Annexin B12 is a sensor of membrane curvature and undergoes major curvature-dependent structural changes

The regulation of membrane curvature plays an important role in many membrane trafficking and fusion events. Recent studies have begun to identify some of the proteins involved in controlling and sensing the curvature of cellular membranes. A mechanistic understanding of these processes is limited, however, as structural information for the membrane-bound forms of these proteins is scarce. Here, we employed a combination of biochemical and biophysical approaches to study the interaction of annexin B12 with membranes of different curvatures. We observed selective and Ca(2+)-independent binding of annexin B12 to negatively charged vesicles that were either highly curved or that contained lipids with negative intrinsic curvature. This novel curvature-dependent membrane interaction induced major structural rearrangements in the protein and resulted in a backbone fold that was different from that of the well characterized Ca(2+)-dependent membrane-bound form of annexin B12. Following curvature-dependent membrane interaction, the protein retained a predominantly alpha-helical structure but EPR spectroscopy studies of nitroxide side chains placed at selected sites on annexin B12 showed that the protein underwent inside-out refolding that brought previously buried hydrophobic residues into contact with the membrane. These structural changes were reminiscent of those previously observed following Ca(2+)-independent interaction of annexins with membranes at mildly acidic pH, yet they occurred at neutral pH in the presence of curved membranes. The present data demonstrate that annexin B12 is a sensor of membrane curvature and that membrane curvature can trigger large scale conformational changes. We speculate that membrane curvature could be a physiological signal that induces the previously reported Ca(2+)-independent membrane interaction of annexins in vivo.     

A novel calcium-independent peripheral membrane-bound form of annexin B12

Annexins are soluble proteins that can interact with membranes in a Ca2+-dependent manner. Recent studies have shown that they can also undergo Ca2+-independent membrane interactions that are modulated by pH and phospholipid composition. Here, we investigated the structural changes that occurred during Ca2+-independent interaction of annexin B12 with phospholipid vesicles as a function of pH. Electron paramagnetic resonance analysis of a helical hairpin encompassing the D and E helices in the second repeat of the protein showed that this region refolded and formed a continuous amphipathic alpha helix following Ca2+-independent binding to membranes at mildly acidic pH. At pH 4.0, this helix assumed a transmembrane topography, but at pH approximately 5.0-5.5, it was peripheral and approximately parallel to the membrane. The peripheral form was reversibly converted into the transmembrane form by lowering the pH and vice versa. Furthermore, analysis of vesicles incubated with annexin B12 using freeze-fracture electron microscopy methods showed classical intramembrane particles at pH 4.0 but none at pH 5.3. Together, these data raise the possibility that the peripheral-bound form of annexin B12 could act as a kinetic intermediate in the formation of the transmembrane form of the protein.     

Calcium-dependent binding of annexin 12 to phospholipid bilayers: stoichiometry and implications

Annexins (ANXs) are a superfamily of proteins whose functional hallmark is Ca2+-dependent binding to anionic phospholipids. Their core domains are usually composed of a 4-fold repeat of a conserved amino acid sequence, with each repeat containing a type II Ca2+ binding site that is generally thought to mediate Ca2+-dependent binding to the membrane. We now report that ANX12 binding to phospholipid vesicles is highly cooperative with respect to Ca2+ concentration (Hill constant approximately 7), thereby suggesting that more than the four well-characterized type II Ca2+ binding sites are involved in phospholipid binding. Two independent approaches, a novel 45Ca2+ copelleting assay and isothermal titration calorimetry, indicate a stoichiometry of approximately 12 mol of Ca2+/mol of ANX12 for binding to phospholipid vesicles. On the basis of the "low-affinity" Ca2+-binding sites in a number of ANX X-ray crystal structures, we propose a model for ANX12 bilayer binding that involves three types of Ca2+ sites in each of the four repeats. In this model, there is a complementarity between the spacing of the ANX12 Ca2+ binding sites and the spacing of the phospholipid headgroups in bilayers. We tested the implications of the model by manipulating the physical state of vesicles composed of phospholipids with saturated acyl chains with temperature and measuring its influence on ANX12 binding. ANX12 bound to vesicles in a Ca2+-dependent manner when the vesicles were in the liquid crystal phase but not when the phospholipid was in the gel phase. Furthermore, ANX12 bound initially to fluid bilayers remained bound when cooled to 4 degrees C, a temperature that should induce the gel phase transition. Overall, these studies suggest that ANX12 is well suited to being a Ca2+ sensor for rapid all-or-none intercellular membrane-related events.     

Calcium- and membrane-induced changes in the structure and dynamics of three helical hairpins in annexin B12

Annexins are a family of soluble proteins that can undergo reversible Ca(2+)-dependent interaction with the interfacial region of phospholipid membranes. The helical hairpins on the convex face of the crystal structure of soluble annexins are proposed to mediate binding to membranes, but the mechanism is not defined. For this study, we used a site-directed spin labeling (SDSL) experimental approach to investigate Ca(2+) and membrane-induced structural and dynamic changes that occurred in the helical hairpins encompassing three of the four D and E helices of annexin B12. Electron paramagnetic resonance (EPR) parameters were analyzed for the soluble and Ca(2+)-dependent membrane-bound states of the following nitroxide scans of annexin B12: a continuous 24-residue scan of the D and E helices in the third repeat (residues 219-242) and short scans encompassing the D-E loop regions of the first repeat (residues 68-74) and the fourth repeat (300-305). EPR mobility and accessibility parameters of most sites were similar when the protein was in solution or in the membrane-bound state, and both sets of data were consistent with the crystal structure of the protein. However, membrane-induced changes in mobility and accessibility were observed in all three loop regions, with the most dramatic changes noted at sites corresponding to the highly conserved serine and glycine residues in the loops. EPR accessibility parameters clearly established that nitroxide side chains placed at these sites made direct contact with the bilayer. EPR mobility parameters showed that these sites were very mobile in solution, but immobilized on the EPR time scale in the membrane-bound state. Since the headgroup regions of bilayer phospholipids are relatively mobile in the absence of annexins, Ca(2+)-dependent binding of annexin B12 appears to form a complex in which the mobility of the D-E loop region of the protein and the headgroup region of the phospholipid are highly constrained. Possible biological consequences of annexin-induced restriction of membrane mobility are discussed.     

Roles of factor Va heavy and light chains in protein and lipid rearrangements associated with the formation of a bovine factor Va-membrane complex.

Factor Va is an essential protein cofactor of the enzyme factor Xa, which activates prothrombin to thrombin during blood coagulation. Peptides with an apparent Mr of approximately 94,000 (heavy chain; HC) and approximately 74,000 or 72,000 (light chain; LC) interact in the presence of Ca2+ to form active Va. The two forms of Va-LC differ in their carboxyl-terminal C2 domain. Using Va reconstituted with either LC form, we examined the effects of the two LC species on membrane binding and on the activity of membrane-bound Va. We found that 1) Va composed of the 72,000 LC bound only slightly more tightly to membranes composed of a mixture of neutral and acidic lipids, the Kd being reduced by a factor of approximately 3 at 5 mM and by a factor of 6 at 2 mM Ca2+. 2) The two forms of Va seemed to undergo different conformational changes when bound to a membrane. 3) The activity of bovine Va varied somewhat with LC species, the difference being greatest at limiting Xa concentration. We have also addressed the role of the two Va peptides in membrane lipid rearrangements and binding: 1) Va binding increased lateral packing density in mixed neutral/acidic lipid membranes. In the solid phase, Va-HC had no effect, whereas Va-LC and whole Va had similar but small effects. In the fluid phase, Va-HC and whole Va both altered membrane packing, with Va-HC having the largest effect. 2) Va-HC bound reversibly and in a Ca2+-independent fashion to membranes composed of neutral phospholipid (Kd, approximately 0.3 microM; stoichiometry approximately 91). High ionic strength had little effect on binding. 3) The substantial effect of Va on packing within neutral phospholipid membranes was mimicked by Va-HC. 4) Based on measurements of membrane phase behavior, binding of Va or its peptide components did not induce thermodynamically discernible lateral membrane domains. These results suggest that the membrane association of factor Va is a complex process involving both chains of Va, changes in lipid packing, and changes in protein structure.     

Novel organization and properties of annexin 2-membrane complexes

Annexin 2 belongs to the annexin family of proteins that bind to phospholipid membranes in a Ca(2+)-dependent manner. Here we show that, under mild acidic conditions, annexin 2 binds to and aggregates membranes containing anionic phospholipids, a fact that questions the mechanism of its interaction with membranes via Ca(2+) bridges only. The H(+) sensitivity of annexin 2-mediated aggregation is modulated by lipid composition (i.e. cholesterol content). Cryo-electron microscopy of aggregated liposomes revealed that both the monomeric and the tetrameric forms of the protein form bridges between the liposomes at acidic pH. Monomeric annexin 2 induced two different organizations of the membrane junctions. The first resembled that obtained at pH 7 in the presence of Ca(2+). For the tetramer, the arrangement was different. These bridges seemed more flexible than the Ca(2+)-mediated junctions allowing the invagination of membranes. Time-resolved fluorescence analysis at mild acidic pH and the measurement of Stokes radius revealed that the protein undergoes conformational changes similar to those induced by Ca(2+). Labeling with the lipophilic probe 3-(trifluoromethyl)-3-(m-[(125)I]iodophenyl)diazirine indicated that the protein has access to the hydrophobic part of the membrane at both acidic pH in the absence of Ca(2+) and at neutral pH in the presence of Ca(2+). Models for the membrane interactions of annexin 2 at neutral pH in the presence of Ca(2+) and at acidic pH are discussed.     

Annexins V and XII insert into bilayers at mildly acidic pH and form ion channels

The functional hallmark of annexins is the ability to bind to the surface of phospholipid membranes in a reversible, Ca(2+)-dependent manner. We now report that human annexin V and hydra annexin XII reversibly bound to phospholipid vesicles in the absence of Ca(2+) at low pH; half-maximal vesicle association occurred at pH 5.3 and 5. 8, respectively. The following biochemical data support the hypothesis that these annexins insert into bilayers at mildly acidic pH. First, a photoactivatable reagent (3-trifluoromethyl)-3-(m-[(125)I]iodophenyl)diazirine) which selectively labels proteins exposed to the hydrophobic domain of bilayers reacted with these annexins at pH 5.0 and below but not at neutral pH. Second, in a Triton X-114 partitioning assay, annexins V and XII act as integral membrane proteins at low pH and as hydrophilic proteins at neutral pH; in the presence of phospholipids half-maximal partitioning into detergent occurred at pH approximately 5.0. Finally, annexin V or XII formed single channels in phospholipid bilayers at low pH but not at neutral pH. A model is discussed in which the concentrations of H(+) and Ca(2+) regulate the reversible conversion of three forms of annexins-soluble, peripheral membrane, and transmembrane.     

Phospholipid binding properties of human placental anticoagulant protein-I, a member of the lipocortin family

Human placental anticoagulant protein-I (PAP-I) is a member of the lipocortin/calpactin/annexin family of Ca2+-dependent phospholipid binding proteins. PAP-I was labeled with fluorescein 5-isothiocyanate (1 mol/mol); this derivative had anticoagulant activity identical to the unlabeled protein and could be used to measure Ca2+-dependent binding to phospholipid vesicles through changes in fluorescence quenching. At 1.2 mM Ca2+, 0.50 M ionic strength, pH 7.4, 25 degrees C, fluorescein-labeled PAP-I bound to phospholipid vesicles containing 80% phosphatidylcholine, 20% phosphatidylserine with a Kd of 1.2 +/- 0.2 nM (mean +/- S.D.). At an ionic strength of 0.15 M, the Kd decreased to less than 0.1 nM. Prothrombin and factor Xa both competed with fluorescein-labeled PAP-I for binding to anionic phospholipid vesicles, but with affinities at least 1000-fold weaker than PAP-I. PAP-I bound only weakly (Kd greater than 2 x 10(-5) M) to neutral or anionic phospholipid monomers, and this binding was not calcium-dependent. These results show that the affinity of PAP-I for anionic phospholipid surfaces is sufficient to explain its potency as an in vitro anticoagulant.     

Expression of annexin A2 in GABAergic interneurons in the normal rat brain

Expression of the Ca(2+)-dependent phospholipids binding protein annexin A2 (ANX2) in the brain is thought to be largely associated with brain pathological conditions such as tumor, inflammation, and neurodegeneration. The recent findings that ANX2 heterotetramer is involved in learning and neuronal activities necessitates a systematic investigation of the physiological expression of ANX2 in the brain. With combination of in situ hybridization and immunohistochemistry, ANX2 mRNA and protein were specifically detected in a group of GABAergic interneurons throughout the brain. Although ANX2 was absent from the interior of pyramidal neurons, it was found on the membrane and seemly the extracellular space of those neurons, where they closely co-localized with glutamate decarboxylase terminals. In cultured developing neurons, ANX2 was present at high concentrations in the growth cones co-distributing with several growth-associated proteins such as growth associated protein 43 (GAP43), turned on after division/Ulip/CRMP (TUC-4), tubulin, and tissue-plasminogen activator. It then became predominantly distributed on the membrane and mostly in axonal branches as neurons grew and extended synaptic networks. ANX2 was also secreted from cultured neurons, in a membrane-bound form that was Ca(2+)-dependent, which was significantly increased by neuronal depolarization. These results may have implications in the function and regulatory mechanism of ANX2 in the normal brain.     

Membrane-bound and soluble extracellular alpha-amylase from Bacillus subtilis.

Extracellular alpha-amylase was purified to homogeneity from a Marburg strain of Bacillus subtilis. The enzyme is a single polypeptide chain of molecular weight approximately 67,000. Its NH2-terminal amino acid sequence is Leu-Thr-Ala-Pro-Ser-Ile-Lys. A membrane-derived alpha-amylase was solubilizing from membrane vesicles by treatment with Triton X-100 and was highly purified by chromatography on an anti-alpha-amylase-protein A-Sepharose column. Membrane-derived alpha-amylase was indistinguishable from the soluble extracellular enzyme by sodium dodecyl sulfate-gel electrophoresis and radioimmunoassay. The membrane-derived enzyme contains phospholipid. Approximately 30 to 80% of the phospholipid was extracted from the purified enzyme by chloroform:methanol. The extracted phospholipid was predominately phosphatidylethanolamine. Treatment with phospholipase D released phosphatidic acid. Membrane-bound alpha-amylase was latent in membrane vesicles. Release of membrane-bound alpha-amylase from vesicles by an endogenous enzyme was maximal at pH 8.5, was inhibited by metal chelators and diisopropyl fluorophosphate and was stimulated by Ca2+ and Mg2+. The amount of membrane-bound alpha-amylase was related to the level of secretion.     

Phospholipid membranes affect tertiary structure of the soluble cytochrome b5 heme-binding domain

The influence of charged phospholipid membranes on the conformational state of the water-soluble fragment of cytochrome b5 has been investigated by a variety of techniques at neutral pH. The results of this work provide the first evidence that aqueous solutions with high phospholipid/protein molar ratios (pH 7.2) induce the cytochrome to undergo a structural transition from the native conformation to an intermediate state with molten-globule like properties that occur in the presence of an artificial membrane surface and that leads to binding of the protein to the membrane. At other phospholipid/protein ratios, equilibrium was observed between cytochrome free in solution and cytochrome bound to the surface of vesicles. Inhibition of protein binding to the vesicles with increasing ionic strength indicated for the most part an electrostatic contribution to the stability of cytochrome b5-vesicle interactions at pH 7.2. The possible physiological role of membrane-induced conformational change in the structure of cytochrome b5 upon the interaction with its redox partners is discussed.     

Optical response of the indicator chlortetracycline to membrane potential

Chlortetracycline is a fluorescent, Ca2+ indicator commonly used to monitor the internal Ca2+ concentration of membrane vesicles and organelles. We have found that the intensity of chlortetracycline fluorescence in the presence of Ca2(+)-loaded liposomes is dependent on the membrane potential of the vesicles as well as the intravesicular Ca2+ concentration. The fluorescence of chlortetracycline was lower when an inside-negative membrane potential was placed across the liposome membrane. Since chlortetracycline diffuses across the membrane in the zwitterionic form, the distribution of chlortetracycline across the membrane should not be strongly dependent on the membrane potential. However, because the proton permeability of phospholipid vesicles is relatively high, the intravesicular proton concentration is dependent on the membrane potential. The binding of Ca2+ to chlortetracycline is dependent on pH in the range of pH 6 to pH 8. Therefore, changes in the intravesicular pH as a result of a change in the membrane potential causes relatively large changes in the chlortetracycline fluorescence signal even when there isn't a change in the Ca2+ concentration.     

Constitutive activity of membrane-inserted protein kinase C

Incubation of purified protein kinase C (PKC) with phospholipid vesicles produced two populations of membrane-bound PKC: one population was dissociated by calcium chelation and the other was not. The second population appeared to be inserted into the membrane. The activity of membrane-inserted PKC was Ca2+-independent and was only modestly sensitive to phorbol esters. Insertion was caused by high calcium concentrations or by phorbol esters plus low calcium. These conditions correlated with those needed to activate PKC; insertion into the membrane may be a primary mechanism of PKC activation. PKC may be a long-term cell regulator which becomes inserted into the membrane upon appearance of the second messengers, calcium and diacylglycerol, and remains in an active membrane-bound state when the second messengers have been removed.     

Characteristics of sarcoplasmic reticulum from normal and denervated rat skeletal muscle.

Denervation of rat skeletal muscle produces after 14 days a decrease in Ca2+ uptake of a heterogeneous population of sarcoplasmic-reticulum vesicles, when measured in the presence of oxalate. The Mg2+-dependent ATPase (Ca2+-independent) activity increased after the same period and the Ca2+ + Mg2+-dependent ATPase activity decreased. Concomitant with these changes, there was an increase in vesicle size and calcium content. The observations are discussed in terms of changes in altered membrane structure, manifested in the shift of the equilibrium of the ATPase from an enzyme involved in calcium transport to a phosphoenzyme giving rise to an increase in the Mg2+-dependent ATPase activity.     

Interaction of human protein Z with thrombin: evaluation of the species difference in the interaction between bovine and human protein Z and thrombin

Protein Z is a vitamin K-dependent protein of unknown function present in normal human and bovine plasma. Binding and kinetic studies showed that bovine protein Z interacts with bovine thrombin with a dissociation constant of 0.11 microM in a Ca(2+)-independent fashion and that thrombin becomes associated with phospholipid vesicles in the presence of protein Z but not in its absence (Hogg, P. J. and Stenflo, J. (1991) J. Biol. Chem., in press). In the present study the interaction of human protein Z with human thrombin and the influence of human protein Z on the association of thrombin with phospholipid vesicles was evaluated. In contrast to bovine protein Z, human protein Z bound human DIP-thrombin with a 20-fold weaker affinity at 1.5 mM Ca2+ and in a Ca(2+)-dependent fashion. Human protein Z was also less effective than bovine protein Z in promoting the association of thrombin with phospholipid vesicles. Also, bovine protein Z cleaved by thrombin at Arg-365 bound DIP-thrombin with a 10-fold weaker affinity than did native bovine protein Z. The data suggest that the species difference in the interaction between protein Z and thrombin can be explained by a difference in the COOH-terminal region of bovine protein Z versus human protein Z.     

Interaction of phospholipids with Ca2+ ions. On the role of the phospholipid head groups

Neutral phospholipids play an important role in Ca2+ binding to biomembranes, in particular if the membrane carries a net negative surface charge due to charged lipids or proteins. The concentration of Ca2+ ions in the plane of the phospholipid head groups can be enhanced by at least two orders of magnitude compared to bulk solution. Ca2+ binding furthermore changes the orientation of the phospholipid head groups which is accompanied by variations of the local membrane dipole potential of the order of 10(5) V/cm. Such high electric fields could entail conformational changes of membrane-bound proteins and the Ca2(+)-induced reorientation of the lipid dipoles could thus play a regulatory role in membrane function.     

Ca2+- and Mg2+-dependent association of phosphorylase kinase with human erythrocyte membranes

The interaction of rabbit muscle phosphorylase kinase (EC 2.7.1.38) with human erythrocyte membranes was investigated. It was found that at pH 7.0 the kinase binds to the inner face of the erythrocyte membrane (inside-out vesicles) and that this binding is Ca2+- and Mg2+-dependent. The sharpest increase in the binding reaction occurs at concentrations between 70 and 550 nM free Ca2+. Erythrocyte ghost or right-side out erythrocyte vesicles showed a significantly lower capacity to interact with phosphorylase kinase. Autophosphorylated phosphorylase kinase shows a similar Ca2+-dependent binding profile, while trypsin activation of the kinase and calmodulin decrease the original binding capacity by about 50%. Heparin (200 micrograms/ml) and high ionic strength (50 mM NaCl) almost completely blocks enzyme-membrane interaction; glycogen does not affect the interaction.     

The Ca2+ -activated protease (calpain) modulates Ca2+/calmodulin dependent activity of smooth muscle myosin light chain kinase

The Ca2+ -activated neutral protease can proteolyze both Ca2+ -dependent cyclic nucleotide phosphodiesterase and smooth muscle myosin light chain kinase. Ca2+ -dependent cyclic nucleotide phosphodiesterase from rat brain was converted to the Ca2+ -independent active form by Ca2+ -activated protease. The proteolytic effects on myosin light chain kinase of Ca2+-activated protease differed in the presence and absence of the Ca2+-calmodulin (CaM) complex. In the presence of bound CaM, myosin light chain kinase (130k dalton) was degradated to a major fragment of 62 kDa, which had Ca2+/CaM-dependent enzyme and CaM-binding activity. When digestion occurred in the absence of bound CaM, myosin light chain kinase cleaved to a fragment of 60 kDa. This peptide had no enzymatic activity in the presence or absence of the Ca2+-CaM complex. Available evidence suggests that the Ca2+-activated proteases may recognize the conformational change of smooth muscle myosin light chain kinase induced by Ca2+-CaM complex.     

Interaction of S-100b protein with cardiolipin vesicles as monitored by electron spin resonance, pyrene fluorescence and circular dichroism

The interaction of S-100b protein with cardiolipin (CL) vesicles has been studied by electron spin resonance, pyrene fluorescence, and circular dichroism. Electron spin resonance and pyrene fluorescence data indicate that S-100b binds to the polar surface of vesicles Ca2+-independently. In the presence of Ca2+, S-100b potentiates the Ca2+-induced clustering of the polar headgroups of CL molecules and causes a further reduction in the Ca2+-dependent decrease in the lateral mobility of the pyrene inserted into the lipid bilayer, which points to an effect of the protein on the hydrophobic core of the lipid bilayer through a larger perturbation of its polar surface. Circular dichroism analyses indicate that CL vesicles cause a decrease in the alpha-helical content of S-100b, analogous to that produced by Ca2+ and that the effects of CL vesicles and of Ca2+ on the secondary structure of the protein are supra-additive. By this technique, we found that the affinity of Ca2+ for S-100b increases substantially in the presence of CL vesicles, even in the presence of physiologic concentrations of KCl, suggesting that once S-100b had interacted with CL vesicles it assumes a new conformation in which its Ca2+-binding properties are greatly enhanced. These results are discussed in relation to binding of S-100b proteins to natural membranes, and to a possible involvement of S-100b in the regulation of membrane structural organization.     

Characterization of CrATP-induced calcium occlusion in membrane-bound and soluble monomeric sarcoplasmic reticulum Ca2+-ATPase

Occlusion of Ca2+ induced by beta, gamma-bidentate CrATP in membrane bound and in soluble monomeric sarcoplasmic reticulum Ca2+-ATPase was studied by previously developed filtration and HPLC techniques (Vilsen and Andersen (1986) Biochim. Biophys. Acta 855, 429-431). Activation of Ca2+ occlusion occurred at micromolar free Ca2+ and depended on the concentration of Ca2+, H+ and Mg2+ in a similar way as activation of Ca2+ transport and equilibrium Ca2+ binding to high-affinity Ca2+ transport sites. The slopes of the Ca2+ titration curves indicated that Ca2+ binding is a cooperative process both in membraneous and in soluble monomeric enzyme. At alkaline pH and absence of Mg2+, occlusion of Ca2+ was inhibited by 1 mM Ca2+ in membrane-bound, but not in soluble monomeric Ca2+-ATPase. Parallel studies of phosphorylation from [gamma-32P]CrATP indicated a stoichiometry of 2 mol Ca2+ occluded per mol Ca2+-dependent EP formed, at saturating as well as at desaturating Ca2+ concentrations. Tryptic digestion of the CrATP induced Ca2+ occluded complex indicated that it belongs to the E1 conformational class (E1P). In the absence of Ca2+ and Mg2+, but presence of CrATP the conformational state was E2. When Mg2+ was added together with CrATP at alkaline pH the conformation was shifted in direction of E1.