Comment on Zhao et al, page 907
Stimulation of the PI3-kinase/AKT pathway by the ligand-bound erythropoietin receptor leads to phosphorylation and activation of the transcription factor GATA-1.
Normal erythroid differentiation requires the cytokine erythropoietin (Epo) and its membrane-bound receptor (EpoR). Signals emerging from this receptor are transmitted to the nucleus by a branched network of downstream effectors and ultimately regulate gene transcription. Arguably, the best-studied erythroid transcription factor is the zinc finger protein GATA-1, which regulates all erythroid-specific genes studied to date. GATA-1 is an obvious candidate for regulation by Epo, but direct evidence for this has been lacking—until now. Two reports, one in this issue of Blood and one in Molecular and Cellular Biology1 show for the first time that Epo stimulates phosphorylation of GATA-1 via AKT. AKT phosphorylates GATA-1 in vitro and in vivo near the second zinc finger at a serine residue (S310), which is embedded in an AKT consensus phosphorylation site and is conserved in the related hematopoietic proteins GATA-2 and GATA-3. Prior studies recognized S310 of GATA-1 as a target for phosphorylation, but the relevant kinases and functional consequences have been elusive.2 Despite the absence of experiments showing directly that AKT is required for Epo-induced GATA-1 phosphorylation in erythroid cells, several gain-of-function studies build a strong case that AKT is indeed the relevant GATA-1 S310 kinase. Both of the current studies show that activated AKT stimulates GATA-1 activity in transfection assays and enhances GATA-1's ability to trigger erythroid differentiation in a GATA-1-null erythroid cell line. The 2 reports complement each other in examining the role of GATA-1 phosphorylation in vivo. The report by Zhao and colleagues shows that nonphosphorylatable GATA-1 mutant (S310A) dominantly inhibits differentiation of primary fetal liver erythroblasts. Kadri et al1 show that erythroid cells from mice bearing an S310A knock-in mutation fail to express the TIMP-1 (tissue inhibitor of matrix metalloproteinase) gene, which is normally activated by Epo and is a direct GATA-1 target.
Together, these studies fill an important gap in our understanding of how Epo promotes erythroid maturation. However, phosphorylation of GATA-1 is clearly not sufficient to account for the broad biologic effects of Epo or AKT.3 Mutation of S310 in GATA-1 affects its activity in selective and subtle ways, with a phenotype much milder than that caused by loss of Epo. Therefore, the search for additional AKT targets in erythroid cells is warranted. One likely group of candidates is the FOXO transcription factors, known AKT targets that regulate a variety of cellular processes and have already been implicated in erythroid differentiation.4,5 In addition, whether AKT is the only kinase targeting S310 of GATA-1 remains to be investigated.
Since GATA-1 plays a role in the transcription of the EpoR gene, the current findings suggest a positive feedback loop between GATA-1 and EpoR. A teaser to support this was provided by transient transfection experiments in which GATA-1 with the S310A mutation failed to activate an EpoR promoter construct in heterologous cells while it activated other targets normally.1 However, expression levels of the EpoR were not examined in the S310A knock-in mice.
How might S310 phosphorylation regulate GATA-1 activity? It does not appear to significantly alter DNA binding, protein stability, and nuclear localization, consistent with previous work.2 Therefore, regulation of GATA-1 protein interactions is a distinct possibility. Another important question raised by these studies concerns the role of TIMP-1 in erythropoiesis. For example, it will be of great interest to examine whether its protease inhibitory activity controls the erythropoietic microenvironment in hematopoietic tissues.
Finally, various myeloproliferative disorders are associated with activating mutations in JAK2 leading to stimulation of downstream effectors including PI3-kinase/AKT.6 This raises the interesting possibility that the pathogenesis of these disorders might involve misregulation of GATA-1 or GATA-2 phosphorylation in myeloid precursors. ▪