The microtubule-dependent formation of a tubulovesicular network with characteristics of the ER from cultured cell extracts

The formation of a dynamic tubulovesicular membrane network that resembles the endoplasmic reticulum (ER) has been observed in extracts of cultured chick embryo fibroblasts (CEF cells) using video-enhanced differential interference contrast microscopy. Initially, membranes in the CEF extracts appeared amorphous and aggregated, but with time, membrane tubules moved out along stationary microtubules. The membrane tubules formed new branches on intersecting microtubules and fused with other branches to form a network of interconnected polygons. The tubulovesicular network was solubilized by detergent and took on a beaded morphology in a hypotonic buffer. Formation of the tubulovesicular network required ATP and microtubules. The network did not contain elements of the plasma membrane, Golgi apparatus, or mitochondria but could be labeled with ER markers. We suggest that the tubulovesicular network contains components from the ER and is formed by membrane associated motors moving upon microtubules in a process we call microtubule-dependent tethering.     

Endoplasmic reticulum-targeted GFP reveals ER remodeling in Mesorhizobium-treated Lotus japonicus root hairs during root hair curling and infection thread formation

The endoplasmic reticulum (ER) of the model legume Lotus japonicus was visualized using green fluorescent protein (GFP) fused with the KDEL sequence to investigate the changes in the root hair cortical ER in the presence or absence of Mesorhizobium loti using live fluorescence imaging. Uninoculated root hairs displayed dynamic forms of ER, ranging from a highly condensed form to an open reticulum. In the presence of M. loti, a highly dynamic condensed form of the ER linked with the nucleus was found in deformed, curled, and infected root hairs, similar to that in uninoculated and inoculated growing zone I and II root hairs. An open reticulum was primarily found in mature inoculated zone III root hairs, similar to that found in inactive deformed/curled root hairs and infected root hairs with aborted infection threads. Co-imaging of GFP-labeled ER with light transmission demonstrated a correlation between the mobility of the ER and other organelles and the directionality of the cytoplasmic streaming in root hairs in the early stages of infection thread formation and growth. ER remodeling in root hair cells is discussed in terms of possible biological significance during root hair growth, deformation/curling, and infection in the Mesorhizobium–L. japonicus symbiosis.     

Construction of the endoplasmic reticulum

To study the construction of the ER, we used the microtubule-disrupting drug nocodazole to induce the complete breakdown of ER structure in living cells followed by recovery in drug-free medium, which regenerates the ER network within 15 min. Using the fluorescent dye 3,3'-dihexyloxacarbocyanine iodide to visualize the ER, we have directly observed the network construction process in living cells. In these experiments, the ER network was constructed through an iterative process of extension, branching, and intersection of new ER tubules driven by the ER motility previously described as tubule branching. We have tested the cytoskeletal requirements of this process. We find that newly formed ER tubules are aligned with single microtubules but not actin fibers or vimentin intermediate filaments. Microtubule polymerization preceded the extension of ER tubules and, in experiments with a variety of different drugs, appeared to be a necessary condition for the ER network formation. Furthermore, perturbations of the pattern of microtubule polymerization with microtubule-specific drugs caused exactly correlated perturbations of the pattern of ER construction. Induction of abnormally short, nonintersecting microtubules with 20 microM taxol prevented the ER network formation; ER tubules only extended along the few microtubules contacting the aggregated ER membranes. This requirement for a continuous network of intersecting microtubules indicates that ER network formation takes place through the branching and movement of ER membranes along microtubules. Cytochalasin B had no apparent effect on the construction of the ER network during recovery, despite apparently complete disruption of actin fibers as stained by phalloidin. Blockage of protein synthesis and disorganization of intermediate filaments with cycloheximide pretreatment also failed to perturb ER construction.     

Insights into the physiological role of pig liver esterase: Isoenzymes show differences in the demethylation of prenylated proteins

The possible physiological role of PLE (E.C. 3.1.1.1) located in the endoplasmic reticulum (ER) of pig liver cells in the conversion of endogenous compounds was investigated as it was reported, that PLE acts as prenylated methylated protein methyl esterase (PMPMEase) hydrolysing methylesters of prenylated proteins. Using the specific PMPMEase substrate benzoyl–glycyl–farnesyl–cysteine methyl ester (BzGFCM), six different PLE isoenzymes expressed recombinantly in the yeast Pichia pastoris were found active. Activities ranged from 1.6–15.6 mU per mg protein and it is suggested that Pro285 has a major influence on high activity. In addition, the role of the C-terminal HAEL retention signal for translocation of pig liver esterase (PLE) in the endoplasmic reticulum (ER) of eukaryotic cells was studied using the γ-isoenzyme of PLE expressed in Pichia pastoris. Using truncated versions (HAE, HA, H and without retention signal) of the enzyme it was found that in contrast to earlier reports no influence of the signal peptide on the expression rate of PLE was found. However, higher enzyme activities were obtained in the periplasmatic fraction compared to the supernatant irrespective of the presence or absence of HAEL and the trimeric formation seems to occur in the supernatant of P. pastoris X33 enabling an easier transition of monomeric forms through cell membranes.     

Brefeldin A effects on the trans-Golgi network and Golgi bodies inBotryococcus braunii are not uniform during the cell cycle

Summary Using cryo-fixation and freeze-substitution electron microscopy, the effects of brefeldin A (BFA) on the structure of the trans-Golgi network (TGN), the endoplasmic reticulum (ER), and Golgi bodies in the unicellular green algaBotryococcus braunii were examined at various stages of the cell cycle. In the presence of BFA, all the TGNs of interphase and dividing cells aggregated to form a single tubular mass. In contrast, the TGNs decomposed just after cell division and disappeared during cell wall formation. Throughout the cell cycle, the TGN produced at least six kinds of vesicles, of which two were not formed in the presence of BFA: vesicles with a diameter of 200 nm and fibrillar substances, which formed in interphase cells; and vesicles with a diameter of 180–240 nm, which may participate in septum formation. In addition, the number of clathrin-coated vesicles attaching to the TGN decreased. In interphase cells, BFA induced the disassembly of Golgi bodies and an increase in the smooth-ER cisternae at the cis-side of Golgi bodies. This result may suggest the existence of retrograde transport from the Golgi bodies to the ER in the presence of BFA. These drastic structural changes in the Golgi bodies and the ER of interphase cells were not observed in BFA-treated dividing cells.Abbreviations BFA brefeldin A - ER endoplasmic reticulum - TGN trans-Golgi network     

Observation and quantification of individual microtubule behavior in vivo: microtubule dynamics are cell-type specific

Recent experiments have demonstrated that the behavior of the interphase microtubule array is cell-type specific: microtubules in epithelial cells are less dynamic than microtubules in fibroblasts (Pepper-kok et al., 1990; Wadsworth and McGrail, 1990). To determine which parameters of microtubule dynamic instability behavior are responsible for this difference, we have examined the behavior of individual microtubules in both cell types after injection with rhodamine-labeled tubulin subunits. Individual microtubules in both cell types were observed to grow, shorten, and pause, as expected. The average amount of time microtubules remained within the lamellae of CHO fibroblasts, measured from images acquired at 10-s intervals, was significantly shorter than the average amount of time microtubules remained within lamellae of PtK1 epithelial cells. Further analysis of individual microtubule behavior from images acquired at 2-s intervals reveals that microtubules in PtK1 cells undergo multiple brief episodes of growth and shortening, resulting in little overall change in the microtubule network. In contrast, microtubules in lamellae of CHO fibroblasts are observed to undergo fewer transitions which are of longer average duration, resulting in substantial changes in the microtubule network over time. A small subset of more stable microtubules was also detected in CHO fibroblasts. Quantification of the various parameters of dynamic instability behavior from these sequences demonstrates that the average rates of both growth and shortening are significantly greater for the majority of microtubules in fibroblasts than for microtubules in epithelial cells (19.8 +/- 10.8 microns/min, 32.2 +/- 17.7 microns/min, 11.9 +/- 6.5 microns/min, and 19.7 +/- 8.1 microns/min, respectively). The frequency of catastrophe events (1/interval between catastrophe events) was similar in both cell types, but the frequency of rescue events (1/time spent shrinking) was significantly higher in PtK1 cells. Thus, individual microtubules in PtK1 lamellae undergo frequent excursions of short duration and extent, whereas most microtubules in CHO lamellae undergo more extensive excursions often resulting in the appearance or disappearance of microtubules within the field of view. These observations provide the first direct demonstration of cell-type specific behavior of individual microtubules in living cells, and indicate that these differences can be brought about by modulation of the frequency of rescue. These results directly support the view that microtubule dynamic instability behavior is regulated in a cell-type specific manner.     

Endoplasmic reticulum–endosome contact increases as endosomes traffic and mature

The endosomal pathway is responsible for plasma membrane cargo uptake, sorting, and, in many cases, lysosome targeting. Endosome maturation is complex, requiring proper spatiotemporal recruitment of factors that regulate the size, maturity, and positioning of endosomal compartments. In animal cells, it also requires trafficking of endosomes on microtubules. Recent work has revealed the presence of contact sites between some endosomes and the endoplasmic reticulum (ER). Although these contact sites are believed to have multiple functions, the frequency, dynamics, and physical attributes of these contacts are poorly understood. Here we use high-resolution three-dimensional electron microscopy to reveal that ER tubules wrap around endosomes and find that both organelles contact microtubules at or near membrane contact sites. As endosomes traffic, they remain bound to the ER, which causes the tubular ER to rearrange its structure around dynamic endosomes at contact sites. Finally, as endosomes transition through steps of maturation, they become more tightly associated with the ER. The major implication of these results is that endosomes mature and traffic while coupled to the ER membrane rather than in isolation.     

MAP6-F Is a Temperature Sensor That Directly Binds to and Protects Microtubules from Cold-induced Depolymerization

Microtubules are dynamic structures that present the peculiar characteristic to be ice-cold labile in vitro. In vivo, microtubules are protected from ice-cold induced depolymerization by the widely expressed MAP6/STOP family of proteins. However, the mechanism by which MAP6 stabilizes microtubules at 4 °C has not been identified. Moreover, the microtubule cold sensitivity and therefore the needs for microtubule stabilization in the wide range of temperatures between 4 and 37 °C are unknown. This is of importance as body temperatures of animals can drop during hibernation or torpor covering a large range of temperatures. Here, we show that in the absence of MAP6, microtubules in cells below 20 °C rapidly depolymerize in a temperature-dependent manner whereas they are stabilized in the presence of MAP6. We further show that in cells, MAP6-F binding to and stabilization of microtubules is temperature- dependent and very dynamic, suggesting a direct effect of the temperature on the formation of microtubule/MAP6 complex. We also demonstrate using purified proteins that MAP6-F binds directly to microtubules through its Mc domain. This binding is temperature-dependent and coincides with progressive conformational changes of the Mc domain as revealed by circular dichroism. Thus, MAP6 might serve as a temperature sensor adapting its conformation according to the temperature to maintain the cellular microtubule network in organisms exposed to temperature decrease.     

Dynamic instability and motile events of native microtubules from squid axoplasm

Native microtubules from extruded axoplasm of squid giant axons were used as a paradigm to characterize the motion of organelles along free microtubules and to study the dynamics of microtubule length changes. The motion of large round organelles was visualized by AVEC-DIC microscopy and analyzed at a temporal resolution of 10 frames per second. The movements were smooth and showed no major changes in velocity or direction. During translocation, the organelles paused very rarely. Superimposed on the rather constant mean velocity was a velocity fluctuation, which indicated that the organelles are subject to considerable thermal motion during translocation. Evidence for a regular low-frequency oscillation was not found. The thermal motion was anisotropic such that axial motion was less restricted than lateral motion. We conclude that the crossbridge connecting the moving organelle to the microtubule has a flexible region that behaves like a hinge, which permits preferential movement in the direction parallel to the microtubule. The dynamic changes in length of native microtubules were studied at a temporal resolution of 1 Hz. About 98% of the native microtubules maintained their length ("stable" microtubules), while 2% showed phases of growing and/or shrinking typical for dynamic instability ("dynamic" microtubules). Gliding and organelle motion were not influenced by dynamic length changes. Transitions between growing and shrinking phases were low-frequency events (1-10 minutes per cycle). However, a new type of microtubule length fluctuation, which occurred at a high frequency (a few seconds per cycle), was detected. The length changes were in the 1-3 micron range. The latter events were very prominent at the (+) ends. It appears that the native axonal microtubules are much more stable than the purified microtubules and the microtubules of cultured cells that have been studied thus far. Potential mechanisms accounting for the three states of microtubule stability are discussed. These studies show that the native microtubules from squid giant axons are a very useful paradigm for studying microtubule-related motility events and microtubule dynamics.     

Microtubule- and microfilament-based dynamic activities of the endoplasmic reticulum and the cell surface in epithelial cells ofSpongilla lacustris (Porifera,Spongillidae)

Summary Dynamic activities of the endoplasmic reticulum (ER) and of the cell surface were analyzed in living epithelial cells (pinacocytes) ofSpongilla lacustris by contrast-enhanced video microscopy with the AVEC-DIC or the ACE equipment. Long term sequences revealed the ER to be a highly unstable system undergoing permanent alterations of the reticular patterns in that tubules merge and split or polygons open and close again. Treatment with colcemid or colchicine causes distinct changes of the typical motile phenomena, whereas cytochalasin D has no influence. On the other hand, the dynamic behavior of the cell surface is characterized by distinct ruffling activities as well as the formation and retraction of spiky filopodia. In contrast to the described ER dynamics, cell surface phenomena are clearly influenced by cytochalasin D but not by colcemid or colchicine. Altogether, results of the present paper are similar to correspondent observations on mammalian cells and point to microtubules and microfilaments as the cytoskeletal elements being responsible for ER and cell surface dynamics, respectively.     

Role of P30 in replication and spread of TMV

The P30 movement protein (MP) of tobacco mosaic virus is essential for distribution of sites of replication within infected cells and for cell–cell spread of infection. MP is an integral membrane protein and in early and mid-stages of infection causes severe disruption of the cortical endoplasmic reticulum (ER). MP also associates with microtubules, and in late stages is targeted for degradation by the 26S proteosome. During these stages, the ER regains its normal pre-infection configuration. Viral RNA is associated with ER and microtubules in the presence of MP. The MP is phosphorylated and mutation of the phosphorylated amino acid reduced association of MP with the ER, plasmodesmata, and microtubules, and altered the stability of the MP. The nature of the association of MP with vRNA and ER and microtubules, and the role of phosphorylation of MP in each of these functions, if any, remains to be determined.     

Microtubules and the endoplasmic reticulum are highly interdependent structures

The interrelationships of the endoplasmic reticulum (ER), microtubules, and intermediate filaments were studied in the peripheral regions of thin, spread fibroblasts, epithelial, and vascular endothelial cells in culture. We combined a fluorescent dye staining technique to localize the ER with immunofluorescence to localize microtubules or intermediate filaments in the same cell. Microtubules and the ER are sparse in the lamellipodia, but intermediate filaments are usually completely absent. These relationships indicate that microtubules and the ER advance into the lamellipodia before intermediate filaments. We observed that microtubules and tubules of the ER have nearly identical distributions in lamellipodia, where new extensions of both are taking place. We perturbed microtubules by nocodazole, cold temperature, or hypotonic shock, and observed the effects on the ER distribution. On the basis of our observations in untreated cells and our experiments with microtubule perturbation, we conclude that microtubules and the ER are highly interdependent in two ways: (a) polymerization of individual microtubules and extension of individual ER tubules occur together at the level of resolution of the fluorescence microscope, and (b) depolymerization of microtubules does not disrupt the ER network in the short term (15 min), but prolonged absence of microtubules (2 h) leads to a slow retraction of the ER network towards the cell center, indicating that over longer periods of time, the extended state of the entire ER network requires the microtubule system.     

Movement and Remodeling of the Endoplasmic Reticulum in Nondividing Cells of Tobacco Leaves

Using a novel analytical tool, this study investigates the relative roles of actin, microtubules, myosin, and Golgi bodies on form and movement of the endoplasmic reticulum (ER) in tobacco (Nicotiana tabacum) leaf epidermal cells. Expression of a subset of truncated class XI myosins, which interfere with the activity of native class XI myosins, and drug-induced actin depolymerization produce a more persistent network of ER tubules and larger persistent cisternae. The treatments differentially affect two persistent size classes of cortical ER cisternae, those >0.3 μm² and those smaller, called punctae. The punctae are not Golgi, and ER remodeling occurs in the absence of Golgi bodies. The treatments diminish the mobile fraction of ER membrane proteins but not the diffusive flow of mobile membrane proteins. The results support a model whereby ER network remodeling is coupled to the directionality but not the magnitude of membrane surface flow, and the punctae are network nodes that act as foci of actin polymerization, regulating network remodeling through exploratory tubule growth and myosin-mediated shrinkage.     

An intact microtubule cytoskeleton is not necessary for interfacial matrix formation in orchid protocorm mycorrhizas

 This paper reports the changes that occur in the microtubule cytoskeleton of cells of orchid protocorms during infection by a compatible mycorrhizal fungus. In cells of protocorms uninfected by a mycorrhizal fungus, microtubules occurred in regular arrays. In contrast, the cells of orchid protocorms with established mycorrhizas appeared to contain irregularly arranged microtubules. Double labelling with anti-β-tubulin and rhodamine-labelled wheat-germ agglutinin demonstrated that these irregularly arranged microtubules occurred only inside fungal hyphae and that microtubules were absent from host cells containing mycorrhizal fungi. Microtubule depolymerisation was shown to occur at the early stages of fungal infection. There was neither loss of nor obvious organisational change in microtubules in cells adjacent to others containing fungal hyphae. Electron microscopy confirmed the presence of an interfacial matrix between the host plasma membrane and the hyphal wall. The loss of microtubules from cells infected by mycorrhizal fungi suggests that an intact host microtubule cytoskeleton is not necessary for the formation of the interfacial matrix in mycorrhizas of orchid protocorms. Accepted: 9 November 1995     

The alpha subunit of sea urchin sperm outer arm dynein mediates structural and rigor binding to microtubules

Glass-adsorbed intact sea urchin outer arm dynein and its beta/IC1 subunit supports movement of microtubules, yet does not form a rigor complex upon depletion of ATP (16). We show here that rigor is a feature of the isolated intact outer arm, and that this property subfractionates with its alpha heavy chain. Intact dynein mediates the formation of ATP-sensitive microtubule bundles, as does the purified alpha heavy chain, indicating that both particles are capable of binding to microtubules in an ATP-sensitive manner. In contrast, the beta/IC1 subunit does not bundle microtubules. Bundles formed with intact dynein are composed of ribbon-like sheets of parallel microtubules that are separated by 54 nm (center-to-center) and display the same longitudinal repeat (24 nm) and cross-sectional geometry of dynein arms as do outer doublets in situ. Bundles formed by the alpha heavy chain are composed of microtubules with a center-to-center spacing of 43 nm and display infrequent, fine crossbridges. In contrast to the bridges formed by the intact arm, the links formed by the alpha subunit are irregularly spaced, suggesting that binding of the alpha heavy chain to the microtubules is not cooperative. Cosedimentation studies showed that: (a) some of the intact dynein binds in an ATP-dependent manner and some binds in an ATP-independent manner; (b) the beta/IC1 subunit does not cosediment with microtubules under any conditions; and (c) the alpha heavy chain cosediments with microtubules in the absence or presence of MgATP2-. These results suggest that the structural binding observed in the intact arm also is a property of its alpha heavy chain. We conclude that whereas force-generation is a function of the beta/IC1 subunit, both structural and ATP-sensitive (rigor) binding of the arm to the microtubule are mediated by the alpha subunit.     

Evidence for a role of capillary pericytes in vascular growth of the developing ovine corpus luteum

The role of microfilaments, microtubules, and mitogen-activated protein (MAP) kinase in regulation of several important dynamic events of porcine oocyte maturation and fertilization is described. Fluorescently labeled microfilaments, microtubules, and cortical granules were visualized using either epifluorescence microscopy or laser scanning confocal microscopy. Mitogen-activated protein kinase phosphorylation was revealed by Western immunoblotting. We showed that 1) microfilament disruption did not affect meiosis resumption and metaphase I meiotic apparatus formation but inhibited further cell cycle progression (chromosome separation) even though MAP kinase was phosphorylated; 2) cortical granule (CG) migration was driven by microfilaments (but not microtubules), and once the chromosomes and CGs were localized beneath the oolemma their anchorage to the cortex was independent of either microfilaments or microtubules; 3) neither microfilaments nor microtubules were involved in CG exocytosis during oocyte activation; 4) sperm incorporation was mediated by microfilaments, while pronuclear (PN) syngamy was controlled by microtubules rather than microfilaments; 5) spindle microtubule organization was temporally correlated with MAP kinase phosphorylation, while the extensive microtubule organization in the sperm aster that is required for PN apposition and syngamy occurred in the absence of MAP kinase activation; and 6) MAP kinase phosphorylation did not change either when microtubules were disrupted by nocodazole or when cytoplasmic microtubule asters were induced by taxol. The present study suggests that the role of the cytoskeleton during porcine oocyte maturation is similar to that of rodents, while the mechanisms of fertilization in pig resemble those of lower vertebrates.     

Pollen exine development precedes microtubule rearrangement inVigna unguiculata (Fabaceae): A model for pollen wall patterning

Summary Although patterns on pollen exines are highly conserved, heritable traits, there is no prevailing explanation for control of pattern development. InVigna unguiculata (Fabaceae), the exine reticulum is unusually coarse so that development of the reticulum can be followed by light microscopy. Developing exine patterns were compared with the arrangement of microtubules in the microspore using immunofluorescence labeling of microtubules. Exine pattern developed in microspores at stages with a regular microtubule pattern. At later stages of exine formation, microtubules were arranged in patches under the lumina of the reticulum. We conclude that microtubules do not determine exine pattern. The developing exine appears to rearrange microtubules. We interpret this as evidence for the selfpatterning of exine based on intrinsic properties of the matrix between the microspore and the callose wall.Abbreviations DIC differential interference contrast - ECM(s) extracellular matrix(ces) - MT(s) microtubule(s) - PBS phosphate buffered saline - SEM scanning electron microscopy     

Alteration of microtubule dynamic instability during preprophase band formation revealed by yellow fluorescent protein-CLIP170 microtubule plus-end labeling

At the onset of mitosis, plant cells form a microtubular preprophase band that defines the plane of cell division, but the mechanism of its formation remains a mystery. Here, we describe the use of mammalian yellow fluorescent protein-tagged CLIP170 to visualize the dynamic plus ends of plant microtubules in transfected cowpea protoplasts and in stably transformed and dividing tobacco Bright Yellow 2 cells. Using plus-end labeling, we observed dynamic instability in different microtubular conformations in live plant cells. The interphase plant microtubules grow at 5 micro m/min, shrink at 20 micro m/min, and display catastrophe and rescue frequencies of 0.02 and 0.08 events/s, respectively, exhibiting faster turnover than their mammalian counterparts. Strikingly, during preprophase band formation, the growth rate and catastrophe frequency of plant microtubules double, whereas the shrinkage rate and rescue frequency remain unchanged, making microtubules shorter and more dynamic. Using these novel insights and four-dimensional time-lapse imaging data, we propose a model that can explain the mechanism by which changes in microtubule dynamic instability drive the dramatic rearrangements of microtubules during preprophase band and spindle formation in plant cells.     

ATP-dependent dissociation of non-disulfide-linked aggregates of coagulation factor VIII is a rate-limiting step for secretion

Deficiency in coagulation factor VIII leads to the bleeding disorder hemophilia A. Previous studies demonstrated that factor VIII secretion is limited due to an ATP-requiring step early in the secretory pathway. In this report, we identified that this ATP-dependent rate-limiting step involves the dissociation of non-disulfide-linked aggregates within the endoplasmic reticulum (ER). In contrast to the numerous examples of interchain disulfide-linked aggregates, factor VIII is the first protein characterized to form non-disulfide-linked high molecular weight aggregates within the ER. Approximately a third of newly synthesized factor VIII was detected in high molecular weight aggregates. These aggregates disappeared over time as functional factor VIII appeared in the medium. The aggregated complexes did not require proteasomal degradation for clearance. Aggregate formation was enhanced by ATP depletion, and upon restoration of metabolic energy, these aggregates were dissociated and secreted. With the coexpression of von Willebrand factor (vWF), a small portion of vWF coaggregated with factor VIII. However, vWF dissociated from the aggregates more rapidly than factor VIII, supporting that these aggregates are dynamic. An increase in the factor VIII expression level elicited a corresponding increase in the fraction of factor VIII that was aggregated. In addition, a 110 amino acid sequence containing a hydrophobic beta-sheet within factor VIII was identified that may predispose factor VIII to aggregation. These data show that formation and ATP-dependent dissolution of nondisulfide-linked factor VIII aggregates is a dynamic, rate-limiting step during the folding process in the early secretory pathway. In summary, we have identified an unprecedented requirement for protein transport out of the ER that involves an ATP-dependent dissociation of non-disulfide-linked aggregates within the ER.