- Email: [email protected]
Discordance of UPR signaling by ATF6 and Ire1p-XBP1 with levels of target transcripts
BBRC Biochemical and Biophysical Research Communications 317 (2004) 390–396 www.elsevier.com/locate/ybbrc
Discordance of UPR signaling by ATF6 and Ire1p-XBP1 with levels of target transcripts Jie Shang and Mark A. Lehrman* Department of Pharmacology, University of Texas, Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9041, USA Received 5 March 2004
Abstract Accumulation of misfolded proteins within the lumen of the mammalian endoplasmic reticulum (ER) activates the unfolded protein response (UPR). ATF6 and Ire1p are ER-associated proteins that control UPR-specific transcription systems in mammals; UPR signaling involves cleavage of ATF6 and splicing of XBP1 mRNA initiated by Ire1p. We tested the hypothesis that activation of ATF6 and/or Ire1p determines the levels of mRNAs derived from target genes encoding GRP78/BiP and EDEM. By subjecting dermal fibroblasts to multiple stresses, strong correlations were found between ATF6 activation and XBP1 splicing, and between GRP78/BiP mRNA and EDEM mRNA accumulation. Surprisingly, there was no reasonable correlation between activation of either signal transducer with accumulation of either target transcript. Thus, ATF6 and Ire1p signaling do not define the magnitude of UPR-dependent mRNA increases, even though they may be necessary for gene activation, suggesting the existence of additional stress-sensitive factors acting as “coincidence detectors” for transcript accumulation. Ó 2004 Elsevier Inc. All rights reserved. Keywords: ATF6; Ire1p; XBP1; GRP78; BiP; EDEM; Unfolded protein response
Accumulation of misfolded protein in the endoplasmic reticulum (ER) can trigger signaling events, collectively known as the unfolded protein response (UPR), that act to return ER function to normal. In mammals, in response to ER stress, two UPR-specific signaling pathways involving the ATF6 and Ire1p families of ER transmembrane proteins generate critical transcription factors that activate UPR-responsive genes [1,2]. Uncleaved ATF6 (ATF6U ) translocates to the Golgi apparatus and is activated by regulated intramembrane proteolysis [3], releasing a soluble cytoplasmic fragment (ATF6C ) that enters the nucleus and activates genes controlled by ERSE promoter elements. Dimerization and autophosphorylation of Ire1p activates the endonuclease activity of its cytoplasmic domain [4], cleaving unspliced mRNA (XBP1U ) encoded by the XBP1 gene [5,6] and excising a 26-nt intervening sequence. Upon ligation the spliced form (XBP1S ) encodes a functional transcription factor that activates genes controlled by *
Corresponding author. Fax: 1-214-648-8626. E-mail address: [email protected] Lehrman).
(M.A.
0006-291X/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2004.03.058
UPRE and ERSE promoter elements. The XBP1 gene contains an ERSE element responsive to both ATF6C and the XBP1S protein [1,5,7,8]. Each system responds to diminution of free GRP78/ BiP chaperone in the ER lumen [9,10], so ATF6 cleavage should correlate with XBP1 splicing under a variety of ER stresses. The GRP78/BiP promoter contains an ERSE element responsive to ATF6C and XBP1S while the EDEM gene, which encodes a mannosidase-like protein involved in ER-associated protein degradation, contains a UPRE promoter element responsive to XBP1S but not ATF6C [7]. Thus, GRP78/BiP mRNA amounts should correlate with Ire1p and/or ATF6 activation, EDEM mRNA amounts should correlate with Ire1p activation, and amounts of the two mRNAs should correlate with each other. These predictions could be tested in a system with variable responses to ER stress inducers, followed by quantitative analysis of UPR events. Since its UPR exhibits such behavior for the ER stress agents dithiothreitol (DTT), thapsigargin (TG), azetidine-2-carboxylic acid (AZC), tunicamycin (TN), and castanospermine (CSN) [11], the dermal fibroblast was used to test these predictions.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Materials and methods Cell culture. Normal human dermal fibroblasts (ATCC CRL-1904) were grown and exposed to ER stress as described [11,12]. RT-PCR analysis of XBP1 mRNA splicing. RNA was harvested (RNA-Bee RNA Isolation Solvent, TEL-TEST, or RNeasy mini kit, Qiagen) immediately after completion of stress treatments. First-strand cDNA synthesis was performed with the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). To amplify XBP1 (NM_005080) mRNA, PCR was for 30 cycles [94 °C for 30 s; 58 °C for 30 s; and 72 °C for 1 min (but 10 min in the final cycle)] using 50 -CTGG AACAGCAAGTGGTAGA-30 and 50 -CTGGGTCCTTCTGGGTAG AC-30 with Taq DNA Polymerase (#1815105; Roche). 398 and 424 bp fragments representing spliced (XBP1S ) and unspliced (XBP1U ) XBP1, plus a hybrid (XBP1H ) migrating as a fragment of approximately 450 bp, were documented after staining 2% agarose gels with ethidium bromide, and scanning photographs (Bio-Rad Fluor-S Multimager). In some experiments fragments were confirmed with a second primer pair [5], 50 -CCTTGTAGTTGAGAACCAGG-30 and 50 -GGGGCTT GGTATATATGTGG-30 , which yielded a 442 bp fragment for XBPU . Immunoblotting. Cells were solubilized in 0.5% Triton X-100 with 50 mM Na–Hepes (pH 7.6), 100 mM NaF, 150 mM NaCl, 1 mM Na3 VO4 , 10.5 lg/ml aprotinin, 1 lg/ml pepstatin A, 1 lg/ml leupeptin, and 1 mM PMSF. Eighty micrograms of cellular protein was resolved by 7.5, 8, or 10% SDS–PAGE and transferred to nitrocellulose membranes. ATF6 was identified with a rabbit polyclonal antibody (gift of Dr. Kazutoshi Mori) for the a and b forms [13], followed by incubation with peroxidase-labeled anti-rabbit antibody (#NA934VS, Amersham), detection by chemiluminescence (ECL kit, Amersham), and exposure to X-ray film. Signals were measured with a Fluor-S scanner. Northern blots for GRP78/BiP and EDEM. Probe templates were prepared from human fibroblast RNA by RT-PCR with primers 50 -TT GCTTATGGCCTGGATAAGAGGG-30 and 50 TGTACCCTTGTC TTCAGCTGTCAC-30 for human GRP78/BiP (AJ271729), for a 935 bp fragment, and primers 50 -TCATCCGAGTTCCAGAAAGC AGTC-30 and 50 -TTGACATAGAGTGGAGGGTCTCCT-30 for human EDEM (NM_014674), for a 671 bp fragment. Probes were labeled with 32 P, and RNA samples (RNeasy mini kit, Qiagen) were analyzed by electrophoresis and blotting as described [11]. Signals were
391
quantified with a phosphorimager (Fuji) and normalized to actin mRNA probed on the same blots. Calculations of transcript accumulation, signal activation, and correlation. GRP78/BiP and EDEM mRNA signals (normalized to actin) for stress conditions were divided by corresponding values for untreated cells to determine fold increases. Signal activations were determined from Fluor-S readings: percentage activation of ATF6 was calculated as 100 [ATF6C ]/[ATF6C + ATF6U ]; percentage activation of XBP1 was calculated as 100 [XBP1S + 0.5 XBP1H ]/[XBP1S + XBP1H + XBP1U ]. A correlation coefficient (corrl) was determined in Figs. 3–5 by linear regression analysis (goodness-of-fit statistics, PSIPlot version 7, Poly Software International). Figs. 2, 3, and 5 show standard error bars.
Results Quantification of UPR signaling Dermal fibroblasts were treated with UPR inducers, and components of the ATF6 and Ire1p systems (Fig. 1) were quantified. For ATF6, the cleaved (approximately 60 kDa) and uncleaved (approximately 90 kDa) proteins were identified by immunoblotting and autoradiography (Fig. 1A). To detect Ire1p activation (panel B), RT-PCR was used to amplify fragments representing both the unspliced (XBPU ) and spliced (XBPS ) forms of XBP1 mRNA, differing by 26 nt [5,6]. Their relative percentages were independent of PCR efficiency since they were amplified with the same primer pair. A third slowly migrating species (XBPH ) was also detected with two independent sets of primers. XBP1H was more abundant with milder stresses than robust stresses. Isolation of XBPH followed by additional PCR generated all three fragments, but additional PCR of either XBPS or XBPU regenerated only the starting fragment. ApaLI and PstI,
Fig. 1. ATF6 and Ire1p-XBP1 signaling in dermal fibroblasts. (A) Immunoblot detection of 90 kDa (uncleaved) ATF6 in unstressed cells, and both 90 and 60 kDa (cleaved) ATF6 in cells stressed with 2 mM DTT. (B) RT-PCR of XBP1 mRNA shows XBP1U in control samples, and XBP1S after strong stress with 2 mM DTT, and a mixture of XBPU , XBPS , and XBPH after moderate stress with 100 nM TG.
392
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
enzymes that cleave within the intron or at the exon border, respectively, cleaved XBPU but not XBPS or XBPH . Isolation and sequencing of XBPS or XBPU gave the expected sequences lacking or containing intron, but sequencing of XBPH in either orientation gave composites of the XBPS and XBPU sequences. We conclude that XBPH is a mixture of two hybrid structures (Fig. 1B), each structure containing one strand from XBPS and XBPU , that form during annealing in the final PCR step. Thus, XBPH is greatest with mild stresses because spliced and unspliced strands are both present. All attempts to eliminate XBPH failed, for example by optimizing PCR conditions, replacing reagents, using more primer, transferring trace amounts of product to new reactions, and refreshing PCRs with polymerase and nucleotides during the final two cycles. To quantify splicing, we assigned 0, 0.5, and 1.0 U to XBPU , XBPH , and XBPS . Correlation between the ATF6 and Ire1p responses to various ER stresses Increasing concentrations of AZC, TG, and DTT caused graded increases of GRP78/BiP mRNA [11]. As shown in Fig. 2, a similar result was obtained for cleavage of ATF6 and splicing of XBP1 mRNA. Pla-
teaus for XBP1 splicing were reached with TG (40%; panel B) and DTT (60%; panel C). Plateaus were not apparent for ATF6 cleavage, and the highest degree (50%) was achieved with DTT. The effect of DTT on XBP1 splicing was fully reversed after 120 min (panel D). CSN (0.2 mg/ml) completely inhibits glycoprotein processing [14]. Its effect on XBP1 splicing peaked by 2 h of treatment and then declined (data not shown). CSN was thus used for 24 h, in which case it behaved as a stable, mild ER stress [11]. Specific stress conditions were then chosen for comparison of UPR signaling and target mRNA accumulation (Table 1). Fig. 3 shows that activation of ATF6 and Ire1p was highly correlated (corrl ¼ 0:99) in a comparison involving AZC (10 and 60 mM), TG (10 and 100 nM), DTT (0.4 mM), and CSN (0.2 mg/ml). Results with 2 mM DTT, which had greater ATF6 cleavage than 0.4 mM DTT, were not included in the correlation because maximum splicing of XBP1 was reached with 0.4 mM DTT (Fig. 2). Detection of ATF6C after 0.4 or 2 mM DTT treatment was not affected by 2 h treatment with 0.1 mM ALLN, a proteasome inhibitor (data not shown). This correlation was consistent with the ATF6 and Ire1p systems both responding to the concentration of unbound GRP78/BiP.
Fig. 2. Dependence of UPR signaling upon concentration of ER stress inducers. Percent activation of Ire1p-XBP1 (circles) and ATF6 (squares) is described under Materials and methods. (A) AZC for 60 min. (B) TG for 30 min. (C) DTT for 20 min. In (D), cells were stressed with 2 mM DTT for 20 min and changed to DTT-free medium for the indicated times. Data are SE.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
393
Table 1 Treatments used to cause ER stress Treatments
Control DTT, 0.4 mM, 20 min DTT, 2 mM, 20 min TG, 10 nM, 30 min TG, 100 nM, 30 min AZC, 10 mM, 1 h AZC, 60 mM, 1 h TN, 5 lg/ml, 1 h TN, 5 lg/ml, 5 h CSN, 200 lg/ml, 2 h
Graph symbols
I j s d n m } r ~
Number of determinations (n) in Figs. 3 and 5 ATF6 cleavage
XBP1 splicing
EDEM mRNA
GRP78 mRNA
15 7 6 3 3 2 2 0 0 2
13 4 3 5 7 2 2 0 0 4
10 7 5 6 6 2 2 2 3 2
10 7 5 6 6 2 2 2 3 2
20-fold range of GRP78/BiP mRNA and a 4.5-fold range of EDEM mRNA. This result suggests that similar mechanisms control the extents to which the GRP78/BiP and EDEM promoters respond to ER stress. Lack of correlation between activation of ATF6 and XBP1 with accumulation of GRP78/BiP and EDEM transcripts
Fig. 3. ATF6 and Ire1p signaling are concordant under multiple ER stress conditions. Stress conditions, symbols, and numbers of repetitions are listed in Table 1. The corrl value (Materials and methods) for the line shown, not including 2 mM DTT, is 0.99.
Correlation between accumulation of GRP78/BiP and EDEM mRNAs Accumulation of GRP78/BiP and EDEM transcripts correlated strongly with each other (corrl ¼ 0:93; Fig. 4). This correlation involved 35 independent pairwise measurements involving five ER stress inducers (i.e., those used for Fig. 3 as well as TN), and held over a
Fig. 4. Concordance between GRP78/BiP and EDEM transcript accumulation under multiple ER stresses. Fold increases of mRNA were determined and normalized (Materials and methods). Each data point represents a single experiment. The corrl value was 0.93.
Since EDEM gene transcription depends upon an XBP1S -sensitive UPRE promoter element, EDEM mRNA accumulation should correlate with signaling by Ire1p. Similarly, since the GRP78/BiP gene contains a ERSE promoter element, which responds to both XBP1S and ATF6C , accumulation of GRP78/BiP mRNA should correlate with Ire1p and/or ATF6 signaling. Surprisingly, multiple determinations with seven different ER stress conditions (Figs. 5A–D) failed to show reasonable correlation between the increases for GRP78 mRNA or EDEM mRNA with either ATF6 cleavage or XBP1 splicing. For example, similar ATF6 cleavage was associated with 2-fold (0.4 mM DTT) and 18-fold (100 nM TG) increases of GRP78/BiP mRNA, and the same XBP1 splicing was associated with a 4.4fold change (100 nM TG) and a 1.2-fold change (0.4 mM DTT) of EDEM mRNA. In no case was the value of corrl greater than 0.1. The strong mutual correlations between the two signaling events and between the two transcripts, but the lack of correlation between either signaling event with either transcript, suggest that an additional stress-dependent factor(s) controls the extent of accumulation of both transcripts, as explained by the model in Fig. 6. Thus, ATF6 cleavage and XBP1 splicing are considered necessary for transcription, but beyond a certain threshold the actual amounts of ATF6C and XBP1S are not important. In the model, the additional factor(s) determines the accumulation of both GRP78/BiP and EDEM transcripts because of their strong mutual correlation. Further, the factor is stress-dependent. Since
394
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Fig. 5. Discordance between signal activation and transcript accumulation. Data sets from Figs. 3 and 4 were used for XBP1 splicing in (A,B), ATF6 activation in (C,D), GRP78/BiP mRNA in (A,C), and EDEM mRNA in (B,D). Values of corrl did not exceed 0.1. Numbers of repetitions are listed in Table 1.
Fig. 6. Model for an additional factor(s) determining the extent of ER stress-dependent accumulation of transcripts. ER stress is proposed to affect at least one component (“?”) other than GRP78/BiP to signal accumulation of misfolded protein. The signal would then activate a factor that acts similarly upon both targets, either by increasing transcription (as indicated) or by affecting mRNA stability. ATF6U , uncleaved ATF6; ATF6C , cleaved ATF6; XBP1S , spliced XBP1 mRNA; XBP1U , unspliced XBP1 mRNA; ERSE, ER stress element; and UPRE, UPR element. Thus, ATF6C and XBP1S are necessary for target gene transcription, but they do not control the actual amounts of mRNA that accumulate.
0.4 mM DTT was more effective for signal activation than transcript accumulation when compared with 10 and 100 nM TG, TG is proposed to activate the factor more strongly than DTT. However, the effects of 0.4 mM DTT were additive, not synergistic, with either 10 nM TG or 100 nM TG for accumulation of EDEM (Fig. 7A) and GRP78/BiP (panel B) transcripts, which in
Fig. 7. Lack of synergy between DTT and TG stresses. Cells were treated with 0.4 mM DTT, 10 nM TG, or 100 nM TG, alone or in combination as indicated, and the fold increases for EDEM (A) and GRP78/BiP (B) mRNA were determined. Results are typical of two experiments. XBP1 splicing was assessed from corresponding sets of dishes as shown in (C); H, U, and S indicate hybrid, unspliced, and spliced XBP1.
turn were inconsistent with XBP1 splicing (panel C). Since XBP1 splicing was maximal with 0.4 mM DTT (panel C), no additional splicing was expected by adding TG. These results might indicate TG- and DTT-specific factors. However, it is plausible that the postulated factor does not behave in a synergistic manner.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Discussion With multiple forms of ER stress ATF6 and Ire1p signaling were mutually correlated (Fig. 3), as was accumulation of transcripts from the ERSE-driven gene GRP78/BiP and the UPRE-driven gene EDEM (Fig. 4), but there was no correlation between either signaling system and either transcript (Fig. 5). These results reinforce the idea that UPR transcriptional responses can be highly variable, and are consistent with previous studies on stress-dependent effects [11], dependence upon dosage and time course of stress treatments [15], and timing of UPR signals [7]. Thus, signaling by ATF6 or Ire1p/XBP1 is necessary, but the extent of target transcript accumulation is highly dependent upon the particular form of stress. The data suggest that there are additional stress-dependent factors that act as transcriptional “coincidence detectors” for ER stress, that may guard against spurious ATF6 or Ire1p signaling. Our studies measured total mRNA levels, and did not distinguish between mRNA synthesis and stability. Although UPR alteration of mRNA stability has not been reported, it remains a possibility. These data may reflect recent reports showing how UPR-dependent transcription can be modulated. For example, the GRP78/BiP promoter contains a functional ATF4-responsive element [16]. This is significant because the ATF4 transcription factor is produced during PERK-dependent translation attenuation [17] as part of the integrated stress response [18], and PERK function itself appears to be regulated by several proteins including GADD34 [19]. Additionally, ATF6 undergoes stress-dependent interactions with NF-Y and YY1 that appear to increase its transcriptional activity [20]. Stress-dependent modulation of the signaling sensors themselves is also possible: Ire1a is stabilized by HSP90 [21] and hypoglycosylation (a general cause of ER stress) of ATF6 enhances its cleavage with greater downstream transcriptional activation [22]. Another factor is the dependence of transcription of the XBP1 gene upon ATF6 activation [5,7]. However, this synergistic relationship is unlikely to explain the absence of correlation between signal activation and target transcript levels. ER stress causes dissociation of GRP78/BiP from ATF6, Ire1p, and PERK, resulting in signal activation [10,9]. Since ATF6 activation and Ire1p activation correlated with each other, but not with target transcript accumulation, some other ER signal and signal transducer may be involved. However, the general mechanism involving dissociation of an ER chaperone from a docking site on a signaling molecule (due to competition with misfolded protein) might be retained. The SERCA 2b pump, which replenishes ER calcium stores (and is the target of TG), is regulated by association with calnexin (CNX) and calreticulin (CRT)–oxidoreductase
395
ERp57 complexes [23]. Thus, it is interesting to speculate that ER stress causes dissociation of the CNX and CRT chaperones from SERCA 2b. This would alter the frequency of calcium oscillations in the cytoplasm, with the potential to cause myriad effects on cytoplasmic processes involved in the propagation of ATF6 and Ire1p signals. In summary, our data with the dermal fibroblast appear to refute the hypothesis that activation of ATF6 and/or Ire1p determines the levels of mRNAs derived from target transcripts. Rather, the results imply the existence of other ER stress-sensitive factors that serve this function.
Acknowledgments This work was supported by NIH Grant GM38545 and Welch Grant I-1168. We thank Andreas Reimold for comments and Biswanath Pramanik for technical support.
References [1] K. Mori, Frame switch splicing and regulated intramembrane proteolysis: key words to understand the unfolded protein response, Traffic 4 (2003) 519–528. [2] C. Patil, P. Walter, Intracellular signaling from the endoplasmic reticulum to the nucleus: the unfolded protein response in yeast and mammals, Curr. Opin. Cell Biol. 13 (2001) 349–356. [3] J. Ye, R.B. Rawson, R. Komuro, X. Chen, U. Dave, R. Prywes, M.S. Brown, J.L. Goldstein, ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs, Mol. Cell 6 (2000) 1355–1364. [4] C. Sidrauski, P. Walter, The transmembrane kinase Ire1p is a sitespecific endonuclease that initiates mRNA splicing in the unfolded protein response, Cell 90 (1997) 1031–1039. [5] H. Yoshida, T. Matsui, A. Yamamoto, T. Okada, K. Mori, XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor, Cell 107 (2001) 881–891. [6] M. Calfon, H. Zeng, F. Urano, J.H. Till, S.R. Hubbard, H.P. Harding, S.G. Clark, D. Ron, IRE1 couples endoplasmic reticulum load to secretory capacity by processing the XBP-1 mRNA, Nature 415 (2002) 92–96. [7] H. Yoshida, T. Matsui, N. Hosokawa, R.J. Kaufman, K. Nagata, K. Mori, A time-dependent phase shift in the mammalian unfolded protein response, Dev. Cell 4 (2003) 265–271. [8] K. Lee, W. Tirasophon, X. Shen, M. Michalak, R. Prywes, T. Okada, H. Yoshida, K. Mori, R.J. Kaufman, IRE1-mediated unconventional mRNA splicing and S2P-mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response, Genes Dev. 16 (2002) 452–466. [9] J. Shen, X. Chen, L. Hendershot, R. Prywes, ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals, Dev. Cell 3 (2002) 99– 111. [10] A. Bertolotti, Y. Zhang, L.M. Hendershot, H.P. Harding, D. Ron, Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response, Nat. Cell Biol. 2 (2000) 326–332. [11] J. Shang, C. Korner, H. Freeze, M.A. Lehrman, Extension of lipid-linked oligosaccharides is a high-priority aspect of the unfolded protein response: endoplasmic reticulum stress in Type
396
[12] [13]
[14]
[15] [16]
[17]
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396 I congenital disorder of glycosylation fibroblasts, Glycobiology 12 (2002) 307–317. J. Shang, M.A. Lehrman, Metformin-stimulated mannose transport in dermal fibroblasts, J. Biol. Chem. 279 (2004) 9703–9712. K. Haze, H. Yoshida, H. Yanagi, T. Yura, K. Mori, Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress, Mol. Biol. Cell 10 (1999) 3787–3799. M.A. Lehrman, Y. Zeng, Pleiotropic resistance to glycoprotein processing inhibitors in Chinese hamster ovary cells, J. Biol. Chem. 264 (1989) 1584–1593. K. Imaizumi, T. Katayama, M. Tohyama, Presenilin and the UPR (letter to the editor), Nat. Cell Biol. 3 (2001) E104. S. Luo, P. Baumeister, S. Yang, S.F. Abcouwer, A.S. Lee, Induction of Grp78/BiP by translational block. Activation of the Grp78 promoter by ATF4 through an upstream ATF/CRE site independent of the endoplasmic reticulum stress elements, J. Biol. Chem. 278 (2003) 37375–37385. H.P. Harding, I. Novoa, Y. Zhang, H. Zeng, R. Wek, M. Schapira, D. Ron, Regulated translation initiation controls stressinduced gene expression in mammalian cells, Mol. Cell 6 (2000) 1099–1108.
[18] H.P. Harding, Y. Zhang, H. Zeng, I. Novoa, P.D. Lu, M. Calfon, N. Sadri, C. Yun, B. Popko, R. Paules, D.F. Stojdl, J.C. Bell, T. Hettmann, J.M. Leiden, D. Ron, An integrated stress response regulates amino acid metabolism and resistance to oxidative stress, Mol. Cell 11 (2003) 619–633. [19] I. Novoa, H. Zeng, H.P. Harding, D. Ron, Feedback inhibition of the unfolded protein response by GADD34-mediated dephosphorylation of eIF2a, J. Cell Biol. 153 (2001) 1011–1021. [20] M.Q. Li, P. Baumeister, B. Roy, T. Phan, D. Foti, S.Z. Luo, A.S. Lee, ATF6 as a transactivator of the endoplasmic reticulum stress element: thapsigargin stress-induced changes and synergistic interactions with NF-Y and YY1, Mol. Cell. Biol. 20 (2000) 5096–5106. [21] M.G. Marcu, D. Doyle, A. Bertolotti, D. Ron, L. Hendershot, L. Neckers, Heat shock protein 90 modulates the unfolded protein response by stabilizing IRE1 alpha, Mol. Cell. Biol. 22 (2002) 8506–8513. [22] M. Hong, S. Luo, P. Baumeister, J.-M. Huang, R.K. Gogia, M. Li, A.S. Lee, Underglycosylation of ATF6 as a novel sensing mechanism for activation of the unfolded protein response, J. Biol. Chem. 279 (2004) 11354–11363. [23] Y. Li, P. Camacho, Ca2þ -dependent redox modulation of SERCA 2b by ERp57, J. Cell Biol. 164 (2004) 35–46.
Discordance of UPR signaling by ATF6 and Ire1p-XBP1 with levels of target transcripts Jie Shang and Mark A. Lehrman* Department of Pharmacology, University of Texas, Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9041, USA Received 5 March 2004
Abstract Accumulation of misfolded proteins within the lumen of the mammalian endoplasmic reticulum (ER) activates the unfolded protein response (UPR). ATF6 and Ire1p are ER-associated proteins that control UPR-specific transcription systems in mammals; UPR signaling involves cleavage of ATF6 and splicing of XBP1 mRNA initiated by Ire1p. We tested the hypothesis that activation of ATF6 and/or Ire1p determines the levels of mRNAs derived from target genes encoding GRP78/BiP and EDEM. By subjecting dermal fibroblasts to multiple stresses, strong correlations were found between ATF6 activation and XBP1 splicing, and between GRP78/BiP mRNA and EDEM mRNA accumulation. Surprisingly, there was no reasonable correlation between activation of either signal transducer with accumulation of either target transcript. Thus, ATF6 and Ire1p signaling do not define the magnitude of UPR-dependent mRNA increases, even though they may be necessary for gene activation, suggesting the existence of additional stress-sensitive factors acting as “coincidence detectors” for transcript accumulation. Ó 2004 Elsevier Inc. All rights reserved. Keywords: ATF6; Ire1p; XBP1; GRP78; BiP; EDEM; Unfolded protein response
Accumulation of misfolded protein in the endoplasmic reticulum (ER) can trigger signaling events, collectively known as the unfolded protein response (UPR), that act to return ER function to normal. In mammals, in response to ER stress, two UPR-specific signaling pathways involving the ATF6 and Ire1p families of ER transmembrane proteins generate critical transcription factors that activate UPR-responsive genes [1,2]. Uncleaved ATF6 (ATF6U ) translocates to the Golgi apparatus and is activated by regulated intramembrane proteolysis [3], releasing a soluble cytoplasmic fragment (ATF6C ) that enters the nucleus and activates genes controlled by ERSE promoter elements. Dimerization and autophosphorylation of Ire1p activates the endonuclease activity of its cytoplasmic domain [4], cleaving unspliced mRNA (XBP1U ) encoded by the XBP1 gene [5,6] and excising a 26-nt intervening sequence. Upon ligation the spliced form (XBP1S ) encodes a functional transcription factor that activates genes controlled by *
Corresponding author. Fax: 1-214-648-8626. E-mail address: [email protected] Lehrman).
(M.A.
0006-291X/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2004.03.058
UPRE and ERSE promoter elements. The XBP1 gene contains an ERSE element responsive to both ATF6C and the XBP1S protein [1,5,7,8]. Each system responds to diminution of free GRP78/ BiP chaperone in the ER lumen [9,10], so ATF6 cleavage should correlate with XBP1 splicing under a variety of ER stresses. The GRP78/BiP promoter contains an ERSE element responsive to ATF6C and XBP1S while the EDEM gene, which encodes a mannosidase-like protein involved in ER-associated protein degradation, contains a UPRE promoter element responsive to XBP1S but not ATF6C [7]. Thus, GRP78/BiP mRNA amounts should correlate with Ire1p and/or ATF6 activation, EDEM mRNA amounts should correlate with Ire1p activation, and amounts of the two mRNAs should correlate with each other. These predictions could be tested in a system with variable responses to ER stress inducers, followed by quantitative analysis of UPR events. Since its UPR exhibits such behavior for the ER stress agents dithiothreitol (DTT), thapsigargin (TG), azetidine-2-carboxylic acid (AZC), tunicamycin (TN), and castanospermine (CSN) [11], the dermal fibroblast was used to test these predictions.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Materials and methods Cell culture. Normal human dermal fibroblasts (ATCC CRL-1904) were grown and exposed to ER stress as described [11,12]. RT-PCR analysis of XBP1 mRNA splicing. RNA was harvested (RNA-Bee RNA Isolation Solvent, TEL-TEST, or RNeasy mini kit, Qiagen) immediately after completion of stress treatments. First-strand cDNA synthesis was performed with the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). To amplify XBP1 (NM_005080) mRNA, PCR was for 30 cycles [94 °C for 30 s; 58 °C for 30 s; and 72 °C for 1 min (but 10 min in the final cycle)] using 50 -CTGG AACAGCAAGTGGTAGA-30 and 50 -CTGGGTCCTTCTGGGTAG AC-30 with Taq DNA Polymerase (#1815105; Roche). 398 and 424 bp fragments representing spliced (XBP1S ) and unspliced (XBP1U ) XBP1, plus a hybrid (XBP1H ) migrating as a fragment of approximately 450 bp, were documented after staining 2% agarose gels with ethidium bromide, and scanning photographs (Bio-Rad Fluor-S Multimager). In some experiments fragments were confirmed with a second primer pair [5], 50 -CCTTGTAGTTGAGAACCAGG-30 and 50 -GGGGCTT GGTATATATGTGG-30 , which yielded a 442 bp fragment for XBPU . Immunoblotting. Cells were solubilized in 0.5% Triton X-100 with 50 mM Na–Hepes (pH 7.6), 100 mM NaF, 150 mM NaCl, 1 mM Na3 VO4 , 10.5 lg/ml aprotinin, 1 lg/ml pepstatin A, 1 lg/ml leupeptin, and 1 mM PMSF. Eighty micrograms of cellular protein was resolved by 7.5, 8, or 10% SDS–PAGE and transferred to nitrocellulose membranes. ATF6 was identified with a rabbit polyclonal antibody (gift of Dr. Kazutoshi Mori) for the a and b forms [13], followed by incubation with peroxidase-labeled anti-rabbit antibody (#NA934VS, Amersham), detection by chemiluminescence (ECL kit, Amersham), and exposure to X-ray film. Signals were measured with a Fluor-S scanner. Northern blots for GRP78/BiP and EDEM. Probe templates were prepared from human fibroblast RNA by RT-PCR with primers 50 -TT GCTTATGGCCTGGATAAGAGGG-30 and 50 TGTACCCTTGTC TTCAGCTGTCAC-30 for human GRP78/BiP (AJ271729), for a 935 bp fragment, and primers 50 -TCATCCGAGTTCCAGAAAGC AGTC-30 and 50 -TTGACATAGAGTGGAGGGTCTCCT-30 for human EDEM (NM_014674), for a 671 bp fragment. Probes were labeled with 32 P, and RNA samples (RNeasy mini kit, Qiagen) were analyzed by electrophoresis and blotting as described [11]. Signals were
391
quantified with a phosphorimager (Fuji) and normalized to actin mRNA probed on the same blots. Calculations of transcript accumulation, signal activation, and correlation. GRP78/BiP and EDEM mRNA signals (normalized to actin) for stress conditions were divided by corresponding values for untreated cells to determine fold increases. Signal activations were determined from Fluor-S readings: percentage activation of ATF6 was calculated as 100 [ATF6C ]/[ATF6C + ATF6U ]; percentage activation of XBP1 was calculated as 100 [XBP1S + 0.5 XBP1H ]/[XBP1S + XBP1H + XBP1U ]. A correlation coefficient (corrl) was determined in Figs. 3–5 by linear regression analysis (goodness-of-fit statistics, PSIPlot version 7, Poly Software International). Figs. 2, 3, and 5 show standard error bars.
Results Quantification of UPR signaling Dermal fibroblasts were treated with UPR inducers, and components of the ATF6 and Ire1p systems (Fig. 1) were quantified. For ATF6, the cleaved (approximately 60 kDa) and uncleaved (approximately 90 kDa) proteins were identified by immunoblotting and autoradiography (Fig. 1A). To detect Ire1p activation (panel B), RT-PCR was used to amplify fragments representing both the unspliced (XBPU ) and spliced (XBPS ) forms of XBP1 mRNA, differing by 26 nt [5,6]. Their relative percentages were independent of PCR efficiency since they were amplified with the same primer pair. A third slowly migrating species (XBPH ) was also detected with two independent sets of primers. XBP1H was more abundant with milder stresses than robust stresses. Isolation of XBPH followed by additional PCR generated all three fragments, but additional PCR of either XBPS or XBPU regenerated only the starting fragment. ApaLI and PstI,
Fig. 1. ATF6 and Ire1p-XBP1 signaling in dermal fibroblasts. (A) Immunoblot detection of 90 kDa (uncleaved) ATF6 in unstressed cells, and both 90 and 60 kDa (cleaved) ATF6 in cells stressed with 2 mM DTT. (B) RT-PCR of XBP1 mRNA shows XBP1U in control samples, and XBP1S after strong stress with 2 mM DTT, and a mixture of XBPU , XBPS , and XBPH after moderate stress with 100 nM TG.
392
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
enzymes that cleave within the intron or at the exon border, respectively, cleaved XBPU but not XBPS or XBPH . Isolation and sequencing of XBPS or XBPU gave the expected sequences lacking or containing intron, but sequencing of XBPH in either orientation gave composites of the XBPS and XBPU sequences. We conclude that XBPH is a mixture of two hybrid structures (Fig. 1B), each structure containing one strand from XBPS and XBPU , that form during annealing in the final PCR step. Thus, XBPH is greatest with mild stresses because spliced and unspliced strands are both present. All attempts to eliminate XBPH failed, for example by optimizing PCR conditions, replacing reagents, using more primer, transferring trace amounts of product to new reactions, and refreshing PCRs with polymerase and nucleotides during the final two cycles. To quantify splicing, we assigned 0, 0.5, and 1.0 U to XBPU , XBPH , and XBPS . Correlation between the ATF6 and Ire1p responses to various ER stresses Increasing concentrations of AZC, TG, and DTT caused graded increases of GRP78/BiP mRNA [11]. As shown in Fig. 2, a similar result was obtained for cleavage of ATF6 and splicing of XBP1 mRNA. Pla-
teaus for XBP1 splicing were reached with TG (40%; panel B) and DTT (60%; panel C). Plateaus were not apparent for ATF6 cleavage, and the highest degree (50%) was achieved with DTT. The effect of DTT on XBP1 splicing was fully reversed after 120 min (panel D). CSN (0.2 mg/ml) completely inhibits glycoprotein processing [14]. Its effect on XBP1 splicing peaked by 2 h of treatment and then declined (data not shown). CSN was thus used for 24 h, in which case it behaved as a stable, mild ER stress [11]. Specific stress conditions were then chosen for comparison of UPR signaling and target mRNA accumulation (Table 1). Fig. 3 shows that activation of ATF6 and Ire1p was highly correlated (corrl ¼ 0:99) in a comparison involving AZC (10 and 60 mM), TG (10 and 100 nM), DTT (0.4 mM), and CSN (0.2 mg/ml). Results with 2 mM DTT, which had greater ATF6 cleavage than 0.4 mM DTT, were not included in the correlation because maximum splicing of XBP1 was reached with 0.4 mM DTT (Fig. 2). Detection of ATF6C after 0.4 or 2 mM DTT treatment was not affected by 2 h treatment with 0.1 mM ALLN, a proteasome inhibitor (data not shown). This correlation was consistent with the ATF6 and Ire1p systems both responding to the concentration of unbound GRP78/BiP.
Fig. 2. Dependence of UPR signaling upon concentration of ER stress inducers. Percent activation of Ire1p-XBP1 (circles) and ATF6 (squares) is described under Materials and methods. (A) AZC for 60 min. (B) TG for 30 min. (C) DTT for 20 min. In (D), cells were stressed with 2 mM DTT for 20 min and changed to DTT-free medium for the indicated times. Data are SE.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
393
Table 1 Treatments used to cause ER stress Treatments
Control DTT, 0.4 mM, 20 min DTT, 2 mM, 20 min TG, 10 nM, 30 min TG, 100 nM, 30 min AZC, 10 mM, 1 h AZC, 60 mM, 1 h TN, 5 lg/ml, 1 h TN, 5 lg/ml, 5 h CSN, 200 lg/ml, 2 h
Graph symbols
I j s d n m } r ~
Number of determinations (n) in Figs. 3 and 5 ATF6 cleavage
XBP1 splicing
EDEM mRNA
GRP78 mRNA
15 7 6 3 3 2 2 0 0 2
13 4 3 5 7 2 2 0 0 4
10 7 5 6 6 2 2 2 3 2
10 7 5 6 6 2 2 2 3 2
20-fold range of GRP78/BiP mRNA and a 4.5-fold range of EDEM mRNA. This result suggests that similar mechanisms control the extents to which the GRP78/BiP and EDEM promoters respond to ER stress. Lack of correlation between activation of ATF6 and XBP1 with accumulation of GRP78/BiP and EDEM transcripts
Fig. 3. ATF6 and Ire1p signaling are concordant under multiple ER stress conditions. Stress conditions, symbols, and numbers of repetitions are listed in Table 1. The corrl value (Materials and methods) for the line shown, not including 2 mM DTT, is 0.99.
Correlation between accumulation of GRP78/BiP and EDEM mRNAs Accumulation of GRP78/BiP and EDEM transcripts correlated strongly with each other (corrl ¼ 0:93; Fig. 4). This correlation involved 35 independent pairwise measurements involving five ER stress inducers (i.e., those used for Fig. 3 as well as TN), and held over a
Fig. 4. Concordance between GRP78/BiP and EDEM transcript accumulation under multiple ER stresses. Fold increases of mRNA were determined and normalized (Materials and methods). Each data point represents a single experiment. The corrl value was 0.93.
Since EDEM gene transcription depends upon an XBP1S -sensitive UPRE promoter element, EDEM mRNA accumulation should correlate with signaling by Ire1p. Similarly, since the GRP78/BiP gene contains a ERSE promoter element, which responds to both XBP1S and ATF6C , accumulation of GRP78/BiP mRNA should correlate with Ire1p and/or ATF6 signaling. Surprisingly, multiple determinations with seven different ER stress conditions (Figs. 5A–D) failed to show reasonable correlation between the increases for GRP78 mRNA or EDEM mRNA with either ATF6 cleavage or XBP1 splicing. For example, similar ATF6 cleavage was associated with 2-fold (0.4 mM DTT) and 18-fold (100 nM TG) increases of GRP78/BiP mRNA, and the same XBP1 splicing was associated with a 4.4fold change (100 nM TG) and a 1.2-fold change (0.4 mM DTT) of EDEM mRNA. In no case was the value of corrl greater than 0.1. The strong mutual correlations between the two signaling events and between the two transcripts, but the lack of correlation between either signaling event with either transcript, suggest that an additional stress-dependent factor(s) controls the extent of accumulation of both transcripts, as explained by the model in Fig. 6. Thus, ATF6 cleavage and XBP1 splicing are considered necessary for transcription, but beyond a certain threshold the actual amounts of ATF6C and XBP1S are not important. In the model, the additional factor(s) determines the accumulation of both GRP78/BiP and EDEM transcripts because of their strong mutual correlation. Further, the factor is stress-dependent. Since
394
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Fig. 5. Discordance between signal activation and transcript accumulation. Data sets from Figs. 3 and 4 were used for XBP1 splicing in (A,B), ATF6 activation in (C,D), GRP78/BiP mRNA in (A,C), and EDEM mRNA in (B,D). Values of corrl did not exceed 0.1. Numbers of repetitions are listed in Table 1.
Fig. 6. Model for an additional factor(s) determining the extent of ER stress-dependent accumulation of transcripts. ER stress is proposed to affect at least one component (“?”) other than GRP78/BiP to signal accumulation of misfolded protein. The signal would then activate a factor that acts similarly upon both targets, either by increasing transcription (as indicated) or by affecting mRNA stability. ATF6U , uncleaved ATF6; ATF6C , cleaved ATF6; XBP1S , spliced XBP1 mRNA; XBP1U , unspliced XBP1 mRNA; ERSE, ER stress element; and UPRE, UPR element. Thus, ATF6C and XBP1S are necessary for target gene transcription, but they do not control the actual amounts of mRNA that accumulate.
0.4 mM DTT was more effective for signal activation than transcript accumulation when compared with 10 and 100 nM TG, TG is proposed to activate the factor more strongly than DTT. However, the effects of 0.4 mM DTT were additive, not synergistic, with either 10 nM TG or 100 nM TG for accumulation of EDEM (Fig. 7A) and GRP78/BiP (panel B) transcripts, which in
Fig. 7. Lack of synergy between DTT and TG stresses. Cells were treated with 0.4 mM DTT, 10 nM TG, or 100 nM TG, alone or in combination as indicated, and the fold increases for EDEM (A) and GRP78/BiP (B) mRNA were determined. Results are typical of two experiments. XBP1 splicing was assessed from corresponding sets of dishes as shown in (C); H, U, and S indicate hybrid, unspliced, and spliced XBP1.
turn were inconsistent with XBP1 splicing (panel C). Since XBP1 splicing was maximal with 0.4 mM DTT (panel C), no additional splicing was expected by adding TG. These results might indicate TG- and DTT-specific factors. However, it is plausible that the postulated factor does not behave in a synergistic manner.
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396
Discussion With multiple forms of ER stress ATF6 and Ire1p signaling were mutually correlated (Fig. 3), as was accumulation of transcripts from the ERSE-driven gene GRP78/BiP and the UPRE-driven gene EDEM (Fig. 4), but there was no correlation between either signaling system and either transcript (Fig. 5). These results reinforce the idea that UPR transcriptional responses can be highly variable, and are consistent with previous studies on stress-dependent effects [11], dependence upon dosage and time course of stress treatments [15], and timing of UPR signals [7]. Thus, signaling by ATF6 or Ire1p/XBP1 is necessary, but the extent of target transcript accumulation is highly dependent upon the particular form of stress. The data suggest that there are additional stress-dependent factors that act as transcriptional “coincidence detectors” for ER stress, that may guard against spurious ATF6 or Ire1p signaling. Our studies measured total mRNA levels, and did not distinguish between mRNA synthesis and stability. Although UPR alteration of mRNA stability has not been reported, it remains a possibility. These data may reflect recent reports showing how UPR-dependent transcription can be modulated. For example, the GRP78/BiP promoter contains a functional ATF4-responsive element [16]. This is significant because the ATF4 transcription factor is produced during PERK-dependent translation attenuation [17] as part of the integrated stress response [18], and PERK function itself appears to be regulated by several proteins including GADD34 [19]. Additionally, ATF6 undergoes stress-dependent interactions with NF-Y and YY1 that appear to increase its transcriptional activity [20]. Stress-dependent modulation of the signaling sensors themselves is also possible: Ire1a is stabilized by HSP90 [21] and hypoglycosylation (a general cause of ER stress) of ATF6 enhances its cleavage with greater downstream transcriptional activation [22]. Another factor is the dependence of transcription of the XBP1 gene upon ATF6 activation [5,7]. However, this synergistic relationship is unlikely to explain the absence of correlation between signal activation and target transcript levels. ER stress causes dissociation of GRP78/BiP from ATF6, Ire1p, and PERK, resulting in signal activation [10,9]. Since ATF6 activation and Ire1p activation correlated with each other, but not with target transcript accumulation, some other ER signal and signal transducer may be involved. However, the general mechanism involving dissociation of an ER chaperone from a docking site on a signaling molecule (due to competition with misfolded protein) might be retained. The SERCA 2b pump, which replenishes ER calcium stores (and is the target of TG), is regulated by association with calnexin (CNX) and calreticulin (CRT)–oxidoreductase
395
ERp57 complexes [23]. Thus, it is interesting to speculate that ER stress causes dissociation of the CNX and CRT chaperones from SERCA 2b. This would alter the frequency of calcium oscillations in the cytoplasm, with the potential to cause myriad effects on cytoplasmic processes involved in the propagation of ATF6 and Ire1p signals. In summary, our data with the dermal fibroblast appear to refute the hypothesis that activation of ATF6 and/or Ire1p determines the levels of mRNAs derived from target transcripts. Rather, the results imply the existence of other ER stress-sensitive factors that serve this function.
Acknowledgments This work was supported by NIH Grant GM38545 and Welch Grant I-1168. We thank Andreas Reimold for comments and Biswanath Pramanik for technical support.
References [1] K. Mori, Frame switch splicing and regulated intramembrane proteolysis: key words to understand the unfolded protein response, Traffic 4 (2003) 519–528. [2] C. Patil, P. Walter, Intracellular signaling from the endoplasmic reticulum to the nucleus: the unfolded protein response in yeast and mammals, Curr. Opin. Cell Biol. 13 (2001) 349–356. [3] J. Ye, R.B. Rawson, R. Komuro, X. Chen, U. Dave, R. Prywes, M.S. Brown, J.L. Goldstein, ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs, Mol. Cell 6 (2000) 1355–1364. [4] C. Sidrauski, P. Walter, The transmembrane kinase Ire1p is a sitespecific endonuclease that initiates mRNA splicing in the unfolded protein response, Cell 90 (1997) 1031–1039. [5] H. Yoshida, T. Matsui, A. Yamamoto, T. Okada, K. Mori, XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor, Cell 107 (2001) 881–891. [6] M. Calfon, H. Zeng, F. Urano, J.H. Till, S.R. Hubbard, H.P. Harding, S.G. Clark, D. Ron, IRE1 couples endoplasmic reticulum load to secretory capacity by processing the XBP-1 mRNA, Nature 415 (2002) 92–96. [7] H. Yoshida, T. Matsui, N. Hosokawa, R.J. Kaufman, K. Nagata, K. Mori, A time-dependent phase shift in the mammalian unfolded protein response, Dev. Cell 4 (2003) 265–271. [8] K. Lee, W. Tirasophon, X. Shen, M. Michalak, R. Prywes, T. Okada, H. Yoshida, K. Mori, R.J. Kaufman, IRE1-mediated unconventional mRNA splicing and S2P-mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response, Genes Dev. 16 (2002) 452–466. [9] J. Shen, X. Chen, L. Hendershot, R. Prywes, ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals, Dev. Cell 3 (2002) 99– 111. [10] A. Bertolotti, Y. Zhang, L.M. Hendershot, H.P. Harding, D. Ron, Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response, Nat. Cell Biol. 2 (2000) 326–332. [11] J. Shang, C. Korner, H. Freeze, M.A. Lehrman, Extension of lipid-linked oligosaccharides is a high-priority aspect of the unfolded protein response: endoplasmic reticulum stress in Type
396
[12] [13]
[14]
[15] [16]
[17]
J. Shang, M.A. Lehrman / Biochemical and Biophysical Research Communications 317 (2004) 390–396 I congenital disorder of glycosylation fibroblasts, Glycobiology 12 (2002) 307–317. J. Shang, M.A. Lehrman, Metformin-stimulated mannose transport in dermal fibroblasts, J. Biol. Chem. 279 (2004) 9703–9712. K. Haze, H. Yoshida, H. Yanagi, T. Yura, K. Mori, Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress, Mol. Biol. Cell 10 (1999) 3787–3799. M.A. Lehrman, Y. Zeng, Pleiotropic resistance to glycoprotein processing inhibitors in Chinese hamster ovary cells, J. Biol. Chem. 264 (1989) 1584–1593. K. Imaizumi, T. Katayama, M. Tohyama, Presenilin and the UPR (letter to the editor), Nat. Cell Biol. 3 (2001) E104. S. Luo, P. Baumeister, S. Yang, S.F. Abcouwer, A.S. Lee, Induction of Grp78/BiP by translational block. Activation of the Grp78 promoter by ATF4 through an upstream ATF/CRE site independent of the endoplasmic reticulum stress elements, J. Biol. Chem. 278 (2003) 37375–37385. H.P. Harding, I. Novoa, Y. Zhang, H. Zeng, R. Wek, M. Schapira, D. Ron, Regulated translation initiation controls stressinduced gene expression in mammalian cells, Mol. Cell 6 (2000) 1099–1108.
[18] H.P. Harding, Y. Zhang, H. Zeng, I. Novoa, P.D. Lu, M. Calfon, N. Sadri, C. Yun, B. Popko, R. Paules, D.F. Stojdl, J.C. Bell, T. Hettmann, J.M. Leiden, D. Ron, An integrated stress response regulates amino acid metabolism and resistance to oxidative stress, Mol. Cell 11 (2003) 619–633. [19] I. Novoa, H. Zeng, H.P. Harding, D. Ron, Feedback inhibition of the unfolded protein response by GADD34-mediated dephosphorylation of eIF2a, J. Cell Biol. 153 (2001) 1011–1021. [20] M.Q. Li, P. Baumeister, B. Roy, T. Phan, D. Foti, S.Z. Luo, A.S. Lee, ATF6 as a transactivator of the endoplasmic reticulum stress element: thapsigargin stress-induced changes and synergistic interactions with NF-Y and YY1, Mol. Cell. Biol. 20 (2000) 5096–5106. [21] M.G. Marcu, D. Doyle, A. Bertolotti, D. Ron, L. Hendershot, L. Neckers, Heat shock protein 90 modulates the unfolded protein response by stabilizing IRE1 alpha, Mol. Cell. Biol. 22 (2002) 8506–8513. [22] M. Hong, S. Luo, P. Baumeister, J.-M. Huang, R.K. Gogia, M. Li, A.S. Lee, Underglycosylation of ATF6 as a novel sensing mechanism for activation of the unfolded protein response, J. Biol. Chem. 279 (2004) 11354–11363. [23] Y. Li, P. Camacho, Ca2þ -dependent redox modulation of SERCA 2b by ERp57, J. Cell Biol. 164 (2004) 35–46.