Ankylosing spondylitis, late osteoarthritis, vascular calcification, chondrocalcinosis and pseudo gout: toward a possible drug therapy.

In this review we consider diseases associated with pathological mineralization/ossification, namely, ankylosing spondylitis (AS), osteoarthritis (OA), generalized artery calcification of infancy (GACI), vascular calcification as well as chondrocalcinosis (CC) and pseudo gout. Deciphering the key enzymes implicated in the calcification process is an objective of prime importance and the ultimate goal is to synthesize inhibitors of these enzymes in order to provide efficient alternate therapeutic strategies that will slow down the pathologic mineralization and complement the arsenal of anti-inflammatory drugs. One of the difficulties in the definition of diseases associated with pathologic mineralization/ossification lies in the controversial relationship between the type of calcification and the nature of the disease. Here, we propose to clarify this relationship by making a distinction between diseases associated with hydroxyapatite (HA) and calcium pyrophosphate dihydrate (CPPD) deposits. AS, OA, GACI and vascular calcification are usually characterized by mineralization/ossification associated with HA deposits, while CC and pseudo gout are mostly characterized by CPPD deposits. Although both HA and CPPD deposits may occur concomitantly, as in chronic pyrophosphate arthritis or in OA with CPPD, they are formed as a result of two antagonistic processes indicating that treatment of distinct diseases can be only achieved by disease-specific drug therapies. The hydrolysis of PPi, an inhibitor of HA formation, is mostly controlled by tissue non-specific alkaline phosphatase TNAP, while PPi production in the extracellular medium is controlled by ANK, a PPi transporter, and/or NPP1 which generates PPi from nucleotide triphosphates. Low PPi concentration may lead to a preferential deposition of HA while high PPi concentration will favor the formation of CPPD deposits. Thus, HA and CCPD deposition cannot occur concomitantly because they are determined by the Pi/PPi ratio which, in turn, depends on the relative activities of antagonistic enzymes, TNAP hydrolyzing PPi or ANK and NPP1 producing PPi. TNAP inhibitors could prevent HA formation in AS, in late OA, in GACI, as well as in vascular calcifications, while ANK or NPP1 inhibitors could slow down CCPD deposition in CC and pseudo gout.

Inorganic pyrophosphate as a regulator of hydroxyapatite or calcium pyrophosphate dihydrate mineral deposition by matrix vesicles.

OBJECTIVE: Pathological mineralization is induced by unbalance between pro- and anti-mineralization factors. In calcifying osteoarthritic joints, articular chondrocytes undergo terminal differentiation similar to that in growth plate cartilage and release matrix vesicles (MVs) responsible for hydroxyapatite (HA) or calcium pyrophosphate dihydrate (CPPD) deposition. Inorganic pyrophosphate (PP(i)) is a likely source of inorganic phosphate (P(i)) to sustain HA formation when hydrolyzed but also a potent inhibitor preventing apatite mineral deposition and growth. Moreover, an excess of PP(i) can lead to CPPD formation, a marker of pathological calcification in osteoarthritic joints. It was suggested that the P(i)/PP(i) ratio during biomineralization is a turning point between physiological and pathological mineralization. The aim of this work was to determine the conditions favoring either HA or CPPD formation initiated by MVs. METHODS: MVs were isolated from 17-day-old chicken embryo growth plate cartilages and subjected to mineralization in the presence of various P(i)/PP(i) ratios. The mineralization kinetics and the chemical composition of minerals were determined, respectively, by light scattering and infrared spectroscopy. RESULTS: The formation of HA is optimal when the P(i)/PP(i) molar ratio is above 140, but is completely inhibited when the ratio decreases below 70. The retardation of any mineral formation is maximal at P(i)/PP(i) ratio around 30. CPPD is exclusively produced by MVs when the ratio is below 6, but it is inhibited for the ratio exceeding 25. CONCLUSIONS: Our findings are consistent with the P(i)/PP(i) ratio being a determinant factor leading to pathological mineralization or its inhibition.

A comparative analysis of strategies for isolation of matrix vesicles.

Matrix vesicles (MVs) are extracellular organelles involved in the initial steps of mineralization. MVs are isolated by two methods. The first isolation method of MVs starts with collagenase digestion of osseous tissues, followed by two differential centrifugations. The second isolation method does not use proteases but rather starts with differential centrifugation, followed by a fractionation on a sucrose gradient. The first method results in a homogeneous population of MVs with higher cholesterol/lipid content, alkaline phosphatase activity, and mineral formation rate as compared with MVs isolated by the second method. The second method leads to higher protein diversity as compared with MVs isolated according to the first method. Due to their distinct protein composition, lipid-to-protein and cholesterol-to-phospholipid ratios, and differences in rates of mineral formation, both types of isolated MVs are crucial for proteomic analysis and for understanding the regulation of mineralization process at the molecular level.

Purification and functional reconstitution of intact ral-binding Gtpase activating protein, RLIP76, in artificial liposomes.

We have recently shown that RLIP76, a ral-binding GTPase activating protein, mediates ATP-dependent transport of glutathione-conjugates (GS-E) and doxorubicin (DOX) (S. Awasthi et al., Biochemistry 39,9327,2000). Transport function of RLIP76 was found to be intact despite considerable proteolytic fragmentation in preparations used for those studies, suggesting either that the residual intact RLIP76 was responsible for transport activity, or that the transport activity could be reconstituted by fragments of RLIP76. If the former were true, intact RLIP76 would have a much higher specific activity for ATP-hydrolysis than the fragmented protein. We have addressed this question by comparing transport properties of recombinant RLIP76 and human erythrocyte membrane RLIP76 purified in buffers treated with either 100 or 500 microM serine protease inhibitor, PMSF. The purity and identity of recombinant and human erythrocyte RLIP76 was established by SDS/PAGE and Western-blot analysis. These studies confirmed the origin of the 38 kDa protein, previously referred to as DNP-SG ATPase, from RLIP76. Higher PMSF concentration resulted in lower yield of the 38 kDa band and higher yield of intact RLIP76 from both human and recombinant source. In contrast, the substrate-stimulated ATPase activity in presence of DNP-SG, doxorubicin, daunorubicin, or colchicine were unaffected by increased PMSF; similarly, ATP-dependent transport of doxorubicin in proteoliposomes reconstituted with RLIP76 was unaffected by higher PMSF. These results indicated that limited proteolysis by serine proteases does not abrogate the transport function of RLIP76. Comparison of transport kinetics for daunorubicin between recombinant vs human erythrocyte RLIP76 revealed higher specific activity of transport for tissue purified RLIP76, indicating that additional factors present in tissue purified RLIP76 can modulate its transport activity.

N- and C-terminal halves of human annexin VI differ in ability to form low pH-induced ion channels.

Human recombinant annexin VI (AnxVI) or its N- (AnxVIA) and C-terminal (AnxVIB) fragments were expressed in E. coli. Their ability to form voltage-dependent ion channels in membranes, induced by low pH, was measured to verify the hypothesis that, upon acidification, the hydrophobicity of AnxVI at a specific domain significantly increases allowing the AnxVI interaction with lipids in a Ca(2+)-independent manner. By theoretically analyzing changes in protein hydrophobicity, we found that hydrophobicity of AnxVIA significantly differed from that of AnxVIB at low pH. These predictions were confirmed experimentally by using planar lipid bilayers and liposome pull-down assay. We found striking difference between AnxVIA and AnxVIB in the ion channel activity, as well as in the membrane binding, suggesting that the halves of AnxVI maybe functionally different. Moreover, we calculated and predicted that the ion channel activity at low pH should appear in other human annexins, as AnxII, AnxV (as known), AnxVIII, and AnxXIII. The possibility that AnxVI acts as cytosolic component of a transmembrane pH-sensing mechanism is proposed.

Conformational states of annexin VI in solution induced by acidic pH.

Acidic pH-induced folding of annexin (Anx)VI in solution was investigated in order to study the mechanism of formation of ion channels by the protein in membranes. Using 2-(p-toluidino)naphthalene-6-sulfonic acid as a hydrophobic probe, it was demonstrated that AnxVI exerts a large change in hydrophobicity at acidic pH. Moreover, circular dichroism spectra indicated that the native state of AnxVI changes at acidic pH towards a state characterized by a significant loss of alpha-helix content and appearance of new beta-structures. These changes are reversible upon an increase of pH. It is postulated that the structural folding of AnxVI could explain how a soluble protein may undergo transition into a molecule able to penetrate the membrane hydrophobic region. The physiological significance of these observations is discussed.

UDP hydrolase activity associated with the porcine liver annexin fraction.

In the crude fraction of porcine liver annexins, we identified annexin IV (AnxIV), AnxII and AnxVI of MW (molecular weight) of 32, 36 and 68 kDa, respectively, an albumin of MW of 61.5 kDa and an UDP hydrolase (UDPase) of MW of 62 kDa, related to the human UDPase from Golgi membranes. The latter enzyme exhibits its highest specificity towards UDP and GDP but not ADP and CDP, and it is stimulated by Mg(2+) and Ca(2+). AnxVI itself, although it binds purine nucleotides, does not exhibit hydrolytic activity towards nucleotides. Taken together, these results suggest that AnxVI may interact in vivo with a nucleotide-utilizing enzyme, UDPase. This is in line with observations made by other investigators that various annexins are able to interact with nucleotide-utilizing proteins, such as protein kinases, GTPases, cytoskeletal proteins and p120(GAP). Such interactions could be of particular importance in modulating the biological activities of these proteins in vivo.

Annexins as nucleotide-binding proteins: facts and speculations.

Annexins are ubiquitous multifunctional Ca2+ and phospholipid-binding proteins whose mechanism of function remains largely unknown. The accumulated in vitro experimental evidence indicates that ATP and GTP are functional ligands for nucleotide-sensitive annexin isoforms. Such nucleotide binding could modulate Ca2+ homeostasis, vesicular transport and/or signal transduction pathways and link them to cellular energy metabolism. Alternatively, since annexins are able to interact with other nucleotide-utilizing proteins, such as various kinases, GTPases and structural proteins, these proteins could influence the guanine nucleotide exchange metabolism and/or control the activity of various G proteins. The nucleotide-binding properties of annexins may affect the development or maintenance of some pathologies and diseases in which changes in physiological concentrations of purine nucleotides or disruption of Ca2+ homeostasis are crucial targets.

GTP-binding properties of the membrane-bound form of porcine liver annexin VI.

Annexin VI (AnxVI) of molecular mass 68-70 kDa belongs to a multigenic family of ubiquitous Ca2+- and phospholipid-binding proteins. In this report, we describe the GTP-binding properties of porcine liver AnxVI, determined with a fluorescent GTP analogue, 2'-(or 3')-O-(2,4,6-trinitrophenyl)guanosine 5'-triphosphate (TNP-GTP). The optimal binding of TNP-GTP to AnxVI was observed in the presence of Ca2+ and asolectin liposomes, as evidenced by a 5.5-fold increase of TNP-GTP fluorescence and a concomitant blue shift (by 17 nm) of its maximal emission wavelength. Titration of AnxVI with TNP-GTP resulted in the determination of the dissociation constant (Kd) and binding stoichiometry that amounted to 1.3 microM and 1:1 TNP-GTP/AnxVI, mole/mole, respectively. In addition, the intrinsic fluorescence of the membrane-bound form of AnxVI was quenched by TNP-GTP and this was accompanied by fluorescence resonance energy transfer (FRET) from AnxVI Trp residues to TNP-GTP. This indicates that the GTP-binding site within the AnxVI molecule is probably located in the vicinity of a Trp-containing domain of the protein. By controlled proteolysis of human recombinant AnxVI, followed by purification of the proteolytic fragments by affinity chromatography on GTP-agarose, we isolated a 35 kDa fragment corresponding to the N-terminal half of AnxVI containing Trp192. On the basis of these results, we suggest that AnxVI is a GTP-binding protein and the binding of the nucleotide may have a regulatory impact on the interaction of annexin with membranes, e.g. formation of ion channels by the protein.

Novel function of human RLIP76: ATP-dependent transport of glutathione conjugates and doxorubicin.

Active transport of conjugated and unconjugated electrophiles out of cells is essential for cellular homeostasis. We have previously identified in human tissues a transporter, DNP-SG [S-(2, 4-dinitrophenyl)glutathione] ATPase, capable of carrying out this function [Awasthi et al. (1998) Biochemistry 37, 5231-5238, 5239-5248]. We now report the cloning of DNP-SG ATPase. The sequence of the cDNA clone was identical to that of human RLIP76, a known Ral-binding protein. RLIP76 expressed in E. coli was purified by DNP-SG affinity chromatography. Purified recombinant RLIP76: (1) had ATPase activity stimulated by DNP-SG or doxorubicin (DOX), and the K(m) values of RLIP76 for ATP, DOX, and DNP-SG were similar to those reported for DNP-SG ATPase; (2) upon reconstitution with asolectin as well as with defined lipids, catalyzed ATP-dependent transport of DNP-SG and DOX with kinetic parameters similar to those of DNP-SG ATPase; (3) when transfected into K562 cells, resulted in increased resistance to DOX, and increased ATP-dependent transport of DNP-SG and DOX by inside-out membrane vesicles from transfected cells; (4) direct uptake of purified RLIP76 protein into mammalian cells from donor proteoliposomes confers DOX resistance. These results indicate that RLIP76, in addition to its role in signal transduction, can catalyze transport of glutathione conjugates and xenobiotics, and may contribute to the multidrug resistance phenomenon.

Lipid metabolism as a target for potassium channel effectors.

K(+) channel effectors are widely used in the treatment of various diseases, including diabetes mellitus type II, hypertension, and cardiac arrhythmia. In addition, a constantly growing body of literature reveals that some of these substances, despite their direct effect on K(+) channels, may influence cellular lipid metabolism. As a result, membrane lipid content and cellular concentrations of lipid messengers are changed. Due to the dependence of K(+) channel activity on membrane lipids, these observations seem to be of particular importance not only to characterize secondary effects of K(+) channel effectors but also to understand the long-term effects of these agents on K(+) channel activity.

Alzheimer's disease: its origin at the membrane, evidence and questions.

Numerous results on membrane lipid composition from different regions of autopsied Alzheimer's disease brains in comparison with corresponding fractions isolated from control brains revealed significant differences in serine- and ethanolamine-containing glycerophospholipid as well as in glycosphingolipid content. Changes in membrane lipid composition are frequently accompanied by alterations in membrane fluidity, hydrophobic mismatch, lipid signaling pathways, transient formation and disappearance of lipid microdomains, changes in membrane permeability to cations and variations of other membrane properties. In this review we focus on possible implications of altered membrane composition on beta-amyloid precursor protein (APP) and on proteolysis of APP leading eventually to the formation of neurotoxic beta-amyloid (A beta) peptides, the major proteinaceous component of extracellular senile plaques, directly involved in Alzheimer's disease pathogenesis.

Affinity labeling of annexin VI with a triazine dye, Cibacron blue 3GA. Probable interaction of the dye with C-terminal nucleotide-binding site within the annexin molecule.

Annexin VI (AnxVI) from porcine liver, a member of the annexin family of Ca(2+)- and membrane-binding proteins, has been shown to bind ATP in vitro with a K(d) in the low micromolar concentration range. However, this protein does not contain within its primary structure any ATP-binding consensus motifs found in other nucleotide-binding proteins. In addition, binding of ATP to AnxVI resulted in modulation of AnxVI function, which was accompanied by changes in AnxVI affinity to Ca2+ in the presence of ATP. Using limited proteolytic digestion, purification of protein fragments by affinity chromatography on ATP-agarose, and direct sequencing, the ATP-binding site of AnxVI was located in a C-terminal half of the AnxVI molecule. To further study AnxVI-nucleotide interaction we have employed a functional nucleotide analog, Cibacron blue 3GA (CB3GA), a triazine dye which is commonly used to purify multiple ATP-binding proteins and has been described to modulate their activities. We have observed that AnxVI binds to CB3GA immobilized on agarose in a Ca(2+)-dependent manner. Binding is reversed by EGTA and by ATP and, to a lower extent, by other adenine nucleotides. CB3GA binds to AnxVI also in solution, evoking reversible aggregation of protein molecules, which resembles self-association of AnxVI molecules either in solution or on a membrane surface. Our observations support earlier findings that AnxVI is an ATP-binding protein.

ATP-Binding site of annexin VI characterized by photochemical release of nucleotide and infrared difference spectroscopy.

Structural changes induced by nucleotide binding to porcine liver annexin VI (AnxVI) were probed by reaction-induced difference spectroscopy (RIDS). Photorelease of the nucleotide from ATP[Et(PhNO2)] produced RIDS of AnxVI characterized by reproducible changes in the amide I region. The magnitude of the infrared change was comparable to RIDS of other ATP-binding proteins, such as Ca(2+)-ATPase and creatine and arginine kinases. Analysis of RIDS revealed the existence of ATP-binding site(s) (K(d) < 1 microM) within the AnxVI molecule, comprising five to six amino acid residues located in the C-terminal portion of the protein molecule. The binding stoichiometry of ATP:AnxVI was determined as 1:1 (mol/mol). ATP, in the presence of Ca2+, induced changes in protein secondary structure reflected by a 5% decrease in alpha-helix content of the protein in favor of unordered structure. Such changes may influence the affinity of AnxVI for Ca2+ and modulate its interaction with membranes.

Annexin VI interacts with adenine nucleotides and their analogs.

Annexin VI (AnxVI), a member of the annexin family of Ca2+- and membrane-binding proteins, has been shown to interact in vitro with adenine nucleotides. Furthermore, it has been proposed that within the AnxVI molecule a nucleotidde-binding domain exists, which is located in the C-terminal half of the protein, in the vicinity of Trp343. By comparison of exposure of tryptophan and multiple tyrosine residues upon nucleotide binding, as revealed by quenching of intrinsic fluorescence of AnxVI by ATP, ADP or cAMP, it can be concluded that the binding of nucleotides evokes changes in the protein tertiary structure. Moreover, in the course of present study we have found that AnxVI binds to a non-hydrolysable analog of ATP, the triazine dye Cibacron blue 3GA (CB3GA), immobilized on agarose. Binding reveals negative cooperativity with respect to protein concentration and is Ca2+-dependent. Binding is prevented by ATP. CB3GA binds to AnxVI also in solution, evoking the formation of annexin multimers. On the basis of this observation it can be suggested that interaction of CB3GA with AnxVI is useful to examine, with some limitations, the self-association of annexin molecules implying to play a role in interacting of AnxVI with biological membranes.

10-Undecynoic acid, an inhibitor of cytochrome P450 4A1, inhibits ethanolamine-specific phospholipid base exchange reaction in rat liver microsomes.

1,12-Dodecanedioic acid, the end-product of omega-hydroxylation of lauric acid, stimulates in a concentration dependent manner, phosphatidylethanolamine synthesis via ethanolamine-specific phospholipid base exchange reaction in rat liver endoplasmic reticulum. On the other hand, administration to rats of 10-undecynoic acid, a specific inhibitor of omega-hydroxylation reaction catalyzed by cytochrome P450 4A1, inhibits the ethanolamine-specific phospholipid base exchange activity by 30%. This is accompanied by a small but significant decrease in phosphatidylethanolamine content in the endoplasmic reticulum and inhibition of cytochrome P450 4A1. On the basis of these results it can be proposed that a functional relationship between cytochrome P450 4A1 and phosphatidylethanolamine synthesis exists in rat liver. Cytochrome P450 4A1 modulates the cellular level of lauric acid, an inhibitor of phospholipid synthesis. In turn, ethanolamine-specific phospholipid base exchange reaction provides molecular species of phospholipids, containing mainly long-chain polyunsaturated fatty acid moieties, required for the optimal activity of cytochrome P450 4A1.

Annexin VI: an intracellular target for ATP.

Annexin VI (AnxVI), an Ca2+- and phospholipid-binding protein, interacts in vitro with ATP in a calcium-dependent manner. Experimental evidence indicates that its nucleotide-binding domain which is localized in the C-terminal half of the protein differs structurally from ATP/GTP-binding motifs found in other nucleotide-binding proteins. The amino-acid residues of AnxVI directly involved in ATP binding have not been yet defined. Binding of ATP to AnxVI induces changes in the secondary and tertiary structures of protein, affecting the affinity of AnxVI for Ca2+ and, in consequence, influencing the Ca2+-dependent activities of AnxVI: binding to F-actin and to membranous phospholipids, and self-association of the annexin molecules. These observations suggest that ATP is a functional ligand for AnxVI in vivo, and ATP-sensitive AnxVI may play the role of a factor coupling vesicular transport and calcium homeostasis to cellular metabolism.