| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
* Department of Neurology, Ernest Gallo Clinic and Research Center, University of California San Francisco, San Francisco, California 94110-3518, USA; and
Center for Experimental Therapeutics and Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-6160, USA
1Correspondence: M.G.K., Center for Experimental Therapeutics and Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104-6160, USA. E-mail: marcelo{at}spirit.gcrc.upenn.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
Key Words: PKC • signal transduction • phorbol ester • chimaerin • anchoring proteins
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
. Over the last two decades many
laboratories have contributed to the elucidation of the molecular basis
of PKC activation and regulation and, in an attempt to elucidate its
function, have provided a wealth of information about PKC. PKC has been
identified as the cellular receptor for the lipid second messenger
diacylglycerol (DAG), and is therefore a key enzyme in the signaling
mechanisms by activation of receptors coupled to phospholipase C, which
leads to a transient elevation in DAG levels. The discovery that PKC is
a high-affinity receptor for the phorbol ester tumor promoters
established the basis of its involvement in multistage carcinogenesis
and has provided us with powerful pharmacological tools with which to
manipulate PKC both in vitro and in cellular systems
(2)
. Initial purification experiments and subsequent
molecular cloning studies revealed that PKC is a multifamily of
lipid-regulated serine-threonine kinases that phosphorylates a variety
of cellular proteins and plays an essential role in signal transduction
mechanisms. The identification of specific substrates for PKC isozymes
and the characterization of its functional role form an active area of
research. The early dogma for the regulation of PKC activity and function has been challenged in the last few years. The initial concepts were as follows: 1) Ca2+ and/or lipids are the only regulators of PKC; 2) PKC isozymes translocate upon activation from the cytosol to the plasma membrane; and 3) PKC isozymes are the only cellular receptors for the phorbol esters and the second messenger DAG. These concepts were insufficient to explain the differences observed in cellular localization and function of individual PKC isozymes, nor did they explain the heterogeneity observed in the cellular responses of the phorbol esters.
This review will summarize the current status of the regulation of PKC
activity, including structural aspects and the classical model of
activation, and emphasize recent findings that have increased our
knowledge of PKC regulation and phorbol ester actions. First, we will
focus on the regulation of PKC activity by serine/threonine and
tyrosine phosphorylation. Second, the regulation of PKC activity and
function by subcellular localization via protein–protein interactions,
and the newly identified domains that mediate these interactions, will
be discussed. Finally, we will focus on the novel phorbol ester/DAG
receptors lacking kinase activity (
- and ß-chimaerin isoforms,
Unc-13/Munc13 isoforms, Ras-GRP) as potential mediators of phorbol
ester/DAG actions.
|
STRUCTURE OF PKC ISOZYMES |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
4
5
).
PKC isozymes have been grouped into three subclasses according to their
regulatory properties, which are conferred by specific domains in the
proteins (Fig. 1
). The `conventional' or `classical' PKCs (cPKCs) include PKC ,
ßI, ßII, and 
. These isoforms can be activated by
Ca2+ and/or by DAG and phorbol esters. The
`novel' PKCs (nPKCs) 
, 
, 
, and 
can also be activated by
DAG and phorbol esters but are Ca2+ independent.
Finally, the atypical PKCs, which include PKC
and PKC
(its mouse
homologue has been named PKC
), are unresponsive to
Ca2+ and DAG/phorbol esters. A related enzyme,
PKCµ or PKD, displays multiple unique features that makes it a
distant relative of the PKC isozymes (6
, 7)
. Each PKC
isozyme is the product of a separate gene with the exception of PKCßI
and ßII, which are alternative spliced variants of the same gene.
|
Each PKC isozyme consists of a single polypeptide chain having two structurally well-defined domains: the amino-terminal regulatory domain and the carboxyl-terminal catalytic domain. The regulatory region possesses the motifs involved in the binding of the phospholipid cofactors and Ca2+ and participates in protein–protein interactions that regulate PKC activity and localization. The carboxyl-terminal region is the kinase domain and includes motifs involved in ATP and substrate binding. The regulatory and the catalytic domains are connected by a hinge region that is highly sensitive to proteolytic cleavage by cellular proteases.
The enzymes possess regions that are highly conserved between different
PKC isozymes (regions C1 to C4) and variable regions (regions V1 to
V5). The C1 region is present in all PKC isozymes. It contains an
autoinhibitory domain or pseudosubstrate that binds to the substrate
binding site in the catalytic domain and keeps the enzyme in an
inactive state in the absence of cofactors and activators
(8)
. The amino acid sequence of the pseudosubstrate
resembles that of the phosphorylation motifs in PKC substrates, but
possesses a nonphosphorylated amino acid (i.e., alanine) instead of
serine or threonine (9)
. A distinct feature of the C1
region is the presence of the cysteine-rich domains, which are involved
in binding of the second messenger DAG and phorbol esters in cPKCs and
nPKCs (10
11
12
13)
.
While the cPKCs and nPKCs have two copies of these motifs in tandem,
only a single copy is found in the aPKCs. A single cysteine-rich motif
is also present in other proteins, including - and ß-chimaerins,
Unc-13, Munc13 isoforms, Ras-GRP, Vav, and Raf (14
15
16
17
18
19)
(Fig. 1)
.
The cPKCs possess a C2 region involved in Ca2+
binding immediately carboxyl-terminal to the cysteine-rich domains. A
C2-like domain is present close to the amino-terminal end in nPKCs
(20)
, although this domain is unable to bind
Ca2+ (see below).
PKCµ/PKD and PRKs: kinases related to PKC
A novel serine-threonine kinase regulated by DAG and phorbol
esters was simultaneously identified by two groups and named PKCµ
(human) or PKD (mouse) (6
, 7)
. Although originally
considered a new member of the PKC family, PKCµ/PKD differs from PKC
isozymes in its regulation and substrate selectivity. PKCµ/PKD
contains an amino-terminal putative transmembrane domain, a C1 region
with two cysteine-rich domains that bind phorbol esters and DAGs
(6
, 7
, 21)
, and a Pleckstrin homology (PH) domain (Fig. 1)
. No pseudosubstrate domain has been identified in PKCµ/PKD. The
catalytic domain exhibits some degree of similarity with members of the
PKC family, but is related more to the kinase domains of myosin
light-chain kinase of Dictyostelium and
Ca2+/calmodulin-dependent kinase II
(7)
. PKCµ/PKD does not catalyze significant
phosphorylation of PKC substrates, including PKC pseudosubstrate-based
peptides, histone, myelin basic protein, or protamine. The synthetic
peptide syntide 2, a substrate of calmodulin-dependent kinases, is
efficiently phosphorylated by PKCµ/PKD (22)
. This unique
pattern of substrate recognition and its Golgi localization
(23)
suggest important functional differences between this
novel kinase and the PKCs.
PKCµ/PKD is activated in cells through PKC-mediated phosphorylation
and can therefore function downstream of PKCs (24)
. An
additional form of regulation is through its PH domain; its deletion
results in a marked increase in the basal activity of the enzyme,
suggesting that the PH domain of PKCµ/PKD plays an inhibitory role in
the regulation of its enzymatic activity (22)
. PKCµ/PKD
may act as a scaffold protein for enzymes engaged in phosphoinositide
synthesis, an effect that requires an intact amino-terminal region,
including its transmembrane domain (25)
.
A second group of kinases with homology to PKC, named PRKs (PKC-related
kinases), was recently described (3
, 26
, 27)
. PRK isoforms
possess a kinase domain with homology to the kinase domain in PKCs
(Fig. 1)
. Although these distant relatives of PKC isoforms, which lack
a C1 domain, do not respond to phorbol esters or
Ca2+, they can be activated by the acidic
phospholipids phosphatidylinositol 4,5-bisphosphate and
phosphatidylinositol 3,4,5-triphosphate (3
, 27)
. A
distinct feature of PRKs is that their kinase activity is directly
regulated by small GTP binding proteins. PRKs possess a binding site
for the Rho small GTP binding proteins and belong to the family of
Rho-activated kinases that includes rhotekin and rhophilin (3
, 28
, 29)
.
|
PROTEIN KINASE C DOMAINS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
31
32
33
34)
.
The work pioneered by Blumberg and colleagues has provided important
insights into the phorbol ester binding site in PKC isozymes and its
interaction with ligands. Early experiments using
[3H]PDBu as a radioligand revealed that PKC
isozymes bind phorbol esters with high affinity in the presence of
phosphatidylserine as a cofactor (35)
and that DAG and
phorbol esters bind to the same site (36)
. Recombinant
cPKC and nPKC isoforms expressed in Sf9 insect cells show similar low
nanomolar affinities for [3H]PDBu, suggesting a
high degree of conservation in the binding site between the different
PKC isozymes (37)
.
The first implication that cysteine-rich domains act as phorbol ester
binding sites originated from experiments performed in Nishizuka's
laboratory, in which either deletion of those motifs in PKC or
mutation of conserved cysteines resulted in loss of phorbol ester
binding (13)
. A deletion analysis by Quest et al. revealed
a minimum domain of 43 amino acids required for phorbol ester binding
(38)
. Each of these cysteine-rich domains possesses the
motif
HX12CX2CX13/14CX2CX4HX2CX7C,
where H is histidine, C is cysteine, and X is any other amino acid.
Each domain tightly binds two atoms of Zn2+,
resulting in a stoichiometry of 4 Zn2+ per
molecule of cPKC or nPKC (39)
.
Through a combination of structural analysis and site-directed
mutagenesis, the molecular basis of phorbol ester binding to PKC has
been established (12
, 40
41
42)
. Each cysteine-rich domain
in PKC folds into a globular structure, and phorbol esters bind in a
groove formed by pulling apart two ß-sheets. Ligand binding does not
induce significant changes in the conformation of the cysteine-rich
domain, but rather `caps' a hydrophilic site at the top of the
structure forming a contiguous hydrophobic region that promotes
insertion of the domain into the lipid bilayer (Fig. 2
) (42)
. Essential residues for phorbol ester binding are
missing in PKC and other cysteine-rich domain-containing proteins
such as Raf or Vav, which compromise ligand binding (17
, 43)
. Although initial reports have suggested that Vav is a
phorbol ester receptor (44)
, it is now clear that Vav,
although it preserves the Zn2+ binding
characteristics, is unable to bind phorbol esters or related ligands
even with low affinity (17)
.
|
The issue of equivalence between the cysteine-rich motifs in PKC is a
subject of controversy and deserves careful attention. Several
laboratories have reported that the stoichiometry of phorbol ester
binding is ~1 mol of phorbol ester/mol of PKC (35
, 45)
.
These results contrast with experiments that show that each individual
cysteine-rich domain is able to bind phorbol esters
(10
11
12
13)
. Whether each cysteine-rich domain binds phorbol
esters with similar affinity remains controversial.
In a series of papers, Stubbs and co-workers have proposed that PKC
contains two distinct binding sites with low and high affinity for the
fluorescent ligand sapintoxin D and that distinct ligands
differentially compete for each site (46
47
48)
. The level
of PKC activation that can be achieved by DAG together with phorbol
ester is greater than that observed with each individual activator at
maximally stimulating concentrations. The nonequivalency of the
cysteine-rich domains was also observed in yeast proliferation assays
using deletion mutants of PKC (49)
. Studies with PKC,
where each individual cysteine-rich domain has been mutated (Pro to Gly
11 in the motif), have revealed a differential pattern of
translocation from cytosol to particulate fraction for each mutant
(50)
. Moreover, these mutants show striking differences in
their ability to translocate in response to different PKC ligands
(51)
. Each cysteine-rich domain may have a different
contribution to the activation of the enzyme. The dissociation between
the affinity for binding and the potency for translocation suggests a
complex mechanism for regulating membrane association and requires
further analysis.
The C2 domain: binding site for Ca2+, lipids, and
proteins
The C2 domain in cPKCs is found immediately
carboxyl-terminal to the cysteine-rich domains and is a
Ca2+ binding site. Although first described for
the cPKCs, the C2 domain has been recognized as a widespread domain. A
great number of proteins containing C2 domains have been identified to
date, most of them related to signal transduction mechanisms or
membrane trafficking. Representative proteins containing C2 domains
include synaptotagmins, phospholipase C isozymes, cytosolic
phospholipase A2, rabphilin-3A, Unc-13, and
Munc13 isoforms. Despite a marked variation in primary sequences,
structural analysis has revealed that all C2 domains fold similarly
into a structure consisting of two four-stranded antiparallel
ß-sheets connected by variable loops at the end of each strand, with
the Ca2+ binding site located at one end of the
domain (52
53
54)
. It has been shown for PKCß and
synaptotagmin that five conserved Asp residues are involved in the
coordination of two Ca2+ ions (53)
.
The C2 domain acts as a membrane docking module, where the
Ca2+ ions and basic residues contribute to
electrostatic membrane binding.
Although not initially recognized, it has now been shown that nPKCs
possess a C2-like domain that is unable to bind
Ca2+ (20)
. Important Asp residues
required for Ca2+ binding are not present in the
C2-like domain of nPKCs. It is believed that this domain in nPKCs may
be involved in phospholipid binding and regulates lipid activation of
nPKCs. A related lipid binding domain has also been reported in PRKs
(3
, 27)
.
The C2 domain of PKCs has a dual role in the regulation of PKC activity. In addition to its proposed lipid or Ca2+/lipid binding sites, this domain regulates protein–protein interactions, as described later.
The kinase domain
The catalytic domain of PKC isozymes includes the C3 and C4
domains. The C3 domain possesses the binding site for ATP, the
phosphate donor for phosphotransferase activity; the C4 domain
possesses the binding site for substrates. When PKC is maintained in an
inactive state, the pseudosubstrate occupies this site, blocking the
binding of substrates. Structural information is not available for the
PKC kinase domain. However, a modeled structure of the catalytic domain
based on the 3-dimensional structure of cyclic AMP-dependent kinase
(PKA) has been postulated (4
, 55)
.
Although many substrates for PKC have been identified, synthetic
peptides based on the pseudosubstrate region are efficiently
phosphorylated by PKCs (37
, 56)
. In elegant studies using
peptide libraries, Cantley and co-workers have provided important
information on the substrate selectivity for individual PKC isozymes.
For most PKC isozymes, the predicted optimal sequences from position
-3 to +2 relative to the phosphoserine were shown to be similar to the
corresponding endogenous pseudosubstrate sequences, with some
variations. In the library screening, all PKC isozymes (except PKCµ)
show high selectivity for peptides with basic amino acids at position
-6, -4, and -2 and variable residues in other positions
(56)
.
|
REGULATION OF PKC ACTIVITY BY COFACTORS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
, 57
58
59
). Experiments using mixed micelles or
lipid bilayers have determined that acidic phospholipids are efficient
cofactors for PKC activation, with a remarkable selectivity for
phosphatidylserine. In addition to serving as an anchoring molecule
through its binding to the cysteine-rich domains, DAG causes a dramatic
increase in the affinity of cPKCs and nPKCs for phosphatidylserine
(5
, 58)
. In addition, binding of DAG to PKC confers
selectivity for the phosphatidylserine headgroup among the different
acidic phospholipids (60
61
62)
.
The accepted model of activation of PKC by lipids is that on binding of
DAG (or phorbol esters) in the presence of the phospholipid cofactor, a
conformational change in PKC results in the removal of the
pseudosubstrate from its binding site and in the activation of the
enzyme. It is believed that the cysteine-rich and C2 domains are not
the only regions involved in phospholipid binding: the pseudosubstrate
domain, once removed from its binding site, may also contribute to
membrane binding through its basic residues (8
, 57
, 61
, 62)
. Membrane association is reflected as a shift in subcellular
localization or `translocation' of cytosolic PKC to membrane
compartments in cellular systems, a process that is also tightly
controlled by protein–protein interactions (see below).
In addition to phospholipids, other lipids can stimulate PKC
activity. Free fatty acids synergize with DAG for PKC activation
(63)
, and lipids such as short-chain phosphatidylcholine
derivatives, lysophosphatidic acid and phosphatidylinositol
3,4,5-triphosphate can activate PKC isozymes. It is likely that
differential lipid requirements for individual PKC isoforms represent a
potential mechanism to control isozyme specific functions in cells. One
exciting finding supporting this concept is that the acidic
phospholipid phosphatidylglycerol stimulates PKCßII in nuclear
membranes, leading to phosphorylation of nuclear substrates by this
isozyme. Identification of the phosphatidylglycerol binding site at the
carboxyl-terminal region of PKCßII provides the first evidence of a
phospholipid binding site in the catalytic domain, and suggests a high
degree of complexity in the regulation of PKC activity by lipids
(64)
.
Regulation of PKC activity by Ca2+
Although the molecular events mediating
Ca2+-induced activation of cPKCs are not fully
understood, experimental evidence supports a model of allosteric
interaction between Ca2+ and phospholipids.
Ca2+ increases the affinity of cPKCs for anionic
phospholipids (65)
. Association to membranes and
activation of kinase activity are differentially regulated by the
cation, and the concentration of Ca2+ required
for membrane binding is substantially lower than that required for
activation (66)
. The model postulated by Newton suggests
that low concentrations of the cation promote a weak membrane
interaction, which is accompanied by conformational changes that are
not sufficient to promote activation of the enzyme. Higher
Ca2+ concentrations, however, produce a
conformational change in PKC that results in the release of the
pseudosubstrate from its binding site in the catalytic domain, leading
to enzyme activation (66)
. In support of this hypothesis,
a mutation analysis of the C2 domain of PKC shows that each
individual Ca2+ ion and the Asp residues involved
in their binding may have different roles in membrane binding and
activation (67)
.
|
PKC PHOSPHORYLATION |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
Serine/threonine phosphorylation
An emerging new theme in the regulation of PKC function by
phosphorylation involves phosphoinositide-dependent protein
kinase-1 (PDK1). Recently, studies by the laboratories of Parker
(68)
and Toker (69)
, have identified PKC
isozymes as substrates for PDK1. Both groups have elegantly
demonstrated that PDK1 phosphorylates PKC within the activation
loop, which leads to the direct activation of this isozyme (68
, 69)
. Furthermore, recent reports by these two groups, as well as
from Newton and co-workers, suggest that PDK1 may be a universal PKC
kinase, since PDK1 can phosphorylate cPKCs, nPKCs, and aPKCs. PKC
(68)
, PKC (68
, 69)
, PKC, and PKCßII
(70)
are phosphorylated by PDK1 within the activation loop
of the PKC kinase domain. This phosphorylation is an important
regulatory step for PKC activation, since coexpression of PKCßII and
an inactive mutant of PDK1 results in an inactive PKC
(70)
. Furthermore, Thr to Ala mutation on the PDK1
phosphorylation site in PKC, PKC, and PKC results in
kinase-deficient, dominant negative mutants (71)
.
Phosphorylation by PDK1 is followed by autophosphorylation of two
additional sites within the carboxyl terminus of the sequence, namely,
Thr-638 and Ser-657 for PKC (72
, 73)
and Thr-641 and
Ser-660 for PKCßII (74
, 75)
. Mutagenesis of the
autophosphorylation sites to Ala or Glu and dephosphorylation of these
residues have revealed different functions for each site. The first
autophosphorylation occurs on Thr-641 for PKCßII (75)
.
This site is fundamental to the catalytic activity of PKCßII
(75)
, PKCßI (76)
and PKC
(77)
. The importance of Thr-641 in the release of PKCß
from detergent-insoluble fraction to the cytosol has been demonstrated
by Newton and co-workers (78)
. Phosphorylation on the same
site in PKC (Thr-638) is not required for its catalytic activity,
but is important for the duration of activation and the rate of
dephosphorylation of the enzyme (73)
. The second
autophosphorylation site in PKCßII (Ser-660) is important for the
correct folding of the enzyme and plays a regulatory role in substrate,
ATP, and Ca2+ binding (74)
.
Autophosphorylation on the same residue in PKC (Ser-657), however,
is important to lock the fully phosphorylated enzyme in a close
conformation that is resistant to phosphatases (72)
. In
summary, the phosphorylation of the activation loop by PDK1, followed
by autophosphorylation on two additional sites are necessary for the
localization of PKC isozymes and to obtain a catalytically competent
conformation. Recently, Hannun and co-workers reported another function
for PKC autophosphorylation. Using GFP-tagged PKCßII and mutants,
they found that autophosphorylation of PKCßII on residues Thr-641 and
Ser-660 is important for membrane dissociation of activated PKC from
the plasma membrane and return of PKC to the cytoplasm after its
activation (79)
.
Tyrosine phosphorylation
PKC is phosphorylated on tyrosine residues within the
regulatory domain in response to a variety of stimuli, including the
engagement of the FcRI receptors in mast cells (80)
and
PDGF in 32D cells (81)
. Src (82
, 83)
and Lyn
(82)
have been identified as the tyrosine kinases
responsible for this phosphorylation. The role of tyrosine
phosphorylation for PKC activation, however, remains controversial.
Whereas some reports show an absence of an effect on PKC activity
(81)
, others have demonstrated an increase
(84)
as well as a decrease in activity (85)
.
Nishizuka and co-workers have recently reported that PKC, as well as
other PKC isozymes, were phosphorylated on tyrosine residues in
response to stress response such as
H2O2. The
H2O2-induced tyrosine
phosphorylation was mapped to the catalytic domain of PKC. Tyrosine
phosphorylation induced by
H2O2 was sufficient to
induce prolonged PKC activation. (86)
.
|
PKC BINDING PROTEINS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
and the yeast two-hybrid system.
Table 1
summarizes the current information on PKC binding proteins. These
proteins may be grouped together on the basis of the conformation of
the PKC required for binding, cofactor requirement, and whether or not
the proteins are PKC substrates. PKC in an active conformation will
bind to receptors for activated C kinase (RACKs) or substrates that
interact with C kinase (STICKs), whereas inactive PKCs will interact
with A kinase anchoring protein (AKAPs) and 14–3-3 (Fig. 3
A, B). PKC binding proteins may or may not be PKC substrates
(for example STICKs and RACKs, respectively) and may require cofactors
for binding to PKC (STICKS require phosphatidylserine; GAP-43 requires
Ca2+). Such proteins serve many functions (Fig. 3)
, which include localizing inactive (AKAPs) or active (RACKs) PKC
isozymes to specific intracellular sites or serving as substrates
(STICKs), shuttling proteins (RACK1), PKC activators (syndecan-4,
F-actin), or PKC inhibitors (14–3-3, Nef).
|
|
RACKs
RACKs were originally identified by Mochly-Rosen and co-workers,
who in the early 1990s proposed that the intracellular localization of
PKC isozymes upon activation is mediated via interaction with
isozyme-specific anchoring proteins (88
, 89)
. RACKs were
first characterized as Triton-X-100 insoluble proteins that bind PKC
isozymes only in the presence of PKC activators (88
, 89)
.
The PKC–RACK interaction is mediated, at least in part, by the C2
region in cPKCs and the C2-like region (within the V1 region) in nPKCs
(90
91
92
93
94
95
96)
. Two RACKs have been identified to date by using
overlay assays: RACK1, which specifically interacts with PKCßII
(97
, 98)
, and RACK2 (ß'-COP), which specifically
interacts with PKC (90)
. Both RACKs bind PKC in its
active conformation. Despite the differences in sequence between RACK1
and RACK2, the two proteins share common features. Although RACK1 and
RACK2 are not PKC substrates, both increase PKC phosphorylation of
substrates (90
, 97)
. RACKs contain `WD40 repeats', a
motif known to mediate protein–protein interactions. The role of WD40
repeats in PKC binding proteins is discussed below. It was recently
observed that RACK1 translocates in response to PKC activation. RACK1
translocates to the same site as activated PKCßII and specifically
associates with this isozyme upon activation. These findings suggest a
potential role for RACK1 as a PKC shuttling protein (93)
.
STICKs
STICKs are an important group of PKC binding proteins that were
discovered by Jaken and co-workers. STICKs require phosphatydilserine
for interaction and are PKC substrates. Phosphorylation of STICKs
regulates their association to PKC. The identified STICKs include
MARCKs, MacMARCKs, -adducin, ß-adducin, -adducin, and clone 72
(SseCKs) (99–103). STICKs are involved in a variety of functions. For
example, adducins are cytoskeletal proteins involved in the interaction
between actin and spectrin. Expression, localization, and
phosphorylated states of -adducin and -adducin have been
correlated with renal tumor progression (99
100
101)
. Clone
72, a major PKC binding protein in REF52 fibroblasts, is involved in
cytoskeleton remodeling and cell growth (102)
. Another
STICK is the serum deprivation response protein (Sdr) that binds and
localizes PKC within the microdomain of caveolae (103)
.
Scaffolding proteins
Scaffolding proteins cluster signaling proteins, thus allowing a
tight control of cellular pathways as well as cross talk between
different cascades. Scaffolding proteins such as caveolin, AKAPs, and
INAD cluster PKC to specific intracellular sites. PKC scaffolding
proteins that have been identified to date bind to PKC in its inactive
conformation.
Caveolin
Caveolin targets a variety of signaling proteins such as
-subunits of G-proteins, Src, and EGF receptor to the caveolae
(104
105
106)
. PKC localizes to the microdomain of the
caveolae (103
, 106)
. PKC, in addition to PKC, also
associates with caveolin, whereas PKC was not found to associate
with this protein (106)
. Caveolin interaction with PKC
results in the inhibition of PKC activity. Moreover, a short peptide
derived from caveolin interacts directly with PKC and inhibits its
kinase activity (106)
.
AKAPs
Two members of the AKAP family of proteins that target PKA to a
specific intracellular site are also scaffolding proteins for PKC.
Elegant work by Scott and co-workers has demonstrated that AKAP79
assembles PKA, phosphatase 2B, and PKC at postsynaptic dendritic
fractions (107
108
109)
. AKAP79 inhibits PKC activity
(108)
. Ca2+-calmodulin as well as
DAG release PKC from AKAP79 (109)
. In HEK-293 cells,
AKAP79 is phosphorylated after PMA stimulation (110)
.
Phosphorylation of AKAP79 by PKC regulates its subcellular
compartmentalization (110)
. Thus, there is a reciprocal
regulation of AKAP79 and PKC with regard to subcellular localization.
Another AKAP identified by Scott and co-workers, AKAP250 (Gravin),
assembles both PKA and PKC to the filopodia in human erythroleukemia
cells (111)
.
p62/ZIP (PKCz interacting protein)
p62/ZIP was isolated as a binding protein for the atypical PKCs
(112
, 113)
. p62/ZIP is a poor PKC substrate
(112)
. This protein has been implicated in targeting
atypical PKC isozymes to a lysosome-targeted endosomal compartment
(113)
. Since p62/ZIP is also associated with other
proteins such as the cytokine receptor EBI-3 and p56Lck, it is possible
that p62/ZIP is yet another example of an isozyme-selective scaffolding
protein.
INAD
Zucker and co-workers have identified the Drosophila
INAD as a scaffolding protein that assembles signaling proteins that
participate in phototransduction. These proteins include the
light-sensitive TRP ion channel, calmodulin, rhodopsin, phospholipase
C, and PKC (114
, 115)
. INAD contains PDZ domains that
mediate its interaction with PKC (see below).
14–3-3
14–3-3 is a family of highly homologous proteins that are
ubiquitously expressed. The role of 14–3-3 proteins in PKC signal
transduction remains controversial. 14–3-3 isozymes may either inhibit
(116
, 117)
or enhance PKC activity (118
, 119)
. The dimeric structure of 14–3-3 enables the protein to
serve as an adaptor or scaffold for a variety of signaling proteins,
including PKC (120)
. Meller et al. have reported a
specific interaction between a PKC isozyme (PKC) and a 14–3-3
isozyme (14–3-3
) in T cells that inhibits PKC translocation and
function (121)
. This interaction only occurs with inactive
PKC, which may help keep this PKC isozyme in its inactive
conformation. 14–3-3 may represent an example of a RICK (a receptor
for inactive C kinase), which binds to PKC and targets the inactive
isozyme to specific intracellular sites.
Direct interaction of PKC isozymes with cytoskeletal proteins
PKC isozymes associate with cytoskeletal proteins, and a review on
the subject has recently been published elsewhere (122)
.
The interaction between PKCs and cytoskeletal proteins is at least in
part isozyme-selective. An example of this isozyme specificity is
PKC, which associates with tubulin via the pseudosubstrate region
(123)
. Terrian and co-workers have demonstrated that
PKC specifically binds F-actin via an actin binding site within the
C1 region (124
, 125)
. F-actin activates PKC in the
absence of phospholipids (124)
. PKCßII (but not PKCßI)
also interacts with F-actin via its V5 domain and translocates to the
actin cytoskeleton upon activation (126)
. PKCßII
selectively phosphorylates actin, although actin is a poor substrate.
The interaction of PKCßII with actin results in a significant
enhancement in autophosphorylation and in an alteration in
substrate specificity (126)
. Furthermore, the interaction
between PKCßII and actin protects PKC from degradation and
down-regulation (126)
. The interaction between PKC and
F-actin was also observed in Aplysia (127)
.
Other PKC binding proteins
A number of PKC binding proteins do not share sequence homology or
functional characteristics with the PKC binding proteins described
previously. One such protein that interacts with PKC/ and was
cloned in a yeast two-hybrid screening is LIP (PKC/ interacting
protein). LIP binds to the cysteine-rich domain of PKC/ but not
to PKC, and serves as an activator of the enzyme (128)
.
A second atypical PKC binding protein is PAR-4, which interacts with
the cysteine-rich domains of atypical PKCs and is involved in apoptotic
responses (129)
. Parker and co-workers characterized
another protein, GAP43 (growth-associated protein, neuromodulin), which
is a major PKC substrate in neurons. GAP43 is a PKC binding protein
and interacts with the C2-like domain of this nPKC (130)
,
suggesting that GAP43 may play a role in the subcellular localization
of PKC.
The human immunodeficiency virus protein Nef interacts with PKC in T
cells and inhibits PKC translocation and activity
(131)
. Recently, the nonreceptor tyrosine kinase Fyn was
found to be directly associated with PKC in T cells
(132)
. Since PKC translocates to the T cell receptor
after T cell receptor activation (133)
, it may be possible
that Fyn, which is also known to associate with the T cell receptor,
plays a role as an anchoring protein for activated PKC. The direct
interaction between PKC isozymes and nonreceptor tyrosine kinases is
not restricted to PKC and Fyn. PKC associates with the
nonreceptor tyrosine kinase Src via tyrosine phosphorylation-dependent
and -independent mechanisms (82
, 83)
.
Another PKC interacting protein is syndecan-4, a member of the
transmembrane matrix binding proteoglycans. The cytoplasmic tail of
syndecan-4 interacts with the kinase domain of PKC, which results in
the localization of PKC to focal contacts and in the activation of
the isozyme (134
, 135)
. Furthermore, syndecan-4 is
phosphorylated in response to PKC activation (136)
. The
phosphorylation status of syndecan-4 does not affect its binding to
PKC, although it regulates its activity (137
, 138)
.
Protein interacting domains in PKC binding proteins
Figure. 3B
and Fig. 4
A summarize the current knowledge on the protein interacting
domains in PKC binding proteins.
|
WD40 repeats
WD40 repeats are sequences of approximately 40 amino acids having
conserved amino acids at the amino-terminal (usually Trp or W, and Asp
or D) and at the carboxyl-terminal of the repeat. WD40 repeats have
been found in more than 100 proteins (139)
. The seven WD40
repeats of the ß-subunit of G-proteins make a seven-bladed
ß-propeller, which explains the protein–protein interaction
properties of this motif (140)
. WD40 repeats are found in
RACKs (90
, 97)
. The entire sequence of RACK1 consists of
seven continuous WD40 repeats (97)
. WD40 repeats in RACK2
account for only 40% percent of its sequence (90)
.
Another WD40-containing protein, the telomerase associate protein 1
(TEP1), associates specifically with the PKC isozyme in human breast
cancer cells (141)
. Since RACKs are isozyme-selective PKC
anchoring proteins, it may be possible that TEP1 is a RACK for PKC.
PH domain
Plecktrin homology domains consist of ~100 amino acids, which
bind polyphosphoinositides as well as proteins and are present in a
wide variety of signaling proteins (142)
. Several reports
have shown an association of PKC isozymes with PH domain-containing
proteins (143
144
145
146)
. For example, the PH domain of rac
protein kinase associates with the regulatory domain of PKC
(143
, 144)
. PKCßI associates with the tyrosine kinase
Btk in mast cells via its PH domain, which results in the
phosphorylation of Btk and inhibition of its kinase activity
(146)
.
PDZ domain
The PDZ (PSD-95, Dlg, and Zo-1) domain is found in a
variety of scaffolding proteins that mediate the organization of
signaling networks (147)
. Thus far, three PKC binding
proteins that contain PDZ domains have been identified, namely, PICK1,
INAD, and ASIP, which interact with the V5 region of PKC isozymes.
PICK1 was isolated in a yeast two-hybrid screening, using the catalytic
domain of PKC as bait (148)
. PICK1 localizes PKC to
a perinuclear region (148
, 149)
. The Drosophila
scaffolding protein INAD contains five PDZ domains, two of which
interact directly with Drosophila eye PKC (114
, 115)
. This association occurs in the absence of PKC activation.
The PKC binding site for INAD has been mapped to the carboxyl-terminal
region of PKC, although sites outside the V5 region may also be
involved (115)
. Finally, the third PDZ-containing protein
ASIP, which was recently identified by Ohno and co-workers, is a novel
atypical PKC-specific interacting protein. This protein contains three
PDZ domains. ASIP localizes atypical PKCs to tight junctions and may
play a role in the maintenance of epithelial cell polarity
(150)
.
LIM domain
This domain is a cysteine-rich protein–protein interaction domain
that has been found in more than 60 proteins (151)
. The
LIM domain-containing protein ENH (Enigma homologue) was cloned using
the yeast two-hybrid system and the regulatory domain of PKCß as the
bait (152)
. ENH contains three LIM domains, one of which
interacts with the V1 region of PKCß and PKC (152)
.
Although PKCß phosphorylates ENH, the interaction is not dependent on
the phosphorylation state of ENH (152)
.
Ring finger-containing proteins
Two proteins possessing ring finger zinc binding domains
have recently been identified as PKC binding proteins
(153
154
155)
. RBCK1 (RBCC protein interacting with PKC) was
cloned using the yeast two-hybrid system and the regulatory domain of
PKCß as a bait (153)
. RBCK1 interacts with PKC and
PKCß, and acts as a transcription factor (153
, 154)
.
XAP3 (X-associated protein 3), a protein that binds to the
transcriptional transactivator hepatitis B virus X protein, is another
candidate as an isozyme selective PKC anchoring protein. XAP3 shares
85% homology with RBCK1, and interacts specifically with PKC,
allowing it to come into close proximity with its substrate protein X.
PKC in turn phosphorylates protein X that results in activation of
transcription (155)
.
Protein interacting regions in PKC
The identification of PKC interacting proteins suggests that
specific regions within the structure of PKC should serve as binding
motifs. These motifs, present in either the regulatory or catalytic
domain of PKC isozymes, are depicted in Fig. 4B
. Short
sequences within the kinase domain of PKC have been identified as
interacting motifs for the PDZ-containing proteins PICK1
(149)
and ASIP (150)
, as well as for caveolin
(106)
, F-actin (125)
, and syndecan-4
(134)
.
Using chimeras of the regulatory and catalytic domains of different PKC
isozymes, the laboratories of Blumberg (156)
and Fields
(157)
have provided evidence for a role of the catalytic
domain in targeting PKC isozymes to specific cellular sites. However,
most of the protein binding sequences are localized to the regulatory
domain. The pseudosubstrate sequence has been reported to mediate, at
least in part, the interaction with STICKs (158)
and to
mediate PKC binding to tubulin (123)
. Short sequences
within the C1 region of PKC were identified as F-actin binding
motifs (125)
. The C1 region in atypical PKCs was
identified as the PAR4 binding site (129)
. The C2 region
contains the RACK binding site (97
, 98)
and the GAP43
binding site in PKC (130)
. This C2 region in PKCß
also possesses the pseudo-RACK binding site (159)
, which
is a short sequence that shares high homology to RACK1. This motif may
bind to the RACK binding site, thereby keeping PKC in an inactive
conformation (159)
. Another autoregulatory sequence has
recently been identified in the C1 domain of PKC and PKC. This
C1-derived sequence interacts with the kinase domain of PKC in a
lipid-independent manner (160)
. It is likely that this
motif also contributes to maintaining PKC in an inactive conformation.
|
NOVEL `NONKINASE' PHORBOL ESTER/DAG RECEPTORS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURE OF PKC ISOZYMESPROTEIN KINASE C DOMAINSREGULATION OF PKC ACTIVITY...PKC PHOSPHORYLATIONPKC BINDING PROTEINSNOVEL `NONKINASE' PHORBOL...CONCLUSIONREFERENCES |
|---|
- and ß-chimaerins: phorbol ester receptors with Rac-GAP
activity- and ß-chimaerins, Ras-GRP, and
Caenorhabditis elegans Unc-13 (Fig. 1
and Table 2
). Each of these novel phorbol ester receptors possess a single copy of
the cysteine-rich domain, suggesting that, as previously observed with
PKC deletion mutants and isolated domains, only one copy of the motif
is sufficient for high-affinity binding of phorbol esters and DAG
(11
, 38
, 161)
.
|
The first evidence for a nonkinase phorbol ester receptor was the
cloning of 1-chimaerin (formerly known as n-chimaerin). This 38 kDa
protein is highly expressed in the brain and resembles a `chimera'
between the regulatory domain of PKC and BCR, the breakpoint cluster
region protein involved in the translocation of Philadelphia chromosome
in chronic myelogenous leukemia. Sequence alignment of 1-chimaerin
with known phorbol ester receptors revealed ~40% homology at the
cysteine-rich domain with PKC isozymes (16)
. The initial
report using bacterially expressed 1-chimaerin showed a
Kd of 30–50 nM for [3H]PDBu
in the presence of phosphatidylserine (162)
. The
Kd for Sf9-expressed 1-chimaerin, however, is
substantially lower (Kd=0.2 nM); it is within the
same range of affinities for cPKCs and nPKCs under similar experimental
conditions (163)
. As expected, deletion of the
cysteine-rich domain of 1-chimaerin resulted in complete loss of
phorbol ester binding (162)
, and mutation of cysteines in
chimaerins (which were shown to be essential for phorbol ester binding
in PKC isozymes) completely abolished [3H]PDBu
binding (M. J. Caloca and M. G. Kazanietz, unpublished
results).
A comparison of phorbol ester binding properties of 1-chimaerin with
PKC shows that both receptors are virtually indistinguishable in
ligand binding structure–activity and phospholipid requirement for
phorbol ester binding. Like PKCs, acidic phospholipids reconstitute
binding activity, with phosphatidylserine as the most efficient
phospholipid. The PKC inhibitors calphostin C and sphingosine inhibit
[3H]PDBu binding to 1-chimaerin and PKC
with similar potencies, suggesting that PKC inhibitors acting on the
cysteine-rich domain may not be suitable tools for selective PKC
inhibition (163)
.
The chimaerin family has expanded with the cloning of novel isoforms:
2-, ß1-, and ß2-chimaerins. The different chimaerin isoforms
come from alternative splicing of the - and ß-chimaerin genes
(18
, 164
, 165)
. Using Sf9-expressed ß2-chimaerin,
Caloca et al. (14)
have shown high-affinity
[3H]PDBu binding for this novel receptor.
ß2-Chimaerin is also a high-affinity receptor for the bryostatins, a
unique class of PKC activators. Similar affinities for DAGs were
observed for 1-chimaerin, ß2-chimaerin, and PKCs (14
, 163)
. The flexibility conferred by the rotatable bonds of the
DAGs may allow a better `accommodation' within the binding site of
the different cysteine-rich domains and therefore allow a
correspondingly positive interaction between the active residues of the
pharmacophore and each cysteine-rich domain. An important observation
is that thymeleatoxin, an analog of the second-stage tumor promoter
mezerein, has a markedly lower affinity for ß2-chimaerin than for
PKCs, providing the first ligand capable of distinguishing between
these two classes of phorbol ester receptors (14)
.
The biological roles for the chimaerins and their relation to DAG
signaling are not yet defined. The chimaerins do not possess a kinase
domain, but rather have a carboxyl-terminal GAP (GTPase-activating
protein) domain. Both - and ß-chimaerins accelerate in
vitro the hydrolysis of GTP from Rac, a member of the of the Rho
family of small GTP binding proteins, with little or no effect on Cdc42
and Rho (164
, 166
, 167)
. Thus, chimaerins down-regulate
Rac function.
Although the Rac-GAP activity of chimaerins in vitro has
been well documented, the cellular effects of these phorbol ester
receptors as regulators of Rac-mediated functions in cells remain
almost unexplored. It may be possible that chimaerins inhibit
Rac-mediated responses, such as cytoskeleton organization, activation
of c-Jun amino-terminal kinase (JNK), regulation of cell growth
and cell cycle, malignant transformation, and control of NADPH oxidase
activity in neutrophils (168
169
170
171)
. The differential
tissue expression of chimaerin isoforms suggests unique roles for each
isozyme. The presence of a putative SH2 domain at the amino-terminal
region of 2- and ß2-chimaerin suggests potential interactions with
phosphotyrosine proteins (18
, 164)
. An attractive but
still untested hypothesis is that these phorbol ester receptors may
integrate signals from GTP binding proteins, tyrosine kinase, and DAG
generation.
An important question is whether phorbol esters and/or DAG can activate
chimaerin Rac-GAP activity. Using in vitro GAP assays, Lim
and co-workers have postulated an allosteric model for activation of
1-chimaerin by phorbol esters. However, the data reported by these
authors show only a 10–15% activation by phorbol esters in the
presence of phosphatidylserine (166)
. In another report
with 2-chimaerin, the same authors did not detect any phorbol ester
effect on GAP activity (164)
. 1-Chimaerin GAP activity
can be regulated by phosphatidylserine and phosphatidic acid. In
contrast to PKC, lysophosphatidic acid and fatty acids inhibit the
catalytic activity of 1-chimaerin (164)
. The issue of a
specific regulation of individual chimaerin isoforms by lipids and
phorbol esters needs to be addressed.
A possible second model of regulation is that DAG and the phorbol
esters target the chimaerins to a cellular compartment shared by Rac or
other chimaerin targets. In support of this positional model is the
observation that ß2-chimaerin, like PKCs, is subjected to
translocation by phorbol esters in cells (14)
. These data
also support the finding that a single cysteine-rich domain is capable
of supporting translocation, as determined with isolated cysteine-rich
domains and mutated PKC isozymes (50
, 51
, 172)
.
Translocation of ß2-chimaerin requires higher concentrations of PMA
and has slower kinetics compared to PKC (14)
. A likely
explanation for the difference between both types of receptors is that
important structural motifs controlling translocation, such as
phospholipid binding sites and/or sites of protein–protein
interactions, are missing in chimaerin isoforms. ß2-Chimaerin
translocates to a per
