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

Expression of epithelial alkaline phosphatase in segmentally iterated bands during grasshopper limb morphogenesis

Although the study of rostral-caudal segmentation of the insect body has been a rich source of information about embryonic pattern formation, relatively little is known of the process of proximal-distal segmentation of insect appendages. Here we demonstrate that during the period of limb segmentation, five segmentally iterated, sharply demarcated bands of cell surface alkaline phosphatase activity are expressed in embryonic grasshopper limbs. These bands span each intersegmental boundary in the limb as well as one boundary within the tarsus. Within appendages, expression is restricted to epithelial cells, where activity is present on both apical and basolateral surfaces. This epithelial alkaline phosphatase remains active at neutral pH, is insensitive to levamisole inhibition, and is strongly inhibited by nucleoside monophosphates. Treatment of embryos with phosphatidylinositol-specific phospholipase C releases almost all visible chromogenic activity, indicating that the epithelial alkaline phosphatase is anchored to the plasma membrane by glycosyl-phosphatidylinositol. When material released by phosphatidylinositol-specific phospholipase C is separated on native polyacrylamide gels, a single broad band of enzymatic activity is detected following incubation with substrate. A polyclonal antiserum raised against a 55 × 10(3) M(r) alkaline phosphatase from shrimp recognizes a single band of 56 × 10(3) M(r) on immunoblots of grasshopper membrane proteins. The spatially restricted expression of epithelial alkaline phosphatase suggests that it may be involved in epithelial cell rearrangements or shape changes associated with limb segmentation and morphogenesis. It also may contribute to definition of axon routes in the limb, since pioneer afferent growth cones turn at, and migrate along, the edge of one alkaline phosphatase-expressing epithelial domain.

Deficiency of the GPI anchor caused by a somatic mutation of the PIG-A gene in paroxysmal nocturnal hemoglobinuria

Paroxysmal nocturnal hemoglobinuria is an acquired hematopoietic disease characterized by abnormal blood cell populations in which the biosynthesis of the glycosylphosphatidylinositol (GPI) anchor is deficient. Deficiency of surface expressions of GPI-anchored complement inhibitors leads to complement-mediated hemolysis. Here we report that PIG-A, which participates in the early step of GPI anchor biosynthesis, is the gene responsible for paroxysmal nocturnal hemoglobinuria. Affected granulocytes and B lymphocytes had the same somatic mutation of PIG-A, indicating their clonal origin from a multipotential hematopoietic stem cell. We localized PIG-A to the X chromosome, which accounts for expression of the recessive phenotype of the somatic mutation and the fact that the same one of the multiple biosynthetic steps is affected in all patients so far characterized.

Actin is a surface component of calf pulmonary artery endothelial cells in culture

An angiogenin binding protein isolated previously from endothelial cells has been shown to be a member of the actin family. Calf pulmonary artery endothelial (CPAE) cells were investigated for the presence of surface actin by immunoblotting of isolated surface proteins and by immunofluorescence. CPAE cell surface proteins were isolated by selective apical biotinylation and recovery of biotinylated proteins by avidin affinity chromatography. Immunoblotting with a specific smooth muscle alpha-actin antibody detected the presence of this type of actin among the isolated cell surface proteins. Immunofluorescence confirmed that smooth muscle alpha-actin is localized at the surface of nonpermeabilized CPAE cells. Exposure of CPAE cells to angiogenin prior to cell surface immunostaining diminished the signal. When CPAE and rat aortic smooth muscle cells were made permeable before staining, stress fibers could be recognized by the antibody in smooth muscle cells but not CPAE cells. The results indicate that a smooth muscle type of alpha-actin is localized specifically on the surface of cultured CPAE cells where it might interact with angiogenin and other actin binding proteins present in the extracellular environment.

Polarized distribution of the viral glycoproteins of vesicular stomatitis, fowl plague and Semliki Forest viruses in hippocampal neurons in culture: a light and electron microscopy study

We have shown previously using immunofluorescence microscopy that upon infection of polarized hippocampal cells in culture with vesicular stomatitis virus (VSV) and fowl plague virus (FPV) the VSV glycoprotein is delivered to the plasma membrane of the dendrites and of the cell body whereas the FPV hemagglutinin is transported to the axonal surface (Cell, 62 (1990) 63-72). In this work electron microscopy of infected rat hippocampal neurons showed that VSV progeny budded from the plasma membrane of the dendrites and the cell body. The location of the budding virions corresponded to the distribution of the VSV glycoprotein which was detected over the somatodendritic plasma membrane by immunoelectron microscopy. In contrast, no FPV formation was seen in the infected neurons although the FPV hemagglutinin was localized to the axonal surface by immunoelectron microscopy. In Semliki Forest virus (SFV) infected hippocampal cells we observed that the viral glycoproteins were exclusively present in the dendrites and cell body but not in axons.

GPI-anchored proteins associate to form microdomains during their intracellular transport in Caco-2 cells

In this study, we have investigated the possibility that glycosyl-phosphatidylinositol (GPI)-anchored proteins form insoluble membrane complexes in Caco-2 cells and that transmembrane proteins are associated with these complexes. GPI-anchored proteins were mainly resistant to Triton X-100 (TX-100) extraction at 4 degrees C but fully soluble in n-octyl-glucoside. Resistance to Triton X-100 extraction was not observed in the endoplasmic reticulum but appeared during transport through the Golgi complex. It was not dependent upon N-glycosylation processing, or pH variation from 6.5 to 8.5, and was not affected by sterol-binding agents. Other apical or basolateral transmembrane proteins were well solubilized in TX-100, with the exception of sucrase-isomaltase, which was partly insoluble. We isolated a membrane fraction from Caco-2 cells that contained GPI-anchored proteins and sucrase-isomaltase but no antigen 525, a basolateral marker, or dipeptidylpeptidase IV, an apical one. These data suggest that GPI-anchored proteins cluster to form membrane microdomains together with an apical transmembrane protein, providing a possible apical sorting mechanism for intestinal cells in vitro that might be related to apical sorting in MDCK cells, and that other mechanisms might exist to sort proteins to the apical membrane.

Are the glypiated adhesion molecules preferentially targeted to the axonal compartment?

The question of how the cell surface molecules may be specifically delivered to subdomains of neurons is of particular interest considering that polarized sorting to the axon could enable adhesion glycoproteins to induce fasciculation of axonal tracts, guidance to the target cell, and the establishment of synaptic contacts. It was recently proposed that GPI-anchored molecules undergo preferential delivery to the axonal surface, implicating a similar polarized sorting of glycoproteins in neurons and epithelial cells (Dotti and Simons, 1990; Dotti et al., 1991). This review focuses on the cellular and subcellular localization of several glypiated adhesion molecules (Thy-1, TAG-1, F3/F11, P-31) in the developing and adult cerebellar cortex of the mouse. We conclude that the cellular distribution of GPI-anchored adhesion molecules within neurons is very complex and depends on: 1. The neuronal cell types, for example, F3/F11 is localized in axons in granule cells but is present in all compartments of Golgi cells. 2. The molecule itself: Thy-1, TAG-1, and P-31 are present on the granule cell body, whereas at the same developmental stage, F3/F11 is restricted to the axon. 3. The differentiation state: Thy-1 delivery to the axon correlates with postsynaptic target maturation.

Correctly sorted molecules of a GPI-anchored protein are clustered and immobile when they arrive at the apical surface of MDCK cells

Glycosyl-phosphatidylinositol (GPI)-anchored proteins are sorted to the apical surface of many epithelial cell types. To better understand the mechanism for apical segregation of these proteins, we analyzed the lateral mobility and molecular associations of a model GPI-anchored protein, herpes simplex virus gD1 fused to human decay accelerating factor (gD1-DAF) (Lisanti, M. P., I. W. Caras, M. A. Davitz, and E. Rodriguez-Boulan. 1989. J. Cell Biol. 109:2145-2156) shortly after arrival and after long-term residence at the surface of confluent, polarized MDCK cells. FRAP measurements of lateral diffusion showed that the mobile fraction of newly arrived gD1-DAF molecules was much less than the mobile fraction of long-term resident molecules (40 vs. 80-90%). Fluorescence resonance energy transfer measurements showed that the newly arrived molecules were clustered, while resident molecules were not. Newly delivered gD1-DAF molecules were clustered but not immobilized in mutant, Concanavalin A-resistant MDCK cells that failed to sort gD1-DAF. Our results indicate that GPI-anchored proteins in MDCK cells are clustered before delivery to the surface. However, clustering alone does not target molecules for apical delivery. The immobilization observed when gD1-DAF is correctly sorted suggests that the clusters must associate some component of the cell's cytoplasm.

Virus infection of polarized epithelial cells

This chapter focuses on the interaction of viruses with epithelial cells. The role of specific pathways of virus entry and release in the pathogenesis of viral infection is examined together with the mechanisms utilized by viruses to circumvent the epithelial barrier. Polarized epithelial cells in culture, which can be grown on permeable supports, provide excellent systems for investigating the events in virus entry and release at the cellular level, and much information is being obtained using such systems. Much remains to be learned about the precise routes by which many viruses traverse the epithelial barrier to initiate their natural infection processes, although important information has been obtained in some systems. Another area of great interest for future investigation is the process of virus entry and release from other polarized cell types, including neuronal cells.

Regulation of cell surface polarity from bacteria to mammals

The generation of unique domains on the cell, cell surface polarity, is critical for differentiation into the diversity of cell structures and functions found in a wide variety of organisms and cells, including the bacterium Caulobacter crescentus, the budding yeast Saccharomyces cerevisiae, and mammalian polarized epithelial cells. Comparison of the mechanisms for establishing polarity in these cells indicates that restricted membrane protein distributions are generated by selective protein targeting to, and selective protein retention at, the cell surface. Initiation of these mechanisms involves reorientation of components of the cytoskeleton and protein transport pathways toward restricted sites at the cell surface and formation of a targeting patch at those sites for selective recruitment and retention of proteins.

New insights into placental iron transport

The transport of iron from mother to fetus presents a number of challenges, none more perplexing than determining how the flow can be unidirectional, yet avoid mixing the mother's serum proteins with the fetal system. Studies with cultured cell lines suggest a possible mechanism by which this may be achieved.

Decay-accelerating factor (DAF, CD55) in normal colorectal mucosa, adenomas and carcinomas

Decay-accelerating-factor (DAF, CD55), a phosphatidyl-inositol anchored glycoprotein, is a member of the cell membrane bound complement regulatory proteins that inhibit autologous complement cascade activation. DAF was found expressed on cells that are in close contact with serum complement proteins, but also on cells outside the vascular space and on tumour cells. Using CD55(BRIC110) and CD55(143-30) we show here that DAF(CD55) is only sporadically expressed on the luminal surface of normal colonic epithelium. However, 5/20 adenomas expressed DAF(CD55) on the cell surface of all tumour cells, 5/20 adenomas were completely negative, 10/20 adenomas expressed DAF(CD55) in various amounts. DAF(CD55) was expressed in various intensities on almost all tumour cells of the colon carcinoma cell line HT29. In 5/88 colorectal carcinomas DAF(CD55) was localised on the apical cell surface of all tumour cells, 31/88 were completely negative, 52/88 expressed DAF(CD55) in parts of their neoplastic populations. There was no correlation between the tumour grading, staging and location and the mode of DAF(CD55) expression, but DAF(CD55) was found more often in mucinous carcinomas (P = 0.007). Although the mode of DAF(CD55) expression is not correlated with tumour prognostic parameters, the upregulation of DAF(CD55) in a subset of adenomas and carcinomas needs further investigation concerning protection of tumour cells against complement cytotoxicity.

The establishment of polarized membrane traffic in Xenopus laevis embryos

Delineation of apical and basolateral membrane domains is a critical step in the epithelialization of the outer layer of cells in the embryo. We have examined the initiation of polarized membrane traffic in Xenopus and show that membrane traffic is not polarized in oocytes but polarized membrane domains appear at first cleavage. The following proteins encoded by injected RNA transcripts were used as markers to monitor membrane traffic: (a) VSV G, a transmembrane glycoprotein preferentially inserted into the basolateral surface of polarized epithelial cells; (b) GThy-1, a fusion protein of VSV G and Thy-1 that is localized to the apical domains of polarized epithelial cells; and (c) prolactin, a peptide hormone that is not polarly secreted. In immature oocytes, there is no polarity in the expression of VSV G or GThy-1, as shown by the constitutive expression of both proteins at the surface in the animal and vegetal hemispheres. At meiotic maturation, membrane traffic to the surface is blocked; the plasma membrane no longer accepts the vesicles synthesized by the oocyte (Leaf, D. L., S. J. Roberts, J. C. Gerhart, and H.-P. Moore. 1990. Dev. Biol. 141:1-12). When RNA transcripts are injected after fertilization, VSV G is expressed only in the internal cleavage membranes (basolateral orientation) and is excluded from the outer surface (apical orientation, original oocyte membrane). In contrast, GThy-1 and prolactin, when expressed in embryos, are inserted or released at both the outer membrane derived from the oocyte and the inner cleavage membranes. Furthermore, not all of the cleavage membrane comes from an embryonic pool of vesicles--some of the cleavage membrane comes from vesicles synthesized during oogenesis. Using prolactin as a marker, we found that a subset of vesicles synthesized during oogenesis was only released after fertilization. However, while embryonic prolactin was secreted from both apical and basolateral surfaces, the secretion of oogenic prolactin was polarized. Oogenic prolactin was secreted only into the blastocoel (from the cleavage membrane), none could be detected in the external medium (from the original oocyte membrane). These results provide the first direct evidence that the oocyte synthesizes a cache of vesicles for specific recruitment to the embryonic cleavage membranes which are polarized beginning with the first cleavage division.

The apical sorting signal on human aminopeptidase N is not located in the stalk but in the catalytic head group

Human aminopeptidase N carries an apical sorting signal on its ectodomain necessary for its correct transport to the apical membrane in Madin-Darby canine kidney cells. To determine whether the apical sorting signal is localized in the serine/threonine rich stalk or in the catalytic head group, anchor/stalk-minus aminopeptidase N, consisting of the hemagglutinin signal peptide and the catalytic head group of human aminopeptidase N, was expressed in MDCK cells. Anchor/stalk-minus aminopeptidase N was secreted mainly to the apical side. The catalytic head group of human aminopeptidase N thus carries an apical sorting signal.

Expression of the influenza A virus M2 protein is restricted to apical surfaces of polarized epithelial cells

The M2 protein of influenza A virus is a small, nonglycosylated transmembrane protein that is expressed on surfaces of virus-infected cells. A monoclonal antibody specific for the M2 protein was used to investigate its expression in polarized epithelial cells infected with influenza virus or a recombinant vaccinia virus that expresses M2. The expression of M2 on the surfaces of influenza virus-infected cells was found to be restricted to the apical surface, closely paralleling that of the influenza virus hemagglutinin (HA). Membrane domain-specific immunoprecipitation indicated that the M2 protein was inserted directly into the apical membrane with transport kinetics similar to those of HA. In polarized cells infected with a recombinant vaccinia virus that expresses M2, we found that 86 to 93% of surface M2 was restricted to the apical domain compared with 88 to 90% of HA in a similar assay. These results indicate that the M2 protein undergoes directional transport in the absence of other influenza virus proteins and that M2 contains the structural features required for apical transport in polarized epithelial cells. The ultrastructural localization of the M2 protein in influenza virus-infected MDCK cells was investigated by immunoelectron microscopy using M2 antibody and a gold conjugate. In cells in which extensive virus budding was occurring, the apical cell membrane was labeled with gold particles evenly distributed between microvilli and the surrounding membrane. In addition, a significant fraction of the M2 label was apparently associated with virions. A monoclonal antibody specific for HA demonstrated a similar labeling pattern. These results indicate that M2 is localized in close proximity to budding and assembled virions.

Carbonic anhydrase IV on brain capillary endothelial cells: a marker associated with the blood-brain barrier

Carbonic anhydrase (CA) activity plays an important role in controlling cerebrospinal fluid production and also influences neuroexcitation and susceptibility to seizures. Until recently, CA II was the only CA demonstrated in brain. Its distribution is limited to the epithelial cells of the choroid plexus and to the myelin-forming cells, the oligodendrocytes. In this report, we present immunoblots, using an antibody raised to CA IV from rat lung, that show that CA IV is also present in rat and mouse brain. Results of immunohistochemistry and immunoelectron microscopy on sections from rat and mouse brain are presented that show the distribution of CA IV to be quite distinct from that of CA II. CA IV is expressed on and is limited to the luminal surface of endothelial cells of cerebral capillaries. These results establish CA IV as a cytochemical marker associated with the blood-brain barrier and suggest an important role for CA IV in CO2 and HCO3- homeostasis in brain.

W(h)ither default? Sorting and polarization in epithelial cells

The discovery of basolateral sorting signals in the past year may leave the default pathway with nowhere to go. With new results suggesting that even more GTP-binding proteins and coatamers might be involved in transport and targeting, it is clear that the age of mapmaking in polarization research is nearly over.

A novel strategy for secreting proteins: use of phosphatidylinositol-glycan-specific phospholipase D to release chimeric phosphatidylinositol-glycan anchored proteins

Phosphatidylinositol-glycan-specific phospholipase D (PI-G PLD) specifically hydrolyzes the inositol-phosphate linkage in phosphatidylinositol-glycan (PI-G) anchored proteins. We recently deduced the primary structure of this enzyme and demonstrated specific enzymatic activity in transfected cells. Co-transfection of PI-G PLD with a natural PI-G anchored protein resulted in the secretion of the PI-G anchored protein via a PI-G PLD specific mechanism. We have taken advantage of these observations to develop an alternative system that may be useful for expressing and secreting proteins not amenable to secretion by conventional methods. Chimeric PI-G anchored proteins were constructed by transferring the COOH-terminal signal peptide for PI-G anchor attachment from placental alkaline phosphatase or from the low affinity IgG receptor, FcGRIIIB, to proteins that are not normally PI-G anchored. This process facilitates the cell surface expression of several proteins including the high affinity IgE receptor alpha subunit, FcERI alpha, which otherwise requires at least one other subunit for surface expression. Co-expression of these chimeric PI-G anchored proteins with PI-G PLD resulted in their secretion via a PI-G PLD specific mechanism.

Phosphatidylinositol-glycan-specific phospholipase D is an amphiphilic glycoprotein that in serum is associated with high-density lipoproteins

Phosphatidylinositol (PtdIns)-glycan-specific phospholipase D was purified from bovine and human serum by phase separation in Triton X-114 and by chromatography on DEAE-cellulose, octyl-Sepharose, concanavalin-A-Sepharose, and hydroxyapatite. The purification of the two enzymes was approximately 1200-fold with a recovery of 3-5%. Bovine serum contained about 40 micrograms/ml of PtdIns-glycan-specific phospholipase D, about 10 times more than the amount determined in human serum. PtdIns-glycan-specific phospholipase D is also present in mammalian cerebrospinal fluid and in mammalian milk but to a much lesser extent than in serum. Enzyme from bovine and human serum displayed amphiphilic properties as revealed by sucrose density gradient centrifugation and gel filtration in the absence and presence of detergent. On density gradient centrifugation, both enzymes sedimented with an apparent sedimentation coefficient of about 6.0 S in the presence of 0.1% Triton X-100, and formed aggregates up to 14.5 S in the absence of detergent. Upon gel filtration, the bovine and human enzymes migrated with a Stokes' radius of 6.5 nm and 6.6 nm, respectively, in the presence of Triton X-100. In the absence of Triton X-100, both enzymes gave a Stokes' radius of 8.8 nm. Serial centrifugation of serum at increasing NaBr concentrations revealed that the majority of the enzyme is contained in the high-density lipoprotein fraction. PtdIns-glycan-specific phospholipase D from bovine and human serum contained 27 and 28 N-acetylglucosamine residues, respectively. Treatment with N-glycosidase F decreased the apparent molecular mass of the bovine and human enzyme from 115 and 123 kDa to 91 and 87 kDa, respectively. Sequence analysis of peptides derived from PtdIns-glycan-specific phospholipase D of bovine serum by CNBr cleavage gave 100% identity to the sequence published for the bovine liver enzyme while there was 83% similarity and 74% identity to the sequence of peptides obtained from the human serum enzyme.

Expression of unique sets of GPI-linked proteins by different primary neurons in vitro

We have surveyed the proteins expressed at the surface of different primary neurons as a first step in elucidating how axons regulate their ensheathment by glial cells. We characterized the surface proteins of dorsal root ganglion neurons, superior cervical ganglion neurons, and cerebellar granule cells which are myelinated, ensheathed but unmyelinated, and unensheathed, respectively. We found that the most abundant proteins are common to all three types of neurons. Reproducible differences in the composition of the integral membrane proteins (enriched by partitioning into a Triton X-114 detergent phase) were detected. These differences were most striking when the expression of glycosylphosphatidyl-inositol (GPI)-anchored membrane proteins by these different neurons was compared. Variations in the relative abundance and degree of glycosylation of several well known GPI- anchored proteins, including Thy-1, F3/F11, and the 120-kD form of the neural cell adhesion molecule (N-CAM), and an abundant 60-kD GPI-linked protein were observed. In addition, we have identified several potentially novel GPI-anchored glycoproteins on each class of neurons. These include a protein that is present only on superior cervical ganglion neurons and is 90 kD; an abundant protein of 69 kD that is essentially restricted in its expression to dorsal root ganglion neurons; and proteins of 38 and 31 kD that are expressed only on granule cell neurons. Finally, the relative abundance of the three major isoforms of N-CAM was found to vary significantly between these different primary neurons. These results are the first demonstration that nerve fibers with diverse ensheathment fates differ significantly in the composition of their surface proteins and suggest an important role for GPI-anchored proteins in generating diversity of the neuronal cell surface.

Partial release of aminopeptidase N from larval midgut cell membranes of the silkworm, Bombyx mori, by phosphatidylinositol-specific phospholipase C

1. The membrane anchor of aminopeptidase N associated with larval midgut cell membranes of the silkworm, Bombyx mori, was investigated by using phosphatidylinositol-specific phospholipase C (PIPLC) and proteases. 2. Aminopeptidase N, which was virtually all localized in the brush border membrane, was solubilized by PIPLC but not by papain or trypsin. 3. Detergent-solubilized amphiphilic aminopeptidase N was converted into a hydrophilic form by PIPLC but not by papain. 4. Either of these effects of PIPLC on aminopeptidase N was maximally 40%. 5. These results suggest that in larval midgut cells of the silkworm, B. mori, at least 40% aminopeptidase N is anchored in the brush border membrane via glycosyl-phosphatidylinositol.

The post-synthetic sorting of endogenous membrane proteins examined by the simultaneous purification of apical and basolateral plasma membrane fractions from Caco-2 cells

A subcellular fractionation method to isolate simultaneously apical and basolateral plasma membrane fractions from the human adenocarcinoma cell line Caco-2, grown on filter supports, is described. The method employs sucrose-density-gradient centrifugation and differential precipitation. The apical membrane fraction was enriched 14-fold in sucrase-isomaltase and 21-fold in 5'-nucleotidase compared with the homogenate. The basolateral membrane fraction was enriched 20-fold relative to the homogenate in K(+)-stimulated p-nitrophenylphosphatase. Alkaline phosphatase was enriched 15-fold in the apical membrane fraction and 3-fold in the basolateral membrane fraction. Analytical density-gradient centrifugation showed that this enzyme was a true constituent of both fractions, and experiments measuring alkaline phosphatase release following treatment with phosphatidylinositol-specific phospholipase C showed that in both membrane fractions the enzyme was glycosyl-phosphatidylinositol-linked. There was very little contamination of either membrane fraction by marker enzymes of the Golgi complex, mitochondria or lysosomes. Both membrane fractions were greater than 10-fold purified with respect to the endoplasmic reticulum marker enzyme alpha-glucosidase. Protein composition analysis of purified plasma membrane fractions together with domain-specific cell surface biotinylation experiments revealed the presence of both common and unique integral membrane proteins in each plasma membrane domain. The post-synthetic transport of endogenous integral plasma membrane proteins was examined using the devised subcellular fractionation procedure in conjunction with pulse-chase labelling experiments and immunoprecipitation. Five common integral membrane proteins immunoprecipitated by an antiserum raised against a detergent extract of the apical plasma membrane fraction were delivered with the same time course to each cell-surface domain.

The 17-residue transmembrane domain of beta-galactoside alpha 2,6-sialyltransferase is sufficient for Golgi retention

beta-Galactoside alpha 2,6-sialyltransferase (ST) is a type II integral membrane protein of the Golgi apparatus involved in the sialylation of N-linked glycans. A series of experiments has shown that the 17-residue transmembrane domain of ST is sufficient to confer localization to the Golgi apparatus when transferred to the corresponding region of a cell surface type II integral membrane protein. Lectin affinity chromatography of chimeric proteins bearing this 17-residue sequence suggests that these chimeric proteins are localized in the trans-Golgi cisternae and/or trans-Golgi network. Further experiments suggest that this 17-residue sequence functions as a retention signal for the Golgi apparatus.

The complement system in human reproduction

Regulation of the complement system in reproduction is unique inasmuch as reproductive tissues represent the only condition where allogeneic interactions occur naturally. Both allogeneic extraembryonic membranes and semen that contact and interact with maternal cells and tissues must avert complement-mediated damage to ensure reproductive success. Several regulators of complement activation exist. Membrane cofactor protein (MCP) and decay accelerating factor (DAF) inactivate C3 and C5 convertases on cell surfaces. In addition, CD59 inhibits the membrane attack complex (MAC) of the complement cascade. Strong expression of these membrane glycoproteins by trophoblast and amniotic epithelium has been observed. MCP, DAF, and CD59 likely safeguard extraembryonic tissues from complement damage originating from maternal and fetal blood or amniotic fluid. Different reproductive tract fluids vary in complement levels. With the exception of ovarian follicular fluid, these levels are generally much less than those in blood. Endometrial and cervical content of C3 appear to be regulated by hormones. These observations suggest that the effects of complement activation may vary in reproductive tissues. MCP is absent from the surfaces of oocytes. Sperm express MCP and DAF in discrete areas that would not be associated with the known complement-regulatory functions of these proteins. Seminal plasma contains MCP and the MAC inhibitor SP-40,40 but not DAF.SP-40,40 may exemplify how complement-regulatory proteins perform alternative functions as it interacts with molecules other than complement components. We have reviewed aspects of the complement system that relate to allogeneic interactions in reproduction and that suggest fruitful areas for further research.

GP-2/THP gene family encodes self-binding glycosylphosphatidylinositol-anchored proteins in apical secretory compartments of pancreas and kidney

A family of homologous genes is shown to encode GP-2, the major glycosylphosphatidylinositol (GPI)-linked glycoprotein of pancreatic zymogen granule membranes, and Tamm-Horsfall protein (THP), a GPI-linked glycoprotein associated with apical vesicles in kidney thick ascending limb of Henle (TALH) cells. The C-terminal regions of GP-2 (Asp54-Phe530) and THP (Asp175-His644) from rat show 53% identity, 86% similarity, and 26 conserved cysteine residues including one epidermal growth factor motif. The unique N-terminal domain of rat THP (unique-THP, Pro29-Gln174) shows four conserved epidermal growth factor motifs, three in tandem and one in reverse orientation. GP-2 homologues are observed in a wide variety of epithelial cells, several of which contain highly regulated secretory processes. GP-2 released from zymogen granule membranes with phosphatidylinositol phospholipase C reacts with anti-cross-reactive determinant antibody (anti-CRD), confirming the GPI nature of the pancreatic homologue. In contrast, GP-2 and THP, released endogenously from pancreas and kidney, respectively, do not react with anti-cross-reactive determinant antibody, suggesting alternative enzymatic mechanisms for their physiological release. Globular domains of GP-2 and THP, but not albumin, show pH- and ion-dependent self-association in vitro. The GP-2/THP family appears to represent a newly discovered class of GPI-anchored proteins, which may utilize pH- and ion-dependent self-association mechanisms for establishing membrane (micro)domains targeted to intracellular secretory compartments.

Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface

We show that a protein with a glycosylphosphatidyl inositol (GPI) anchor can be recovered from lysates of epithelial cells in a low density, detergent-insoluble form. Under these conditions, the protein is associated with detergent-resistant sheets and vesicles that contain other GPI-anchored proteins and are enriched in glycosphingolipids, but do not contain a basolateral marker protein. The protein is recovered in this complex only after it has been transported to the Golgi complex, suggesting that protein-sphingolipid microdomains form in the Golgi apparatus and plasma membrane and supporting the model proposed by Simons and colleagues for sorting of certain membrane proteins to the apical surface after intracellular association with glycosphingolipids.

Expression of the Neuronal Surface Glycoprotein Thy-1 Does Not Follow Appearance of Its mRNA in Developing Mouse Purkinje Cells

In developing rodent nervous system, although the appearance of Thy-1 mRNA, as seen by in situ hybridisation, is in general quickly followed by the appearance of immunohistochemically detectable protein, there are certain sites where a delay of several days occurs between expression of detectable message and protein. Mouse Purkinje cells exemplify this behaviour and are the dominant Thy-1-expressing cell in early postnatal cerebellum, so allowing quantitative, homogenate-based methods to be used to test whether such a lag in protein expression does occur. Measurement of Thy-1 mRNA (by slot blot) and protein (by radioimmunoassay) shows a substantial excess of Thy-1 message, compared to protein accumulating in the tissue, during the first postnatal week, which is not found in tissues (rat cerebellum, and rat or mouse cerebrum) where no lag is apparent in appearance of Thy-1 protein from the section-based methods. The species of Thy-1 mRNA produced by Purkinje cells does not appear to change during development, as assessed either in terms of its size (by northern blotting) or in the heterogeneous pattern of transcription initiation sites used (assessed by S1 nuclease protection analysis). Appearance of Thy-1 protein in these cells, therefore, seems to be regulated posttranscriptionally.

An endogenous MDCK lysosomal membrane glycoprotein is targeted basolaterally before delivery to lysosomes

Using surface immunoprecipitation at 37 degrees C to "catch" the transient apical or basolateral appearance of an endogenous MDCK lysosomal membrane glycoprotein, the AC17 antigen, we demonstrate that the bulk of newly synthesized AC17 antigen is polarly targeted from the Golgi apparatus to the basolateral plasma membrane or early endosomes and is then transported to lysosomes via the endocytic pathway. The AC17 antigen exhibits very similar properties to members of the family of lysosomal-associated membrane glycoproteins (LAMPs). Parallel studies of an avian LAMP, LEP100, transfected into MDCK cells revealed colocalization of the two proteins to lysosomes, identical biosynthetic and degradation rates, and similar low levels of steady-state expression on both the apical (0.8%) and basolateral (2.1%) membranes. After treatment of the cells with chloroquine, newly synthesized AC17 antigen, while still initially targeted basolaterally, appears stably in both the apical and basolateral domains, consistent with the depletion of the AC17 antigen from lysosomes and its recycling in a nonpolar fashion to the cell surface.

An internal deletion in the cytoplasmic tail reverses the apical localization of human NGF receptor in transfected MDCK cells

A cDNA encoding the full-length 75-kD human nerve growth factor receptor was transfected into MDCK cells and its product was found to be expressed predominantly (80%) on the apical membrane, as a result of vectorial targeting from an intracellular site. Apical hNGFR bound NGF with low affinity and internalized it inefficiently (6% of surface bound NGF per hour). Several mutant hNGFRs were analyzed, after transfection in MDCK cells, for polarized surface expression, ligand binding, and endocytosis. Deletionof juxta-membrane attachment sites for a cluster of O-linked sugars did not alter apical localization. A mutant receptor lacking the entire cytoplasmic tail (except for the five proximal amino acids) was also expressed on the apical membrane, suggesting that information for apical sorting was contained in the ectoplasmic or transmembrane domains. However, a 58 amino acid deletion in the hNGFR tail that moved a cytoplasmic tyrosine (Tyr 308) closer to the membrane into a more charged environment resulted in a basolateral distribution of the mutant receptor and reversed vectorial (basolateral) targeting. The basolateral mutant receptor also internalized 125I-NGF rapidly (90% of surface bound NGF per hour), exhibited a larger intracellular fraction and displayed a considerably shortened half-life (approximately 3 h). We suggest that hNGFR with the internal cytoplasmic deletion expresses a basolateral targeting signal, related to endocytic signals, that is dominant over apical targeting information in the ecto/transmembrane domains. These results apparently contradict a current model that postulates that basolateral targeting is a default mechanism.

A quantitative immunoelectronmicroscopic study on soluble, membrane-associated and membrane-bound lysosomal enzymes in human intestinal epithelial cells

We have used quantitative immunoelectronmicroscopy to compare the in situ localization of acid alpha-glucosidase, lysosomal acid phosphatase, beta-hexosaminidase and glucocerebrosidase in intestinal epithelial cells of the human duodenum. Differences between these four lysosomal enzymes were observed with respect to their presence at the apical cell surface. Transport to the apical membrane seems to be a more important intracellular route for lysosomal acid phosphatase and acid alpha-glucosidase than it is for beta-hexosaminidase. The membrane associated lysosomal enzyme glucocerebrosidase is not transported to the microvilli. The studies emphasize that lysosomal enzyme transport pathways are enzyme and cell type specific.

Basolateral sorting in MDCK cells requires a distinct cytoplasmic domain determinant

In MDCK cells, Golgi to basolateral transport of several membrane proteins has been found to involve a cytoplasmic domain determinant. In some cases (Fc receptor, lysosomal glycoprotein Igp120), the determinant appears similar to that required for endocytosis via clathrin-coated pits; for Igp120, elimination of a single cytoplasmic domain tyrosine both blocks internalization and results in apical transport. In other cases (LDL receptor), the determinant does not involve the cytoplasmic domain tyrosine required for endocytosis. Thus, contrary to current models, basolateral transport in MCDK cells occurs not by default but depends on one or more cytoplasmic domain determinants, the precise nature of which is unknown. For some proteins, it is closely related to coated pit determinants. The fact that many membrane proteins can reach the apical surface in the absence of this determinant suggests that signals for apical transport are widely distributed.

Targeting of transmembrane and GPI-anchored forms of N-CAM to opposite domains of a polarized epithelial cell

The calcium-independent neural cell adhesion molecule N-CAM is expressed transiently during development in many tissues, including epithelia. The three naturally occurring principal isoforms of N-CAM differ in the way in which they associate with the membrane and in their cytoplasmic domains. These isoforms are generated by developmentally regulated alternative splicing of a single gene: the large cytoplasmic domain (ld) form (relative molecular mass 180,000 (Mr 180K] is specific for post-mitotic neurons; the 120K small cytoplasmic domain (ssd) and 140K small surface domain (sd) forms also occur on other cell types. One function of the different isoforms could be to specify cellular localization; for example, glycosyl phosphatidyl inositol (GPI)-membrane anchoring acts as a targeting signal for expression on the apical surface of polarized epithelial cells. Neurons and epithelial cells may use similar mechanisms for polarizing their plasma membrane proteins. We have therefore investigated the targeting of GPI-anchored (ssd N-CAM, 120K) and transmembrane forms of N-CAM (sd N-CAM, 140K; ld N-CAM, 180K) by comparing the expression of each after transfection of the appropriate complementary DNAs into polarized epithelial cells. We find that isoforms with alternative modes of membrane association are targeted to different surfaces of polarized epithelial cells: ssd N-CAM is expressed on the apical surface, whereas sd and ld N-CAM are expressed on the basolateral surface. These results suggest that the different isoforms of N-CAM determine their own diverse cellular destinations. They also support the hypothesis that the GPI anchor acts as an apical targeting signal in epithelia.

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.

A single amino acid change in the cytoplasmic domain alters the polarized delivery of influenza virus hemagglutinin

In the polarized kidney cell line MDCK, the influenza virus hemagglutinin (HA) has been well characterized as a model for apically sorted membrane glycoproteins. Previous work from our laboratory has shown that a single amino acid change in the cytoplasmic sequence of HA converts it from a protein that is excluded from coated pits to one that is efficiently internalized. Using trypsin or antibodies to mark protein on the surface, we have shown in MDCK cells that HA containing this mutation is no longer transported to the apical surface but instead is delivered directly to the basolateral plasma membrane. We propose that a cytoplasmic feature similar to an endocytosis signal can cause exclusive basolateral delivery.

An autonomous signal for basolateral sorting in the cytoplasmic domain of the polymeric immunoglobulin receptor

The polymeric immunoglobulin receptor is normally delivered from the Golgi to the basolateral surface of epithelial cells and then transports polymeric IgA and IgM to the apical surface. We now report that a 14 residue segment of the 103 amino acid cytoplasmic domain, proximal to the plasma membrane, directs the receptor to the basolateral surface. A mutant receptor lacking these 14 amino acids is sorted directly to the apical surface from the Golgi. Furthermore, this sequence is sufficient to redirect an apical membrane protein, placental alkaline phosphatase, to the basolateral plasma membrane. We conclude that this sequence contains an autonomous signal, which specifies sorting from the Golgi to the basolateral surface, a process previously postulated to occur by default.