Index Introduction Materials and Methods Cell culture and transfection Immunological techniques Metabolic labeling of ATF6 Discontinuous sucrose gradient centrifugation Results and Discussion Glucose starvation activates ATF6 ATF6 is reduced in response to glucose starvation and reduced monomer ATF6 is transported to the Golgi apparatus Acknowledgements References

Introduction

The endoplasmic reticulum (ER) provides an optimal environment for the folding and assembly of newly synthesized secretory and transmembrane proteins, and ensures the quality of these proteins by allowing the exit of correctly folded molecules to reach their final destination while retaining incorrectly folded molecules in the ER (Gething and Sambrook, 1992; Helenius et al., 1992). In addition, when excessive amounts of unfolded proteins accumulate in the ER, the ER transmits signals toward the cytoplasm and nucleus to cope with this so-called ER stress. This homeostatic signaling from the ER is collectively termed the unfolded protein response (UPR), which consists of translational and transcriptional control. In response to ER stress, translation is generally attenuated to decrease the burden on the ER, while transcription of a particular set of genes is induced to counter various consequences of ER stress. ER-localized molecular chaperones and folding enzymes (collectively referred to as ER chaperones hereafter) are thus major targets of the UPR on account of their ability to deal directly with unfolded proteins accumulated in the ER. Signaling from the ER is mediated by transmembrane proteins in the ER, the number of which has increased with evolution: one (Ire1p) in yeast cells and three (IRE1, PERK and ATF6) in mammalian cells (Harding et al., 2002; Mori, 2000; Patil and Walter, 2001; Schroder and Kaufman, 2005).

ATF6 is embedded in the ER as a type II transmembrane protein under normal conditions (Haze et al., 2001; Haze et al., 1999). When unfolded proteins accumulate in the ER, ATF6 is activated by regulated intramembrane proteolysis (Brown et al., 2000). Thus, ATF6 is transported to the Golgi apparatus via COPII vesicles to be cleaved sequentially by Site-1 and Site-2 proteases (Nadanaka et al., 2004; Okada et al., 2003; Shen et al., 2002; Ye et al., 2000). The cytoplasmic region liberated from the Golgi membrane contains all the domains necessary for the active transcription factor and enters the nucleus to activate transcription (Yoshida et al., 2000; Yoshida et al., 2001). This nuclear and active form is designated as pATF6(N), whereas the ER-localized precursor form is designated pATF6(P). Ectopic expression of pATF6(N) at a physiological level is sufficient to induce the transcription of various ER chaperones (Okada et al., 2002). These observations indicate that ATF6 is involved in all processes in signaling from the ER to the nucleus, beginning with sensing ER stress and culminating in activating transcription. Among these processes, the mechanism by which ER stress is sensed is largely unknown. It is thought that ATF6 is retained in the ER under normal conditions by binding to the major ER chaperone BiP, and that ER stress-induced dissociation of BiP triggers the exit of ATF6 from the ER to reach the Golgi apparatus (Shen et al., 2002).

We recently found that, owing to the presence of intra- and intermolecular disulfide bridges formed between the two conserved cysteine residues in the lumenal domain, ATF6 occurs in unstressed ER in monomer, dimer and oligomer forms (Nadanaka et al., 2007). Disulfide-bonded ATF6 is reduced on treatment of cells with not only the reducing reagent dithiothreitol but also the glycosylation inhibitor tunicamycin. The extent of reduction correlates with that of activation, hence dithiothreitol is a much better inducer of ATF6 cleavage than tunicamycin, although reduction alone is not sufficient for activation of ATF6. Importantly, only the reduced monomer ATF6 reaches the Golgi apparatus in dithiothreitol- or tunicamycin-treated cells, and this reduced monomer ATF6 is a better substrate than disulfide-bonded forms for Site-1 protease, the first cleaving enzyme. We thus proposed that this mechanism ensures the strictness of regulation, in that the cell can only process ATF6 which has experienced the changes in the ER (Nadanaka et al., 2007). However, because reduction was induced by chemical reagents in our previous study, it remained unclear whether reduction of ATF6 also occurs under more physiological conditions.

Here, we investigated whether ATF6 is reduced physiologically using glucose starvation on the basis that this stress occurs under a variety of circumstances in the living organism, and because it was the inducer used in the paper first reporting the UPR in the literature nearly 30 years ago (Shiu et al., 1977).

Materials and Methods

Cell culture and transfection

CHO cells were grown in complete medium [a 1:1 mixture of Ham’s F12 and Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, 2 mM glutamine and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin sulfate)] in a 5% CO₂, 95% air incubator at 37°C. Glucose starvation was induced by culturing cells in glucose-free medium [Dulbecco’s modified Eagle’s medium lacking glucose (GIBCO) but supplemented with 10% dialyzed fetal calf serum, 1 mM pyruvic acid and antibiotics]. Cells were transfected with plasmid DNA as described previously (Nadanaka et al., 2004) using Superfect (Qiagen) basically according to the manufacturer’s instructions and then incubated at 37°C for an appropriate time for expression of the transfected gene. The plasmids pCMVshort-EGFP-ATF6α and pCMVshort-EYFP-ATF6α(S1P-) were constructed previously (Nadanaka et al., 2004).

Immunological techniques

Immunoblotting analysis was carried out according to standard procedure (Sambrook et al., 1989) as described previously (Okada et al., 2002) using Western Blotting Luminol Reagent (Santa Cruz Biotechnology). Chemiluminescence was detected using an LAS-1000plus LuminoImage analyzer (Fuji Film). ATF6α was detected with rabbit anti-ATF6α polyclonal antibody (Haze et al., 1999). Goat anti-ribophorin I polyclonal antibody (C-15) was purchased from Santa Cruz. Mouse anti-GM130 monoclonal antibody was obtained from BD Transduction Laboratories. Anti-KDEL monoclonal antibody was purchased from Stressgen. A.v. peptide antibody against green fluorescent protein (GFP) was obtained from Clonetech. Mouse anti-actin monoclonal antibody was purchased from CHEMICON.

Immunoprecipitation and indirect immunofluorescence were carried out essentially as described previously (Nadanaka et al., 2004). Anti-GFP monoclonal antibody (mixture of clone 7.1 and 13.1) was purchased from Roche.

Metabolic labeling of ATF6

CHO cells grown in 60-mm dishes until approximately 80% confluency were cultured in complete medium or glucose-free medium for 24 h, and then incubated for 30 min in labeling medium (L-methionine- and L-cysteine-free Dulbecco’s modified Eagle’s medium supplemented with 10% dialyzed fetal calf serum, 2 mM glutamine and antibiotics). Cells were pulse labeled for 10 min with 0.14 mCi (5 MBq)/plate EXPRE³⁵S³⁵S protein labeling mix (PerkinElmer) dissolved in 0.5 ml of labeling medium, and then chased in fresh medium when necessary. Because CHO cells are auxotroph for proline (Curriden and Englesberg, 1981), proline was added to all media at a final concentration of 17.25 mg/L. Labeled cells were washed with ice-cold PBS, harvested by scraping, and suspended in 50 μl of lysis buffer [20 mM Tris/HCl, pH 7.4, containing 1% SDS, protease inhibitor cocktail (Nacalai Tesque) and 10 μM MG132]. Immunoprecipiation was carried out as described previously (Haze et al., 1999). Radio-labeled bands were visualized and quantified using an LAS-1000plus LuminoImage analyzer (Fuji Film).

Discontinuous sucrose gradient centrifugation

CHO cells were scraped after being washed once with 1 ml of ice-cold PBS containing protease inhibitor cocktail and 10 μM MG132. Cells of 8 dishes were pooled, suspended in 200 μl of ice-cold sucrose buffer (10 mM HEPES/KOH, pH 7.4, containing 0.25 M sucrose, 1 mM EDTA, protease inhibitor cocktail, and 10 mM NEM), and homogenized with 30 strokes using a dounce tissue homogenizer (1-ml size, Wheaton). After the addition of 200 μl of sucrose buffer, cells were further homogenized with 20 strokes and then centrifuged at 1,000×g for 5 min at 4°C to remove nuclei and cell debris. pATF6α(N), a cleaved and nuclear form, was recovered into precipitates, whereas pATF6α(P), a precursor form, was recovered into supernatant. Two hundred-microliter aliquots of the resulting supernatant were laid on the top of a discontinuous sucrose gradient consisting of five layers of 200 μl each with distinct sucrose concentrations (20%, 30%, 40%, 50%, and 60% from top to bottom) and centrifuged at 4°C at 50,000 rpm for 30 min in a S55S-1096 rotor (Hitachi) using a Himac CS 150 GXL microultracentrifuge (Hitachi), with deceleration performed without braking. Six fractions of 200 μl each were then collected from the top.

Results and Discussion

Glucose starvation activates ATF6

We examined whether glucose starvation activates ATF6. Mammalian ER expresses two types of ATF6 ubiquitously, designated ATF6α and ATF6β (Haze et al., 2001; Haze et al., 1999). Although ATF6β appears less active in transcription than ATF6α (Thuerauf et al., 2004), no significant difference in activation process between them has been noted. We therefore focused on ATF6α in this report. When Chinese hamster ovary (CHO) cells were cultured in the absence of glucose for 24 h, pATF6α(P), the precursor form, was cleaved to produce pATF6α(N), the nuclear and active form of ATF6α, albeit very weakly, leading to the induction of two major ER chaperones, BiP and GRP94 (Fig. 1A, lane 4).

View Details

Fig. 1. Activation of ATF6 by glucose starvation. (A) CHO cells were cultured in glucose-free medium for the indicated periods. Cell lysates were prepared and analyzed by immunoblotting using anti-ATF6α or anti-KDEL antibody. Migration positions of pATF6α(P), pATF6α(N), GRP94 and BiP are indicated. (B) CHO cells transfected with pCMVshort-EGFP-ATF6α were cultured in glucose-free medium for the indicated periods or treated with 2 μg/ml tunicamycin (Tm) for 5 h. Cells were lysed and subjected to immunoprecipitation with anti-GFP antibody. Immunoprecipitates were analyzed by immunoblotting using anti-GFP antibody. Migration positions of GFP-ATF6α(P) and GFP-ATF6α(N) are indicated. GFP-ATF6α(P*) denotes the non-glycosylated form of GFP-ATF6α(P). (C) CHO cells transfected with pCMVshort-EGFP-ATF6α were untreated (panels a–c), incubated in the presence of 2 μg/ml tunicamycin for 5 h (Tm, panels d–f), or cultured in glucose-free medium for 24 h (-Glc, panels g–i). Cells were fixed and stained with anti-GM130 antibody (panels a, d and g). GFP-ATF6α was visualized by its own fluorescence (panels b, e and h). (D) CHO cells transfected with pCMVshort-EYFP-ATF6α(S1P-) were treated with 1 mM dithiothreitol (DTT, lanes 1–4) or cultured in glucose-free medium (-Glc, lanes 5–8) for the indicated periods. Cells were lysed and subjected to immunoprecipitation with anti-GFP antibody. Immunoprecipitates were analyzed by immunoblotting using anti-KDEL or anti-ATF6α antibody. YFP-ATF6α(S1P-)** denotes an endoglycosidase H-resistant form of YFP-ATF6α(S1P-).

We succeeded in expressing green fluorescent protein (GFP)-ATF6α fusion protein at a level comparable with that of endogenous ATF6α using a truncated CMV promoter, and showed that this GFP-ATF6α behaved quite similarly to endogenous ATF6α (Nadanaka et al., 2004). When CHO cells expressing GFP-ATF6α by transfection were starved of glucose for 27 h, GFP-ATF6α(P) was cleaved to produce GFP-ATF6α(N) (Fig. 1B, lane 3), albeit very weakly as compared to the cleavage observed in cells treated with tunicamycin for 5 h (Fig. 1B, lane 4), which causes ER stress by inhibiting protein N-glycosylation (Kaufman, 1999). Accordingly, some GFP-ATF6α was localized in the Golgi apparatus (Fig. 1C, panels g–i) in glucose-starved cells, similarly to that in tunicamycin-treated cells (panels d–f). GM130 is a marker protein for the Golgi apparatus. This indicates that GFP-ATF6α(P) was relocated from the ER to the Golgi apparatus in response to glucose starvation to be cleaved by Site-1 and Site-2 proteases, resulting in the production of GFP-ATF6α(N).

It is known that ATF6 is retained in the ER via binding to BiP and that ER stress-induced dissociation of BiP triggers the export of ATF6 (Shen et al., 2002). This notion is reproduced in Fig. 1D. This experiment employed ATF6α fused to yellow-emitting GFP whose Site-1 protease cleavage site was mutated, which was referred to as YFP-ATF6α(S1P-). CHO cells were transfected to express YFP-ATF6α(S1P-), treated with or without dithiothreitol, and subjected to immunoprecipitation with anti-GFP antibody. Immunoblotting analysis showed that BiP was bound to YFP-ATF6α(S1P-) in unstressed cells (Fig. 1D, lane 1) and dissociated from YFP-ATF6α(S1P-) after dithiothreitol treatment (Fig. 1D, lanes 2–4), which causes ER stress by disrupting disulfide bridges (Kaufman, 1999). After reaching the Golgi apparatus, YFP-ATF6α(S1P-) was not cleaved by Site-1 protease but its carbohydrate moieties were modified to an endoglycosidase H-resistant form, which migrated more slowly in SDS-PAGE (Fig. 1D, lane 4, indicated by YFP-ATF6α(S1P-)^**). Similarly, BiP was dissociated from YFP-ATF6α(S1P-) after glucose starvation, albeit with very low efficiency (Fig. 1D, lanes 5–8).

It should be noted that the band of pATF6α(P) or GFP-ATF6α(P) became broader when cells were starved of glucose for 24–27 h (Fig. 1A, lane 4 or Fig. 1B, lane 3, respectively). Because GFP-ATF6α(P) in cells starved of glucose for 27 h still migrated more slowly than the non-glycosylated form of GFP-ATF6α(P) [referred to as GFP-ATF6α(P^*)] produced in tunicamycin-treated cells (Fig. 1B, compare lane 3 with lane 4), we suspect that glucose starvation results in a decrease in the amount of oligosaccharide intermediates used for protein N-glycosylation and thus less glycosylation of newly-synthesized proteins, leading to the accumulation of unfolded proteins in the ER and activation of ATF6.

We also noted that the level of pATF6α(P) or GFP-ATF6α(P) increased when cells were starved of glucose for 24–27 h (Fig. 1A, lane 4 or Fig. 1B, lane 3, respectively). We therefore examined whether glucose starvation affects the synthesis rate or degradation rate of ATF6α. CHO cells were pulse labeled with ³⁵S-methionine and cysteine for 10 min before and after culture in glucose-free medium for 24 h, followed by cell lysis and immunoprecipitation with anti-ATF6α or anti-actin antibody. Results showed that synthesis of ATF6α was enhanced approximately 2-fold over that of actin after glucose starvation (Fig. 2A). In contrast, pulse-chase experiments showed that although synthesis of ATF6α was enhanced 1.4-fold over that of actin in this experiment (Fig. 2B, compare lane 1 with lane 4), the degradation rate of ATF6α was not significantly affected by glucose starvation (Fig. 2B, compare lanes 2 and 3 with lanes 5 and 6). The half-life of ATF6α was determined to be approximately 2 h in this experiment, which matched well that obtained previously in HeLa cells (Haze et al., 1999). As ATF6α was stabilized by treatment of cells with MG132 (data not shown), ATF6α is considered to be degraded by the proteasome through an ER-associated degradation-related mechanism (Tsai et al., 2002), details of which are currently under investigation. The results shown in Fig. 2 suggest that ATF6α may also be a target of the UPR, and that its transcription may be induced in response to ER stress in CHO cells. Indeed, upregulation of ATF6 mRNA in response to ER stress was reported in mouse embryonic fibroblasts (Lee et al., 2003) and rat neonatal cardiomyocytes (Terai et al., 2005). The ATF6 promoter region should be examined for binding sites of XBP1, ATF4 and ATF6 itself, which are transcription factors activated by ER stress (Schroder and Kaufman, 2005).

View Details

Fig. 2. Effect of glucose starvation on synthesis and degradation of ATF6α. (A) CHO cells were cultured in complete medium or glucose-free medium for 24 h, and then pulse-labeled for 10 min with 35S-methionine and cysteine. Labeled cells were lysed and subjected to immunoprecipitation with anti-ATF6α or anti-actin antibody. The immunoprecipitates were subjected to SDS-PAGE. The intensity of ATF6α band was determined, normalized with the value of corresponding actin, and corrected for the value of ATF6α in unstarved cells. (B) CHO cells treated as in (A) were pulse labeled for 10 min and chased for the indicated periods. Immunoprecipitation was carried out as in (A). (bottom panel) The intensity of each band was determined, normalized with the value of corresponding actin, corrected for the value of ATF6α in the 0-h control, and plotted against chase periods.

ATF6 is reduced in response to glucose starvation and reduced monomer ATF6 is transported to the Golgi apparatus

We then examined whether ATF6α is reduced in response to glucose starvation similarly to its reduction with dithiothreitol or tunicamycin treatment. SDS-PAGE analysis under non-reducing conditions showed that ATF6α was present in unstressed ER as monomer, dimer and oligomer (Fig. 3A, middle panel, lane 1), as we reported recently (Nadanaka et al., 2007). To follow changes in levels of the three forms of ATF6α after glucose starvation quantitatively, we needed to load the same amounts of ATF6α on a gel. Because the total amounts of ATF6α were markedly increased 24 h after glucose starvation, as shown in Fig. 1A, and enhancement was 2-fold in the experiments shown in Fig. 3A, 20 μg of total proteins were loaded for the 24-h glucose starvation sample, whereas 40 μg were loaded for the 0-, 4-, and 8-h samples. Results showed that dimer ATF6α largely disappeared and monomer ATF6α decreased after 24 h glucose starvation (Fig. 3A, middle panel, lane 4), at which time pATF6α(N), the cleaved form, was detected (Fig. 3A, upper panel, lane 4). This result is very similar to our previous result for tunicamycin treatment (Nadanaka et al., 2007).

View Details

Fig. 3. Reduction of ATF6α in response to glucose starvation. (A) CHO cells were cultured in glucose-free medium for the indicated periods. Cells were recovered in PBS containing 10 mM NEM, suspended in Laemmli’s SDS sample buffer containing 10 mM NEM and then boiled for 10 min. NEM was included to avoid post-lysis artifact. Cell lysates were boiled again for 5 min in the presence (reducing) or absence (non-reducing) of 2-mercaptoethanol. Forty micrograms of total protein each of the 0-, 4-, and 8-h samples as well as 20 μg total protein of the 24-h sample were subjected to SDS-PAGE to adjust the amounts of ATF6α to be analyzed. Immunoblotting was carried out using anti-ATF6α or anti-actin antibody. The result of longer exposure of the region around the nuclear form of ATF6α is shown below the top panel. Migration positions of monomer, dimer and oligomer forms of ATF6α are indicated. (B) CHO cells were cultured in complete medium (control) or in glucose-free medium (Glc starvation) for 24 h. Cells were homogenized and fractionated through a sucrose gradient. All buffer contained 10 mM NEM. Aliquots of the six resulting fractions were subjected to SDS-PAGE under non-reducing conditions and analyzed by immunoblotting with anti-ATF6α antibody. Aliquots of each fraction were also analyzed under reducing conditions using anti-ribophorin I (ER marker) and anti-GM130 (Golgi marker) antibodies. (C) An aliquot of fraction 6 of glucose-starved cells was applied to a centrifugation-coupled membrane filter (Amicon Ultra-4, Millipore, molecular weight cut at 10 kDa) to replace buffer with 10 mM Tris/HCl, pH 7.4, containing 1% SDS, and then subjected to SDS-PAGE under non-reducing conditions together with reduced or non-reduced lysates of control cells or glucose-starved cells. Monomer* denotes less-glycosylated form of monomer ATF6α.

To determine the consequence of glucose starvation-induced reduction of ATF6α, we fractionated cell lysates by sucrose gradient centrifugation. As shown in Fig. 3B, the ER marker ribophorin I was recovered in heavy fractions (fractions 1 and 2), whereas the Golgi marker GM130 was recovered in light fractions (fractions 5 and 6). Under the conditions used, all three forms of ATF6α in unstressed cells were recovered in heavy fractions (control, fractions 1 and 2), whereas only monomer ATF6α was detected in the lightest fraction when cells were starved of glucose for 24 h (Glc starvation, fraction 6). To determine whether the band detected in fraction 6 of glucose-starved cells represented reduced or oxidized monomer ATF6α, an aliquot of fraction 6 was applied to centrifugation-coupled membrane filtration to replace the buffer and then subjected to SDS-PAGE under non-reducing conditions, on the basis that high concentrations of sucrose somewhat disturbed the migration of ATF6α during SDS-PAGE. Comparison with the migration positions of standards containing reduced and oxidized monomer ATF6α revealed that the band in fraction 6 of glucose-starved cells was the reduced monomer ATF6α (Fig. 3C). As pointed out in Fig. 1, ATF6α in glucose-starved cells migrated slightly faster than that in control cells under reducing conditions, probably due to the decreased glycosylation in glucose-starved cells (Fig. 3C, compare lane 1 with lane 2). We concluded that ATF6α is reduced in response to glucose starvation, and that the monomer ATF6α so reduced is then transported from the ER to the Golgi apparatus.

The mechanism by which the cell senses ER stress seems much more complicated than has been previously proposed, namely, that all three UPR mediators present in mammalian ER (IRE1, PERK and ATF6) are regulated by simple binding and dissociation of the ER chaperone BiP, thus forming a negative feedback loop (Bertolotti et al., 2000; Okamura et al., 2000; Shen et al., 2002). Under normal conditions, not all of the BiP is necessarily engaged in the folding of newly synthesized proteins and therefore some of the BiP is able to bind to UPR mediators and maintain them in an inactive state. BiP is much more abundant than UPR mediators. Under ER stress conditions, BiP dissociates from UPR mediators to bind to and refold unfolded proteins accumulated in the ER, resulting in the activation of freed UPR mediators. When ER stress is thus dispersed by the actions of induced ER chaperones, BiP is again able to bind to UPR mediators to inactivate them. It is known that IRE1 and PERK are oligomerized after dissociation of BiP, and that their oligomerized forms autophosphorylate each other to transmit a signal across the ER membrane (Bertolotti et al., 2000). In contrast, our previous and current studies show that ATF6 is deoligomerized in response to ER stress through the reduction of disulfide bridges formed in the lumenal domain of ATF6, and that the reduced monomer ATF6 is transported and cleaved in the Golgi apparatus to release an active transcription factor domain. This represents a clear difference in the activation mechanism between ATF6 and IRE1 or PERK, revealing the diversity existing in the ER stress-sensing mechanism. Given that the glucose starvation employed in the current study represents a physiological insult which causes ER stress (Shiu et al., 1977), as compared with the reducing reagent dithiothreitol and protein glycosylation inhibitor tunicamycin employed in our previous study (Nadanaka et al., 2007), we consider the ER stress-induced reduction of ATF6 to be a generally important event that occurs in the process of ATF6 activation. Nevertheless, it is important to note that because the reduction of ATF6 alone is not sufficient for the activation of ATF6 (Nadanaka et al., 2007), activation requires at least two events, namely, the dissociation of BiP and the reduction of the disulfide bridges. Mammalian ER contains dozens of oxidoreductases (Tu and Weissman, 2004) and we suspect that ATF6 is being specifically reduced by the actions of such enzymes. Identification of the enzymes responsible for ER stress-induced reduction of ATF6 and elucidation of their activation mechanisms will provide new insights into the fundamental question of how the cell senses the accumulation of unfolded proteins in the ER.

Acknowledgements

We wish to thank Ms. Kaoru Miyagawa for technical and secretarial assistance. This work was supported, in part, by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (14037233 and 15GS0310 to K. M.).