Targeted Deletion of γ-Adducin in Mice.

The adducins are a family of three closely related proteins (α, β, γ) encoded by distinct genes. α- and γ-adducin are expressed ubiquitously, while β expression is restricted to hematopoietic cells and the brain. In red blood cells (RBCs) adducin localizes to spectrin-actin junctions in the membrane skeleton as αβ heterotetramers. Previously (Gilligan et. al., PNAS, 1999) we showed that deletion of β-adducin results in osmotically fragile, microcytic RBCs and an overall phenotype of hereditary spherocytosis (HS). Notably, α-adducin was significantly reduced in β-adducin null RBCs. We also demonstrated that γ-adducin is present in low amounts in normal mouse RBCs and is upregulated ∼5-fold in β-adducin null RBCs. The increase in γ-adducin suggests that αγ heterotetramers may be compensating for the absence of β-adducin. In an effort to analyze γ-adducin function in RBCs in greater detail, we generated a conditional γ-adducin knockout allele in mice using a Cre-loxP strategy to delete exon 2 containing the start codon. All mice were maintained on a segregating B6.129 genetic background. Western blotting confirmed the absence of γ-adducin in spleen homogenates and RBC ghost preparations from γ-adducin null mice. All other membrane skeleton proteins examined by a combination of SDS-PAGE and western blotting, including α- and β-adducin, are normal in γ-adducin null RBCs (spectrin, ankyrin, band 3, protein 4.1, protein 4.2, dematin). Phenotypically, γ-adducin null mice display normal growth curves and show no overt defects. γ-adducin null RBCs appear normal on Wright’s stained peripheral blood smears and by scanning electron microscopy (SEM). The RBC count, hemoglobin content, hematocrit, MCV, reticulocyte %, osmotic fragility, and all other hematopoietic parameters are normal in γ-adducin null mice vs. wildtype. The apparent compensation by γ-adducin in β-adducin null red cells previously observed was tested by intercrossing mice null for γ- and β-adducin to produce βγ null double homozygotes. The additional loss of γ-adducin did not exacerbate the β-adducin null RBC phenotype as judged by examination of peripheral blood smears and SEM. Moreover, RBC osmotic fragility and complete blood counts in βγ-adducin null mice did not differ from β-adducin null mice. Western blotting of RBC ghost proteins confirmed reduction of α-adducin to ∼20% of normal in β-adducin null mice, as previously described. Strikingly, α-adducin in βγ-null RBC ghosts is reduced to barely detectable levels (<5% of normal). These studies show that (1) loss of γ-adducin alone does not significantly impact RBC membrane skeleton structure and function; (2) α- and β-adducin are stable and present at normal levels in the absence of γ-adducin; (3) the loss of γ-adducin in β-adducin null mice does not further exacerbate the β-adducin null HS phenotype; (4) the exacerbated loss of α-adducin in βγ double null RBCs suggests that up-regulated γ-adducin in β-adducin null mice associates in some way with and stabilizes α-adducin in the RBC membrane, but is unable to compensate functionally for the loss of the β subunit. We conclude that the normal function and stable incorporation of adducin into the peripheral membrane skeleton of red cells requires the presence of heterologous αβ binding subunits. Additional studies of adducin null mouse models, including our recently generated α-adducin null strain, will be useful tools in defining adducin functions and interactions in multiple tissues and organs.