|Home | About | Journals | Submit | Contact Us | Français|
Folding of secretory proteins is associated with the formation and isomerization of disulfide bonds. ERp72, a protein disulfide isomerase (PDI) family member, possesses 3 thioredoxin homology domains, but the participation of each domain in disulfide-bond formation and isomerization remains to be determined. We analyzed the function of individual domains in the insulin reduction assay system by site-directed mutagenesis with cysteine-to-serine replacement. All domains contributed to apparent steady-state binding (Km) and catalysis at saturating substrate concentrations (kcat) but in different manners. A mutant ERp72 with mutations in domains 1 and 2 (ERp72-mut-1+2) exhibited reductions in kcat of 73.9% when compared with wild type, whereas ERp72-mut-1+3 (mutations in domains 1 and 3) and ERp72-mut-2+3 (mutations in domains 2 and 3) exhibited less substantial reductions in kcat. ERp72-mut-1+3 and ERp72-mut-2+3 showed elevations in Km of 89.9% and 96.2%, respectively, when compared with wild type, whereas ERp72-mut-1+2 exhibited smaller elevations in Km. These results suggest that domains 1 and 2 make greater contributions to catalyzing efficacy and domain 3 to binding affinity. Domain 2 is involved in binding affinity, in combination with domain 3, in addition to its own contribution to catalyzing efficacy. This assignment of functions to individual domains is similar to that observed in other PDI domains, which is consistent with the high sequence homology between ERp and PDI domains.
The folding of nascent proteins in cellular secretory pathways is associated with the formation and isomerization of disulfide bonds (Pfeffer and Rothman 1987; Rapoport et al 1999; Trombetta and Parodi 2003). Numerous proteins resident in the endoplasmic reticulum (ER) have been identified as chaperones and folding enzymes that assist in protein folding, provide retention, and enable quality control (Gething and Sambrook 1992; Hammond and Helenius 1995). Protein disulfide isomerase (PDI), one of the first described folding enzymes, contains 2 thioredoxin homology domains and catalyzes the formation and isomerization of disulfide bonds during protein folding (Hillson et al 1984; Freedman et al 1995; Gilbert 1998). The thioredoxin homology domain contains a Trp-Cys-Gly-His-Cys-Lys (WCGHCK) motif that is identical to the active site consensus sequences of PDI (Lee 1981; Mazzarella et al 1990, 1994; Freedman et al 1995; Gilbert 1998). Additional proteins that possess the thioredoxin homology domains and are localized in the ER, such as ERp61and protein disulfide isomerase-related protein (PDIR), comprise the PDI family and are regarded as folding enzymes that participate in the formation and isomerization of disulfide bonds in secretory proteins (Mazzarella et al 1990; Martin et al 1991; Srivastava et al 1991; Hayano and Kikuchi 1995; Otsu et al 1995).
ERp72 that has been identified as a luminal ER protein of murine plasmacytoma cells and, together with its rat liver homologue CaBP2, bears 3 thioredoxin homology domains and an Lys-Glu-Glu-Leu (KEEL) sequence at the C-terminus (Mazzarella et al 1990; Lenny and Green 1991; Van et al 1993). ERp72 shares the sequence (WCGHCK) responsible for its redox activity with other PDI proteins. However, the functional roles of individual thioredoxin homology domains remain to be elucidated.
In this study, we characterized the functional properties of the individual thioredoxin homology domains of ERp72 using mutant domains in which cysteine in the catalytic center was replaced with serine and by assaying insulin reduction activity.
ERp72 complementary deoxyribonucleic acid (cDNA) was kindly provided by Dr M. Green and was amplified by polymerase chain reaction with a primer set (CGGGATCCGGAGATGCACAGGAAGATAC and CGGGATCCAAGCTCTTCCTTGGTCCTGC), Pfu DNA polymerase, and pGEM72-1 as a template (Mazzarella et al 1990). Amplified fragments were digested with BamHI, and the cDNA fragments were subcloned into the BamHI site of pBluescriptIIKS− (Stragaene, La Jolla, CA, USA) (designated as pBS-ERp72). The NcoI fragment was removed from pBS-ERp72, which was then self-ligated (pBS-ERp72ΔNcoI). Alternatively, pBS-ERp72 was digested with a combination of BamHI and SacI, and smaller fragments of ERp72 cDNA were subcloned into the BamHI-SacI site of pBluescriptIIKS− (pBS-ERp72BS). Mutations were introduced into the thioredoxin homology domains using the QuikChange™ site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Primer sets and templates for introducing mutations were as follows: for changing both Cys-60 and Cys-63 to Ser (C60S and C63S) (mut-1), a primer set (CTATGCACCATGGTCTGGACACTCCAAGCAGTTTGCTC and GAGCAAACTGCTTGGAGTGTCCAGACCATGGTGCATAG) and pBS-ERp72BS were used; for changing both Cys-175 and Cys-178 to Ser (C175S and C178S) (mut-2), a primer set (CTATGCCCCGTGGTCTGGACACTCCAAGAAACTTGCC and GGCAAGTTTCTTGGAGTGTCCAGACCACGGGGCATAG) and pBS-ERp72BS were used; and for changing both Cys-524 and Cys-527 to Ser (C524S and C527S) (mut-3), a primer set (CTATGCACCCTGGTCTGGGCACTCCAAGCAGCTAGAG and CTCTAGCTGCTTGGAGTGCCCAGACCAGGGTGCATAG) and pBS-ERp72ΔNcoI were used. The resultant plasmids were designated pBS-ERp72BS-mut-1, pBS-ERp72BS-mut-2, and pBS-ERp72ΔNcoI-mut-3, respectively. Each mutation was confirmed by DNA sequencing.
The SacI fragment from pBS-ERp72 was subcloned into the SacI site of pBS-ERp72BS-mut-1 and pBS-ERp72BS-mut-2 (designated pBS-ERp72-mut-1 and pBS-ERp72-mut-2, respectively). The NcoI fragment of ERp72 was subcloned into the NcoI site of pBS-ERp72ΔNcoI-mut-3 (pBS-ERp72-mut-3), and the NcoI fragment of pBS-ERp72-mut-3 was replaced with the NcoI fragments of pBS-ERp72-mut-1 and pBS-ERp72-mut-2 (pBS-ERp72-mut-1+3 and pBS-ERp72-mut-2+3, respectively). The NheI-HindIII fragment of pBS-ERp72-mut-1 was replaced with the NheI-HindIII fragment of pBS-ERp72-mut-2 (pBS-ERp72-mut-1+2). Finally, the NcoI fragment of pBS-ERp72-mut-3 was replaced with the NcoI fragment of pBS-ERp72-mut-1+2 (pBS-ERp72-mut-1+2+3). All plasmids containing either wild-type or mutant ERp72 were digested with BamHI, and the resultant fragments were separately subcloned into the BamHI-BglII site of pQE16 (Qiagen, Valencia, CA, USA) (pQE-ERp72, pQE-ERp72-mut-1, pQE-ERp72-mut-2, pQE-ERp72-mut-3, ERp72-mut-1+2, ERp72-mut-2+3, and ERp72-mut-1+2+3, respectively). Expression and purification of the wild-type and mutant sequences were carried out, according to the manufacturer's instructions, using Ni-NTA agarose (Qiagen). Further purification of these proteins was performed using a Resource Q column (Pharmacia Biotech AB, Uppsala, Sweden) after dilution with 20 mM phosphate buffer (pH 7.2) until the proteins were bound.
Wild-type ERp72 protein is known to catalyze the reduction of disulfide bonds in proteins such as insulin. The ERp72-catalyzed reduction of insulin by glutathione, reduced form (GSH) (Sigma, St Louis, MO, USA) was monitored by coupling the reduction of the product, oxidized form of glutathione (GSSG), to reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidation with glutathione reductase, as described elsewhere (Gilbert 1998). In brief, glutathione reductase (16 U, Oriental Yeast Co. Ltd, Tokyo, Japan) was preincubated with GSH (3.7 mM, Sigma) and NADPH (0.12 mM, Oriental Yeast ) in reaction buffer (0.2 M phosphate buffer and 5 mM ethylenediamine-tetraacetic acid, pH 7.5) for 30 seconds to remove contaminating GSSG. Insulin (Sigma) was added, and the velocity of nonenzymatic insulin reduction was measured by monitoring the decrease in NADPH absorbance at 340 nm for 2 minutes (velocity of nonenzymatic reaction was 0.0013 to 0.0018 AU/min). The catalytic reaction was then initiated by addition of either recombinant wild-type or mutant ERp72 protein. The enzyme activity in the coupled reaction was determined spectrophotometrically by monitoring the decrease in absorbance at 340 nm for 2 minutes using the kinetics assay mode of an Ultraspec 2000 spectrophotometer (Pharmacia Biotech AB) with a peltier-heated cell holder. After subtracting the background rate, the specific activity was calculated based on a Δ340 of 6.23 mM−1 cm−1.
Experimental data were analyzed by nonlinear least squares fitting with equation . Individual points were means of multiple replications (duplicate or higher) performed independently on different days.
where Km represents the Michaelis constant and Vmax represents the maximum velocity. Calculations of Vmax and Km were performed using SPSS software (SPSS Inc, Chicago, IL, USA). Catalysis at saturating substrate concentrations (kcat) represents turnover number per minute.
We focused on the conserved cysteine residues of thioredoxin homology domains, and each residue was replaced with serine using site-directed mutagenesis. Figure 1 shows the active sites of the thioredoxin domains of PDI and ERp72 and lists the mutant proteins generated. Figure 2 shows the Coomassie brilliant blue staining pattern of the purified recombinant wild-type and mutant ERp72 proteins.
We performed a pilot insulin reduction study using wild-type ERp72 to determine its Km. We then measured insulin reduction by wild-type and mutant ERp72 proteins with insulin concentrations being near Km. The activity traces for wild-type and mutant proteins are shown in Figure 3.
The Km and kcat values for wild-type and mutant ERp72 proteins are shown in Table 1. The Km of ERp72-mut-1+3 and ERp72-mut-2+3 were elevated by 89.9% and 96.2%, respectively, whereas that of ERp72-mut-1+2 was only elevated by 32.7%. In contrast, the kcat of ERp72-mut-1+2 was reduced by 73.9%, whereas those of ERp72-mut-1+3 and ERp72-mut-2+3 were only reduced by 33.2% and 49.1%, respectively. The kcat/Km ratios for ERp72-mut-1+2, ERp72-mut-1+3, and ERp72-mut-2+3 were 20%, 27.2%, and 34.4%, respectively. The assay for ERp72-mut-1+2+3 was performed with a 3-fold higher amount of mutant protein, but activity was not detectable. Therefore, the Km and Vmax for ERp72-mut-1+2+3 were not determined.
The Km of ERp72-mut-1 mutant was almost equal to that of wild type, whereas those of ERp72-mut-2 and ERp72-mut-3 were elevated by 57.0% and 84.4%, respectively. In contrast, the kcat values of ERp72-mut-1 and ERp72-mut-2 were reduced by 57% and 42.3%, respectively, whereas that of ERp72-mut-3 was almost equal to that of wild type.
ERp72, a member of the PDI family, has been identified as a luminal ER protein and is believed to participate in the formation and isomerization of disulfide bonds in cellular secretory proteins (Mazzarella et al 1990; Van et al 1993; Rupp et al 1994; Feng et al 1996). Although ERp72 shares the sequence (WCGHCK) responsible for redox activity with other PDI family proteins, the functional roles of the 3 thioreoxin homology domains of ERp72 have not been elucidated. In this study, we characterized the functional properties of the 3 individual thioredoxin homology domains of ERp72 by focusing on the reduction of disulfide bonds in insulin. Presumably, this reflected one of the steps involved in the oxidative folding process in which “nonnative” bonds are reduced by ERp72 before rearrangement.
Kinetic analysis of ERp72 mutants with Cys to Ser mutations at 2 cysteine residues in the individual thioredoxin homology domains revealed that triple mutants with substitutions in domains 1, 2, and 3 had no insulin reduction activity, indicating that at least one of these thioredoxin homology domains of ERp72 is involved in catalyzing insulin reduction.
Single mutation in domain 1 (intact domains 2 and 3) reduced kcat compared with that of wild type, but hardly altered Km values, indicating that the active site of domain 1 chiefly contributed to catalytic activity. Single mutation in domain 2 (intact domains 1 and 3) reduced kcat and increased Km, indicating that the domain 2 active site contributed to both binding affinity for substrate and catalytic activity. Single mutation in domain 3 (intact domains 1 and 2) doubled Km without altering kcat, indicating that the principal role of the domain 3 active site was apparently to enhance the recognition and binding of substrate in the steady state.
Preservation of intact domain 2 represented by double mutations in domains 1 plus 3 (mut-1+3) and of intact domain 1 represented by double mutations in domains 2 plus 3 (mut-2+3) had 50–60% activity of wild type but showed about 2-fold elevation in Km, indicating that the active sites of domains 1 and 2 chiefly contributed to catalytic activity but not binding affinity for substrate. Whereas, preservation of intact domain 3 by double mutations in domains 1 plus 2 (mut-1+2) had similar Km to that of wild type, indicating that domain 3 made the most significant contribution to binding affinity for substrates among domains 1, 2, and 3. These results of assays using single mutation and double mutations indicated in a consistent manner that the active site of domain 1 contributed to catalytic activity and the active site of domain 3 contributed to binding affinity for substrates. However, the results on domain 2 appeared inconsistent between single mutation study and double mutation study.
We interpret the results on domain 2 as follows. Compared with mut-1+2 (intact domain 3), mut-1 (intact domains 2 and 3) elevated Km, indicating that domain 2 was involved in binding affinity for substrate. However, compared with mut-2+3 (intact domain 1), mut-3 (intact domains 1 and 2) did not elevate Km, indicating that domain 2 did not contribute to binding affinity for substrate. Therefore, domain 2 appeared to be involved in binding affinity in a somewhat indirect manner, that is, only in combination with domain 3. Altogether, all 3 domains contributed to catalyzing insulin reduction, and individual domains acted synergistically rather than having simply additive effects.
The PDI has 5 μM of Km for insulin, and loss of C-terminal cysteines of PDI affects the apparent steady-state binding of substrate as evidenced by a 4-fold increase in Km (Gilbert 1998). Compared with this, elevation in Km caused by double mutations in domains 1 plus 3 or domains 2 plus 3 of ERp72 was relatively small (approximately 2-fold), which could be because of the innate low binding affinity for insulin (36.7 μM for wild-type ERp72). Therefore, it is natural to raise a question whether the differences in function among domains of ERp72 have biological significance. We claim that they do, on the basis of the following discussion. First, a decrease in kcat caused by double mutations in domains 1 plus 2 of ERp72 is comparable with that caused by loss of N-terminal cysteins of PDI (30–40% residual activity for N-terminal mutated PDI). Second, there is no information regarding the accurate concentration of ribonuclease (Rnase) or insulin in the ER, whereas the major folding enzymes and chaperones, PDI, BiP/GRP78, GRP94, and ERp72, are known to be present in millimolar concentrations in the ER (Hillson et al 1984; Lyles and Gilbert 1991). Therefore, we consider that if functional differences between thioredoxin domains of PDI have biological significance in disulfide bond formation and isomerizaton of RNase or insulin in vivo, functional differences among individual domains of ERp72 must also be biologically significant in vivo. We propose in this study that there may be a transfer of substrate to a low affinity site to allow catalysis and that binding of substrate to the higher affinity site may cause a decrease in the folding rate.
The manner of action of the thioredoxin homology domains of ERp72 was similar to that of the 2 thioredoxin homology domains of PDI (Lyles and Gilbert 1994). Both thioredoxin homology domains of PDI have been reported to contribute to some extent to Km and kcat. However, the method of contribution differs in that the C-terminal domain contributes more to Km and the N-terminal domain contributes more to kcat (Lyles and Gilbert 1994). Figure 4 shows the alignment of homologies between the PDI N- and C-terminal thioredoxin homology domains and the ERp72 thioredoxin homology domains 1, 2, and 3. There is complete conservation of the 11 amino acid– sequence EFYAPWCGHCK in the PDI and ERp72 thioredoxin homology domains, but the homologies of the amino acid sequence preceding or following the 11 amino acid–sequence are very different among the thioredoxin homology domains. ERp72 domains 1 and 2 are more homologous with the PDI N-terminal domain (48% and 51% identical, respectively) when compared with the homology between the ERp72 domain 3 and the PDI N-terminal domain (24% identical). In contrast, ERp72 domain 3 is more homologous with the PDI C-terminal domain (46% identical) when compared with the homologies of ERp72 domains 1 and 2 with the PDI N-terminal domain (26% and 23% identical, respectively). Domains 1 and 2 contributed more to catalyzing efficacy, and domain 3 contributed more to binding affinity. This assignment of function to individual domains is similar to that observed in PDI domains, which is consistent with the high homology between the ERp domains 1 and 2 and the PDI N-terminal domain, and that between the ERp72 domain 3 and the PDI N-terminal domain.
Given the general principle that essentially important sequences that exert fundamental biological functions are strictly conserved through evolution, the amino acid sequences flanking the active site cysteines, which are well conserved between ERp72 and PDI regardless of the very different sequences preceding and following this site, must represent essentially important roles, such as enzymatic catalysis of disulfide bond formation or isomerization and formation of secondary or tertiary domain structures. Further characterization of the thioredoxin homology domains of ERp72 requires future studies on oxidative folding and thiol-dependent isomerization of substrates such as reduced or scrambled RNase.
We thank Dr M. Green for providing ERp72 cDNA. This work was supported by Grants-in-Aid 10780452 and 15500250 for scientific research from the Japan Society for the Promotion of Science.