DOMAIN ANALYSIS FOR TWO DISTINCT TYPE II METACASPASES FROM ARABIDOPSIS THALIANA

Programmed cell death (PCD) plays important role in development and in response to environmental stresses. Metacaspases (MCs), like the structurally related caspases of metazoans, are cysteine proteases that have been genetically demonstrated to contribute to PCD control and stress responses in plants, protozoa and fungi. Type I AtMCs, contain an N-terminal pro-domain with zinc-binding motifs and/or a proline-rich region, while type II MCs that are predominantly found in plants, lack this N-terminal extension but contain a variable-sized linker between the putative p20 and p10 domains and a longer C-terminal tail. Although MCs have been discovered for more than 15 years, their regulatory mechanisms remain largely uncharacterized. AtMC4 and AtMC9 are two prototypical type II metacaspases from the model plant Arabidopsis thaliana and they display distinct as well as common biochemical properties. These include different requirements of calcium ion for zymogen activation, their pH optima, and the common characteristic of auto-inactivation. Taking advantage of the common domain structure between these two MCs, we seek to determine the structural basis for these characteristics in order to elucidate the mechanisms that may regulate these important proteases' activities. Using domain-swap strategies, we generated a set of 6 chimeric constructs encoding MC proteins that contain all possible permutations of the three analogous domains in AtMC4 and AtMC9: the p20 domain containing the active site cysteine, the linker region, and the distal p10 domain at the C-terminus. These protein chimeras, and the wild type AtMC4 and AtMC9, were produced in bacteria using inducible pET vectors and purified using His-tag strategy on affinity columns. Comparison of the enzymatic activities of the eight recombinant proteins revealed that the p20 domain at the N-terminus of AtMC4 contains the determinant for calcium dependence. Interestingly, the auto-inactivation property that is observed in both AtMC4 and AtMC9 are not observed in any of the 6 chimeric proteins, thus indicating the likely coupling of multiple structural elements in all three domains is required for this behavior. Lastly, by substitution of the linker region in AtMC9 from that of AtMC4, the apparent pH optimum is shifted from an acidic pH of approximately 5.5 to a nearly neutral pH of 7.5 while the reverse substitution of the linker in AtMC4 did not perturb its pH optimum. This indicates that a combination of structural factors may be required in the linker and p10 domains in order to generate an acidic optimum for enzyme activity in type II MCs. Our work using domain-swap approach thus revealed evidence for a calcium-dependent switch located in the N-terminus domain of AtMC4. Further structural analysis of this p20 domain in comparison to that from AtMC9 should help reveal this novel element. Active chimera MCs that we created in this work may also help to elucidate the functions of these conserved proteases by circumventing their normal regulatory pathways.