Ca2+-dependent redox modulation of SERCA 2b by ERp57

Abstract

We demonstrated previously that calreticulin (CRT) interacts with the lumenal COOH-terminal sequence of sarco endoplasmic reticulum (ER) calcium ATPase (SERCA) 2b to inhibit Ca2+ oscillations. Work from other laboratories demonstrated that CRT also interacts with the ER oxidoreductase, ER protein 57 (also known as ER-60, GRP58; ERp57) during folding of nascent glycoproteins. In this paper, we demonstrate that ERp57 overexpression reduces the frequency of Ca2+ oscillations enhanced by SERCA 2b. In contrast, overexpression of SERCA 2b mutants defective in cysteines located in intralumenal loop 4 (L4) increase Ca2+ oscillation frequency. In vitro, we demonstrate a Ca2+-dependent and -specific interaction between ERp57 and L4. Interestingly, ERp57 does not affect the activity of SERCA 2a or SERCA 2b mutants lacking the CRT binding site. Overexpression of CRT domains that disrupt the interaction of CRT with ERp57 behave as dominant negatives in the Ca2+ oscillation assay. Our results suggest that ERp57 modulates the redox state of ER facing thiols in SERCA 2b in a Ca2+-dependent manner, providing dynamic control of ER Ca2+ homeostasis.

Figure 1.

ERp57 reduces the frequency of Ca ²⁺ oscillations mediated by SERCA 2b. (A) Confocal images of Ca²⁺ oscillations in Xenopus oocytes expressing SERCA 2b alone (n = 22) or coexpressing ERp57 with SERCA 2b (n = 22). Traces represent changes in fluorescence (ΔF/F) as a function of time. Each plot represents two independent experiments with 11 oocytes per group. Individual Ca²⁺ wave images are presented at the indicated time. The white square in each and subsequent images represents a 5 × 5 pixel area used in determining ΔF/F. Histogram plots of the period between oscillations and the decay time (t_1/2) are for individual oscillations at peak activity, defined as the highest frequency for a 100-s window of time. Asterisks indicate statistical significance (P < 0.05, t test). (B) Western blots of SERCA 2b and ERp57. One oocyte equivalent was loaded per lane and proteins were resolved through 10% SDS-PAGE. SERCA 2b migrates ∼110 kD and ERp57 ∼60 kD as expected. The loading control lane in each and subsequent gels represents an invariant Xenopus protein that migrates at ∼42 kD. These gels represent three independent Western blots. Bar, 100 μm.

Figure 2.

L4 cysteine–deficient mutants of SERCA 2b exhibit higher frequencies of Ca ²⁺ oscillations. (A) Western blots of SERCA 2b and L4 mutants were loaded as follows: (lane 1) SERCA 2b, (lane 2) SERCA 2b-C1SC2S and (lane 3) SERCA 2b-C1AC2A. Two oocyte equivalents were loaded per lane and proteins were resolved through 12% SDS-PAGE. Gels represent four independent Western blots. (B) Confocal images of Ca²⁺ oscillations are shown in oocytes expressing SERCA 2b (n = 34) or the L4 cysteine–deficient mutants SERCA 2b-C1SC2AS (n = 40) and SERCA 2b-C1AC2A (n = 38). Traces represent two independent experiments with 15–20 oocytes per group. Histogram plots t_1/2 for (1) SERCA 2b, (2) SERCA 2b-C1SC2S, and (3) SERCA 2b-C1AC2A expressing oocytes. Asterisks indicate statistical significance (P < 0.05, t test) between SERCA 2b and each mutant group. Bar, 100 μm.

Figure 3.

Enzymatic activity of ERp57 and mutants measured in vitro . (A) ERp57 mutants lacking thioredoxin motifs. Two conserved WCGHCK motifs of ERp57 are thought to be the active sites for thiol-dependent oxidoreductase activity. Cysteines in each motif are indicated in black circles. Mutagenesis of relevant cysteines into serines is indicated by white circles. (B) Thiol-dependent catalytic activity of GST-PDI (positive control) and GST-ERp57 or mutant GST-ERp57 fusion proteins measured in vitro by the insulin turbidity assay at 300 μM [Ca²⁺]. GST alone and lack of enzyme input are used as negative controls in this assay. Triplicate absorbance_{OD 650} values were taken in three independent experiments and plotted as a function of time. Input amount of GST fusion proteins was 0.8 μM except for GST (22.04 μM). (C) Coomassie blue staining of GST and GST fusion proteins as labeled after one-step affinity purification. Proteins (5 μg each) were resolved through 12% SDS-PAGE.

Figure 4.

Enzymatic activity of ERp57 plays a critical role in modulating SERCA 2b. (A) Confocal images of Ca²⁺ oscillations shown in oocytes overexpressing proteins as labeled: SERCA 2b alone (n = 23); SERCA 2b + ERp57 (n = 24); SERCA 2b + ERp57-T1 (n = 20); SERCA 2b + ERp57-T2 (n = 23); and SERCA 2b + ERp57-T1T2 (n = 32). Traces represent two independent experiments with 11–15 oocytes per group. (B) Histograms of period and t_1/2 for the experiment in A. Asterisks indicate statistical significance (P < 0.05, t test) between SERCA 2b and SERCA 2b + ERp57 or SERCA 2b + ERp57-T2. (C) Western blots of SERCA 2b and ERp57 from lysates of experimental oocytes. One oocyte equivalent was loaded per lane and proteins were resolved through 12% SDS-PAGE. Gels represent three independent Western blots. (D) In vitro translations of ERp57 and mutants in rabbit reticulocytes in the absence or presence of canine pancreatic microsomes. Note that the signal peptide was proteolytically cleaved in reactions supplemented with microsomes. This gel is representative of four independent experiments. Bar, 100 μm.

Figure 5.

The interaction between ERp57 and L4 is Ca ²⁺ dependent and specific. (A) GST pull-down assays performed at the indicated [Ca²⁺] in micromolars. The gel is loaded as follows: input of in vitro–translated L4 (lane 1); pull downs for 30 μg GST control (lanes 2–5); pull downs for 10 μg GST-ERp57 (lanes 6–9). Proteins were resolved through 15% SDS-PAGE. L4 migrates ∼11 kD (109 aa). Data represent three independent experiments. Histogram depicts densitometric analysis from these experiments. L4 bands were normalized to the intensity of the band at 300 μM Ca²⁺. Asterisks indicate statistical significance (P < 0.05, one-way ANOVA). (B) The interaction between ERp57 and L4 is specific by GST pull-down assay. Lanes were loaded as follows: input of in vitro–translated L4 (lane 1); 30 μg GST negative control at the indicated [Ca²⁺] (lanes 2 and 3); 10 μg GST-ERp57 (lanes 4 and 5); 10 μg GST-PDI (lanes 6 and 7). Proteins were resolved through 15% SDS-PAGE. The gel represents three independent experiments. (Right gel) Coomassie blue stain of 3 μg purified GST (lane 1); 1 μg GST-ERp57 (lane 2), and 1 μg GST-PDI (lane 3). Proteins were resolved through 13% SDS-PAGE. GST migrates ∼27 kD, and GST-ERp57 and GST-PDI migrate ∼80 kD (mature rPDI is 489 aa, mature hERp57 is 481 aa). The arrowheads correspond to L4. (C) The L4 protein localizes to the ER. (Top) Coexpression in Xenopus oocytes of GFP-L4 (green) with DsRed-IP₃R (red). The IP₃R is used as a marker of ER localization. The overlay (yellow) demonstrates that the L4 localizes in the ER. (Bottom) Cytosolic GFP (green) used a negative control, is coexpressed with DsRed-IP₃R (red). Note that the cytosolic GFP expression pattern is more diffused than that of the GFP-L4 and that there is no overlap of the two proteins indicating lack of colocalization. High magnification of the small square regions are shown as insets in each figure. Bar, 5 μm.

Figure 6.

The catalytic activity of ERp57 and PDI is Ca ²⁺ independent. Enzymatic activity is measured using the insulin turbidity assay at the indicated Ca²⁺ concentrations. Absorbance_{OD 650} values were collected in triplicate and plotted. These plots represent four independent experiments. The standard error bars were smaller than the size of the symbols. Input amounts of GST-ERp57 and GST-PDI fusion proteins was 0.8 μM.

Figure 7.

ERp57 does not affect Ca^{2 +} oscillations mediated by SERCA 2a, nor SERCA 2b glycosylation motif mutants. (A) Confocal images of Ca²⁺ oscillations in oocytes overexpressing SERCA 2b (n = 31), SERCA 2b + ERp57 (n = 39), SERCA 2a (n = 25), and SERCA 2a + ERp57 (n = 30). The traces represent two independent experiments with 12–20 oocytes per group. (B) Images of Ca²⁺ oscillations in oocytes overexpressing SERCA 2b-N1036A (n = 34), SERCA 2b-N1036A + ERp57 (n = 15), SERCA 2b-S1038A (n = 37), and SERCA 2b-S1038A + ERp57 (n = 16). The traces represent two independent experiments with 8–19 oocytes per group. (C) Histograms plot period and t_1/2 for experiments in A and B, respectively. Asterisks indicate statistical significance (P < 0.05, t test) between SERCA 2b with SERCA 2b + ERp57 oocytes. (D) Western blots of SERCA 2a, SERCA 2b, or its mutants, and ERp57 from lysates of experimental oocytes as labeled. One oocyte equivalent was loaded per lane and proteins were resolved through 12% SDS-PAGE. Arrowheads indicate the protein as labeled at the top of each gel Bars: (A and B) 100 μm.

Figure 8.

Overexpression of L4, NC, N, or P domains of CRT increase Ca ²⁺ oscillation frequency. Representative confocal images of Ca²⁺ oscillations in oocytes as labeled: H₂O (n = 46); L4 (n = 31); CRT (n = 22); CRT-NP (n = 28); CRT-NC (n = 28); CRT-N domain (n = 45); and CRT-P domain (n = 49). Two independent experiments with 11–25 oocytes per group were performed. Histogram plots t_1/2 of individual waves normalized to the H₂O control group. Asterisks indicate statistical significance (P < 0.05, t test) between L4, CRT, CRT-NP, CRT-NC, CRT-N, CRT-P, and the control group. Bar, 100 μm.

Figure 9.

Dominant negative constructs increase the frequency of Ca ²⁺ oscillations enhanced by SERCA 2b, but not SERCA 2a. (A) Confocal images of Ca²⁺ oscillations in Xenopus oocytes overexpressing SERCA 2b (n = 22); or coexpressing SERCA 2b with L4 (n = 22); CRT (n= 22); CRT-NC (n = 21); CRT-N (n = 21); and CRT-P (n = 21) are shown. Two independent experiments with 10–11 oocytes per group were performed. (B) Histogram plots t_1/2 of individual waves. Asterisks indicate statistically significant difference (P < 0.05, t test) between L4, CRT, CRT-NC, CRT-N, and CRT-P with respect to the SERCA 2b (S2b) overexpression. (C) Histogram shows the result of a similar experiment performed for SERCA 2a. Confocal images of Ca²⁺ oscillations in oocytes overexpressing SERCA 2a (n = 21) or coexpressing SERCA 2a with L4 (n = 20); CRT (n = 20); CRT-N (n = 20); CRT-P (n = 20); and CRT-NC (n = 20) are not depicted. Note that on the same y-axis scale as in experiment in B, none of the tested constructs display a significant difference with respect to SERCA 2a (labeled S2a). Bar, 100 μm.

Figure 10.

Model depicting the functional consequences of the interaction between ERp57–CRT with SERCA 2b. At rest, the ER Ca²⁺ stores are full (high [Ca²⁺]_ER), the N domain of CRT interacts with the COOH-terminal sequence of SERCA 2b, whereas the P domain recruits ERp57 to target L4 in SERCA 2b. At high [Ca²⁺]_ER our data suggest that ERp57 promotes disulfide bond formation between thiol groups in the loop, reducing pump activity (i.e., decreasing the frequency of Ca²⁺ oscillations and t_1/2). This activity of ERp57 favors disulfide bond formation because the prevailing conditions in the ER are relatively oxidizing. As [Ca²⁺]_ER decreases below ∼50 μM, ERp57 dissociates from L4, allowing SERCA 2b in its reduced form (cysteine mutagenesis) to be more active. Thus, at low [Ca²⁺]_ER SERCA 2b favors ER store refilling. The arrow indicates the point of divergence between SERCA 2a and SERCA 2b. NFS is the glycosylation consensus motif at the COOH-terminal sequence of SERCA 2b.