Coated-vesicle formation in vitro: Conflicting results using different assays

FORUM III

I

comment

I

The events leading to the formation of ctathrin-coated vesicles from the plasma membrane during endocytosis in viva have been delineated, although the biochemical mechanisms are poorly understood 1, Coat proteins assemble from cytosolic pools onto the m e m b r a n e to form a planar protein lattice that demarcates the beginning of a coated pit. The planar lattices gain curvature along with the overlying m e m b r a n e to sequester receptor-ligand complexes into deeply invaginated coated pits. Membrane fission leads to formation of a sealed coated vesicle tO complete the internalization process. The principal actors in this scene have been identified. The major subunits of the plasma membrane coat are clathrin triskelions , which consist of three heavy chains and three: tightly associated light chains2, 3, and AP2 complexes, which are heterotetramers consisting of two -100 kDa subunits (~ and [3) and two smaller subunits of -50 kDa and -16 kDa 4-6 (Fig. 1). The cue for efficient recruitm e n t of receptors into coated pits is given by an 'internalization' motif consisting of a tyrosine residue presented in the context of a ~-turn and located on the cytoplasmic domain of receptors7, 8. However, in contrast to our knowledge of the principal actors, proteins that play supporting or regulatory roles have yet to be identified. A variety of in vitro approaches for studying coat assembly b o t h in solution9,10 and on the membrane 11-13 have been developed, and are reviewed elsewhere1, 4-6, The focus here is on two in vitro approaches to study coated-vesicle formation at the cell surface. Coated-vesicle budding from isolated plasma membranes One approach to reconstitute coated-vesicle formation uses isolated plasma membranes in an extension of a previously described system for measuring coated:pit assembly14,15. Suspension cells are attached to poly-L-lysine coated surfaces at 4°C by centrifugation. The cytoplasmic surfaces of the attached plasma membranes are exposed by sonication and then re-incubated under a variety of assay conditions. Coated-vesicle formation is measured indirectly by the disappearance of endogenous coated pits as observed biochemically by using an ELISA-based assay, or morphologically by electron and light microscopy. In some cases, goldlabelled ligands for the low-density lipoprotein (LDL) receptor (either LDL or anti-LDL-receptor antibodies) are bound to cells in suspension at 4°C before their attachment to surfaces so that the disappearance of ligand from the membranes can also be measured. Under appropriate incubation conditions, 60-80% of morphologically recognizable coated pits and >80% of receptor-bound gold particles disappear. Control experiments show that the disappearance of LDL-gold requires the clustering of wild-type LDL receptors and does not occur after membranes are stripped at high pH to remove endogenous clathrin 14. The disappearance of clathrin from the membranes requires incubation at 37°C, the addition of a cytosolic fraction, >100 ~M Ca 2+ and TRENDS IN CELL BIOLOGY VOL. 3 MAY 1993

Coated-vesicle formation in vitro: conflicting results using different assays Cell-free systems provide essential tools for elucidating the molecular mechanisms underlying complex cellular processes such as vesicular transport. The biochemical utility of these model systems is strengthened by assays that allow rapid, quantitative detection of the events being studied. Two model systems have recently been developed to reconstitute coated-vesicle budding, and two different biochemical assays are used to detect this event. Striking differences in the biochemical requirements for 'coated-vesicle budding' are detected by these two assays, suggesting that two distinct events are being measured. These findings have wide implications for the use of cell-free assay systems in cell biology.

the addition of nucleotides (with the specificity ATP = ATP~3 -- ADP > GTP >> AMP-PNP). Clathrin disappearance in the presence of ATP is not inhibited by either ATPyS or AMP-PNP, indicating that ATP hydrolysis is not required. The ability of ATPy5 to support activity could suggest the involvement of protein kinases14,15; however, since ADP also supports clathrin loss this possibility is unlikely. GTPyS had no inhibitory effect, indicating that GTPases do not participate in the events leading to the disappearance of ctathrin. Coated-vesicle formation in perforated A431 cells A second cell-free assay system for reconstituting coated vesicle formation involves the use of perforated A431 cells16,17. In this system adherent cells are scraped, leaving behind small areas of the plasma membrane. As the gaps are too large to reseal, the scraped cells can be washed flee of cytosol by low-speed centrifugation and remain totally accessible to exogenously added reagents such as cytosolic proteins, peptides or antibodies. Receptor-mediated endocyt0sis of transferrin (Tfn) © 1993 ElsevierScience Publishers Ltd (UK) 0962-8924/93/$06.00

Sandra Schmid is at the Department

of Cell Biology, The Scripps Research Institute, 10666 N. Torrey

PinesRd, La]olla, CA, USA.

145

FORUM

initiation/ priming

invagination/I

I coat assembly/ i receptor recruitment

sequestration

I budding/ I internalization

Ligand becomes inaccessible to avidin Ligand becomes inaccessible to MesNa

~.

l\

•!1

¢

I

Stimulated by GTP

Inhibited by GDPI3S Requires ATP hydrolysis Inhibited by GTP~/S

Requires ATP hydrolysis Inhibited by GTP~/S

I

I I

Requires cytosolic factors

and is Ca 2+ independent Requires direct interaction with receptor tails, AP2 complexes and cytosolic clathrin

FIGURE 1

Energy and cytosolic requirements for coated-vesicle formation in perforated A431 cells. Assaysthat measure the sequestration and/or internalization of receptor-bound ligands in perforated A431 cells allow detection of distinct events along the pathway of coated-vesicle formation. MesNa (13-mercaptoethanesulphonate) is a small membrane-impermeant reducing agent. Four stages in coated-vesicleformation can now be biochemically delineated: priming, coated-pit assembly and receptor recruitment, coated-pit invagination and coated-vesicle budding. Thesestage-specific assays have revealed that multiple ATPases, GTPasesand cytosolic factors are differentially required for coated-pit assembly, invagination and coated-vesicle budding.

in perforated cells has been extensively characterized morphologically using peroxidase or goldlabelled ligands. These studies have shown that Tfn is concentrated in deeply invaginated coated pits, internalized in sealed coated vesicles and transported to endosomal structures along a pathway that is morphologically indistinguishable from that observed in intact cells (Refs 16, 17; C. Lamaze, T. Baba, T. E. Redelmeier and S. L. Schmid, unpublished). Events leading to coated-vesicle formation in perforated cells are detected biochemically by the internalization of ligand into membrane-bound vesicles. Intermediate reactions leading to the sequestration of ligand into deeply invaginated coated pits can also be detected 17,18,19. For detedion, Tfn is biotinylated through a cleavable disulphide bond (Tfn labelled in this way is referred to as BSST). The sequestration of BSST into deeply invaginated coated pits with openings narrow enough to exclude large probes is scored by its acquired inaccessibility to avidin. BSST in deeply invaginated pits remains accessible to a small membrane-impermeant reducing agent; however, internalization of BSST into sealed coated vesicles results in its acquired resistance to this small probe (Fig. 1). While 50-70% of ligand can be efficiently sequestered into deeply invaginated coated pits, coated-vesicle formation in perforated cells is less efficient, resulting in the internalization of 25-30% of bound ]igand. 146

Characterization of endocytosis using this model system has revealed that both sequestration and internalization of bound ligand requires incubation at 37°C, ATP hydrolysis, and multiple cytosolic factors (Refs 16-19; C. Lamaze, T. Baba, T. E. Redelmeier and S. L. Schmid, unpublished) (see Fig. 1). Coatedvesicle formation in perforated A431 cells is independent of added Ca 2+ and is insensitive to addition of 5 mM EGTA (C. Lamaze, T. Baba, T. E. Redelmeier and S. L. Schmid, unpublished). Early events leading to ligand sequestration into deeply invaginated coated pits specifically require direct interactions with receptor tails, AP2 complexes and clathrin. AP2 complexes are a limiting component of the cytosol in that ligand sequestration can be stimulated by addition of purified AP218. Interestingly, clathrin isolated by standard procedures from coated vesicles is inactive in this assay, whereas an enriched fraction of clathrin derived from cytoso] is active. These results are consistent with previous findings on the inability of purified clathrin to bind to harshly stripped membranes even after addition of AP2 complexes 13 and indicate that assembled and unassembled pools of clathrin might be functionally distinct TM. Additionally, coated-vesicle formation in this system requires multiple GTP-binding proteins 19. Coated-pit assembly and coated-vesicle budding are inhibited by nonhydrolysable analogues of GTP and by [A1F4]-, whereas coated-pit invagination is TRENDS IN CELLBIOLOGYVOL. 3 MAY 1993

FORUM

unaffected. By contrast, both coated-pit invagination and coated-vesicle budding are stimulated by GTP and inhibited by GDP[~S. These results indicate that multiple GTP-binding proteins participate at distinct stages in coated-vesicle formation 19. Different assays detect different events A comparison of the properties of the two assay systems for coated-vesicle formation in vitro leads to the conclusion that they cannot be measuring the same event. Does either assay measure events involved in coated-vesicle formation? To ensure that the events detected biochemically in these assays are a result of the cellular processes we wish to study, it is critical to apply as many biochemical and morphological criteria~n vitro as have been determined in vivo and to examine the reaction products produced. The physiological properties of coated-vesicle formation in perforated cells suggest that events that occur in this model system mimic the situation in vivo. Morphological studies have shown that Tfn is internalized in perforated cells through coated vesicles and delivered to endosomal compartments (Refs 16, 17; C. Lamaze, T. Baba, T. E. Redelmeier and S. L. Schmid, unpublished). The requirement for ATP hydrolysis in vitro is consistent with results obtained in vivo using ATPdepleted cells 20. Biochemically, the ligand is sequestered into a sealed membrane-bound vesicle as measured by latency to a small membraneimpermeant probe. While further morphological analysis is required to determine the structure of the deeply invaginated coated pit that efficiently sequesters bound ligands, this intermediate has been detected in vivo in ATP-depleted cells 20. Conversely, the products of coated-vesicle formation from isolated plasma membranes have not been characterized. If the loss of clathrin and receptor-bound ligand detected morphologically were due to coated-vesicle formation, then there should be a parallel loss of receptors, AP2 complexes and clathrin. In addition, the requirements for the loss of these other constituents of coated vesicles should be identical to those documented for the loss of clathrin. These criteria have been ohly partially tested and met. For example, there is a corresponding loss of AP2 complexes but this toss appears to require less cytosol (at low cytosol concentrations -30% of AP2 complexes are lost, while 95% of membrane clathrin remains). Similarly, >80% of the clustered ligand-gold complexes are lost under conditions that result in loss of only 60% of the clathrin and only 16-36% of the LDL receptors 14. Furthermore, some properties of clathrin loss from isolated membranes appear nonphysiological. By contrast to in vivo findings 20, clathrin loss does not require either ATP or ATP hydrolysis. The disappearance of clathrin from isolated plasma membranes requires >100 ~M Ca 2+, whereas receptor-mediated endocytosis of EGF in intact cells is unaffected by membrane-permeable Ca 2÷ chelators 21. Recent results have suggested that this in vitro requirement for Ca 2+ reflects the involvement of TRENDS IN CELL BIOLOGY VOL. 3 MAY 1993

comment

annexin VI in coated-vesicle budding is. However, several discrepancies in the data bring this interpretation into question. For example, purified annexin VI supported partial clathrin loss (-30%), but this activity was independent of added Ca 2+. Ca 2+ dependence was observed only upon addition of annexin-depleted cytosol to the reaction mixture, suggesting that other factors confer the Ca 2+ r e q u i r e m e n t . Ca 2+ concentrations tenfold higher than that required for annexin-membrane interactions and 100-1000-fold higher than physiological levels are required for clathrin loss from isolated membranes. While annexin VI might somehow participate in destabilization of clathrin membrane interactions in vitro, evidence suggests it does not participate in coated-vesicle budding in vivo. Ca 2+ chelators do not inhibit coated-vesicle budding in perforated cells 19 or in intact cells 21 and recent results have shown that annexin VI is not expressed in A431 cells (S. Moss and G. Warren, pers. commun.) indicating that it is not an absolute requirement for endocytosis. Coated-vesicle formation in perforated cells is measured directly by the internalization of ligand into a membrane-bounded vesicle where it acquires latency to small membrane-impermeant probes. By contrast, a negative assay is used to measure coatedvesicle formation from isolated membranes. The disappearance of coats or ligands could arise from m a n y unrelated events. For example, are the coat proteins and receptors still intact and immunologically reactive after incubation or have they been digested? Could the activation of abundant Ca2+-dependent proteases be involved? Have the coats and receptor-ligand complexes been destabilized and released in a process that is independent of vesicle budding? Is the ligand that has disappeared now found in a sealed vesicle? Has the ligand become latent or sedimentable? A positive indication of coated-vesicle formation is required in this assay system. Factors that regulate coated-vesicle formation in vitro Although clathrin and AP assembly in solution and onto membranes can occur spontaneously in vitro 9-13, studies on coated-vesicle formation in perforated cells (Refs 16-19; C. Lamaze, T. Baba, T. E. Redelmeier and S. L. Schmid, unpublished) have demonstrated additional cytosolic and energy requirements as diagrammed in Fig. 1. ATP and GTP hydrolysis, for example, are required at multiple stages of coated-vesicle formation including clathrin assembly into functionally active coated pits, the invagination of pits to sequester receptorb o u n d ligands and coated-vesicle budding 17-19. Insight into the ATP- and GTP-dependent events awaits identification of the ATPases and GTPases involved. Vesicular transport by nonclathrin- or COP-coated vesicles requires multiple GTPases including heterotrimeric G proteins, members of the rab family of small GTP-binding proteins and ARF or ADP-ribosylation factor, a major constituent of the coat. COP-coated-pit assembly 14 7

FORUM

comment

appears to be regulated through interactions of ARF with heterotrimeric G proteins, while small GTP-binding proteins appear to participate in later events 22. Both clathrin-coated-pit assembly and coated-vesicle budding in vitro are sensitive to antagonists of heterotrimeric G proteins 19. Furthermore, both rab5 (Ref. 23) and ARF24 have been identified as minor constituents in preparations of clathrincoated vesicles. Direct evidence for the involvement of these GTP-binding proteins in coated-vesicle formation has not been obtained 19,23, but genetic evidence from studies of the shibire mutation in Drosophila has suggested that dynamin (the m a m malian homologue of the shibire gene product) might be one GTPase required for endocytosis 2s-27. Flie s carrying mutations in the shibire gene show a pleiotropic increase in the surface area of the plasma membrane and in the number of coated pits, suggesting that coated-vesicle formation is blocked while m e m b r a n e recycling and secretion conAcknowledgements tinue 2S. Dynamin might play a role in governing I thank members the decision to close a coated pit by triggering of my laboratory invagination or vesicle budding. and W. E. Balch Identification of the cytosolic factors that drive for helpful dis- coated-vesicle formation and elucidation of their cussions and con- mechanism of action will require the judicious use structive criticism. of model cell~free systems and biochemical assays Researchin my that faithfully reconstitute and measure these laboratory has events. The stage-specific assays developed for been supported by measuring coated-vesicle formation in perforated the NIH and by A431 cells provide powerful tools with which to the Lucille P. begin the biochemical dissection of this complex Markey Charitable process. Trust in the form of a Lucille P. Markey Scholarship.This is TSRImanuscript no. 7735-CB.

148

References 1 SCHMID, S. L. (1992) BioEssays 14, 58%596 2 PEARSE,B. M. F. and CROWTHER,R. A. (1987) Annu. Rev. Biophys. Chem. 16, 49-68 3 BRODSKY,F. M. (1988) Science 242, 1396-1402

4 PEARSE,B. M. F. and ROBINSON,M. S. (1990) Annu. Rev. Cell Biol. 6, 151-171 5 KEEN,J. H. (1990) Annu. Rev. Biochem. 59, 415-438 6 ROBINSON,M. S. (1992) Trends CellBioL 2, 293-297 7 TROWBRIDGE,I. S. (1991) Curr. Opin. Cell Biol. 3, 634-641 8 VAUX, D. (1992) Trends Cell Biol. 2, 189-192 9 AHLE,S. and UNGEWlCKELL,E. (1989) J. BioL Chem. 264, 20089-20093 10 PRASAD,K. and KEEN,J. H. (1991 ) Biochemistry 30, 5590-5597 11 MOORE,M. S., MAHAFFEY,D. T., BRODSKY,F. M. and ANDERSON, R. G. W. (1987) Science 236, 558-563 12 MAHAFFEY,D. T., MOORE, M. S., BRODSKY,F. M. and ANDERSON, R. G. W. (1989) J. Cell BioL 108, 1615-1624 13 MAHAFFEY,D. T., PEELER,J. S., BRODSKY,F. M. and ANDERSON, R. G. W. (1990) J. BioL Chem. 265, 16514-16520 14 LIN, H. C., MOORE, M. S., SANAN, D. A. and ANDERSON, R. G. W. (1991) J. CellBioL 114, 881-891 15 LIN, H. C., SUDHOF,T. C. and ANDERSON, R. G. W. (1992) Cell 70, 283-291 16 SMYTHE,E., PYPAERT,M., LUCOCQ, J. and WARREN, G. (1989) J. Cell BioL 108, 843-853 17 SCHMID, S. L. and SMYTHE,E. (1991) J. CellBioL 114, 869-880 18 SMYTHE,E., CARTER,L. L. and SCHMID, S. L. (1992) J. Cell BioL 119, 1163-1171 19 CARTER,L. L., REDELMEIER,T. E., WOOLLENWEBER,L. A. and SCHMID, S. L. (1993)J. Cell. BioL 120, 37-45 20 SCHMID, S. L. and CARTER,L. L. (1990) J. Cell Biol. 111, 2307-2318 21 CHANG,C. P. et aL (1991)J. BioL Chem. 266, 23467-23470 22 ROTHMAN,]. E. and ORCI, L. (1992) Nature 355, 409-415 23 BUCCI,C. et aL (1992) Cell 70, 715-728 24 LENNARD,J. M., KAHN, R. S. and STAHL, P. D. (1992) J. BioL Chem. 267, 13047-13052 25 KOSAKA,T. and IKEDA, K. (1983) J. Cell Biol. 97, 499-507 26 VAN DER BLIEK,A. M. and MEYEROWlTZ,E. M. (1991) Nature 351, 411-414 27 VALLEE,R. D. (1992) J. Muscle Res. Cell. Mot. 13, 493-496

TRENDS IN CELL BIOLOGY VOL. 3 MAY 1993