Proteins expelled from the ER by dislocation are likely to be at least partially unfolded, although an assessment of their true conformational state remains an obvious challenge. Unless tight coupling exists between dislocation and degradation, these (partly) unfolded polypeptides are exposed to an aqueous environment and are prone to aggregation. The latter applies a fortiori
when considering membrane proteins such as Class I MHC products, one of the few examples for which there is evidence of a soluble cytoplasmic intermediate with its TM segment present and intact 
. We show here that the cytoplasmic chaperone BAT3 is recruited to Derlin2 at the ER membrane, can engage ER glycoproteins that are subject to dislocation, and that BAT3 is required for dislocation of TCRα.
A disruption of proteasomal degradation not only halts breakdown of ER-derived proteins, but often leads to their accumulation within the ER lumen. This is complemented by the requirement for poly-ubiquitylation to complete the dislocation reaction, giving rise to the idea that dislocation of a substrate is mostly tightly coupled to its degradation. Examples exist where such coupling is undone, for instance the Human Cytomegalovirus proteins US11 and US2, both of which funnel Class I MHC heavy chains into a dislocation pathway, and result in accumulation of the heavy chain in the cytoplasm when proteasomal proteolysis is blocked. We showed that in the presence of the EBV-DUB, ER-derived glycoproteins accumulate in the cytoplasm but are not degraded 
. Even though these ER proteins accumulate in the cytoplasm, it is unclear how they are maintained in a soluble state once there.
The cytoplasm contains folding machinery that prevents protein aggregation. HSC70 and HSP90, and their heat-shock induced counterparts, are examples of chaperones that engage (partly) unfolded polypeptides in the cytoplasm. Even though these chaperones have been linked to ER protein maturation and dislocation (reviewed in 
), particularly for the example of HSC70 and its involvement in dislocation of the misfolded cystic fibrosis transmembrane conductance regulator ΔF508 
, it remains to be established whether they play a direct role in dislocation and bridge the gap between the chaperone-rich ER lumen and the proteasome.
BAT3 has a role in the degradation of defective polypeptides synthesized at the ribosome, and may shuttle such substrates to the proteasome 
. We now show that BAT3 engages defective ER proteins discharged from the ER, and that BAT3 and its interactors accumulate as a complex when proteasomeal targeting is blocked by expression of the EBV-DUB.
Derlin2 is a small ER membrane protein. Members of the Derlin family have all been implicated in dislocation and may form part of a putative channel (dislocon) that facilitates the passage of misfolded ER proteins to the cytoplasm 
. We now find that BAT3 interacts with Derlin2 and associates with at least a select set of misfolded proteins. BAT3 is at the nucleus of a protein complex that shields the hydrophobic transmembrane domain of tail-anchored (TA) proteins after ribosomal synthesis and prior to engagement by a dedicated ER-targeting chaperone 
. BAT3 could perform a similar task for unfolded ER proteins at a dislocation channel, especially when such substrates contain a hydrophobic transmembrane stretch. Similarly, BAT3 was found to sequester mis-targeted proteins from the translocon 
Upon depletion of BAT3, the dislocation substrate TCRα accumulates primarily inside the ER lumen (). As BAT3 can function as a co-chaperone recruiter, the interactions we observe most likely occur in the context of a multi-protein complex. In fact, the BAT3-nucleated complex responsible for capturing TA proteins fresh off the ribosome has recently been implicated in protein dislocation 
Even though Derlin2 occurs in a complex with both Derlin1 and SEL1L 
, we retrieve BAT3 exclusively with Derlin2 in our immunoprecipitation experiments (). The choice of detergent may contribute to this selectivity: BAT3 may well associate with other components of the dislocation machinery, or the complex may simply not be sufficiently stable to allow retrieval of all of its constituents.
There are two moments during the dislocation process at which a chaperone like BAT3 could be of importance. The first moment is after complete extraction of the polypeptide by p97, but prior to its degradation. We can capture dislocation intermediates at this stage by expression of the EBV-DUB and observe increased association of TCRα with BAT3 under these conditions. In addition, BAT3 associates with the proteasome 
, opening the possibility that it shuttles dislocated proteins to the proteasome, akin to proteasomal shuttling factors such as RAD23 
. The second moment is prior to substrate engagement by p97. P97 can engage a protein ready for extraction through its cofactors UFD1 and NPL4, both of which recruit p97 to poly-ubiquitylated proteins 
. However, the active domains of the E2 and E3 enzymes responsible for substrate ubiquitylation face the cytoplasmic side of the ER. Ubiquitin conjugation to the polypeptide chain thus requires the substrate to at least peek out of the ER prior to engagement of p97. This window is then extended, as substrates require de-ubiquitylation prior to their full extraction from the ER so they may pass through the central pore of p97 
. It is at this moment that chaperone activity could be required to allow a smooth exit from the ER and prevent a blockade induced by exposed hydrophobic domains.