A novel family of phosphatidylinositol 4-kinases conserved from yeast to humans.

Phosphatidylinositolpolyphosphates (PIPs) are centrally involved in many biological processes, ranging from cell growth and organization of the actin cytoskeleton to endo- and exocytosis. Phosphorylation of phosphatidylinositol at the D-4 position, an essential step in the biosynthesis of PIPs, appears to be catalyzed by two biochemically distinct enzymes. However, only one of these two enzymes has been molecularly characterized. We now describe a novel class of phosphatidylinositol 4-kinases that probably corresponds to the missing element in phosphatidylinositol metabolism. These kinases are highly conserved evolutionarily, but unrelated to previously characterized phosphatidylinositol kinases, and thus represent the founding members of a new family. The novel phosphatidylinositol 4-kinases, which are widely expressed in cells, only phosphorylate phosphatidylinositol, are potently inhibited by adenosine, but are insensitive to wortmannin or phenylarsine oxide. Although they lack an obvious transmembrane domain, they are strongly attached to membranes by palmitoylation. Our data suggest that independent pathways for phosphatidylinositol 4-phosphate synthesis emerged during evolution, possibly to allow tight temporal and spatial control over the production of this key signaling molecule.

The mechanism of GTP hydrolysis by dynamin II: a transient kinetic study.

Dynamin II is a 98 kDa protein (870 amino acids) required for the late stages of clathrin-mediated endocytosis. The GTPase activity of dynamin is required for its function in the budding stages of receptor-mediated endocytosis and synaptic vesicle recycling. This activity is stimulated when dynamin self-associates on multivalent binding surfaces, such as microtubules and anionic liposomes. We first investigated the oligomeric state of dynamin II by analytical ultracentrifuge sedimentation equilibrium measurements at high ionic strength and found that it was best described by a monomer-tetramer equilibrium. We then studied the intrinsic dynamin GTPase mechanism by using a combination of fluorescence stopped-flow and HPLC methods using the fluorescent analogue of GTP, mantdGTP (2'-deoxy-3'-O-(N-methylanthraniloyl) guanosine-5'-triphosphate), under the same ionic strength conditions. The results are interpreted as showing that mantdGTP binds to dynamin in a two-step mechanism. The dissociation constant of mantdGTP binding to dynamin, calculated from the ratio of the off-rate to the on-rate (k(off)/k(on)), was 0.5 microM. Cleavage of mantdGTP then occurs to mantdGDP and P(i) followed by the rapid release of mantdGDP and P(i). No evidence of reversibility of hydrolysis was observed. The cleavage step itself is the rate-limiting step in the mechanism. This mechanism more closely resembles that of the Ras family of proteins involved in cell signaling than the myosin ATPase involved in cellular motility.

Regulation of the enzymatic and motor activities of myosin I.

Myosins I were the first unconventional myosins to be purified and they remain the best characterized. They have been implicated in various motile processes, including organelle translocation, ion channel gating and cytoskeletal reorganization but their exact cellular functions are still unclear. All members of the myosin I family, from yeast to man, have three structural domains: a catalytic head domain that binds ATP and actin; a tail domain believed to be involved in targeting the myosins to specific subcellular locations and a junction or neck domain that connects them and interacts with light chains. In this review we discuss how each of these three domains contributes to the regulation of myosin I enzymatic activity, motor activity and subcellular localization.

Correlation between self-association modes and GTPase activation of dynamin.

The GTPase activity of dynamin is obligatorily coupled, by a mechanism yet unknown, to the internalization of clathrin-coated endocytic vesicles. Dynamin oligomerizes in vitro and in vivo and both its mechanical and enzymatic activities appear to be mediated by this self-assembly. In this study we demonstrate that dynamin is characterized by a tetramer/monomer equilibrium with an equilibrium constant of 1.67 x 10(17) M(-3). Stopped-flow fluorescence experiments show that the association rate constant for 2'(3')-O-N-methylanthraniloyl (mant)GTP is 7.0 x 10(-5) M(-1) s(-1) and the dissociation rate constant is 2.1 s(-1), whereas the dissociation rate constant for mantdeoxyGDP is 93 s(-1). We also demonstrate the cooperativity of dynamin binding and GTPase activation on a microtubule lattice. Our results indicate that dynamin self-association is not a sufficient condition for the expression of maximal GTPase activity, which suggests that dynamin molecules must be in the proper conformation or orientation if they are to form an active oligomer.

Essential role of the dynamin pleckstrin homology domain in receptor-mediated endocytosis.

Pleckstrin homology (PH) domains are found in numerous membrane-associated proteins and have been implicated in the mediation of protein-protein and protein-phospholipid interactions. Dynamin, a GTPase required for clathrin-dependent endocytosis, contains a PH domain which binds to phosphoinositides and participates in the interaction between dynamin and the betagamma subunits of heterotrimeric G proteins. The PH domain is essential for expression of phosphoinositide-stimulated GTPase activity of dynamin in vitro, but its involvement in the endocytic process is unknown. We expressed a series of dynamin PH domain mutants in cultured cells and determined their effect on transferrin uptake by those cells. Endocytosis is blocked in cells expressing a PH domain deletion mutant and a point mutant that fails to interact with phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2]. In contrast, expression of a point mutant with unimpaired PI(4,5)P2 interaction has no effect on transferrin uptake. These results demonstrate the significance of the PH domain for dynamin function and suggest that its role may be to mediate interactions between dynamin and phosphoinositides.

Synergistic activation of dynamin GTPase by Grb2 and phosphoinositides.

Hydrolysis of GTP by dynamin is essential for budding clathrin-coated vesicles from the plasma membrane. Two distinct domains of dynamin are implicated in the interactions with dynamin GTPase activators. Microtubules and Grb2 bind to the carboxyl-terminal proline/arginine-rich domain (PRD), whereas phosphoinositides bind to the pleckstrin homology (PH) domain. In this study we tested the effect of different phosphoinositides on dynamin GTPase activity and found that the best activator is phosphatidylinositol 4,5-bisphosphate followed by 1-O-(1, 2-di-O-palmitoyl-sn-glycerol-3-benzyloxyphosphoryl)-D-myo-inositol 3,4,5-triphosphate. Phosphatidylinositol 4-phosphate was a weak activator and phosphatidylinositol 3,4-bisphosphate did not activate GTPase at all. We then addressed the question of whether both domains of dynamin, PRD and PH, can be engaged simultaneously, and determined the effects of dual occupancy on dynamin GTPase activity. We found that Grb2 and phosphatidylinositol 4,5-bisphosphate together increased the dynamin GTPase activity up to 4-fold higher than that obtained by these activators tested separately, and also reduced the dynamin concentration required for half-maximal activities by 3-fold. These results indicate that both stimulators can bind to dynamin simultaneously resulting in superactivation of dynamin GTPase activity. We propose that SH3-containing proteins such as Grb2 bind to the dynamin PRD to target it to clathrin-coated pits and prime it for superactivation by phosphoinositides.

Phosphatidylinositol (4,5)-bisphosphate-dependent activation of dynamins I and II lacking the proline/arginine-rich domains.

Dynamins comprise a family of GTPases that participate in the early stages of endocytosis. The GTPase activity of neuronal specific dynamin I is stimulated by microtubules, negatively charged phospholipid vesicles, and Src homology 3-containing proteins, including Grb2. These activators were previously shown to bind to a proline/arginine-rich domain (PRD) in the carboxyl-terminal region of the enzyme. Dynamin II, which is ubiquitously expressed, had not been purified or characterized previously. In this study, the enzymatic properties of rat dynamin II and of D746, a dynamin II truncation mutant lacking the PRD, have been characterized. Dynamin II has a higher basal activity than dynamin I, but the two types of dynamin are stimulated similarly by microtubules, Grb2, and phospholipids. D746 is not activated by microtubules or Grb2, highlighting the significance of the PRD for these interactions, but it is activated by phospholipid vesicles containing phosphatidylserine or phosphatidylinositol-4,5- bisphosphate. Moreover, in contrast to previous reports, the PRD appears not to be required for phospholipid-stimulated self-assembly of dynamin, which is a key element in the regulation of its activity. Similar results were obtained with bovine brain dynamin I that had been subjected to limited proteolytic digestion to remove the PRD. Our data highlight the potential involvement of dynamin pleckstrin homology domains in the regulation of GTPase activity by phospholipids.

Phosphorylation of dynamin by ERK2 inhibits the dynamin-microtubule interaction.

In the present study we show that purified bovine brain dynamin can be phosphorylated by MAP kinase, ERK2, with a stoichiometry of 1 mol phosphate/mol dynamin. The phosphorylated serine residue is located within the C-terminal 10 kDa of dynamin. Dynamin I phosphorylated by ERK2 can be specifically dephosphorylated by calcineurin but not by protein phosphatase 2A (PP2A). Phosphorylation of dynamin by ERK2 weakens the binding of dynamin to microtubules and inhibits dynamin's microtubule-activated GTPase activity. Stimulation of GTPase activity by either Grb2 or phospholipids was not affected by ERK2 phosphorylation, suggesting that the binding sites for Grb2 and phospholipids do not overlap with that for microtubules.

Domain structure of a mammalian myosin I beta.

We have determined the primary structure of a myosin I (called mammalian myosin I beta, MMI beta) from bovine brain and identified its functional domains. The protein was previously purified from brain and adrenal gland. Several constructs were generated and expressed in Escherichia coli as glutathione S-transferase fusion proteins and the recombinant proteins were recognized by monoclonal antibodies that recognize either "head" or "tail" domains of native myosin I. A gel overlay method was used to confirm that calmodulin binds to the consensus calmodulin-binding sequence in MMI beta. Binding assays were used to detect interaction with anionic phospholipid vesicles. We conclude that MMI beta consists of an amino-terminal 80.5-kDa domain that contains the ATP- and actin-binding sites, followed by an 8.5-kDa domain with three calmodulin-binding sequences and a basic 30-kDa carboxyl-terminal tail segment that binds to anionic phospholipids and membranes.

Tissue distribution and subcellular localization of mammalian myosin I.

Myosin I, a nonfilamentous single-headed actin-activated ATPase, has recently been purified from mammalian tissue (Barylko, B., M. C. Wagner, O. Reizes, and J. P. Albanesi. 1992. Proc. Natl. Acad. Sci. USA. 89:490-494). To investigate the distribution of this enzyme in cells and tissues mAbs were generated against myosin I purified from bovine adrenal gland. Eight antibodies were characterized, five of them (M4-M8) recognize epitope(s) on the catalytic "head" portion of myosin I while the other three (M1-M3) react with the "tail" domain. Immunoblot analysis using antiadrenal myosin I antibody M2 demonstrates the widespread distribution of the enzyme in mammalian tissues. Myosin I was immunolocalized in several cell types including bovine kidney (MDBK), rat kidney (NRK), rat brain, rat phaeochromocytoma (PC12), fibroblast (Swiss 3T3), and CHO cells. In all cases, myosin I was concentrated at the cell periphery. The most intense labeling was observed in regions of the cell usually associated with motile activity (i.e., filopodia, lamellipodia and growth cones). These results are consistent with earlier observations on protozoan myosin I that suggest a motile role for the enzyme at the plasma membrane.

Purification and characterization of a mammalian myosin I.

Myosin I, an actin-dependent force-generating enzyme, has been purified from three mammalian sources: bovine adrenal medulla, adrenal cortex, and brain. The purification procedure includes extraction of tissue with ATP at low ionic strength and coprecipitation with actin, followed by gel filtration on Sepharose 4B, anion-exchange chromatography on Q Sepharose, and affinity chromatography on ATP-agarose. Mammalian myosin I molecules are composed of a heavy chain of 116 kDa and multiple low molecular weight polypeptides identified as calmodulin. The structural and enzymatic properties of adrenal medulla myosin I were further characterized. This enzyme exhibits high K+,EDTA- and Ca(2+)-ATPase specific activities (about 0.2 mumol.min-1 per mg of protein), whereas the Mg(2+)-ATPase activity is very low (1-3 nmol.min-1.mg-1). The Mg(2+)-ATPase of medulla myosin I is activated by F-actin in a Ca(2+)-dependent manner: activity is stimulated 40-fold in the presence of EGTA and 90-fold in the presence of 10 microM Ca2+. Two structural domains of the myosin I heavy chain were identified. A 74-kDa chymotryptic fragment contains the catalytic site, while a 36-kDa polypeptide contains the calmodulin-binding sites. These results indicate that mammalian myosin I is more closely related to myosin I from the avian intestinal brush border than to the enzymes isolated from the protozoans Acanthamoeba and Dictyostelium.

Brain myosin assembly: characterization of aggregation-competent fragments by antibodies.

Six different monoclonal antibodies raised against pig brain myosin were used to characterize aggregation-competent fragments of the rod portion of bovine brain myosin. As a prerequisite, the antibody-binding regions in pig brain myosin were determined, and recognition of the same epitopes in the bovine protein was ascertained. A combination of electron microscopy on rotary shadowed myosin: antibody complexes, immunoblotting of proteolytic rod fragments and immunoelectron microscopy with gold-conjugated antibodies allowed for the following conclusions: (1) Rod fragments lacking as much as 24 kDa at the N-terminal, and approximately 16 kDa at the C-terminal end are still aggregation competent. (2) Brain myosin rods aggregate in an antiparallel fashion. These data contribute to our knowledge on structural features of brain myosin relevant to its presumed functions in brain cells.

The assembly of the rod portion of brain myosin.

Electron microscopy was used to study the structural arrangement of the rod portion of brain myosin under various experimental conditions. At low ionic strength the rod formed spindle-like filaments with continuous 14 nm periodicity. In the presence of KCNS and a high concentration of CaCl2 brain myosin and its rod precipitated in a form of segments displaying both bipolar and unipolar arrangement characteristic of the myosin filaments. Limited proteolytic digestion of the rod with chymotrypsin generated several fragments of molecular masses in the range of 84 kDa to 30 kDa. The 74 kDa fragment appeared to be the shortest one which preserves the ability to form filaments.

Ca2+-calmodulin-dependent polymerization of actin by myelin basic protein.

The interaction between myelin basic protein (MBP) and G-actin was studied under nonpolymerizing conditions, i.e.,2mM HEPES, pH 7.5, 0.1 mM CaCl2 and 0.2 mM ATP. Fluorescence studies using pyrenyl-actin and the measurements of ATP hydrolysis rate show that MBP induces changes in the structure of the actin monomer similar to those occurring during polymerization by salt. Electron microscope observations of the MBP-G-actin complex reveal the presence of filamentous structures which appear as separate filaments or as bundles of filaments in lateral association. These filaments are polar as visualized by attachment of heavy meromyosin. The biochemical data together with electron microscope observations suggest that the binding of MBP to G-actin under non-polymerizing conditions induces an interaction between actin monomers leading to the formation of filamentous structures which may be similar to F-actin filaments. The effects of MBP on G-actin can be reversed by calmodulin in the presence of Ca2+.

Proteolytic fragmentation of brain myosin and localisation of the heavy-chain phosphorylation site.

The heavy chains and the 19-kDa and 20-kDa light chains of bovine brain myosin can by phosphorylated. To localise the site of heavy-chain phosphorylation, the myosin was initially subjected to digestion with chymotrypsin and papain under a variety of conditions and the fragments thus produced were identified. Irrespective of the ionic strength, i.e. whether the myosin was monomeric or filamentous, chymotryptic digestion produced two major fragments of 68 kDa and 140 kDa; the 140-kDa fragment was further digested by papain to yield a 120-kDa and a 23-kDa fragment. These fragments were characterised by (a) a gel overlay technique using 125I-labelled light chains, which showed that the 140-kDa and 23-kDa polypeptides contain the light-chain-binding sites; (b) using myosin photoaffinity labelled at the active site with [3H]UTP, which showed that the 68-kDa fragment contained the catalytic site, and (c) electron microscopy, using rotary shadowing and negative-staining techniques, which demonstrated that after chymotryptic digestion the myosin head remains attached to the tail whereas on papain digestion isolated heads and tails were observed. Thus the 120-kDa polypeptide derived from the 140-kDa fragment is the tail of the myosin, and the 68-kDa fragment containing the catalytic site and the 23-kDa fragment, with the light-chain-binding sites, form the head (S1) portion of the myosin. When [32P]-phosphorylated brain myosin was digested with chymotrypsin and papain it was shown that the heavy-chain phosphorylation site is located in a 5-kDa peptide at the C-terminal end of the heavy chain, i.e. the end of the myosin tail. Using hydrodynamic and electron microscopic techniques, no significant effect of either light-chain or heavy-chain phosphorylation on the stability of brain myosin filaments was observed, even in the presence of MgATP. Brain myosin filaments appear to be more stable than those of other non-muscle myosins. Light-chain phosphorylation did, however, have an effect on the conformation of brain myosin, for example in the presence of MgATP non-phosphorylated myosin molecules were induced to fold into a very compact folded state.

Interaction of tropomyosin with myelin basic protein and its effect on the ATPase activity of actomyosin.

Myelin basic protein (MBP) binds to both skeletal muscle and brain tropomyosin resulting in the formation of paracrystalline tactoids in the absence of divalent cations and at neutral pH. Both types of tropomyosin reduce the inhibition of the ATPase activity of actomyosin caused by MBP. On the other hand, MBP alters the effect of both brain and skeletal muscle tropomyosins on the actomyosin ATPase, even though MBP and tropomyosin bind independently to actin. We conclude that MBP cannot substitute for troponin I in the regulation of the action of tropomyosin on actin.

Ca2+-calmodulin-dependent regulation of F-actin-myelin basic protein interaction.

Myelin basic proteins (MBP) interacts with F-actin resulting in the precipitation of a complex of both proteins. Electron microscope observations of this complex reveal the presence of ordered bundles of F-actin filaments similar to those obtained from F-actin and troponin I. In addition to the bundles, there also appear short fragments of F-actin filaments. In the presence of Ca2+ calmodulin causes a release of MBP from its complex with F-actin, accompanied by dissociation of F-actin bundles into separate filaments. Parallel to the binding of MBP to F-actin the ATPase activity of actomyosin is progressively reduced. This inhibition is reversed by calmodulin but only in the presence of Ca2+. Studies of the binding of S-1 to F-actin and to the F-actin-MBP complex indicate that the interaction sites for MBP and S-1 on the actin molecule are different.