Truth or dare: role of Liar in EPO-dependent signaling

The Lyn tyrosine kinase is an important signaling protein in murine erythropoiesis. In this issue of Blood, Ingley and colleagues identify a novel Lyn interacting protein and characterize its role in red blood cell development.

Identification of a novel Lyn scaffolding protein adds to the complexity, regulation, and understanding of erythropoietin (Epo)–dependent signaling. Our knowledge of the proximal events in Epo-dependent signaling pathways has been advanced by the generation and characterization of gene-targeted mice. Knockout of Epo, Epo receptor, and the tyrosine kinase JAK2 have defined a linear genetic pathway consisting of a single ligand, a cognate receptor, and a critical upstream tyrosine kinase that couples to activation of STAT transcription factors, Akt and MAP kinases (reviewed in Richmond et al¹ ). Other tyrosine kinases, most notably the Src tyrosine kinase family member, Lyn, have been shown to play an important role in Epo-mediated signaling.

Lyn is important for signaling in B cells,^2-4  myelomonocytes, and mast cells. Lyn deficiency results in decreased phosphorylation of Syk as well as increases in Erk and Akt phosphorylation in B cells.^5,6  Enhanced responses to granulocyte-macrophage colony stimulating factor (GM-CSF) and macrophage-CSF (M-CSF) as well as enhanced Akt phosphorylation was observed in Lyn-deficient macrophages.⁷  The Fc-Epsilon Receptor-I expresses ITAM motifs that facilitate binding of Syk and Lyn^8,9  in mast cells. Negative regulation of mast cell signaling is facilitated by Lyn.^10,11 

Three studies have examined erythroid defects in Lyn-deficient mice.^12-14  Hibbs and colleagues reported splenomegaly and enhanced sensitivity to 5-fluorouracil in Lyn^−/− mice.¹⁴  Decreased Epo-dependent phosphorylation of STAT5 was observed in 2 studies.^12,13  Klinken and colleagues also demonstrated that Lyn-deficient erythroblasts had lower levels of the erythroid transcription factors GATA-1 and EKLF (KLF1).¹³  Wojchowski and colleagues demonstrated defective expansion of late-stage Lyn-deficient cells that were also characterized by lower expression of Bcl-XL.¹² 

Ingley and colleagues have performed yeast 2 hybrid screening using catalytically active and inactive forms of Lyn as bait. Previous studies showed that HS1, a Src family kinase substrate, decreased Epo-dependent proliferation and differentiation of J2E cells.¹⁵  In this issue of Blood, Samuels et al report the identification of a novel scaffolding protein termed Lyn-interacting ankyrin repeat (Liar) protein.¹⁶  This highly conserved signaling adaptor expresses a putative ATP-binding loop, a nuclear localization sequence, 4 ankyrin repeats, and a nuclear export sequence. The mouse gene is found on chromosome 15 whereas the human ortholog is located at a syntenic region on human chromosome 22q13.1.

Lyn and Liar constitutively associate when overexpressed in COS cells. This interaction requires the SH3 domain of Lyn and binds to Liar in a novel manner. The Lyn-Liar binding is ankyrin-domain dependent, and no consensus SH3 proline-rich binding motif is expressed on Liar. Of significance, Liar does bind to other SH3 domain–containing proteins including Vav1, HS1, ESE2L, Hip55, and LASP1 as determined by using Liar as bait in a yeast 2 hybrid assay.

Liar is expressed in the erythroid cell line, J2E, and is transiently localized in the nucleus within 30 minutes after Epo stimulation. There is no evidence for Epo-dependent tyrosine phosphorylation of Liar, illustrating the importance of using protein interaction assays to identifying interacting proteins that are not subject to tyrosine phosphorylation.

Overexpression of Liar or a form of Liar with a mutated form of the nuclear localization sequence (Liar-ΔNLS) in primary erythroid cells inhibited burst-forming unit–erythroid and colony-forming unit–erythroid colony formation. Liar and Liar-ΔNLS inhibit erythroid differentiation of J2E transfectants without affecting expression of the critical erythroid regulatory transcription factor, GATA-1. Liar-expressing cells display increased proliferative capacity. Interestingly, overexpression of Liar or Liar-ΔNLS significantly inhibited Akt phosphorylation and J2E cells expressing Liar displayed enhanced sensitivity to serum withdrawal. Overexpression of Liar in J2E cells also reduces Epo-dependent phosphorylation of Lyn and Stat5. Liar-ΔNLS, but not Liar, results in enhanced time-dependent phosphorylation of Erk1/2. The cytoplasmic form of Liar also appears to be enriched at the cell membrane whereas a significant pool of wild-type Liar is found in the nucleus within 60 minutes of cytokine stimulation. This indicates that Liar is a scaffold that is found in both the cytosol and nucleus and may be important in transporting Lyn (which is myristoylated and palmitoylated) from the membrane to intracellular compartments.

The identification of Liar has important implications in signal transduction. It will be interesting to determine whether Liar can bind to other Src family tyrosine kinases. Its binding to other SH3 domain–containing proteins indicates that it could be a generic scaffolding protein that couples to distinct signal transduction pathways. Liar is widely expressed in many tissues so gene targeting experiments may require tissue-specific knockouts. Of significance, Lyn participates in chronic myeloid leukemia,¹⁷  in B-cell acute lymphoid leukemia mediated by BCR-ABL,¹⁸  and in chronic lymphoid leukemia.¹⁹  Liar resides in an area on human chromosome 22q13.1 that has been associated with leukemia and other cancers. Whether a signaling complex involving Lyn-Liar participates in human disease will need to be evaluated. The truth about the function of this interesting novel gene will emerge from future investigations.