Analysis of the signal for attachment of a glycophospholipid membrane anchor

Immunochemical and molecular analysis of antigen binding to lipid anchored and soluble forms of an MHC independent human alpha/beta T cell receptor

We have constructed antigen-specific chimeric human T cell receptor (TCR) molecules deleted of the transmembrane domain and containing the signal sequence for the biosynthesis of the phosphatidyl inositol glycan (GPI) linkage. These membrane-anchored forms of the TCR alpha and beta chains have been expressed in non-T cells, and they are recognized by alpha or beta TCR specific monoclonal antibodies. We have utilized both immunochemical methods and flow cytometry to prove that the enzyme phosphatidylinositol phospholipase C (PI/PLC) is able to cleave the GPI anchored TCR as a heterodimer from the CHO cell surface. We have demonstrated that the alpha/beta TCR heterodimer on the surface of CHO cells will recognize and bind polymers containing fluorescein (FL-polymer), and the binding activity is completely eliminated by the enzyme, PI/PLC. Moreover, soluble forms of the alpha/beta heterodimer will bind tightly to FL substituted sepharose, which demonstrates the retention of biological activity by the TCR after solubilization. Molecular modelling of the putative antigen binding site of the alpha FL beta FL TCR was derived from the known atomic coordinates of eight different hapten or peptide specific antibodies. Mutagenesis of several residues predicted from the model to be important in FL binding gave results consistent with involvement of Ig equivalent CDR2 and CDR3 domains in the antigen binding pocket. Therefore, using a model hapten system in studying recognition of the TCR independent of MHC interactions, we conclude that amino acid residues located in similar positions within CDR domains as compared to the case of MHC restricted TCR recognition are used in the binding of either hapten or peptide antigens.

Uncoupled expression of Moloney murine leukemia virus envelope polypeptides SU and TM: a functional analysis of the role of TM domains in viral entry

Moloney murine leukemia virus ecotropic envelope expression plasmids were used to demonstrate that the synthesis of the retroviral envelope SU and TM polypeptides can be uncoupled with retention of biologic activity. By substituting a glycosyl-phosphatidylinositol (GPI) membrane anchor for part or all of the retroviral envelope transmembrane protein and creating several deletion variants of the TM subunit, we have begun to dissect the role of the TM protein in envelope function. We show that a GPI-anchored envelope can be incorporated into virions and binds receptor. We found that the envelope cytoplasmic tail, while not required, influences the efficiency of retroviral transduction at some step after membrane fusion (possibly by interacting with core). The membrane-spanning domain of TM is involved in membrane fusion, and this function is distinct from its role as a membrane anchor. As few as eight amino acids of the putative membrane-spanning domain are sufficient to achieve membrane anchoring of envelope but not to mediate membrane fusion. In addition, though not required, the membrane-spanning domain may have some direct role in the incorporation of envelope into virions. Finally, the extracellular domain of TM, besides containing the putative fusion domain and interacting with SU, may influence the synthesis or stability and the glycosylation of envelope, possibly by affecting oligomerization of the complex and proper intracellular transit.

Recombinant soluble CD59 inhibits reactive haemolysis with complement

Three soluble forms of membrane attack complex inhibitory factor (MACIF or CD59) were prepared using recombinant baculovirus-infected insect cells. They consisted of 70, 77 and 86 amino acids, starting from the amino terminus of naturally occurring CD59, and were designated recombinant (r) CD59 70, 77 and 86, respectively. All three rCD59 lacked a glycosyl-phosphatidylinositol (GPI) anchor, unlike membrane CD59 which has a GPI anchor at the anchor at the carboxyl terminus (77th amino acid). Their activities in inhibiting complement activation were assayed with C5b-7 intermediate cells and C8 and C9 components. The inhibitory activity of rCD59 70 was as high as that of rCD59 77 and twice that of rCD59 86. In addition, it was one-fourth and one-hundredth lower than the activities of urine and erythrocyte CD59, respectively. However, when assayed in the presence of human serum at a final concentration of 50% (v/v), the activities of both urine and erythrocyte CD59 were greatly decreased to to one-tenth of that of rCD59 70. Purified rCD59 70 molecules were all glycosylated, but rCD59 77 and 86 were mixtures of glycosylated and non-glycosylated molecules. The inhibitory activities of rCD59 77 and 86 were the same for the glycosylated and non-glycosylated forms. These results suggest that the soluble rCD59 provide a means for elucidating the biological roles of CD59.

Requirements for glycosylphosphatidylinositol attachment are similar but not identical in mammalian cells and parasitic protozoa

The general features of the glycosylphosphatidylinositol (GPI) signal have been conserved in evolution. To test whether the requirements for GPI attachment are indeed the same in mammalian cells and parasitic protozoa, we expressed the prototype GPI-linked protein of Trypanosoma brucei, the variant surface glycoprotein (VSG), in COS cells. Although large amounts of VSG were produced, only a small fraction became GPI linked. This impaired processing is not caused by the VSG ectodomain, since replacement of the VSG GPI signal with that of decay accelerating factor (DAF) produced GPI-linked VSG. Furthermore, whereas fusion of the DAF GPI signal to the COOH terminus of human growth hormone (hGH) produces GPI-linked hGH, an analogous hGH fusion using the VSG GPI signal does not, indicating that the VSG GPI signal functions poorly in mammalian cells. By constructing chimeric VSG-DAF GPI signals and fusing them to the COOH terminus of hGH, we show that of the two critical elements that comprise the GPI-signal--the cleavage/attachment site and the COOH terminal hydrophobic domain--the former is responsible for the impaired activity of the VSG GPI signal in COS cells. To confirm this, we show that the VSG GPI signal can be converted to a viable signal for mammalian cells by altering the amino acid configuration at the cleavage/attachment site. We also show that when fused to the COOH terminus of hGH, the putative GPI signal from the malaria circumsporozoite (CS) protein produces low levels of GPI-anchored hGH, suggesting that the CS protein is indeed GPI linked, but that the CS protein GPI signal, like the VSG-signal, functions poorly in COS cells. The finding that the requirements for GPI attachment are similar but not identical in parasitic protozoa and mammalian cells may allow for the development of selective inhibitors of GPI-anchoring that might prove useful as antiparasite therapeutics.

Expression of an enzymically active glycosylphosphatidylinositol-anchored form of neutral endopeptidase (EC 3.4.24.11) in Cos-1 cells

Neutral endopeptidase (EC 3.4.24.11, NEP) is a type-II integral membrane protein found in a wide variety of cell types. We previously produced a secreted form of the enzyme by deletion of the cytoplasmic and transmembrane domains and in-frame fusion of the cleavable signal peptide of pro-opiomelanocortin [Lemay, Waksman, Roques, Crine and Boileau (1989) J. Biol. Chem. 264, 15620-15623]. Here we have used this secreted form of NEP and fused to it the glycosylphosphatidylinositol (GPI)-anchor attachment signal of decay-accelerating factor to produce a GPI-anchored form. Expression of this chimeric form in Cos-1 cells resulted in cell-surface activity. This activity could be released from the cell surface by phosphatidylinositol-specific phospholipase C and radiolabelling studies showed that the protein could incorporate [3H]ethanolamine, indicating that the enzyme was GPI-anchored. The Km value, using [D-Ala2,Leu5]enkephalin as substrate, of GPI-anchored NEP (62 +/- 5 microM) was comparable with that of wild-type NEP (70 +/- 4 microM), as were the sensitivities to the inhibitors phosphoramidon and thiorphan. However, pulse-chase studies showed that the biosynthesis and cell-surface delivery of GPI-anchored NEP was delayed compared with that of the wild-type transmembrane form of NEP. These results suggest a lower rate of biosynthesis and/or cellular transport for GPI-anchored NEP compared with its transmembrane counterpart.

Statistical prediction of the locus of endoproteolytic cleavage of the nascent polypeptide in glycosylphosphatidylinositol-anchored proteins

Existing methods of identifying the cleavage site of the nascent polypeptide and the C-terminal residue to which the glycosylphosphatidylinositol (GPI) anchor is attached in mature GPI-anchored proteins are technically difficult and labour-intensive. We tested the hypothesis that it was possible to predict this locus using data from the cDNA-deduced amino acid sequence and amino acid composition of GPI-anchored proteins. We employed a statistical approach which allowed repeated chi 2 comparisons between the proportions of residual amino acids in the major body of the cDNA-deduced polypeptide (minus the N-terminal signal peptide) after repeated computer-generated progressive exoproteolysis from its C-terminus one amino acid at a time and the fixed proportion of amino acids obtained from amino acid analysis of the mature GPI-anchored protein. Initial comparison of the two parameters invariably revealed a relatively high chi 2 statistic which progressively lowered to a minimum point at which the amino acid proportions of progressively exoproteolysed polypeptide and fixed endoproteolysed polypeptides of the mature GPI-anchored protein were in closest agreement. This objectively defined and unique minimum point of closest agreement accurately identified the locus of post-translational endoproteolytic cleavage of the nascent polypeptide in several tissue-specific single-gene-encoded GPI-anchored proteins. Thus the C-terminal amino acid to which the GPI anchor is attached can be rapidly identified using data from the cDNA sequence and the amino acid composition of proteins suspected to be GPI-anchored.

Axonal glycoproteins with immunoglobulin- and fibronectin type III-related domains in vertebrates: structural features, binding activities, and signal transduction

The L1- and F11-like axonal glycoproteins, implicated in neurite outgrowth and fasciculation, are members of the Ig superfamily comprising multiple fibronectin type III-like domains. Their Ig-like and fibronectin type III-related domains are likely to be composed of seven beta-strands arranged in two opposing beta-sheets of highly similar topology. Whereas the F11-like molecules lack a transmembrane sequence and are anchored in the plasma membrane by a glycosylphosphatidylinositol, the L1-like molecules comprise cytoplasmic domains with highly conserved sequence motifs. Most of the latter proteins occur in different isoforms generated by alternative pre-mRNA splicing, which has not been documented for molecules of the F11 subgroup. L1-like proteins undergo heterophilic as well as homophilic interactions, whereas only the former mode of binding was observed for F11-like proteins. Evidence is accumulating that these Ig superfamily molecules with fibronectin type III-like domains are interacting in a complex manner with each other and molecules of the extracellular matrix. Investigations assigning structure to function reveal that their individual extracellular domains serve distinct binding activities. Recent studies also suggest that L1 and NCAM are implicated in the transduction of transmembrane signals.

Carboxyl terminus structural requirements for glycosyl-phosphatidylinositol anchor addition to cell surface proteins

Glycosyl phosphatidylinositol (GPI)-anchored proteins contain in their COOH-terminal region a peptide segment that is thought to direct glycolipid addition. This signal has been shown to require a pair of small amino acids positioned 10–12 residues upstream of an hydrophobic C-terminal domain. We analysed the contribution of the region separating the anchor acceptor site and the C-terminal hydrophobic segment by introducing amino acid deletions and substitutions in the spacer element of the GPI-anchored Thy-1 glycoprotein. Deletions of 7 amino acids in this region, as well as the introduction of 2 charged residues, prevented the glycolipid addition to Thy-1, suggesting that the length and the primary sequence of the spacer domain are important determinants in the signal directing GPI anchor transfer onto a newly synthesized polypeptide. Furthermore, we tested these rules by creating a truncated form of the normally transmembranous Herpes simplex virus I glycoprotein D (gDI) and demonstrating that when its C-terminal region displays all the features of a GPI-anchored protein, it is able to direct glycolipid addition onto another cell surface molecule.

Membrane proteins in paroxysmal nocturnal haemoglobinuria

A review of recent information on the abnormalities of the blood cell membrane in paroxysmal nocturnal haemoglobinuria (PNH) is presented, with a detailed analysis of biochemical and flow cytometry findings. The complex patterns observed in the various cell lineages of which the PNH clone consists are described, and a simplified monoclonal antibody panel is defined for diagnostic purposes. Available data on in vitro culture of progenitor cells and on the recent establishment of PNH cell lines are summarized. Finally, we discuss speculative hypotheses on the growth advantage of the PNH clone.

Biosynthesis of glycosylphosphatidylinositol (GPI)-anchored membrane proteins in intact cells: specific amino acid requirements adjacent to the site of cleavage and GPI attachment

Mutational studies were previously carried out at the omega site intact cells (Micanovic, R., L. Gerber, J. Berger, K. Kodukula, and S. Udenfriend. 1990. Proc. Natl. Acad. Sci. USA. 87:157-161; Micanovic R., K. Kodukula, L. Gerber, and S. Udenfriend. 1990. Proc. Natl. Acad. Sci. USA: 87:7939-7943) and at the omega + 1 and omega + 2 sites in a cell-free system (Gerber, L., K. Kodukula, and S. Udenfriend. 1992. J. Biol. Chem. 267:12168-12173) of nascent proteins destined to be processed to a glycosylphosphatidyl-inositol (GPI)-anchored form. We have now mutated the omega + 1 and omega + 2 sites in placental alkaline phosphatase (PLAP) cDNA and transfected the wild-type and mutant cDNAs into COS 7 cells. Only glycine at the omega + 2 site yielded enzymatically active GPI membrane-anchored PLAP in amounts comparable to the wild type (alanine). Serine was less active and threonine and valine yielded very low but significant activity. By contrast the omega + 1 site was promiscuous, with only proline being inactive. These and the previous studies indicate that the omega and omega + 2 sites of a nascent protein are key determinants for recognition by COOH-terminal signal transamidase. Comparisons have been made to specific requirements for substitution at the -1, -3 sites of amino terminal signal peptides for recognition by NH2-terminal signal peptidase and the mechanisms of NH2 and COOH-terminal signaling are compared.

Proteins containing an uncleaved signal for glycophosphatidylinositol membrane anchor attachment are retained in a post-ER compartment

Glycophosphatidylinositol (GPI)-anchored membrane proteins are initially synthesized with a cleavable COOH-terminal extension that signals anchor attachment. Overexpression in COS cells of hGH-DAF fusion proteins containing the GPI signal of decay accelerating factor (DAF) fused to the COOH-terminus of human growth hormone (hGH), produces both GPI-anchored hGH-DAF and uncleaved precursors that retain the GPI signal. Using hGH-DAF fusion proteins containing a mutated, noncleavable GPI signal, we show that uncleaved polypeptides are retained inside the cell and accumulate in a brefeldin A-sensitive, Golgi-like juxtanuclear structure. Retention requires the presence of either a functional or a noncleavable GPI signal; hGH-DAF fusion proteins containing only the COOH-terminal hydrophobic domain (a component of the GPI signal) are secreted. Immunofluorescence analysis shows colocalization of the retained, uncleaved fusion proteins with both a Golgi marker and with p53, a marker of the ER-Golgi intermediate compartment. Since N-linked glycosylation is postulated to facilitate the transport of proteins to the cell surface, we engineered a glycosylation site into hGH-DAF. Glycosylation failed to completely override the transport block, but allowed some uncleaved hGH-DAF to pass through the secretory pathway and acquire endoglycosidase H resistance. The retained molecules remained endoglycosidase H sensitive. We suggest that the uncleaved fusion protein is retained in a sorting compartment between the ER and the medial Golgi complex. We speculate that a mechanism exists to retain proteins containing an uncleaved GPI signal as part of a system for quality control.

Structural requirements for glycosyl-phosphatidylinositol-anchor attachment in the cellular receptor for urokinase plasminogen activator

The urokinase-plasminogen-activator receptor (u-PAR) is a glycosyl-phosphatidylinositol(glycosyl-PtdIns)-anchored membrane protein. Using site-directed mutagenesis, we have studied features in the u-PAR sequence important for successful glycosyl-PtdIns attachment. Two critical sequence elements were identified. In the sequence Ser282-Gly283-Ala284, simultaneous substitution of all of these residues prevented membrane anchoring. Individual substitution of each of the residues indicated that Gly283 is the more critical residue and the likely attachment site. However, it was unexpectedly found that mutation of this residue gave rise only to a partial impairment of glycosyl-PtdIns attachment. We therefore propose that more than one residue within this sequence can be utilized as glycosyl-PtdIns-attachment site. In the last eight COOH-terminal amino acids encoded in u-PAR cDNA, deletion of this sequence (residues 306-313) completely prevented glycosyl-PtdIns attachment. However, the remaining COOH-terminal region proved still to possess a potential glycosyl-PtdIns signal activity; it could be converted to a new functional glycosyl-PtdIns signal by substitution of a single positively charged residue (Arg304). Substitution of Arg304 by Leu converted this truntaced u-PAR to a glycosyl-PtdIns-anchored protein, indistinguishable from the wild type. Substitution of Arg304 by a negatively charged residue (Glu) led to a partial acquisition of the glycosyl-PtdIns-anchoring ability. These findings show that charged amino acids placed in the COOH-terminus interfere negatively with glycosyl-PtdIns-anchoring, and, furthermore, that this effect is more pronounced for positively charged than for negatively charged amino acid residues.

Conversion of human interferon-beta from a secreted to a phosphatidylinositol anchored protein by fusion of a 17 amino acid sequence to its carboxyl terminus

A number of cell-surface proteins are anchored in plasma membranes by a glycosylated phosphatidylinositol (PI) moiety that is covalently attached to the carboxyl-terminal amino acid of the mature protein. We have previously reported the construction of a cDNA clone of a truncated Platelet-derived growth factor (PDGF) receptor that consists of the extracellular domain without the transmembrane and cytoplasmic domains. In the construction of the vector, a sequence of 51 base pairs (bp) from the 3'-untranslated region of the receptor cDNA was linked in frame with the external domain coding sequence. The truncated receptor protein with the peptide VTSGHCHEERVDRHDGE fused to its carboxyl terminus was covalently attached to the membrane by a PI linkage and it was released by phosphatidylinositol specific-phospholipase C (PI-PLC). When the 51 bp sequence was deleted, the external domain receptor protein was secreted into the media. To determine whether the PI linkage of the protein was due to the 17 amino acids added, the peptide was fused to the carboxyl terminus of the secreted protein human Interferon-beta (hu-IFN-beta). Chinese hamster ovary (CHO) cells transfected with the hu-IFN-beta cDNA secreted the protein to the conditioned media, whereas CHO cells transfected with the carboxyl terminus modified-hu-IFN-beta cDNA did not secrete detectable levels of protein. CHO cells expressing the carboxyl terminus modified-hu-IFN-beta were treated with PI-PLC, the media and cell lysates were analyzed by SDS-PAGE after immunoprecipitation with antibodies against hy-IFN-beta. The modified protein is anchored to the plasma membrane by a PI linkage and it is specifically released by PI-PLC, whereas a control preparation of CHO cells expressing wild type hu-IFN-beta does not show the same pattern. The 17 amino acid peptide fused to the carboxyl terminus of IFN-beta directs attachment of a PI anchor and targets the fusion protein to the plasma membrane.

A cDNA construct of tissue inhibitor of metalloproteinases (TIMP) linked to the last exon of Thy-1 confers glycophospholipid anchorage on this naturally secreted protein

A naturally secreted protein, tissue inhibitor of metalloproteinases (TIMP), has been transiently expressed on the surface of transfected COS cells and stably on transfected murine BW 5147 thymoma cells, by linkage of the entire coding sequence of the cDNA to the last exon of Thy-1. Thy-1 is a glycophospholipid-linked protein. In COS cells the chimaeric protein can be labelled by [3H]ethanolamine, which is a component of glycophospholipid anchors. Ltk- cells cannot anchor proteins by glycan phosphatidylinositol linkage and were found to be unable to express the engineered protein extracellularly on their plasma membranes. Phosphatidylinositol-specific phospholipase C treatment released 90% of the protein from all BW 5147 cells, but very little from the COS-1 cells. It is concluded that the last exon of Thy-1 has conferred the property of glycophospholipid anchorage on the normally secreted protein TIMP.

Biosynthesis of phosphatidylinositol-glycan (PI-G)-anchored membrane proteins in cell-free systems: cleavage of the nascent protein and addition of the PI-G moiety depend on the size of the COOH-terminal signal peptide

Nascent translation products of PI-G-anchored membrane proteins contain both NH2- and COOH-terminal signal sequences of approximately 15-30 residues that are removed during processing. Removal of the latter occurs concomitant with the addition of the PI-G moiety to the newly formed COOH terminus. In human placental alkaline phosphatase (PLAP) the COOH-terminal signal peptide contains 29 residues. An engineered form of PLAP, miniPLAP 208, containing the same NH2- and COOH-terminal signal peptides as PLAP, was used as a substrate for cell-free processing. A comparison was made with mutants (delta 202, delta 197, delta 184, and delta 179) truncated at the COOH terminus. Intact preprominiPLAP 208 and truncated delta 202 were processed to yield the same mature product which, by size and distribution between Triton X-114 and water before and after treatment with inositol-specific phospholipases, indicates that it contained the PI-G moiety. Mutants that were further truncated at the COOH terminus, miniPLAPs delta 197, delta 184, and delta 179, were processed only at their NH2 termini. Those portions of the COOH-terminal sequence in miniPLAPs delta 197 and delta 1984 that extended beyond residue 179 were not removed during processing.

Site-specific mutations in the COOH-terminus of placental alkaline phosphatase: a single amino acid change converts a phosphatidylinositol-glycan-anchored protein to a secreted protein

Placental alkaline phosphatase (PLAP) is anchored in the plasma membrane by a phosphatidylinositol-glycan moiety (PI-glycan). PI-glycan is added posttranslationally to the nascent peptide chain after the removal of 29 amino acids from the COOH-terminus. The contribution of selected COOH-terminal amino acids to the signal for PI-glycan addition was tested by creating a fusion protein with the COOH-terminus of PLAP and a secreted protein and by mutagenesis of specific PLAP COOH-terminal amino acids. The cDNA encoding the COOH-terminus of PLAP was fused in frame to the cDNA for human clotting Factor X and expressed in transfected COS-1 cells. Fusion proteins containing 32 amino acids of the PLAP COOH-terminus were modified by PI-glycan addition. Thus, the signal for PI-glycan modification must reside in these amino acids. Next, the region between the hydrophobic domain and the cleavage site was examined for additional determinants. Mutations of the hydrophilic residues in the spacer region demonstrated that these amino acids do not contribute to the signal for PI-glycan addition. Deletion of amino acids in the spacer region prevented the addition of PI-glycan suggesting that the length of the spacer domain or the amino acids around the cleavage site are important determinants. Finally, we demonstrated that interruption of the hydrophobic domain by a charged residue prevents PI-glycan addition and results in a protein that is secreted into the medium. The finding that a single Leu to Arg substitution in the hydrophobic domain converts a PI-glycan anchored, membrane protein to a secreted protein suggests that an essential signal for the correct sorting of PI-glycan anchored proteins versus secreted proteins resides in the hydrophobic domain. Substitution of a charged amino acid for a hydrophobic amino acid may be a mechanism for producing membrane bound and secreted forms of the same protein.

Two lipid-anchored cAMP-binding proteins in the yeast Saccharomyces cerevisiae are unrelated to the R subunit of cytoplasmic protein kinase A

We show that the yeast, Saccharomyces cerevisiae, contains two cAMP-binding proteins in addition to the well-characterized regulatory (R) subunit of cytoplasmic cAMP-dependent protein kinase (PKA). We provide evidence that they comprise a new type of cAMP receptor, membrane-anchored by covalently attached lipid structures. They are genetically not related to the cytoplasmic R subunit. The respective proteins can be detected in sral mutants, in which the gene for the R subunit of PKA has been disrupted and a monoclonal antibody raised against the cytoplasmic R subunit does not cross-react with the two membrane-bound cAMP-binding proteins. In addition, they differ from the cytoplasmic species also with respect to their location and the peptide maps of the photoaffinity-labeled proteins. Although they differ from one another in molecular mass and subcellular location, peptide maps of the cAMP-binding domains resemble each other and both proteins are membrane-anchored by lipid structures, one to the outer surface of the plasma membrane, the other to the outer surface of the inner mitochondrial membrane. Both anchors can be metabolically labeled by Etn, myo-Ins and fatty acids. In addition, the anchor structure of the cAMP receptor from plasma membranes can be radiolabeled by GlcN and Man. After cleavage of the anchor with glycosylphosphatidylinositol-specific phospholipase C from trypanosomes, the solubilized cAMP-binding protein from plasma membranes reacts with antibodies which specifically recognize the cross-reacting determinant from soluble trypanosomal coat protein, suggesting similarity of the anchors. Degradation studies also point to the glycosylphosphatidylinositol nature of the anchor from the plasma membrane, whereas the mitochondrial counterpart is less complex in that it lacks carbohydrates. The plasma membrane cAMP receptor is, in addition, modified by an N-glycosidically linked carbohydrate side chain, responsible mainly for its higher molecular mass.

Two alternatively spliced mouse urokinase receptor mRNAs with different histological localization in the gastrointestinal tract

Two mouse urokinase-type plasminogen activator receptor (muPAR) cDNAs were isolated: muPAR1 is homologous to the human urokinase-type plasminogen activator receptor while muPAR2 codes for a 199 residue protein sharing the first 133 residues with muPAR1. Mouse genomic DNA sequencing indicates that the two different mRNAs arise by alternative splicing. In situ hybridization showed differential expression of the two mRNAs in mouse gastric mucosa. muPAR1 mRNA is located in luminal epithelial cells situated close to urokinase-type plasminogen activator-producing connective tissue cells of the lamina propria, pointing to plasmin generation controlled by the cooperation of different cells that may play a role in the release of gastric epithelial cells. muPAR2 mRNA is expressed in the basal epithelial cells, and the deduced protein sequence includes the receptor ligand binding domain, but omits the region involved in glycolipid-mediated membrane anchoring, suggesting that muPAR2 may code for a secreted uPA binding protein.

Fusion of sequence elements from non-anchored proteins to generate a fully functional signal for glycophosphatidylinositol membrane anchor attachment

Glycophosphatidylinositol (GPI) membrane anchor attachment is directed by a cleavable signal at the COOH terminus of the protein. The complete lack of homology among different GPI-anchored proteins suggests that this signal is of a general nature. Previous analysis of the GPI signal of decay accelerating factor (DAF) suggests that the minimal requirements for GPI attachment are (a) a hydrophobic domain and (b) a cleavage/attachment site consisting of a pair of small residues positioned 10-12 residues NH2-terminal to a hydrophobic domain. As an ultimate test of these rules we constructed four synthetic GPI signals, meeting these requirements but assembled entirely from sequence elements not normally involved in GPI attachment. We show that these synthetic signals are able to direct human growth hormone (hGH), a secreted protein, to the plasma membrane via a GPI anchor. Our results indicate that different hydrophobic sequences, derived from either the prolactin or hGH NH2-terminal signal peptide, can be linked to different cleavage sites via different hydrophilic spacers to produce a functional GPI signal. These data confirm that the only requirements for GPI-anchoring are a pair of small residues positioned 10-12 residues NH2 terminal to a hydrophobic domain, no other structural motifs being necessary.

Phosphatidylinositol-glycan linked proteins of the erythrocyte membrane

The human erythrocyte bears a number of proteins anchored to the outer membrane surface via a phosphatidylinositol-glycan linkage. This class of proteins includes several complement regulatory proteins (including decay-accelerating factor, CD59 antigen (protectin), and C8 binding protein) as well as several enzymes and at least one protein important in cell-cell interaction. In addition, a number of blood group antigens have been identified to reside on proteins with phosphatidylinositol anchors. One blood group (Cromer) resides on DAF. Study of variants in this blood group system has led to interesting information about the function and expression of this protein. Several other blood groups, such as JMH and Holley/Gregory, appear to reside on as yet unidentified phosphatidylinositol-linked proteins. In paroxysmal nocturnal haemoglobinuria, a variable proportion of red cells fail to express or express weakly all phosphatidylinositol-linked proteins. The origin of this deficiency is now being worked out. In addition, individuals with inherited deficiency of DAF or CD59 (protectin) have been identified. Only the latter deficiency leads to a PNH-like syndrome.

A cAMP-binding ectoprotein in the yeast Saccharomyces cerevisiae

Purified plasma membranes from the yeast Saccharomyces cerevisiae bind about 1.2 pmol of cAMP/mg of protein with high affinity (Kd = 6 nM). By using photoaffinity labeling with 8-N3-[32P]cAMP, we have identified in plasma membrane vesicles a cAMP-binding protein (Mr = 54,000) that is present also in bcy1 disruption mutants, lacking the cytoplasmic R subunit of protein kinase A (PKA). This argues that it is genetically unrelated to PKA. Neither high salt, nor alkaline carbonate, nor cAMP extract the protein from the membrane, suggesting that it is not peripherally bound. The observation that (glycosyl)phosphatidylinositol-specific phospholipases (or nitrous acid) release the amphiphilic protein from the membrane, thereby converting it to a hydrophilic form, indicates anchorage by a glycolipidic membrane anchor. Treatment with N-glycanase reduces the Mr to 44,000-46,000 indicative of a modification by N-linked carbohydrate side chain(s). In addition to the action of a phospholipase, the efficient release from the membrane requires the removal of the carbohydrate side chain(s) or the presence of high salt or methyl alpha-mannopyranoside, suggesting complex interactions with the membrane involving not only the glycolipidic anchor but also the glycan side chain(s). Topological studies show that the protein is exposed to the periplasmic space, raising intriguing questions for the function of this protein.

A nonfunctional sequence converted to a signal for glycophosphatidylinositol membrane anchor attachment

The COOH terminus of decay-accelerating factor (DAF) contains a signal that directs glycophosphatidylinositol (GPI) membrane anchor attachment in a process involving concerted proteolytic removal of 28 COOH-terminal residues. At least two elements are required for anchor addition: a COOH-terminal hydrophobic domain and a cleavage/attachment site located NH2-terminal to it, requiring a small amino acid as the acceptor for GPI addition. We previously showed that the last 29-37 residues of DAF, making up the COOH-terminal hydrophobic domain plus 20 residues of the adjacent serine/threonine-rich domain (including the anchor addition site), when fused to the COOH terminus of human growth hormone (hGH) will target the fusion protein to the plasma membrane via a GPI anchor. In contrast, a similar fusion protein (hGH-LDLR-DAF17, abbreviated HLD) containing a fragment of the serine/threonine-rich domain of the LDL receptor (LDLR) in place of the DAF-derived serine/threonine-rich sequences, does not become GPI anchored. We now show that this null sequence for GPI attachment can be converted to a strong GPI signal by mutating a pair of residues (valine-glutamate) in the LDLR sequence at a position corresponding to the normal cleavage/attachment site, to serine-glycine, as found in the DAF sequence. A single mutation (converting valine at the anchor addition site to serine, the normal acceptor for GPI addition in DAF) was insufficient to produce GPI anchoring, as was mutation of the valine-glutamate pair to serine-phenylalanine (a bulky residue). These results suggest that a pair of small residues (presumably flanking the cleavage point) is required for GPI attachment. By introducing the sequence serine-glycine (comprising a cleavage-attachment site for GPI addition) at different positions in the LDLR sequence of the fusion protein, HLD, we show that optimal GPI attachment requires a processing site positioned 10-12 residues NH2-terminal to the hydrophobic domain, the efficiency anchor attachment dropping off sharply as the cleavage site is moved beyond these limits. These data suggest that the GPI signal consists solely of a hydrophobic domain combined with a processing site composed of a pair of small residues, positioned 10-12 residues NH2-terminal to the hydrophobic domain. No other structural motifs appear necessary.

Mannosamine, a novel inhibitor of glycosylphosphatidylinositol incorporation into proteins

Mannosamine (2-amino-2-deoxy D-mannose) is shown here to block the incorporation of glycosylphosphatidylinositol (GPI) into GPI-anchored proteins. The amino sugar drastically reduced the surface expression of a recombinant GPI-anchored protein in polarized MDCK cells, converted this apical membrane-bound protein to an unpolarized secretory product and blocked the expression of endogenous GPI-anchored proteins. Furthermore, it specifically inhibited the incorporation of [3H]ethanolamine (a GPI component) into mammalian and trypanosomal GPI-anchored proteins and into a well characterized GPI-lipid of Trypanosoma brucei. These results suggest that mannosamine converted an apical GPI-anchored protein to a non-polarized secretory product by depleting transfer competent GPI-precursor lipids. Our inhibitor studies provide new independent evidence for the apical targeting role of GPI in polarized epithelia and open the way towards a greater understanding of the functional role of GPI in membrane trafficking and cell regulation.

Human immunodeficiency virus infection and syncytium formation in HeLa cells expressing glycophospholipid-anchored CD4

The CD4 molecule, a glycoprotein expressed primarily on the cell surface of specific T lymphocytes, is thought to function in T-cell antigen recognition and activation. In addition, CD4 serves as a receptor for human immunodeficiency virus type 1 (HIV-1) by a direct interaction with the HIV-1 surface glycoprotein (gp120). To further characterize the HIV-1-cell interaction, a HeLa cell line was established that expressed a chimeric molecule of CD4 and decay-accelerating factor (DAF). In the chimeric CD4-DAF molecule the transmembrane and cytoplasmic domains of CD4 were deleted and replaced with the carboxy-terminal 37 amino acids of DAF. This resulted in the anchoring of the extracellular domain of CD4 to the cell membrane via a glycophospholipid linkage. The glycophospholipid-anchored CD4 had a molecular size of approximately 56 to 62 kDa and was released following treatment of the cells with phosphatidylinositol-specific phospholipase C. HeLa cells expressing the CD4-DAF hybrid could be infected with HIV-1, as evidenced by reverse transcriptase activity, p24 core antigen content, and infectious virus production. In addition, transfection of the HeLa CD4-DAF cells with a plasmid that directs the synthesis of HIV-1 envelope glycoproteins or cocultivation with HeLa cells expressing the virus glycoproteins resulted in syncytium formation. These results indicate that the transmembrane and cytoplasmic domains of the CD4 molecule are dispensable for both HIV infection and syncytium formation.

Thy-1 expresses two signals for apical localization in epithelial cells

Recent work has shown that anchoring via glycosyl phosphatidylinositol (GPI) results in apical targeting for a variety of endogenous and transfected plasma membrane proteins expressed in epithelial cells. To further determine the correlation between GPI anchoring and apical localization, we expressed GPI-anchored and secretory forms of Thy-1 in Madin-Darby canine kidney cells by transfection. Native GPI-anchored Thy-1, normally expressed in thymocytes and neurons, was localized to the apical surface. A truncated form of Thy-1, lacking 22 out of 31 hydrophobic amino acids at the COOH-terminus, was also constructed; this deletion blocked attachment of the GPI anchor and resulted in apical secretion of Thy-1. In combination with previous results, our observations indicate that Thy-1 contains apical targeting information in its protein sequence as well as in the GPI anchor.

An internally positioned signal can direct attachment of a glycophospholipid membrane anchor

All known glycophosphatidylinositol (GPI)-anchored membrane proteins contain a COOH-terminal hydrophobic domain necessary for signalling anchor attachment. To examine the requirement that this signal be at the COOH terminus of the protein, we constructed a chimeric protein, DAFhGH, in which human growth hormone (hGH) was fused to the COOH terminus of decay accelerating factor (DAF) (a GPI-anchored protein), thereby placing the GPI signal in the middle of the chimeric protein. We show that the fusion protein appears to be processed at the normal DAF processing site in COS cells, producing GPI-anchored DAF on the cell surface. This result indicates that the GPI signal does not have to be at the COOH terminus to direct anchor addition, suggesting that the absence of a hydrophilic COOH-terminal extension (beyond the hydrophobic domain) is not a necessary requirement for GPI anchoring. A similar DAFhGH fusion, containing an internal GPI signal in which the DAF hydrophobic domain was replaced with the signal peptide of hGH, also produced GPI-anchored cell surface DAF. The signal for GPI attachment thus exhibits neither position specificity nor sequence specificity. In addition, mutant DAF or DAFhGH constructs lacking an NH2-terminal signal peptide failed to produce GPI-anchored protein, suggesting that membrane translocation is necessary for anchor addition.

Determinants for glycophospholipid anchoring of the Saccharomyces cerevisiae GAS1 protein to the plasma membrane

A 125-kDa glycoprotein exposed on the surface of Saccharomyces cerevisiae cells belongs to a class of eucaryotic membrane proteins anchored to the lipid bilayer by covalent linkage to an inositol-containing glycophospholipid. We have cloned the gene (GAS1) encoding the 125-kDa protein (Gas1p) and found that the function of Gas1p is not essential for cell viability. The nucleotide sequence of GAS1 predicts a 60-kDa polypeptide with a cleavable N-terminal signal sequence, potential sites for N- and O-linked glycosylation, and a C-terminal hydrophobic domain. Determination of the anchor attachment site revealed that the C-terminal hydrophobic domain of Gas1p is removed during anchor addition. However, this domain is essential for addition of the glycophospholipid anchor, since a truncated form of the protein failed to become attached to the membrane. Anchor addition was also abolished by a point mutation affecting the hydrophobic character of the C-terminal sequence. We conclude that glycophospholipid anchoring of Gas1p depends on the integrity of the C-terminal hydrophobic domain that is removed during anchor attachment.

Determinants for glycophospholipid anchoring of the Saccharomyces cerevisiae GAS1 protein to the plasma membrane

A 125-kDa glycoprotein exposed on the surface of Saccharomyces cerevisiae cells belongs to a class of eucaryotic membrane proteins anchored to the lipid bilayer by covalent linkage to an inositol-containing glycophospholipid. We have cloned the gene (GAS1) encoding the 125-kDa protein (Gas1p) and found that the function of Gas1p is not essential for cell viability. The nucleotide sequence of GAS1 predicts a 60-kDa polypeptide with a cleavable N-terminal signal sequence, potential sites for N- and O-linked glycosylation, and a C-terminal hydrophobic domain. Determination of the anchor attachment site revealed that the C-terminal hydrophobic domain of Gas1p is removed during anchor addition. However, this domain is essential for addition of the glycophospholipid anchor, since a truncated form of the protein failed to become attached to the membrane. Anchor addition was also abolished by a point mutation affecting the hydrophobic character of the C-terminal sequence. We conclude that glycophospholipid anchoring of Gas1p depends on the integrity of the C-terminal hydrophobic domain that is removed during anchor attachment.

Cell adhesion in transformed D. discoideum cells: expression of gp80 and its biochemical characterization

Dictyostelium discoideum amoebae were transformed with an expression vector for the gp80, a protein believed to mediate EDTA-resistant cell adhesion in developmental cells. Vegetative cells, that do not normally contain gp80, expressed the protein and this expression was correlated with the formation of cell-cell adhesions. These contacts exhibited minimal EDTA-resistance. Biochemical analyses of the protein synthesized by vegetative cells suggested that it is identical to that produced by aggregation-competent cells, including the presence of a glycolipid anchor. Additional experiments indicated that the anchor was insensitive to hydrolysis by exogenous (glycosly)phosphatidylinositol-specific phospholipase C [G)PI-PLCs) but was sensitive to the endogenous anchor degrading enzyme. This enzyme, initially described in aggregating cells (da Silva and Klein, Exp. Cell Res., in press) was found to be present also in vegetative amoebae.

The axonal glycoprotein TAG-1 is an immunoglobulin superfamily member with neurite outgrowth-promoting activity

Pathfinding of axons in the developing nervous system is thought to be mediated by glycoproteins expressed on the surface of embryonic axons and growth cones. One molecule suggested to play a role in axonal growth is TAG-1, a 135 kd glycoprotein expressed transiently on the surface of subsets of neurons in the developing mammalian nervous system. We isolated a full-length cDNA clone encoding rat TAG-1. TAG-1 has six immunoglobulin-like domains and four fibronectin type III-like repeats and is structurally similar to other immunoglobulin-like proteins expressed on developing axons. Neurons maintained in vitro on a substrate of TAG-1 extend long neurites, suggesting that this protein plays a role in the initial growth and guidance of axons in vivo. TAG-1 is anchored to the neuronal membrane via a glycosyl phosphatidylinositol linkage and is also released from neurons, suggesting that TAG-1 also functions as a substrate adhesion molecule when released into the extracellular environment.

Glycophospholipid membrane anchoring provides clues to the mechanism of protein sorting in polarized epithelial cells

The sorting signals that guide proteins to apical and basolateral surfaces of epithelial cells have remained elusive. Current evidence suggests a hierarchy of sorting information with multiple sorting signals (apical and basolateral) present in different domains of a given plasma membrane protein. The observation that covalently attached glycosyl-phosphatidylinositol (GPI) acts as a 'dominant' apical targeting signal is compatible with the involvement of glycolipids in epithelial protein sorting.

Single gene encodes glycophospholipid-anchored and asymmetric acetylcholinesterase forms: alternative coding exons contain inverted repeat sequences

Polymorphic forms of acetylcholinesterase are tethered extracellularly either as dimers membrane-anchored by a glycophospholipid or as catalytic subunits disulfidelinked to a collagen tail that associates with the basal lamina. Genomic clones of acetylcholinesterase from T. californica revealed that individual enzyme forms are encoded within a single gene that yields multiple mRNAs. Each enzyme form is encoded in three exons: the first two exons, bases -22 to 1502 and 1503 to 1669, encode sequence common to both forms, while alternative third exons encode a hydrophobic C-terminal region, to which a glycophospholipid is added upon processing, and a nonprocessed C-terminus, yielding a catalytic subunit that disulfide-links with a collagen-like structural unit. The 3' untranslated region of each alternative exon contains tandem repeat sequences that are inverted with respect to the other exon. This may either dictate alternative exon usage by formation of cis stem-loops or affect the abundance of translatable mRNA by trans-hybridization between the alternative spliced mRNA species.

Selectivity of the cleavage/attachment site of phosphatidylinositol-glycan-anchored membrane proteins determined by site-specific mutagenesis at Asp-484 of placental alkaline phosphatase

Many proteins are now known to be anchored to the plasma membrane by a phosphatidylinositol-glycan (PI-G) moiety that is attached to their COOH termini. Placental alkaline phosphatase (PLAP) has been used as a model for investigating mechanisms involved in the COOH-terminal processing of PI-G-tailed proteins. The COOH-terminal domain of pre-pro-PLAP provides a signal for processing during which a largely hydrophobic 29-residue COOH-terminal peptide is removed, and the PI-G moiety is added to the newly exposed Asp-484 terminus. This cleavage/attachment site was subjected to an almost saturation mutagenesis, and the enzymatic activities, COOH-terminal processing, and cellular localizations of the various mutant PLAP forms were determined. Substitution of Asp-484 by glycine, alanine, cysteine, asparagine, or serine (category I) resulted in PI-G-tailed and enzymatically active proteins. However, not all category I mutant proteins were PI-G tailed to the same extent. Pre-pro-PLAP with other substituents at position 484 (threonine, proline, methionine, valine, leucine, tyrosine, tryptophan, lysine, glutamic acid, and glutamine; category II) were expressed, as well as the category I amino acids, but there was little or no processing to the PI-G-tailed form, and this latter group exhibited very low enzyme activity. The bulk of the PLAP protein produced by category II mutants and some produced by category I mutants were sequestered within the cell, apparently in the endoplasmic reticulum (ER). Most likely, certain amino acids at residue 484 are preferred because they yield better substrates for the putative "transamidating" enzyme. In transfected COS cells, at least, posttranslational PI-G-tail processing does not go to completion even for preferred substrates. Apparently PI-G tailing is a requisite for transport from the ER and for PLAP enzyme activity. Proteins that are not transamidated are apparently retained in the ER in an inactive conformation.