Phosphorylated Serine Marks Proteins for Degradation By the Mitochondrial Matrix Protease, Clpxp

Abstract

Mitochondrial ClpXP is an ATP-dependent protease complex in which the ClpX AAA+ chaperone binds, unfolds, and translocates specific protein substrates into the degradation chamber of ClpP. The serine protease ClpXP is located in the mitochondrial matrix and maintains the integrity of the respiratory chain by degrading damaged and/or misfolded proteins. This protease is overexpressed in 45% of AML patients. Genetic or chemical inhibition or hyperactivation of ClpXP kills leukemic cells and stem cells in vitro and in vivo. Yet, the degron sequences or post-translational modifications that are targeted by mitochondrial ClpXP remain to be elucidated. In Bacillus subtilis, the bacterial ClpXP homologue degrades proteins containing phospho-arginine. To determine if phosphorylated amino acids also mark proteins for mitochondrial ClpXP degradation, we evaluated human ClpXP protease activity in the presence of phosphorylated amino acids. Recombinant human ClpXP was incubated with FITC-tagged casein substrates and increasing concentrations of phospho-serine (pSer), phospho-threonine (pThr), phospho-arginine (pArg), or phospho-tyrosine (pTyr). The release of fluorogenic FITC-tag was measured to assess ClpXP protease activity. In a dose-dependent manner, pSer and pThr free amino acids inhibited casein cleavage by ClpXP. Similar inhibition was seen with short peptides containing pSer and pThr. In contrast, pTyr, pArg, free phosphates, and the dephosphorylated amino acids or peptides did not affect ClpXP protease activity in this assay. Although pSer and pThr inhibited the protease activity of ClpXP, they had no impact on ClpX ATPase activity nor the peptidase activity of isolated ClpP. We characterized ClpP peptidase activity and ClpX ATPase activity using the Ac-Trp-Leu-Ala-AMC (Ac-WLA-AMC) peptidase activity assay and the Malachite Green Phosphate assay, respectively. pSer and pThr had no effect on ClpP peptidase activity nor ClpX ATPase activity. Therefore, pSer and pThr specifically inhibited ClpXP's ability to degrade full-length proteins. Next, we investigated if pSer and pThr could bind to ClpX or ClpP. Thermal shift binding assays indicated that pSer and pThr but not pTyr and pArg bound to ClpX and none of the phosphorylated amino acids bound to ClpP. Similarly, Ala-pSer-Ala (ApSA), Arg-Arg-Ala-pSer-Val-Ala (RRApSVA) and Ala-pThr-Ala (ApTA) peptides bound to ClpX, while the non-phosphorylated peptides did not. Thus, taken together, pSer and pThr interact with ClpX at a site outside its ATP binding pocket. Previously we showed that ClpX and ClpP interact with respiratory chain complex II subunit SDHA. Moreover, inhibition of ClpP in AML cells increased misfolded or damaged SDHA protein and impaired respiratory chain complex II activity. Therefore, we asked if inhibition of ClpP and ClpX impacted levels of SDHA serine phosphorylation. We knocked down ClpX and ClpP individually in OCI-AML2 cells using shRNA. Then, mitochondrial lysates were prepared, serine phosphorylated proteins were immunoprecipitated and probed for SDHA. Increased abundance of serine phosphorylated SDHA was observed in AML cells after ClpP and ClpX knockdown. We also demonstrated that addition of recombinant ClpXP to the immunoprecipitates selectively degraded the fraction of serine phosphorylated SDHA. We then showed that serine phosphorylated SDHA was present only in the digitonin insoluble fraction of mitochondria proteins, indicating it is only found on damaged protein aggregates. In summary, we discovered that pSer and pThr promote the degradation of damaged respiratory chain proteins by mitochondrial ClpXP. Thus, this work highlights new mitochondrial biology and identifies a potential new strategy to develop ClpXP inhibitors for the treatment of AML.