ARF: a key regulatory switch in membrane traffic and organelle structure.

The small GTP-binding protein ADP ribosylation factor (ARF) regulates, through a GTP cycle, the reversible binding of cytosolic coat proteins to Golgi membranes. By determining the binding and release of coat proteins from membranes, ARF controls the production and lifetime of coated-membrane structures. In the past year, studies suggesting a role for ARF in phospholipid metabolism have broadened our perspective on ARF function within the cell.

Regulation of T cell receptor expression in immature CD4+CD8+ thymocytes by p56lck tyrosine kinase: basis for differential signaling by CD4 and CD8 in immature thymocytes expressing both coreceptor molecules.

Signals transduced through the T cell antigen receptor (TCR) are modulated by the src family tyrosine kinase p56lck (lck), which associates in mature T cells with the coreceptor molecules CD4 and CD8. Here we describe a novel function of lck in immature CD4+CD8+ thymocytes, that of regulating TCR expression. Activation of lck in immature CD4+CD8+ thymocytes by intrathymic engagement of CD4 maintains low TCR expression by causing most TCR components to be retained and degraded within the endoplasmic reticulum. Importantly, activation of lck in immature CD4+CD8+ thymocytes results from engagement of surface CD4 molecules, but not surface CD8 molecules, despite the nearly fourfold greater surface expression of CD8 than CD4. The competence of CD4 to activate lck in CD4+CD8+ thymocytes relates to the fact that a relatively large fraction of surface CD4 molecules (25-50%) are associated with intracellular lck molecules, whereas only 2% of surface CD8 molecules are associated with lck. The amount of lck associated with CD4 in CD4+CD8+ thymocytes is diminished by chronic CD4 engagement in the thymus, as activated lck molecules subsequently dissociate from CD4. Indeed, the amount of lck associated with CD4 in CD4+CD8+ thymocytes is markedly increased in major histocompatibility complex (MHC) class II- mice that lack the intrathymic ligand for CD4 and in which surface CD4 molecules are consequently not engaged. Thus, the present study demonstrates that (a) activation of lck in CD4+CD8+ thymocytes regulates distribution and expression of TCR components; (b) unlike CD4 molecules, CD8 molecules on CD4+CD8+ thymocytes cannot efficiently activate lck despite their significantly greater surface expression; and (c) the amount of lck associated with CD4 in the CD4+CD8+ thymocytes is inversely related to the extent of CD4 engagement by MHC class II molecules in the thymus.

Aluminum fluoride stimulates surface protrusions in cells overexpressing the ARF6 GTPase.

To study the effector function of the ADP- ribosylation factor (ARF) 6 GTP-binding protein, we transfected HeLa cells with wild-type, epitope-tagged ARF6. Previously shown to indirectly activate the ARF1 GTPase, aluminum fluoride (AIF) treatment of ARF6-transfected cells resulted in a redistribution of both ARF6 and actin to discrete sites on the plasma membrane, which became increasingly protrusive over time. The effects of AIF were reversible, specific to cells transfected with wild-type ARF6, and resembled the cellular protrusions observed in cells expressing the GTPase defective mutant of ARF6. Importantly, the protrusions observed in cells transfected with ARF6 were distinct from the enhanced stress fibers and membrane ruffles observed in cells transfected with RhoA and Rac1, respectively. In cells forming protrusions, there was an apparent stimulation of macropinocytosis and membrane recycling within the protrusive structures. In contrast, no block in transferrin uptake or alteration of the distribution of clathrin AP-2 complexes was detected in these cells. The AIF-induced, ARF6- dependent formation of protrusive structures was blocked by cytochalasin D and inhibitors of the lipoxygenase pathway. These observations support a novel role for the ARF6 GTPase in modeling the plasma membrane and underlying cytoskeleton.

Guanine nucleotides modulate the effects of brefeldin A in semipermeable cells: regulation of the association of a 110-kD peripheral membrane protein with the Golgi apparatus.

The release of a 110-kD peripheral membrane protein from the Golgi apparatus is an early event in brefeldin A (BFA) action, preceding the movement of Golgi membrane into the ER. ATP depletion also causes the reversible redistribution of the 110-kD protein from Golgi membrane into the cytosol, although no Golgi disassembly occurs. To further define the effects of BFA on the association of the 110-kD protein with the Golgi apparatus we have used filter perforation techniques to produce semipermeable cells. All previously observed effects of BFA, including the rapid redistribution of the 110-kD protein and the movement of Golgi membrane into the ER, could be reproduced in the semipermeable cells. The role of guanine nucleotides in this process was investigated using the nonhydrolyzable analogue of GTP, GTP gamma S. Pretreatment of semipermeable cells with GTP gamma S prevented the BFA-induced redistribution of the 110-kD protein from the Golgi apparatus and movement of Golgi membrane into the ER. GTP gamma S could also abrogate the observed release of the 110-kD protein from Golgi membranes which occurred in response to ATP depletion. Additionally, when the 110-kD protein had first been dissociated from Golgi membranes by ATP depletion, GTP gamma S could restore Golgi membrane association of the 110-kD protein, but not if BFA was present. All of these effects observed with GTP gamma S in semipermeable cells could be reproduced in intact cells treated with AlF4-. These results suggest that guanine nucleotides regulate the dynamic association/dissociation of the 110-kD protein with the Golgi apparatus and that BFA perturbs this process by interfering with the association of the 110-kD protein with the Golgi apparatus.

Monensin inhibits intracellular dissociation of asialoglycoproteins from their receptor.

Treatment of short-term monolayer cultures of rat hepatocytes with the proton ionophore, monensin, abolishes asialoglycoprotein degradation, despite little effect of the drug on either surface binding of ligand or internalization of prebound ligand. Centrifuging cell homogenates on Percoll density gradients indicates that, as a result of monensin treatment, ligand does not enter lysosomes but sediments instead in a lower density subcellular fraction that is likely an endocytic vesicle. Analyzing the degree of receptor association of intracellular ligand revealed that monensin prevents the dissociation of the receptor-ligand complex that normally occurs subsequent to endocytosis. The weak base, chloroquine, also blocks this intracellular dissociation. Evidence from sequential substitution experiments is presented, indicating that monensin and chloroquine act at the same point in the sequence of events leading to ligand dissociation. These data are discussed in terms of a pH-mediated dissociation of the receptor-ligand complex within a prelysosomal endocytic vesicle.

Intracellular segregation of asialoglycoproteins and their receptor: a prelysosomal event subsequent to dissociation of the ligand-receptor complex.

Rat hepatocytes in monolayer culture rapidly internalized asialoglycoproteins and the receptors to which they are bound. Subsequent to endocytosis, the receptor-ligand complex is dissociated within an acidic endosome (Harford, J., K. Bridges, G. Ashwell, and R. D. Klausner, 1983, J. Biol. Chem. 258:3191-3197; Harford, J., A. W. Wolkoff, G. Ashwell, and R. D. Klausner, 1983, J. Cell Biol. 96:1824-1828). Here we show that addition of the proton ionophore monensin to the cells after dissociation has occurred results in intracellular rebinding of ligand molecules. With increasing time inside the cell, the ability of ligand to reassociate with receptor progressively decreases consistent with a segregation of receptor and ligand. The combination of colchicine and cytochalasin B appears to retard the process of segregation. In contrast, removal of sodium from the medium, while inhibiting degradation of ligand, does not affect the decrease in monensin-mediated rebinding. Nonetheless, both sodium deprivation and treatment with colchicine plus cytochalasin B result in the ligand remaining in a low density, nonlysosomal subcellular fraction. Thus, segregation, like dissociation, appears to occur in a pre-lysosomal endocytic compartment. Perturbation of the endocytic pathway by reduced temperature (18 degrees C) was also explored. Our data are consistent with two temperature-sensitive steps: receptor-ligand dissociation is inhibited and there is an independent temperature-sensitive step involved in delivery of ligand to lysosomes. This second effect was localized as being beyond the point in the pathway sensitive to sodium deprivation.

Study of the transit of an integral membrane protein from secretory granules through the plasma membrane of secreting rat basophilic leukemia cells using a specific monoclonal antibody.

The monoclonal antibody 5G10 reacted specifically with an 80-kD integral membrane protein in rat basophilic leukemia (RBL) cells. Immunofluorescence microscopy studies of RBL cells, fixed and permeabilized, revealed that the 80-kD protein was located in the membrane of cytoplasmic vesicles. The vesicles were identified as secretory granules by their content in immunoreactive serotonin. Expression of the 5G10 antigen on the surface of unstimulated RBL cells was low. However, RBL cells stimulated to secrete with anti-dinitrophenyl IgE followed by dinitrophenyl-bovine serum albumin or with the Ca2+ ionophore A-23187 displayed an increased expression of the antigen on their surface. Surface exposure of the 5G10 antigen was maximal at 5 min after stimulation of secretion. Removal of dinitrophenyl-bovine serum albumin from the incubation medium resulted in internalization of 50% of the antigen within 10 min.

Exposure of K562 cells to anti-receptor monoclonal antibody OKT9 results in rapid redistribution and enhanced degradation of the transferrin receptor.

When the human erythroleukemia cell line K562 is treated with OKT9, a monoclonal antibody against the transferrin receptor, effects on receptor dynamics and degradation ensue. The apparent half-life of the receptor is decreased by greater than 50% as a result of OKT9 treatment. The transferrin receptor is also rapidly redistributed in response to OKT9 such that a lower percentage of the cellular receptors are displayed on the cell surface. OKT9 treatment also leads to a decrease in the total number of receptors participating in the transferrin cycle for cellular iron uptake. The reduction in iron uptake that results from the loss of receptors from the cycle leads to enhanced biosynthesis of the receptor. Receptors with bound OKT9 continue to participate in multiple cycles of iron uptake. However, OKT9 treatment appears to result in a relatively small increase per cycle in the departure of receptors from participation in iron uptake to a pathway leading to receptor degradation. Radiolabeled OKT9 is itself degraded by K562 cells and this degradation is inhibitable by leupeptin or chloroquine. In the presence of leupeptin, OKT9 treatment results in the enhanced intracellular accumulation of transferrin. Because the time involved in the transferrin cycle is shorter (12.5 min) than the normal half-life of the receptor (8 h), a small change in recycling efficiency caused by OKT9 treatment could account for the marked decrease in receptor half-life. In this paper the implications of these findings are discussed as they relate to systems in which receptor number is regulated by ligand.

Selective degradation of T cell antigen receptor chains retained in a pre-Golgi compartment.

We have examined the fate of newly synthesized T cell antigen receptor (TCR) subunits in a T cell hybridoma deficient in expression of the clonotypic beta chain. Synthesis and assembly of the remaining chains proceed normally but surface expression of TCR chains is undetectable in these cells. A variety of biochemical and morphological techniques has been used to show that the TCR chains in these cells fail to be transported to any of the Golgi cisternae. Instead, they are retained in a pre-Golgi compartment which is either part of or closely related to the endoplasmic reticulum. The CD3-delta chain is degraded by a non-lysosomal process that is inhibited at temperatures at or below 27 degrees C. By contrast, the remaining chains (CD3-epsilon, CD3-gamma, and zeta) are very stable over 7 h. We propose possible mechanisms that may explain the differential fate of TCR chains retained in a pre-Golgi compartment.

Role of microtubules in the organization and localization of the Golgi apparatus.

Normal interphase PtK2 and A549 cells display long microtubules radiating from the microtubule-organizing center (MTOC) to the plasma membrane. Both MTOC and Golgi apparatus are contained in the same perinuclear area. Treatment of cells with 1 microM colcemid for 2 h results in microtubule depolymerization and fragmentation of the Golgi apparatus into elements scattered throughout the cytoplasm. Both normal microtubules and the Golgi apparatus assemble again following removal of colcemid. Injection of the alpha, beta-nonhydrolyzable GTP analog, guanosine 5'(alpha, beta-methylene)diphosphate [pp(CH2)pG], into interphase cells growing in normal medium results in the formation of microtubule bundles resistant to colcemid and prevents the fragmentation of the Golgi apparatus. Injection of pp(CH2)pG into cells incubated with colcemid results in substitution of tubulin ribbons for microtubules and has no effect on the Golgi-derived elements scattered throughout the cytoplasm. Removal of colcemid 1 h after the injection of pp(CH2)pG results in polymerization of large numbers of short, single randomly oriented microtubules, whereas the Golgi apparatus remains fragmented. Treatment of cells with 10 microM taxol for 3 h results both in polymerization of microtubule bundles without relation to the MTOC in the cell periphery and fragmentation of the Golgi apparatus. The Golgi-derived fragments are present exclusively in regions of the peripheral cytoplasm enriched in microtubules. The codistribution of microtubules and Golgi elements can be reversed in taxol-treated cells by injection of a monoclonal (YL 1/2) antibody reacting specifically with the tyrosylated form of alpha-tubulin. Cells incubated with colcemid after treatment with taxol have large numbers of Golgi-derived elements in close association with colcemid-resistant microtubule bundles. Incubation of cells with 50 microM vinblastine for 90 min results in microtubule dissembly, formation of tubulin paracrystals, and fragmentation of the Golgi apparatus into elements without relation to the tubulin paracrystals.

Regulating the retention of T-cell receptor alpha chain variants within the endoplasmic reticulum: Ca(2+)-dependent association with BiP.

Immunoglobulin heavy chain binding protein (BiP, GRP 78) coprecipitates with soluble and membrane-associated variants of the T-cell antigen receptor alpha chain (TCR-alpha) which are stably retained within the ER. Chelation of Ca2+ during solubilization of cells leads to the dissociation of BiP from the TCR-alpha variants, which is dependent upon the availability of Mg2+ and hydrolyzable ATP; this suggests that Ca2+ levels can serve to modulate the association/dissociation of these proteins with BiP. In vivo treatment of cells expressing either the soluble or membrane-anchored TCR-alpha variants with the Ca2+ ionophore, A23187, or an inhibitor of an ER Ca(2+)-ATPase, thapsigargin, or the membrane-permeant Ca2+ chelator BAPTA-AM, results in the redistribution of these proteins out of the ER and their subsequent secretion or cell surface expression. Under the same assay conditions, no movement of BiP out of the ER is observed. Taken together, these observations indicate that decreased Ca2+ levels result in the dissociation of a protein bound to BiP, leading to its release from ER retention. These data suggest that the intracellular fate of newly synthesized proteins stably associated with BiP can be regulated by Ca2+ levels in the ER.

Pre-Golgi degradation of newly synthesized T-cell antigen receptor chains: intrinsic sensitivity and the role of subunit assembly.

The T cell antigen receptor (TCR) is a multisubunit complex composed of at least seven transmembrane chains. The predominant species in most T cells has the composition alpha beta gamma delta epsilon zeta 2. The roles of subunit assembly in transport out of the ER and in the recently described process of pre-Golgi degradation of newly synthesized TCR chains were analyzed in a T-cell line deficient in the synthesis of delta chains (delta 2) and in COS-1 fibroblasts transfected with genes encoding individual TCR chains. Studies with the delta-deficient T-cell line showed that, in the absence of delta, the other TCR chains were synthesized at normal rates, but, instead of being transported to the cell surface, they were retained in the ER. Analysis of the fate of TCR chains retained in the ER showed that they were degraded at vastly different rates by a nonlysosomal pathway. Whereas the alpha chains were degraded rapidly, gamma, zeta, and epsilon were relatively long-lived. To analyze whether this selective degradation was because of different intrinsic susceptibility of the individual chains to degradation or to the formation of resistant oligomers, TCR chains were expressed alone or in combinations in COS-1 fibroblasts. These studies showed that (a) individual TCR chains were degraded at different rates when expressed alone in COS-1 cells, and (b) sensitive chains could be stabilized by coexpression with a resistant chain. Taken together, these observations indicate that both intrinsic sensitivity and subunit assembly play a role in determining the rates at which newly synthesized TCR chains are degraded in the ER.

Dissociation of a 110-kD peripheral membrane protein from the Golgi apparatus is an early event in brefeldin A action.

Brefeldin A (BFA) has a profound effect on the structure of the Golgi apparatus, causing Golgi proteins to redistribute into the ER minutes after drug treatment. Here we describe the dissociation of a 110-kD cytoplasmically oriented peripheral membrane protein (Allan, V. J., and T. E. Kreis. 1986. J. Cell Biol. 103:2229-2239) from the Golgi apparatus as an early event in BFA action, preceding other morphologic changes. In contrast, other peripheral membrane proteins of the Golgi apparatus were not released but followed Golgi membrane into the ER during BFA treatment. The 110-kD protein remained widely dispersed throughout the cytoplasm during drug treatment, but upon removal of BFA it reassociated with membranes during reformation of the Golgi apparatus. Although a 30-s exposure to the drug was sufficient to cause the redistribution of the 110-kD protein, removal of the drug after this short exposure resulted in the reassociation of the 110-kD protein and no change in Golgi structure. If cells were exposed to BFA for 1 min or more, however, a portion of the Golgi membrane was committed to move into and out of the ER after removal of the drug. ATP depletion also caused the reversible release of the 110-kD protein, but without Golgi membrane redistribution into the ER. These findings suggest that the interaction between the 110-kD protein and the Golgi apparatus is dynamic and can be perturbed by metabolic changes or the drug BFA.

Failure to release iron from transferrin in a Chinese hamster ovary cell mutant pleiotropically defective in endocytosis.

A Chinese hamster ovary cell mutant defective in the receptor-mediated endocytosis of several unrelated ligands (Robbins, A. R., S. S. Peng, and J. L. Marshall, 1983, J. Cell Biol., 96:1064-1071) failed to accumulate iron provided in the form of diferric transferrin. Analysis of the steps of the transferrin cycle indicated that binding and internalization of transferrin proceeded normally in mutant cells. However, the mutant appeared unable to dissociate iron from transferrin, as evidenced by release of diferric transferrin from the mutant versus apotransferrin from the parent. Uptake of ferric ions from the growth medium was enhanced in the mutant.

Forskolin inhibits and reverses the effects of brefeldin A on Golgi morphology by a cAMP-independent mechanism.

Brefeldin A (BFA) causes rapid redistribution of Golgi proteins into the ER, leaving no definable Golgi apparatus, and blocks transport of proteins into post-Golgi compartments in the cell. In this study we follow the disassembly of the Golgi apparatus in BFA-treated, living cells labeled with NBD-ceramide and demonstrate that forskolin can both inhibit and reverse this process. Long, tubular processes labeled with NBD-ceramide were observed emerging from Golgi elements and extending out to the cell periphery in cells treated with BFA for 5 min. With longer incubations in BFA, the NBD label was dispersed in a fine reticular pattern characteristic of the ER. Treatment with forskolin inhibited these effects of BFA as well as BFA's earliest morphologic effect on the Golgi apparatus: the redistribution to the cytosol of a 110-kD Golgi peripheral membrane protein. In addition, forskolin could reverse BFA's block in protein secretion. Forskolin inhibition of BFA's effects was dose dependent and reversible. High concentrations of BFA could overcome forskolin's inhibitory effect, suggesting forskolin and BFA interact in a competitive fashion. Remarkably, in cells already exposed to BFA, forskolin could reverse BFA's effects causing the 110-kD Golgi peripheral membrane protein to reassociate with Golgi membrane and juxtanuclear Golgi complexes to reassemble. Neither membrane permeant cAMP analogues nor cAMP phosphodiesterase inhibitors could replicate or enhance forskolin's inhibition of BFA. 1,9-Dideoxyforskolin, which does not activate adenylyl cyclase, was equally as effective as forskolin in antagonizing BFA. A derivative of forskolin, 7-HPP-forskolin, that is less potent than forskolin at binding to adenylyl cyclase, was also equally effective as forskolin in antagonizing BFA. In contrast a similar derivative, 6-HPP-forskolin, that is equipotent with forskolin at binding to adenylyl cyclase, did not inhibit BFA's effects. These results suggest that forskolin acts as a competitive antagonist to BFA, using a cAMP-independent mechanism to prevent and reverse the morphologic effects induced by BFA.

Overexpression of wild-type and mutant ARF1 and ARF6: distinct perturbations of nonoverlapping membrane compartments.

The ARF GTP binding proteins are believed to function as regulators of membrane traffic in the secretory pathway. While the ARF1 protein has been shown in vitro to mediate the membrane interaction of the cytosolic coat proteins coatomer (COP1) and gamma-adaptin with the Golgi complex, the functions of the other ARF proteins have not been defined. Here, we show by transient transfection with epitope-tagged ARFs, that whereas ARF1 is localized to the Golgi complex and can be shown to affect predictably the assembly of COP1 and gamma-adaptin with Golgi membranes in cells, ARF6 is localized to the endosomal/plasma membrane system and has no effect on these Golgi-associated coat proteins. By immuno-electron microscopy, the wild-type ARF6 protein is observed along the plasma membrane and associated with endosomes, and overexpression of ARF6 does not appear to alter the morphology of the peripheral membrane system. In contrast, overexpression of ARF6 mutants predicted either to hydrolyze or bind GTP poorly shifts the distribution of ARF6 and affects the structure of the endocytic pathway. The GTP hydrolysis-defective mutant is localized to the plasma membrane and its overexpression results in a profound induction of extensive plasma membrane vaginations and a depletion of endosomes. Conversely, the GTP binding-defective ARF6 mutant is present exclusively in endosomal structures, and its overexpression results in a massive accumulation of coated endocytic structures.

Activation of T cells by a tyrosine kinase activation domain in the cytoplasmic tail of CD3 epsilon.

The multichain T cell antigen receptor functions by interacting with and activating one or more nonreceptor tyrosine kinases. The cytoplasmic tail of the zeta chain can activate T cells independently of the rest of the receptor complex. The function of the remaining invariant CD3 chains remains unknown. A 22-amino acid region of the cytoplasmic tail of CD3 epsilon was also able to independently activate T cells. Stimulation of T cells by means of the cytoplasmic tails of either zeta or CD3 epsilon resulted in quantitatively distinct patterns of tyrosine phosphorylation, suggesting activation of different biochemical pathways.

Transmembrane helical interactions and the assembly of the T cell receptor complex.

Studies of the subunit interactions of the multicomponent T cell antigen receptor (TCR) revealed that specific pairs of chains have the ability to assemble after transfection into fibroblasts. For one such pair, TCR-alpha and CD3-delta, their ability to assemble was encoded by their transmembrane domains. The specificity of this interaction suggests that well-defined helical interactions in the membrane can explain the assembly of some multichain membrane complexes.

High-efficiency expression and solubilization of functional T cell antigen receptor heterodimers.

The T cell receptor (TCR) zeta chain was attached to the TCR alpha and beta extracellular domains to induce efficient expression of alpha beta heterodimers that can recognize complexes of antigen with major histocompatibility complex (MHC) molecules. Chimeric constructs expressed in RBL-2H3 cells were efficiently transported to the cell surface uniquely as disulfide-linked heterodimers. Transfectants were activated by specific antigen-MHC complexes, which demonstrated that the expressed alpha beta was functional and that CD3 was not required for antigen-MHC binding. Constructs with thrombin cleavage sites were efficiently cleaved to soluble disulfide-linked heterodimers. Thus, attachment of TCR zeta domains and protease cleavage sites to TCR alpha and beta induces expression of demonstrably functional heterodimers that can be solubilized.

A peptide sequence confers retention and rapid degradation in the endoplasmic reticulum.

A nonlysosomal pathway exists for the degradation of newly synthesized proteins retained within the endoplasmic reticulum (ER). This pathway is extremely selective: whereas some proteins are rapidly degraded, others survive for long periods in the ER. The question of whether this selectivity is due to the presence within the sensitive proteins of definable peptide sequences that are sufficient to target them for degradation has been addressed. Deletion of a carboxyl-terminal sequence, comprising the transmembrane domain and short cytoplasmic tail of the alpha chain of the T cell antigen receptor (TCR-alpha), prevented the rapid degradation of this polypeptide. Fusion of this carboxyl-terminal sequence to the extracellular domain of the Tac antigen, a protein that is normally transported to the cell surface where it survives long-term, resulted in the retention and rapid degradation of the chimeric protein in the ER. Additional mutagenesis revealed that the transmembrane domain of TCR-alpha alone was sufficient to cause degradation within the ER. This degradation was not a direct consequence of retention in the ER, as blocking transport of newly synthesized proteins out of the ER with brefeldin A did not lead to degradation of the normal Tac antigen. It is proposed that a 23-amino acid sequence, comprising the transmembrane domain of TCR-alpha, contains information that determines targeting for degradation within the ER system.