Myristoylation-dependent and electrostatic interactions exert independent effects on the membrane association of the myristoylated alanine-rich protein kinase C substrate protein in intact cells

Protein kinase C regulates the nuclear localization of diacylglycerol kinase-zeta

Diacylglycerol kinases (DGKs) terminate signalling from diacylglycerol by converting it to phosphatidic acid. Diacylglycerol regulates cell growth and differentiation, and its transient accumulation in the nucleus may be particularly important in this regulation. Here we show that a fraction of DGK-zeta is found in the nucleus, where it regulates the amount of nuclear diacylglycerol. Reducing nuclear diacylglycerol levels by conditional expression of DGK-zeta attenuates cell growth. The nuclear-localization signal of DGK-zeta is located in a region that is homologous to the phosphorylation-site domain of the MARCKS protein. This is, to our knowledge, the first evidence that this domain, which is a major target for protein kinase C, can localize a protein to the nucleus. Two isoforms of protein kinase C, but not others, regulate the localization of DGK-zeta. Our results define a cycle in which diacylglycerol activates protein kinase C, which then regulates the metabolism of diacylglycerol by alternating the intracellular location of DGK-zeta. This may be a general mechanism to control mitogenic signals that depend on nuclear diacylglycerol.

Acylation of Escherichia coli hemolysin: a unique protein lipidation mechanism underlying toxin function

The pore-forming hemolysin (HlyA) of Escherichia coli represents a unique class of bacterial toxins that require a posttranslational modification for activity. The inactive protoxin pro-HlyA is activated intracellularly by amide linkage of fatty acids to two internal lysine residues 126 amino acids apart, directed by the cosynthesized HlyC protein with acyl carrier protein as the fatty acid donor. This action distinguishes HlyC from all bacterial acyltransferases such as the lipid A, lux-specific, and nodulation acyltransferases, and from eukaryotic transferases such as N-myristoyl transferases, prenyltransferases, and thioester palmitoyltransferases. Most lipids directly attached to proteins may be classed as N-terminal amide-linked and internal ester-linked acyl groups and C-terminal ether-linked isoprenoid groups. The acylation of HlyA and related toxins does not equate to these but does appear related to a small number of eukaryotic proteins that include inflammatory cytokines and mitogenic and cholinergic receptors. While the location and structure of lipid moieties on proteins vary, there are common effects on membrane affinity and/or protein-protein interactions. Despite being acylated at two residues, HlyA does not possess a "double-anchor" motif and does not have an electrostatic switch, although its dependence on calcium binding for activity suggests that the calcium-myristoyl switch may have relevance. The acyl chains on HlyA may provide anchorage points onto the surface of the host cell lipid bilayer. These could then enhance protein-protein interactions either between HlyA and components of a host signal transduction pathway to influence cytokine production or between HlyA monomers to bring about oligomerization during pore formation.

Membrane-targeting sequences on AKAP79 bind phosphatidylinositol-4, 5-bisphosphate

Protein kinases and phosphatases are targeted through association with anchoring proteins that tether the enzymes to subcellular structures and organelles. Through in situ fluorescent techniques using a Green Fluorescent Protein tag, we have mapped membrane-targeting domains on AKAP79, a multivalent anchoring protein that binds the cAMP-dependent protein kinase (PKA), protein kinase C (PKC) and protein phosphatase 2B, calcineurin (CaN). Three linear sequences termed region A (residues 31-52), region B (residues 76-101) and region C (residues 116-145) mediate targeting of AKAP79 in HEK-293 cells and cortical neurons. Analysis of these targeting sequences suggests that they contain putative phosphorylation sites for PKA and PKC and are rich in basic and hydrophobic amino acids similar to a class of membrane-targeting domains which bind acidic phospholipids and calmodulin. Accordingly, the AKAP79 basic regions mediate binding to membrane vesicles containing acidic phospholipids including phosphatidylinositol-4, 5-bisphosphate [PtdIns(4,5)P2] and this binding is regulated by phosphorylation and calcium-calmodulin. Finally, AKAP79 was shown to be phosphorylated in HEK-293 cells following stimulation of PKA and PKC, and activation of PKC or calmodulin was shown to release AKAP79 from membrane particulate fractions. These findings suggest that AKAP79 might function in cells not only as an anchoring protein but also as a substrate and effector for the anchored kinases and phosphatases.

The calpain-calpastatin system and protein degradation in fusing myoblasts

Calpain (Ca(2+)-activated cysteine protease) induced proteolysis has been suggested to play a role in myoblast fusion. We previously found that calpastatin (the endogenous inhibitor of calpain) diminishes markedly in myoblasts during myoblast differentiation just prior to the start of fusion, allowing Ca(2+)-induced calpain activation at that stage. Here, we show that a limited degradation of some proteins occurs within the myoblasts undergoing fusion, but not in proliferating myoblasts. The protein degradation is observed at the stage when calpastatin is low. Protein degradation within the myoblasts and myoblast fusion are inhibited by EGTA, by the cysteine protease inhibitors calpeptin and E-64d and by calpastatin. The degradation appears to be selective for certain myoblast proteins. Integrin beta 1 subunit, talin and beta-tropomyosin are degraded in the fusing myoblasts, whereas alpha-actinin, beta-tubulin and alpha-tropomyosin are not. A similar pattern of degradation is observed in lysates of proliferating myoblasts when Ca2+ and excess calpain are added, a degradation that is inhibited by calpastatin. The results support the notion that degradation of certain proteins is required for myoblast fusion and that calpain participates in the fusion-associated protein degradation. Participation of calpain is made possible by a change in calpain/calpastatin ratio, i.e., by a diminution in calpastatin level from a high level in the proliferating myoblasts to a low level in the differentiating myoblasts. Degradation of certain proteins, known to be responsible for the stability of the membrane-skeleton organization and for the interaction of the cell with the extracellular matrix, would allow destabilization of the membrane and the creation of membrane fusion-potent regions.

Binding of basic peptides to membranes produces lateral domains enriched in the acidic lipids phosphatidylserine and phosphatidylinositol 4,5-bisphosphate: an electrostatic model and experimental results

Direct fluorescence digital imaging microscopy observations demonstrate that a basic peptide corresponding to the effector region of the myristoylated alanine-rich C kinase substrate (MARCKS) self-assembles into membrane domains enriched in the acidic phospholipids phosphatidylserine (PS) and phosphatidylinositol 4,5-bisphosphate (PIP2). We show here that pentalysine, which corresponds to the first five residues of the MARCKS effector region peptide and binds to membranes through electrostatic interactions, also forms domains enriched in PS and PIP2. We present a simple model of domain formation that represents the decrease in the free energy of the system as the sum of two contributions: the free energy of mixing of neutral and acidic lipids and the electrostatic free energy. The first contribution is always positive and opposes domain formation, whereas the second contribution may become negative and, at low ionic strength, overcome the first contribution. Our model, based on Gouy-Chapman-Stern theory, makes four predictions: 1) multivalent basic ligands, for which the membrane binding is a steep function of the mole fraction of acidic lipid, form domains enriched in acidic lipids; domains break up at high concentrations of either 2) basic ligand or 3) monovalent salt; and 4) if multivalent anionic lipids (e.g., PIP2) are present in trace concentrations in the membrane, they partition strongly into the domains. These predictions agree qualitatively with experimental data obtained with pentalysine and spermine, another basic ligand.

Interaction of cytoskeletal proteins with membrane lipids

Rapid and significant progress has been made in understanding lipid/protein interactions involving cytoskeletal components and the plasma membrane. Covalent and noncovalent lipid modifications of cytoskeletal proteins mediate their interaction with lipid bilayers. The application of biophysical techniques such as differential scanning colorimetry, neutron reflection, electron spin resonance, CD spectroscopy, nuclear magnetic resonance, and hydrophobic photolabeling, allow various folding stages of proteins during electrostatic adsorption and hydrophobic insertion into lipid bilayers to be analyzed. Reconstitution of proteins into planar lipid films and liposomes help to understand the architecture of biological interfaces. During signaling events at plasma membrane interfaces, lipids are important for the regulation of catalytic protein functions. Protein/lipid interactions occur selectively and with a high degree of specificity and thus have to be considered as physiologically relevant processes with gaining impact on cell functions.

Determinants of calcineurin binding to model membranes

The biochemical factors that lead to membrane targeting of the Ser/Thr protein phosphatase calcineurin were examined using model phospholipid membranes. The interaction of myristoyl- and non-myristoylcalcineurin with lipid surfaces was investigated as a function of negatively charged phospholipids, diacylglycerol, Ca2+, and calmodulin. The data indicate that calcineurin binding to phospholipid monolayers both is myristoyl-independent and is mediated by anionic phospholipids and/or diacylglycerol. Although the effect of Ca2+ on calcineurin-lipid binding is minor, calmodulin altered the binding of calcineurin to the lipid membrane in a Ca2+-dependent manner. Experiments with a constitutively active form of calcineurin that does not bind calmodulin indicated that the effect required the interaction of calcineurin with calmodulin. Our results suggest that phosphatidylserine, diaclyglycerol, and calmodulin may mediate the lipid binding properties of calcineurin in vivo.

Kinetics of interaction of the myristoylated alanine-rich C kinase substrate, membranes, and calmodulin

Membrane binding of the myristoylated alanine-rich C kinase substrate (MARCKS) requires both its myristate chain and basic "effector" region. Previous studies with a peptide corresponding to the effector region, MARCKS-(151-175), showed that the 13 basic residues interact electrostatically with acidic lipids and that the 5 hydrophobic phenylalanine residues penetrate the polar head group region of the bilayer. Here we describe the kinetics of the membrane binding of fluorescent (acrylodan-labeled) peptides measured with a stopped-flow technique. Even though the peptide penetrates the polar head group region, the association of MARCKS-(151-175) with membranes is extremely rapid; association occurs with a diffusion-limited association rate constant. For example, kon = 10(11) M-1 s-1 for the peptide binding to 100-nm diameter phospholipid vesicles. As expected theoretically, kon is independent of factors that affect the molar partition coefficient, such as the mole fraction of acidic lipid in the vesicle and the salt concentration. The dissociation rate constant (koff) is approximately 10 s-1 (lifetime = 0.1 s) for vesicles with 10% acidic lipid in 100 mM KCl. Ca2+-calmodulin (Ca2+.CaM) decreases markedly the lifetime of the peptide on vesicles, e.g. from 0.1 to 0.01 s in the presence of 5 micrM Ca2+.CaM. Our results suggest that Ca2+.CaM collides with the membrane-bound MARCKS-(151-175) peptide and pulls the peptide off rapidly. We discuss the biological implications of this switch mechanism, speculating that an increase in the level of Ca2+-calmodulin could rapidly release phosphatidylinositol 4, 5-bisphosphate that previous work has suggested is sequestered in lateral domains formed by MARCKS and MARCKS-(151-175).

Identification and characterization of cathepsin B as the cellular MARCKS cleaving enzyme

The importance of regulating the cellular concentrations of the myristoylated alanine-rich C kinase substrate (MARCKS), a major cellular substrate of protein kinase C, is indicated by the fact that mice lacking MARCKS exhibit gross abnormalities of central nervous system development and die shortly after birth. We previously identified a novel means of regulating cellular MARCKS concentrations that involved a specific proteolytic cleavage of the protein and implicated a cysteine protease in this process (Spizz, G., and Blackshear, P. J. (1996) J. Biol. Chem. 271, 553-562). Here we show that p40, the carboxyl-terminal fragment resulting from this cleavage of MARCKS, was associated with the mitochondrial/lysosomal pellet fraction of human diploid fibroblasts and that its generation in cells was sensitive to treatment with NH4Cl. These data suggest the involvement of lysosomes in the generation and/or stability of p40. The MARCKS-cleaving enzyme (MCE) activity was peripherally associated with a 10,000 x g pellet fraction from bovine liver, and it co-purified with the activity and immunoreactivity of a lysosomal protease, cathepsin B. Cathepsin B catalyzed the generation of p40 from MARCKS in a cell-free system and behaved similarly to the MCE with respect to mutants of MARCKS previously shown to be poor substrates for the MCE. Treatment of fibroblasts with a cell-permeable, specific inhibitor of cathepsin B, CA074-Me, resulted in parallel time- and concentration-dependent inhibition of cathepsin B and MCE activity. Incubation of a synthetic MARCKS phosphorylation site domain peptide with purified cathepsin B resulted in cleavage of the peptide at sites consistent with preferred cathepsin B substrate sites. These data provide evidence for the identity of the MCE as cathepsin B and suggest that this cleavage most likely takes place within lysosomes, perhaps as a result of specific lysosomal targeting sequences within the MARCKS primary sequence. The data also suggest a direct interaction between MARCKS and cathepsin B in cells and leave open the possibility that MARCKS may in some way regulate the protease for which it is a substrate.

Electrostatic interaction of myristoylated proteins with membranes: simple physics, complicated biology

Cell membrane association by several important peripheral proteins, such as Src, MARCKS, HIV-1 Gag, and K-Ras, requires nonspecific electrostatic interactions between a cluster of basic residues on the protein and acidic phospholipids in the plasma membrane. A simple theoretical model based on the nonlinear Poisson-Boltzmann equation describes well the experimentally measured electrostatic association between such proteins and the cell membrane.

MARCKS regulates membrane ruffling and cell spreading

The dynamic rearrangement of the actin cytoskeleton is fundamental to most biological processes including embryogenesis, morphogenesis, cell movement, wound healing and metastasis [1]. Membrane ruffling and reversible cell-substratum interactions underlie actin-driven cell movement. Protein kinase C (PKC) stimulates membrane ruffling and adhesion [2], but the mechanism by which this occurs is unknown. Myristoylated alaninerich C kinase substrate (MARCKS) is a PKC substrate that cycles on and off membranes by a mechanism termed the myristoyl-electrostatic switch [3-6]. While at the membrane, MARCKS binds to and sequesters acidic phospholipids including phosphatidyl-inositol-4,5-bisphosphate (PIP2) [7]. MARCKS also binds and cross-links filamentous actin, an activity which is regulated by PKC-dependent phosphorylation and calcium-calmodulin [3]. In this report, we demonstrate that expression, in fibroblasts, of MARCKS containing a mutation which abrogates the myristoyl-electrostatic switch prevents cell spreading. The MARCKS mutant arrests the cell during an early stage of spreading, characterized by profuse membrane blebbing, and prevents the formation of membrane ruffles and lamellae usually found at the leading edge of spreading cells. This defect in the regulation of the actin cytoskeleton is accompanied by a decrease in cell-substratum adhesion. Our results provide direct evidence that MARCKS and PKC regulate actin-dependent membrane ruffling and cell adhesion, perhaps via a PIP2-dependent mechanism.

Understanding covalent modifications of proteins by lipids: where cell biology and biophysics mingle

Much effort has been expended on the in vitro characterization of enzymes that covalently attach lipids to proteins. Less information is available about properties conferred on modified proteins by their attached lipid groups, but biophysical studies of simple model systems have begun to shed light on this issue. Recent evidence suggests that the specificity of lipid modifications may be dependent upon the intracellular compartmentalization of the lipid and protein substrates of lipidating enzymes. The function and targeting of their lipidated products appear to be regulated dynamically through addition or subtraction of lipid moieties, other covalent or noncovalent modifications, as well as several devices that at this point can only be inferred. This field of research illustrates the necessity of integrating cell-biological and biophysical perspectives.