Activation of the UPR sensor ATF6α is regulated by its redox-dependent dimerization and ER retention by ERp18

ER Stress Promotes Redox-Dependent Formation of a Hyperglycosylated ATF6α C467 Dimer.

The ATF6α luminal domain contains two cysteine residues (C467 and C618) that either form an intrachain disulfide (C467-618) in the monomer or interchain disulfide bonded dimers (C467-467 or C618-618) (16) (Fig. 1A). To investigate any changes to the redox status of ATF6α during ER stress, we used a previously characterized HEK293T cell line stably expressing epitope tagged ATF6α containing a HA-tag in the cytosolic and a V5-tag in the luminal domain (16). To prevent ATF6α cleavage by S1P during ER stress, cells were treated with an S1P inhibitor (S1Pi) (23) before treatment with various stressors. We have shown previously that S1P inhibition prior to induction of ER stress prevents ATF6α cleavage allowing it to accumulate in the Golgi and become hyperglycosylated (16). Initially, we analyzed the redox status of ATF6α after chemotoxic stress induced by thapsigargin (TG). When cell lysates were separated under nonreducing SDS/PAGE conditions and immunoblotted with anti-HA, we detected three bands: ATF6α monomer (designated M) and two dimers, designated 618D and 467D (Fig. 1A), as reported previously (16, 21). In unstressed cells, ATF6α consisted mainly of the monomer (Fig. 1A, lane 1). However, following ER stress, the intensity of the ATF6α monomer becomes significantly reduced, while the intensity of the ATF6α 467D increased to become the predominant form (Fig. 1A, lane 2). The ATF6α 618D remained mostly unchanged. Furthermore, in the TG-treated cells, we detected a slower-migrating band above the ATF6α 467D, (indicated with an asterisk in Fig. 1A), which was absent in unstressed cells. This band indicates further glycosylation modification of ATF6α 467D, consistent with Golgi-localization. We found the same redox change and hyperglycosylation with endogenous ATF6α visualized with an ATF6α-specific antibody (SI Appendix, Fig. S1) even at 10-fold lower concentrations of added TG, indicating the effects were not specific to ectopically expressed protein or to the TG concentration used to induce the UPR.

Previous reports have shown that hyperglycosylation of ATF6α occurs during ER stress, which results in reduced SDS/PAGE mobility (15, 16). We carried out a time course of TG treatment to follow the kinetics of accumulation of ATF6α 467D and its subsequent hyperglycosylation (Fig. 1B). After 30 min, we observed a reduction in the intensity of the ATF6α monomer and an increase of the ATF6α 467D, including the appearance of the hyperglycosylated ATF6α 467D (Fig. 1B, lane 2). Additional faint bands were detected between the 467D, 618D, and monomer with increased intensity after TG treatment (depicted here and in subsequent experiments where apparent with arrowheads). These varied in intensity between experiments, were particularly apparent when the V5-antibody was used, and were seen in the presence of S1Pi, suggesting they are not generated due to S1P cleavage. They may represent alternate redox forms of ATF6α and require further investigation for their identification. Over 70% of the ATF6α detected after 30 min of TG treatment was ATF6α 467D (Fig. 1C), and this level remained unchanged for the duration of the time course (Fig. 1B, lanes 2 to 5, and Fig. 1C). The intensities of the ATF6α 618D remained mostly unaffected (Fig. 1B, lanes 1 to 5, and Fig. 1C). These results demonstrate that the redox status of ATF6α changes following chemotoxic stress, consistent with a reduction of the C467-C618 disulfide in the monomer and formation of a C467-467 disulfide bonded dimer.

To examine the redox changes to ATF6α 467D following ER stress induced by different reagents, we carried out time courses in the presence of dithiothreitol (DTT), which disrupts protein disulfide formation, or tunicamycin, a glycosylation inhibitor. After 30 min of DTT treatment, we did not detect any significant changes in the levels of ATF6α monomer and 467D; however, we detected the appearance of the hyperglycosylated form of 467D (Fig. 1D, lane 2). The hyperglycosylated ATF6α 467D reached its peak after 90 min and was still detectable after 120 min (Fig. 1D, lanes 4 and 5). The apparent lack of an increase in the levels of ATF6α 467D after 30 min of DTT treatment might be due to reduction of the interchain disulfide by DTT itself or the rapid conversion of ATF6α 467D to the hyperglycosylated form. At higher concentrations, DTT can reduce the ATF6α intra- and intermolecular disulfides (21) (SI Appendix, Fig. S2). Here we used a relatively low concentration of DTT to minimize ATF6α disulfide reduction but still allow a robust UPR. Even in the presence of the reducing agent, the main species that becomes hyperglycosylated is 467D.

In tunicamycin-treated cells, we also detected a decrease in the ATF6α monomer with a concomitant increase in the 467D, although the accumulation of 467D in these cells exhibited a slower appearance compared to the TG-treated cells (compare Fig. 1 B and E). This is consistent with tunicamycin displaying a slower kinetics of activation of ATF6α compared with TG (21). We also detected unglycosylated ATF6α (ung) after 90 and 120 min (Fig. 1E, lanes 4 and 5), consistent with tunicamycin inhibition (21, 24). At the later time point, some hyperglycosylated 467D is apparent (Fig. 1E, lane 5). Hence, for the chemotoxic agents studied that induce ER stress there is a specific hyperglycosylation of the 467D, suggesting that this form is selectively transported to the Golgi.

The commonly used UPR inducers, DTT, TG, and tunicamycin exert multiple effects on cellular homeostasis, in addition to activating the UPR pathways. It is apparent that chemical induction of the UPR confers different genetic and protein expression signatures in comparison to the UPR induced by expression of misfolded proteins in the ER (25). To investigate changes to the redox status of ATF6α under proteotoxic induction of the UPR, we generated a Flp-In T-Rex doxycycline-inducible HeLa cell line for expression of HA-tagged mutant lipoprotein receptor (LDLR), HA-LDLR C646Y. This mutation causes a defect in LDLR folding and trafficking and consequently the misfolded protein accumulates in the ER (26, 27). To investigate if expression of misfolded HA-LDLR C646Y would elicit a UPR response, cells were treated with doxycycline for various times and HA-tagged LDLR was detected by immunoblot using anti-HA. Expression of HA-LDLR was seen after 2 h, with a steady increase over the time course (SI Appendix, Fig. S3). Immunoblot analyses of endogenous ATF6α in these cells revealed the presence of two forms, monomer and 467D, with the 618D, barely detectable, mirroring that seen with exogenously expressed ATF6α (Fig. 1 A and F). Interestingly, there was a shift in the redox status of ATF6α from predominantly the monomer to the 467 dimer, with maximal levels of the dimer seen at 8 h (Fig. 1F, lane 5). We also detected some hyperglycosylation of the 467D at later time points (Fig. 1F, lanes 4 and 5). Note the time course of induction was carried out in the presence of S1Pi to prevent ATF6α processing. The change from the monomer to the 467D coincides with the increase in expression of the misfolded HA-LDLR (SI Appendix, Fig. S3), suggesting that accumulation of this protein in the ER is sufficient to induce a change in the redox status of ATF6α. The expression of misfolded LDLR activated the IRE1 pathway, as indicated by the splicing of XBP1-u to produce XBP1-s, thereby confirming a UPR response (SI Appendix, Fig. S4). In addition, levels of BiP increased after induction of misfolded LDLR in the absence of S1Pi (SI Appendix, Fig. S3A). This increase was attenuated in the presence of S1Pi (SI Appendix, Fig. S3B). Importantly, the predominant form of ATF6α during ER stress produced by proteotoxic stress is the 467D, which becomes hyperglycosylated indicating selective transport to the Golgi. These results are consistent with our finding with chemotoxic stressors and confirm that the redox changes seen with exogenously expressed ATF6α also occur with endogenous ATF6α.

ATF6α 467D Is Preferentially Trafficked to the Golgi.

The ER stress induced hyperglycosylation of ATF6α 467D would suggest that it is transported from the ER to the Golgi during the UPR. To further elucidate the location of ATF6α 467D hyperglycosylation, we analyzed the sensitivity of ATF6α monomer and dimers to endoglycosidase H (endo H) digestion. Cells pretreated with S1Pi were left untreated or treated with TG, to induce ER stress and cell lysates incubated with or without endo H. In unstressed cells, ATF6α monomer and dimers were sensitive to endo H digestion, as demonstrated by their increased SDS/PAGE mobilities (Fig. 2A, lane 2), indicating ER localization. Only the hyperglycosylated C467 dimeric form generated after ER stress was resistant to endo H digestion, consistent with trafficking to the Golgi (compare lanes 3 and 4, band with asterisk, in Fig. 2A) (15).

To determine if hyperglycosylated ATF6α 467D is O-linked glycosylated, we treated cells with the O-linked glycosylation inhibitor, benzyl-2-acetamido-2-deoxy-α-d-galactopyranoside (BADAP) (28, 29). Cells were pretreated with S1Pi, and then with 10 mM DTT for 30 min. We favored a high DTT concentration as the stressor for this experiment since it causes a robust activation of ATF6α and results in a maximal build-up of hyperglycosylated ATF6α (16) (SI Appendix, Fig. S2). Treatment with BADAP resulted in increased SDS/PAGE mobilities of the hyperglycosylated ATF6α before or after endo H treatment suggesting inhibition of O-linked glycosylation (Fig. 2B, lanes 1 to 4). Additionally, whereas hyperglycosylated ATF6α was resistant to endo H digestion (Fig. 2B, lanes 1 and 3), incubation with PNGase resulted in increased mobility due to removal of N-linked oligosaccharides (Fig. 2B, lanes 1 and 5). Interestingly, treatment with BADAP followed by PNGase caused a further increased mobility of hyperglycosylated ATF6α (Fig. 2B, lane 6). Together, these results demonstrate the reduced mobility form of ATF6α has modifications to the N-linked glycans conferring endo H resistance as well as O-linked glycosylation that can be inhibited by BADAP.

Previously, it was suggested that only monomeric ATF6α traffics to the Golgi based upon stress induction with DTT or tunicamycin, followed by fractionation of the ER and Golgi by discontinuous sucrose gradient. We also fractionated lysates from unstressed and stressed cells in a discontinuous sucrose density gradient, to partially separate ER and Golgi components. In unstressed cells, all forms of ATF6α were detected in the heavier fractions (Fig. 2C, lanes 1 to 5), consistent with its ER localization. Western blotting demonstrated the presence of the ER membrane protein, ribophorin, in fractions 1 to 5, confirming the presence of ER components in the heavier fractions. GM130, used as a Golgi marker, partially overlaps with ribophorin, but was detected primarily in the lighter fractions (Fig. 2C, lanes 4 to 9). In the TG-treated cells, we also detected ATF6α monomer and dimers in the heavier ER fractions (Fig. 2D, lanes 1 to 5). The hyperglycosylated 467D was predominantly found in lighter fractions (Fig. 2D, lanes 4 and 5). Note the presence of a cleaved from of ATF6α (ATF6-P) that migrates at about 75 kDa that we have previously reported and is often seen after treatment with the S1Pi (16). Sucrose gradient fractionation of cells treated with tunicamycin also showed that hyperglycosylated ATF6α 467D transitioned to lighter fractions (SI Appendix, Fig. S5, lanes 4 and 5). These results indicate stress-induced trafficking of ATF6α 467D from ER fractions to lighter fractions consistent with the early Golgi compartment.

The trafficking of ATF6α 467D from the ER to Golgi should result in the formation of a disulfide bonded luminal domain dimer following cleavage by S1P. To determine if this was the case, we immunoblotted with anti-V5, cell lysates from cells treated with TG in the absence of the S1Pi (Fig. 3). The ATF6α construct contains a V5-epitope at the C terminus allowing detection of any cleaved luminal domain. After 30 min of TG treatment, a V5 reactive band was detected migrating at about 90 kDa (Fig. 3A). When cells were pretreated with ammonium chloride, before treatment with TG, the intensity of the ATF6α luminal domain dimer was increased (Fig. 3B and SI Appendix, Fig. S6A), suggesting that inhibition of lysosomal degradation (30, 31) stabilizes the cleaved ATF6α luminal domain. Similar stabilization of the luminal domain was seen in the presence of another lysosomal degradation inhibitor, Bafilomycin A1 (SI Appendix, Fig. S6A). This band was absent when the samples were immunoblotted with an anti-HA antibody (SI Appendix, Fig. S6B) whose epitope is located at the N-terminal cytosolic region of ATF6α. We confirmed this band to be a result of cleavage by S1P, as it was undetected in the presence of S1Pi (Fig. 3 C and D and SI Appendix, Fig. S6A) and migrated at about 45 kDa when samples were analyzed under reducing SDS/PAGE (Fig. 3C). No monomeric luminal domain was detected during the time course of TG treatment (Fig. 3 A and B), suggesting that dimerization occurs prior to Golgi trafficking and cleavage. These results demonstrate that ATF6α 467D is preferentially trafficked to the Golgi during stress and suggest that the released ATF6α luminal domain localizes to the lysosome for degradation.

Role of Interchain Disulfide Formation in ATF6 Trafficking.

To determine the potential role of the interchain disulfides within the ATF6α luminal domain in activation and trafficking, we generated stable cells individually expressing ATF6α C467A and C618A mutants in a HT1080 fibrosarcoma cell line where the ATF6α gene was deleted. We used cells lacking endogenous ATF6α for these experiments to exclude the possibility of dimerization between endogenous and exogenously expressed ATF6α, which could lead to artifactual results. We confirmed the knockout (KO) of endogenous ATF6α in these cells (SI Appendix, Fig. S7) by immunoblot analyses (SI Appendix, Fig. S7A) and by comparing the induction of BiP expression upon TG-induced ER stress with wild-type cells (SI Appendix, Fig. S7 B and C). Cell lysates from either the C467A or C618A cells were treated with or without TG and separated under nonreducing SDS/PAGE (Fig. 4A). Immunoblotting with anti-ATF6α revealed the presence of monomers and dimers in unstressed cells (Fig. 4A, lanes 1 and 4), with the expected mobilities of the dimeric forms as previously reported (16, 21). Treatment with TG led to hyperglycosylation of the 467D but not the 618D (Fig. 4A, lanes 3 and 6). No hyperglycosylation was seen after TG treatment in the absence of S1Pi (Fig. 4A, lanes 2 and 5), indicating that trafficked ATF6α was rapidly cleaved in the Golgi. These results suggest that for the wild-type ATF6α and C618A mutant, it is the 467D form that is preferentially transported to the Golgi during ER stress where it becomes hyperglycosylated.

To identify which dimeric form of ATF6α gives rise to the luminal domain dimer, we transiently transfected the ATF6α KO cell line with wild-type, the C618A or C467A mutants of ATF6α, and determined the presence of the luminal domain dimer after ER stress and stabilization with ammonium chloride (Fig. 4B). The luminal domain dimer was only observed with the wild-type and C618A mutant after ER stress (Fig. 4B, lanes 2 and 4), but not the C467A mutant, though a small amount of monomeric luminal domain was observed (Fig. 4B, lane 6). These experiments with the cysteine mutants of ATF6α, further substantiate that 467D predominates after ER stress and it is this form of ATF6 that is preferentially transported to the Golgi.

To determine the functional consequence of this preferential transport, we also determined whether preventing formation of 467D would affect ATF6α cleavage or transcriptional activation. The cleavage of wild-type and the cysteine mutants of ATF6 was determined by preparing a nuclear extract and evaluating the appearance of the cleaved nuclear form of ATF6 (ATF6-N) (Fig. 4 C and D). For these experiments we used stable cell lines expressing exogenous ATF6α in the ATF6-KO cell line. Upon TG activation, ATF6-N was formed after 60 min with both the wild-type (Fig. 4C) and the C618A mutant, but not the C467A mutant (Fig. 4D). In addition, there was an attenuation of induction of the ATF6α target BiP with the C467A mutant, but the induction of expression of BiP with the C618A mutant was equivalent to that seen with the wild-type after TG treatment (Fig. 4E). The presence of ATF6-N and induction of BiP verifies trafficking and cleavage of the 467D in the Golgi, confirming the results with hyperglycosylation and sucrose gradient fractionation.

Specificity of S1P for the Redox Forms of ATF6.

The presence of the S1P-generated luminal domain dimer with wild-type and C618A mutant, but not the C467A mutant of ATF6α, could be caused by the inability of S1P to cleave the 618D. To determine whether this was the case, we treated cells with brefeldin A and determined the appearance of the ATF6α luminal domain before or after ER stress (Fig. 5). Brefeldin A causes a merger of ER and Golgi components (32) and results in the cleavage of ATF6α by S1P and S2P in the absence of additional ER stress (16, 21). Lysates from brefeldin A-treated cells were separated under nonreducing SDS/PAGE followed by V5 immunoblot analysis. In wild-type ATF6α cells (Fig. 5A), treatment with brefeldin A alone did not result in any detectable cleavage product of ATF6α (Fig. 5A, lane 2). However, in the presence of TG and brefeldin A, we detected a dimeric ATF6α luminal domain (Fig. 5A, lane 4). The ATF6α luminal domain was absent in cells treated with S1Pi prior to brefeldin A and TG treatments (Fig. 5A, lane 3), confirming that its appearance was dependent upon S1P activity. The fact that brefeldin A treatment alone did not yield any detectable monomeric ATF6α luminal domain would suggest that the monomer present in these conditions is resistant to cleavage. TG treatment led to a switch from an intra- to an interchain disulfide resulting in enhanced S1P cleavage of ATF6α. In contrast, in the ATF6α C618A cells, treatment with brefeldin A alone resulted in a cleavage product of about 45 kDa, consistent with the ATF6α luminal domain monomer (Fig. 5B, lane 2). No intrachain disulfide can form with this mutant indicating the ability of S1P to cleave the reduced monomer. Interestingly, treatment of these cells with TG and brefeldin A led to the production of both monomer and dimeric forms of the ATF6α luminal domain (Fig. 5B, lane 4); these cleavage products were undetected in the presence of the S1Pi (Fig. 5B, lane 3). This result confirms that the 467D can be cleaved by S1P. In the ATF6α C467A cells, treatment with either brefeldin alone or in combination with TG resulted in the production of a monomeric ATF6α luminal domain (Fig. 5C, lanes 2 and 4), suggesting that the 618D is not a substrate for S1P. These results indicate that the redox status of ATF6α regulates the specificity of S1P, which preferentially cleaves the 467D or reduced monomer.

ERp18 but Not PDIR Reduces the C467 Dimer and Promotes Retention of ATF6α in the ER.

We have previously shown that ERp18 and protein disulfide isomerase-related (PDIR) are catalytically active toward ATF6α (16). We demonstrated that mutants of ERp18 or PDIR lacking the second cysteine in the CXXC active site can form mixed disulfides with ATF6α, demonstrating their ability to catalyze disulfide reduction (33). PDIR also has been implicated in the process of ATF6α activation (18), and is known to catalyze reduction of ATF6 disulfides in the absence of ER stress (16). As ATF6α switched its redox status following stress from the monomer to the 418D, we assessed here the specificity of these two enzymes toward either the 467-467 or 618-618 disulfide. When the substrate-trapping mutant of PDIR (Fig. 6A) was expressed in the presence of the C467A or C618A ATF6α mutants, mixed disulfides were seen with the C467A but not the C618A mutant (Fig. 6B, lanes 1 and 3). In contrast, the ERp18 substrate-trapping mutant formed mixed disulfides with both ATF6α mutants (Fig. 6B, lanes 2 and 4). These results indicate that ERp18 can reduce both disulfides, whereas PDIR specifically reduces the 618-618 disulfide. Indeed, we have previously shown that during ER stress, ERp18 can catalyze the reduction of the 467-467 disulfide indicating this enzyme antagonizes 467D formation.

Given the ability of ERp18 to reduce the 467-467 disulfide, we questioned whether overexpression of this oxidoreductase would lead to an attenuation of ATF6α activation. Our previous work had highlighted the role of ERp18 in regulating trafficking of ATF6α (16). Deletion of ERp18 resulted in faster trafficking of ATF6α due to a physical interaction between these two proteins following ER stress, indicating a role for ERp18 in retention. Therefore, we investigated the consequence of overexpression of ERp18 on both dimer formation, Golgi trafficking, S1P cleavage, and BiP induction. We created a stable cell line expressing ATF6α as well as overexpressing ERp18 (Fig. 6C). When a UPR was induced in these cells, the formation of hyperglycosylated ATF6α 467D was dramatically reduced (Fig. 6D, compare lanes 2 and 4). In addition, the appearance of ATF6-N was attenuated in this cell line when compared to wild-type (Fig. 6 E and F, compare 60-min time points). Finally, fold-induction of BiP following activation was reduced upon overexpression of ERp18 (Fig. 6G). Taken together, these results indicate a reduction of ATF6α trafficking to the Golgi and processing to generate an active transcription factor following overexpression of ERp18.

Cell Lines.

All HEK293T, HT1080, and the Flp-In T-Rex doxycycline-inducible HeLa cells were maintained in DMEM supplemented with 10% FBS and 100 units/mL penicillin and 100 µg/mL streptomycin at 37 °C in a 5% CO₂ incubator.

DNA Constructs.

The pcDNA 3.1-ATF6α plasmid containing an N-terminal HA-tag and a C-terminal V5-tag followed by a KDEL sequence and stop codon, and the single cysteine to alanine mutations in the luminal domain of ATF6α (ATF6 C467A and ATF6 C618A) have been described previously (16). The pcDNA 3.1/HA-LDLR C646Y has been described previously (27). The human PDIR CXXA substrate-trapping mutant, ERp18 wild-type and ERp18 CXXA substrate-trapping mutant have been described previously (27, 33).

Cell Culture and Transfections.

HEK293T and HT1080 were either singly or cotransfected with plasmid DNA at 80 to 90% confluence using MegaTran (Origene) transfection reagent. To generate stable overexpression cells, the transfected cells were placed on antibiotic selection for ∼3 to 4 wk until colonies appeared. Screening for positive clones was carried out by Western blotting.

Cell Lysis, Immunoisolation, and Western Blots.

Cells were harvested by treatment with trypsin-EDTA (Gibco – 25300-054) or by scrapping using a rubber policeman (Greiner bio-one, Cat. No. 541 080). They were collected by centrifugation at 250 × g for 5 min and then washed twice with ice-cold PBS (Gibco). The cells were then resuspended in cell lysis buffer (CLB) [1% (vol/vol) Triton X-100, 50 mM Tris⋅HCl, pH 7.4, 150 mM NaCl, 2 mM ethylenediamine tetra-acetic acid (EDTA) and 0.5 mM phenylmethylsulphonyl fluoride (PMSF)], incubated on ice for 10 min, followed by centrifugation at 16,100 × g to obtain the postnuclear supernatant. Prior to immunoisolation the postnuclear supernatant was precleared by incubating with protein A Sepharose beads (Generon) for 30 min at 4 °C. The mixture was then cleared by centrifugation at 16,100 × g for 1 min and the supernatant incubated with protein A Sepharose beads and the appropriate antibody for 16 h at 4 °C. Immunoisolates were washed three times in CLB. Samples were then boiled at 95 °C for 5 min in 2× Laemmli sample buffer (200 mM Tris-Cl, pH 6.8, 3% SDS, 10% glycerol, 1 mM EDTA, and 0.004% Bromophenol blue) prior to SDS/PAGE either under reducing (treated with 50 mM DTT) or nonreducing conditions.

For Western blotting, proteins were transferred to nitrocellulose membrane (Li-Cor Biosciences), which were blocked in 5% (wt/vol) nonfat dried skimmed milk in TBST [Tris-buffered saline with Tween 20: 10 mM Tris, 150 mM NaCl, pH 7.5, and 0.1% (vol/vol) Tween 20] for at least 60 min. Primary antibodies were diluted in TBST and incubations were carried out for ∼16 h at either 4 °C or at room temperature. Li-Cor IRDYE fluorescent secondary antibodies were used for detection, typically at 1:5,000 dilutions. Blots were scanned using an Odyssey SA imaging system (Li-Cor Biosciences).

Determination of ATF6α Redox Status and Trafficking during ER Stress.

Confluent HEK/ATF6α cells were pretreated with 30 µM PF429242 (S1Pi) for 60 min before treatment with various ER stressors including DTT (at either 1 mM or 10 mM), 5 µM TG, or 5 µg/mL tunicamycin. Whole-cell lysates were subjected to nonreducing SDS/PAGE, followed by Western blotting. Quantification of band intensities were carried out with the LiCor Image Studio v4 software and histograms prepared using Origin 6.0 professional software.

Generation of Flp-In T-Rex Doxycycline-Inducible HeLa Cell Line.

The Flp-In T-Rex doxycycline-inducible HeLa/HA-LDLR C646Y cell line was generated by Invitrogen Flip-in T-REx kit according to the manufacturer’s protocol. Briefly, pOG44 and pcDNA3.1/HA-LDLR vector were cotransfected into the cells and selected for by hygromycin B (200 μg/mL) and blasticidin (4 μg/mL). The successfully transfected cells were induced by 2 μg/mL doxycyclin.

Analyses of ER to Golgi Trafficking of ATF6–O-glycosylation Status.

Cells were treated with 30 µM S1Pi, and either left untreated or treated with 2 mM BADAP for 60 min. The cells were then treated with 10 mM DTT for 30 min. Whole-cell lysates were prepared and incubated either without or with endoH and Peptide: N-glycosidase F (PNGase). Samples were separated under reducing SDS/PAGE conditions and ATF6α detected using rabbit anti-HA.

Sucrose Gradient Fractionation of ATF6α.

HEK/ATF6 cells in 15-cm dishes were either untreated or treated with the S1Pi and ER stressors (5 µM TG for 90 min, 5 µg/mL TM for 2 h). Cells were harvested and washed twice in 1 mL ice-cold PBS/NEM (20 mM NEM, buffer supplemented with EDTA-free protease inhibitor tablet), followed by centrifugation at 1,400 rpm for 5 min. PBS was completely removed, and cell pellet was resuspended in 400 µL of ice-cold buffer A (10 mM Hepes pH 7.4, 250 mM Sucrose, 10 mM KCl, 1.5 mM MgCl₂, 1 mM EDTA, and 1 mM EGTA) supplemented with protease inhibitor and left on ice for 10 min. Cells were lysed by passing through a 23-gauge needle (Fine-Ject 23G, Henke Sass Wolf) attached to a 1-mL syringe, 20 times, followed by centrifugation at 1,000 × g, for 7 min at 4 °C. Next, 200 µL of the resulting supernatant were laid on top a discontinuous sucrose gradient consisting of eight layers, each containing 200 µL of distinct sucrose densities of 20%, 25%, 30%, 35%, 40%, 45%, 50%, and 60%, from top to bottom, in a polycarbonate tubes (11 × 34 mm; Beckman Coulter). The samples were centrifuged at 214,000 × g for 60 min at 4 °C, in a TLS 55 rotor using a Beckman Coulter Optima MAX-XP ultracentifuge. Nine fractions of 200 µL each were carefully taken from the top using a pipette. Samples were analyzed on either reducing or nonreducing SDS/PAGE, followed by Western blotting.

Analyses of ATF6α Cleavage Products.

Cells in 6-cm dishes were either left untreated (−) or treated (+) with either 30 µM S1Pi (60 min), 20 mM ammonium chloride (45 min), 100 nM bafilomycin A1 (4 h), or 5 µg/mL brefeldin A (60 min) treatment, before addition of 5 µM TG to induce ER stress. After treatment, whole-cell lysates were prepared and separated under reducing or nonreducing SDS/PAGE, followed by Western blotting.

ER Stress-Induced Cleavage of ATF6α.

Wild-type or ERp18 KO HEK293T/ATF6α overexpression cells in 6-cm dishes were either left untreated (−) or treated (+) for at least 1 h with 30 µM S1Pi, before addition of 10 mM DTT or 5 µM TG to induce ER stress. For the brefeldin A (5 µg/mL) treatment, cells were incubated with the drug for at least 1 h prior to harvest.

Isolation and analyses of nuclear ATF6α, ATF6-N, has been described previously (16, 37). Briefly, cells in 6-cm dishes were washed once with 3 mL PBS supplemented with protease inhibitor tablet. Cells were then scraped off the dish in 2.9 mL ice-cold PBS, using a rubber policeman (cell scraper) and transferred into a 15-mL tube. Samples were centrifuged at 1,619 × g for 5 min, and supernatant completely removed. The cell pellets were resuspended in 1 mL ice-cold buffer A (10 mM Hepes pH 7.4, 250 mM sucrose, 10 mM KCl, 1.5 mM MgCl₂, 1 mM EDTA, and 1 mM EGTA) supplemented with protease inhibitor, transferred to 1.5-mL tube, and left on ice for 10 min. Cells were lysed by passing through a 23-gauge needle attached to a 1-mL syringe, 20 times, followed by centrifugation at 1,000 × g, for 7 min at 4 °C to obtain the nuclear pellet. The pellet was then resuspended in 100 μL buffer B (10 mM Hepes pH 7.6, 2.5% Glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 1 mM EDTA, and 1 mM EGTA) supplemented with protease inhibitor, followed by incubation for 60 min on a 4 °C rocker. The samples were centrifuged at 100,000 × g for 30 min, 4 °C, to obtain a supernatant containing the nuclear extract. The samples were separated on reducing SDS/PAGE gels, followed by Western blotting.

Analyses of BiP Induction during ER Stress.

Confluent cells in 6-cm dishes were either left untreated or treated with 1 µM TG for 16 h. Whole-cell lysates were subjected to reducing SDS/PAGE, followed by Western blotting. Quantification of band intensities were carried out with the LiCor Image Studio v4 software and histograms prepared using Origin 6.0 professional software.

Deglycosylation Assays.

HEK/ATF6 wild-type cells were harvested and washed in PBS then lysed in CLB. Endo H and PNGase reactions were set up according to the manufacturer’s protocol (New England Biolabs). The digests were carried out overnight at 37 °C and samples were resolved on either reducing or nonreducing SDS/PAGE.

XBP-I Splicing Assay.

Total RNA from TG-treated cells was extracted using TRIzol Reagent (Ambion), according to the manufacturer’s protocol. RT-PCR was carried out using the SuperScript II Reverse Transcriptase (Invitrogen) with oligo dTs (Invitrogen), according to the manufacturer’s protocol. cDNA for endogenous XBP1 was amplified using the following primers pairs: 5′- GAATGAAGTGAGGCCAGTGG -3′, 5′- GGGGCTTCCTATATATGTGG -3′. Actin cDNA was amplified using the following primer pairs: 5′-CCACACCTTCTACAATGAGC-3′, 5′-ACTCCTGCTTGCTGATCCAC-3′. PCR parameters include an initial denaturation step at 95 °C for 5 min; then 35 cycles of 95 °C for 45 s, an annealing step of 60 °C for 45 s, and 72 °C for 45 s; followed by a final elongation step of 72 °C for 10 min.

Alt-R CRISPR-Cas 9 KO of ATF6.

Using the Integrated DNA Technologies (IDT) Alt-R CRISPR-Cas 9 System, the predesigned guide RNA (gRNA) Hs.Cas9.ATF6.1.AB was selected. We used a single gRNA (cRNA) duplexed with a tracrRNA (IDT, 1072532) and Alt-R S.p. HiFi Cas9 Nuclease V3 (IDT, 1081060) for genome editing. To generate an ATF6 KO, Cas9 nuclease was added to a duplex of 1AB crRNA:tracrRNA in Optimem to form the RNP complex (Fisher, 31985062), which was then transfected into HT1080 cells using Lipofectamine CRISPRMAX transfection reagent (Thermo Fisher, CMAX00001). After 48 h, cells were harvested for genomic DNA using QuickExtract DNA Extraction Solution (Lucigen, QE0905T). Genomic PCR using primers designed for Exon 1 of ATF6α was performed to create template DNA for the T7 Endonuclease assay (New England Biolabs, M0302s). Briefly, the wild-type and 1.AB DNA were hybridized slowly and then digested using T7 endonuclease, to detect any cleavage events by mismatching of the DNA via genome editing events. Samples were analyzed on an agarose gel and by DNA sequencing. After cleavage was confirmed, cells were transferred to 15-cm dishes until colonies appeared (about 10 to 12 d later). Positive ATF6α KO cells were identified by Western blotting using the mouse anti-ATF6 antibody.