(Received for publication, March 20, 1997, and in revised form, April 23, 1997)
From the Division of Tumor Cell Regulation, German
Cancer Research Center, D-69009 Heidelberg, Germany and the
¶ Institute of Cell Biology and Immunology, University of
Stuttgart, 70569 Stuttgart, Germany
Protein kinase Cµ is a novel member of the protein kinase C (PKC) family that differs from the other isoenzymes in structural and enzymatic properties. No substrate proteins of PKCµ have been identified as yet. Moreover, the regulation of PKCµ activity remains obscure, since a structural region corresponding to the pseudosubstrate domains of other PKC isoenzymes has not been found for PKCµ.
Here we show that aldolase is phosphorylated by PKCµ in vitro. Phosphorylation of aldolase and of two substrate peptides by PKCµ is inhibited by various proteins and peptides, including typical PKC substrates such as histone H1, myelin basic protein, and p53. This inhibitory activity seems to depend on clusters of basic amino acids in the protein/peptide structures. Moreover, in contrast to other PKC isoenzymes PKCµ is activated by heparin and dextran sulfate. Maximal activation by heparin is about twice and that by dextran sulfate four times as effective as maximal activation by phosphatidylserine plus 12-O-tetradecanoylphorbol-13-acetate, the conventional activators of c- and nPKC isoforms.
We postulate that PKCµ contains an acidic domain, which is involved in the formation and stabilization of an active state and which, in the inactive enzyme, is blocked by an intramolecular interaction with a basic domain. This intramolecular block is thought to be released by heparin and possibly also by 12-O-tetradecanoylphorbol-13-acetate/phosphatidylserine, whereas basic peptides and proteins inhibit PKCµ activity by binding to the acidic domain of the active enzyme.
Protein kinase Cµ
(PKCµ)1 is a
serine/threonine protein kinase that is
phospholipid-dependent and activated by diacylglycerol and
the phorbol ester TPA (1-3). In this respect, PKCµ behaves similarly
as most PKC isoenzymes (cPKCs and nPKCs, for reviews see Refs. 4 and 5)
of the PKC family. However, PKCµ differs in some structural and
enzymatic features from the other PKC isotypes known so far (1-3, 6),
indicating that it may represent a novel subfamily of PKC. Thus, the
two cysteine-rich domains, which serve as binding sites for cofactors
and activators, are much further apart in PKCµ than in all the other
PKCs and in contrast to the other PKC isoenzymes, PKCµ contains a
pleckstrin homology domain and lacks a region corresponding to
pseudosubstrate regions of the PKC family members. Moreover, PKCµ is
not inhibited by a PKC-specific inhibitor (7) and is not down-regulated
upon prolonged TPA treatment of murine keratinocytes and epidermis (8).
As yet, no substrate proteins of PKCµ have been found (3, 9), neither
in vivo nor in vitro, even though numerous proteins are known that are phosphorylated by the other PKC isoenzymes. Recently, Sidorenko et al. (10) claimed that the tyrosine
kinase Syk and the phospholipase C1 were substrates of PKCµ.
However, incorporation of phosphate into these proteins was lower than into myelin basic protein, which was shown by us and others to be an
extremely poor substrate of PKCµ (3, 9). As PKCµ substrates are not
known so far, the function of PKCµ in cellular signaling is for the
most part obscure. Very recently, data were presented suggesting that
PKCµ is located at the Golgi apparatus and is involved in basal
transport processes (11). Moreover, it was suggested that PKCµ
regulates lymphocyte signaling (10).
Here, we report on a possibly differential regulation of the kinase activities of PKCµ and other PKC isoenzymes. In contrast to other PKCs, PKCµ is inhibited by various proteins and peptides, most likely due to clusters of basic residues in their structure, and is activated by heparin and dextran sulfate. Based on these results, the putative role of an acidic domain in the activation of PKCµ is discussed.
TPA was supplied by Prof. Dr. E. Hecker, German
Cancer Research Center, Heidelberg, Germany. Gö6976 and
Gö6983 were kindly provided by Gödecke, A.G., Freiburg,
Germany. The tumor suppressor protein p53 was given by Dr. M. Frey,
German Cancer Research Center, Heidelberg, Germany. Syntide 2, µ-peptide 1, µ-peptide 2, p53-peptide, and the Ser-pseudosubstrate
peptides of PKC,
, and
(
-peptide, [
]-peptide,
[
]-peptide-1, [
]-peptide-2, [
]-peptide) were synthesized by Dr. R. Pipkorn, German Cancer Research Center, Heidelberg, Germany).
Other materials were bought from companies as indicated:
[-32P]ATP (specific activity, 5000 Ci/mmol), Hartman
Analytic (Braunschweig, Germany); aldolase, Boehringer (Mannheim,
Germany); heparin, phosphatidylserine (PS), protamine sulfate, dextran
sulfate, histone H1 (III-S), myelin basic protein,
poly-L-lysine (Mr 15,000-60,000),
poly-L-lysine (Mr 1,000-4,000)
poly-L-arginine (Mr 15,000-60,000),
L-lysine, histamine, quinine, Sigma (Munich, Germany).
Sf 158 cells were infected with recombinant PKCµ baculovirus, and cell extracts were prepared and used as source for PKCµ as described previously (3, 7).
Protein Kinase Cµ AssayPhosphorylation reactions were
carried out in a total volume of 100 µl containing buffer I (50 mM Tris-HCl, pH 7.5, 10 mM -mercaptoethanol), 4 mM MgCl2, 5 µl of a
Sf 158 cell extract containing recombinant PKCµ, 35 µM
ATP containing 1 µCi of [
-32P]ATP and 5 µg of
syntide 2 or µ-peptide 1 as substrates. PS, TPA, heparin,
Gö6976, Gö6983, or various other compounds (see Table I)
were added at concentrations indicated in the legends of the figures
and Table I. After incubation for 7 min at 30 °C, the reaction was
terminated by transferring 50 µl of the assay mixture onto a 20-mm
square piece of phosphocellulose paper (Whatman p81), which was washed
three times in deionized water and twice in acetone. The radioactivity
on each paper was determined by liquid scintillation counting.
Phosphate incorporated into the substrate peptide was obtained by
subtracting values determined in the absence of kinase.
|
These phosphorylations were carried out essentially as
described for the protein kinase Cµ assay. However, no substrate was added for the autophosphorylation, and for the phosphorylation of
aldolase or histone H1, these proteins instead of the substrate peptides were added, at the concentrations indicated in the text. The
assay contained 7 µCi of [-32P]ATP. Proteins of the
reaction mixture were separated by SDS-polyacrylamide gel
electrophoresis and visualized by autoradiography.
No substrate protein of PKCµ, neither in vitro nor
in vivo, has been found so far. In accordance with previous
reports (3, 9), we observed that typical PKC substrates, such as
histone H1, myelin basic protein, and protamine sulfate, were
phosphorylated very weakly by PKCµ and, therefore, cannot be
considered as PKCµ substrates. Recently, the tyrosine kinase Syk and
the phospholipase C1 were claimed to be substrates of PKCµ.
However, phosphorylation by PKCµ in vitro of these
proteins was even weaker than that of myelin basic protein (10). Here
we show that aldolase can serve as a substrate for PKCµ in
vitro. Aldolase was phosphorylated by PKCµ much more effectively
than histone H1 (Fig. 1). To our surprise, when aldolase (5 µg) and histone (5 or 10 µg) were both present in the kinase assay, histone H1 suppressed the phosphorylation of aldolase almost completely and also autophosphorylation of PKCµ
was inhibited (Fig. 1).
To determine the dose dependence of the inhibitory effect of histone H1
on PKCµ, we used syntide 2 as well as a novel synthetic peptide that
we termed µ-peptide 1 as substrates for PKCµ. The µ-peptide 1 with the amino acid sequence RKRYSVDKTLSHPWL, corresponding to the
sequence 825-839 of human PKCµ, proved to be a potent PKCµ substrate. It incorporated around 30% more phosphate than syntide 2 on
phosphorylation with PKCµ.2
Phosphorylation of syntide 2 and µ-peptide 1 by PKCµ was inhibited by histone H1 depending on the concentration of the inhibitor (Fig.
2). Half-maximal inhibition was reached
with 0.2-0.3 µg of histone H1/100 µl (IC50).
To prove the hypothesis that the basic properties of histone H1 were
responsible for the suppression of PKCµ activity, we tested various
proteins and peptides containing basic regions and other basic
compounds (5 µg/100 µl each) for their inhibitory capacity (Table
I). All proteins tested suppressed the
PKCµ-catalyzed phosphorylation, with the exception of the elongation
factor EF-1. Protamine sulfate and histone H1 were most effective in
this respect, followed by myelin basic protein and the human tumor
suppressor protein p53. The p53-peptide with the amino acid sequence
SHLKSKKGQSTSRHKK, corresponding to sequence 367-382 of human p53,
was similarly active as the p53 protein. Peptides derived from the
pseudosubstrate domain of PKC
(IYRRGSRRWRKL) and
(RKRQRSMRRRVH),
which contain serine instead of alanine and therefore serve as
substrates for several PKC isoenzymes, were also found to effectively
inhibit PKCµ. However, the respective peptide derived from the
pseudosubstrate of PKC
(MNRRGSIKQAKI) as well as the µ-peptide 2 (GVRRRRL), corresponding to the amino acid sequence 198-204 of human
PKCµ, did not show such an inhibitory effect. A major difference
between the two peptides and the inhibitory peptides (
-peptide,
-peptide, myristoylated alanine-rich protein kinase C
substrate-peptide, and p53-peptide) exists in the total number and
clustering of basic amino acids (Arg/Lys). The
-peptide and the
µ-peptide 2 contain four basic amino acids and one cluster of two or
four basic amino acids, respectively, whereas the inhibitory peptides
have at least six basic amino acids arranged in two or three clusters.
The myristoylated alanine-rich protein kinase C substrate-peptide
(KKKKKRFSFKKSFKLSGFSFKKSK) with 12 basic residues and three clusters
was the most effective inhibitor peptide. Thus, a peptide might require
a minimal positive net charge and/or specific clusters of basic amino
acids to be able to inhibit PKCµ. In fact, an exchange of one or two
basic amino acids in the
-peptide for neutral residues resulting in
-peptide-1 (IYRRGSIRWRKL) and
-peptide-2 (IYRRGSIRWAKL) caused a
gradual loss of inhibitory activity (Table I).
-peptide-2 has a
similar arrangement of basic amino acids as the
-peptide. EF-1
protein, which does not contain any cluster of basic amino acids even
though it is basic (pI = 9), did not inhibit PKCµ thus further
supporting our notion. The strongly basic polypeptides poly-L-arginine and poly-L-lysine (molecular
weights of 15-60 kDa), and even the smaller poly-L-lysine
(molecular mass of 1-4 kDa) inhibited PKCµ effectively.
L-Lysine and other basic low molecular weight compounds,
such as histamine and quinine, were on the other hand unable to inhibit
PKCµ activity. This indicates that structural features, such as the
above mentioned clusters of basic amino acids, rather than a positive
net charge, determine the suitability of a compound to act as PKCµ
inhibitor, thus pointing to some specificity of the interaction with
the kinase. Most of the proteins and peptides inhibiting PKCµ are
substrates rather than inhibitors of the other PKC isoenzymes, and some
of them, such as protamine and poly-L-arginine, have been
found to activate other PKC isoenzymes (13). On the other hand, none of
the inhibitory proteins and peptides was significantly phosphorylated
by PKCµ. Thus, inhibition of PKCµ was not likely to be due to a
competition of the inhibitory compound with the substrate syntide 2 for
ATP. Inhibition was not reduced by increasing substrate concentrations, as demonstrated in Fig. 3 for the
inhibition by protamine sulfate of syntide 2 phosphorylation by PKCµ.
For comparative purposes, Fig. 3 shows also the inhibition of PKC
by
the pseudosubstrate peptide. In this case, inhibition decreased upon
increasing the concentration of the substrate syntide 2. This clearly
demonstrates that the PKCµ-inhibiting peptides do not act like the
well known pseudosubstrate peptides that inhibit other PKC isoenzymes
by competing with the PKC substrate for its binding site (14). Therefore, we postulate that PKCµ contains an acidic domain,
different from the acidic substrate-binding motif of other PKCs (see
Ref. 15), which is involved in enzyme activation or stabilization of
the active state of the kinase. In the active state PKCµ is inhibited
by proteins and peptides containing clusters of basic residues probably
due to an interaction with this "activating" domain. In the
inactive state the acidic domain might not be accessible, due to an
interaction with an autoregulatory basic domain of the enzyme. Indeed,
PKCµ exhibits a highly acidic domain (amino acid sequence 336-391 of
human PKCµ) in the regulatory part close to the C terminus of the
cysteine-rich regions. This domain contains 40% acidic and just 2%
basic residues and, in a smaller region (342-362), even 48% acidic
residues. It is intriguing that the other PKC isoenzymes lack a
comparable domain.
Our hypothesis would imply that polyanions are able to break up the
autoinhibitory interaction between the acidic and the basic domain. In
fact, the highly sulfated polysaccharides heparin and dextran sulfate
were found to function as potent activators of PKCµ. Maximal
activation of PKCµ by heparin alone, i.e. in the absence
of any other cofactor, was about twice and that by dextran sulfate four
times as effective as maximal activation by PS/TPA (Fig.
4). Dextran sulfate contains more sulfate
groups than heparin and is, therefore probably, more active than
heparin in stimulating PKCµ. Application of PS/TPA together with
heparin or dextran sulfate did not further increase the activity of
PKCµ. Activation of PKCµ by heparin or dextran sulfate was
saturable at low concentrations (Fig. 5).
The Ka values for heparin and dextran sulfate, as
determined by a Lineweaver-Burk plot, were approximately 0.36 and 0.05 µM (based on an average molecular weight of 20,000 and
500,000, respectively, as given by the supplier). Thus, heparin and
dextran sulfate are very effective activators as compared for instance
with diacylglycerol (e.g. the Ka value of
dioctanoylglycerol for PKC is 10 µM, see Ref. 16). As
shown in Fig. 6, A and
B, the maximal velocities (Vmax) of the heparin- and dextran sulfate-activated syntide 2 phosphorylations (23.3 pmol/min and 41.7 pmol/min, respectively) were much higher than
that of the PS/TPA-activated phosphorylation (9.5 pmol/min). On the
other hand, the affinity of the enzyme for the substrate was rather
lower upon activation with the two polyanions (same Km for both: 9.5) than with PS/TPA
(Km: 4.8 µM). Thus, the much more
effective incorporation of phosphate into syntide 2 by the
polyanion-activated PKCµ than by the PS/TPA-activated kinase is due
to an increase in the maximal velocity of the phosphorylation reaction.
As the mechanisms of action of heparin and dextran sulfate are likely
to be identical, we will in the following just deal with heparin.
Autophosphorylation of PKCµ was also more efficiently stimulated by
heparin alone than by PS/TPA (Fig. 7).
Both, heparin- and PS/TPA-activated (7) autophosphorylation could be
strongly suppressed by 1 µM of the PKC inhibitor
Gö6976, but not at all by 1 µM of Gö6983, an
effective inhibitor of the other PKC isoenzymes. These inhibitors are
known to interact with the ATP binding site of PKC.
TPA is generally thought to activate PKC by a conformational change
that results from its binding to the zinc finger regions of the enzymes
(4). As a consequence, an inhibitory pseudosubstrate domain is removed
from the substrate binding site. Whether this mechanism can explain the
activation by TPA/PS of PKCµ remains an open question, since a domain
corresponding to the pseudosubstrate regions of other PKC isoenzymes
has not been found in the PKCµ structure (7). This does not exclude,
however, that upon identification of bona fide in vivo
substrates of PKCµ, a specific pseudosubstrate sequence will be
identified in the future. Within the PKC family the activation by
heparin may, on the other hand, turn out to be a specific feature of
PKCµ, since PKC activity was not affected by heparin (data not
shown) and a PKC preparation from rat brain (containing mainly PKC
,
,
) was even inhibited by heparin (17). The latter result is in
agreement with the finding that heparin might block smooth muscle cell
proliferation by inhibition of PKC
(18). Several other protein
kinases, such as casein kinase 1 and 2, nuclear kinases, the tyrosine
kinase Syk, and G-protein-coupled receptor kinases are inhibited by
heparin (17, 19, 20, and references in Ref. 17). On the other hand,
activation by heparin was reported for instance for a RNA-activated
protein kinase (21) and a Lyn-related tyrosine protein kinase (22). Activation of each of the two kinases by heparin was shown to occur
through mechanisms different from those of other known activators of
these kinases, thus resembling the activation of PKCµ by heparin. Moreover, many growth factors are known to bear specific
heparin-binding sites that contain a cluster of basic amino acid
residues (23, 24).
The activation by heparin or dextran sulfate of PKCµ appears to be rather specific, as other acidic compounds, such as chondroitin sulfate, cholesterol sulfate, double-stranded polyinosinic-polycytidylic acid, DNA (calf thymus), poly-L-aspartic acid, and poly-L-glutamic acid, did not or just very weakly activate PKCµ (data not shown). This supports the notion that heparin and dextran sulfate specifically break up the intramolecular interaction of basic residues with an acidic domain of PKCµ. The apparent specificity of the stimulatory effect may be taken as an indication for a physiological function of heparin or heparin-like compounds in the control of PKCµ activity. Heparin has been shown to affect various intracellular signaling pathways, including PKC-dependent pathways, and to be a potent proliferation inhibitor of several cell types (Refs. 25 and 26 and references therein). However, these effects are thought to be mediated by binding of heparin to cell surface binding sites or growth factors (23, 24, 27). Little is known about possibly direct actions of heparin on signaling pathways inside the cell. As heparin is synthesized in the Golgi complex and PKCµ has recently been shown to be localized there (11), a direct action of heparin on PKCµ in this cellular compartment is conceivable. Alternatively, heparin might mimic the effects of heparin-like factors in vivo. Such factors are produced, for instance, by endothelial and smooth muscle cells and are growth-inhibitory for these cells (28, 29).