PtdSer (1 mg/ml), 1,2-dioctanoyl-sn-glycerol (20 µg/ml), PKC-
(0.1 unit), 300 µM dye-labeled PKC substrate, and varying
amounts of PtdIns(3,4,5)P3. After incubating at 30 °C for 20 min, 20 µl of the mixture were applied to a separation unit
containing an affinity membrane that would retain the phosphorylated peptide. The membrane was washed under reduced pressure with 750 µl
of a phosphopeptide-binding buffer consisting of 0.1 M
sodium citrate, pH 5.0, 0.5 M NaCl, and 0.02% sodium azide
to remove the unreacted peptide. The phosphopeptide was then eluted by
washing the membrane with 600 µl of 15% formic acid. Quantitation of
the phosphorylated product was accomplished by measuring its absorbance at 570 nm in reference to a standard curve constructed from known amounts of the phosphopeptide generated from the reaction.
The bacterial
expression vector for rat neurogranin pDGRC3 (a kind gift from Dr. Dan
D. Gerendasy) was transformed into Escherichia coli BL21
(DE3) (pLysS), and recombinant neurogranin was produced and purified
according to a procedure described by Gerendasy et al. (26).
In brief, The bacterial culture was grown in LB broth containing
ampicillin (100 µg/ml) and chloramphenicol (30 µg/ml) to
A600 nm = 0.4, isopropyl--D-thiogalactoside was added to a final
concentration of 0.4 mM, and the cells were grown for another 6 h. The cells were collected by centrifugation and were suspended in 10 ml of cold lysis buffer consisting of 50 mM
Tris/HCl, pH 7.5, 100 mM NaCl, 2 mM EDTA, 1 mM EGTA, and 50 mM dithiothreitol. The
suspension was frozen and thawed, and sonicated. After removing cell
debris by centrifugation, the crude homogenate was treated with
perchloric acid (final concentration, 2.5% v/v), followed by
trichloroacetic acid (final concentration, 15% w/v). The perchloric acid-soluble, trichloroacetic-insoluble material was dissolved in 50 mM Tris/HCl, pH 7.5, containing 200 mM NaCl, 2 mM EDTA, 1 mM EGTA, and 50 mM
dithiothreitol (washing buffer), and applied to a column containing 15 ml of calmodulin-Sepharose 4B. The column was washed with 5 column
volumes of washing buffer and the adsorbed protein was eluted with 50 ml of elution buffer consisting of 50 mM Tris/HCl, pH 7.5, 200 mM NaCl, 7 mM CaCl2, and 50 mM dithiothreitol. Fractions were collected throughout, and
the pooled solution was subjected to the aforementioned acid
treatments. The resulting pellet was washed twice with ethanol-ether
(1:1) and dissolved in 10 mM Tris/HCl, 75 mM
KCl, and 1 mM dithiothreitol. The homogeneity of the
purified protein was indicated by a single band on SDS-PAGE with silver
staining, and the identity was confirmed by N-terminal sequencing. The
sequence of the first 15 amino acids at the N terminus was
MDCCTESACSKPDDD, which was identical to that reported in the literature
(26). Concentrations of neurogranin were determined by the BCA method
with bovine serum albumin as a standard.
In this study, a gel filtration-based binding
assay was employed to search for PtdIns(3,4,5)P3-binding
peptides. Individual peptides were incubated with micelles consisting
of pure PtdIns(3,4,5)P3, and the mixture was applied onto a
short path Sephacryl S-200 column. In principle, binding to
PtdIns(3,4,5)P3 would be indicated by the co-elution of the
tested peptide with the micelles in the void volume. This expedient
assay obviated the use of radioligands and laborious separation
procedures to detect ligand binding. Numerous unrelated peptides with
lengths ranging from 10- to 20-amino acid residues were subjected to
this analysis. Of more than 100 peptides evaluated, only one peptide
W-NG28-43 displayed tight binding to
PtdIns(3,4,5)P3-containing micelles. Fig. 1
shows the elution profile of the peptide in the presence of
PtdIns(3,4,5)P3 at an 1:1 molar ratio (solid
line). As shown, the peptide existed exclusively in the
micelle-bound form, indicating the high affinity with
PtdIns(3,4,5)P3.
W-NG28-43 was a synthetic peptide with the amino acid sequence of WAAKIQASFRGHMARKK (single-letter amino acid code). It is worthy to note that after deleting the N-terminal Trp, this PtdIns(3,4,5)P3-binding peptide would be identical to NG28-43 that constituted the PKC phosphorylation and calmodulin-binding domain of neurogranin (27). W-NG28-43 was originally designed in our laboratory for the kinetic study of PKC-mediated phosphorylation of the parent peptide by fluorescence spectrophotometry.
In light of the polycationic nature of W-NG28-43 (net
charge, +6), the binding to the acidic phospholipid might be caused by
nonspecific electrostatic interactions. To clarify this speculation, a
number of unrelated Arg/Lys-rich peptides were subjected to the same
binding assay, which included epidermal growth factor receptor peptide
(RKRTLRRL; net charge, +5); peptide (ERMRPRKRQGSVRRRV; net charge,
+7); myelin basic protein peptide4-14 (EKRPSQRSKYL; net
charge, +3); PKC pseudosubstrate peptide (RFARKGSLRQKNV; net charge,
+5); and Kemptide (LRRASLG; net charge, +2). However, none of these
peptides showed appreciable binding to micellar PtdIns(3,4,5)P3, thereby excluding the possibility of
nonspecific ionic interactions.
This gel filtration analysis was repeated at different PtdIns(3,4,5)P3/W-NG28-43 ratios, and the amount of remaining free peptide was plotted against the theoretical concentration of PtdIns(3,4,5)P3 (Fig. 1, inset). Accordingly, the apparent molecular stoichiometry was calculated from the dose-dependence curve to be 1:1, indicating that each W-NG28-43 was associated with 1 PtdIns(3,4,5)P3 at saturation. In addition, from the data of individual experiments, the apparent dissociation constant (Kd) was estimated to be 2 µM.
Further evidence that W-NG28-43 exhibited high affinity
with PtdIns(3,4,5)P3 was provided by the fluorescence
titration experiment (Fig. 2A). As this
peptide contained a tryptophan residue, its interaction with the
phospholipid could be assessed by monitoring the fluorescence with
excitation wavelength set to 292 nm.
As shown in Fig. 2A, PtdIns(3,4,5)P3 quenched the fluorescence of W-NG28-43 in a dose-dependent and saturable manner accompanied by a rapid shift of the fluorescence maximum from 350 nm to 328 nm. Saturation of the fluorescence titration was attained at 1 molar equivalent of PtdIns(3,4,5)P3, which is consistent with that noted in the gel filtration study. Furthermore, preincubation of W-NG28-43 with Ins(1,3,4,5)P4 or Ins(1,4,5)P3, even at a molar ratio of 1:50, did not perturb the fluorescence titration with PtdIns(3,4,5)P3, suggesting that the affinity between W-NG28-43 and inositol phosphates was negligible.
Enhancement of PKC-mediated Phosphorylation of W-NG28-43 by PtdIns(3,4,5)P3Another important issue that warranted investigation was the mode of complex formation between W-NG28-43 and PtdIns(3,4,5)P3. It has been reported that polycationic polypeptides such as histone III-S and neurogranin peptide analog29-47 formed aggregates with mixed micelles containing PtdSer and diacylglycerol in a time-dependent manner (28). Concomitant with the formation of these aggregates there was a progressive loss of substrate phosphorylation by Ca2+-dependent PKC. Conceivably, embedment of the basic peptide into the micelles due to aggregation blocked its access to PKC, thereby preventing the subsequent phosphorylation.
To distinguish between specific peptide-phospholipid binding and
phospholipid aggregation, fully activated PKC was used to examine
the phosphorylation of NG28-43 in the presence of varying
amounts of PtdIns(3,4,5)P3. A number of PKC substrates that
lacked affinity with PtdIns(3,4,5)P3 were also tested as control, which included myelin basic protein peptide4-14, glycogen synthase peptide, PKC pseudosubstrate, epidermal growth factor
receptor peptide, and
peptide. As mentioned, all these polycationic
peptides failed to bind micellar PtdIns(3,4,5)P3 according
to the gel filtration assay.
Fig. 3 shows the dose dependence of PKC activity on
PtdIns(3,4,5)P3 with myelin basic protein
peptide4-14 (open bars) and
NG28-43 (shaded bars) as substrates. The
initial rate of PKC phosphorylation of NG28-43 showed a
significant increase with elevated levels of
PtdIns(3,4,5)P3, up to 3.3-fold at a lipid/peptide ratio of
0.6, while no appreciable modulatory effect was noted with myelin basic
protein peptide4-14 or other peptide substrates (data not
shown) at comparable concentrations. It is worth mentioning that the
enhancement of PKC activity was independent of the preparation of the
phospholipid vesicle/micelle. PtdIns(3,4,5)P3 could be externally added to PtdSer/1,2-dioctanoyl-sn-glycerol
vesicles in a micelle form or subjected to sonication with PtdSer and
1,2-dioctanoyl-sn-glycerol to form mixed vesicles. The
results obtained with these two preparations were virtually
identical.
Clearly, the Ser residue of NG28-43 was accessible to PKC after binding to the phospholipid, indicating that the interaction between these two molecules was highly specific.
Differential Recognition of Phospholipids by W-NG28-43To examine the binding specificity of W-NG28-43, various phospholipids were subjected to the aforementioned analyses, including PtdIns(4,5)P2, PtdIns(3,4)P2, PtdIns, PtdSer, PtdCho, and PtdEA. The respective molecular stoichiometry and Kd values were estimated by the gel filtration assay and summarized in Table I.
|
Among these phospholipids, PtdIns(4,5)P2 cross-reacted with
W-NG28-43, however, with an affinity 5-fold lower than
PtdIns(3,4,5)P3. This cross-reactivity was presumably due
to their largely shared structural motifs. In contrast, the affinity
with PtdIns(3,4)P2 was 40-fold weaker than that of
PtdIns(3,4,5)P3. This structure-activity correlation
suggests the importance of the 5-phosphate in the peptide binding,
which was supported by the 31P NMR examination of the
interaction between W-NG28-43 and PtdIns(3,4,5)P3 (Fig. 4).
The 31P NMR signals for the phosphates at positions 1, 3, 4, and 5 of PtdIns(3,4,5)P3 were noted at 2.92, 3.08, 3,88, and 3.49 ppm, respectively (spectrum a). Due to the size of the peptide, changes in the chemical shifts upon binding were small. Among these phosphate functions, the 5-phosphate exhibited the highest sensitivity to the peptide-induced change in chemical shift, which underscored the involvement of the 5-phosphate in the ionic interactions. For instance, when 0.75 molar equivalents of W-NG28-43 were added, a downfield shift of 0.12 ppm was noted for the phosphate, vis à vis 0.02 ppm (upfield), 0.09 ppm (downfield), and 0.02 ppm (upfield) for 1-, 3-, and 4-phosphates, respectively. No such changes were noted when an equivalent amount of PtdSer was added to the peptide.
Further assessment of the interfacial interactions between W-NG28-43 and various phosphoinositides was conducted by fluorescence and CD spectroscopy. Fig. 2, A-C, compares the fluorescence spectral changes of W-NG28-43 by PtdIns(3,4,5)P3, PtdIns(4,5)P2, and PtdIns(3,4)P2. No significant change in the fluorescence spectrum was noted with other phospholipids examined such as PtdCho, PtdIns, and PtdEA at comparable concentrations. Modifications of the spectrum by phosphoinositides appear to be structurally dependent. In comparison, the molar equivalencies of PtdIns(3,4,5)P3, PtdIns(4,5)P2, and PtdIns(3,4)P2 to attain saturation in the fluorescence titration were 1, 1.6, and 6, respectively, which were in accord with that determined by gel filtration. Although PtdIns(3,4,5)P3 binding caused a significant change in the emission maximum, especially between the molar ratios of 0.46:1 and 0.65:1, no appreciable blue shift was noted for PtdIns(4,5)P2 (Fig. 2B) and PtdIns(3,4)P2 (Fig. 2C). Moreover, the extent of fluorescence quenching by PtdIns(4,5)P2 was greater than the other two counterparts. These results suggest that the modes of binding with W-NG28-43 among these three phosphoinositides differed.
The same conclusion could be drawn by the results from CD spectroscopy.
CD spectra were recorded in the far UV range between 200 and 250 nm (Fig. 5) for W-NG28-43 alone
(curve a) or in the presence of PtdIns(3,4)P2
(curve b), PtdIns(4,5)P2 (curve c),
or PtdIns(3,4,5)P3 (curve d). It is worth
mentioning that these phosphoinositides alone did not give any
appreciable spectra.
As shown, the CD spectrum of the free peptide is characterized by a
weak, negative band centered at 204 nm, which might be attributed to a
random coil conformation (29). The spectra underwent marked changes
when phosphoinositides were added. This result indicates that
W-NG28-43 exhibited different structural behaviors when
moving from a free state to a micelle-bound state at the interface and
that the ligand-induced conformational change was dependent upon the
phosphoinositide structure. However, lack of significant absorption at
222 nm in these spectra indicates that the peptide did not acquire
-helical structures after interfacial binding. When inositol
phosphates or other phospholipids that lacked affinity with the peptide
(e.g. PtdCho and PtdEA) were added, no change in the CD
spectrum was observed.
In view of
the fact that W-NG28-43 represented the PKC
recognition/calmodulin-binding domain of neurogranin, the interaction between PtdIns(3,4,5)P3 and the full-length protein was
assessed by the gel filtration assay. Fig. 6 illustrates
the representative elution profiles of recombinant neurogranin (21.3 µM) alone (A) and with increasing amounts of
PtdIns(3,4,5)P3. By using data determined at six different
lipid/protein ratios, the Kd value and binding
stoichiometry were estimated to be 2.2 ± 0.5 µM and
10, respectively.
In this study we demonstrate the selective recognition of PtdIns(3,4,5)P3 by a synthetic peptide, W-NG28-43. These two molecules form a 1: 1 complex of which the binding affinity (Kd = 2 µM) is in line with that observed in many protein-phosphoinositide interactions (23, 24, 30). Moreover, this small peptide exhibited high degree of stereochemical selectivity in phospholipid binding. Despite largely shared structural motifs, PtdIns(4,5)P2 and PtdIns(3,4)P2 displayed substantially lower affinity (5-fold and 40-fold, respectively) with W-NG28-43 vis à vis PtdIns(3,4,5)P3. Other phospholipids examined, including PtdIns, PtdSer, PtdEA, and PtdCho, showed low or negligible binding affinity. Several lines of evidence rule out the possibility that the binding was due to nonspecific electrostatic interactions or phospholipid aggregation: 1) many unrelated polybasic peptides failed to exert appreciable binding to PtdIns(3,4,5)P3; 2) inositol phosphates such as Ins(1, 3,4,5)P4 and Ins(1,4,5)P3 do not affect the binding; and 3) the Ser residue of W-NG28-43 was accessible to PKC phosphorylation after binding to the phospholipid.
Conceivably, this peptide model provides a useful tool to study the
biomolecular recognition of phosphoinositides and to delineate PtdIns(3,4,5)P3-binding motifs in putative targets such as
the nonconventional PKC isozymes and the SH2 domains on the p85
subunit. Structurally, this 17-mer peptide can be divided into two
discrete regions composed largely of apolar and basic residues,
respectively. It is noteworthy that the sequence of the C-terminal
polybasic segment (i.e. RGHMARKK) bears resemblance to the
consensus sequences for PtdIns(4,5)P2-binding motifs
deduced by Yin and co-workers (24),
(K/R)XXXKX(K/R)(K/R) and
(K/R)XXXXKX(K/R)(K/R), which are present in
phospholipase C isozymes (24, 32) and various actin-regulating proteins
such as profilin, gelsolin, cofilin, and villin (Table
II). More recently, Fukami et al. (33)
reported another consensus sequence by comparing the putative
PtdIns(4,5)P2-binding site of -actinin (Table II) with
the homologous sequences on spectrin, and the pleckstrin homology (PH)
domain of various proteins such as phospholipase C
1,
pleckstrin, Grb 7, Ras-GAP, and racK
. Again, the sequence
RXXXXXXX(H/R/K)XX(X)W(K/R) is
analogous to that of W-NG28-43 with regard to the spacing
of the basic residues.
|
Most of these putative binding peptides listed in Table II were able to
block binding of the parent proteins to PtdIns(4,5)P2 and inhibit the consequent physiological responses. For example, the
two synthetic peptides from gelsolin, gelsolin-(35-149), and gelsolin-(150-169), prevented PtdIns(4,5)P2 from
inhibiting actin filament severing by gelsolin (24, 31). Also, the
putative PtdIns(4,5)P2-binding peptide from -actinin
inhibited phospholipase C activity by competing for substrate binding
(33). Nevertheless, its phospholipase C
2 counterpart
showed a stimulatory effect on the enzyme activity (32). Furthermore,
binding studies indicated that synthetic peptide from the binding
motifs of actin-binding proteins, such as gelsolin-(150-169), bound
PtdIns(4,5)P2 with an affinity comparable to that of the
native proteins (24).
However, no information is available concerning the differential affinity of these PtdIns(4,5)P2-binding motifs toward different phosphoinositides. Our recent study showed that PtdIns(3,4,5)P3 was able to bind actin-regulating proteins such as profilin and gelsolin with an affinity greater than or at least in the same order of magnitude as PtdIns(4,5)P2 (19). Taken together with the fact that these two phosphoinositides are closely related in their structures, it is plausible that the binding motif for PtdIns(3,4,5)P3 and that for PtdIns(4,5)P2 have similar structural features, which may account for the cross-reactivity between the two phosphoinositides. Although principles governing the ligand selectivity remain unclear, it is plausible that the number and composition of amino acid residues located between those basic residues affects the phosphoinositide preference.
To assess the role of the polybasic domain of W-NG28-42 in interacting with the acidic phospholipid, a truncated peptide, WFRGHMARKK, was prepared by deleting a hydrophobic segment from the N-terminal half. It, however, showed extremely low affinity with PtdIns(3,4,5)P3. This finding suggests that the binding is not simply attributed to electrostatic interactions between the basic residues and the charged head group of the phospholipid, and requires a cooperative action between hydrophobic and electrostatic motifs for interfacial recognition. This premise is further collaborated by the spectroscopic data. First, the CD spectra indicate that the peptide undergoes conformational change from a random coil to an ordered structure when it binds to micellar phosphoinositides. Secondly, a significant blue shift coupled with considerable attenuation of the fluorescence intensity shows that the N-terminal Trp physically interacts with the apolar environment at the interface. Consequently, it is reasonable to postulate that the PtdIns(3,4,5)P3-binding motifs consists of two contiguous segments, a polybasic region flanked by a hydrophobic segment for interfacial recognition. Previously, Cantley and co-workers reported that PtdIns(3,4,5)P3 interacted with the SH2 domain on the regulatory subunit of PI 3-kinase (18). It is interesting that the N-terminal SH2 domain on the p85 subunit of human PI 3-kinase (35) contains an internal peptide sequence that bears some resemblance to W-NG28-42 in the terms of the spacing of the stretch of basic residues (Scheme 1).
![]() |
![]() |
![]() |
![]() |
As mentioned, W-NG28-43 corresponds to the phosphorylation
domain of neurogranin and closely resembles that of neuromodulin (ATKIQASFRGHITRKK) (20). Both proteins are physiological relevant PKC
substrates in the hippocampal region of central nerve system, and are
not known to bind phospholipids. It is worthy to note that besides this
homology, these two brain-specific proteins are not related over the
rest of their sequences. Here, we provide evidence that neurogranin
binds PtdIns(3,4,5)P3 with affinity comparable to that of
the partial peptide, though with much larger binding stoichiometry.
Thus, the physiological implication of this selective recognition is
suggested by the dose-dependence of PKC activity on
PtdIns(3,4,5)P3. This enhancement is thought to result from
the substrate-concentrating effect of
PtdIns(3,4,5)P3-containing micelles based on the following
rationales. First, it is known that PKC
is not susceptible to
PtdIns(3,4,5)P3 activation (15). Secondly, the observed
stimulatory effect was specific for W-NG28-43 and was not
noted with other PKC substrates that lacked affinity with
PtdIns(3,4,5)P3. This finding raises a possibility that
PtdIns(3,4,5)P3 serves as a targeting site on the plasma
membrane for neurogranin and neuromodulin to facilitate their
phosphorylation. The investigation of this hypothesis is currently
underway in this laboratory.
We gratefully acknowledge Dr. Dan D. Gerendasy (Scripps Research Institute) for the generous gift of the
bacterial expression vector for neurogranin and Drs. Alex Toker and
Kiyotaka Nishikawa (Harvard Medical School) for providing PKC .