Expression of surface lymphotoxin and tumor necrosis factor on activated T, B, and natural killer cells.

The expression of membrane-associated forms of lymphotoxin (LT) and TNF were examined on cell lines of T, B, and myeloid origin, IL-2 dependent T cell clones, and peripheral blood lymphocytes. Inducible and constitutive patterns of surface LT expression were found on T cells as exemplified by the II-23.D7, a CD4+T cell hybridoma, and HUT-78, a T cell lymphoma. Phorbol ester induced surface LT expression on Ramos, an EBV transformed B cell line, but at a slower rate of appearance when compared to the II-23.D7. Secretion of LT was rapidly inducible by phorbol ester in II-23.D7 and also in HUT-78 but with slower kinetics; surface LT expression continued in both lines after secretion had ceased. Low levels of membrane TNF were transiently induced on II-23.D7 and HUT-78, but none was observed on Ramos. Peripheral blood monocytes and some myeloid tumor lines did not express surface LT. Several T cell clones expressed surface LT after Ag-specific stimulation, and expression persisted several days. Stimulation through the TCR or by IL-2 rapidly induced surface LT on resting peripheral T cells and CD56+ NK cells; pokeweed mitogen activation induced expression on CD20+ B cells. Consistent with previous results, immunoprecipitation with anti-LT mAb showed that LT was complexed with a distinct 33 kDa glycoprotein (p33) on cells that expressed surface LT, whereas secreted LT was not associated with p33. Surface and secreted modes of LT expression by activated T, B, and NK cells suggests that LT can be utilized as either a localized or diffusible mediator in immune responses.     

Lymphotoxin and an associated 33-kDa glycoprotein are expressed on the surface of an activated human T cell hybridoma.

A human T cell hybridoma, II-23.D7, was induced with phorbol ester to express a surface form of lymphotoxin (LT, TNF-beta) and an associated 33-kDa glycoprotein. The LT epitopes were detected by surface immunofluorescence staining and by immunoprecipitation from radioiodinated or biosynthetically labeled cells with the use of anti-rLT polyclonal and monoclonal antibodies. The epitopes detected by the antibody were related to LT because adsorption of the anti-rLT with PMA-activated II-23.D7 cells resulted in the removal of the neutralizing titer of the anti-rLT antiserum. Immunoprecipitation of surface radioiodinated II-23.D7 cells revealed two bands of 25 kDa and 33 kDa that were specifically precipitated with anti-rLT, but not anti-rTNF antibodies. Enzymatic digestion with glycanases showed both proteins to have N-linked carbohydrate, with O-linked sugar limited to the 25-kDa protein. To determine the biochemical relationship between these proteins, the two LT-like forms were purified from detergent-solubilized II-23.D7 cells by immunoaffinity chromatography. Peptide mapping using CNBr cleavage showed the 25-kDa surface form to be identical to rLT, whereas the 33-kDa protein was different. Biosynthetic labeling studies showed that p33 contained both methionine and cysteine, whereas the p25 contained only methionine. Thus, the surface LT form lacks a leader peptide indicating an anchoring mechanism distinct from that described for membrane TNF. The nature of the attachment of this LT form to the membrane surface is not clear, however, neither TNF receptor binding nor lipid linkages appear to be involved. The accessory protein, p33, may anchor LT to the surface. These findings identify a new characteristic of LT and point toward an additional pathway by which T lymphocytes may mediate cytolytic activity and regulate inflammatory processes.     

Further characterization of the membrane anchor found on the tissue-specific class I molecule Qa2

Previous studies have determined that various Qa2 serologic determinants can be removed from the surface of spleen cells by treatment with a phospholipase C. Our studies have determined that the class I molecule Qa2, expressed on the surface of spleen cells and activated T cells, behaves as an integral membrane protein based on its ability to associate with detergent micelles. Studies utilizing two purified phospholipase C have revealed that although most (90 to 95%) of the Qa2 molecules expressed on the surface of resting spleen cells are released as intact 40-kDa polypeptides associated with beta 2-microglobulin, activated T cells contain a major cell subpopulation expressing lipase-resistant Qa2 molecules. Flow cytometric analysis revealed that L3T4+-activated T cells expressed lipase-sensitive Qa2 molecules, whereas Lyt-2+ cells express lipase-resistant forms of the Qa2 molecule. The relationship between the secreted form of the Qa2 molecule and the lipase-generated soluble Qa2 molecule was investigated. Based on SDS-PAGE analysis, the secreted Qa2 molecules has a Mr of 39 kDa whereas the cell surface form released from either resting spleen or activated T cells by phosphatidylinositol-specific phospholipase C has a Mr of approximately equal to 40 kDa. Furthermore, the secreted Qa2 molecule lacks an epitope, cross-reacting determinant, often present on lipase-solubilized cell surface molecules. Thus, based on serologic and biochemical criteria, the soluble Qa2 molecules generated by an exogenous phospholipase C and the secreted Qa2 molecule are structurally distinct.     

Bacterial surface association of heat-labile enterotoxin through lipopolysaccharide after secretion via the general secretory pathway

Heat-labile enterotoxin (LT) is an important virulence factor expressed by enterotoxigenic Escherichia coli. The route of LT secretion through the outer membrane and the cellular and extracellular localization of secreted LT were examined. Using a fluorescently labeled receptor, LT was found to be specifically secreted onto the surface of wild type enterotoxigenic Escherichia coli. The main terminal branch of the general secretory pathway (GSP) was necessary and sufficient to localize LT to the bacterial surface in a K-12 strain. LT is a heteromeric toxin, and we determined that its cell surface localization was mediated by the its B subunit independent of an intact G(M1) ganglioside binding site and that LT binds lipopolysaccharide and G(M1) concurrently. The majority of LT secreted into the culture supernatant by the GSP in E. coli associated with vesicles. Only a mutation in hns, not overexpression of the GSP or LT, caused an increase in vesicle yield, supporting a specific vesicle formation machinery regulated by the nucleoid-associated protein HNS. We propose a model in which LT is secreted by the GSP across the outer membrane, secreted LT binds lipopolysaccharide via a G(M1)-independent binding region on its B subunit, and LT on the surface of released outer membrane vesicles interacts with host cell receptors, leading to intoxication. These data explain a novel mechanism of vesicle-mediated receptor-dependent delivery of a bacterial toxin into a host cell.     

Genomic characterization of LIGHT reveals linkage to an immune response locus on chromosome 19p13.3 and distinct isoforms generated by alternate splicing or proteolysis.

LIGHT is a member of the TNF cytokine superfamily that signals through the lymphotoxin (LT)beta receptor and the herpesvirus entry mediator. LIGHT may function as a costimulatory factor for the activation of lymphoid cells and as a deterrent to infection by herpesvirus, which may provide significant selective pressure shaping the evolution of LIGHT. Here, we define the molecular genetics of the human LIGHT locus, revealing its close linkage to the TNF superfamily members CD27 ligand and 4-1BB ligand, and the third complement protein (C3), which positions LIGHT within the MHC paralog on chromosome 19p13.3. An alternately spliced isoform of LIGHT mRNA that encodes a transmembrane-deleted form is detected in activated T cells and gives rise to a nonglycosylated protein that resides in the cytosol. Furthermore, membrane LIGHT is shed from the cell surface of human 293 T cells. These studies reveal new mechanisms involved in regulating the physical forms and cellular compartmentalization of LIGHT that may contribute to the regulation and biological function of this cytokine.     

Regulation of lymphotoxin production by the p21ras-raf-MEK-ERK cascade in PHA/PMA-stimulated Jurkat cells

Although the production of lymphotoxin (LT) from activated Th1 lymphocytes has been reported extensively, the intracellular signaling mechanisms that regulate this T cell function remain totally undefined. We have examined whether the p21ras-raf-1-mitogen-activated protein kinase/extracellular signal-regulated protein kinase (ERK) kinase (MEK)-ERK cascade plays a role in regulating the production of LT, because the activity of these signaling molecules is up-regulated in activated T lymphocytes. Transfection of Jurkat leukemic T cells with a dominant negative mutant of p21ras (ras17N or ras15A), raf-1 (raf 1-130), or ERK1 (Erk1-K71R) resulted in the suppression of the mitogen/phorbol ester-stimulated production/secretion of LT. This suppression was accompanied by a parallel inhibition of mitogen-stimulated ERK activation. The selective antagonist of MEK1 activation, PD98059, also attenuated the mitogen-stimulated or anti-CD3 Ab and phorbol ester-stimulated production of LT from Jurkat cells or peripheral blood T lymphocytes. This study provides, for the first time, direct evidence that the p21ras-raf-MEK-ERK cascade plays a vital role in regulating the production of LT.     

Biochemical study of a recombinant soluble interleukin-2 receptor. Evidence for a homodimeric structure

A truncated soluble form of the human interleukin-2 receptor p55 chain (T-S-IL-2R) was expressed to high levels in RODENT (mammalian) cells and affinity-purified. Its biochemical behavior was analyzed by polyacrylamide gel electrophoresis (PAGE), gel filtration, and sucrose gradient centrifugation. It migrated as a single 40-kDa band on sodium dodecyl sulfate-PAGE (reducing or nonreducing conditions), whereas it ran as a 80-kDa component on native PAGE or as a 86-kDa component on gel filtration. The combination of gel filtration and density gradient sedimentation gave a Stokes radius of 4.0 nm and a sedimentation coefficient of 3.72 S. The deduced molecular mass was 67 kDa, and the fractional ratio was 1.516. These data therefore indicated that the T-S-IL-2R was secreted as an homodimer of two noncovalently associated 40-kDa subunits. Cross-linking experiments using bifunctional reagents enabled the materialization of the dimeric structure on sodium dodecyl sulfate-PAGE. Stoichiometric binding studies using two monoclonal antibodies (mAbs 33B3.1 and 11H2) reacting with separate epitopes on the p55 chain also strongly supported the dimeric structure. Indeed, there was one binding site for the 33B3.1 mAb (and Fab fragment) per T-S-IL-2R 40-kDa subunit, whereas the 11H2 mAb (or Fab fragment) could bind only half a site per subunit, a result which could only be explained when assuming more than one subunit for the native T-S-IL-2R. Soluble interleukin-2 receptor species were also purified from culture supernatants of either L cells transfected with the full-length p55 cDNA or a normal alloreactive T cell clone. Similar biochemical behavior and reactivities with the two mAbs were found. Finally, cell-surface p55 chains expressed either by pgL21 or 4AS cells bound the 33B3.1 and 11H2 mAbs in a 2:1 ratio, suggesting that the p55 chains are also associated as homodimers when imbedded in the membrane.     

Production of transmembrane and secreted forms of tumor necrosis factor (TNF)-alpha by HIV-1-specific CD4+ cytolytic T lymphocyte clones. Evidence for a TNF-alpha-independent cytolytic mechanism.

Candidate AIDS vaccines consisting of recombinant forms of the HIV-1 envelope glycoprotein induce, in seronegative human volunteers, an env-specific T cell response that includes CD4+, MHC class II-restricted CTL capable of lysing HIV-1-infected target cells. In this study, we have analyzed the production of the cytokines TNF-alpha and lymphotoxin (LT) by a set of env-specific CD4+ human CTL clones. TNF-alpha and LT are of interest because of their potential role in target cell destruction by CD4+ CTL. Our studies focused on the possibility that a cell surface form of TNF-alpha expressed by CTL after physiologic activation with target APC might participate in the cytolytic reactions mediated by these clones. We found that, upon interaction with target cells expressing env epitopes in the context of the appropriate MHC class II molecules, CD4+ CTL released TNF-alpha with kinetics that were rapid, compared with other cytokines, and that were generally similar to the kinetics of target cell destruction. LT secretion was not detected during the time course of the cytolytic reactions. A novel flow cytometric assay was used to show that physiologic activation of CD4+ CTL with target APC induced expression by the CTL of cell surface forms of TNF-alpha. Immunoprecipitations from activated, surface-iodinated CTL clones revealed two forms of surface TNF-alpha, a 26-kDa form, representing the transmembrane precursor of secreted TNF-alpha, as well as the 17-kDa secreted form bound to the cell surface. For a subset of CD4+ CTL, we found that treatment of CTL with cyclosporin A inhibited Ag-induced production of both transmembrane and secreted forms of TNF-alpha but had no effect on cytolysis. Thus, although transmembrane and secreted TNF-alpha produced by HIV-1-specific CD4+ CTL may have important effects in vivo, the rapid destruction of target APC by the set of CD4+ CTL clones described here occurs through a TNF-alpha-independent mechanism.     

Identification of membrane-associated lymphotoxin (LT) on mitogen-activated human lymphocytes using heterologous anti-LT antisera in vitro.

Surface-associated lymphotoxin (LT) molecules have been identified on mitogen-activated human lymphocytes employing heterologous anti-α-LT serum in vitro. These membrane-associated LT molecules are present on PHA- or Con A-activated lymphocytes but do not appear to be expressed on unstimulated cells. Furthermore, these molecules were detected primarily on activated T lymphocytes, with little detectable on activated B- or null-cell populations. The removal of surface LT-bearing lymphocytes, using anti-α-LT serum + C′, does not dramatically affect the capacity of the remaining cells to release LT after mitogen restimulation. In addition, the presence of toxic molecules on the surface of activated lymphocytes suggests that these materials may be expressed in an inactive, noncytotoxic form.     

GM1 binding-deficient exotoxin is a potent noninflammatory broad spectrum intradermal immunoadjuvant

Intradermal (i.d.) immunization is a promising route of vaccine administration. Suitable i.d. adjuvants are important to increase vaccine efficacy in poorly responding populations such as the elderly or for dose-sparing strategies in the face of vaccine shortages. Bacterial exotoxins, such as Escherichia coli heat-labile enterotoxin (LT), exert strong immunostimulatory effects through binding to monosialoganglioside (GM1) cell surface receptors; however, injection is hampered by local inflammation. We demonstrate that the injection of LT formulations deficient in GM1 binding by mutation (LT(G33D)) or in vitro ligand coupling does not cause localized edema and inflammation in mice, yet these formulations retain potent adjuvant activity by enhancing functional Ab and cellular immune responses to coadministered Ags. Complete protection against in vivo lethal tetanus toxin challenge and the induction of Ag-specific CTL responses capable of killing target cells in vivo indicated in vivo efficacy of the induced immune responses. LT(G33D) proved superior to standard alum adjuvant regarding the magnitude and breadth of the induced immune responses. Immunizations in complex ganglioside knockout mice revealed a GM1-independent pathway of LT adjuvanticity. Immunostimulation by i.d. LT(G33D) is explained by its ability to induce migration of activated APCs to the proximal draining lymph nodes. LT(G33D) is a promising candidate adjuvant for human trials of parenteral vaccines in general and for current i.d. vaccine development in particular.     

Identification of a cell surface component of Swiss 3T3 cells associated with an inhibition of cell division

Polyclonal IgG, prepared to a purified bovine cell surface sialoglycopeptide (SGP) inhibitor of cell division, was used to identify an antigenically related molecule on the surface of Swiss 3T3 cells. SDS-PAGE and Western analyses showed that the anti-SGP antibody was monospecific and primarily recognized a 66-kDA protein of 3T3 cell membranes. Treatment of intact 3T3 cells or 3T3 cell membranes with either broad and phosphatidylinositol-specific phospholipase C enzymes suggested that the antigenic material most likely existed as an integral membrane molecule, or associated as a multimeric complex, and was not anchored at the cell surface by a phospholipid. The addition of anti-SGP IgG to 3T3 cell monolayer cultures was shown to promote cell division, suggesting a regulatory function for the membrane-associated molecule.     

Both the p33 and p55 Subunits of the Helicobacter pylori VacA Toxin Are Targeted to Mammalian Mitochondria

Helicobacter pylori infection causes peptic ulcers and gastric cancer. A major toxin secreted by H. pylori is the bipartite vacuolating cytotoxin A, VacA. The toxin is believed to enter host cells as two subunits: the p55 subunit (55 kDa) and the p33 subunit (33 kDa). At the biochemical level, it has been shown that VacA forms through the assembly of large multimeric pores composed of both the p33 subunit and the p55 subunit in biological membranes. One of the major target organelles of VacA is the mitochondria. Since only the p33 subunit has been reported to be translocated into mitochondria and the p55 subunit is not imported, it has been contentious as to whether VacA assembles into pores in a mitochondrial membrane. Here we show the p55 protein is imported into the mitochondria along with the p33 protein subunit. The p33 subunit integrally associates with the mitochondrial inner membrane, and both the p33 subunit and the p55 subunit are exposed to the mitochondrial intermembrane space. Their colocalization suggests that they could reassemble and form a pore in the inner mitochondrial membrane.     

Characterization of the receptor for tumor necrosis factor (TNF) and lymphotoxin (LT) on human T lymphocytes. TNF and LT differ in their receptor binding properties and the induction of MHC class I proteins on a human CD4+ T cell hybridoma

TNF-alpha and lymphotoxin (LT or TNF-beta) are structurally related cytokines that share several proinflammatory and immunomodulatory activities. The shared biologic activities of TNF and LT have been attributed to their binding to a common cell surface receptor(s). We observed that rTNF enhanced the expression of MHC class I proteins on the human T cell hybridoma, II-23.D7, however LT was largely unable to regulate MHC expression. To determine the molecular basis of this disparity between LT and TNF the receptor binding characteristics of rTNF and rLT were investigated by direct and competitive radioligand assays on the II-23.D7 T hybridoma, and for comparison, anti-CD3 activated human T lymphocytes. Specific 125I-rTNF binding to the II-23.D7 line revealed a single class of sites with a Kd = 175 pM and 3000 sites/cell; anti-CD3 activated T cells exhibited specific TNF binding with similar properties. The relationship of receptor occupancy to the induction of MHC class I Ag yielded a hyperbolic curve indicating a complex relationship between rTNF binding and biologic response. LT appeared to function like a partial agonist in that rLT was 10- to 20-fold less effective than rTNF in competitively inhibiting 125I-rTNF binding on the II-23.D7 line. Scatchard type analysis revealed a single class of low affinity binding sites for 125I-rLT. No differences in the competitive binding activity of rTNF and rLT were observed on the anti-CD3-activated T cells. Receptors for rTNF and rLT were immunoprecipitated from the II-23.D7 and activated T cells with anticytokine antibodies after cross-linking of radioiodinated rTNF or rLT to intact cells by using chemical cross-linking reagents. Analysis of the cross-linked adducts by SDS-PAGE and autoradiography indicated a major adduct of 92 kDa for rTNF and 104 kDa for rLT. Enzymatic digestion with neuraminidase or V8 protease revealed a unique structure to these adducts consistent with the cross-linking of a single chain of cytokine to a cell surface glycoprotein. rTNF inhibited the formation of the 104-kDa adduct formed with 125I-rLT on the II-23.D7 line, indicating these two cytokines bind to the same receptor of approximately 80 kDa. These results suggest that the disparate activities of LT and TNF to induce MHC class I proteins on the II-23.D7 cells are, in part, associated with a modified state of a common receptor.     

Cloning and expression of a novel MAPKK-like protein kinase, lymphokine-activated killer T-cell-originated protein kinase, specifically expressed in the testis and activated lymphoid cells

A novel protein kinase, TOPK (T-LAK cell-originated protein kinase), was isolated from a lymphokine-activated killer T (T-LAK) cell subtraction cDNA fragment library. The open reading frame of the TOPK gene encodes a protein of 322 amino acids, possessing a protein kinase domain profile. The cap site analysis of the 5'-end of TOPK mRNA revealed two forms, a major full-length form and a minor spliced form at the 5'-site, both encoding the same protein. A BLAST homology search and phylogenetic analysis indicated that TOPK is related to dual specific mitogen-activated protein kinase kinase (MAPKK). The transfection of the TOPK gene to COS-7 cells up-regulated a phosphorylation of p38 MAPK but not ERK1/2 or SAPK/JNK. Gel precipitation study indicated that TOPK protein can be associated with p38 in vitro. Tissue distribution of TOPK mRNA expression was specific for the testis, T-LAK cells, activated lymphoid cells, and lymphoid tumors. On the other hand, deactivated T-LAK cells did not show TOPK mRNA expression. These data suggest that TOPK is a newly identified member of a novel MEK3/6-related MAPKK that may be enrolled in the activation of lymphoid cells and support testicular functions.     

beta-Chemokine production in CD40L-stimulated monocyte-derived macrophages requires activation of MAPK signaling pathways

CD40 ligand is a cell surface molecule on CD4(+) T cells that interacts with its receptor, CD40, on antigen presenting cells to mediate humoral and cellular immune responses. Our previous studies demonstrated that a trimeric soluble form of CD40L (CD40LT) activates macrophages to produce beta-chemokines and decrease CCR5 and CD4 cell surface expression, thus inducing resistance to HIV-1 infection. However, the mechanism(s) by which CD40LT mediates these effects in primary macrophages remains unclear. In this report, we demonstrate that CD40LT induces synthesis of beta-chemokines through the activation of MAPK signaling pathways. Treatment of macrophages with CD40LT results in a rapid activation of p38 and ERK1/2 mitogen-activated protein kinases. Inhibitors of these MAPKs blocked beta-chemokine production, while protein kinase A and C inhibitors had little or no effect. We also provide evidence that CD40LT stimulates beta-chemokine production directly, as well as indirectly via a TNF-alpha-dependent mechanism. At the early time points, CD40LT directly stimulated beta-chemokine production, whereas at later time points the effect was mediated to some extent by TNF-alpha. In conclusion, our results suggest that CD40-CD40L interactions are important for the activation of monocyte-derived macrophage antiviral response affecting both viral replication and the recruitment of immune cells.     

Fusion of a cleavable signal peptide to the ectodomain of neutral endopeptidase (EC 3.4.24.11) results in the secretion of an active enzyme in COS-1 cells

Neutral endopeptidase (EC 3.4.24.11) is an integral membrane protein found at the plasma membrane of many cell types and is especially abundant at the apical "brush border" membrane of the kidney proximal tubules. The enzyme consists of a short amino-terminal cytosolic domain of 27 amino acids, a single hydrophobic sequence which is believed to be responsible for anchoring the enzyme in the plasma membrane, and a large extracellular domain containing the active site. This model is consistent with the proposed function of neutral endopeptidase, which is believed to play an important role in the inactivation of small regulatory peptides at the cell surface. Site-directed mutagenesis has allowed the identification of 1 glutamic acid and 2 histidine residues essential for catalysis. All are located near the COOH terminus of the protein. We do not know, however, whether other segments of the protein are involved in the structure of the active site. The exact role of the cytosolic and transmembrane domains is also unknown. In this report, we have induced the secretion of a soluble form of recombinant neutral endopeptidase in COS-1 cells by fusing in-frame, the cDNA encoding the signal peptide of a secreted protein (pro-opiomelanocortin) to the cDNA sequences of the complete ectodomain of neutral endopeptidase. Characterization of the secreted recombinant protein indicated that it has the same catalytic properties as the membrane-bound recombinant enzyme or as the enzyme extracted from kidney brush border membranes. Thus the extracellular domain alone is sufficient for conferring full catalytic activity to neutral endopeptidase.     

Expression of the functional soluble form of human fas ligand in activated lymphocytes.

Fas is a type I membrane protein which mediates apoptosis. Fas ligand (FasL) is a 40 kDa type II membrane protein expressed in cytotoxic T cells upon activation that belongs to the tumor necrosis factor (TNF) family. Here, we found abundant cytotoxic activity against Fas-expressing cells in the supernatant of COS cells transfected with human FasL cDNA but not with murine FasL cDNA. Using a specific polyclonal antibody against a peptide in the extracellular region of human FasL, a protein of 26 kDa was detected in the supernatant of the COS cells. The signal sequence of granulocyte colony-stimulating factor was attached to the extracellular region of human FasL. COS cells transfected with the cDNA coding for the chimeric protein efficiently secreted the active soluble form of human FasL (sFasL). Chemical crosslinking and gel filtration analysis suggested that human sFasL exists as a trimer. Human peripheral T cells activated with phorbol myristic acetate and ionomycin also produced functional sFasL, suggesting that human sFasL works as a pathological agent in systemic tissue injury.     

Biosynthesis of a phosphatidylinositol-glycan-linked membrane protein: signals for posttranslational processing of the Ly-6E antigen.

The Ly-6E/A protein is a murine cell surface protein expressed at high levels on activated peripheral T cells. The only linkage known to be responsible for its association with the plasma membrane is a phosphatidylinositol-glycan (PI-G) moiety. To examine the biosynthesis of this structure, we constructed a series of mutants of Ly-6E that were expressed in COS cells by using transient-transfection procedures. When 12 or 20 carboxy-terminal residues were deleted from the primary translation product, the PI-G modification was completely abolished and the mutant proteins became secreted. Addition of the PI-G tail was partially inhibited when the charged 12-amino-acid peptide found as a cytoplasmic tail on the transmembrane form of LFA-3 was added to the COOH terminus of the Ly-6E protein. Proteolytic cleavage occurred on this mutant protein, but the PI-G moiety was added to only 50% of the molecules. Changing an Asn residue to a Lys at the hypothetical cleavage site resulted in a PI-G-linked protein having a detectable alteration in electrophoretic mobility. This finding raises the possibility that proteolytic cleavage at other amino acid sites may occur and that PI-G attachment can occur at this new site. A model identifying two regions that may act as necessary signals for the biosynthesis of the PI-G tail is presented.