Analysis of the signal for attachment of a glycophospholipid membrane anchor

The COOH terminus of decay accelerating factor (DAF) contains a signal that directs attachment of a glycophospholipid (GPI) membrane anchor. To define this signal we deleted portions of the DAF COOH terminus and expressed the mutant cDNAs it CV1 origin-deficient SV-40 cells. Our results show that the COOH-terminal hydrophobic domain (17 residues) is absolutely required for GPI anchor attachment. However, when fused to the COOH terminus of a secreted protein this hydrophobic domain is insufficient to direct attachment of a GPI anchor. Additional specific information located within the adjacent 20 residues appears to be necessary. We speculate that by analogy with signal sequences for membrane translocation, GPI anchor attachment requires both a COOH-terminal hydrophobic domain (the GPI signal) as well as a suitable cleavage/attachment site located NH2 terminal to the signal.     

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.     

Glycophospholipid membrane anchor attachment. Molecular analysis of the cleavage/attachment site.

The COOH terminus of decay accelerating factor (DAF) contains a signal that directs attachment of a glycophosphatidylinositol (GPI) membrane anchor in a process involving proteolytic removal of 17-31 COOH-terminal residues. Previous work suggested that two elements are required for anchor addition, a COOH-terminal hydrophobic domain (the GPI signal) and an element located NH2-terminal to it, postulated to be the cleavage/attachment site. Using [3H]ethanolamine (a component of the anchor) to tag the COOH terminus, we isolated and sequenced a COOH-terminal tryptic peptide, thereby identifying Ser-319 as the COOH-terminal residue attached to the GPI anchor. This indicates that a 28-residue peptide is removed during processing and localizes the cleavage/attachment site precisely to the region previously shown to be required for anchor attachment (between 10 and 20 residues NH2-terminal to the hydrophobic domain). Since DAF contains multiple cryptic cleavage/attachment sites, we used a GPI-linked human growth hormone-DAF fusion to study the structural requirements for cleavage/attachment. Our results show that while sequences immediately NH2-terminal to the attachment site are not required for anchor addition, deletion of Ser-319 abolishes both anchor attachment and transport to the cell surface. Systematic replacement of the attachment site serine with all possible amino acids indicated that alanine, aspartate, asparagine, glycine, or serine efficiently support GPI anchor attachment while valine and glutamate are partially effective. All other substitutions including cysteine (permitted at the attachment site in other GPI-anchored proteins) abolish both GPI anchor attachment and transport to the cell surface, resulting in accumulation of uncleaved fusion protein in internal compartments (endoplasmic reticulum and Golgi). These results support the general rule that the residue at the cleavage/attachment site must be small. Further, addition of a GPI anchor appears to be necessary for transport to the cell surface in transfected COS cells.     

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.     

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.     

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.     

The endoplasmic reticulum (ER) translocon can differentiate between hydrophobic sequences allowing signals for glycosylphosphatidylinositol anchor addition to be fully translocated into the ER lumen

The signal sequence within polypeptide chains that designates whether a protein is to be anchored to the membrane by a glycosylphosphatidylinositol (GPI) anchor is characterized by a carboxyl-terminal hydrophobic domain preceded by a short hydrophilic spacer linked to the GPI anchor attachment (omega) site. The hydrophobic domain within the GPI anchor signal sequence is very similar to a transmembrane domain within a stop transfer sequence. To investigate whether the GPI anchor signal sequence is translocated across or integrated into the endoplasmic reticulum membrane we studied the translocation, GPI anchor addition, and glycosylation of different variants of a model GPI-anchored protein. Our results unequivocally demonstrated that the hydrophobic domain within a GPI signal cannot act as a transmembrane domain and is fully translocated even when followed by an authentic charged cytosolic tail sequence. However, a single amino acid change within the hydrophobic domain of the GPI-signal converts it into a transmembrane domain that is fully integrated into the endoplasmic reticulum membrane. These results demonstrated that the translocation machinery can recognize and differentiate subtle changes in hydrophobic sequence allowing either full translocation or membrane integration.     

Identification of a novel functional specificity signal within the GPI anchor signal sequence of carcinoembryonic antigen

Exchanging the glycophosphatidylinositol (GPI) anchor signal sequence of neural cell adhesion molecule (NCAM) for the signal sequence of carcinoembryonic antigen (CEA) generates a mature protein with NCAM external domains but CEA-like tumorigenic activity. We hypothesized that this resulted from the presence of a functional specificity signal within this sequence and generated CEA/NCAM chimeras to identify this signal. Replacing the residues (GLSAG) 6-10 amino acids downstream of the CEA anchor addition site with the corresponding NCAM residues resulted in GPI-anchored proteins lacking the CEA-like biological functions of integrin modulation and differentiation blockage. Transferring this region from CEA into NCAM in conjunction with the upstream proline (PGLSAG) was sufficient to specify the addition of the CEA anchor. Therefore, this study identifies a novel specificity signal consisting of six amino acids located within the GPI anchor attachment signal, which is necessary and sufficient to specify the addition of a particular functional GPI anchor and, thereby, the ultimate function of the mature protein.     

Yeast Gaa1p is required for attachment of a completed GPI anchor onto proteins

Anchoring of proteins to membranes by glycosylphosphatidylinositols (GPIs) is ubiquitous among all eukaryotes and heavily used by parasitic protozoa. GPI is synthesized and transferred en bloc to form GPI- anchored proteins. The key enzyme in this process is a putative GPI:protein transamidase that would cleave a peptide bond near the COOH terminus of the protein and attach the GPI by an amide linkage. We have identified a gene, GAA1, encoding an essential ER protein required for GPI anchoring. gaal mutant cells synthesize the complete GPI anchor precursor at nonpermissive temperatures, but do not attach it to proteins. Overexpression of GAA1 improves the ability of cells to attach anchors to a GPI-anchored protein with a mutant anchor attachment site. Therefore, Gaa1p is required for a terminal step of GPI anchor attachment and could be part of the putative GPI:protein transamidase.     

Recognition of the carboxyl-terminal signal for GPI modification requires translocation of its hydrophobic domain across the ER membrane

A carboxyl-terminal hydrophobic domain is an essential component of the processed signal for attachment of the glycosyl-phosphatidylinositol (GPI) membrane anchor to proteins and it is linked to the site (omega) of GPI modification by a spacer domain. This study was designed to test the hypothesis that the hydrophobic domain interacts with the lipid bilayer of the endoplasmic reticulum (ER) membrane to optimally position the omega site for GPI modification. The hydrophobic domain of the GPI signal in the human folate receptor (FR) type alpha was substituted with the carboxyl-terminal segment of the low-density lipoprotein receptor (LDLR), including its membrane spanning region, without altering either the spacer or the omega site. The FR-alpha/LDLR chimera was not GPI modified but was attached to the plasma membrane by a polypeptide anchor. When the carboxyl-terminal half of the hydrophobic transmembrane polypeptide in the FR-alpha/LDLR chimera was altered by introduction of negatively charged (Asp) residues, or when the cytosolic domain in the chimera was deleted, the mutated proteins became GPI-anchored. On the other hand, attachment of a carboxyl-terminal segment of LDLR including the entire cytosolic domain to FR-alpha converted it into a transmembrane protein. The results indicate that in the FR-alpha/LDLR chimera the inability of the cellular machinery for GPI modification to recognize the hydrophobic domain is not due to the intrinsic nature of the peptide, but is rather due to the retention of the peptide within the lipid bilayer. It follows that the hydrophobic domain in the signal for GPI modification must traverse the ER membrane prior to recognition of the omega site by the GPI-protein transamidase. The results thus establish a critical topographical requirement for recognition of the GPI signal in the ER.     

A glycophospholipid membrane anchor acts as an apical targeting signal in polarized epithelial cells

Glycosyl-phosphatidylinositol- (GPI) anchored proteins contain a large extracellular protein domain that is linked to the membrane via a glycosylated form of phosphatidylinositol. We recently reported the polarized apical distribution of all endogenous GPI-anchored proteins in the MDCK cell line (Lisanti, M. P., M. Sargiacomo, L. Graeve, A. R. Saltiel, and E. Rodriguez-Boulan. 1988. Proc. Natl. Acad. Sci. USA. 85:9557-9561). To study the role of this mechanism of membrane anchoring in targeting to the apical cell surface, we use here decay-accelerating factor (DAF) as a model GPI-anchored protein. Endogenous DAF was localized on the apical surface of two human intestinal cell lines (Caco-2 and SK-CO15). Recombinant DAF, expressed in MDCK cells, also assumed a polarized apical distribution. Transfer of the 37-amino acid DAF signal for GPI attachment to the ectodomain of herpes simplex glycoprotein D (a basolateral antigen) and to human growth hormone (a regulated secretory protein) by recombinant DNA methods resulted in delivery of the fusion proteins to the apical surface of transfected MDCK cells. These results are consistent with the notion that the GPI anchoring mechanism may convey apical targeting information.     

The carboxyl terminus of Dictyostelium discoideum protein 1I encodes a functional glycosyl-phosphatidylinositol signal sequence

The 1I gene is expressed in the prespore cells of culminating Dictyostelium discoideum. The open reading frame of 1I cDNA encodes a protein of 155 amino acids with hydrophobic segments at both its NH(2)- and COOH-termini that are indicative of a glycosyl-phosphatidylinositol (GPI)-anchored protein. A hexaHis-tagged form of 1I expressed in D. discoideum cells appeared on Western blot analysis as a doublet of 27 and 24 kDa, with a minor polypeptide of 22 kDa. None of the polypeptides were released from the cell surface with bacterial phosphatidylinositol-specific phospholipase C, although all three were released upon nitrous acid treatment, indicating the presence of a phospholipase-resistant GPI anchor. Further evidence for the C-terminal sequence of 1I acting as a GPI attachment signal was obtained by replacing the GPI anchor signal sequence of porcine membrane dipeptidase with that from 1I. Two constructs of dipeptidase with the 1I GPI signal sequence were constructed, one of which included an additional six amino acids in the hydrophilic spacer. Both of the resultant constructs were targeted to the surface of COS cells and were GPI-anchored as shown by digestion with phospholipase C, indicating that the Dictyostelium GPI signal sequence is functional in mammalian cells. Site-specific antibodies recognising epitopes either side of the expected GPI anchor attachment site were used to determine the site of GPI anchor attachment in the constructs. These parallel approaches show that the C-terminal signal sequence of 1I can direct the addition of a GPI anchor.     

Proper protein folding in the endoplasmic reticulum is required for attachment of a glycosylphosphatidylinositol anchor in plants

Endoplasmic reticulum (ER) quality control processes recognize and eliminate misfolded proteins to maintain cellular protein homeostasis and prevent the accumulation of defective proteins in the secretory pathway. Glycosylphosphatidylinositol (GPI)-anchored proteins carry a glycolipid modification, which provides an efficient ER export signal and potentially prevents the entry into ER-associated degradation (ERAD), which is one of the major pathways for clearance of terminally misfolded proteins from the ER. Here, we analyzed the degradation routes of different misfolded glycoproteins carrying a C-terminal GPI-attachment signal peptide in Arabidopsis thaliana. We found that a fusion protein consisting of the misfolded extracellular domain from Arabidopsis STRUBBELIG and the GPI-anchor attachment sequence of COBRA1 was efficiently targeted to hydroxymethylglutaryl reductase degradation protein 1 complex-mediated ERAD without the detectable attachment of a GPI anchor. Non-native variants of the GPI-anchored lipid transfer protein 1 (LTPG1) that lack a severely misfolded domain, on the other hand, are modified with a GPI anchor and targeted to the vacuole for degradation. Impaired processing of the GPI-anchoring signal peptide by mutation of the cleavage site or in a GPI-transamidase-compromised mutant caused ER retention and routed the non-native LTPG1 to ERAD. Collectively, these results indicate that for severely misfolded proteins, ER quality control processes are dominant over ER export. For less severely misfolded proteins, the GPI anchor provides an efficient ER export signal resulting in transport to the vacuole.

Severely misfolded proteins carrying a glycosylphosphatidylinositol (GPI)-anchor attachment sequence undergo a stringent quality control process in the endoplasmic reticulum that prevents GPI anchoring.     

The specificity for the differentiation blocking activity of carcinoembryonic antigen resides in its glycophosphatidyl-inositol anchor

Ectopic expression of various members of the human carcinoembryonic antigen (CEA) family of intercellular adhesion molecules in murine myoblasts either blocks (CEA, CEACAM6) or allows (CEACAM1) myogenic differentiation. These surface glycoproteins form a subset of the immunoglobulin (Ig) superfamily and are very closely related, but differ in the precise sequence of their external domains and in their mode of anchorage to the cell membrane. CEA and CEACAM6 are glycophosphatidyl-inositol (GPI) anchored, whereas CEACAM1 is transmembrane (TM) anchored. Overexpression of GPI-linked neural cell adhesion molecule (NCAM) p125, also an adhesion molecule of the Ig superfamily, accelerates myogenic differentiation. The molecular requirements for the myogenic differentiation block were investigated using chimeric constructs in which the COOH-terminal hydrophobic domains of CEA, CEACAM1, and NCAM p125 were exchanged. The presence of the GPI signal sequence specifically from CEA in the chimeras was sufficient to convert both CEACAM1 and NCAM into differentiation-blocking proteins. Conversely, CEA could be converted into a neutral protein by exchanging its GPI anchor for the TM anchor of CEACAM1. Since the external domains of CEA, CEACAM1, and NCAM can all undergo homophilic interactions, and mutations in the self-adhesive domains of CEA abrogate its differentiation-blocking activity, the structural requirements for differentiation-inhibition are any self-adhesive domains attached to the specific GPI anchor derived from CEA. We therefore suggest that biologically significant functional information resides in the processed extreme COOH terminus of CEA and in the GPI anchor that it determines.     

Glycosylphosphatidylinositol-anchored proteins are preferentially targeted to the basolateral surface in Fischer rat thyroid epithelial cells

Glycosylphosphatidylinositol (GPI) acts as an apical targeting signal in MDCK cells and other kidney and intestinal cell lines. In striking contrast with these model polarized cell lines, we show here that Fischer rat thyroid (FRT) epithelial cells do not display a preferential apical distribution of GPI-anchored proteins. Six out of nine detectable endogenous GPI-anchored proteins were localized on the basolateral surface, whereas two others were apical and one was not polarized. Transfection of several model GPI proteins, previously shown to be apically targeted in MDCK cells, also led to unexpected results. While the ectodomain of decay accelerating factor (DAF) was apically secreted, 50% of the native, GPI-anchored form, of this protein was basolateral. Addition of a GPI anchor to the ectodomain of Herpes simplex gD-1, secreted without polarity, led to basolateral localization of the fusion protein, gD1-DAF. Targeting experiments demonstrated that gD1-DAF was delivered vectorially from the Golgi apparatus to the basolateral surface. These results indicate that FRT cells have fundamental differences with MDCK cells with regard to the mechanisms for sorting GPI-anchored proteins: GPI is not an apical signal but, rather, it behaves as a basolateral signal. The "mutant" behavior of FRT cells may provide clues to the nature of the mechanisms that sort GPI-anchored proteins in epithelial cells.     

Aberrant processing of alkaline phosphatase precursor caused by blocking the synthesis of glycosylphosphatidylinositol.

Alkaline phosphatase is anchored to the membrane via glycosylphosphatidylinositol (GPI). Mannose residues of the GPI glycan are suggested to be derived from dolichol-P-mannose. In the present study we examined the effect of 2-fluoro-2-deoxy-D-glucose (F-Glc), an inhibitor of dolichol-P-mannose synthesis, on the biosynthesis and processing of alkaline phosphatase in JEG-3 cells. In control cells, a proform precursor (64.5 kDa) with a hydrophobic peptide domain at the COOH terminus was immediately processed into an intermediate form (63 kDa) by proteolytic removal of the COOH-terminal extension and replacement with the GPI anchor, and then to a mature form (66 kDa) by terminal glycosylation of its N-linked oligosaccharides. In contrast, when cells were treated with F-Glc (1 mM), the protein was synthesized as a proform of 61 kDa. The reduction in its molecular mass was mostly due to the inhibition in maturation of N-linked oligosaccharides by F-Glc. The 61-kDa proform identified by antibodies to the COOH-terminal peptide was detectable even at 3 h after the synthesis, and was gradually processed to doublet forms of 58-59 kDa which were finally secreted into the medium. None of these forms were labeled with [3H]ethanolamine and [3H]stearic acid, components of the GPI anchor, and expressed on the cell surface as a membrane-bound form. Taken together, these results suggest that the inhibition of the GPI synthesis causes a prolonged accumulation of the proform, which is then gradually processed into secretory forms by proteolytic removal of the COOH-terminal hydrophobic peptide.     

Early events in glycosylphosphatidylinositol anchor addition. substrate proteins associate with the transamidase subunit gpi8p

The addition of glycosylphosphatidylinositol (GPI) anchors to proteins occurs by a transamidase-catalyzed reaction mechanism soon after completion of polypeptide synthesis and translocation. We show that placental alkaline phosphatase becomes efficiently GPI-anchored when translated in the presence of semipermeabilized K562 cells but is not GPI-anchored in cell lines defective in the transamidase subunit hGpi8p. By studying the synthesis of placental alkaline phosphatase, we demonstrate that folding of the protein is not influenced by the addition of a GPI anchor and conversely that GPI anchor addition does not require protein folding. These results demonstrate that folding of the ectodomain and GPI addition are two distinct processes and can be mutually exclusive. When GPI addition is prevented, either by synthesis of the protein in the presence of cell lines defective in GPI addition or by mutation of the GPI carboxyl-terminal signal sequence cleavage site, the substrate forms a prolonged association with the transamidase subunit hGpi8p. The ability of the transamidase to recognize and associate with GPI anchor signal sequences provides an explanation for the retention of GPI-anchored protein within the ER in the absence of GPI anchor addition.     

Defining the boundaries of species specificity for the Saccharomyces cerevisiae glycosylphosphatidylinositol transamidase using a quantitative in?vivo assay

In eukaryotes, GPI (glycosylphosphatidylinositol) lipid anchoring of proteins is an abundant post-translational modification. The attachment of the GPI anchor is mediated by GPI-T (GPI transamidase), a multimeric, membrane-bound enzyme located in the ER (endoplasmic reticulum). Upon modification, GPI-anchored proteins enter the secretory pathway and ultimately become tethered to the cell surface by association with the plasma membrane and, in yeast, by covalent attachment to the outer glucan layer. This work demonstrates a novel in vivo assay for GPI-T. Saccharomyces cerevisiae INV (invertase), a soluble secreted protein, was converted into a substrate for GPI-T by appending the C-terminal 21 amino acid GPI-T signal sequence from the S. cerevisiae Yapsin 2 [Mkc7p (Y21)] on to the C-terminus of INV. Using a colorimetric assay and biochemical partitioning, extracellular presentation of GPI-anchored INV was shown. Two human GPI-T signal sequences were also tested and each showed diminished extracellular INV activity, consistent with lower levels of GPI anchoring and species specificity. Human/fungal chimaeric signal sequences identified a small region of five amino acids that was predominantly responsible for this species specificity.     

Efficient glycosylphosphatidylinositol (GPI) modification of membrane proteins requires a C-terminal anchoring signal of marginal hydrophobicity

Many plasma membrane proteins are anchored to the membrane via a C-terminal glycosylphosphatidylinositol (GPI) moiety. The GPI anchor is attached to the protein in the endoplasmic reticulum by transamidation, a reaction in which a C-terminal GPI-attachment signal is cleaved off concomitantly with addition of the GPI moiety. GPI-attachment signals are poorly conserved on the sequence level but are all composed of a polar segment that includes the GPI-attachment site followed by a hydrophobic segment located at the very C terminus of the protein. Here, we show that efficient GPI modification requires that the hydrophobicity of the C-terminal segment is "marginal": less hydrophobic than type II transmembrane anchors and more hydrophobic than the most hydrophobic segments found in secreted proteins. We further show that the GPI-attachment signal can be modified by the transamidase irrespective of whether it is first released into the lumen of the endoplasmic reticulum or is retained in the endoplasmic reticulum membrane.