The erythrocyte receptor for the channel-forming toxin aerolysin is a novel glycosylphosphatidylinositol-anchored protein

The plasma membrane of rat erythrocytes contains a 47-kDa glycoprotein that binds the channel-forming toxin aerolysin with high affinity and accounts for the sensitivity of these cells to the toxin. The receptor was purified so that its N-terminal sequence could be determined after Western blotting. The sequence did not match any sequences in the databases, indicating that the receptor is a novel erythrocyte surface protein. However, it exhibited considerable homology to the N-termini of a group of membrane proteins that are thought to be involved in ADP-ribosyl transfer reactions. A common property of these proteins is that they are attached to plasma membranes by C-terminal glycosylphosphatidylinositol (GPI) anchors. The aerolysin receptor was shown to be anchored in the same way by treating rat erythrocytes with phosphatidylinositol-specific phospholipase C. This caused the selective release of the receptor and a reduction in the rodent cells' sensitivity to aerolysin. Human and bovine erythrocytes were shown to contain an aerolysin-binding protein with similar properties to the rat erythrocyte receptor. Proteins with GPI anchors are thought to have unusually high lateral mobility, and this may be an advantage for a toxin, such as aerolysin, which must oligomerize after binding to become insertion competent.     

Phosphatidylethanolamine is the donor of the ethanolamine residue linking a glycosylphosphatidylinositol anchor to protein.

Numerous cell surface glycoproteins from eukaryotic organisms including African trypanosomes and budding yeast (Saccharomyces cerevisiae), are anchored to the lipid bilayer by a glycophospholipid, glycosylphosphatidylinositol, covalently linked to the carboxyl terminus of the protein via a phosphoethanolamine bridge. In this paper we describe metabolic labeling experiments aimed at identifying the biosynthetic origin of the ethanolamine residue in the phosphoethanolamine bridge. Using yeast mutants generated by disruption of the ethanolaminephosphotransferase (EPT1) and cholinephosphotransferase (CPT1) genes, we report data consistent with the proposal that the ethanolamine residue is derived from phosphatidylethanolamine.     

A human monoclonal antibody specific to placental alkaline phosphatase,a marker of ovarian cancer

Placental alkaline phosphatase (PLAP) is a promising ovarian cancer biomarker. Here, we describe the isolation, affinity-maturation and characterization of two fully human monoclonal antibodies (termed B10 and D9) able to bind to human PLAP with a dissociation constant (K_d) of 10 and 30 nM, respectively. The ability of B10 and D9 antibodies to recognize the native antigen was confirmed by Biacore analysis, FACS and immunofluorescence studies using ovarian cancer cell lines and freshly-frozen human tissues. A quantitative biodistribution study in nude mice revealed that the B10 antibody preferentially localizes to A431 tumors, following intravenous administration. Anti-PLAP antibodies may serve as a modular building blocks for the development of targeted therapeutic products, armed with cytotoxic drugs, radionuclides or cytokines as payloads.     

Mutation of a single amino acid converts germ cell alkaline phosphatase to placental alkaline phosphatase

Human placental and germ cell alkaline phosphatases (PLAP and GCAP, respectively), are characterized by their differential sensitivities to inhibition by L-leucine, EDTA, and heat. Yet, they differ by only 7 amino acids at positions 15, 67, 68, 84, 241, 254, and 429 within their respective 484 residues. To determine the structural basis and the amino acid(s) involved in these physicochemical differences, we constructed three GCAP mutants by site-directed mutagenesis and six GCAP/PLAP chimeras and then expressed these alkaline phosphatase mutants in COS-1 cells. We report that the differential reactivity of PLAP and GCAP depends critically on a single amino acid at position 429. GCAP with Gly-429 is strongly inhibited by L-leucine, EDTA, and heat, whereas PLAP with Glu-429 is resistant. By substituting Gly-429 of GCAP with a series of amino acids, we demonstrate that the relative sensitivities of these mutants to L-leucine, EDTA, and heat inhibition are, in general, parallel. Mutants in the order of resistance to these treatments are: Glu (most resistant), Asp/Ile/Leu, Gln/Val/Lys, Ser/His, and Arg/Thr/Met/Cys/Phe/Trp/Tyr/Pro/Asn/Ala/Gly (least resistant). However, the Ser-429 and His-429 mutants were more resistant to EDTA and heat inhibition than the wild-type GCAP, but were equally sensitive to L-leucine inhibition. Structural analysis of mammalian alkaline phosphatase modeled on the refined crystal structure of Escherichia coli alkaline phosphatase indicates that the negative charge of Glu-429 of PLAP, which simultaneously stabilizes the protein as a whole and the metal binding specifically, probably acts through interactions with the metal ligand His-320 (His-331 in E. coli alkaline phosphatase). Replacement of codon 429 with Gly in GCAP leads to destabilization and loosening of the metal binding. The data suggest that the natural binding site for L-leucine may be near position 429, with the amino and carboxyl groups of L-leucine interacting with bound phosphate and His-432 (His-412 in E. coli alkaline phosphatase), respectively.     

The glycosylphosphatidylinositol (GPI) signal sequence of human placental alkaline phosphatase is not recognized by human Gpi8p in the context of the yeast GPI anchoring machinery

Biosynthesis of glycosylphosphatidylinositol (GPI)-anchored proteins involves the action of a GPI trans-amidase, which replaces the C-terminal GPI signal sequence (GPI-SS) of the primary translation product with a preformed GPI lipid. The transamidation depends on a complex of four proteins, Gaa1p, Gpi8p, Gpi16p and Gpi17p. Although the GPI anchoring pathway is conserved throughout the eukaryotic kingdom, it has been reported recently that the GPI-SS of human placental alkaline phosphatase (hPLAP) is not recognized by the yeast transamidase, but is recognized in yeast that contain the human Gpi8p homologue. This finding suggests that Gpi8p is intimately involved in the recognition of GPI precursor proteins and may also be responsible for the subtle taxon-specific differences in transamidase specificity that sometimes prevent the efficient GPI anchoring of heterologously expressed GPI proteins. Here, we confirm that the GPI signal sequence of hPLAP is indeed not recognized by the yeast GPI-anchoring machinery. However, in our hands, GPI attachment cannot be restored by the co-expression of human Gpi8p in yeast cells under any circumstances.     

Specificity of allozyme-absorbed antisera to human placental alkaline phosphatase for individual phenotypes

Antisera raised against the rare FD phenotype enzyme were exhaustively absorbed with SS and FF phenotype enzyme immobilized on agarose gels. When it was absorbed with the FF phenotype enzyme, the antiserum no longer reacted with the F-variant enzyme, but did with the S-, D-, and I-variants, as determined by electrophoretic retardation experiments and precipitation of antigen-antibody complexes using staphylococcal protein A. When the antiserum was absorbed with SS phenotype enzyme, it no longer reacted with S-, D-, or I-variant enzyme, but did have some reactivity with the F-variant, as seen in the protein A assay. Based upon the IgG concentration, which bound 40% of the appropriate enzyme, 1/20 of the antiserum preparation was specific for the S-, D-, and I-variant shared specificity, and 1/400 was specific for the F-variant alone.     

l-Tryptophan. A non-allosteric organspecific uncompetitive inhibitor of human placental alkaline phosphatase

l-Tryptophan, but not d-tryptophan, inhibits human placental and intestinal alkaline phosphatases, but not those of liver and bone. The nature of this stereospecific organ-specific inhibition has been elucidated. Thus, from a study of the effect of substrate concentration on inhibition in which double-reciprocal plots of 1/v versus 1/s at various inhibitor concentrations were made, this inhibition is judged to be ;uncompetitive'. That the inhibition is non-allosteric is an opinion based on (1) hyperbolic curves obtained from plotting the percentage inhibition against inhibitor concentration; (2) the independence of the inhibition to heat denaturation and urea treatment; (3) the relatively low value of entropy change; and (4) a value close to unity for n, the number of l-tryptophan molecules that combine with one molecule of enzyme. Finally, a homosteric mechanism is further postulated for the inhibition by l-tryptophan based on the increase of optimum temperature for maximum velocity and the decrease of this inhibition with increasing temperature. The mechanism of this inhibition is discussed.     

Chemical characterization of the membrane-anchoring domain of human placental alkaline phosphatase

Membrane and soluble forms of alkaline phosphatase (ALP) were selectively prepared from human placental microsomes by treatment with 1-butanol at pH 8.5 and 5.5, respectively. The purified membrane (mALP) and soluble (sALP) forms were analyzed for chemical compositions. mALP was found to contain 1 mol each of palmitate, stearate, and glycerol/subunit of ALP, which were absent in sALP. Both the forms contained 1 mol of inositol and 2 mol of ethanolamine/subunit. However, none of these compounds was detectable in another soluble form prepared by treatment with papain, which is known to cleave the carboxyl-terminal region. The results suggest that mALP contains diacylglycerol, the removal of which results in its conversion to sALP. We then prepared [3H]ethanolamine-labeled ALP by incubating choriocarcinoma cells (JEG-3) with the isotope. 3H-Labeled sALP was mixed with unlabeled sALP and treated with papain. A 3H-labeled single component was purified from the digests by sequential chromatography through anti-ALP-IgG-Sepharose, concanavalin A-Sepharose, Bio-Gel P-6, and TSK G-2000 columns. Chemical analyses revealed that the purified sample contains the tripeptide Thr-Thr-Asp, ethanolamine, glucosamine, mannose, inositol, and phosphate. Molar ratios of the latter five compounds were calculated to be 2, 1, 3, 1, and 2, respectively, by taking Asp as 1 mol. The tripeptide sequence was identified at positions 482-484 in the primary structure deduced from the cDNA sequence, which predicts a further extension to position 513, containing a hydrophobic amino acid sequence. Taken together, these results suggest that the mature ALP molecule lacks the predicted carboxyl-terminal peptide extension and is attached at Asp484 with a glycosylphospholipid, the components of which are characterized above. The glycosylphospholipid thus attached is considered to function as the membrane anchor of ALP.     

Structural study on the carbohydrate moiety of human placental alkaline phosphatase

Alkaline phosphatase purified from human placenta contains a single asparagine-linked sugar chain in one molecule. The sugar chain was quantitatively liberated as radioactive oligosaccharides from the polypeptide moiety by hydrazinolysis followed by N-acetylation and NaB3H4 reduction, and separated by paper electrophoresis into one neutral and two acidic fractions. By a combination of sequential exoglycosidase digestion and methylation analysis, the structures of oligosaccharides in the neutral fraction were confirmed to be as follows: Gal beta 1----4GlcNAc beta 1----2Man alpha 1----6(Gal beta 1----4GlcNAc beta 1----2Man alpha 1----3)Man beta 1----4GlcNAc beta 1----4(+/- Fuc alpha 1----6)GlcNAc. The acidic oligosaccharide fractions were mixtures of mono- and disialyl derivatives of the neutral fraction. All the sialic acid residues of the sugar chains occur as the NeuAc alpha 2----3Gal group. In the case of monosialyl derivatives, the N-acetylneuraminic acid was exclusively linked to the Man alpha 1----3 arm.     

A highly sensitive assay method for human placental alkaline phosphatase involving a monoclonal antibody bound to a paper disk

A monoclonal antibody which is specific for human placental alkaline phosphatase and does not cross-react at all with intestinal alkaline phosphatase was prepared, and a procedure for the determination of placental alkaline phosphatase activity in serum was developed involving this monoclonal antibody bound to a paper disk. The minimum amount of placental alkaline phosphatase detectable by this method is 0.0025 King-Armstrong unit. Good correlation with the heat-treatment method was obtained. Therefore this proposed method can be used as a routine clinical test for the determination of serum placental alkaline phosphatase.     

Ethanolamine phosphate linked to the first mannose residue of glycosylphosphatidylinositol (GPI) lipids is a major feature of the GPI structure that is recognized by human GPI transamidase

Glycosylphosphatidylinositol (GPI) anchoring of proteins is catalyzed by GPI transamidase (GPIT), a multisubunit, endoplasmic reticulum (ER)-localized enzyme. GPIT recognizes ER-translocated proteins that have a GPI-directing C-terminal signal sequence and replaces this sequence with a preassembled GPI anchor. Although the GPI signal sequence has been extensively characterized, little is known about the structural features of the GPI lipid substrate that enable its recognition by GPIT. In a previous study we showed that mature GPIs could be co-immunoprecipitated with GPIT complexes containing functional subunits (Vainauskas, S., and Menon, A. K. (2004) J. Biol. Chem. 279, 6540-6545). We now use this approach, as well as a method that reconstitutes the interaction between GPIs and GPIT, to define the basis of the interaction between GPI and human GPIT. We report that (i) human GPIT can interact with GPI biosynthetic intermediates, not just mature GPIs competent for transfer to protein, (ii) the ethanolamine phosphate group on the third mannose residue of the GPI glycan is not critical for GPI recognition by GPIT, (iii) the ethanolamine phosphate residue linked to the first mannose of the GPI structure is a major feature of GPIs that is recognized by human GPIT, and (iv) the simplest GPI recognized by human GPIT is EtN-P-2Manalpha1-4GlcN-(acyl)-phosphatidyl-inositol. These studies define the molecular characteristics of GPI that are recognized by GPIT and open the way to identifying GPIT subunits that are involved in this process.     

Spectroscopic study of the activation and oligomerization of the channel-forming toxin aerolysin: identification of the site of proteolytic activation.

The channel-forming protein aerolysin is secreted as a protoxin which can be activated by proteolytic removal of a C-terminal peptide. The activation and subsequent oligomerization of aerolysin were studied using a variety of spectroscopic techniques. Mass spectrometric determination of the molecular weights of proaerolysin and aerolysin permitted identification of the sites at which the protoxin is processed by trypsin and chymotrypsin. The results of far- and near-UV circular dichroism measurements indicated that processing with trypsin does not lead to major changes in secondary or tertiary structure of the protein. An increase in tryptophan fluorescence intensity and a small red shift in the maximum emission wavelength of tryptophans could be observed, suggesting that there is a change in the environment of some of the tryptophans. There was also a dramatic increase in the binding of the hydrophobic fluorescent probe 1-anilino-8-naphthalenesulfonate during activation, leading us to conclude that a hydrophobic region in the protein is exposed by trypsin treatment. Using measurements of light scattering, various parameters influencing oligomerisation of trypsin-activated aerolysin were determined. Oligomerization rates were found to increase with the concentration of aerolysin, whereas they decreased with increasing ionic strength.     

Molecular requirements for attachment of the glycosylphosphatidylinositol anchor to the human alpha folate receptor

The alpha isoform of the folate receptor (FR) is a 38-KDa glycosylphosphatidylinositol (GPI) protein which mediates the internalization of folates. The FR amino acid sequence has features typical of GPI-linked proteins, including the presence of a hydrophobic carboxyl-terminus, a hinge region, and a stretch of small and uncharged amino acids. Substitution of predicted cleavage/attachment Ser234 with arginine or threonine, or replacement of Gly235 with proline by site-directed mutagenesis had no effect on GPI processing. In fact, CHO cells transfected with each of the three cDNA variants or with FR wild-type showed comparable amounts of phosphatidylinositol-specific phospholipase C-resistant FR in double-determinant radioimmunoassay. Western blot analysis of total cell lysates from all transfectants consistently revealed the 38-KDa FR band. Deletion of residues 233-237 in the amino-terminal portion of the FR cDNA constructs derived by a polymerase chain reaction strategy abrogated GPI processing, with only a small proportion of the FR remaining in the cytoplasm in four of the five clones tested. This finding suggests that FR residues 233-237 are essential in properly juxtaposing the FR hydrophobic domain. Together, these data support the hypothesis that the postulated Ser234 is not the only potential cleavage/attachment site of the alpha isoform of FR.     

Immunoelectron microscopic analysis of the binding of monoclonal antibodies to molecular variants of human placental alkaline phosphatase

Three monoclonal antibodies with distinct antigenic specificities were examined by electron microscopy for their binding to three common genetic variants (SS, FS, and FF) of human placental alkaline phosphatase. In the reaction with the monoclonal antibody H5, all three variants of human placental alkaline phosphatase preferentially formed circular immune complexes composed of two antibodies and two enzyme molecules. In separate reactions with the F11 and B2 monoclonal antibodies, the SS variant formed circular complexes and the FS variant formed Y-shaped complexes composed of one antibody and two enzyme molecules, whereas the FF variant scarcely reacted. These results confirm immunochemical data showing that H5 binds to both S and F subunits with similar affinities, whereas F11 and B2 bind the S subunit with markedly higher affinity than they do the F subunit. Furthermore, the formation of circular complexes in the reaction of the mixture of the two antibodies, F11 and B2, with FS molecules suggests that these two antibodies bind to different sites on the S subunit. Therefore, the F and S subunits differ from one another at more than one site. This is the first indication that alleles of human placental alkaline phosphatase may result from more than just single point mutations in the gene encoding them.     

Site-directed mutagenesis of a single tryptophan near the middle of the channel-forming toxin aerolysin inhibits its transfer across the outer membrane of Aeromonas salmonicida

The channel-forming protein aerolysin must cross both the inner and outer bacterial membranes during its secretion from Aeromonas hydrophila or from Aeromonas salmonicida containing the cloned structural gene. We examined the fate of three mutant proteins in which Trp-227, near the middle of the amino acid chain, was replaced with glycine, leucine, or phenylalanine by site-directed mutagenesis. All three proteins crossed the inner membrane and entered the periplasm in the same way as wild-type, and in each case the signal sequence was removed correctly. Little or none of the proaerolysin substituted with glycine or leucine was released into the culture supernatant. Instead, significant amounts became associated with the outer membrane. The Phe-227 protoxin was secreted by the bacteria but at a reduced rate. The leucine and phenylalanine mutant proteins were purified and compared with native proaerolysin. They were processed correctly to the mature forms by treatment with trypsin, and like native aerolysin, both were resistant to further proteolysis. In each case, processing was followed by the formation of oligomers similar to those produced by native toxin. The hemolytic activity of the processed Phe-227 mutant was one-quarter that of wild-type toxin whereas Leu-227 aerolysin had less than one-hundredth the wild-type activity. These results are further evidence that aerolysin is secreted in at least two steps. As well, they show that the last step, crossing the outer membrane, can be blocked by an apparently small change in the structure of the protein.