Identification of ERSE-II, a New cis-Acting Element Responsible for the ATF6-dependent Mammalian Unfolded Protein Response*

  • Koichi Kokame
    Correspondence
    To whom correspondence should be addressed
    Affiliations
    National Cardiovascular Center Research Institute, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan
    Search for articles by this author
  • Hisao Kato
    Affiliations
    National Cardiovascular Center Research Institute, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan
    Search for articles by this author
  • Toshiyuki Miyata
    Affiliations
    National Cardiovascular Center Research Institute, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan
    Search for articles by this author
  • Author Footnotes
    * This work was supported in part by grants-in-aid from the Ministry of Health and Welfare of Japan; from the Ministry of Education, Science, Sports, and Culture of Japan; from the Japan Society for the Promotion of Science; and by the Program for Promotion of Fundamental Studies in Health Sciences of the Organization for Pharmaceutical Safety and Research of Japan.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.The nucleotide sequence reported in this paper has been submitted to the GenBank™/EMBL/DDBJ Data Bank with accession number .
      Herp is a 54-kDa membrane protein in the endoplasmic reticulum (ER). The mRNA expression level of Herp is increased by the accumulation of unfolded proteins in the ER. Transcriptional changes designed to deal with this type of ER stress is called the unfolded protein response (UPR). Most mammalian UPR-target genes encode ER-resident molecular chaperones: GRP78, GRP94, and calreticulin. The promoter regions of these genes contain acis-acting ER stress response element, ERSE, with the consensus sequence of CCAAT-N9-CCACG. Under conditions of ER stress, p50ATF6 (the active form of the transcription factor, ATF6) binds to CCACG when CCAAT is bound by the general transcription factor, NF-Y/CBF. Here, we report the genomic structure of human Herp and the presence of a new ER stress response element, ERSE-II, in its promoter region. The gene for Herp consists of eight exons, localized to chromosome 16q12.2–13. The promoter region contains a single ERSE-like sequence. In reporter gene assays, disruption of thiscis-element resulted in a partial reduction of the transcriptional response to ER stress, suggesting that the element is functional for the UPR. These results also suggest the involvement of additional elements in the UPR. Further analysis, using an optimized plasmid containing an mRNA-destabilizing sequence, revealed ERSE-II (ATTGG-N-CCACG) as the second ER stress response element. Interestingly, ERSE-II was also dependent on p50ATF6, in a manner similar to that of ERSE, despite the disparate structure. The strong induction of Herp mRNA by ER stress would be achieved by the cooperation of ERSE and ERSE-II.
      ER
      endoplasmic reticulum
      UPR
      unfolded protein response
      ERSE
      ER stress response element
      FISH
      fluorescence in situ hybridization
      HUVEC
      human umbilical vein endothelial cell
      UTR
      untranslated region
      GM-CSF
      granulocyte-monocyte colony-stimulating factor
      MMS
      methyl methanesulfonate
      bp
      base pair(s)
      PCR
      polymerase chain reaction
      PBS
      phosphate-buffered saline
      The expression of Herp, a 54-kDa endoplasmic reticulum (ER)1-resident protein, is induced by the unfolded protein response (UPR) (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ). The UPR is a transcriptional response to remedy the accumulation of unfolded proteins in the ER. Most previously identified UPR-target genes encode ER-resident molecular chaperones and folding enzymes, such as GRP78/BiP, GRP94, protein-disulfide isomerase, and calreticulin. In contrast to these ER luminal proteins, Herp is an integral membrane protein, both N and C termini of which face the cytoplasmic side of the ER. This membrane topology makes it unlikely that Herp acts as a molecular chaperone for proteins in the ER. Herp plays an unknown role in the cellular survival response to the accumulation of unfolded proteins in the ER.
      The ER provides an optimal environment for the synthesis, folding, and assembly of membrane and secreted proteins. The accumulation of unfolded or misfolded proteins in the ER under conditions of “ER stress” threatens the normal functioning of eukaryotic cells. Although the physiological conditions inducing ER stress are not fully understood, the cellular response to the stress is essential for homeostasis (comprehensively reviewed by Kaufman (Ref.
      • Kaufman R.J.
      )). The ER-stress responses are currently categorized to three mechanisms: transcriptional induction, translational attenuation, and degradation (reviewed by Mori (Ref. )). In addition, ER stress activates c-Jun N-terminal kinases (
      • Urano F.
      • Wang X.
      • Bertolotti A.
      • Zhang Y.
      • Chung P.
      • Harding H.P.
      • Ron D.
      ) and induces caspase-12-mediated apoptosis (
      • Nakagawa T.
      • Zhu H.
      • Morishima N.
      • Li E.
      • Xu J.
      • Yankner B.A.
      • Yuan J.
      ).
      The molecular mechanism of the UPR is extensively defined in the yeast,Saccharomyces cerevisiae. The ER luminal domain of Ire1p, an ER-resident type I transmembrane protein, senses the accumulation of unfolded proteins in the ER, activating its cytoplasmic endoribonuclease domain through homo-oligomerization and trans-autophosphorylation (
      • Cox J.S.
      • Shamu C.E.
      • Walter P.
      ,
      • Mori K.
      • Ma W.
      • Gething M.J.
      • Sambrook J.
      ). Activated Ire1p triggers the production of a transcription factor, Hac1p, through an unusual mRNA splicing mechanism (
      • Cox J.S.
      • Walter P.
      ,
      • Kawahara T.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Mori K.
      • Ogawa N.
      • Kawahara T.
      • Yanagi H.
      • Yura T.
      ). Hac1p binds to the UPR-dependent cis-acting element, UPRE, in the promoter regions of UPR-target genes to activate their transcription (
      • Cox J.S.
      • Walter P.
      ,
      • Mori K.
      • Kawahara T.
      • Yoshida H.
      • Yanagi H.
      • Yura T.
      ,
      • Mori K.
      • Ogawa N.
      • Kawahara T.
      • Yanagi H.
      • Yura T.
      ). In mammals, two Ire1p orthologs, IRE1α and IRE1β, have been identified (
      • Tirasophon W.
      • Welihinda A.A.
      • Kaufman R.J.
      ,
      • Wang X.Z.
      • Harding H.P.
      • Zhang Y.
      • Jolicoeur E.M.
      • Kuroda M.
      • Ron D.
      ), although the substrates of their endoribonuclease activities are unknown. ATF6 has been identified as the transcription factor responsible for the UPR (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). The quiescent form of ATF6 (p90ATF6), a type II-transmembrane protein, is embedded in the ER membrane and proteolyzed in an ER stress-dependent manner (
      • Haze K.
      • Yoshida H.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). The liberated N-terminal fragment (p50ATF6) translocates to the nucleus, binding to the mammalian UPR-dependentcis-acting element, designated the ER stress response element (ERSE) (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ).
      ERSE, possessing a consensus sequence of CCAAT-N9-CCACG, is necessary and sufficient for the induction of at least three major ER chaperones (GRP78, GRP94, and calreticulin) (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). This sequence is present in the proximal promoter regions of many ER stress-responsive proteins (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Roy B.
      • Lee A.S.
      ). The general transcription factor, NF-Y/CBF, binds to the CCAAT motif of ERSE (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ,
      • Roy B.
      • Lee A.S.
      ,
      • Roy B.
      • Li W.W.
      • Lee A.S.
      ). Under conditions of ER stress, p50ATF6 binds to the CCACG motif of ERSE, resulting in the transcriptional induction of ER chaperones (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). Multiple ER stress-responsive genes, however, possess proximal promoter regions without an ERSE sequence, such as FKBP13 (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ), asparagine synthetase (
      • Barbosa-Tessmann I.P.
      • Chen C.
      • Zhong C.
      • Siu F.
      • Schuster S.M.
      • Nick H.S.
      • Kilberg M.S.
      ), ATF3 (
      • Cai Y.
      • Zhang C.
      • Nawa T.
      • Aso T.
      • Tanaka M.
      • Oshiro S.
      • Ichijo H.
      • Kitajima S.
      ), and RTP/NDRG1.
      K. Kokame, H. Kato, and T. Miyata, unpublished data.
      2K. Kokame, H. Kato, and T. Miyata, unpublished data.
      In the present paper, we demonstrate the existence of a new ER stress response element, ERSE-II, found in the Herp promoter region. In a manner similar to ERSE, ERSE-II mediates the ATF6-dependent UPR.

      EXPERIMENTAL PROCEDURES

       Cloning and Sequencing of the Human Herp Gene

      A human whole blood, λ genomic library (Stratagene) was screened to obtain genomic clones encoding Herp using the PCR-based screening method described previously (
      • Kokame K.
      • Kato H.
      • Miyata T.
      ). Positive phages were cloned utilizing theEscherichia coli XL1-Blue MRA strain as host. PCR was performed using 5′-TGGTTTCTCCGGTTACAC-3′ and 5′-AGAGACCACAGGTATCTC-3′ as primers with the plate lysates as templates. Two positive phages were cloned by limiting serial dilution. The insert DNAs isolated from these clones (∼17 kilobases each) were sequenced using a BigDye Terminator Cycle Sequencing FS Ready Reaction Kit (PerkinElmer Life Sciences).

       Fluorescence in Situ Hybridization (FISH) Analysis

      The P1-derived artificial chromosome clone containing the Herp gene was isolated from a human PAC DNA library (GenomeSystems) using Herp cDNA as a probe. DNA from the clone, labeled by nick translation with digoxigenin dUTP, was hybridized to normal metaphase chromosomes derived from phytohemagglutinin-stimulated peripheral blood lymphocytes. Specific hybridization signals were detected using fluorescein-conjugated anti-digoxigenin antibodies, followed by counterstaining with 4′,6-diamidino-2-phenylindole dihydrochloriden-hydrate.

       Construction of Plasmids

      Progressive deletion fragments of the Herp gene 5′-flanking region were PCR-amplified using sense primers containing an additional 5′-BglII site and an antisense primer containing an original NcoI site at the initial Met of Herp (5′-TTCGGTCTCGGACTCCATGGC-3′). After digestion withBglII and NcoI, the fragments were inserted between the BglII and NcoI sites of the firefly luciferase reporter plasmid, pGL3-Basic (Promega). Site-directed mutations were introduced into the inserts by using oligonucleotide primers containing the desired mutations, according to the QuikChange site-directed mutagenesis kit protocol (Stratagene). All inserts were sequenced to confirm the desired sequence.

       Cell Culture and Luciferase Assay

      Human umbilical vein endothelial cells (HUVECs, Clonetics) were cultured on 24-well plates coated with type I collagen (Sumitomo Bakelite) in MCDB131 medium (Life Technologies, Inc.), supplemented with 10 mm glutamine (Life Technologies, Inc.), 20 mm Hepes-NaOH (pH 7.4), 2% fetal bovine serum (Life Technologies, Inc.), and 10 ng/ml human basic fibroblast growth factor (R&D Systems). Using 1.05 μl/well FuGENE6 transfection reagent (Roche Molecular Biochemicals), we transfected HUVECs with 0.5 μg/well amounts of either the pGL3-Basic-derived plasmid described above or control plasmid (pGL3-Control, firefly luciferase vector with SV40 promoter and enhancer sequences, Promega) together with 0.025 μg/well of the internal control plasmid (pRL-SV40, Renilla reniformis luciferase vector with SV40 promoter and enhancer sequences, Promega). Following a 23-h incubation, cells were incubated for 6 h in either 1 μmthapsigargin (Sigma), 10 μg/ml tunicamycin (Sigma), or 10 mm 2-mercaptoethanol (Nacalai Tesque). Cells were then washed with Dulbecco's PBS (Life Technologies, Inc.) and harvested in 100 μl of Passive Lysis Buffer (Promega). We measured the firefly andRenilla luciferase activities of 20 μl of each lysate using a Dual-Luciferase reporter assay system (Promega). Bioluminescence was detected using a LUMINOUS CT-9000 luminometer (Dia-Iatron). After dividing luminescence intensity of firefly luciferase by that of Renilla luciferase, we determined the “relative luciferase activity” to be the ratio of the value obtained from each test plasmid to that of the pGL3-Control. In each assay, the values were averaged from four independent wells.

       Insertion of AT-rich Sequence to the 3′-Untranslated Region (UTR) of the Luciferase Gene

      To make the luciferase mRNA unstable, we inserted a synthetic double-stranded oligonucleotide, 5′-TAATATTTATATATTTATATTTTTAAAATATTTATTTATTTATTTATTTAA-3′, into the XbaI sites of both the pGL3-Basic and pGL3-Control plasmids. This AT-rich sequence was derived from 3′-UTR of granulocyte-monocyte colony-stimulating factor (GM-CSF).

       Transient Expression of ATF6(366)

      The cDNA encoding human ATF6 was the kind gift of Dr. Hiderou Yoshida and Dr. Kazutoshi Mori (Kyoto University, Kyoto, Japan). The partial open reading frame corresponding to Met1–Asn366 was PCR-amplified; the product was inserted into the mammalian expression vector, pcDNA3.1(+) (Invitrogen). To express ATF6(366) transiently, the resultant plasmid pcDNA3ATF6(366) (0.01 μg/well) was transfected to HUVECs together with the luciferase plasmids (0.5 μg/well). As a negative control, the mock vector pcDNA3.1(+) (0.01 μg/well) was cotransfected in place of pcDNA3ATF6(366).

       Fluorescent Immunocytochemistry

      We constructed an expression plasmid coding for a FLAG-tagged version of ATF6 Met1–Asn366 and transfected this plasmid into HUVECs. After a 21-h incubation, cells were rinsed with Dulbecco's PBS, fixed in 2% paraformaldehyde for 15 min, and permeabilized with 0.05% Triton X-100 for 2 min. Following an incubation in 5% normal goat serum and 5% fish gelatin for 30 min, we detected endogenous Herp and transiently expressed FLAG-ATF6(366) simultaneously in 1-h incubation of 20 μg/ml anti-Herp rabbit polyclonal antibody (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ) and 10 μg/ml anti-FLAG M2 mouse monoclonal antibody (Eastman Kodak Co.). Cells were then incubated with Oregon Green 514-conjugated goat anti-rabbit IgG (Molecular Probes) and Rhodamine Red-X-conjugated goat anti-mouse IgG (Molecular Probes) for 1 h. After washing with PBS, fluorescence was visualized using a confocal laser-scanning microscope with FLUOVIEW (Olympus).

      RESULTS

       Genomic Structure of Human Herp

      We obtained a complete sequence of the human Herp gene (GenBank™ accession no. AB034990) by comparison of two partially overlapping clones, isolated from the human genomic DNA library, to the Herp cDNA sequence (GenBank™ accession no. AB034989). The gene was officially designatedHERPUD1 by the HUGO Gene Nomenclature Committee. The schematic structure, sequences across the exon-intron junctions, and the sizes of exons and introns are shown in Fig.1 (A and B). The Herp gene contains eight exons and spans 11,738 bp in length. The identified exon-intron junctions agreed with the intron 5′-GT and 3′-AG consensus sequences. The 5′-terminal transcription start site had been previously determined by cap-site hunting (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ). Exon 1 encoded the 5′-UTR and the first 49 N-terminal residues including the initial Met codon. The stop codon and the 3′-UTR were encoded by exon 8.
      Figure thumbnail gr1
      Figure 1Human Herp gene. A, genomic organization. Exons are indicated by black boxes andnumbered. The initial Met and stop codons appear in exons 1 and 8, respectively. B, exon-intron boundaries. The exon sequences, deduced from comparison to the Herp cDNA, are given inuppercase letters, whereas the intron sequences are inlowercase letters. All introns start with GT and end with AG. The potential polyadenylation signal is underlined, and the polyadenylation site at the 3′ end of exon 8 is indicated by anarrowhead. The complete sequence data have been submitted to the GenBank™/EMBL/DDBJ data bases under the accession numberAB034990. The gene has been officially designated HERPUD1 by the HUGO Gene Nomenclature Committee. C, FISH analysis. Labeled DNA containing the Herp gene was hybridized to metaphase chromosomes. The hybridization signals on chromosome 16 are indicated by arrows. D, sequence of 5′-flanking region. Exon 1 isshaded. Negative numbers indicate the distance of the nucleotide from the transcription start site. Putative regulatory motifs are labeled with lines. The motif corresponding to ERSE is shown in black-boxed white letters.
      To localize the Herp gene on human chromosomes, we performed FISH analysis utilizing DNA from the P1-derived artificial chromosome clone containing the Herp gene. Labeled Herp DNA was hybridized to chromosomes derived from peripheral blood lymphocytes. Eighty metaphase cells were analyzed; 73 exhibited specific labeling of the 16q12.2–13 region (Fig. 1 C).

       Sequence of the 5′-Flanking Promoter Region

      We sequenced the ∼6-kilobase pair 5′-flanking region of the Herp gene. Computer analysis by TFSEARCH using the TRANSFAC data base (
      • Heinemeyer T.
      • Wingender E.
      • Reuter I.
      • Hermjakob H.
      • Kel A.E.
      • Kel O.V.
      • Ignatieva E.V.
      • Ananko E.A.
      • Podkolodnaya O.A.
      • Kolpakov F.A.
      • Podkolodny N.L.
      • Kolchanov N.A.
      ) revealed many potential transcription factor-binding sites within the sequence. The proximal 200-bp sequence upstream of the transcriptional start site, including several putative cis-acting regulatory elements, is shown in Fig. 1 D. The canonical TATA box, specifying the transcriptional start site, is found in close proximity to exon 1. Two CAAT boxes were also identified. The 5′-flanking region contained several GC boxes (GGCG), suggesting multiple Sp1-binding sites.
      The Herp promoter region contains one ERSE-like sequence,−88CCAATGGGCGGCAGCCACA−70, located upstream of the TATA box (Fig. 1 D). ERSE is a cis-acting regulatory element identified in the promoters of mammalian UPR target genes (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). ERSE, with a consensus of CCAAT-N9-CCACG, is necessary and sufficient for the induction of the ER-resident molecular chaperones, GRP78, GRP94, and calreticulin. Although the G nucleotide at the 3′ end of the consensus sequence is replaced by an A in the Herp ERSE, we predict this sequence functions in the UPR-dependent induction of Herp expression at the transcriptional level.

       Functional Mapping of the Herp Promoter

      A series of reporter plasmids containing sense fragments of the Herp 5′-flanking region (from nucleotide −5000 to −200) upstream of the firefly luciferase gene were transfected into HUVECs. The firefly luciferase activity in each assay was normalized to a cotransfected Renillaluciferase plasmid, pRL-SV40, to compensate for a varied efficiency of transfection.
      The basal luciferase activity of plasmid containing the longest 5′-flanking sequence (−5000/+98) exhibited approximately half the activity of an SV40 promoter control (Fig.2). Thapsigargin, an inhibitor of ER-resident Ca2+-ATPase, is used experimentally to activate the UPR. Following a 6-h treatment with 1 μmthapsigargin, luciferase activity increased significantly (∼4.3-fold) over basal activity, consistent with previous results demonstrating the induction of Herp mRNA by thapsigargin (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ). The SV40 promoter encoded by the pGL3-Control vector did not respond to thapsigargin treatment. Removal of the −5000 to −1800 region of Herp resulted in an increase of basal activity, suggesting the existence of silencing element in the region; little effect, however, was observed in the response to thapsigargin. Both the basal and thapsigargin-treated activities of plasmids containing −1000/+98, −800/+98, −600/+98, and −400/+98 were similar in magnitude to −1800/+98. Although removal of the −400 to −200 region resulted in a reduction of basal activity, the strong induction of luciferase activity in response to thapsigargin remained intact. We, therefore, concluded that thecis-elements responsible for the response to thapsigargin treatment would lie within the region 200 bp upstream of the transcription start site. In the following experiments, we used a plasmid containing the −200/+98 region to analyze this hypothesis in detail.
      Figure thumbnail gr2
      Figure 2Map of the functional elements in the Herp promoter. HUVECs were transfected with luciferase reporter plasmids containing progressive 5′ end deletions from nucleotide −5000, made from pGL3-Basic and the Herp gene. After 23 h, the cells were incubated with or without 1 μm thapsigargin for 6 h and assayed for luciferase activity. Basal and thapsigargin-treated activities are shown by open andclosed bars, respectively. Values averaged from four independent experiments are shown with standard deviations relative to the basal activity of pGL3-Control containing the SV40 promoter.Tg, thapsigargin.

       Disruption of ERSE in the Herp Promoter

      One ERSE-like sequence, −88CCAATgggcggcagCCACA−70, is contained within the Herp 5′-flanking region. As the A nucleotide at the 3′ end was different from a G in the ERSE consensus, CCAAT-N9-CCACG, we examined the transcriptional effect of this nucleotide difference. Both basal and thapsigargin-treated activities of the plasmid containing CCACg (Fig. 3 A,line 2) were similar to those of the original plasmid (line 1), suggesting that the A nucleotide functions similarly to a G nucleotide in the response to thapsigargin. We performed site-directed mutagenesis on two motifs of ERSE and examined the effects on the transcriptional induction following thapsigargin treatment. Throughout this paper, the term “mutation” is defined as the substitution of A, C, G, and T for C, A, T, and G, respectively. Disruptive mutation of either of the two motifs, CCAAT or CCACA, resulted in a partial reduction of the thapsigargin-dependent induction of luciferase activity (Fig. 3 A, lines 3 and 4), indicating their involvement in the induction. Mutation of both motifs, however, did not completely abrogate the response to thapsigargin (line 5). These results suggest that other cis-elements are involved in thapsigargin-dependent transcriptional induction. Under these experimental conditions, however, the observed inducibilities were not high enough to define the elements. We, therefore, modified the plasmid DNAs to effectively monitor the difference in activity with or without thapsigargin treatment.
      Figure thumbnail gr3
      Figure 3Disruption of ERSE and optimization of the reporter plasmid. A, mutational effects of ERSE on Herp promoter activity. The ERSE-like sequence in the Herp 5′ region, CCAATgggcggcagCCACA, was mutated to CCAATgggcggcagCCACg(line 2), aaccggggcggcagCCACA (line 3), CCAATgggcggcagaacac (line 4),aaccggggcggcagaacac (line 5); the luciferase activities of each were then measured. Basal and thapsigargin-treated activities are shown in open andclosed bars, respectively. B, effect of an mRNA-destabilizing sequence on induction. The AT-rich sequence in the 3′-UTR of GM-CSF was inserted downstream of the luciferase gene of both the plasmid containing the −200/+98 region of Herp (line 2) and the pGL3-Control plasmid (line 4). We compared the luciferase activities of these plasmids to the parent plasmids (lines 1 and 3) in the absence (open bars) or presence (closed bars) of thapsigargin.Tg, thapsigargin.

       Optimization of the Reporter Plasmid to Monitor the Induction Effectively

      Observation of the effects of stimulants on transcriptional induction in reporter gene assays is contingent on a faster turnover of mRNA produced from the test plasmid DNA. We, therefore, introduced an AT-rich sequence into the 3′-UTR of the firefly luciferase plasmids. The 51-nucleotide stretch (TAATATTTATATATTTATATTTTTAAAATATTTATTTATTTATTTATTTAA), known to selectively destabilize mRNA, was identified from the 3′-UTR of GM-CSF cDNA (
      • Shaw G.
      • Kamen R.
      ). Insertion of this sequence into the luciferase 3′-UTR of the plasmid containing the Herp −200/+98 region resulted in dramatic reduction of the basal activity; the activity in the presence of thapsigargin was relatively unchanged (compare lines 1and 2 in Fig. 3 B). As a result, the induction rate of thapsigargin treatment increased from 2.8 to 7.7 in this assay. Insertion of the AT-rich sequence into the control pGL3-Control plasmid, containing the SV40 promoter, had little effect on the ratio of basal to thapsigargin-treated activities (1.1 to 1.4, lines 3 and 4), although the luciferase activities were reduced in both cases. We utilized this optimized plasmid to identify additional transcriptional control elements in the Herp promoter region.

       Identification of ERSE-II

      To identify additionalcis-elements involved in thapsigargin-induction, we made a series of mutant plasmids that also contained the disrupted ERSE. First, we searched the region from nucleotide −196 to −89, making 11 sets of consecutive 10-bp mutations. After measuring the resulting luciferase activities (Fig.4 A), we found that basal activities were reduced when two regions,−186GCGGGTTGCA−177 and−176TCAGCCCGTG−167 were mutated, although the induction by thapsigargin treatment remained intact (lines 4and 5). Mutation of−126GCCGATTGGG−117 or−116CCACGTTGGG−107, however, resulted in a significant decrease of luciferase activity upon thapsigargin treatment, despite little effect on basal activity (lines 10and 11). To identify the nucleotides involved in the thapsigargin response, we assessed the effects of 14 nucleotide mutations crossing these two regions on luciferase activity (Fig.4 B). Mutations at −122, −121, −120, −119, −118, −116, −115, −114, −113, and −112 demonstrated inhibitory effects on the thapsigargin-induced response of luciferase activity (lines 3–13 except line 8). These results indicate that the 11-bp stretch, −122ATTGGgCCACG−112, in the Herp promoter region is responsible for the transcriptional response to thapsigargin. This 11-bp sequence contains two motifs forming the ERSE consensus, CCAAT (complementary to ATTGG) and CCACG, although the orientation of the first sequence is inverted. We termed thiscis-element, ERSE-II.
      Figure thumbnail gr4
      Figure 4Identification of ERSE-II. A, effect of 10-bp mutations on Herp promoter activity. Each of 10-bp units upstream of the ERSE was disrupted; the effect on luciferase activities was analyzed. Luciferase induction by thapsigargin treatment was reduced when nucleotides −126/−117 (line 10) and −116/−107 (line 11) were mutated. B, effect of 1-bp mutations on Herp promoter activity. Nucleotides from −123 to −110, upstream of the disrupted ERSE, were separately mutated; the effect on luciferase activity was then analyzed. Luciferase induction by thapsigargin treatment was reduced when nucleotides −122 (line 3), −121 (line 4), −120 (line 5), −119 (line 6), −118 (line 7), −116 (line 9), −115 (line 10), −114 (line 11), −113 (line 12), and −112 (line 13) were mutated. Tg, thapsigargin.

       Functional Contribution of ERSE and ERSE-II to the UPR

      To compare the activity of ERSE and ERSE-II, we measured the luciferase activity of plasmids containing combination of mutations in these twocis-elements. We utilized not only thapsigargin but also tunicamycin (N-glycosylation inhibitor) and mercaptoethanol (reducing agent) as ER-stress inducers to see specific induction by the UPR. Plasmid DNA containing the 5′-flanking region (−200/+98) of the Herp gene demonstrated enhanced activity in the presence of all the reagents used (Fig. 5, line 1), in contrast to the control plasmid containing the SV40 promoter (line 5). Disruption of the original ERSE resulted in decrease of the response to the ER-stress inducers but not in a complete loss (line 2). In a similar way, disruption of the novel ERSE-II also exhibited a weakened response (line 3). When both elements were disrupted, the transcriptional induction by ER stress was abrogated (line 4). These results suggest that ERSE and ERSE-II would function independently as cis-acting elements, contributing equally to the UPR-dependent induction of Herp mRNA.
      Figure thumbnail gr5
      Figure 5Comparative analysis of ERSE and ERSE-II ascis-elements for the UPR. The luciferase activity of plasmids under the control of the Herp −200/+98 region containing either the disrupted ERSE (line 2), the disrupted ERSE-II (line 3), or both (line 4) were compared with the parent plasmid (line 1). Activity was measured in the absence (open bars) or presence of either 1 μmthapsigargin (closed bars), 10 μg/ml tunicamycin (light gray bars), or 10 mm mercaptoethanol (dark gray bars). Tg, thapsigargin;Tm, tunicamycin; Me, mercaptoethanol.

       Effect of ATF6 Overexpression on the Herp Promoter Activity

      The general transcription factor, NF-Y, constitutively binds the CCAAT motif of ERSE (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ,
      • Roy B.
      • Lee A.S.
      ). The transcription factor, ATF6 (p90ATF6), on the ER membrane is activated by proteolysis in response to ER stress; the resultant N-terminal soluble form (p50ATF6) moves into nuclei to bind directly to the CCACG motif (
      • Haze K.
      • Yoshida H.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). We, therefore, examined the effect of p50ATF6 overexpression on the induction of Herp expression. As the cleavage site involved in conversion from p90ATF6 to p50ATF6 is unknown, we utilized ATF6(366), an N-terminal soluble fragment containing the entire basic region and majority of the leucine zipper region of ATF6. ATF6(366) translocates to the nucleus to enhance the levels of GRP78 mRNA (
      • Haze K.
      • Yoshida H.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). Upon transfection of the expression plasmid encoding FLAG-tagged ATF6(366) into HUVECs, a fraction of transfected cells possessed nuclei recognized by an anti-FLAG-tag antibody, indicating that the expressed ATF6(366) was present in nuclei (Fig.6 A, red signal). Cells with immunonegative nuclei were also observed, likely due to a failure of transfection. Following staining of cells with an anti-Herp antibody, immunopositive signals of the ER in ATF6(366)-expressing cells were stronger than those in cells without ATF6(366) (Fig.6 A, green signal). This suggests that overexpressed ATF6(366) functions in vivo to induce the expression of Herp in the ER.
      Figure thumbnail gr6
      Figure 6Effect of ATF6 overexpression on Herp promoter activity. A, fluorescent immunocytochemical analysis. FLAG-tagged ATF6(366) was transiently expressed in HUVECs and double-stained with both anti-FLAG-tag and anti-Herp antibodies. Bar, 20 μm. B, reporter gene assays. Plasmid DNA encoding ATF6(366) was cotransfected into HUVECs together with plasmids containing luciferase under the control of the Herp −200/+98 region with various mutations. The activity of each was measured (closed bars). The mock vector, pcDNA3.1(+), was used as a control (open bars).
      To demonstrate that p50ATF6 induces the transcriptional activity of the Herp promoter, the plasmid containing the −200/+98 region of Herp was cotransfected into HUVECs in conjunction with the ATF6(366)-expression plasmid. As expected, coexpression of ATF6(366) resulted in an enhancement of luciferase activity (Fig. 6 B, line 1). The induction was partially reduced when the two motifs, CCAAT and CCACA, of ERSE were disrupted (line 2), indicating both that the effect of ATF6(366) is dependent on the cis-element and that other elements are involved in this induction. Disruption of both motifs, ATTGG (complementary to CCAAT) and CCACA, of ERSE-II also demonstrated a partial reduction in induction (line 3). Disruption of both elements, ERSE and ERSE-II, resulted in a complete loss of the ATF6 effect (line 4). These data suggest that both ERSE and ERSE-II are involved in the ATF6-dependent UPR.
      p50ATF6 binds directly to the CCACG portion of ERSE to exert its ability as a trans-factor (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). Mutation of CCACA/G motifs of both ERSE and ERSE-II in the Herp promoter abrogated the inducible effect of ATF6(366) (Fig. 6 B, line 5), suggesting that the enhancer activity of p50ATF6 requires the CCACG sequences of both ERSE-II and ERSE. p50ATF6 binds to CCACG only when CCAAT is bound by NF-Y, exactly 9 bp upstream of CCACG (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). Mutation of the CCAAT motifs of both ERSE and ERSE-II abrogated the ATF6 effect as well (line 6). The indispensability of NF-Y binding is also applicable to ERSE-II as well as ERSE, despite the differences in both the direction and interval of CCAAT and CCACG in ERSE-II from those in ERSE.

      DISCUSSION

      We identified two cis-acting elements responsible for the UPR-dependent transcriptional induction in the proximal promoter region of the Herp gene. CCAATgggcggcagCCACA is almost identical to the 19-nucleotide consensus sequence of ERSE, CCAAT-N9-CCACG (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ). The other, ATTGG-N-CCACG, is a new element, termed ERSE-II. ERSE-II also contains two motifs, CCAAT (complementary to ATTGG) and CCACG, although the orientation and the interval between them are different from ERSE. Moreover, ERSE-II functions as an ER stress response element in an ATF6-dependent fashion, in the same manner as the original ERSE.
      The A nucleotide at position 19 in ERSE of the Herp promoter differs from a G of the ERSE consensus. Our data, however, could not demonstrate a significant functional difference in response to thapsigargin treatment between A and G at this position (Fig.3 A). Yoshida et al. (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ) demonstrated that substitution of the nucleotide G to T was a crucial mutation, impairing the UPR; they did not, however, examine the effect of substitution to A. Furthermore, ERSE-like sequences also appear in the human ER stress-responsive genes, GRP58 (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ) and SERCA2(
      • Caspersen C.
      • Pedersen P.S.
      • Treiman M.
      ), with a sequence of CCAAT-N9-CCACA. Therefore, the ERSE consensus should be described as containing the sequence: CCAAT-N9-CCACG/A.
      It has been reported that the transcription factors, NF-Y and ATF6, simultaneously bind to the CCAAT and CCACG portions of ERSE, respectively (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). The former is considered to bind in a constitutive manner, independent of the UPR. The latter binds only when it is converted from the ER membrane-embedded p90ATF6 to the soluble p50ATF6 by processing induced by ER stress (
      • Haze K.
      • Yoshida H.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). Binding of ATF6 to CCACG requires the binding of NF-Y to the CCAAT sequence at a position exactly 9 bp upstream of CCACG (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). This 9-bp distance is critical; neither 8 nor 10 bp is acceptable (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). The unidirectional necessity, however, of CCAAT and CCACG was not investigated. Our data indicate that the role of ERSE-II as a cis-acting element was exerted by ATF6 as was the case with that of ERSE (Fig. 6). Both the CCACG and the ATTGG (complementary to CCAAT) sequences of ERSE-II were critical for ATF6-mediated transcription. Although direct evidence is not available, it is likely that both NF-Y and p50ATF6 bind to ERSE-II to enhance transcription (Fig. 7). We observed specific binding of NF-Y to the CCACG sequence of ERSE-IIin vitro (data not shown). If our model is correct, the inverse direction of two motifs may be necessary when the distance between them is 1 bp, not 9. A study of the steric structure of protein-DNA interaction will help determine the validity of this argument. By analogy to ERSE, other transcription factors, such as CREB-RP (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ) and XBP-1 (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ), may bind to the CCACG sequence of ERSE-II.
      Figure thumbnail gr7
      Figure 7Model for UPR-dependent transcriptional induction of Herp. The up-regulation of Herp mRNA expression resulting from ER stress is regulated by twocis-acting elements, ERSE and ERSE-II. NF-Y binds constantly to the CCAAT sequence in these elements. Upon ER stress, p90ATF6 in the ER membrane is activated to form p50ATF6, a soluble molecule that translocates to the nucleus and binds to the CCACG/A sequence of ERSE and ERSE-II, activating transcription. TBP, TATA-binding protein.
      Most UPR-target genes, such as GRP78, GRP94, and the gene for calreticulin, are fully activated by multiple copies of ERSE (
      • Yoshida H.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Mori K.
      ,
      • Roy B.
      • Lee A.S.
      ). Despite a single ERSE, however, Herp mRNA induction by the UPR is very strong as compared with other ER chaperones (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ). ERSE-II may cooperate with ERSE to facilitate the strong induction of Herp in response to ER stress. We searched for ERSE-II in other ER stress-responsive genes to demonstrate the function of this sequence in the response to cellular stress. ORP150 is an ER-resident protein whose expression is induced by hypoxia; three distinct mRNA species are produced by alternative promoters (
      • Kaneda S.
      • Yura T.
      • Yanagi H.
      ). One of them was preferentially induced by hypoxia and ER stress, in a manner dependent on a single ERSE-like sequence (−93CCAATgagcgcccgCCgCG−75) in the promoter region (
      • Kaneda S.
      • Yura T.
      • Yanagi H.
      ). We found two ERSE-II-like sequences,−267ATTGGaCCACG−277 and−160ATTGGaCCACG−170, upstream of the ERSE. They also might be involved in the UPR.
      Recently, van Laar et al. (
      • van Laar T.
      • Schouten T.
      • Hoogervorst E.
      • van Eck M.
      • van der Eb A.J.
      • Terleth C.
      ) identified a human methyl methanesulfonate (MMS)-inducible gene, Mif1, identical to Herp. The mRNA is also induced by tunicamycin, osmotic shock, and UV irradiation. Although they demonstrated that onecis-element, ERSE, was involved in the response to tunicamycin, ERSE-II was not mentioned. The induction ofMif1 by MMS was mediated by neither ERSE nor ERSE-II but by a 122-bp fragment (−257 to −136). As MMS also induces the mRNA expression of GRP78 (
      • van Laar T.
      • Schouten T.
      • Hoogervorst E.
      • van Eck M.
      • van der Eb A.J.
      • Terleth C.
      ,
      • Wang X.Z.
      • Lawson B.
      • Brewer J.W.
      • Zinszner H.
      • Sanjay A.
      • Mi L.J.
      • Boorstein R.
      • Kreibich G.
      • Hendershot L.M.
      • Ron D.
      ), GRP94 (
      • van Laar T.
      • Schouten T.
      • Hoogervorst E.
      • van Eck M.
      • van der Eb A.J.
      • Terleth C.
      ), andCHOP (
      • Wang X.Z.
      • Lawson B.
      • Brewer J.W.
      • Zinszner H.
      • Sanjay A.
      • Mi L.J.
      • Boorstein R.
      • Kreibich G.
      • Hendershot L.M.
      • Ron D.
      ), known UPR-target genes, these genes and Herp may share an additional cis-acting MMS response element.
      The function of Herp is still unknown. It was believed that all proteins encoded by UPR-target genes functioned as molecular chaperones and folding enzymes to relieve the disturbance of the ER. As the majority of the molecule is exposed to the cytoplasm, Herp may play a role independent of molecular chaperones (
      • Kokame K.
      • Agarwala K.L.
      • Kato H.
      • Miyata T.
      ). ER stress also up-regulates the mRNA expression of the transcription factors, CHOP (
      • Wang X.Z.
      • Lawson B.
      • Brewer J.W.
      • Zinszner H.
      • Sanjay A.
      • Mi L.J.
      • Boorstein R.
      • Kreibich G.
      • Hendershot L.M.
      • Ron D.
      ), ATF3 (
      • Cai Y.
      • Zhang C.
      • Nawa T.
      • Aso T.
      • Tanaka M.
      • Oshiro S.
      • Ichijo H.
      • Kitajima S.
      ), and XBP-1 (
      • Yoshida H.
      • Okada T.
      • Haze K.
      • Yanagi H.
      • Yura T.
      • Negishi M.
      • Mori K.
      ). In addition, a large number of UPR-target genes have been identified in yeast using microarray techniques that are not limited to proteins involved in ER folding (
      • Travers K.J.
      • Patil C.K.
      • Wodicka L.
      • Lockhart D.J.
      • Weissman J.S.
      • Walter P.
      ). Further research from a wide viewpoint will be required to determine the physiological function of Herp.
      To facilitate our study, we modified the plasmid DNA for reporter gene assays. We reduced basal luciferase activity in cells by preventing the accumulation of superfluous mRNA and enzyme prior to stimulation by destabilizing the luciferase mRNA. This technique allowed us to identify a new cis-element responsible for stimulation of gene expression. Although we used this technique to detect response to ER stress, the destabilization of reporter gene plasmids will be widely applicable to the search for cis-elements responsible for other conditions.

      Acknowledgments

      We thank Akemi Fukumoto and Chikako Yasuda for technical assistance, and Dr. Hiderou Yoshida and Dr. Kazutoshi Mori for their kind donation of the ATF6 cDNA.

      REFERENCES

        • Kokame K.
        • Agarwala K.L.
        • Kato H.
        • Miyata T.
        J. Biol. Chem. 2000; 275: 32846-32853
        • Kaufman R.J.
        Genes Dev. 1999; 13: 1211-1233
        • Mori K.
        Cell. 2000; 101: 451-454
        • Urano F.
        • Wang X.
        • Bertolotti A.
        • Zhang Y.
        • Chung P.
        • Harding H.P.
        • Ron D.
        Science. 2000; 287: 664-666
        • Nakagawa T.
        • Zhu H.
        • Morishima N.
        • Li E.
        • Xu J.
        • Yankner B.A.
        • Yuan J.
        Nature. 2000; 403: 98-103
        • Cox J.S.
        • Shamu C.E.
        • Walter P.
        Cell. 1993; 73: 1197-1206
        • Mori K.
        • Ma W.
        • Gething M.J.
        • Sambrook J.
        Cell. 1993; 74: 743-756
        • Cox J.S.
        • Walter P.
        Cell. 1996; 87: 391-404
        • Kawahara T.
        • Yanagi H.
        • Yura T.
        • Mori K.
        Mol. Biol. Cell. 1997; 8: 1845-1862
        • Mori K.
        • Ogawa N.
        • Kawahara T.
        • Yanagi H.
        • Yura T.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 4660-4665
        • Mori K.
        • Kawahara T.
        • Yoshida H.
        • Yanagi H.
        • Yura T.
        Genes Cells. 1996; 1: 803-817
        • Mori K.
        • Ogawa N.
        • Kawahara T.
        • Yanagi H.
        • Yura T.
        J. Biol. Chem. 1998; 273: 9912-9920
        • Tirasophon W.
        • Welihinda A.A.
        • Kaufman R.J.
        Genes Dev. 1998; 12: 1812-1824
        • Wang X.Z.
        • Harding H.P.
        • Zhang Y.
        • Jolicoeur E.M.
        • Kuroda M.
        • Ron D.
        EMBO J. 1998; 17: 5708-5717
        • Yoshida H.
        • Haze K.
        • Yanagi H.
        • Yura T.
        • Mori K.
        J. Biol. Chem. 1998; 273: 33741-33749
        • Haze K.
        • Yoshida H.
        • Yanagi H.
        • Yura T.
        • Mori K.
        Mol. Biol. Cell. 1999; 10: 3787-3799
        • Yoshida H.
        • Okada T.
        • Haze K.
        • Yanagi H.
        • Yura T.
        • Negishi M.
        • Mori K.
        Mol. Cell. Biol. 2000; 20: 6755-6767
        • Roy B.
        • Lee A.S.
        Nucleic Acids Res. 1999; 27: 1437-1443
        • Roy B.
        • Lee A.S.
        Mol. Cell. Biol. 1995; 15: 2263-2274
        • Roy B.
        • Li W.W.
        • Lee A.S.
        J. Biol. Chem. 1996; 271: 28995-29002
        • Barbosa-Tessmann I.P.
        • Chen C.
        • Zhong C.
        • Siu F.
        • Schuster S.M.
        • Nick H.S.
        • Kilberg M.S.
        J. Biol. Chem. 2000; 275: 26976-26985
        • Cai Y.
        • Zhang C.
        • Nawa T.
        • Aso T.
        • Tanaka M.
        • Oshiro S.
        • Ichijo H.
        • Kitajima S.
        Blood. 2000; 96: 2140-2148
        • Kokame K.
        • Kato H.
        • Miyata T.
        J. Biol. Chem. 1996; 271: 29659-29665
        • Heinemeyer T.
        • Wingender E.
        • Reuter I.
        • Hermjakob H.
        • Kel A.E.
        • Kel O.V.
        • Ignatieva E.V.
        • Ananko E.A.
        • Podkolodnaya O.A.
        • Kolpakov F.A.
        • Podkolodny N.L.
        • Kolchanov N.A.
        Nucleic Acids Res. 1998; 26: 362-367
        • Shaw G.
        • Kamen R.
        Cell. 1986; 46: 659-667
        • Caspersen C.
        • Pedersen P.S.
        • Treiman M.
        J. Biol. Chem. 2000; 275: 22363-22372
        • Kaneda S.
        • Yura T.
        • Yanagi H.
        J. Biochem. ( Tokyo ). 2000; 128: 529-538
        • van Laar T.
        • Schouten T.
        • Hoogervorst E.
        • van Eck M.
        • van der Eb A.J.
        • Terleth C.
        FEBS Lett. 2000; 469: 123-131
        • Wang X.Z.
        • Lawson B.
        • Brewer J.W.
        • Zinszner H.
        • Sanjay A.
        • Mi L.J.
        • Boorstein R.
        • Kreibich G.
        • Hendershot L.M.
        • Ron D.
        Mol. Cell. Biol. 1996; 16: 4273-4280
        • Travers K.J.
        • Patil C.K.
        • Wodicka L.
        • Lockhart D.J.
        • Weissman J.S.
        • Walter P.
        Cell. 2000; 101: 249-258