Herp, a New Ubiquitin-like Membrane Protein Induced by Endoplasmic Reticulum Stress*

Hyperhomocysteinemia, a risk factor for vascular disease, injures endothelial cells through undefined mechanisms. We previously identified several homocysteine-responsive genes in cultured human vascular endothelial cells, including the endoplasmic reticulum (ER)-resident molecular chaperone GRP78/BiP. Here, we demonstrate that homocysteine induces the ER stress response and leads to the expression of a novel protein, Herp, containing a ubiquitin-like domain at the N terminus. mRNA expression of Herp was strongly up-regulated by inducers of ER stress, including mercaptoethanol, tunicamycin, A23187, and thapsigargin. The ER stress-dependent induction of Herp was also observed at the protein level. Immunochemical analyses using Herp-specific antibodies indicated that Herp is a 54-kDa, membrane-associated ER protein. Herp is the first integral membrane protein regulated by the ER stress response pathway. Both the N and C termini 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, in contrast to GRP78 and other ER stress-responsive proteins. Herp may, therefore, play an unknown role in the cellular survival response to stress.

thrombosis in patients with hyperhomocysteinemia is still poorly understood (4).
To investigate the gene expression changes induced by homocysteine, we performed a differential display analysis that identified six up-regulated and one down-regulated genes in human umbilical vein endothelial cells (HUVECs) 1 (5). GRP78/ BiP, a gene up-regulated by homocysteine, is an endoplasmic reticulum (ER)-resident molecular chaperone. The induction of GRP78 is a consequence of the cellular response to perturbations of the ER, suggesting that homocysteine may serve to induce the ER stress response. Another group also showed that homocysteine caused ER stress and growth arrest in HUVECs (6). GRP78 mRNA levels in the livers of hyperhomocysteinemic mice lacking the cystathionine ␤-synthase gene were increased over wild-type mice (7).
The ER in eukaryotic cells is optimized for synthesizing, folding, and assembling membrane proteins and soluble proteins destined for secretion or trafficking to lysosomes. Environmental changes leading to the accumulation of unfolded proteins in the ER trigger a stress response, referred to as the ER stress response. As one aspect of the cellular response, the unfolded protein response (UPR) was characterized in terms of the transcriptional induction of a set of mRNAs encoding ERresident molecular chaperones and folding enzymes including GRP78, GRP94, PDI, ERp72, and calreticulin. The signaling pathway from the ER to the nucleus has been extensively studied in the yeast, Saccharomyces cerevisiae. Perturbation of the ER is sensed by an ER-resident transmembrane kinase/ endoribonuclease, Ire1p (8,9). Activated Ire1p produces transcription factor, Hac1p, through unusual mRNA splicing (10 -13). Hac1p binds to the promoter cis-acting element, UPRE, activating the transcription of ER-resident chaperones (10,14,15).
Accumulating evidence supports a similar mechanism for the ER stress response in mammalian cells. A cis-element, designated ERSE, is present in the promoter regions of mammalian ER-resident proteins (16,17). ATF6 has been identified as a trans-acting ERSE-binding protein (16,18). Mammalian orthologues for the yeast UPR sensor Ire1p, IRE1␣ (19) and IRE1␤ (20), have been cloned and characterized, although their substrates are not known. As recently reviewed (21,22), however, the mammalian UPR is considerably more diverse than the yeast UPR. The mammalian ER stress response seems to coordinate both transcriptional and translational controls involved in the cellular survival, inflammation, immune response, and apoptosis.
Here, we report the characteristics of a novel protein induced by the ER stress. In contrast to the previous features, this new UPR target protein, with an N-terminal ubiquitin-like domain, is present on the cytoplasmic face of the ER membrane.
cDNA Cloning of Herp-A cDNA library was constructed from homocysteine-treated HUVECs (5). Positive phages were cloned by PCRbased screening as described previously (5). Approximately 1.5 ϫ 10 4 phages were plated at a density of 300 -400 plaque-forming units/dish with Escherichia coli XL1-Blue MRFЈ strain as the host. PCR was performed using 5Ј-CTGGGAAGCTGTTGTTGG-3Ј and 5Ј-CATGTAG-TACTGTCGTGC-3Ј as primers with the plate lysates as templates. After several sequential dilutions, 10 positive phages were cloned, and their plasmid DNAs were prepared. Insert DNAs were sequenced in both directions using a BigDye Terminator Cycle Sequencing FS Ready Reaction Kit (PerkinElmer Life Sciences).
Cap Site Hunting of Human Herp-To determine the transcriptional start site of human Herp, we performed cap site hunting using human heart and brain Cap Site cDNA dT kits (Nippon Gene) according to the manufacturer's instructions. We used 5Ј-ACGCGGACGCTCGGGGTA-GACG-3Ј and 5Ј-TGACGGGCTCGGGTTCGGTCTC-3Ј as primers for the initial PCR and the nested PCR, respectively. After inserting the PCR products into the plasmid vector, pCR II (Invitrogen), 10 clones derived from each of the heart and brain cDNA libraries were sequenced.
Preparation of Antiserum Specific to Herp-Two oligonucleotides, 5Ј-CGGGATCCATGGAGTCCGAGACCGAACC-3Ј and 5Ј-CCGCTCGA-GTCAGTTTGCGATGGCTGGG-3Ј, were used to amplify the entire open reading frame of Herp. The PCR product digested with BamHI and XhoI was ligated into the corresponding site of pGEX4T-3 (Amersham Pharmacia Biotech), a Schistosoma japonicum glutathione Stransferase fusion expression vector. Glutathione S-transferase-Herp fusion protein was purified according to the manufacturer's instructions. Recombinant Herp was purified following digestion of the fusion protein with thrombin. We raised antiserum against recombinant Herp in rabbits. Fusion protein (1 mg) was emulsified in adjuvant TiterMax (CytRx) and injected intradermally. Every 2 weeks, rabbits were boosted with 0.5 mg of fusion protein in the same adjuvant. The blood was taken 2 weeks after the third injection. Antibodies specific for Herp (anti-Herp) were purified by recombinant Herp-immobilized affinity chromatography.
Western Blot Analysis-HUVECs were incubated with a final concentration of 10 mM 2-mercaptoethanol, 10 g/ml tunicamycin, or 1 M thapsigargin for 6 h. After washing, cells were lysed with SDS sample buffer (10 mM Tris-HCI, 2% SDS, 50 mM dithiothreitol, 2 mM EDTA, 0.02% bromphenol blue, 6% glycerol, pH 6.8) and boiled for 7 min. After SDS-polyacrylamide gel electrophoresis, proteins in the gel were transferred to an Immun-Blot polyvinylidene difluoride membrane (Bio-Rad). Remaining binding sites on the membrane were blocked with 3% skim milk in T-PBS (10 mM sodium phosphate, 150 mM NaCl, 0.05% Tween 20, pH 7.4) for 1 h. The membrane was then incubated with 1 g/ml anti-Herp or anti-KDEL (StressGen) in 3% skim milk for 1 h. After washing with 3% skim milk, the membrane was incubated with 0.1 g/ml peroxidase-labeled goat anti-rabbit IgG or anti-mouse IgG (Kirkegaard and Perry Laboratories) in 3% skim milk for 1 h. The membrane was thoroughly washed with T-PBS and with PBS (10 mM sodium phosphate, 150 mM NaCl, pH 7.4), followed by chemiluminescent detection using the Renaissance Western blot Chemiluminescence Reagent Plus (PerkinElmer Life Sciences). Chemiluminescence was detected by an image analyzer, LAS-1000plus.
In Vitro Synthesis of Herp-Two oligonucleotides, 5Ј-GGGGTACCA-TGGAGTCCGAGACCGAAC-3Ј and 5Ј-CGGAATTCTCAGTTTGCGAT-GGCTGGG-3Ј, were used to PCR amplify the entire open reading frame of Herp. The product, digested with KpnI and EcoRI, was ligated into the corresponding site of a plasmid vector, pZeoSV2(ϩ) (Invitrogen), with a T7 priming site. Using the resultant plasmid DNA, pZeoSV2Herp, as a template, Herp was synthesized in vitro by the TNT T7 Quick Coupled Transcription/Translation Systems (Promega).
Subcellular Fractionation-HUVECs were treated with 1 M thapsigargin for 6 h and removed from culture dishes by scraping. Cells were sonicated in Dulbecco's PBS supplemented with a protease inhibitors mixture tablet (Roche Molecular Biochemicals). In the absence or presence of either 1 M NaCl, 0.1 M Na 2 CO 3 (pH 11) or 1% Triton X-100, lysates were centrifuged at 100,000 ϫ g for 1 h at 4°C and separated into supernatant and pellet fractions. The fractions were subjected to Western blot analysis using an anti-Herp antibody.
Fluorescent Immunocytochemistry-HUVECs were incubated with or without 1 M thapsigargin for 6 h. 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, followed by an incubation in 5% normal goat serum and 5% fish gelatin for 30 min. To detect Herp and GRP78/ GRP94 simultaneously, 50 g/ml anti-Herp rabbit antibody and 5 g/ml anti-KDEL mouse monoclonal antibody were added to cells for 1 h. Then, cells were incubated for 1 h with both Rhodamine Red-X-conjugated goat anti-rabbit IgG (Molecular Probes) and Oregon Green 514conjugated goat anti-mouse IgG (Molecular Probes). After washing with PBS, fluorescence was visualized by a confocal laser scanning microscope, FLUOVIEW (Olympus).
Protease Protection Assay-The open reading frame of Herp tagged with the N-terminal Myc (EQKLISEEDL) and C-terminal FLAG (DYKDDDDK) peptides was PCR-amplified with the two oligonucleotides 5Ј-CGGGATCCACCATGGAGCAGAAGCTGATCTCCGAG-GAGGACCTGATGGAGTCCGAGACCGAAC-3Ј and 5Ј-GGAATTCTC-ACTTATCGTCATCGTCCTTGTAGTCGTTTGCGATGGCTGGGGGGC-3Ј. The product was digested with both BamHI and EcoRI and ligated into the vector, pcDNA3.1(ϩ) (Invitrogen). HUVECs were transfected in a 90-mm dish with 25 g of the Myc-Herp-FLAG expression plasmid, pcDNA3mHerpf, or the mock vector pcDNA3.1(ϩ) mixed with 50 l of FuGENE6 (Roche Molecular Biochemicals). After a 24-h incubation, cells were removed from culture dishes, placed in Dulbecco's PBS, and centrifuged at 1,000 ϫ g for 5 min at 4°C. The cell pellet from each dish was resuspended in 0.4 ml of buffer (10 mM Hepes-KOH, 250 mM sucrose, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, pH 7.4), passed through a 26-gauge needle 30 times, and centrifuged at 1,000 ϫ g for 5 min at 4°C. The supernatant, containing microsomal membranes, was treated with variable amounts of Proteinase K (Qiagen) for 5 min on ice. Reactions were stopped by the addition of (p-amidinophenyl)methanesulfonyl fluoride hydrochloride (Wako). After centrifugation at 100,000 ϫ g for 5 min at 4°C, the pellet was subjected to Western blot analysis. Myc and FLAG tags were detected by anti-Myc tag (Medical and Biological Laboratories) and anti-octapeptide epitope tag (Zymed Laboratories Inc.) antibodies, respectively. The N and C termini of calnexin were detected by anti-Calnexin NT and CT (St-ressGen), respectively.
Streptolysin O Treatment-For determination of the membrane topology of Herp, HUVECs expressing Myc-Herp-FLAG were treated with 0.5 units/ml streptolysin O (SLO, gifted from Dr. Murata) for 10 min on ice, followed by an incubation in buffer (20 mM Hepes-KOH, 110 mM CH 3 COOK, 5 mM CH 3 COONa, 2 mM (CH 3 COO) 2 Mg, 1 mM EGTA, 1 mM dithiothreitol, pH 7.3) containing propidium iodide at 32°C for 5 min. After fixation with 2% paraformaldehyde, a portion of the total cells were permeabilized with 0.05% Triton X-100. Cells were then stained with 10 g/ml of anti-Myc tag, anti-FLAG M2 (Eastman Kodak), or anti-KDEL as primary antibodies and Oregon Green 514conjugated goat anti-mouse IgG as secondary antibodies. Fluorescence was observed by a confocal laser-scanning microscope.

RESULTS
Previously, we identified several transcribed fragments of novel genes that were responsive to a 4-h treatment of HUVEC with homocysteine (5). The CA13 fragment demonstrated the strongest induction following treatment (data not shown). The sequence, containing a typical polyadenylation signal near the 3Ј terminus, is considered to be a 3Ј region of cDNA. We termed the encoded protein Herp.
Induction of the mRNA Expression by Homocysteine-Northern blot analysis revealed a faint 2.2-kilobase mRNA in untreated HUVECs. mRNA levels increased dramatically following an incubation with homocysteine (Fig. 1). The mRNA induction peaked after 4 h of treatment increasing in intensity approximately 50-fold over levels seen in untreated cells. Previous work demonstrated the dependence of the induction of GRP78, an ER-resident molecular chaperone, on homocysteine (5). Enhanced expression of GRP78 is a consequence of the cellular responses to ER stress. To examine the relationship of homocysteine to ER stress induction, we analyzed the expression of various stress proteins after homocysteine treatment. Homocysteine treatment of HUVECs enhanced the mRNA abundance of multiple molecular chaperones or folding enzymes in the ER: GRP78, GRP94, PDI, ERp72, and calnexin ( Fig. 1). mRNA induction was not observed for HSP70 and HSP90, molecular chaperones in the cytoplasm/nucleus, or HSP60, in mitochondria. These data suggest that homocysteine specifically induces the cellular ER stress response. Notably, mRNA induction of Herp is faster and of a greater magnitude than other ER-resident stress proteins examined.
To examine the regulation of Herp expression by the ER stress response, we investigated the effects on Herp mRNA levels of additional agents known to induce the ER stress response (Fig. 2). Herp mRNA increased in response to treatment with 2-mercaptoethanol (reducing agent), tunicamycin (N-glycosylation inhibitor), A23187 (calcium ionophore), and thapsigargin (ER-resident Ca 2ϩ -ATPase inhibitor). Thus, Herp expression is regulated by the ER stress response referred to as the UPR.
Nucleotide Sequence of Herp cDNA-We isolated 10 independent clones with full-length cDNAs containing the CA13 fragment sequence from a homocysteine-treated HUVEC cDNA library. All clones contained an identical open reading frame encoding the Herp protein. We performed cap site hunting to determine the 5Ј-terminal transcription start site of the full-length cDNA. All 20 clones sequenced contained the identical starting sequence of 5Ј-AGAGACG, which we concluded to be the 5Ј terminus of the full-length Herp cDNA.
The complete nucleotide sequence of human Herp cDNA (Fig. 3) has been submitted to the GenBank TM /EMBL/DDBJ data bases under accession number AB034989. The 1,176-base pair open reading frame, beginning at an ATG start codon and ending at a TGA stop codon, produces a protein sequence of 391 amino acid residues (designated as Herp). The presumed initiating ATG, assigned to the first methionine codon, perfectly matches the consensus sequence for initiation of translation in vertebrates, GCCGCC(A/G)CCATGG (24). Homology searches of the GenBank TM /EMBL/DDBJ data bases revealed that human Herp cDNA is 99.9% identical to KIAA0025 (GenBank TM accession number D14695), a randomly sampled human cDNA clone (25), and 95.5% identical to a stretch of 0041AS (Gen-Bank TM accession number M29512), a pig hepatic cDNA clone (26) as Mif1 (GenBank TM accession number NM014685), a gene induced by methyl methanesulfonate. They also showed that Mif1 expression was induced by tunicamycin treatment, osmotic shock, and UV light irradiation (27).
Amino Acid Sequence of Herp-Many homologous clones were found when the human Herp nucleotide sequence was searched against the mouse expressed sequence tag data base. We have submitted the accurate sequence to the GenBank TM / EMBL/DDBJ data bases under the accession number AB034991. The deduced amino acid sequences of human and mouse Herp share 88.7% identity (aligned in Fig. 4A). Human Herp, with a calculated molecular mass of 43,719 Da, is predicted to be a membrane protein with one transmembrane domain (Ser 285 -Trp 307 ; Fig. 4A, underline) by the SOSUI, the secondary structure prediction system (28). A PATTERN search of the PROSITE data base (29) did not discover a motif match that might divulge the functional properties of Herp; a PROFILE search, however, demonstrated a significant match to ubiquitin (Fig. 4B). Ubiquitin is a highly conserved, 76amino acid protein with 100% identity between the human and the mouse. The stretch of amino acids between Val 14 and Val 85 of the human and mouse Herps share 32% and 33% identity with the Val 5 -Val 70 stretch of ubiquitin, respectively.
Constitutive Expression of Herp in Human Organs-To examine the distribution of Herp in human organs, we performed a Northern blot analysis using a Herp-specific probe. Human Herp was ubiquitously expressed in all organs tested as a single species of mRNA at 2.2 kilobases (Fig. 5), consistent with the band appearing in cultured HUVECs (Fig. 1). Herp mRNA was most abundant in pancreas, suggesting a role for Herp in that organ.
Identification of Endogenous Herp-We immunized rabbits against bacterial-expressed glutathione S-transferase-Herp fusion protein to obtain Herp-specific polyclonal antibodies (anti-Herp). Western blot analysis following the in vitro transcrip-tion/translation of the Herp expression plasmid, pZeoSV2Herp, revealed a single, 54-kDa protein (Fig. 6A, lane 1); no signal was observed using the mock vector, pZeoSV2(ϩ) (data not shown). Anti-Herp recognized a single 54-kDa band by Western blot analysis of HUVEC lysates (lane 2), considered to be the endogenous Herp. The 10-kDa difference between the apparent mass of the endogenous 54-kDa protein and the calculated mass (43, 719 Da) may be due to acidic characteristics resulting from a low pI of 5.0.
To examine the effects of ER stress on the expression of the 54-kDa protein, HUVECs were incubated with 10 mM 2-mercaptoethanol, 10 g/ml tunicamycin, or 1 M thapsigargin for 6 h. When the lysates were subjected to Western blot analysis, the immunoreactive 54-kDa protein levels increased as a result of each treatment (Fig. 6A, lanes 3-5). This observation was consistent with the expression pattern of Herp mRNA (Fig. 2). These results suggest that the 54-kDa protein recognized by anti-Herp is the bona fide product of the Herp gene.
Membrane Association of Herp-To confirm the transmembrane nature of Herp, we separated HUVEC lysates into supernatant and pellet fractions by centrifugation. In isotonic buffer, Herp was observed only in the pellet (Fig. 6B, lanes 2  and 3). In the presence of 1 M NaCl, Herp was also present only in the pellet (lanes 4 and 5). This localization pattern indicated that Herp is strongly bound to membranes. A similar distribution was observed in the presence of 0.1 M Na 2 CO 3 (pH 11) (lanes 6 and 7), which permeabilizes microsome membranes to extract both peripheral and luminal proteins. In contrast, treatment with 1% Triton X-100 resulted in the efficient solubilization of Herp (lanes 8 and 9). These results suggest that Herp is an integral membrane protein.
Subcellular Localization of Herp-To elucidate the subcellular localization of Herp, we performed indirect fluorescent immunocytochemistry. We observed an intense perinuclear staining and a uniform staining of a peripheral lace-like network in thapsigargin-treated HUVEC stained with anti-Herp (Fig. 7D). In untreated cells, the fluorescence intensity signals was faint (Fig. 7A), consistent with the protein levels observed by Western blot (Fig. 6A, lanes 2 and 5). Because the perinuclear staining and peripheral lace-like network suggested Herp may associate with the ER, we performed co-localization studies with marker proteins of the ER. The anti-KDEL staining pattern of Fig. 7F, reacting to GRP78 and GRP94, was nearly identical to the staining pattern of anti-Herp (Fig. 7D). These data suggest that Herp is associated with the ER. Therefore, Herp was named the homocysteine-responsive ER-resident protein.
Membrane Topology of Herp-Knowledge of the protein orientation of Herp in the ER membrane may facilitate an understanding of its function. We investigated the membrane topology of Herp by protease protection assay. We expressed a doubly epitope-tagged Herp, Myc-Herp-FLAG, in HUVECs by transient transfection of the expression plasmid, pcDNA3mHerpf. Following homogenization and centrifugation, both anti-Myc tag and anti-FLAG tag recognized a single 61-kDa protein in the membrane fraction (Fig. 8A, lanes 1 and  4), considered to be Myc-Herp-FLAG. This band was absent from microsomes prepared with the mock vector pcDNA3.1(ϩ)-transfected cells (data not shown). As a control, we monitored endogenous calnexin, a type I transmembrane protein in the ER, by anti-Calnexin NT and CT, recognizing the luminal N terminus and the cytoplasmic C terminus of calnexin, respectively. Following digestion with levels of proteinase K that diminish the full-length calnexin band, anti-Calnexin NT (Fig. 8A, lane 9), not CT (lane 12), detected a lower band corresponding to the N-terminal luminal domain of calnexin. Therefore, polypeptides in the lumen were protected from proteinase K activities under these experimental conditions. In the same microsomes, neither the anti-Myc tag (lane 3) nor the anti-FLAG tag (lane 6) antibodies detected fragments derived from the parent 61-kDa band, suggesting that both the N and C termini are exposed to the cytoplasm.
We next performed immunocytochemistry on permeabilized HUVECs transiently transfected with the Myc-Herp-FLAG expression plasmid. SLO is a bacterial, pore-forming toxin that requires cholesterol binding in the bilayer, selectively permeabilizing the plasma membrane without affecting intracellular membranes. Thus, the immunoreactivity of a given epitope is dependent upon both the accessibility of the antibody and the orientation (e.g. cytosolic or luminal) of a membrane protein.
Without the SLO treatment, fluorescence could not be detected after staining with anti-Myc tag, anti-FLAG tag, or anti-KDEL (Fig. 8B, panels a, d, and g). Following treatment with Triton X-100, permeabilizing both the plasma membrane and intracellular membranes, anti-KDEL, recognizing the ER-luminal proteins GRP78 and GRP94, stained a lace-like pattern (panel i). Although we detected nuclear staining with propidium iodide in SLO-treated cells (panel h), anti-KDEL signals were not observed under these conditions, indicating an intact ER membrane. Under the same condition, both the anti-Myc tag (panel b) and anti-FLAG tag (panel e) antibodies stained the ER. Therefore, both the N and C termini of Herp are exposed to the cytoplasm. DISCUSSION Perturbations of the ER in mammalian cells result in the activation of a stress signaling pathway out of the ER, leading to transcriptional and translational regulation. Following the accumulation of unfolded proteins in the ER, two protein kinases, IRE1␣/␤ (19,20) and PERK (30), are activated. Active IRE1 is believed to be the most proximal factor for the transcriptional induction of molecular chaperones and folding enzymes in the ER. Activated PERK contributes to ER stressinduced translational attenuation by phosphorylating the ␣-subunit of eukaryotic initiation factor 2.
In this report, we demonstrate a new target gene for UPRinduced transcription, Herp. Herp is localized to the ER (Fig. 7) in a manner similar to other genes induced by ER stress: GRP78, GRP94, GRP170, calreticulin, FKBP13, PDI, and ERp72. In contrast to the proteins described above present in the ER lumen, Herp is an integral membrane protein (Fig. 6B).  8 and 9). These fractions were analyzed by Western blot using anti-Herp.
Although calnexin is a transmembrane protein in the ER, it demonstrates weak induction by the UPR (Fig. 1). UPR-induced ER-luminal soluble proteins have an ER-retrieval signal motif, KDEL or KDEL-like sequence, at the C terminus (31,32), absent from the Herp sequence. Herp also lacks a Cterminal K(X)KXX motif, facilitating the interaction of the protein with the coatomer (COP1) complex for the retrograde transport of type I membrane proteins back to the ER (33). Other motifs such as a diphenylalanine sequence (34,35), a ␦L sequence (36), and transmembrane domain structure (37)(38)(39) may act as the signal for ER retrieval or retention; the mechanism by which Herp is localized to the ER, however, remains unknown.
At present, all UPR-inducible proteins in the ER are molecular chaperones or folding enzymes assisting protein folding. It is doubtful that Herp conforms to this category of ER stressinduced protein judging from the membrane topology demon-strated from several experiments: 1) The protease protection assay shown (Fig. 8A). 2) Immunocytochemistry of selectively permeabilized cells shown (Fig. 8B). 3) The lack of N-glycosylation despite a potential site for N-glycosylation (Asn 141 ). In vitro synthesized or tunicamycin-induced Herp was the same size as other stimulant-induced Herp (Fig. 6A); PNGase F treatment did not alter the size of Herp on SDS-PAGE (data not shown). 4) When the activation peptide sequence of blood coagulation factor IX, known to be glycosylated (40), was added to either the N or C terminus of Herp, glycosylation did not occur (data not shown). Although the tagged protein may have inserted in the membrane incorrectly, these data strongly suggest that both the N and C termini of Herp face the cytoplasm. The SOSUI system secondary structure prediction (28) suggests that Herp is a membrane protein with one transmembrane domain (Ser 285 -Trp 307 ). If this is correct, the topology of Herp is depicted like model 1 in Fig. 9. A hydropathy profile based on the algorithm of Kyte and Doolittle (41), however, predicts an additional hydrophobic stretch at the extreme C terminus (Fig. 9). If this region is embedded in the lipid bilayer, Herp may assume a multi-pass topology with both termini exposed to the cytoplasm. The Arg 311 -Ser 365 loop would face the ER lumen (Fig. 9, model 2), with the majority of the molecule exposed to the cytoplasm. Thus, Herp may play a role separate from the molecular chaperones for proteins in the ER. Herp may interact with other ER-resident chaperones to assist their function in the UPR.
Herp has an unusual N-terminal domain similar to ubiquitin (Fig. 4B). Ubiquitin is a small, highly conserved protein present in all eukaryotic cells. The modification of cellular proteins with ubiquitin targets them for degradation by a large, multisubunit protease, the 26 S proteasome. Many proteins with a structural similarity to ubiquitin present in cells, Ubls (ubiquitin-like proteins), are divided into two subclasses: small, type-1 Ubls, such as SUMO-1 and NEDD8, that are ligated to target proteins in a similar manner to ubiquitin, and type-2 Ubls containing ubiquitin-like structures within a variety of large proteins having distinct functions, such as Elongin B, Rad23, and Parkin. Although ubiquitin and type-1 Ubls are central players in post-translational protein modification, the significance of type-2 Ubls remains obscure (42,43). The lack of a diglycine motif at the C-terminal end of the ubiquitin domain suggests that Herp is a type-2 Ubl. The ubiquitin-like domain of Elongin B serves a chaperone-like function, facilitating the assembly and enhancing the stability of the Elongin A/B/C complex (44,45). Rad23 interacts with the 26 S proteasome through an N-terminal ubiquitin-like domain (46,47). ERresident molecular chaperones and folding enzymes participate in ER-associated degradation in a manner dependent on the proteasome (48). From the viewpoint of quality control (49), it is interesting to consider that UPR-inducible Herp could possibly interact with the proteasome. The identification of molecules interacting with the Herp ubiquitin-like domain may promote a functional understanding of Herp.
The proximal promoter region of UPR target genes such as GRP78, GRP94, and calreticulin contains an ER stress response cis-element, ERSE (16,17). van Laar et al. (27) recently demonstrated a functional ERSE in the 5Ј-flanking region of Mif1/KIAA0025, a gene identical to Herp. To obtain the high levels of Herp induced by the UPR, an additional cis-element may be involved in Herp mRNA up-regulation. In addition to transcriptional regulation, cells under ER stress immediately control the cellular, translational capacity, protecting against further accumulation of unfolded proteins in the ER. This translational inhibition occurs mainly through phosphorylation of the ␣-subunit of eukaryotic initiation factor 2, catalyzed by PERK (30). It remains unknown how UPR-inducible proteins such as Herp and GRP78 can be efficiently translated during transcriptional repression. GRP78 mRNA can be translated in poliovirus-infected cells when general, cap-dependent translation of host cell mRNAs is inhibited, indicating that cap-independent translation initiated at an internal ribosomebinding site is utilized (50). Interestingly, the induction of Herp at the protein level is greater than GRP78 in our experiments (Fig. 6A). It will be interesting to address the molecular mechanism facilitating the translation of UPR-inducible proteins.
At present, the role of the activation of the UPR response in patients with hyperhomocysteinemia remains unknown. Development of a sensitive method to monitor ER stresses is required. The UPR-sensitive transcriptional regulation and undefined translational system of Herp may be useful in the understanding and treatment of this disease condition. FIG. 9. Herp on the ER membrane. The present study demonstrates that the majority of the Herp molecule on the ER membrane faces the cytoplasm. Hydropathic profile of the amino acid sequence of human Herp was obtained according to the algorithm of Kyte and Doolittle (41). Positive values represent increased hydrophobicity.