Induction of Endoplasmic Reticulum-induced Stress Genes in Panc-1 Pancreatic Cancer Cells Is Dependent on Sp Proteins*

Endoplasmic reticulum (ER) stress plays a critical role in multiple diseases, and pharmacologically active drugs can induce cell death through ER stress pathways. Stress-induced genes are activated through assembly of transcription factors on ER stress response elements (ERSEs) in target gene promoters. Gel mobility shift and chromatin immunoprecipitation assays have confirmed interactions of NF-Y and YY1 with the distal motifs of the tripartite ERSE from the glucose-related protein 78 (GRP78) gene promoter. The GC-rich nonanucleotide (N9) sequence, which forms the ER stress response binding factor (ERSF) complex binds TFII-I and ATF6; however, we have now shown that in Panc-1 pancreatic cancer cells, this complex also binds Sp1, Sp3, and Sp4 proteins. Sp proteins are constitutively bound to the ERSE; however, activation of GRP78 protein (or reporter gene) by thapsigargin or tunicamycin is inhibited after cotransfection with small inhibitory RNAs for Sp1, Sp3, and Sp4. This study demonstrates that Sp transcription factors are important for stress-induced responses through their binding to ERSEs.

The endoplasmic reticulum (ER) 1 plays a critical role in protein folding, and diseases such as Parkinson, Alzheimer, and Huntington disease and obesity are linked to misfolded proteins (1)(2)(3)(4)(5). Various chemical and biological stressors also induce ER stress, resulting in the unfolded protein response, which both activates and deactivates gene/protein expression to restore the cell to homeostasis (6 -11). Failure to counteract induced ER stress can result in activation of apoptosis (12), and several pharmacologically active drugs act through this pathway (13)(14)(15). The unfolded protein response is accompanied by activation of several possible intracellular signaling pathways (16,17), which include the cleavage of the inactive form of the transmembrane protein ATF6 to give the N-terminal fragment which acts as a nuclear transcription factor (18 -22). ATF6 binds the ER stress response element (ERSE), and the resulting trans-acting protein complex activates expression of several ER stress-responsive genes including glucose-related protein 78 (GRP78) and CHOP (GADD153) (23)(24)(25)(26)(27).
The ERSE serves as a critical cis-acting motif, which mediates the transcriptional response to ER stress. The consensus ERSE is a tripartite nucleotide sequence (CCAAT(N 9 )CCACG), which contains binding sites for the transcription factors NFY/ CBF and YY1 (16,17). The ER stress response element binding factor (ERSF) complex also interacts with the ERSE nonanucleotide sequence and contains ATF6 and TFII-I proteins (21,22,28). This study now demonstrates that activation of the stress response by thapsigargin (Tg) and tunicmycin (Tm) (8,9,29,30) is attenuated by Sp family proteins, and using a combination of gel mobility shift and chromatin immunoprecipitation (ChIP) and transactivation assays, we now demonstrate a role for Sp1, Sp3, and Sp4 in mediating ER stress-induced gene expression.
Transfection of Panc-1 Pancreatic Cancer Cells and Preparation of Nuclear Extracts-Cells were cultured in 6-well plates in 2 ml of Dulbecco's modified Eagle's medium/F-12 medium supplemented with 5% fetal bovine serum. After 16 -20 h when cells were 50 -60% confluent, siRNA duplexes and/or reporter gene constructs were transfected using Oligofectamine Reagent (Invitrogen). The effects of Tg and Sp siRNAs on transactivation was investigated in Panc-1 cells cotransfected with (500 ng) GRP78 constructs. Briefly, the siRNA duplex was transfected into each well to give a final concentration of 70 nM. Cells were harvested after 48 h, and luciferase activity of lysates (relative to ␤-galactosidase activity) was determined. For the ER stress study, cells were treated with Me 2 SO (control) or with the indicated concentration of Tg or Tm for 18 h. For electrophoretic mobility shift assay (EMSA), nuclear extracts from Panc-1 cells were isolated as previously described and aliquots were stored at Ϫ80°C until used (31,32).
Western Immunoblot-Cells were washed once with phosphate-buffered saline and collected by scraping in 200 l of lysis buffer (50 mM * This work was supported by National Institutes of Health Grants CA108178 and ES09106, M. D. Anderson Cancer Center Grant P20-CA-10193, and by the Texas Agricultural Experiment Station. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
FIG. 1. Activation of GRP78 by ER stress requires Sp family proteins. A, Tm and Tg induction of GRP78. Panc-1 cells were treated with Tg (0.5 M) or Tm (0.5 g/ml) for 20 h, and GRP78 expression was determined by Western blot analysis of whole cell lysates as described under "Materials and Methods." Results are expressed as means Ϯ S.D. for at least three separate determinations for each treatment group and significant (p Ͻ 0.05) induction of GRP78 protein (normalized to ␤-tubulin protein) is indicated (*). B, Effects of RNA interference of Sp1, Sp3 and Sp4 protein on induction of GRP78 by Tg and Tm. These experiments were carried out essentially as described for A except that Panc-1 cells were transfected with iScr (nonspecific control), iSp1, iSp3, or iSp4. Tg and Tm significantly induced GRP78 expression in all treatment groups; however, significant (p Ͻ 0.05) decreases after transfection with iSp1, iSp3, or iSp4 are indicated (*). The specificity of iSp3 is indicated in the Western blot for Sp3 protein in the various treatment groups. C, specificities of iSp3 and iSp4. This experiment was identical to that described in Fig. 1B and shows Western blot analysis of Sp4 (left) and Sp1 (right) proteins in cells transfected with iSp1 or iSp4. D, Sp protein knockdown decreases Tg-induced activation of pGRP78. Panc-1 cells were treated with Tg (0.5 M); cotransfected with iScr, iGL2 (luciferase), iSp1, iSp3, or iSp4; and luciferase activity determined as described under "Materials and Methods." Results are expressed as means Ϯ S.D. for three separate determinations for each treatment group, and significant (p Ͻ 0.05) induction by Tg (*) and inhibition by siRNAs (**) are indicated.

FIG. 2. Protein interactions with ERSE and related oligonucleotides.
A, 32 P-labeled ERSE binding to nuclear extracts from Panc-1 cells treated with Tg. Nuclear extracts from Me 2 SO or Tg-treated cells were incubated with 32 P-labeled ERSE alone or in combination with 100-fold excess unlabeled ERSE (100X) oligonucleotide, YY1 (␣YY1), or TFII-I (␣TFII-I) antibodies and separated in a gel mobility shift assay as described under "Materials and Methods." The specifically bound ERSF, NF-Y, and YY1 complexes are indicated by arrows. B, induction of TFII-I by Tg and gel mobility shift assay. Whole cell lysates from untreated (Ϫ) or Tg-treated Panc-1 cells were analyzed by Western blot analysis for TFII-I expression as described under "Materials and Methods." Nuclear extracts from Tg-treated Panc-1 cells were analyzed by gel mobility shift assays as described in Fig. 2A; however, increasing amounts of TFII-I antibody were used. C, supershift of ERSE-protein complex by Sp protein antibodies. Panc-1 cells were treated with Me 2 SO and Tg. Nuclear extracts were incubated with 32 P-labeled ERSE and Sp1 (␣Sp1), Sp3 (␣Sp3), or Sp4 (␣Sp4) antibodies and separated in a gel mobility shift assay as described under "Materials and Methods." ERSF, NF-Y, YY1, and Sp protein complexes HEPES, 0.5 M sodium chloride, 1.5 mM magnesium chloride, 1 mM EGTA, 10% (v/v) glycerol, 1% Triton X-100, and 5 l/ml of protease inhibitor mixture (Sigma)). The lysates from the cells were incubated on ice for 1 h with intermittent vortexing followed by centrifugation at 40,000 ϫ g for 10 min at 4°C. Equal amounts of protein from each treatment group were diluted with loading buffer, boiled, and loaded onto 10% SDS-polyacrylamide gel. Samples were electrophoresed, and proteins were detected by incubation with polyclonal primary antibodies Sp1 (PEP2), Sp3 (D-20), Sp4 (V-20), TFII-I, GRP78 (H-129), and ␤-tubulin (H-235) followed by blotting with appropriate horseradish peroxidase-conjugated secondary antibody as previously described (31,32). After autoradiography, band intensities were determined by a scanning laser densitometer (Sharp Electronics Corp., Mahwah, NJ) using Zero-D Scanalytics software (Scanalytics Corp., Billerica, MA).
EMSA-Wild-type ERSE (wtERSE), mutant ERSE (mERSE), and GC-rich oligonucleotides were synthesized and annealed, and 5-pmol aliquots were 5Ј-end-labeled using T4 kinase and [␥-32 P]ATP. A 30-l EMSA reaction mixture contained ϳ100 mM potassium chloride, 3 g of crude nuclear protein, or 1-2 band-forming units of human recombinant Sp proteins, 1 g of poly(dI-dC), with or without unlabeled competitor oligonucleotide, and 10 fmol of radiolabeled probe. After incubation for 20 min on ice, antibodies against selected proteins were added and incubated another 20 min on ice. Protein-DNA complexes were resolved by 5% polyacrylamide gel electrophoresis as previously described (31,32). Specific DNA-protein and antibody-supershifted complexes were observed as retarded bands in the gel. wtERSE and mERSE sequence used in gel shift analysis are given below. The NF-Y/CBF and YY1 motifs are underlined, and the mutation is in bold: human GRP78, GGG CCA ATG AAC GGC CTC CAA CGA (Ϫ94 ERSE (wt)) and GGG CCA ATG AAC TTA CTC CAA CGA (Ϫ94 ERSE (m)).
Chromatin Immunoprecipitation (ChIP) Assay-Panc-1 cells (2 ϫ 10 7 cells) were treated with Me 2 SO (time 0) or 1 g/ml Tm for the indicated times. Cells were then fixed with 1.5% formaldehyde, and the crosslinking reaction was stopped by addition of 0.125 M glycine. After washing twice with phosphate-buffered saline, cells were scraped and pelleted. Collected cells were hypotonically lysed, and nuclei were collected. Nuclei were then sonicated to the desired chromatin length (500 -1000 bp). The chromatin was precleared by addition of protein A-conjugated beads and then incubated at 4°C for 1 h with gentle agitation. The beads were pelleted, and the precleared chromatin supernatant was immunoprecipitated with antibodies to IgG, Sp1, Sp3, Sp4, YY1, NF-Y, TFII-I, TFIIB, and ATF6 at 4°C overnight. The protein-antibody complexes were collected by addition of protein A-conjugated beads at room temperature for 1 h; the beads were extensively washed, and protein-DNA cross-links were reversed. DNA was purified by phenol/chloroform extraction and ethanol precipitation, and PCR was performed on the purified material. The GRP78 primers were as follows: 5Ј-CGG AGC AGT GAC GTT TAT TG-3Ј (forward) and 5Ј-ACC TCA CCG TCG CCT ACT C-3Ј (reverse); and they amplified a 165-bp region of the human GRP78 promoter, which contains the ERSEs. The positive control primers were 5Ј-TAC TAG CGG TTT TAC GGG CG-3Ј (forward) and 5Ј-TCG AAC AGG AGG AGC AGA GAG CGA-3Ј (reverse) and amplified a 167-bp region of the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene. The negative control primers were 5Ј-ATG GTT GCC ACT GGG GAT CT-3Ј (forward) and 5Ј-TGC CAA AGC CTA GGG GAA GA-3Ј (reverse) and amplified a 174-bp region of human CNAP1 exon. PCR products were resolved on a 2% agarose gel in the presence of 0.5 g/ml ethidium bromide as described (33).
Statistical Analysis-Statistical significance was determined by analysis of variance and Scheffe's test, and the levels of probability are noted. The results are expressed as means Ϯ S.D. for at least three separate (replicate) experiments for each treatment.

RESULTS
ERSEs from various ER stress-responsive genes exhibit a tripartite structure, which includes two motifs that bind NF-Y/CBF and YY1 and flank a central nonanucleotide (N 9 ) sequence that shows some variability in different ERSEs (28). N 9 motifs from most ERSEs are GC-rich, and we hypothesized that Sp family proteins may directly or indirectly influence ER stress-induced gene/protein expression. Results summarized in Fig. 1 show that Tm (0.5 g/ml) and Tg (0.5 M) induced GRP78 protein in Panc-1 pancreatic cancer cells, whereas levels of ␤-tubulin were unaffected by ER stress (Fig. 1A). The potential role of Sp proteins on ER stress-induced GRP78 protein expression was determined in Panc-1 cells transfected with small inhibitory RNAs for Sp1 (iSp1), Sp3 (iSp3), and Sp4 (iSp4) and a nonspecific scrambled (iScr) RNA. Western blot analysis of whole cell lysates from the various treatment groups showed that iScr did not affect Tg-/Tm-induced GRP78 protein, whereas significant inhibition of GRP78 protein levels was observed in lysates from cells transfected with iSp1, iSp3, and iSp4 (Fig. 1B). The order of effectiveness of these siRNAs in decreasing GRP78 protein levels was iSp4 Ͼ iSp1 Ͼ iSp3. The results in Fig. 1, B and C, also demonstrate that neither Tg or Tm affected Sp protein levels and that the RNA interference assay for Sp protein knockdown is highly selective for Sp1, Sp3, and Sp4 as reported previously (31)(32)(33). These data clearly demonstrate that ER stress-induced GRP78 protein expression is dependent on Sp1, Sp3, and Sp4.
The direct effects of Sp proteins on GRP78 was further investigated in Panc-1 cells transfected with the plasmid pGRP78, which contains a Ϫ374-bp insert derived from the proximal region of the human GRP78 gene promoter. This promoter sequence contains three ERSEs, and treatment with Tg significantly induced transactivation (Ͼ3-fold) (Fig. 1D), and this paralleled the induction of GRP78 protein (Fig. 1A). In cells treated with Tg and cotransfected with pGRP78 and iSp1, iSp3, or iSp4, both basal and induced luciferase activities were significantly decreased (Fig. 1D). These results further confirm that Sp1, Sp3, and Sp4 play an important role in regulation of a prototypical stress induced gene/protein GRP78. We further examined the role of Sp proteins in stress responses by gel mobility shift assays using the 32 P-labeled ERSE, which was identical to ERSE-2 from the GRP78 promoter. The pattern of three retarded bands associated with the YY1 and NF-Y or ERSF-DNA complexes in extracts from control (Me 2 SO) or Tg ( Fig. 2A) was similar to that reported previously (28). All three bands were decreased after coincubation with 100-fold excess of unlabeled ERSE (lane 5) and YY-1 antibodies immunodepleted the YY-1-ERSE complex (lane 7). Previous studies report that TFII-I was induced by ER stress and bound the ERSF-ERSE complex, and this contributed to the enhanced intensity of the retarded band (28). Incubation of stress-induced Panc-1 cell lysate with 32 P-labeled ERSE and TFII-I antibodies decreased intensity of the ERSF-ERSE band (immunodepletion), although a supershifted band was not detected. Tg also induces TFII-I protein levels in Panc-1 cells as reported previously in other cell lines (28), and gel mobility shift assays show that the intensity of the Tg-induced ERSF-DNA complex (lane 6) is significantly decreased after coincubation with increasing amounts of TFII-I antibody (lanes 3-5) (Fig. 2B). Even under these conditions, a supershifted complex was not observed.
Sp proteins interact directly or indirectly with other transcription factors to modulate gene/protein expression (34). Direct interaction of Sp1 protein with the ERSE was investigated in a gel mobility shift assays using 32 P-labeled ERSE extracts from Tg-treated Panc-1 cells and Sp1, Sp3, and Sp4 antibodies (Fig. 2C, lanes 4 -6). Distinct supershifted bands were observed using Sp1, Sp3, and Sp4 antibodies (lanes 4 -6), and the most pronounced decrease in the intensity of the ERSF-ERSE complex was observed with the Sp1 antibody suggesting that Sp1 are indicated by arrows, and the supershifted Sp protein complexes are indicated. D, binding of Sp1 protein to different radiolabeled oligonucleotides. Recombinant Sp1 protein (1 or 2 band-forming units) was incubated with 32 P-labeled ERSE, mERSE, or GC oligonucleotides alone or in combination with 100-fold excess unlabeled ERSE oligonucleotide, ␣IgG, or Sp1 (␣Sp1) antibodies and separated in a gel mobility shift assay as described under "Materials and Methods." Sp1-DNA complexes and supershifted complexes are indicated by arrows.
protein is the major Sp protein bound to the ERSF-ERSE complex. These results indicate that Sp proteins may directly bind the ERSE, and this was further confirmed in gel mobility shift assays using a series of radiolabeled oligonucleotides and recombinant Sp1 protein. Previous studies indicate that the consensus GC-rich oligonucleotide binds Sp1, Sp3, and Sp4 with high affinity (31)(32)(33)(34), and the results illustrated in Fig. 2D compare the binding of recombinant human Sp1 protein with 32 P-labeled GC (lanes 1-4), ERSE (lanes 7, 8, and 10 -12), and mutant ERSE (lane 9). Using the same amount of Sp1 protein, it was clear that the Sp1-GC retarded band was significantly more intense than the Sp1-ERSE band reflecting different Sp protein-DNA binding affinities. However, the results show that Sp1 directly binds the GC and ERSE oligonucleotides, and both retarded bands were supershifted by Sp1 antibody (lanes 4 and 10) but not by nonspecific IgG (lane 11). The mutant ERSE, which did not bind Sp1 (lane 9), contained three specific mutations (underlined) in the nonanucleotide sequence (GAA CTT ACT) in which a TTA sequence replaced the wild-type GGC. Parker et al. (28) have previously demonstrated the high conservation of the GC-rich sequences in ERSEs from GRP78 and other stress-related genes, and mutation of the GGC sequence greatly decreased stress-induced responses in transactivation assays (16,35). The transactivation studies correlated with the loss of Sp protein-ERSE binding observed in this study (Fig.  2D) and complement the functional studies showing that Sp proteins are critical for activation of stress-induced GRP78 (Fig. 1).
Previous studies indicate that induction of ER stress is accompanied by cleavage of p90 ATF-6 to p60 ATF-6, a nuclear transcription factor that interacts with NF-Y proteins (21,22). TFII-I is also induced by ER stress and interacts with ATF-6 to form part of the ERSE-protein complex (28). We examined Tm-induced protein interactions with the ERSEs of the GRP78 gene promoter by ChIP (Fig. 3A). As a positive control for this experiment, we show that TFIIB interacts with a specific region of the human GAPDH promoter but not the CNAP1 gene exon (Fig. 3B) (33). Treatment of Panc-1 cells with Tm for 0 (Me 2 SO), 15, 30, and 60 min, followed by ChIP assay, showed that Sp1, Sp3, Sp4, NF-Y, and YY1 were constitutively bound to this region of the GRP78 promoter. There were only minimal temporal increases in Sp1 and YY1 binding. In contrast, TFII-I and ATF-6 were not constitutively bound to the ERSE region of the GRP78 promoter but were rapidly recruited to the promoter after treatment with Tm, and both proteins remained bound for up to 60 min. The induction of ATF6 and TFII-I by ER stress (Tm) and their recruitments to the GRP78 promoter in Panc-1 cells are consistent with the important role of these proteins in mediating a stress-induced response (21,22,28). We also investigated the ERSE binding of nuclear extracts from untreated (lane 1) and Tg-treated (lanes 2-5) Panc-1 cells cotransfected with nonspecific scramble RNA (iScr, lanes 1 and 2) and iSp1, iSp3, or iSp4 (lanes 3-5, respectively) (Fig. 3C). The results showed that Sp protein knockdown decreased but did not eliminate intensity of the ERSF-DNA complex, suggesting that stress-induced proteins are retained as part of the complex and do not require Sp proteins for binding. DISCUSSION ER stress has been linked to various diseases, and several drugs act through activating ER stress pathways, which can lead to growth inhibition and apoptosis (1)(2)(3)(4)(5)(13)(14)(15). Activation of ER stress pathways is accompanied by modulation of several transcriptional and translational pathways, which include induction of several stress-response genes including GRP78 and GADD153 (CHOP). The ERSE is a major cis-element that is involved in activation of stress-responsive genes, and interac- Interaction of transcription factors with the GRP78 promoter was determined in Panc-1 cells treated with Tm for 15, 30, and 60 min and Me 2 SO, which serves as a control. Sonicated nuclear extracts were immunoprecipitated with IgG (nonspecific control) and antibodies for Sp1, Sp3, Sp4, NF-Y and YY1, TFII-I and ATF6, and specific protein interactions with the ERSE region of GRP78 were determined by PCR as described under "Materials and Methods." B, TFIIB binding to GAPDH. Using the same ChIP assay procedure (outlined in Fig. 3A), the control experiment determined interactions of TFIIB with the GAPDH promoter, but not the CNAP1 exon, as described (33). C, effects of Sp protein knockdown. Nuclear extracts from cells treated with iScr alone, Tg ϩ iScr, iSp1, iSp3, or iSp4 were analyzed in a gel mobility shift assay using 32 P-labeled ERSE as described under "Materials and Methods." D, proposed model for ER stress-dependent transactivation. Sp proteins facilitate interactions with basal transcription factors (BTF) and other nuclear factors. tion of trans-acting nuclear transcription factors with the ERSE is responsible for target gene activation. Interaction of NF-Y/CBF and YY1 with the two end-flanking motifs of the ERSE has been well characterized (16,17), and both ATF-6 and TFII-I have also been identified as components of the ERSEprotein complex (19 -22, 28). The inner N 9 sequence in most ERSEs are GC-rich and are required for maximal stress-dependent transactivation (28,35). Sp family proteins Sp1, Sp3, and Sp4 are highly expressed in Panc-1 and other cancer cell lines (31), and we hypothesized that one or more Sp proteins, which bind GC-rich elements, may be required for ER stress-induced responses. RNA interference with siRNAs for Sp1, Sp3, and Sp4 clearly demonstrate that these proteins are involved in Tg-/Tminduced GRP78 protein and reporter gene expression in Panc-1 cells transfected with pGRP78 (Fig. 1). Moreover, the results of both ChIP and electrophoretic mobility shift assays demonstrate that Sp proteins interact with the ERSE, and this is observed in both stress and unstressed cells (Figs. 2 and 3).
These results demonstrate for the first time that Sp1, Sp3, and Sp4 proteins play a key role in mediating stress-dependent activation of GRP78 in Panc-1 cells. Sp family proteins form a critical part of the ERSF, binding directly to the ERSE and cooperatively activating ER stress-dependent responses. RNA interference studies suggest some redundancy for Sp1, Sp3, and Sp4 in mediating stress-induced GRP78 expression, indicating that constitutive expression of one or more of these transcription factors may be sufficient for enhanced transactivation. A model for the responses observed in this study is summarized in Fig. 3D. Sp proteins bind the nonanucleotide region of the ERSE and interact with GC-rich motifs in untreated and stress-induced cells. Previous studies show that ATF6 and TFII-I also bind directly to the ERSE (19 -22, 28), and it is possible that Sp proteins may function to enhance interactions of the complex with basal transcription factors or other nuclear factors that induce transcriptional activation of stress genes. Current studies are focused on the cell contextdependent role of specific Sp proteins in activating stressinduced genes with variable ERSEs and determining interactions with other transcription factors on these motifs.