Cardiolipin is the membrane receptor for mitochondrial creatine phosphokinase

Treatment of rat heart mitochondria with phosphate or mersalyl releases a number of proteins, including the mitochondrial creatine kinase (mt-CK). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the released proteins showed that phosphate is more selective than mersalyl in releasing mt-CK. The rebinding of mt-CK to mitochondria was selectively inhibited by adriamycin, which complexes membrane-bound cardiolipin. mt-CK activity and binding experiments have shown that intact mitochondria are able to bind approximately twice the amount of mt-CK they originally contain. Liver mitochondria bound heart mitochondria mt-CK to the same extent as creatine kinase-depleted heart mitochondria. mt-CK was bound by liposomes but only if they contained cardiolipin. The binding of mt-CK to cardiolipin-containing liposomes was inhibited by adriamycin. Phosphatidylcholine liposomes reconstituted with the purified ADP/ATP translocator failed to bind mt-CK.     

Electron spin resonance studies on the lipid-protein interaction between cardiolipin and anti-cardiolipin antibodies.

Electron spin resonance measurements were performed in order to investigate the influence of anti-cardiolipin antibodies on cardiolipin-containing liposomes. The physical state of the lipid structures and the alterations caused by the interaction with specific antibody were determined by measuring the freedom of motion of spin-labeled stearic acid derivatives incorporated into the lipid structures. The interaction of the cardiolipin-containing liposomes with the anti-cardiolipin antibodies reduced the mobility of the spin-labeled stearic acid probe I (12, 3), whose nitroxide group is assumed to be located near the polar region of the lipid bilayer. The restricted mobility, which qualitatively resembles the interaction of cardiolipin liposomes with calcium ions, is probably the result of a tighter packing of the polar groups in their crystalline array. The binding sites of the cardiolipin structures for anti-cardiolipin antibodies and Ca2 ions seem to be identical. As indicated by the spin-labeled stearic acid probe I (1, 14), the apolar region of the lipid bilayer is not affected by the interaction of the cardiolipin-containing liposomes with the anti-cardiolipin antibodies.     

Isolation and identification by sequence homology of a putative cytosine methyltransferase from Arabidopsis thaliana.

A plant cytosine methyltransferase cDNA was isolated using degenerate oligonucleotides, based on homology between prokaryote and mouse methyltransferases, and PCR to amplify a short fragment of a methyltransferase gene. A fragment of the predicted size was amplified from genomic DNA from Arabidopsis thaliana. Overlapping cDNA clones, some with homology to the PCR amplified fragment, were identified and sequenced. The assembled nucleic acid sequence is 4720 bp and encodes a protein of 1534 amino acids which has significant homology to prokaryote and mammalian cytosine methyltransferases. Like mammalian methylases, this enzyme has a C terminal methyltransferase domain linked to a second larger domain. The Arabidopsis methylase has eight of the ten conserved sequence motifs found in prokaryote cytosine-5 methyltransferases and shows 50% homology to the murine enzyme in the methyltransferase domain. The amino terminal domain is only 24% homologous to the murine enzyme and lacks the zinc binding region that has been found in methyltransferases from both mouse and man. In contrast to mouse where a single methyltransferase gene has been identified, a small multigene family with homology to the region amplified in PCR has been identified in Arabidopsis thaliana.     

Adriamycin inhibits the formation of non-bilayer lipid structures in cardiolipin-containing model membranes

The effect of adriamycin on cardiolipin-containing model membrane systems have been studied by ³¹P-NMR, freeze-fracture electron microscopy and binding experiments. Adriamycin effectively inhibits the formation of non-bilayer lipid structures induced by Ca²⁺ and cytochrome c in cardiolipin-containing liposomes. This drug also strongly inhibits the uptake of Ca²⁺ by cardiolipin into an organic phase. These results are discussed in relation to the cardiotoxic effect of adriamycin and the possible importance of non-bilayer lipid structures for the functioning of the mitochondrion.     

Relating the multi-functionality of cytochrome c to membrane binding and structural conversion

Cytochrome c is known as an electron-carrying protein in the respiratory chain of mitochondria. Over the last 20 years, however, alternative functions of this very versatile protein have become the focus of research interests. Upon binding to anionic lipids such as cardiolipin, the protein acquires peroxidase activity. Multiple lines of evidence suggest that this requires a conformational change of the protein which involves partial unfolding of its tertiary structure. This review summarizes the current state of knowledge of how cytochrome c interacts with cardiolipin-containing surfaces and how this affects its structure and function. In this context, we delineate partially conflicting results regarding the affinity of cytochrome c binding to cardiolipin-containing liposomes of different size and its influence on the structure of the protein and the morphology of the membrane.     

Molecular cloning and sequence analysis of human placental alkaline phosphatase

The complete amino acid sequence of the precursor and mature forms of human placental alkaline phosphatase have been inferred from analysis of a cDNA. A near full-length PLAP cDNA (2.8 kilobases) was identified upon screening a bacteriophage lambda gt11 placental cDNA library with antibodies against CNBr fragments of the enzyme. The precursor protein (535 amino acids) displays, after the start codon for translation, a hydrophobic signal peptide of 21 amino acids before the amino-terminal sequence of mature placental alkaline phosphatase. The mature protein is 513 amino acids long. The active site serine has been identified at position 92, as well as two putative glycosylation sites at Asn122 and Asn249 and a highly hydrophobic membrane anchoring domain at the carboxyl terminus of the protein. Significant homology exists between placental alkaline phosphatase and Escherichia coli alkaline phosphatase. Placental alkaline phosphatase is the first eukaryotic alkaline phosphatase to be cloned and sequenced.     

Covalent structural analysis of yeast inorganic pyrophosphatase: Location of groups implicated in the catalytic function; placement of the NH2-terminal 100 residues in the subunit

Covalent structural analysis of two of the three cyanogen bromide fragments from yeast inorganic pyrophosphatase (EC 3.6.1.1, pyrophosphate phosphohydrolase) was undertaken by a strategy involving both automated Edman degradation and conventional sequence analysis. Automated degradation of intact, reduced and carboxymethylated pyrophosphatase provided the sequence of the first 34 residues in the NH₂-terminal 45-residue peptide, CNBr VI, in addition to a partial sequence through 50 cycles which confirmed the overlap into the internal fragment, CNBr III. The sequence of CNBr VI was completed through analysis of peptides derived from hydrolysis of the fragment with trypsin and chymotrypsin. Structural analysis of CNBr III has provided the sequence of the first 55 amino acids in this 103-residue fragment. The sequence was established by conventional and automated procedures applied to the analysis of tryptic peptides generated from the citraconylated fragment. These findings constitute the sequence of the first 100 residues in the pyrophosphatase subunit and, together with structural information obtained earlier, define over half of the covalent structure of the molecule. Moreover, the sequence derived thus far permits the placement of a number of amino acids that are of importance relative to studies of the enzyme mechanism, and with regard to analysis of its three-dimensional structure.     

Carboxyacyl derivatives of cardiolipin as four-tailed hydrophobic anchors for the covalent coupling of hydrophilic proteins to liposomes.

Two carboxyacyl derivatives of cardiolipin, O-succinyl- and O-glutarylcardiolipin, were synthesized with the aim of using them as artificial membrane anchors for the immobilization of hydrophilic proteins to liposomes. Four adjacent fatty acid residues can be introduced into a protein with only one single amino group being blocked, by reacting the cardiolipin derivatives with the protein amino groups after carbodiimide activation. alpha-Chymotrypsin, used as a model protein, and modified with on average two molecules of O-succinylcardiolipin was incorporated into liposomes, which had been prepared by different methods, with very high yield. If incorporated in preformed liposomes, the carboxyacyl cardiolipin anchors were also efficient in binding proteins to liposomal surfaces. Up to 350 micrograms chymotrypsin/mumol lipid were coupled to small unilamellar vesicles, preserving reactivity of the enzyme towards specific macromolecular inhibitors. Human IgG could also be bound to anchor-containing liposomes with high protein to lipid coupling ratio as well as high coupling yield.     

Regulatory effect of pig heart phospholipids on heart muscle AMP-deaminase

AMP-deaminase purified from pig heart has been found to be activated by liposomes prepared from phospholipids extracted from pig heart mitochondria, microsomes and cytoplasm, as well as by intact microsomes. The activation by phospholipids occurred only in the presence of ATP and after the enzyme had been preincubated with liposomes for 30 min. Liposomes prepared from cardiolipin displayed an inhibitory effect on pig heart AMP-deaminase.     

The primary structure of Lactobacillus casei thymidylate synthetase. I. The isolation of cyanogen bromide peptides 1 through 5 and the complete amino acid sequence of CNBr 1, 2, 3, and 5.

Thymidylate synthetase from Lactobacillus casei was S-carboxymethylated and degraded by treatment with cyanogen bromide. Although the protein contains 6 methionine residues, only 5 cyanogen bromide peptides were obtained due to the presence of 1 methionine on the NH2 terminus and another adjacent to a threonine residue which was resistant to cleavage. The peptides were isolated by differential extraction, first with ammonium acetate, then pyridine acetate, and finally the residue was solubilized with 50% acetic acid. Each peptide was further purified to homogeneity by Bio-Gel chromatography. The size of the peptides from the amino to carboxyl end of the enzyme subunit was CNBr 1, 4,100; CNBr 2, 10,300; CNBr 3, 8,100; CNBr 4, 11,800; CNBr 5, 2,200. The sum of the amino acid residues of the peptides is equal to the sum of the residues in an enzyme subunit, indicating that all of the CNBr peptides have been isolated. The CNBr-resistant methionine was located in CNBr 2 and the 5-fluoro-2'-deoxyuridine 5'-monophosphate binding site in CNBr 4. The holoenzyme molecular weight, based on the residue weights of the amino acids in the two equivalent subunits, is equal to 73,176. The complete sequence of each of the CNBr peptides, except for CNBr 4, which is presented in the following paper, is described.     

Binding of cardiolipin/lecithin and monoamine oxidase to lipid-depleted mitochondrial membrane structure.

Lipid-depleted pig liver mitochondrial residues were incubated with different proportions of the acidic phospholipid cardiolipin and the zwitterionic phospholipid lecithin in either separate or mixed liposomes. When cardiolipin and lecithin were present in separate liposomes all of the cardiolipin but no lecithin bound to the residues. When present in the same liposomes, cardiolipin also caused binding of lecithin to the mitochondrial residues. When monoamine oxidase solubilized from pig liver mitochondria by extraction of the phospholipids was included in the incubation, binding of the enzyme to the residues occurred in the presence of cardiolipin. The percentage of enzyme bound followed the same trend as the binding of phospholipids to the mitochondrial residues.     

Recombinant hepatitis B surface antigen and anionic phospholipids share a binding region in the fifth domain of beta2-glycoprotein I (apolipoprotein H)

Human beta2-glycoprotein I (beta 2GPI) binds to recombinant hepatitis B surface antigen (rHBsAg), but the location of the binding domain on beta 2GPI is unknown. It has been suggested that the lipid rather than the protein moiety of rHBsAg binds to beta 2GPI. Since beta 2GPI binds to anionic phospholipids (PL) through its lipid-binding region in the fifth domain of beta 2GPI, we predicted that this lipid-binding region may also be involved in binding rHBsAg. In this study, we examined rHBsAg binding to two naturally occurring mutants of beta 2GPI, Cys306Gly and Trp316Ser, or evolutionarily conserved hydrophobic amino acid sequence, Leu313-Ala314-Phe315 in the fifth domain of beta 2GPI. The two naturally occurring mutations and two mutagenized amino acids, Leu313Gly or Phe315Ser, disrupted the binding of recombinant beta 2GPI (rbeta 2GPI) to both rHBsAg and cardiolipin (CL), an anionic PL. These results suggest that rHBsAg and CL share the same region in the fifth domain of beta2GPI. Credence to this conclusion was further provided by competitive ELISA, where CL-bound rbeta 2GPI was incubated with increasing amounts of rHBsAg. As expected, pre-incubation of rbeta 2GPI with CL precluded binding to rHBsAg, indicating that CL and rHBsAg bind to the same region on beta 2GPI. Our data provide evidence that the lipid (PL) rather than the protein moiety of rHBsAg binds to beta 2GPI and that this binding region is located in the fifth domain of beta 2GPI, which also binds to anionic PL.     

Localization of the amino acid substitution site in a fast migrating variant of human serum albumin

Albumin Mi/Fg is an Italian genetic variant of human serum albumin arising from a Lys----Glu substitution which has been located in a CNBr fragment (CNBr VII) corresponding to the -COOH terminal portion of the molecule [(1984) J. Chromatogr. 298, 336-344]. Tryptic peptides of CNBr VII from normal and Mi/Fg albumin have been purified by reverse-phase high-performance liquid chromatography (RP-HPLC) and submitted to comparative structural studies. The amino acid sequence of the tryptic peptide of Mi/Fg variant that differs from the corresponding fragment of the normal serum albumin shows that the Lys----Glu substitution responsible for this variant is located at position 573. This region of the albumin molecule is involved in the binding of long chain fatty acids.     

The receptor-binding domain of human apolipoprotein E. Binding of apolipoprotein E fragments

To identify the domain of apolipoprotein E (apo-E) involved in binding to low density lipoprotein (LDL) receptors on cultured human fibroblasts, apo-E was cleaved and the fragments were tested for receptor binding activity. Two large thrombolytic peptides (residues 1-191 and 216-299) of normal apo-E3 were combined with the phospholipid dimyristoylphosphatidylcholine (DMPC) and tested for their ability to compete with 125I-LDL for binding to the LDL (apo-B,E) receptors on human fibroblasts. The NH2-terminal two-thirds (residues 1-191) of apo-E3 was as active as intact apo-E3 . DMPC, while the smaller peptide (residues 216-299) was devoid of receptor-binding activity. When apo-E3 was digested with cyanogen bromide (CNBr) and the four largest CNBr fragments were combined with DMPC and tested, only one fragment competed with 125I-LDL for binding to cultured human fibroblasts (CNBr II, residues 126-218). This fragment possessed binding activity similar to that of human LDL. The 125I-labeled CNBr II . DMPC complex also demonstrated high affinity, calcium-dependent saturable binding to solubilized bovine adrenal membranes. The binding of CNBr II . DMPC was inhibited by 1,2-cyclohexanedione modification of arginyl residues or diketene modification of lysyl residues. In addition, the CNBr II had to be combined with DMPC before it demonstrated any receptor-binding activity. Pronase treatment of the membranes abolished the ability of this fragment to bind to the apo-B,E receptors. This same basic region in the center of the molecule has been implicated as the apo-B,E receptor-binding domain not only by this study but also by other studies showing that 1) natural mutants of apo-E that display defective binding have single amino acid substitutions at residues 145, 146, or 158; and 2) the apo-E epitope of the monoclonal antibody 1D7, which inhibits apo-E binding, is centered around residues 139-146.     

Deletion mutagenesis using an 'M13 splint': the N-terminal structural domain of tyrosyl-tRNA synthetase (B. stearothermophilus) catalyses the formation of tyrosyl adenylate.

The X-ray crystallographic structure of tyrosyl-tRNA synthetase (TyrTS) comprises only the N-terminal 320 amino acids of the molecule as the C-terminal 99 amino acids are poorly ordered in the crystal. A new technique, employing a single-stranded M13 splint, has been used to direct a deletion in the cloned gene of TyrTS so as to remove the disordered C-terminal region. We find that the truncated enzyme catalyses the formation of tyrosyl adenylate with unchanged Kcat and Km values and the crystallographic model must therefore include all the binding and catalytic residues involved in tyrosine activation. However, the truncated enzyme no longer binds tRNATyr or transfers tyrosine to tRNATyr. This indicates that the structural division of TyrTS is equally a functional one: the N-terminal structural domain catalyses tyrosine activation while the disordered C-terminal domain carries major determinants in tRNA binding.     

Substitution of basic amino acids in the basic region stabilizes DNA binding by E12 homodimers.

The E2A gene encodes two alternatively spliced products, E12 and E47. The two proteins differ in their basic helix-loop-helix motifs (bHLH), responsible for DNA binding and dimerization. Although both E12 and E47 can bind to DNA as heterodimers with tissue-specific bHLH proteins, E12 binds to DNA poorly as homodimers. An inhibitory domain in E12 has previously been found to prevent E12 homodimers from binding to DNA. By measuring the dissociation rates using filter binding and electrophoretic mobility shift assays, we have shown here that the inhibitory domain interferes with DNA binding by destabilizing the DNA-protein complexes. Furthermore, we have demonstrated that substitution of basic amino acids (not other amino acids) in the DNA-binding domain of E12 can increase the intrinsic DNA-binding activity of E12 and stabilize the binding complexes, thus alleviating the repression from the inhibitory domain. This ability of basic amino acids to stabilize DNA-binding complexes may be of biological significance in the case of myogenic bHLH proteins, which all possess two more basic amino acids in their DNA binding domain than E12. To function as heterodimers with E12, the myogenic bHLH proteins may need stronger DNA binding domains.     

Interaction of adriamycin with cardiolipin-containing vesicles. Evidence of an embedded site for the dihydroanthraquinone moiety

In the presence of cardiolipin-containing small unilamellar vesicles, the antitumor compound adriamycin loses its ability to catalyse the flow of electrons from NADH to molecular oxygen through NADH dehydrogenase. The data strongly suggest that in the presence of cardiolipin the dihydroanthraquinone moiety is embedded in the phospholipid bilayer and thus inaccessible to the enzyme.     

The interaction of adriamycin with small unilamellar vesicle liposomes. A fluorescence study

The interaction of the antineoplastic agent adriamycin with sonicated liposomes composed of phosphatidylcholine alone and with small amounts (1-6%) of cardiolipin has been studied by fluorescence techniques. Equilibrium binding data show that the presence of cardiolipin increases the amount of drug bound to liposomes when the bilayer is below its phase transition temperature and when the ionic strength is relatively low (0.01 M). At higher ionic strength (0.15 M) and above the Tm (i.e. conditions which are closer to the physiological state) the binding of the drug to the two liposome types is nearly the same. Thus the differences in the interactions of adriamycin with cardiolipin-containing membranes, as opposed to those composed of phosphatidylcholine alone, are not due simply to increased binding but rather to an altered membrane structure when this lipid is present. Quenching of adriamycin fluorescence by iodide shows that bound drug is partially, but not completely, buried in the liposomal membrane. Both in the presence and absence of cardiolipin the bulk of the adriamycin is more accessible to the quencher below the Tm than above it; that is, a solid membrane tends to exclude the drug from deep penetration. Above the Tm, the presence of cardiolipin alters the nature of liposome-adriamycin interaction. Here the fluorescence quenching data suggest that the presence of small amounts of cardiolipin (3%) in a phosphatidylcholine matrix creates two types of binding environments for drug, one relatively exposed and the other more deeply buried in the membrane. The temperature dependence of the adriamycin fluorescence and the liposome light scattering reveal that cardiolipin alters the thermal properties of the bilayer as well as its interaction with adriamycin. At low ionic strength lateral phase separations may occur with both pure phosphatidylcholine and when 3% cardiolipin is present; under these conditions the bound adriamycin exists in two kinds of environment. It is notable that only adriamycin fluorescence reveals this phenomenon; thebulk property of liposome light scattering reports only on the overall membrane phase change. These data suggest that under certain conditions the drug binding sites in the membranes are decoupled from the bulk of the lipid bilayer.     

A hydrophobic sequence at position 313-316 (Leu-Ala-Phe-Trp) in the fifth domain of apolipoprotein H (beta2-glycoprotein I) is crucial for cardiolipin binding.

Apolipoprotein H (apoH, protein; APOH, gene) binds to negatively charged phospholipids, which triggers the production of a subset of autoantibodies against phospholipid in patients with autoimmune diseases. We have demonstrated that two naturally occurring missense mutations in the fifth domain of apoH, Trp316Ser and Cys306Gly, disrupt the binding of native apoH to phosphatidylserine [Sanghera, D. K., Wagenknecht, D. R., McIntyre, J. A. & Kamboh, M. I. (1997) Hum. Mol. Genet. 6, 311-316]. To confirm whether these are functional mutations, we mutagenized APOH cDNAs and transiently expressed them in COS-1 cells. The cardiolipin ELISA of wild-type and mutant recombinant apoH confirmed that the Gly306 and Ser316 mutations are responsible for abolishing the binding of recombinant apoH to cardiolipin. These mutations, however, had no effect on the levels of expression or secretion of recombinant apoH in transfected COS-1 cells. While the Cys306Gly mutation disrupts a disulfide bond between Cys306 and Cys281, which appears to be critical for clustering positively charged amino acids, the Trp316Ser mutation affects the integrity of an evolutionarily conserved hydrophobic sequence at position 313-316 (Leu-Ala-Phe-Trp), which is hypothesized to interact with anionic phospholipid. To test this hypothesis, we exchanged the remaining three hydrophobic amino acids with neutral amino acids by site-directed mutagenesis (Leu313Gly, Ala314Ser and Phe315Ser). Binding of the Leu313Gly and Phe315Ser mutants to cardiolipin was significantly reduced to 25% and 13%, respectively, of that of the wild-type. On the other hand, the Ala314Ser mutation showed normal cardiolipin binding. Taken together with our previous findings, these results strongly suggest that the configuration of the fifth domain of apoH, as well as the integrity of the highly conserved hydrophobic amino acids at positions 313-316, is essential for the binding of apoH to anionic phospholipid.