Zinc-Dependent Homomultimerization of Mutant Calreticulin Is Required for MPN Pathogenesis

Recurrent mutations in the ER chaperone calreticulin (CALR) are found in ~30% of MPNs. All known CALR mutations result in a +1-frameshift of the CALR reading frame that retains the N-domain and P-arm of wild-type CALR and generates a novel mutant-specific C-terminal domain encoded by an alternative reading frame. Multiple groups have previously demonstrated that mutant CALR (CALR^MUT) physically binds to and activates the thrombopoietin receptor MPL and causes inappropriate receptor activation and downstream JAK-STAT signaling in a ligand-independent manner. These oncogenic activities of CALR^MUT have been shown to be dependent on a threshold of basic amino acid residues in the mutant-specific C-terminus and on the lectin motif within the N-domain.

In this study, we extended the analysis of the role of the N-domain in CALR^MUT activity by performing a comprehensive alanine mutagenesis screen of all 177 non-alanine residues within the N-domain and assessed the ability of alanine-mutants to confer cytokine independent growth in Ba/F3-MPL cells. We identified two functional groups of residues as being important for CALR^MUT activity: (i) Residues associated with CALR lectin motif. Consistent with previous work which had shown that Asp135 and Asp317 was essential for CALR^MUT activity, we observed that loss of any of 4 residues within the lectin motif (Cys105, Lys111, Gly133 and Asp135) led to total impairment in the ability of CALR^MUT to confer cytokine independence, bind MPL or elicit JAK-STAT signaling. Surprisingly, other residues implicated in lectin activity (Tyr109, Tyr128, Met131) had either only a partial effect or no effect on CALR^MUT oncogenic activity, suggesting that different regions within the lectin motif of CALR have differential effects in facilitating MPL binding. (ii) Residues associated with zinc binding. We also identified a group of zinc-binding histidine residues (His99, His145 and His170) as being important for modulating CALR^MUT activity. We observed that loss of any single histidine residue (1xHis) led to partial abrogation of cytokine independence in Ba/F3-MPL cells. Strikingly, combined loss of two (or more) histidines (2xHis) led to complete impairment in the capacity to confer cytokine independent growth or elicit JAK-STAT signaling, in conjunction with impaired ability to physically associate with Zn²⁺ by affinity chromatography. In addition, 2xHis-CALR^MUT failed to bind MPL in pulldown assays or colocalize with MPL in FRET-based immunocytochemical analysis.

We next investigated in greater detail how the zinc binding capacity of CALR^MUT regulates its oncogenic activity. Previous studies have shown that CALR^MUT forms homomultimers (Araki, Leukemia, 33:122), and that zinc is able to increase homomultimerization of recombinant wild-type CALR in vitro (Li, Biochemistry, 40:11193). We therefore tested the multimerization capacity of lectin-deficient or zinc binding-deficient CALR^MUT by assessing the ability of heterologously-expressed FLAG-tagged CALR^MUT species to complex and co-immunoprecipitate with a V5-tagged CALR^MUT species in 293T cells. We observed that lectin-deficient CALR^MUT and 1xHis-CALR^MUT retained the capacity to form homomultimers, but the 2xHis- CALR^MUT could not, suggesting that zinc binding was required for homomultimer formation. Treatment with ZnCl₂ which increases free intracellular Zn²⁺ resulted in increased CALR^MUT homomultimerization in a dose-dependent manner, as well as increased binding to MPL within heteromeric complexes. Finally, treatment with the zinc chelator TPEN led to decreased CALR^MUT homomultimer formation, decreased CALR^MUT-MPL heteromultimeric complexes and attenuated JAK-STAT signaling. Cumulatively, our data suggest that zinc is required for homomultimerization of CALR^MUT, and is a necessary prerequisite event to facilitate binding to and activation of MPL in CALR-mutated MPNs.