Nucleobindin 1 Controls the Unfolded Protein Response by Inhibiting ATF6 Activation*

In response to endoplasmic reticulum (ER) stress, activating transcription factor 6 (ATF6), an ER membrane-anchored transcription factor, is transported to the Golgi apparatus and cleaved by site-1 protease (S1P) to activate the unfolded protein response (UPR). Here, we identified nucleobindin 1 (NUCB1) as a novel repressor of the S1P-mediated ATF6 activation. NUCB1 is an ER stress-inducible gene with the promoter region having functional cis-elements for transcriptional activation by ATF6. Overexpression of NUCB1 inhibits S1P-mediated ATF6 cleavage without affecting ER-to-Golgi transport of ATF6, whereas knock-down of NUCB1 by siRNA accelerates ATF6 cleavage during ER stress. NUCB1 protein localizes in the Golgi apparatus, and disruption of the Golgi localization results in loss of the ATF6-inhibitiory activity. Consistent with these observations, NUCB1 can suppress physical interaction of S1P-ATF6 during ER stress. Together, our results demonstrate that NUCB1 is the first-identified, Golgi-localized negative feedback regulator in the ATF6-mediated branch of the UPR.

X-box-binding protein 1 (XBP1) is induced and forms a complex with active XBP1, a strong UPR transcription factor that is produced by IRE1-mediated unconventional splicing of XBP-1 mRNA (15). The complex formation leads to rapid degradation of both XBP-1 proteins (16). These findings raise the possibility that there are some genes that have the ability to modulate the UPR among uncharacterized ER stress-inducible genes.
Herein we focus on NUCB1, one of ER stress-inducible genes that we identified using a cDNA microarray system. NUCB1, also known as nucleobindin or calnuc, was first found as a secretory protein that possessed a signal sequence, a leucine zipper structure, two EF-hand motifs and a basic amino acidrich region (17,18,19). Indeed, NUCB1 has been shown to be secreted in culture supernatant of a murine B cell line established from mice prone to the autoimmune disorder systemic lupus erythematosus (17,20,21). Interestingly, further studies have also demonstrated that NUCB1 is expressed ubiquitously and localizes in the Golgi apparatus of intact cells (19,22). Although Golgi-localized NUCB1 has been suggested to be involved in Ca 2ϩ homeostasis (23), the function of NUCB1 is largely unknown. In this study, we demonstrate that Golgi-localized NUCB1 suppresses ATF6 activation by inhibiting S1Pmediated cleavage during ER stress.
Cell Culture and Treatments-Cells were maintained in RPMI 1640 medium (for HT1080) or Dulbecco's modified Eagle's medium (for 293T and HeLa) supplemented with 10% heat-inactivated fetal bovine serum and 100 g/ml of kanamycin and were cultured at 37°C in a humidified atmosphere containing 5% CO 2 . Glucose-free RPMI 1640 medium was obtained from Invitrogen and supplemented with 10% heat-inactivated fetal bovine serum for experimental use, as described (24). To induce UPR, cells underwent glucose deprivation for varying times. This was achieved by replacing the medium with the glucose-free medium, or chemical stressors 2DG, tunicamycin, thapsigardin, or DTTcontaining medium (24). In some experiments, we treated cells with the proteasome inhibitor MG132 at 10 M during exposure to the UPR inducers.
RT-PCR Analysis-RT-PCR was performed as previously described (24). Briefly, total RNA was isolated from cells using the RNeasy Mini Kit with DNase digestion (Qiagen, Tokyo, Japan) and converted to cDNA with SuperScript II reverse transcriptase (Invitrogen). The cDNAs for NUCB1, GRP78, and G3PDH were then amplified by PCR with PfuTurbo DNA polymerase (Stratagene). PCR products were separated by electrophoresis on 1.2% agarose gels and visualized by ethidium bromide staining.
Immunoblot Analysis-Cells were lysed in 1ϫ SDS sample buffer, and protein concentrations of the lysates were measured with a Bio-Rad protein assay kit. For analysis of secreted proteins, culture medium (serum-free) was concentrated with Microcon YM-30 (Millipore; Billerica, MA) and then boiled in 1ϫ SDS buffer. Equal amounts of proteins were resolved on a 10% SDS-polyacrylamide gel and transferred by electroblotting to a nitrocellulose membrane. Membranes were probed with antibodies, as indicated, and the specific signals were detected using an enhanced chemiluminescence detection system (GE Healthcare Bio-Sciences Corp) (24).
Plasmids-pcFLAG vector, pATF6wt, pATF6act, pXBP1sp, pGRP78pro160-Luc have been described previously (24). The plasmid of 7ϫ FLAG-tagged ATF6 was produced by tandem ligation of an oligonucleotide DNA sequence encoding 3ϫ FLAG epitope to the HindIII site of pATF6wt. Expression vectors for non-tagged, V5-tagged and FLAG-tagged NUCB1 proteins were constructed by inserting each cDNA amplified by RT-PCR into pcDNA3 (Invitrogen), pcDNA3.1 TOPO V5/His (Invitrogen) and pFLAG-CMV-5c (Sigma), respectively. Expression vectors for IRE1 and S1P were constructed by inserting each cDNA amplified by RT-PCR (for IRE1) or by PCR using a KIAA0091 clone (kindly provided from Kazusa DNA Research Institute (28)) as a template (for S1P) into the pcDNA3.1 TOPO V5/His vector (Invitrogen). Reporter plasmids containing a promoter region of NUCB1 were generated by ligating products amplified by PCR, using genomic DNA from HT1080 cells as a template, into pGL3-basic vector (Promega). Site-directed mutagenesis was carried out using a QuikChange mutagenesis kit (Stratagene) for the NUCB1 mutant, P28A plasmid, which was exchanged Pro to Ala at position 28. The proper construction of plasmids was confirmed by DNA sequencing.
Transfection-Transient transfections were performed using either FuGENE6 Transfection Reagent (Roche Applied Sciences, Indianapolis, IN) for HT1080 cells or Lipofectamine 2000 (Invitrogen) for 293T, according to each manufacturer's protocol.
Reporter Assay-HT1080 cells (2 ϫ 10 5 in each well of 12-well plates) were cultured overnight under normal growth conditions. After changing the medium to antibiotic-free RPMI 1640 medium supplemented with 5% fetal bovine serum, we incubated the cells for 6 h at 37°C in a CO 2 incubator with transfection mixtures containing the indicated amounts of NUCB1 plasmids and 375 ng of the firefly luciferase-containing reporter plasmids pGRP78pro160-Luc with 125 ng of plasmid pRL-TK (Promega) (in which Renilla luciferase expression is under the control of the herpes simplex virus thymidine kinase promoter) as an internal control. The medium was then replaced with fresh medium lacking plasmids, and the cells were incubated under the same conditions for another 6 h. The cells were then treated for 12 h with UPR inducers. Relative firefly-to-Renilla luciferase activity (means with S.D. values of triplicate determinations) was determined using the dual luciferase kit (Promega). For NUCB1 promoter analysis, HT1080 cells (1 ϫ 10 5 in each well of 24-well plates) were transiently transfected with 500 ng of pNUCB1-Luc (firefly) plasmids, which contained different lengths (Ϫ584 to ϩ112, Ϫ184 to ϩ16, Ϫ134 to ϩ16, Ϫ24 to ϩ112) or point mutants of the NUCB1 promoter region (see Fig. 2) together with 1 ng of pRL-hCMV plasmid (Renilla) (Promega) as an internal control and then assayed as described above.
RNA Interference-Double-stranded RNA duplexes corresponding to human NUCB1 (1: 5Ј-AAGAGTTCAAGCGC-TACGAGA-3Ј and 2: 5Ј-CTCAGTGATCGGCTTAACTTA-3Ј), as well as Control siRNA were purchased from Qiagen. For transient transfection of siRNA, cells were seeded at a density of 0.5 ϫ 10 5 /well in a 12-well plate and were cultured overnight. The cells were transfected for 6 h with siRNA (80 nM), using Lipofectamine 2000 reagent according to the manufacturer's protocol. Two days after transfection, the cells were used for experimentation.
Sialidase Treatment-NUCB1 proteins prepared from cells were immunoprecipitated with anti-NUCB1. After elution using synthetic peptide, each sample was digested with sialidase II (Clostridium perfringens) at 37°C for 3 h.

RESULTS
NUCB1, an ER Stress-responsive Gene-We conducted a microarray analysis of gene expression to identify novel genes induced by glucose starvation, one of the physiological ER stress conditions. For this purpose, we used mRNAs from human ovarian cancer A2780 and colon cancer HT29 cells that had been cultured for 18 h under normal or glucose starvation conditions. As a result, NUCB1 was identified as a glucose starvation-inducible gene. To validate the finding, we analyzed mRNA expression of NUCB1 by RT-PCR in human fibrosarcoma HT1080 cells (Fig. 1A). Enhanced NUCB1 expression, as well as GRP78 expression, was seen in cells starved of glucose for 18 h. NUCB1 was also induced by treatment with 2-deoxyglucose (2DG) and tunicamycin, representative chemical ER stressors. Immunoblot analysis revealed that NUCB1 protein was also up-regulated by glucose withdrawal or 2DG treatment in HT1080 and HeLa cells (Fig. 1B).
As previously reported (19), immunostaining demonstrated that NUCB1 localized at the Golgi apparatus (see Fig. 5C). However, NUCB1 was also reported as a secretory protein in mouse pituitary AtT20 and normal rat kidney cells (21). To verify secretion of NUCB1, we cultured HT1080 cells in serumfree medium in the presence of 2DG or tunicamycin and then recovered the culture medium for immunoblot analysis. NUCB1 was detected in the culture medium under both normal and stress conditions, and the secretion was increased by the chemical stressors within 3 h (Fig. 1C). At that time point, GRP78 induction was marginal, suggesting that the increased NUCB1 secretion occurred at a relatively early phase of the UPR. These results indicate that NUCB1 can be Golgi-resident inside the cells and can be secreted outside the cells.
NUCB1 has been shown to be modified with sialylated O-linked oligosaccharides (21). To determine sialylated status of NUCB1, we immunopurified NUCB1 proteins and digested them with sialidase (Fig. 1D). The sialidase treatment resulted in an approximate 1-2-kDa shift in extracellular pool mobility of NUCB1 proteins, indicating that secreted NUCB1 proteins were indeed sialylated. A similar mobility shift was seen in intracellular pools of NUCB1 proteins in both unstressed and 2DG-stressed cells, which had been detected as two bands and resulted in the disappearance of the upper band (Fig. 1D). Together, these results suggested that sialylation might be an essential modification for secretion of NUCB1.
ER Stress-responsive Elements in NUCB1 Promoter-We cloned an approximate 700-bp putative promoter region of the human NUCB1 gene, including the transcriptional initiation site (Ϫ584 to ϩ112), into a luciferase reporter vector (pNUCB1-Luc). The pNUCB1-Luc or the SV40 promoterdriven pGL3 (control) was transfected into HT1080 cells. Treating the cells with tunicamycin for 18 h led to activation of the NUCB1 promoter, but not the SV40 promoter, in a dosedependent manner (Fig. 2A). We also found that co-transfection of pATF6act or pXBP1sp, the active forms of UPR transcription factors ATF6 and XBP1, respectively (8,15), resulted in a strong activation of NUCB1 promoter activity ( Fig. 2A).  OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40

JOURNAL OF BIOLOGICAL CHEMISTRY 29267
Examination of the NUCB1 promoter region revealed one ERSE II (Ϫ124ϳϪ114, inverted) and one unfolded protein response element (UPRE)-like sequence (Ϫ152ϳϪ144) as potential ER stress-responsive cis-elements (Fig. 2B). We tested the role of the regions in promoter activity using pNUCB1-Luc1 (Ϫ184ϳϩ16), a deletion version of pNUCB1-Luc that still contains both the UPRE-like and the ERSE II regions. The pNUCB1-Luc1 was strongly activated by tunicamycin treatment (18 h) and by co-transfection with pATF6act or pXBP1sp (Fig. 2C). Deletion of the UPRE-like region (pNUCB1-Luc2: Ϫ134ϳϩ16) decreased the promoter activation without affecting the basal promoter activity under normal conditions. Further bulky deletion (pNUCB1-Luc3: Ϫ24ϳϩ112) led to promoter activation loss and remarkable decrease in the basal promoter activity. Consistently, each mutation of UPRE-like (Mut1) and ERSE II (Mut2 and 3) in pNUCB1-Luc1 decreased promoter activation by tunicamycin treatment and by pATF6act or pXBP1sp co-transfection (Fig. 2D). Some variations were seen in the inhibitory effect of each mutation, depending on the promoter activation conditions. Combined mutation of both elements resulted in further promoter activation decrease. Interestingly, the point mutations had a marginal effect on the basal promoter activity. These results demonstrated that both ERSE II and UPRE-like regions mediate NUCB1 promoter activation during ER stress.
NUCB1 Represses UPR with Inhibiting ATF6 Activation-To examine the role of NUCB1 in the UPR, we used pGRP78-Luc, which contained a GRP78 promoter region (Ϫ160ϳϩ7) immediately upstream of the firefly luciferase. As previously reported (3), the GRP78 promoter reporter was activated by treating the transfected cells with 2DG or tunicamycin in a dose-dependent manner (Fig. 3A). We found that co-transfection of NUCB1 significantly attenuated the GRP78 promoter activity induced by each stressor (Fig. 3A). Consistently, immunoblot analysis of lysates from 293T cells revealed that NUCB1 overexpression significantly attenuated induction of endogenous UPR target proteins under stress conditions (Fig. 3B).
GRP78 promoter activity was also enhanced in proportion to the transfected amount of pATF6wt (Fig. 4A, left). Under the experimental conditions, the ATF6wt-induced GRP78 promoter activation was decreased remarkably by co-transfection of NUCB1. In contrast, NUCB1 had little effect on the promoter activity induced by the active form of ATF6 (Fig. 4A,  middle). As to the IRE1-XBP1 pathway, NUCB1 did not affect IRE1-or XBP1sp-mediated activation of the GRP78 promoter (Fig. 4A, right).
We next examined the effects of NUCB1 on the activation process of ATF6. For this purpose, we used a construct of 7ϫ FLAG-tagged ATF6 to increase the detection sensitivity to the small amount of ATF6 because extremely overexpressed ATF6 can be readily activated by proteolytic cleavage independently of ER stress (11). When the plasmid of 7ϫ FLAG-tagged ATF6 was transfected into HT1080 cells, the ATF6 activation process was detected in an ER stress-dependent manner (Fig. 4B). Indeed, in addition to p90ATF6, an intermediate product (S1P cleavage product) and p50ATF6/active form (sequential S2P cleavage product) emerged after thapsigargin treatment. We found that the production of processed ATF6 was decreased by co-transfection of NUCB1 (Fig. 4B, left). Conversely, siRNAmediated NUCB1 knockdown enhanced the processed form production during tunicamycin or thapsigargin treatment (Fig.  4B, right).
We also examined the effects of NUCB1 overexpression on endogenous ATF6 processing (Fig. 4C, left). When 293T cells were treated with tunicamycin for 8 -12 h, p50ATF6 as well as the underglycosylated form of p90ATF6 were produced. The production of p50ATF6 was attenuated by overexpression of NUCB1. In agreement with this finding, silencing of NUCB1 in HT1080 cells enhanced the production of endogenous p50ATF6 during tunicamycin treatment for 8 h or 12 h, as compared with non-silencing cells (Fig. 4C, right and supplemental  Fig. S5). These results indicate that NUCB1 can negatively regulate ATF6 processing for its activation during ER stress.
NUCB1 Represses ATF6 Activation in the Golgi Apparatus-We examined whether or not NUCB1 overexpression influenced the subcellular localization of ATF6 (7ϫ FLAG tag) (Fig.  5A). At 20 min after DTT addition, ATF6 was concentrated in Golgi-like structures either with or without NUCB1 overexpression (Fig. 5A, arrow). At 40 min, some populations of the cells transfected with pATF6 alone took up nucleus staining (Fig. 5A, arrowhead) while cells co-transfected with ATF6 and FIGURE 3. Effects of NUCB1 on ER stress response. A, HT1080 cells were transiently co-transfected with pNUCB1 (V5 tag) and pGRP78-Luc together with pRL-TK as the internal control. The cells were then treated with the indicated doses of 2DG or tunicamycin for 12 h. The luciferase activity was measured using a dual luciferase assay kit. Each activity of firefly or Renilla luciferase is shown separately in supplemental Fig. S2. B, 293T cells were transiently transfected with pNUCB1 (FLAG tag) and were then treated with the indicated doses of 2DG or tunicamycin for 14 h. The cells lysates were subjected to immunoblot analysis with anti-KDEL, anti-calnexin, anti-ERp72, anti-FLAG, and anti-␤-actin antibodies. NUCB1 remained Golgi staining (Fig. 5A, arrow), indicating that the nucleus translocation of ATF6 was repressed by NUCB1 at the Golgi apparatus. Quantifying more than 100 transfected cells after 30 min of DTT treatment revealed that the nuclear translocation of ATF6 was decreased by half when NUCB1 was co-expressed (Fig. 5B). In contrast, the NUCB1 co-expression had no effect on ATF6 translocation from the ER to the Golgi apparatus (Fig. 5B).
We have screened several NUCB1 mutants at the N terminus and found that P28A mutant localized in the ER (Fig. 5C). Although the mechanism of ER localization of this mutant is not known at present, we used it to confirm the finding that NUCB1 suppresses ATF6 processing in the Golgi apparatus.
We compared the effects of P28A and WT on ATF6 processing. HT1080 cells were co-transfected with pATF6wt (7ϫ FLAG-tag) and WT or P28A and then treated with thapsigargin (Fig. 5D). We found that P28A was scarcely able to repress thapsigargin-induced ATF6 processing (Fig. 5D). Consistently, P28A did not suppress nuclear translocation of ATF6, while, for unknown reasons, P28A stimulated it (Fig. 5E). Together, the Golgi localization of NUCB1 is likely required for inhibition of ATF6 processing.
Impairment of Interaction between ATF6 and S1P by NUCB1-We found that interaction between ATF6 and S1P was detected during ER stress by a co-immunoprecipitation assay. Indeed, co-transfection of pS1P (V5 tag) and pATF6wt (7ϫ FLAG tag) made it possible to detect the interaction between ATF6 and S1P under conditions of 2DG, tunicamycin, and thapsigargin (Fig. 6A). We then examined the effects of NUCB1 on the stress-induced interaction between ATF6 and S1P. As shown in Fig. 6B, thapsigargin-induced interaction between ATF6 and S1P was clearly diminished by NUCB1 overexpression. At the same time, NUCB1 overexpression inhibited production of the active form of ATF6. Thus, NUCB1 prevented the interaction between ATF6 and S1P leading to ATF6 cleavage during ER stress.
In agreement with these findings, we also found that overexpression of S1P dose-dependently suppressed the inhibitory activity of NUCB1 against ATF6-mediated GRP78 promoter activation (Fig. 6C). Together, these results support the notion that the interaction between ATF6 and S1P is an important target for inhibition of ATF6 activation by NUCB1.

DISCUSSION
The ATF6 activation system must contain a negative feedback regulation mechanism because ATF6 processing was temporarily increased and then gradually attenuated during ER stress (3,29). However, regulatory proteins for the process are largely unknown. We present evidence that NUCB1 can function as an ER stress-responsive negative regulator of ATF6 activation. Indeed, both mRNA and protein of NUCB1 are induced during ER stress. The promoter region of the NUCB1 gene con- . NUCB1 represses UPR with inhibiting ATF6 activation. A, HT1080 cells were transiently co-transfected with pNUCB1 (V5 tag), pGRP78-Luc, and pRL-TK (internal control) together with the indicated amounts of pATF6wt (left), pATF6act (middle), or 50 ng of pIRE1 or pXBP1sp (right). After 24 h, the luciferase activity was measured using a dual luciferase assay kit. Each activity of firefly or Renilla luciferase is shown separately in supplemental Fig. S3. B, HT1080 cells were transiently co-transfected with 50 ng of pATF6wt (7ϫ FLAG tag) and 50 ng of pNUCB1 (V5 tag) and were then treated with 300 nM of thapsigargin for the indicated time periods (left). HT1080 cells were transiently transfected with non-silencing control or NUCB1 siRNA (1, 2) (right). After 24 h, the cells were further transiently transfected with 50 ng of pATF6wt (7ϫ FLAG tag), and the next day, the cells were treated with 2 g/ml of tunicamycin for 9 h or 300 nM thapsigargin for 6 h. The cell lysates were subjected to immunoblot analysis with anti-FLAG, anti-V5, anti-NUCB1, and anti-␤-actin antibodies. C, 293T cells were transiently transfected with 500 ng of pNUCB1 (FLAG tag). After 24 h, the cells were treated with 4 g/ml of tunicamycin for the indicated time periods (left). HT1080 cells were transiently transfected with non-silencing control or NUCB1 siRNA (1,2). After 48 h, the cells were treated with 2 g/ml of tunicamycin for 8 h (right). Six hours before harvest in each experiment, MG132 (final 10 M) was added to the culture medium. The cells lysates were subjected to immunoblot analysis with anti-ATF6, anti-KDEL, anti-NUCB1, and anti-␤actin antibodies.
tains ERSE II and UPRE-like regions, which can be functional as cis-elements for transcriptional activation by ATF6 and by UPR-inducing stressors. NUCB1 protein localizes in the Golgi apparatus, the intracellular site where ATF6 is cleaved by S1P/ S2P for nuclear translocation. In agreement with intracellular localization, NUCB1 suppresses physical interaction of S1P-ATF6 and subsequent S1P-mediated cleavage of ATF6 during ER stress. Collectively, these findings demonstrate that NUCB1 can be involved in a negative feedback loop of the ATF6-mediated branch of the UPR.
We showed herein that ATF6 activation is suppressed by overexpression of NUCB1, whereas it is enhanced by knockdown of NUCB1. Thus, the increased expression of NUCB1 during ER stress can be an important mechanism to sup-press ATF6 activation. As observed for NUCB1, certain UPR-inducible genes have also been reported to exhibit UPR-modulating activity (9,13,14,16). As to ATF6 regulation, GRP78, a major UPR target chaperone protein, has been shown to down-regulate the ATF6 activation process by preventing transport of newly synthesized ATF6 from the ER to the Golgi apparatus (9). Thus, the ATF6 activation process can be regulated by GPR78 in the ER and by NUCB1 in the Golgi. In addition to expression levels, subcellular localization of NUCB1 may contribute to ATF6 regulation during ER stress. Indeed, we observed a tendency that increased secretion of NUCB1 occurred sooner than increased intracellular expression of NUCB1 during ER stress (Fig. 1C). As secreted NUCB1 is incapable of preventing ATF6 activation in the Golgi apparatus, increased secretion of NUCB1, seen at the early phase of the UPR, may play a role in facilitating the ATF6 activation process by weakening the NUCB1mediated negative regulation. On the other hand, at the UPR late phase, the ATF6 activation process may be fine-tuned through the increased intracellular expression levels of NUCB1 together with GRP78.
While previous studies demonstrated that extracellular NUCB1 can stimulate B cell growth (20), we observed no effect of silencing NUCB1 on HT1080 fibrosarcoma cell growth (data not shown), which suggests that the growth-promoting activity of NUCB1 may be cell typedependent. Alternatively, the secreted form may contribute to modulating the stressed cell microenvironment. Actually, extracellular NUCB1 has been suggested to serve as a modulator of matrix maturation in bone, based on the observations that NUCB1 is secreted by osteoblasts and osteocytes and can indeed be detected in the osteoid extracellular matrix (18,30). Thus, NUCB1 secretion might have a role not only in facilitating ATF6 activation inside stressed cells but also in modulating stressful microenvironment outside the cells.
We showed that as an inhibition mechanism of ATF6 activation, NUCB1 prevented S1P-ATF6 interaction during ER stress. However, we were unable to detect a binding between NUCB1 and either ATF6 or S1P under the same experimental conditions that the S1P-ATF6 interaction was easily seen. This . NUCB1 represses ATF6 activation in the Golgi. A, HT1080 cells were transiently transfected with pATF6wt (7ϫ FLAG tag) alone or together with pNUCB1 (V5 tag) and then were treated with 2 mM DTT for 0, 20, or 40 min. The cells were fixed and stained with anti-FLAG and anti-V5 antibodies. In ATF6 localization, arrow indicates Golgi localization and arrowhead indicates nucleus localization. B, after transfection, as described in A, cells were treated with 1 mM DTT for 30 min. The ratio of ATF6 translocation to the Golgi apparatus or subsequent translocation was determined by counting more than 100 FLAG-positive (ATF6) cells. The Golgi column represents the Golgi translocation cell ratio in FLAG-positive (ATF6) cells. The nucleus column represents the nucleus translocation cell ratio in Golgi-translocated, FLAG-positive (ATF6) cells. Complete translocation of ATF6 to the nucleus was not observed with 1 mM DTT treatment for 30 min. C, after transfection with each of the indicated constructs (V5 tag), HT1080 cells were fixed and double-stained with anti-V5, anti-golgin 97 (Golgi marker), or anti-KDEL (GRP78) antibodies. Nuclei were counterstained with Hoechst 33342. D, HT1080 cells co-transfected with pATF6wt (7ϫ FLAG-tag) and each of the indicated pNUCB1 mutants (V5 tag) were treated with 300 nM thapsigargin for the indicated time periods. The cell lysates were subjected to immunoblot analysis with anti-FLAG, anti-V5, and anti-␤-actin antibodies. E, ratio of ATF6 nuclear translocation was determined as in B.
may have been the result of technical problems because we could detect an interaction between NUCB1 and ATF6 when both proteins were extremely overexpressed by transfection (data not shown). Thus, a complex between NUCB1 and ATF6 or S1P may be too sensitive to detergents or too transient to be detected under our experimental conditions. We speculate that NUCB1 requires cooperating protein(s) to bind to ATF6 or S1P and to prevent S1P-ATF6 interaction. In other words, the S1P-ATF6 interaction during ER stress may be an event regulated by several factors containing NUCB1 but not be a mere consequence of ATF6 translocation from ER to the Golgi apparatus. In this context, it may be worth noting that NUCB1 has been reported to be able to interact physically with other proteins, such as cyclooxygenase I and a G protein, Gi␣3 (31,32). It would be interesting to examine whether such proteins are involved in regulation of the UPR, especially the ATF6-mediated branch.
In conclusion, our studies identified NUCB1 as a novel UPR negative regulatory protein that suppresses the S1P-mediated cleavage activation of ATF6 in the Golgi apparatus. In addition to ATF6, there are several members of ER-anchored transcriptional factors that are structurally similar to ATF6 (33,34,35). These members, including SREBP2, OASIS and CREBH, all have been shown to be activated by S1P/S2P-mediated cleavage, the so-called, RIP. However, it is unlikely that RIP is generally regulated by NUCB1 because overexpression or siRNA-mediated knockdown of NUCB1 had little effect on SREBP2 processing (data not shown). Thus, these findings provide a clue to elucidating regulation mechanisms of RIP for each transcriptional factor at the Golgi apparatus. Besides, the finding of enhanced NUCB1 secretion during ER stress may also be intriguing, as increased NUCB1 secretion has reportedly been associated with induction or enhancement of autoimmunity in murine models of lupus (20). Thus, our findings could offer information for studying the relationship between the UPR and autoimmune response, as well as other UPR-associated pathophysiological states, such as tumor development, neurodegenerative disorders, and diabetes.