If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* The work was supported by Biocenter Oulu, the Wellcome Trust (United Kingdom), and the Academy of Finland. 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.
Native disulfide bond formation in the endoplasmic reticulum is a critical process in the maturation of many secreted and outer membrane proteins. Although a large number of proteins have been implicated in this process, it is clear that our current understanding is far from complete. Here we describe the functional characterization of a new 18-kDa protein (ERp18) related to protein-disulfide isomerase. We show that ERp18 is located in the endoplasmic reticulum and that it contains a single catalytic domain with an unusual CGAC active site motif and a probable insertion between β3 and α3 of the thioredoxin fold. From circular dichroism and NMR measurements, ERp18 is well structured and undergoes only a minor conformational change upon dithioldisulfide exchange in the active site. Guanidinium chloride denaturation curves indicate that the reduced form of the protein is more stable than the oxidized form, suggesting that it is involved in disulfide bond formation. Furthermore, in vitro ERp18 possesses significant peptide thiol-disulfide oxidase activity, which is dependent on the presence of both active site cysteine residues. This activity differs from that of the human PDI family in that under standard assay conditions it is limited by substrate oxidation and not by enzyme reoxidation. A putative physiological role for Erp18 in native disulfide bond formation is discussed.
Disulfide bonds are covalent linkages formed between two cysteine residues in proteins, whose primary function is to stabilize the folded structure of the protein. A large number of secreted proteins, including many high value proteins targeted by the biotechnology industry, contain disulfide bonds. Since any two cysteine residues in a protein have the potential to form a disulfide bond, the correct formation of native disulfide bonds is not trivial. Hence, it is not surprising that native disulfide bond formation is often the rate-limiting step in the folding of proteins in vitro and in vivo (
). However, whereas many of the participants in the cellular process are known, their individual roles are still largely confused. Furthermore, it is likely that not all of the participants in the process have yet been identified. This situation not only inhibits our understanding of the biogenesis of a range of important proteins but also prevents the effective manipulation of the cellular environment by the biotechnology industry for the efficient production of therapeutic proteins.
Native disulfide bond formation usually occurs via multiple parallel pathways (see Ref.
as an example). Each step in these pathways can be considered to be an oxidation or isomerization reaction. Oxidative reactions result in the formation of a disulfide bond from two free thiols. Isomerization reactions do not alter the overall number of disulfides, but they result in a rearrangement in their position within the substrate protein. Isomerization reactions are required due to the formation of non-native disulfide bonds in the substrate protein and may either occur directly or via a linked reduction-oxidation cycle. It is unclear to what extent, if any, reduction reactions (i.e. the formation of two free thiols from a disulfide) play a role in the formation of native disulfide bonds under physiological conditions. In vitro the rate-limiting step for native disulfide bond formation in proteins that contain multiple disulfides is late stage isomerization reactions, where disulfide bond formation is linked to conformational changes in protein substrates with substantial regular secondary structure (
). Whereas PDI is an excellent catalyst of isomerization reactions, it is a relatively poor catalyst of oxidative processes in vitro. Furthermore, the participation of PDI in disulfide bond formation requires the presence of a system for reoxidizing PDI. For many years, it was presumed that the primary reoxidant in the ER was oxidized glutathione. However, it has become clear that there are parallel pathways for providing the oxidizing equivalents required for native disulfide bond formation. ERo1 (
) and those of the other components including the glutathione biosynthetic pathway are not, it is still unclear to what extent the three possible oxidative pathways contribute to native disulfide bond formation under physiologically normal conditions.
A further complicating factor in understanding the physiological process of native disulfide bond formation is the presence of a family of proteins related to PDI. In S. cerevisiae, there are five proteins that are related to PDI (Pdi1p, Eug1p, Mpd1p, Mpd2p, and Eps1p) (see Refs.
for characterization). Whereas of these five proteins only Pdi1p can be considered to be an efficient catalyst of disulfide bond isomerization, it is clear that in higher eukaryotes this is not the case. Instead, a range of PDI homologues have been found, including ERp72, ERp57, P5, PDIp, and PDIr (see Refs.
), the function of both of which is currently unresolved.
Here we report a new addition to the ER-located PDI-related family of proteins, which we call ERp18. ERp18 contains a single catalytic domain with an unusual CGAC active site motif and a probable insertion between β3 and α3 of the thioredoxin fold. Although it is listed in several data bases as a thioredoxin-like protein, the reduced form of the protein is more stable than the oxidized form, suggesting that it is involved in disulfide bond formation and not reduction. Furthermore, in vitro ERp18 possesses significant peptide thiol-disulfide oxidase activity. A putative physiological role for Erp18 in native disulfide bond formation is discussed.
Generation of Expression Vectors—Plasmids encoding for full-length and mature (Ser24–Leu172) human ERp18 were generated by PCR from a liver cDNA library (Clontech) using primers that included an in frame NdeI site 5′ to the first codon of the gene and a SalI site after a TAA stop codon at the 3′-end. The inserts were cloned into pLWRP51 (
), a modified version of pET23b (Novagen), which encodes for a N-terminal His tag (MHHHHHHM) prior to the first amino acid of the protein sequence. The vector encoding mature human ERp18 (pHIA128) was used as the template to PCR-amplify gene fragments encoding C-terminal truncations (Ser24–Gln146, Ser24–Gly150, Ser24–Arg157, and Ser24–Leu168), which were also cloned into NdeI/SalI-digested pLWRP51. The vector encoding full-length human ERp18 (pHIA131) was used to PCR-amplify ERp18 without a stop codon, which was cloned in frame into BglII/PinAI-digested p-EGFPN1, generating a ERp18-GFP chimera with a GPVAT linker. Since ERp18 contains a putative C-terminal ER retention signal, pHIA131 was also used to clone residues Met1-Leu168 into pHIA135 (a modified version of p-EGFPN1 in which synthetic oligonucleotides had been used to add Leu168–Leu172 of ERp18 in frame C-terminally with GFP). The resulting plasmid encoded ERp18 (Met1–Leu168)-GPVAT linker-GFP-ERp18 (Leu168–Leu172). A plasmid encoding for full-length Escherichia coli thioredoxin was generated by PCR using an XL1-blue E. coli colony as a template and primers that included an in frame NdeI site 5′ to the first codon of the gene and a BamHI site after a TAA stop codon at the 3′-end. The insert was cloned into pLWRP51, and the resulting gene product included the sequence MHHHHHHM- prior to the first amino acid of the domain sequence. All plasmids generated were sequenced to ensure that there were no errors in the cloned genes (see Table I for plasmid names).
Protein Expression and Purification—Protein production was carried out in E. coli strain BL21 (DE3) pLysS. Strains were grown in LB medium at 37 °C and 200 rpm and induced at an A600 of 0.3 for 4 h with 1 mm isopropyl β-d-thiogalactoside. Cells were pelleted by centrifugation (6,500 rpm for 10 min), and the pellet was resuspended in one-tenth volume of buffer A (20 mm sodium phosphate, pH 7.3) and one-one thousandth volume of 10 mg/ml DNase (Roche Applied Science). The cells were lysed by freeze-thawing twice, and the cell debris were removed by centrifugation (12,000 rpm for 20 min). The supernatant was filtered through a 0.45-μm filter before being applied to an immobilized metal affinity chromatography column (chelating Sepharose fast flow; Amersham Biosciences), precharged with Ni2+ and equilibrated in buffer A. After loading, the column was washed in 20 mm sodium phosphate, 50 mm imidazole, 0.5 m NaCl, pH 7.3, and then in buffer A, before the His-tagged proteins were eluted using 20 mm sodium phosphate, 50 mm EDTA, pH 7.3. The eluant was diluted 5× into buffer A and then applied to a Resource Q anion exchanger, from which the proteins were eluted with a linear gradient from buffer A to buffer A containing 0.5 m NaCl over 9 column volumes. Eluted fractions were checked for purity by SDS-polyacrylamide gel electrophoresis, and fractions containing pure protein were pooled and buffer-exchanged into 20 mm sodium phosphate, pH 7.3, using an Amicon ultra 15 centrifugal filter device (10-kDa nominal molecular weight limit membrane filter). The concentration of the protein was determined spectrophotometrically using a calculated absorption coefficient of 16,680 m–1 cm–1 at 280 nm. 15N-Labeled ERp18 was produced by growing the expressing strain in M9 medium using 15N-labeled NH4Cl (Cambridge Isotopes) with protein purification as described for unlabeled protein. To ensure full oxidation of the active site CXXC motif prior to NMR studies, 15N-labeled ERp18 was incubated with 0.5 mm oxidized glutathione for 15 min at room temperature, and then the glutathione was removed by buffer exchanging using an Amicon ultra 15 centrifugal filter device (10-kDa nominal molecular weight limit membrane filter) into 20 mm sodium phosphate, 150 mm sodium chloride, pH 6.5. The a-domain of human PDI was purified as per human ERp18 (calculated absorption coefficient of 19,720 m–1 cm–1 at 280 nm). E. coli thioredoxin was purified as per human ERp18 except that elution from the chelating Sepharose column was with 25 mm EDTA, and the anion exchange column was run in 20 mm Tris buffer, pH 8.6, instead of buffer A.
Cell Transfections—COS-7 cells (ATCC, Manassas, VA) were grown on 30-mm diameter Petri dishes with or without glass coverslips in Dulbecco's modified Eagle's medium-high glucose medium supplemented with Glutamax (Invitrogen), 10% fetal calf serum, and penicillin/streptomycin. Cells seeded 1 day earlier were transfected with the ERp18-GFP plasmid using 0.5–1 μg/plate and the Fugene6™ transfection reagent (Roche Applied Science) as suggested by the manufacturer. After 24 h, cells were rinsed with phosphate-buffered saline, fixed with 4% p-formaldehyde for 20 min, and processed for indirect immunofluorescence as described earlier (
). Monoclonal antibodies against protein-disulfide isomerase (PDI, Dako A/S, Glostrup, Denmark) and the Golgi matrix protein, Gm130 (BD Biosciences, Lexington, KY) were used as the ER and Golgi markers, respectively, to allow localization of the expressed ERp18-GFP in transfected cells. Fixed and stained cells were examined using an epifluorescence microscope (Olympus BX61) and photographed with a CCD camera.
Biophysical Analysis—Far UV CD spectra were recorded on a Jasco J600 spectrophotometer. All scans were collected at 25 °C as an average of eight scans, using a cell with a path length of 0.1 cm, scan speed 50 nm/min, a spectral bandwidth of 1.0 nm, and a time constant of 0.5 s. The maximal HT voltage was 750 V.
Fluorescence spectra were collected on a PerkinElmer Life Sciences LS50 spectrophotometer using a 1-ml cuvette. All scans were collected at 25 °C as an average of four scans, excitation at 280 nm, emission at 300–400 nm, slit widths of 5 nm, and scan speed of 200 nm/min. Fully oxidized and reduced proteins were generated immediately prior to use by preincubating the protein stock in 10 mm oxidized glutathione or 20 mm reduced glutathione for 15 min at room temperature. Protein stocks were diluted at least 200-fold into 0.2 m phosphate buffer, pH 7.0, containing 0–6 m guanidinium chloride and equilibrated for 5 min at 25 °C before fluorescence spectra were recorded. All spectra were corrected for the blank spectra with no protein added. The fluorescence parameter examined to investigate the effects of guanidinium chloride on protein structure was the ratio of the average fluorescence intensity 2 nm to either side of the λmax for native protein (337 nm for oxidized ERp18; 336 nm for reduced ERp18; 340 nm for PDI a-domain) to the average fluorescence intensity over the range 320–400 nm. This parameter was chosen because it is independent of concentration and less dependent on the direct effects of guanidinium chloride on tryptophan fluorescence.
NMR spectra were collected on a Varian Inova 600-MHz spectrometer from samples of uniformly 15N-labeled ERp18 (0.3 mm) in 20 mm sodium phosphate buffer (pH 6.5) containing 150 mm NaCl and 10% (v/v) D2O. Reduced ERp18 was prepared by preincubating the oxidized protein with 3 mm dithiothreitol for 30 min at room temperature. 1H/15N HSQC spectra were collected at 25 °C over 6 h with acquisition times of 148 ms in 1H and 71 ms in 15N and with water suppression using the WATERGATE sequence.
) using a fluorescent decapeptide was used to determine the oxidase activity of each of the purified human PDI family members. McIlvaine buffer (0.2 m disodium hydrogen phosphate, 0.1 m citric acid, pH 3.0–7.5) to give a final assay volume of 1 ml was placed in a fluorescence cuvette. Except where noted, to this was added 10 μl of oxidized glutathione (50 mm stock solution in buffer A), 20 μl of reduced glutathione (100 mm stock solution in buffer A), and 5–10 μl of enzyme. After mixing, the cuvette was placed in a PerkinElmer Life Sciences LS50 spectrophotometer for 5 min to allow thermal equilibration of the solution. 6.3 μl of substrate peptide (539 μm in 30% acetonitrile, 0.1% trifluoroacetic acid) was added and mixed, and the change in fluorescence intensity (excitation at 280 nm, emission at 350 nm, slit widths at 5 nm) was monitored over an appropriate time (usually 15 min) with 600–1800 data points being collected. In the absence of substrate or in the absence of oxidized glutathione, no significant change in fluorescence occurred (e.g. for ERp18 alone, the change in fluorescence over 30 min was only 0.65%, which probably represents photobleaching).
Identification of a Novel Thioredoxin Superfamily Member—A data base search for novel human PDI family members identified a previously uncharacterized open reading frame (NP_056997.1; Q96H50; ENSG00000117862). The gene is located at bp 51374826–51410866 on chromosome 1 and has seven exons. The mature form (see below) of the predicted protein has a molecular mass of 17.8 kDa and is expected to be localized in the endoplasmic reticulum (see below), and accordingly we have named it ERp18.
) indicates that ERp18 is widely expressed in humans including in the kidney, brain, prostate, lung, liver, heart, spinal cord, mammary gland, ovary, colon, and vascular epithelium. Unigene (J. U. Pontius, L. Wagner, and G. D. Schuler; available on the World Wide Web at www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) cDNA sources include liver, stomach, uterus, bone, placenta, brain, adrenal gland, ovary, prostate, lung, testis, pancreas, and kidney. The ERp18 coding region used in these studies was PCR-amplified from a liver cDNA library.
Homologues of ERp18 can be found in mouse, rat, Xenopus, and Caenorhabditis elegans. Multiple alignments of these four proteins were performed using ClustalW (
). These alignments showed considerable consensus, varying only in the extreme N- and C-terminal regions. A consensus alignment that takes into account the probable structural alignment with the thioredoxin fold (see below) is shown in Fig. 1. Over the mature protein (see below), the human and mouse proteins show 94.6% identity, human and rat proteins show 94.0% identity, human and Xenopus proteins show 81.2% identity, and human and C. elegans proteins show 41.6% identity.
Analysis of the Domain Structure of ERp18 —Analysis of the sequence of ERp18 using PSORT (available on the World Wide Web at psort.ims.u-tokyo.ac.jp/) and TargetP (
) indicated that the protein would be targeted by a cleavable N-terminal signal sequence to the secretory pathway, with the mature protein starting at Ser24 (PSORT) or His27 (TargetP). The C-terminal amino acids are EDEL, which probably acts as an ER retrieval motif (usual consensus KDEL).
If the mature protein starts at Ser24, then it is 149 amino acids long, with a molecular mass of 16.7 kDa and a theoretical pI of 5.1 (ProtParam). The protein contains one CXXC thioredoxin superfamily active site motif, located at Cys66–Cys69, and no other cysteine residues. The nature of the two residues between the active site cysteines plays a key role in determining the redox potential of the thioredoxin superfamily and hence their biological activity. In the thioredoxins, which act as reductants, the motif is CGPC, whereas in the PDI family, which act as oxidants and isomerases, it is usually CGHC, and in the DsbA family, which act as oxidants, it is usually CPHC. The sequence in ERp18 is CGAC, which does not directly fit into any of these families.
Multiple alignments of the catalytic domains of the human PDI family with ERp18 were performed using ClustalW (
), and alignments of pairs of proteins were performed with a wide range of programs. These alignments showed some degree of consensus at the N terminus of the proteins, indicating that the probable residue starting a putative thioredoxin fold in human ERp18 was Ile38 (equivalent to Val26 in human PDI) but showed very significant differences for the alignment for the C-terminal part of ERp18. ERp18 is clearly longer than a single PDI family member catalytic domain, and the different alignments indicated that the additional sequence could either occur (i) at the C terminus (i.e. beyond the end of the thioredoxin domain), (ii) as an insert between secondary structure elements β3 and α3 or α3 and β4, or (iii) as a combination of these. To help to distinguish between these three possibilities, four C-terminal deletion constructs were generated that truncated the protein at a point where an alignment with the PDI catalytic domain ended. If the alignment was correct, then it would be reasonable to expect that the truncated protein might retain the ability to fold and form stable structures, since they would still encompass the entire thioredoxin-like fold. The constructs terminating with Gln146 or Gly150 were expressed insolubly in the cytoplasm of E. coli under all conditions tested, whereas a construct terminating with Arg157 generated a small percentage of soluble material, which was clearly unstable. Only constructs expressing mature human ERp18 (Ser24–Leu172) or with the putative ER retention signal deleted (Ser24–Leu168) were solubly and stably expressed in the cytoplasm of E. coli (data not shown).
Localization of ERp18 —ERp18 contains a putative secretory pathway signal sequence and a putative ER retention signal; hence, it is most probably an ER-resident protein. To investigate the subcellular localization of ERp18, two ERp18-GFP chimeras were constructed. The first had the whole ERp18 gene fused in frame N-terminally to GFP; the second had amino acids Met1–Leu168 of ERp18 fused in frame to the N terminus of GFP and Leu168-Leu172 of ERp18 fused in frame C-terminal to GFP. Direct microscopic examination of the transfected cells showed that the first ERp18-GFP fusion protein could not be visualized intracellularly. The second chimera localized in a fine reticular network and the nuclear envelope (Fig. 2), suggesting that the fusion protein localizes mainly in the ER. This was confirmed by staining of the transfected cells with antibodies against the known ER and Golgi markers (PDI and Gm130). The chimeric protein, with ERp18 (including the signal sequence) N-terminal to GFP and the retrieval signal (EDEL) C-terminal to GFP, co-localized well with PDI but not with the Golgi marker Gm130. The chimeric protein appears also to be efficiently retained in the ER (via the EDEL), since no marked accumulation of the protein in the Golgi region was observed (Fig. 2, right, ERp18-GFP).
Biophysical Analysis of ERp18 —Mature ERp18 (Ser24–Leu172), with an N-terminal hexa-His tag to aid purification, was expressed solubly in the cytoplasm of E. coli, and the construct was purified to apparent homogeneity by a combination of immobilized metal affinity chromatography and anion exchange chromatography (data not shown). Mass spectrometric analysis of the purified protein by MALDI mass spectrometry (mass accuracy 0.1%) gave a mass of 17,768.6, close to the calculated mass for oxidized ERp18 of 17,768.9. Purified ERp18 was then analyzed by a variety of techniques.
The far UV CD spectra of purified ERp18 (Fig. 3A) indicated that the protein was well structured and contained both α-helix and β-sheet. All members of the thioredoxin superfamily including thioredoxin and the catalytic a-domain of human PDI share the same α/β fold (see Refs.
as examples). Since ERp18 showed considerable homology with the N-terminal region of the catalytic domains of the PDI family, the far UV CD spectra of purified ERp18 was compared with those of the a-domain of human PDI and E. coli thioredoxin. The resulting spectra are significantly different for all three proteins (see Fig. 3A).
Fluorescence spectra of ERp18 under nondenaturing conditions gave a peak with a λmax of 337 nm and a shoulder at around 351 nm, indicating that one of the two tryptophans of ERp18 is in a hydrophobic environment, whereas the other is in a hydrophilic environment (see Fig. 3B). Similar observations were made for the a-domain of human PDI, which also contains 2 tryptophan residues (λmax 340 nm; data not shown). Upon the addition of 6 m guanidinium chloride, the fluorescence spectra of ERp18 had a λmax of 356.5 nm, indicative of a denatured protein (see Fig. 3B).
The redox potential of members of the thioredoxin superfamily is dependent on the relative stability of the oxidized and reduced forms of the active site CXXC motif, and this determines whether individual proteins act as oxidants, reductants, and/or isomerases. Guanidinium chloride denaturation curves for oxidized and reduced ERp18 (Fig. 4A) and the a-domain of human PDI (Fig. 4B) indicated that for both proteins the reduced form of the protein is significantly more stable than the oxidized. This is consistent with an oxidative function for both proteins. Whereas the midpoint for denaturation for both proteins in the oxidized or reduced states is approximately equal, denaturation of ERp18 occurs over a narrower concentration range of guanidinium chloride. This indicates that the stability of ERp18 under nondenaturing conditions is higher than that of the a-domain of human PDI. Using the six-component equation for denaturant-dependent changes in ΔG (
), midpoints for denaturation, ΔG0, and ΔΔG0 between the oxidized and reduced forms were calculated (see Table II). These results indicate that ERp18 is more stable than PDI a-domain in both the oxidized and reduced forms and that it has a nearly equivalent redox potential. However, it should be noted that ΔG0 values calculated from denaturation curves, especially curves with sharp transitions, are prone to error, and hence these values should be treated with some caution.
Table IIThermodynamic parameters derived from the line of best fit to the guanidinium chloride denaturation curves for oxidized and reduced human ERp 18 and human PDI a-domain
Structural Measurements of ERp18 —To further analyze the structure of ERp18, 15N-labeled material was generated. MALDI mass spectrometry indicated a 95% labeling efficiency. The resolution and dispersion of the 1H/15N HSQC spectra for both oxidized and reduced ERp18 (Fig. 5) suggested that in both states the protein existed as a fully folded and compact molecule. The cross-peaks represent N-H pairs in the protein and derive mainly from the backbone, but they also include some side chain NH2 groups. A second HSQC experiment where the side chain NH2 signals were selectively suppressed (data not shown) allowed the number of backbone amide signals to be counted. The number was estimated to be 141, which is in good agreement with the number of nonproline residues (n = 147) in the His-tagged ERp18 construct.
A comparison of the oxidized (Fig. 5A) and reduced (Fig. 5B) spectra of ERp18 showed that the majority of the backbone amide resonances remained unchanged in position upon reduction of the Cys66–Cys69 active site disulfide bond. This finding suggests that the overall structure of the protein is unperturbed by a change in oxidation state. However, significant changes in position were seen for 5 or 6 residues, and about a further 25 residues shifted position slightly. This degree of change is consistent with a local structural rearrangement, presumably centered on the CGAC redox active motif.
Activity Measurements of ERp18 —To measure the relative peptide oxidase activity of ERp18, a simple fluorescence-quenching assay was used (
). In this assay, the formation of a disulfide bond in the substrate peptide can be monitored in real time, since this brings an arginine residue in the peptide into close proximity with the single tryptophan, resulting in quenching of the intrinsic fluorescence. Under standard assay conditions at pH 6.5, the ERp18 showed significant peptide oxidase activity; however, this was considerably lower than that of the a-domain of human PDI (see Table III and Fig. 6). The rate of reaction was directly proportional to enzyme concentration (data not shown). Varying the concentration of oxidized glutathione in the assay indicated that the rate-limiting step for ERp18 peptide oxidase activity was oxidation of the substrate and not reoxidation of reduced enzyme by oxidized glutathione as is the norm for other PDI family members (e.g. see Ref.
). To further investigate this, the peptide substrate concentration was varied to estimate the Km and Vmax for ERp18. The results (see Table III) indicate that the Km of ERp18 for the substrate peptide is greater than can be determined by the dynamic range of the fluorescence-based assay used ([substrate] is up to 8 μm; an estimate can be made of Km of the order of 25 μm, which is 1 order of magnitude greater than the apparent Km of the other human PDI family members, excluding PDIr, under standard assay conditions).
All thioredoxin superfamily members characterized to date have a CXXC active site motif, with both cysteine residues being required for oxidative activity. Single point mutants were generated in both active site cysteines of ERp18, C66S and C69S. The corresponding proteins were purified, the correct mass was confirmed by MALDI mass spectrometry, and the proteins were tested for peptide oxidase activity. Both mutants gave no detectable activity in the peptide oxidase assay (see Table II).
ERp18 is listed in several data bases as thioredoxin-like protein p19 (TLP19). Whereas it is clearly a member of the thioredoxin superfamily, the data presented here suggest that it is involved in disulfide bond formation and not reduction (i.e. that it is not a thioredoxin but it is instead more closely related to the PDI family).
Three motifs can be used as a diagnostic of which family in the thioredoxin superfamily a protein is likely to belong to: the CXXC active site motif, the residues from the end of β2 to the end of α2 (which include this motif), and the residues around the conserved cis-proline (Pro100 in the a-domain of human PDI). In the thioredoxins, the conserved active site motif is CGPC, in the PDI family it is usually CGHC, and in the DsbA family it is usually CPHC. The CGAC sequence found in ERp18 does not belong to any of these families but is closer to the motif found in PDI and thioredoxin than to the DsbA motif. The residues around the active site motif in ERp18, WCGACKALK-PKF, are the same as the conserved motif of the human PDIs (excluding the uncharacterized protein PDIr), WCGHCKX(L/M/F)XPX(Y/W/F) and the thioredoxin consensus sequence, WCGPC(K/R)X(I/F/L)XP. However, since it contains an aromatic residue +2 from the fully conserved proline residue in the PDI sequence, which forms an important structure-defining kink in α2 of the thioredoxins and in human PDI a-domain (
), the motif in ERp18 is marginally closer to that of the PDIs. The third diagnostic motif is centered on the conserved proline residue (Pro100 in the a-domain of human PDI), which in all known superfamily structures is in the cis conformation. The sequence in human ERp18, GNPSYKY, does not directly fit into the consensus sequence of any of the thioredoxin superfamily families. However, it most closely resembles that of the human PDI family, G(F/Y)PT(I/L)X(F/Y/I), with X often representing lysine, since it has a Gly at the –2-position (a feature shared with the DsbAs but not the thioredoxins or glutaredoxins).
Whereas the comparison of the consensus sequences suggests that, although not a member of the PDI family, ERp18 more closely resembles the PDIs than the thioredoxins, it does not provide strong or direct evidence of function. However, the results reported in this study on the ER localization of ERp18 and the greater stability of the reduced over oxidized form of the enzyme (a property shared with PDI and DsbA but not with thioredoxin) are indicative of a role for ERp18 in disulfide bond formation. This is further reinforced by the significant peptide oxidase activity of ERp18, for which both active site cysteine residues were found to be required. The oxidase assay was carried out in a glutathione redox buffer that mimics the redox potential of the ER, the compartment in which ERp18 is localized, and hence indicates a physiologically relevant function.
It is clear that the sequence of mature ERp18 (149 amino acids) is significantly longer than the catalytic domain of thioredoxin or a catalytic domain of PDI (around 110 amino acids). It is possible that ERp18 contains either an N-terminal or C-terminal extension, as is seen in all human PDIs (
) or that it contains an insertion in the thioredoxin fold as is seen in the DsbAs. All of the alignment programs used indicated that ERp18 has a 14-amino acid N-terminal extension, which is larger than that found in human PDI or ERp57 but shorter than that found in human PDIp or ERp72 (
). One of the four alignment programs used placed another long extension at the C terminus of the protein, whereas the others placed an insertion between either β3-α3 or α3-β4. The later programs align Pro135 with the conserved cis-proline found in the thioredoxin superfamily, and the region around this proline (GNPSY) does align better with the highly conserved sequence, G(F/Y)PT(I/L), than the region that aligns by the insertion being at the C terminus (Pro113, YIPRI). An insertion into the thioredoxin fold at this proposed position would not be unique, since the DsbA family have an all α-helical domain inserted between β3 and α3 of this fold (
). The nature of the insertion and the C-terminal extension varied between alignment programs, and C-terminal truncations were made to elucidate which alignment was most probable. The inability for C-terminal truncations deleted from Val147, Met151, and Leu158 to fold in vivo, whereas a truncation at Leu168 could still fold, suggests that the most probable alignment is that which inserts the sequence EEEPKDEDFSPDGGYIPRILFLD into the loop between β3 and α3 (Fig. 7). This alignment also places the insertion in the thioredoxin fold at the same location as the large insertion (∼75 amino acids) found in the thioredoxin fold in the DsbAs and puts the insertions/deletions found in the homologous Xenopus and C. elegans sequences (Fig. 1) outside the limits of the thioredoxin fold or in the loops between the secondary structural elements. The insertion in human ERp18 being both acidic and containing a significant number of P/G residues is reminiscent of the interdomain linkers found in PDIr (
). Whereas the CD spectra of ERp18 is significantly different from that of the a-domain of human PDI, so is the spectra of E. coli thioredoxin, which shares the same fold; hence, nothing can be inferred from this other than the fact that human ERp18 is well structured. It is clear from the 15N/1H HSQC NMR spectra that there are no significant unstructured regions in ERp18 and that the overall structure of ERp18 is unperturbed by the redox state of the active site (as seen for all other superfamily members studied to date) (see Refs.
Although the multiple alignments of the catalytic domains of the PDI family with ERp18 differed significantly, the consensus regions for all indicated that ERp18 does not possess the conserved (except in PDIr) glutamic acid (Glu47 in the a-domain of human PDI), which is important for the catalytic redox cycle of the PDI family
). In addition, the unusual active site motif, CGAC, does not contain a histidine residue found in both the DsbA (CPHC) and PDI (CGHC) families, which has been implicated in helping to generate an oxidizing redox potential for the active site (
). However, ERp18 has significant oxidase activity (which is limited by substrate binding), and it shows nearly identical guanidinium chloride stability curves compared with the a-domain of human PDI for both the oxidized and reduced states. This indicates that ERp18 must stabilize the reduced state of the enzyme by a different method from that used by PDI. We have recently reported that modulations in His94 in the all-helical domain insert in Vibrio cholerae DsbA, which spatially is located close to the active site, modulates the activity of the enzyme (
), and it is possible that the putative insert in the thioredoxin fold in ERp18 plays a similar role.
Whereas the data presented here suggest a role for ERp18 in disulfide bond formation in the ER, it does not define what that function might be. The 186-amino acid C. elegans homologue of human ERp18 (Y57A10A.23) shows no observable RNA interference phenotype (WormBase Web site, available at www.wormbase.org, release WS99), implying that the function of this gene product is not essential. However, C. elegans also contains a 257-amino acid protein (F49H12.5) whose N-terminal region shows 44.3% identity with Y57A10A.23 and whose C-terminal 88 amino acids are composed almost entirely (94.3%) of Lys, Glu, and Asp. F49H12.5 also shows no observable RNA interference phenotype (WormBase Web site, available at www.wormbase.org, release WS99). It is possible that there may be functional complementation between these proteins. The results from the oxidase activity of ERp18 are unusual in that the rate-limiting step in the process is not reoxidation of the enzyme by oxidized glutathione, but instead oxidation is limited by interactions of ERp18 with the substrate. Whereas it is possible that ERp18 has a different substrate specificity than PDI, it should be noted that with this peptide substrate all human PDI family members (except PDIr) have nearly equivalent molar active site activities, and all are rate-limited by reoxidation by glutathione.2 Furthermore, to our knowledge, the oxidation of all reported substrates by PDI is limited by reoxidation (see Refs.
for examples). If ERp18 has a thioredoxin fold with an insert between β3 and α4 (as seen in DsbA), then this insert will be spatially close to the active site and may regulate access to it. The rapid reoxidation of ERp18 in vitro by glutathione and the low affinity for peptide substrates may suggest a role for ERp18 in shuffling oxidizing equivalents from ERo1 to luminally located proteins, a role played in S. cerevisiae by Mpd2p (