Interaction of the Thiol-Dependent Reductase ERp57 with Nascent Glycoproteins

See allHide authors and affiliations

Science  03 Jan 1997:
Vol. 275, Issue 5296, pp. 86-88
DOI: 10.1126/science.275.5296.86


Calnexin and calreticulin interact specifically with newly synthesized glycoproteins in the endoplasmic reticulum (ER) and function as molecular chaperones. The carbohydrate-specific interactions between ER components and glycoproteins synthesized in isolated canine pancreatic microsomes were analyzed using a cross-linking approach. A carbohydrate-dependent interaction between newly synthesized glycoproteins, the thiol-dependent reductase ERp57, and either calnexin or calreticulin was identified. The interaction between ERp57 and the newly synthesized glycoproteins required trimming of the N-linked oligosaccharide side chain. Thus, it is likely that ERp57 functions as part of the glycoprotein-specific quality control machinery operating in the lumen of the ER.

The lumen of the ER contains a number of molecular chaperones that assist in the later stages of protein biosynthesis and folding (1, 2). A number of studies have highlighted specific interactions between newly synthesized glycoproteins and the putative chaperones calnexin and calreticulin (3, 4, 5, 6). The binding of calnexin and calreticulin to newly synthesized proteins is normally characterized by a specific requirement for correctly processed, asparagine-linked (N-linked), carbohydrate side chains. In combination with uridine 5′-diphosphate (UDP)-glucose:glycoprotein glucosyltransferase (7), calnexin and calreticulin are thought to mediate a quality control cycle for newly synthesized glycoproteins (2, 8, 9). The function of this cycle is to ensure that only correctly folded and assembled proteins exit the ER and gain access to later compartments of the secretory pathway (2, 10).

Here, we used model substrates derived from the secretory protein preprolactin (PPL) (11) to determine the effect of N-linked glycosylation on the interactions between newly synthesized polypeptides and ER proteins. When the PPL92.CHO transcript was translated in vitro in the presence of canine pancreatic microsomes, a glycosylated 62-amino acid prolactin fragment (PL62.CHO) was generated. In contrast, translation of the PPL92.Con transcript generated a nonglycosylated 62-amino acid fragment (PL62.Con). The interactions between both PL62.CHO and PL62.Con and ER components were analyzed with the use of the membrane-permeable cross-linking reagent succinimidyl 4-(N-maleimidomethyl) cyclohexane carboxylate (SMCC), a heterobifunctional reagent that principally cross-links lysines to cysteines. Two groups of cross-linking partners could be identified: (i) ER proteins that interacted with both glycosylated and nonglycosylated polypeptides, exemplified by protein disulfide isomerase (PDI) (Fig. 1A); and (ii) ER proteins that interacted only with glycosylated polypeptides. Calnexin and calreticulin, which interact principally with glycoproteins, were observed to cross-link to PL62.CHO (Fig. 1B) but not to PL62.Con (Fig. 1A). A weaker 160-kD product (Fig. 1B) presumably represented a ternary cross-linking adduct of calnexin, PL62.CHO, and a third unidentified component.

Fig. 1.

Glycosylation of imported polypeptides leads to specific interactions with ER proteins. PPL92.CHO and PPL92.Con mRNAs (11) were translated in a wheat germ lysate supplemented with canine pancreatic microsomes and [35S]methionine (20, 21). The microsomes were isolated and incubated with 1 mM SMCC (19, 22). (A) PL62.Con cross-linking products were immunoprecipitated before (native) or after (denaturing) denaturation with 1% SDS (6). Affinity-purified anti-ERp57 (23), a nonrelated control serum (NRS), and rabbit antisera (α) specific for prolactin (PL), calnexin, calreticulin, and PDI were used. (B) PL62.CHO cross-linking products were analyzed exactly as above. Samples were resolved by electrophoresis on 12% SDS-polyacrylamide gels and visualized with a Fujix BAS-2000 bioimaging system and software. The apparent molecular masses of calnexin and calreticulin are 88 and 60 kD, respectively. The prominent 97-kD cross-linking product immunoprecipitated by the anti-calnexin serum is a PL62.CHO-calnexin adduct. Stars denote the 71-kD calreticulin cross-linking product, the arrowheads show the CAP-60 product, and the solid circle is a minor 160-kD calnexin-containing adduct. In this and subsequent figures, molecular size markers (in kilodaltons) are at the sides of lanes.

In addition to antisera that recognized calnexin, calreticulin, and PDI, a number of antisera to other ER luminal proteins were screened for immunoprecipitation of glycosylation-dependent (that is, PL62.CHO-specific) cross-linking products (12). We were able to identify ERp57 (13), a thiol-dependent reductase (14) and putative cysteine protease (15), as a strong cross-linking partner of PL62.CHO but not of PL62.Con (compare Figs. 1A and 1B). This result suggested that, like calnexin and calreticulin, ERp57 interacts specifically with glycoproteins.

When immunoprecipitation was performed under “native” conditions (6), a 69-kD cross-linking product was coprecipitated with both calnexin and calreticulin (Fig. 1B). This 69-kD product was absent when samples were denatured with SDS before immunoprecipitation (Fig. 1B). The glycosylated PL62.CHO product had an apparent molecular mass of 9 kD (11), which implied that a 60-kD cross-linking partner was coprecipitated with the calnexin and calreticulin cross-linking products. A similar 60-kD calnexin-associated protein, denoted CAP-60, coprecipitates with adducts of the Glut-1 glucose transporter and calnexin (6).

The 69-kD cross-linking product obtained with PL62.CHO had a similar mobility to those obtained with both PDI and ERp57 (Fig. 1B). To establish whether CAP-60 was actually one of these components, we performed sequential immunoprecipitations. After native immunoprecipitation, samples were denatured with SDS and reprecipitated. The 97- and 160-kD calnexin-derived cross-linking products were reprecipitated with anti-calnexin serum, but CAP-60 was no longer observed and could not be reprecipitated with anti-PDI serum (Fig. 2A). Likewise, the 71-kD calreticulin cross-linking product was reprecipitated by the anti-calreticulin serum, but no products were recognized by anti-PDI serum (Fig. 2A). In a control experiment, anti-PDI serum worked well in sequential immunoprecipitations of PDI cross-linking products (Fig. 2A). A similar experiment showed that the 69-kD cross-linking product, which associated with both calnexin and calreticulin, was reprecipitated with antibodies specific for ERp57 (Fig. 2B) (16). Thus, CAP-60 was in fact ERp57 and interacted with glycoproteins in combination with calnexin and calreticulin.

Fig. 2.

Identification of CAP-60 by sequential immunoprecipitation using PPL92.CHO mRNA. (A) After immunoprecipitation with anti-calnexin, anti-calreticulin, or anti-PDI sera, samples were denatured with SDS and reprecipitated with antibodies specific for calnexin, calreticulin, PDI, or the nascent chain (PL). Control lanes show products obtained after the first round of immunoprecipitation. (B) After immunoprecipitation with anti-calnexin, anti-calreticulin, and anti-PDI sera, samples were denatured with SDS and reprecipitated with sera specific for calnexin, calreticulin, or PDI, or with affinity-purified anti-ERp57 (23). Control lanes show products obtained after the first round of immunoprecipitation. The symbols in the lanes have the same meanings as in Fig. 1.

To investigate further the interaction between ERp57 and glycoproteins, we analyzed two authentic glycosylated secretory proteins (17) by cross-linking. Both yeast pro-α factor and human interferon-γ (IFN-γ) were found to be cross-linked to ERp57 and calreticulin (Fig. 3, B and C). Thus, ERp57 could be cross-linked to a variety of glycoproteins.

Fig. 3.

The interaction of ERp57 with glycoproteins requires trimming of the oligosaccharide side chains. mRNAs encoding PPL92.CHO, S. cerevisiae prepro-α factor (PPα), and human IFN-γ (17) were translated in the presence of canine pancreatic microsomes, and imported polypeptides were cross-linked to interacting components with SMCC (22). The glucosidase inhibitor castanospermine (CST) was included at 1 mM during the translation reaction as indicated (-, no CST added). (A to C) Cross-linking products were denatured with 1% SDS and then analyzed by immunoprecipitation with anti-calnexin, anti-calreticulin, or anti-PDI sera or affinity-purified anti-ERp57 (23). (D) Castanospermine treatment inhibits glucose trimming, resulting in a reduced mobility for each glycoprotein examined. The major glycosylated forms of each polypeptide after trimming are indicated by stars; the number of stars indicates the number of N-linked carbohydrate side chains present (24).

Calnexin and calreticulin interact preferentially with glycoprotein substrates that have been enzymatically processed to bear the monoglucosylated form of the N-linked carbohydrate side chain (2). To determine the role of glucose trimming in promoting the interaction of glycoproteins with ERp57, we used castanospermine to specifically inhibit the glucosidases responsible for this processing (9). The efficacy of the castanospermine treatment was established by comparing the mobility of the glycoproteins with and without the treatment (5). In each case, castanospermine caused a reduction in the mobility of the glycosylated proteins due to the increase in the number of glucose residues present on the carbohydrate side chain (Fig. 3D). The interactions of calnexin (Fig. 3A), calreticulin (Fig. 3, A to C), and ERp57 (Fig. 3, A to C) were all substantially inhibited by castanospermine treatment. In contrast, cross-linking to PDI was either unaffected (Fig. 3, A and B) or only partially inhibited (50%) (Fig. 3C).

Thus, like calnexin and calreticulin, ERp57 binding required glucose trimming of the N-linked carbohydrate side chains. We propose that ERp57 functions in combination with calnexin and calreticulin as a molecular chaperone of glycoprotein biosynthesis. We observed a time-dependent decrease in the amount of the PL62.CHO cross-linking products with calnexin, calreticulin, and the associated ERp57 (18). This suggested that the interaction between ERp57 and nascent glycoproteins was transient, like other molecular chaperone-substrate interactions (3, 5, 19). We believe a specific modulation of glycoprotein folding could be achieved by coupling the lectin-like properties of calnexin and calreticulin (2) with the thiol-dependent reductase activity of ERp57 (14).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
View Abstract

Stay Connected to Science

Navigate This Article