Erythroid-Specific Apoptosis Induced by Jun Transcription Factors Employs a Novel Pathway Involving Recruitment of COP1 and Itch Ubiquitin Ligases.

Upregulation of the transcription factor c-Jun has been identified in acute myeloid leukemia, chronic myelogenous leukemia, myelodysplastic syndromes, and in iron deficiency. These conditions all share as a common feature an impairment of erythropoiesis, in some instances manifested by apoptosis of erythroid progenitors. Depending on cellular context, c-Jun may promote either survival or apoptosis, the latter occurring primarily through regulation of target genes such as Bim and FasL. Transcriptional activation by its kinase JNK typically enhances the apoptotic effects of c-Jun. In primary human erythroid progenitors, we have identified a novel Jun-mediated apoptotic pathway that does not depend on transcriptional activation and is associated with retrograde inhibition of JNK signaling. To determine the nature of this pathway, we have employed a system of retroviral transduction of human CD34+ peripheral blood mobilized progenitors, which undergo culture in serum free erythroid medium containing erythroipoietin and stem cell factor. Wild type c-Jun and JunB, as well as c-Jun mutants lacking transactivation and DNA binding functions, all efficiently induced caspase-dependent apoptosis in early erythroid precursors within the CD36-positive GPA-dim compartment. By contrast, transduced cells undergoing megakaryocytic and myeloid differtentiation, as indicated by CD41 and CD13 expression, showed minimal apoptosis regardless of construct. Deletion or replacement of the c-Jun leucine zipper (LZ) dimerization domain completely eliminated the induction of erythroid apoptosis. Introduction of an LZ point mutation E281K, which alters dimerization specificity, also eliminated induction of apoptosis. These results demonstrated the dispensibility of direct target gene regulation but indicated an absolute requirement for specific LZ protein-protein interactions. Because JunD showed relatively inefficient apoptotic induction, we considered a role for recruitment of ubiquitin ligases such as Itch, which selectively binds c-Jun and JunB. Indeed, the c-Jun Y170F mutation, which ablates the PPXY binding site for Itch (and possibly other WW-domain ubiquitin ligases), completely abrogated apoptosis. Surprisingly, the c-Jun mutant VP232/233AA, which eliminates recruitment of the COP1 RING finger ubiquitin ligase, also manifested major defects in apoptosis induction. Biochemical analysis of the transduced cells showed that induction of apoptosis positively correlated with repression of JNK phosphorylation. Wild type c-Jun, JunB, and mutants that retained induction of apoptosis blocked JNK phosphorylation. All mutants tested that were defective in apoptosis induction caused an enhancement of JNK phosphorylation. We therefore postulate that this novel pathway of erythroid apoptosis results from c-Jun nucleation of a protein complex involving an LZ partner and multiple ubquitin ligases. Assembly of this complex then leads to blockade of JNK phosphorylation and induction of apoptosis.