Introduction

Calreticulin (CRT) is an endoplasmic reticulum (ER) luminal protein that plays key roles in the assembly and folding of N-linked glycoproteins. CRT has a highly acidic C-terminal domain that contains multiple low affinity calcium binding sites (Baksh and Michalak, 1991, Wijeyesakere et al., 2011). As an ER chaperone, CRT is thought to play a critical role in maintaining cellular calcium homeostasis (Venkatesan et al., 2021b, Michalak, 2024). The ER has a relatively high free calcium concentration of 0.5-1 μM through the presence of various low affinity calcium-binding proteins that buffer calcium, whereas cytosolic calcium levels are maintained within a narrow range of around 100 nM (Michalak, 2024). The maintenance of calcium homeostasis within cellular compartments is essential for cell function and survival (Bagur and Hajnoczky, 2017). Early studies have shown that the C-domain of CRT contributes to ER calcium storage, although free ER luminal concentrations are unchanged (Nakamura et al., 2001).

Somatic mutations of exon 9 of the CALR gene including a 52-base pair deletion (Del52; type 1 mutation) and a 5-base pair insertion (Ins5; type 2 mutation), are driver mutations in some myeloproliferative neoplasms (MPN), specifically in essential thrombocythemia (ET) and primary and secondary myelofibrosis (PMF) (Klampfl et al., 2013, Nangalia et al., 2013). The MPN mutations change the CRT C-domain residues from being predominantly acidic to a high prevalence of basic amino acids and result in a loss of the ER-retention KDEL sequence (Klampfl et al., 2013, Nangalia et al., 2013). Although all MPN CRT mutants change the sequence and charge of the C-domain, a greater number of acidic residues are retained in MPN type 2 mutant (CRT_Ins5), compared to the type I mutant (CRT_Del52). The loss of more acidic residues from the C-terminal domain of CRT_Del52 has been reported to impair its calcium binding and reduce ER calcium levels (Ibarra et al., 2022). Store-operated calcium entry (SOCE) is a physiological process that allows for the entry of extracellular calcium and ER calcium restoration following ER calcium depletion (Putney, 2011, Zhang et al., 2020). Altered calcium release from the ER and SOCE was reported for MPN type I but not type 2 mutants in cultured megakaryocytes from MPN patients (Pietra et al., 2016). In another study, spontaneous calcium spikes and increased SOCE activation were observed in cultured megakaryocytes from MPN type I and type 2 patients when compared to healthy donors (Di Buduo et al., 2020).

Although global calcium binding defects were predicted or qualitatively measured for CRT_Del52 (Shivarov et al., 2014, Ibarra et al., 2022) and altered cellular calcium signaling is also reported (Pietra et al., 2016, Di Buduo et al., 2020), it remains unclear whether MPN-linked CRT_Del52 has any low affinity calcium binding capacity, given that a few conserved acidic residues are retained in its C-domain (Figure 1). In fact, our previous studies showed that a truncated version of murine wild-type CRT (mCRT) containing the non-mutated part of CRT_Del52 and two additional C-terminal residues (mCRT_1-351) showed similar low affinity calcium binding characteristics as the wild-type mCRT (Wijeyesakere et al., 2011). On the other hand, further truncation of mCRT (mCRT_1-339) was needed to disrupt calcium binding (Wijeyesakere et al., 2011, Wijeyesakere et al., 2016). The goal of the current study was to quantitatively measure low affinity calcium binding by CRT_Del52 compared with wild-type human CRT and also quantify ER and cytosolic calcium levels and calcium responsiveness in cells expressing wild-type CRT or CRT_Del52.

Figure 1:

Material and methods

Protein expression and purification

B1 domain of streptococcal protein G (GB1) and his-tagged versions of wild-type CRT, CRT_{(1- 351)}, CRT_(1-339) and CRT_Del52 were purified following expression in E. coli cells using a nickel resin and gel filtration chromatography as described previously (DelProposto et al., 2009, Del Cid et al., 2010, Wijeyesakere et al., 2011) (also see supplemental methods). The monomer fractions were pooled, dialyzed and concentrated to 4 mg/mL in buffer containing 20 mM HEPES, 10 mM NaCl 1 mM CaCl₂ pH 7.5 using Amicon ultra concentrators with a molecular weight cut-off of 30,000 Dalton (Millipore) and frozen in 10% glycerol for further use.

Isothermal titration calorimetry

Isothermal titration calorimetry (ITC) measurements were done using the Nano-ITC SV (TA Instruments). Purified CRT monomers were dialyzed into 20 mM HEPES, 10 mM NaCl pH 7.5, brought to a concentration of 100 µM, and then pre-incubated with 50-100 µM CaCl₂ in 20 mM HEPES, 10 mM NaCl pH 7.5 for 30 minutes to block the high affinity calcium binding sites. This was followed by 25 sequential 10µL injections of 20 mM HEPES, 10 mM NaCl, 10 mM CaCl₂ pH 7.5 into the protein solutions maintained at 37°C. The heat changes measured for the sequential injections were curve-fitted to calculate the enthalpies, stoichiometries, and the dissociation constants using NanoAnalyze software.

Cytosolic calcium measurement and store-operated calcium entry (SOCE) induction

MEG-01 cells and their CALR-modified versions (see supplemental methods) were plated on glass coverslips pretreated with 0.01% poly-L-lysine to facilitate cell attachment. Cells were loaded with 2.5 µM Fura-2/AM dye (Ion Biosciences) supplemented with 0.1% pluronic acid in a physiological salt solution (PSS) containing 140 mM NaCl, 2 mM CaCl₂, 5 mM KCl, 2 mM MgCl₂, 10 mM HEPES and 11 mM glucose, and incubated for 60 minutes at 37°C. Live single-cell imaging was done at 20X magnification using an Olympus IX71 inverted microscope equipped with a TILL Polychrome V monochromator light source, a Dual-View emission beam splitter, and a cooled Photometrics Quant-EM camera. Cells were transferred to a perfusion chamber that was mounted on the microscope stage, and PSS with or without 2 mM CaCl₂ was flowed over cells using a gravity-fed perfusion system. Cytosolic calcium measurements were performed by perfusing Fura-2 loaded cells in the chamber with 2 mM Ca²⁺, followed by the addition of 25 µM cyclopiazonic acid (CPA) (Cayman) into the imaging buffer. To determine cytosolic calcium levels after activating SOCE, cells were initially perfused with Ca²⁺-free PSS containing 0.1 mM EGTA, followed by the addition of 25 µM CPA in Ca²⁺-free PSS and subsequent addition of 2 mM Ca²⁺ and 25 µM CPA in PSS (without EGTA) to reestablish calcium influx. The emission signals corresponding to the 340 and 380 nm excitation were recorded, and the signals resulting from 340/380 nm excitation were calculated using the Meta Fluor software. The basal cytosolic calcium was defined as the 340/380 nm signal ratio before CPA treatment (measured over a 4-or 5-minute interval immediately before CPA addition). The fractional increase in cytosolic calcium levels after ER Ca²⁺ depletion with CPA was calculated as the difference between peak signal (single time point) and baseline (averaged over a 4-or 5-minute interval before CPA addition), as a ratio relative to the baseline. The fractional increase in cytosolic calcium after addition of 2 mM extracellular Ca²⁺ and CPA was calculated as the difference between the peak signal following Ca²⁺ and CPA addition (averaged over a 1-minute time interval following the peak signal) and the baseline (the signal averaged over a 0.8-1-minute interval immediately before the Ca²⁺ and CPA addition) relative to the baseline. Total calcium release after ER Ca²⁺ depletion was reported as the total area under the curve (AUC). Fractional increases were calculated as (Peak-Baseline)/Baseline x 100. This was calculated as peak signals after CPA addition relative to the baseline before CPA (measured over a 4-or 5-minute interval immediately before CPA addition).

ER calcium measurements in HEK293T cells

CALR-modified HEK cells were prepared as described in supplemental methods. Cells were transiently transfected with pCIS-GEM-CEPIA1er (Suzuki et al., 2014); Addgene plasmid # 58217) using Fugene HD reagent (Promega). The transfected cells were harvested after 72 hours. Cells were washed twice with FACS buffer (PBS + 2% FBS) supplemented with 2 mM Ca²⁺ and resuspended in the same FACS buffer. The live-dead stain 7-AAD (BD) was added at 1:200 dilution and the cells were analyzed by flow cytometry using the BD LSRFortessa cell analyzer. The signal for the Ca²⁺-bound GEM-CEPIA1er probe was recorded in the Pacific Blue channel (excitation at 405 nm and emission at 452-455 nm) and the signal for the unbound probe was recorded in the AmCyan channel (excitation at 405 and emission at 498 nm). The analysis of flow cytometry data was done in the FlowJo software (v10.10.00). The cells were gated based on the forward (FSC-A) and side scatter (SSC-A), followed by gating on the 7AAD-negative (live) cells, and subsequent gating on the cells positive for the expression of GEM-CEPIA1er probe measured in the Pacific Blue and AmCyan channels. Basal signals were initially recorded, followed by the addition of 5 µM thapsigargin (Tg) and subsequent recording of signals for the treated and untreated cells at four different time points between 2 – 10 minutes. The ratio of mean fluorescence intensity (MFI) of the bound to the unbound signal was used for calculations.

RNA-sequencing

The parental and CRT-KO MEG-01 cells were grown in complete media for 12-18 hours before harvesting. Total RNA was isolated using the RNeasy Plus Mini Kit (Qiagen). This was followed by poly-A selection and enrichment of mRNA using NEB Next Ultra Express RNA kit and a NEB Poly A kit (New England Biolabs). Sequencing was done on the NovaSeq at PE150, targeting 30-40 million reads/sample after cDNA library preparation. Data were pre-filtered to remove genes with 0 counts in all samples. Differential gene expression analysis was performed using DESeq2. The differentially expressed genes were selected as those with a log2 fold-change of <-0.49 or >0.49 and p values <0.05. The results from RNA sequencing were further analyzed in the iPathwayGuide to identify significantly altered proteins related to calcium signaling. The volcano plot was plotted in GraphPad Prism as Log2 fold-change (x-axis) vs. –Log10 adjusted p-value.

Real-time polymerase chain reaction

Total RNA was isolated using the RNeasy Plus Mini Kit (Qiagen) followed by synthesis of complementary DNA using the Superscript III First-Strand Synthesis Kit (ThermoFisher Scientific) following the manufacturer’s protocol. Quantitative RT-PCR was performed on the ABI 7500 FAST Real Time PCR machine (ThermoFisher Scientific, USA) using the SYBR Green PCR Master mix (Applied Biosystems, UK) and gene-specific primers (Supplementary Table 2). Measurement was done in technical duplicates. The expression of target genes was quantified using the 2^−ΔΔCt method and normalized to the expression of either one or multiple endogenous control genes including beta-actin (ACTB), glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and hypoxanthine phosphoribosyltransferase 1 (HPRT1). The expression of the target gene in MEG-01 CRT-KO and wild type and CRT_Del52 reconstitution cells was further normalized to the mean expression of parental MEG-01 or CRT-KO-vec. The calculated relative mRNA transcript expression values (RQ) were plotted using GraphPad Prism.

Results

MPN-linked CRT_Del52 retained at least some low affinity calcium binding sites

Previous studies (Wijeyesakere et al., 2011) have implicated residues 340 to 351 of murine CRT, which are identical in wild-type human CRT and fully conserved with the 340 to 349 residues of CRT_Del52, to be involved in low affinity calcium binding. Further truncation of the C-domain to the residue 339 (mCRT_1-339) led to the loss of calcium binding measured by ITC (Wijeyesakere et al., 2011). Figure 1 shows a sequence alignment of murine CRT, human CRT, CRT_Ins5, and CRT_Del52 proteins and indicates that the residues previously implicated in low affinity calcium binding are also shared between wild-type CRT and CRT_Del52. Recombinant his and GB1 tagged purified proteins were used to compare calcium binding by human CRT and its variants. Chromatograms for human wild-type CRT, CRT_Del52, CRT_(1-339) and CRT_(1-351) are shown in Figure 2A. SDS-PAGE gels corresponding to the purified proteins are shown in Figure 2B. CRT_Ins5 was not purified in sufficient quantities needed for ITC experiments and was not further investigated in this study.

Figure 2:

For the ITC measurements, calcium titrations were performed after pre-blocking the high affinity calcium binding sites with 50-100 µM CaCl₂. Similar binding profiles and properties were observed with wild-type CRT, CRT_Del52, and the CRT_(1-351) mutant (Figures 3A, C and D and Table 1). However, consistent with previous studies with mCRT_(1–339), endothermic calcium binding by human CRT_(1–339) was of lower magnitude compared to other constructs, with a K_D value that was too low to be reliably fit, suggesting the loss of low affinity calcium binding sites in this mutant (Figure 3B and Table 1). There are exothermic peaks observed with human CRT_(1–339) at the beginning of the calcium titration, which may be attributed to altered low affinity sites retained within the mutant or reduced affinity of the high affinity sites. We concluded that CRT_Del52 retains low affinity calcium binding that is measurable by ITC.

Figure 3:

Table 1:

Similar ER calcium levels in CRT-KO HEK cells reconstituted with wild-type CRT or CRT_Del52

While the CRT_Del52 mutant retains some ability to bind calcium (Figure 3 and Table 1), the loss of the ER-retention KDEL sequence induced by the frameshift mutation, which results in the secretion of the protein from the ER (Araki et al., 2016, Chachoua et al., 2016, Han et al., 2016, Arshad and Cresswell, 2018, Liu et al., 2020, Masubuchi et al., 2020, Venkatesan et al., 2021a, Pecquet et al., 2023) might be expected to cause ER calcium insufficiency in cells expressing CRT_Del52 compared to wild-type CRT. We investigated ER calcium signals in CRT knock-out HEK293T cells generated by CRISPR/Cas9 gene editing. A clonal KO line was generated and subsequently reconstituted with constructs encoding wild-type CRT, CRT_Del52, or CRT_Del52 with a KDEL sequence added at the C-terminus (CRT_Del52-KDEL). Protein expression was assessed by immunoblotting with anti-CRT(N), which detects an N-domain epitope present in all CRT constructs, and anti-CRT(C-mut), which was specifically raised against the mutant CRT C-terminus (Venkatesan et al., 2021a). While the recognition with anti-CRT (N) was very weak in lysates of cells expressing CRT_Del52, which is secreted, CRT_Del52-KDEL is readily detectable with both antibodies. Both wild-type CRT and CRT_Del52-KDEL were reconstituted to higher levels than in the parental cells (Figure 4A, lanes 1 and 2 compared to 7-8 and 11-12). The KO and reconstituted cells were further transfected with pCIS-GEM-CEPIA1er, which encodes a ratiometric ER calcium probe (Suzuki et al., 2014). Emission signals for calcium-bound (452-455 nm) and unbound (498 nm) probe were measured using flow cytometry in the Pacific Blue and AmCyan channels, respectively, following excitation with the 405 nm (violet) laser. Flow cytometric analysis was undertaken in a buffer containing 2 mM extracellular Ca²⁺ (Figures 4B and 4C). There was a significant decrease in basal free ER calcium, measured by the bound/unbound ratios in CRT-KO cells compared to the parental cells, (Figure 4D, left panel). ER calcium levels were partially restored, although non-significantly, when the CRT-KO cells were reconstituted with wild-type CRT. CRT-KO cells reconstituted with CRT_Del52 and CRT_{Del52- KDEL} showed small but non-significant decreases in ER calcium levels compared to those reconstituted with wild-type CRT (Figure 4D, right panel). Changes to ER calcium levels in response to thapsigargin (Tg), a SERCA inhibitor, were similar in the parental vs. CRT-KO cells, and in the CRT-KO cells compared with those reconstituted with different CRT proteins (Figure 4E).

Figure 4:

Similar cytosolic calcium and SOCE signals in a CRT-KO megakaryocyte cell line reconstituted with wild-type CRT or CRT_Del52

MPN CRT mutations cause thrombopoietin-independent differentiation of megakaryocytes to platelets via the thrombopoietin receptor (MPL) (Araki et al., 2016, Chachoua et al., 2016, Marty et al., 2016, Li et al., 2018, Masubuchi et al., 2020). We, therefore, used a megakaryoblast cell line, MEG-01, to assess the role of CRT deficiency and the CRT_Del52 mutant on basal cytosolic calcium levels and SERCA inhibition-induced calcium influx into the cytosol. As with the HEK cells (Figure 4), we created a stable CRT knockout clone of MEG-01 cells by CRISPR/Cas9 editing for cytosolic calcium measurements. The cells were subsequently reconstituted with wild-type CRT or CRT_Del52. Representative immunoblots are shown to demonstrate the successful knockout and reconstitution (Figure 5A). Wild-type CRT reconstitution in MEG-01 CRT-KO cells resulted in expression levels higher than those in the parental cells (Figure 5A, lanes 7 and 8 compared with lanes 1 and 2). In the presence of 2 mM Ca²⁺ in the media, we observed no significant differences in the basal cytosolic calcium levels or CPA-induced calcium release from ER in CRT-KO cells compared to the parental MEG-01 cells (Figure 5C and 5D). Additionally, no significant differences in basal cytosolic calcium were observed in CRT-KO cells compared to those reconstituted with wild-type CRT or CRT_Del52 (Figure 5E). Furthermore, cytosolic calcium levels in CRT-KO and the two reconstituted cell lines were not statistically different following CPA treatments.

Figure 5:

SOCE is an important process for restoring ER calcium and maintaining intracellular calcium levels. In contrast to cytosolic calcium signals in the presence of 2 mM Ca²⁺ in the media, in calcium-free saline, CRT-KO cells exhibited a small but significant increase in basal cytosolic calcium levels relative to the parental cells (Figure 6A, left panel and 6B). Upon ER calcium depletion with CPA, the mean fractional increase in the cytosolic calcium measured in the Ca²⁺-free saline did not differ significantly in parental cells vs. CRT-KO cells. The areas under the curve were also not different for the two cell lines (Figures 6A and 6C). However, consistent with the higher basal cytosolic calcium measured in CRT-KO cells in calcium-free saline, the measured SOCE signal (following the addition of 2 mM Ca²⁺ and CPA) was significantly reduced in CRT-KO cells compared to parental cells (Figure 6D). Reconstitution of the CRT-KO cells with either wild-type CRT or CRT_Del52 tended to reduce basal cytosolic calcium and induce SOCE (Figure 6A, right panel, 6E and 6G), but both trends were non-significant compared to CRT-KO cells. Importantly, there were no significant differences between the wild-type CRT and CRT_Del52 reconstituted cells, in either their basal or CPA-induced cytosolic calcium signals or in SOCE (Figures 6E-G).

Figure 6:

CRT deficiency induced numerous transcriptional changes to genes within the calcium signaling pathway

As noted in Figures 4 and 6, some of the measured differences between CRT-KO and parental cells are not fully reversed by reconstitution with wild-type CRT, although the trends are maintained. We predicted that there may be other indirect effects of CRT-KO and reconstitution that affect cellular calcium levels and calcium signaling that could explain some of these measured differences. As proof of concept, we undertook bulk RNA-sequencing to determine differential gene expression in MEG-01 parental vs. CRT-KO cells. There were 3078 upregulated and 1790 downregulated genes. The volcano plot and heat map show identified genes that were differentially expressed in CRT-KO MEG-01 compared to parental cells (Figure 7A). Based on pathway analysis, numerous genes involved in cellular calcium signaling were identified. Of specific note is the upregulation of GRP94 encoded by the HSP90B1 gene, BiP encoded by HSPA5 gene, and PDIA6 encoded by PDIA6 gene, which are known to have low affinity calcium binding sites (Lievremont et al., 1997, Macer and Koch, 1988, Biswas et al., 2007, Marzec et al., 2012, Okumura et al., 2021) (Figures 7A and 7B). Other studies have reported similar upregulation of these genes in CRT-KO (Tang et al., 2025) as well as in heterozygous CRT_Del52 knock-in cells (Figure 7C) (Fosselteder et al., 2023). Upregulation of these proteins may serve a compensatory role in buffering calcium in CRT-KO cells. Additionally, PDIA3, which is shown to regulate the activities of stromal interaction molecule 1 (STIM1) (Prins et al., 2011) and Sarco/endoplasmic reticulum Ca²⁺ ATPase (SERCA2b) (Li and Camacho, 2004), was consistently upregulated in all the studies.

Figure 7:

We further used real-time PCR with target gene primers (Supplementary Table 2) to validate the expression of the key ER resident proteins identified from the RNA-sequencing data which have roles in calcium binding and regulation of ER calcium homeostasis. Consistent with the RNA-seq data, the expression of HSPA5, HSP90B1, PDIA6 and PDIA3 mRNA transcripts was significantly increased in CRT-KO MEG-01 compared to parental cells. Reconstitution of CRT-KO cells with wild-type CRT but not CRT_Del52 showed a significant decrease in the expression of HSPA5, HSP90B1, PDIA6 and PDIA3 (Figures 7D-G). Protein levels of BiP, GRP94, PDIA6 and PDIA3 were also assessed in the indicated cells by immunoblots using specific antibodies (supplementary methods). Representative immunoblots show increased expression of BiP (Figure 7D, right panel), GRP94 (Figure 7E, right panel), PDIA6 (Figure 7F, right panel) and PDIA3 (Figure 7G, right panel) in CRT-KO cells compared to parental MEG-01 cells as well as increased expression in CRT-KO-Vec cells compared to CRT-KO cells reconstituted with wild type CRT. The effects of CRT_Del52 in reducing expression were generally less robust than wild type CRT.

Discussion

Calreticulin is a ubiquitously expressed ER luminal calcium-binding chaperone, with a role in ER calcium storage (Venkatesan et al., 2021b, Michalak, 2024). Previous studies using dye-based assays reported impaired calcium binding by the MPN-associated CRT_Del52 mutant (Ibarra et al., 2022). Yet direct and quantitative analyses of its calcium-binding properties have been lacking. In this study, we investigated low affinity calcium binding by wild-type human CRT, CRT_Del52, and the truncation mutants CRT_(1-351) and CRT_(1-339) using ITC (Figure 3 and Table 1). Our studies show that wild-type CRT, CRT_Del52, and CRT_(1-351) retain low affinity calcium binding sites, exhibiting average K_D values of 574.24, 938.82, and 444.34 µM, respectively. On the other hand, low affinity calcium binding by CRT_(1-339) could not be reliably fit. This is the first quantitative characterization of the low affinity calcium binding profiles of wild-type human CRT, CRT_Del52, and functionally informative truncation mutants.

Although CRT_Del52 retains at least partial low affinity calcium binding capability, loss of its KDEL retention sequence (Figure 1) could still disrupt ER calcium storage and signaling. Interestingly, the cellular assays reveal that ER and cytosolic calcium levels, as well as SOCE, are similar between CRT-KO cells and those reconstituted with either wild-type CRT or CRT_Del52 (Figures 4-6). One study has reported decreased ER calcium levels in a human osteosarcoma cell line expressing CRT_Del52 compared to those expressing wild-type CRT, following transient transfections (Ibarra et al., 2022). The same study suggested the contributions of the P and C-domain of wild-type in restoring ER calcium levels in cells expressing CRT_Del52 (Ibarra et al., 2022). While the conditions of those experiments differ from the studies described here which use a CRT-KO background, our findings are consistent with the prediction that any excess free ER calcium generated in cells lacking CRT or sufficient ER calcium binding sites would rapidly diffuse out of the ER and be pumped out of the cell.

Related to cytosolic calcium levels and SOCE activation, consistent with our findings (Figure 6G), a previous study found no differences in cytosolic calcium levels or SOCE induction in myeloid progenitor 32D cells expressing wild-type CRT and CRT_Del52 both under conditions of cytokine deprivation and thrombopoietin (TPO) stimulation (Bhuria et al., 2024). While cultured megakaryocytes from MPN patients with CALR mutations have SOCE abnormalities compared to healthy control cells (Pietra et al., 2016, Di Buduo et al., 2020) sample to sample heterogeneities in primary cells make it difficult to attribute measured differences solely to calcium binding by CRT or its mutants.

We found that CRT-KO HEK cells displayed lower ER calcium compared with the parental HEK cells (Figure 4D, left panel) and also that CRT-KO Meg-01 cells displayed increased cytosolic calcium in the absence of extracellular calcium (Figure 6B). However, reconstitutions with wild-type CRT did not reverse these effects (Figure 4D, right panel and Figure 6E). Transcriptomic analysis identified significant alterations in the expression of genes related to calcium storage and signaling between parental and CRT-KO Meg-01 cells (Figure 7). These changes highlight adaptive or compensatory pathways induced by CRT-KO and emphasize that, when comparing wild-type and CRT-KO cell lines for subcellular calcium levels or responses to SERCA inhibitors, the measured effects might not be based solely on the abilities of CRT to maintain ER and cytosolic Ca²⁺ levels via its calcium binding, storage and buffering capacity. Rather, many other compensatory cellular changes could affect subcellular calcium levels and signaling. In this study, various ER proteins with known calcium binding activities were found to be upregulated in CRT-KO cells (Figure 7A and 7B). GRP94 (Macer and Koch, 1988, Biswas et al., 2007, Marzec et al., 2012), BiP ((Lievremont et al., 1997) and our unpublished findings), and PDIA6 (Okumura et al., 2021) are each known to contain low affinity calcium binding sites. CRT-KO cells have significantly induced expression of the genes encoding these proteins. Wild type CRT reconstitution into the CRT-KO cells significantly reduced the expression of these ER proteins, whereas CRT_Del52 reconstitution only partly and non-significantly reduced expression of HSP90B1 and PDIA6 and did not reduce the expression of HSPA5 (Figures 7D-F). GRP94, BiP and PDIA6 have functions in ER protein folding and quality control (Hendershot et al., 2024, Okumura et al., 2021). It is known that CRT_Del52 has impaired chaperone activity towards its substrates compared with wild type CRT (Arshad and Cresswell, 2018, Schurch et al., 2022, Kaur et al., 2025). This could account for some of these measured differences between wild type CRT and CRT_Del52 in altering the expression of the ER chaperones and folding factors. However, their induction in knock-in cells with heterozygous expression of wild type CRT and CRT_Del52 (Fosselteder et al., 2023), in which the expressed wild type CRT appears sufficient for multiple protein folding activities (Kaur et al., 2025, Schurch et al., 2022) could point to compensatory effects relevant to calcium signaling. Regardless of the precise mechanisms relevant to induction of the calcium binding proteins, their combined expression, despite protein compositional differences between the KO and reconstituted cells, appears sufficient to maintain ER calcium levels (Figure 4D, right panel), calcium release from the ER (Figure 4E, right panel, and 5F) and cytosolic calcium levels (Figure 5E).

PDIA3 (ERp57) forms a complex with CRT and interacts with and regulates the activities of the ubiquitously expressed SERCA2b (Li and Camacho, 2004). Its expression is also induced by the CRT-KO and suppressed by wild type CRT expression but non-significantly by CRT_Del52 expression. Altered SERCA2b activity could also affect basal ER calcium levels, contributing to similarities between CRT-KO, and wild type CRT or CRT_Del52 reconstituted cells.

Collectively, our findings reveal that CRT_Del52 retains some ability to bind calcium at low affinity sites. CRT-KO cells reconstituted with wild-type or CRT_Del52 have comparable ER and cytosolic calcium levels and store-operated calcium entry (SOCE). CRT-KO induces broad changes in ER calcium binding protein expression as well as in the expression of ERp57. Overall, we can conclude that a number of factors that contribute to the maintenance of ER calcium levels and regulation of calcium transport into the ER are altered in CRT-KO cells. Thus, any disease-related changes in calcium signaling and homeostasis are likely multifactorial, encompassing transcriptional regulation and altered expression of critical mediators of calcium storage, transport and signaling pathways. These studies provide a more nuanced understanding of CRT’s role in cellular calcium signaling and may inform future therapeutic strategies targeting calcium signaling pathways in hematological malignancies.

Acknowledgements

We thank Grace Pagnucco for her contribution to creating the CRISPR/Cas9-engineered HEK293T CRT-KO cells and Arunkumar Venkatesan for the MEG-01 CRT-KO cells. We would like to extend special thanks and appreciation to Chante Liu and Mathewos Biniam Tebeje from the Satin lab for their support with microscopic calcium measurements. We thank Luis Teran-Rodriguez for assistance with data analysis and Dr. Sivaraj Sivaramakrishnan for many helpful suggestions.

We thank the Advanced Genomic Core (ACG) at the University of Michigan for conducting the library preparation and next-generation sequencing, as well as Rebecca Tagett and the Bioinformatics Core of the University of Michigan Medical School for conducting the RNA sequencing data analysis. We thank Dr. Arthur Sherman for many insightful comments. This work was funded by the National Institute of Health grants (R01 AI123957 to MR and R01DK46409 to LSS) and by the University of Michigan Fast Forward Protein Folding Diseases Initiative.

Additional files

Supplemental Methods

Additional information

Funding

HHS | NIH | National Institute of Allergy and Infectious Diseases (NIAID) (AI123957)

• Malini Raghavan

HHS | NIH | National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (DK46409)

• Leslie S Satin