Lectin-deficient calreticulin retains full functionality as a chaperone for class I histocompatibility molecules

Abstract

Calreticulin is a molecular chaperone of the endoplasmic reticulum that uses both a lectin site specific for Glc(1)Man(5-9)GlcNAc(2) oligosaccharides and a polypeptide binding site to interact with nascent glycoproteins. The latter mode of substrate recognition is controversial. To examine the relevance of polypeptide binding to protein folding in living cells, we prepared lectin-deficient mutants of calreticulin and examined their abilities to support the assembly and quality control of mouse class I histocompatibility molecules. In cells lacking calreticulin, class I molecules exhibit inefficient loading of peptide ligands, reduced cell surface expression and aberrantly rapid export from the endoplasmic reticulum. Remarkably, expression of calreticulin mutants that are completely devoid of lectin function fully complemented all of the class I biosynthetic defects. We conclude that calreticulin can use nonlectin-based modes of substrate interaction to effect its chaperone and quality control functions on class I molecules in living cells. Furthermore, pulse-chase coimmunoisolation experiments revealed that lectin-deficient calreticulin bound to a similar spectrum of client proteins as wild-type calreticulin and dissociated with similar kinetics, suggesting that lectin-independent interactions are commonplace in cells and that they seem to be regulated during client protein maturation.

Figure 1.

Wild-type and mutant Crt are expressed in Crt-deficient cells and associate with newly synthesized proteins. (A) Crt-deficient murine fibroblasts (K42 cells) were infected with virus encoding the indicated wild-type (Wt) and oligosaccharide-binding deficient Crt proteins (Y128A and D317A). In addition, cells designated K42 Crt^−/− and K41 Crt^+/+ were infected with virus packaged with an empty expression vector. Cells (10⁶) were lysed in NP-40 lysis buffer, proteins were separated by SDS-PAGE (10% gel), and Crt was detected by immunoblotting. Actin was also immunoblotted to serve as a gel loading control. (B) Cells transfected with various Crt constructs (or empty vector) were radiolabeled for 10 min with [³⁵S]Met, lysed in digitonin lysis buffer, and subjected to immunoisolation with anti-Crt antiserum. Crt-associated proteins were separated by SDS-PAGE and visualized by fluorography. (C) The experiment in B was repeated except that before immunoisolation, the lysates were incubated on ice for 45 min. Exposures, 30 h (B) and 72 h (C).

Figure 2.

Interaction kinetics of wild-type and mutant Crt with newly synthesized proteins. K41 and K42 cells expressing the indicated Crt proteins were radiolabeled with [³⁵S]Met for 10 min, and then they were chased in medium containing unlabeled Met for the times shown. At each time point, cells were lysed in digitonin lysis buffer, and then Crt and associated proteins were recovered by immunoisolation with anti-Crt antiserum. Radiolabeled proteins were separated by SDS-PAGE and visualized by fluorography. The results depicted in the top two panels were obtained from a single experiment and those in the bottom two panels were obtained in another experiment.

Figure 3.

Wild-type and lectin-deficient Crt normalize surface expression of MHC class I molecules in Crt^−/− K42 cells. (A) K42 cells stably expressing wild-type Crt or lectin-deficient mutants were incubated with mAbs Y-3 or B22.249R1 to detect cell surface H-2K^b and H-2D^b molecules, respectively. Cells were then incubated with phycoerythrin-conjugated goat anti-mouse IgG and analyzed by flow cytometry. K41 and K42 cells transfected with empty vector were also subjected to this analysis. As a negative control, cells were stained with secondary antibody alone. (B) Mean fluorescence values are plotted as percentages relative to Crt^+/+ K41 cells. Error bars represent the SE values from three independent experiments.

Figure 4.

Peptide loading of class I molecules is promoted by wild-type and lectin-deficient Crt. Stably transfected K41 and K42 cells expressing the indicated Crt proteins were incubated with the K^b binding peptide SIINFEKL and the D^b binding peptide NP366-374 overnight and subsequently evaluated for cell surface class I expression by flow cytometry by using mAbs Y3 and B22-249.R1. Results are shown as the -fold increase in expression (mean fluorescence value) relative to cells incubated without peptide. Error bars represent the SE values from three independent experiments.

Figure 5.

Wild-type and lectin-deficient Crt enhance the presentation of a specific ovalbumin peptide by H-2K^b molecules. K41 cells and K42 cells expressing the indicated Crt proteins were transiently transfected with a plasmid encoding GFP fused to ubiquitin followed by the ovalbumin-derived sequence SIINFEKL. After 24 h, cells were analyzed by flow cytometry by using mAb 25D1.16, which recognizes the SIINFEKL peptide in complex with H-2K^b molecules, followed by Alexa Fluor 642-conjugated anti-mouse Ab. Samples were gated against low, medium, and high GFP fluorescence (GFP gates 1–4) and plotted against the corresponding mean 25D1.16 fluorescence. Two separate experiments are shown and include controls lacking plasmid or in which an isotype-matched primary antibody replaced mAb 25D1.16.

Figure 6.

Association of lectin-deficient Crt with members of the peptide loading complex. (A) The indicated cell lines were lysed in digitonin lysis buffer and then subjected to immunoisolation with either anti-tapasin antiserum (lanes 1) or preimmune serum (lanes 2). Immune complexes adsorbed on protein A beads were dissociated with 100 μM of the C-terminal tapasin peptide used to raise the antiserum. Eluted proteins were separated by SDS-PAGE and immunoblotted with antibodies directed against Crt, ERp57, and the TAP peptide transporter. Note that the high background in the Crt panel is due to residual Ig heavy chain from the anti-tapasin antiserum (50 kDa), which migrates just below Crt (60 kDa). (B) The indicated cells were radiolabeled with [³⁵S]Met for 30 min, incubated on ice for 10 min with 20 mM N-ethylmaleimide in PBS, lysed in digitonin lysis buffer, and then subjected to immunoisolation with anti-tapasin antiserum. Isolated proteins were resolved on a nonreducing 10–15% gradient SDS-PAGE gel.

Figure 7.

Both wild-type and lectin-deficient Crt retard ER-to-Golgi transport of assembling class I molecules. (A) The indicated cell lines were radiolabeled for 10 min with [³⁵S]Met and chased with unlabeled Met for various times. Cells were then lysed and H-2K^b and D^b molecules were immunoisolated sequentially first with anti-8 antiserum followed by a combination of mAbs 28-14-8S and B22-249.R1. Isolated proteins were digested with endo H before analysis by SDS-PAGE. The mobilities of endo H-sensitive (s) and -resistant (r) heavy chains are indicated. Note that in Crt^−/− K42 cells two bands are designated as endo H resistant for D^b. These appear variably between experiments; the slower band corresponds to D^b molecules with all three oligosaccharides processed to complex forms whereas the faster band corresponds to molecules that possess 1 immature and 2 complex oligosaccharides. Both represent mature, Golgi-processed molecules and they are combined when calculating the percentage of endo H-resistant heavy chains. Asterisks denote nonspecific bands. (B) Fluorograms in A were scanned and endo H-sensitive and endo H-resistant band intensities were quantified using NIH ImageJ software (http://rsb.info.nih.gov/ij/). The endo H-resistant heavy chain was then calculated as a percentage of the total heavy chain signal at each time point. Error bars represent the SE values of three independent experiments for K^b and two independent experiments for D^b.