Rac GTPases regulate the morphology and deformability of the erythrocyte cytoskeleton

Abstract

Actin oligomers are a significant structural component of the erythrocyte cytoskeleton. Rac1 and Rac2 GTPases regulate actin structures and have multiple overlapping as well as distinct roles in hematopoietic cells; therefore, we studied their role in red blood cells (RBCs). Conditional gene targeting with a loxP-flanked Rac1 gene allowed Crerecombinase-induced deletion of Rac1 on a Rac2 null genetic background. The Rac1(-/-);Rac2(-/-) mice developed microcytic anemia with a hemoglobin drop of about 20% and significant anisocytosis and poikilocytosis. Reticulocytes increased more than 2-fold. Rac1(-/-);Rac2(-/-) RBCs stained with rhodamine-phalloidin demonstrated F-actin meshwork gaps and aggregates under confocal microscopy. Transmission electron microscopy of the cytoskeleton demonstrated junctional aggregates and pronounced irregularity of the hexagonal spectrin scaffold. Ektacytometry confirmed that these cytoskeletal changes in Rac1(-/-);Rac2(-/-) erythrocytes were associated with significantly decreased cellular deformability. The composition of the cytoskeletal proteins was altered with an increased actin-to-spectrin ratio and increased phosphorylation (Ser724) of adducin, an F-actin capping protein. Actin and phosphorylated adducin of Rac1(-/-);Rac2(-/-) erythrocytes were more easily extractable by Triton X-100, indicating weaker association to the cytoskeleton. Thus, deficiency of Rac1 and Rac2 GTPases in mice alters actin assembly in RBCs and causes microcytic anemia with reticulocytosis, implicating Rac GTPases as dynamic regulators of the erythrocyte cytoskeleton organization.

Figure 1.

Cre-mediated deletion of Rac1 sequence after pI-pC treatment. (A) Rac1 protein as detected by immunoblotting in WT and Rac1^–/–;Rac2^–/– washed RBC samples obtained weekly over the period of 2 to 5 weeks after induction. GAPDH used as loading control. Data representative of 12 samples tested per week. (B) Active, GTP-bound Rac1 determined by PBD pull-down assay 3 weeks after pI-pC treatment. Genotypes shown above. GAPDH used as loading control. Data representative of 3 experiments with similar results.

Figure 2.

Abnormal morphology of Rac1^–/–;Rac2^–/– erythrocytes. (A) Wright-Giemsa staining of blood smears of WT, Rac1^–/–, Rac2^–/–, and Rac1^–/–;Rac2^–/– mice. Rac1^–/– and Rac2^–/– blood smears show mild poikilocytosis and anisocytosis. The Rac1^–/–;Rac2^–/– blood smear demonstrates severe poikilocytosis; arrows indicate the presence of hypochromia (1), polychromasia (2), and fragmented cells (3). Scale bar represents 10 μm. (B) RBC volume and hemoglobin concentration histograms obtained by automated complete blood count analysis of WT and Rac1^–/–;Rac2^–/– whole blood with an Advia Hematology Analyzer (images representative of 6 samples tested for each genotype). (C) Three-dimensional reconstruction of glutaraldehyde-fixed erythrocyte images. Images were obtained with a × 100 oil-immersed objective lens, numerical aperture 1.45. Arrows show Rac1^–/–;Rac2^–/– erythrocytes with thinned appearance especially at the central area (1), and a high frequency of bizarre microspherocytes with punctuate lesions (2). One unit represents 9.2 μm. (D) Increased iron deposits in the spleen of Rac1^–/–;Rac2^–/– mice compared with WT. Scale bar represents 100 μm.

Figure 3.

Reduced deformability index of Rac1^–/–;Rac2^–/– erythrocytes. Ektacytometry of Rac1^–/–;Rac2^–/– RBCs was used to determine the maximum value of deformability index (DImax). Results are expressed as mean ± SD of 12 Rac1^–/–;Rac2^–/– blood samples and 6 WT, Rac1^–/–, Rac2^–/– blood samples obtained 2 weeks after completion of treatment with pI-pC. P < .001 by t test for the difference between Rac1^–/–;Rac2^–/– and WT. Rac1^–/– and Rac2^–/– blood samples had no statistically significant difference from WT.

Figure 4.

Cytoskeleton structure of Rac1^–/–;Rac2^–/– erythrocytes. (A) Erythrocytes fixed with acrolein and stained with rhodamine-phalloidin for F-actin. Arrowheads in magnified inset show meshwork gaps and aggregates of F-actin in Rac1^–/–;Rac2^–/– erythrocytes. (B) RBCs stained with anti–band 3 antibody to visualize band 3 distribution. Images were obtained with a × 63 oil-immersed objective lens, software zoom × 2, numerical aperture 1.4; 1 unit represents 7.3 μm.

Figure 5.

Organization of the membrane skeletons of WT and Rac1^–/–;Rac2^–/– erythrocytes. Micrographs of representative metal cast of a WT and a Rac1^–/–;Rac2^–/– unroofed erythrocyte. (A) The membrane skeleton of the WT erythrocyte is composed of a uniform lattice of interconnected strands, as previously reported. Rac1^–/–;Rac2^–/– membrane skeletons revealed irregularity of the hexagonal spectrin scaffold and a paucity of membrane “endovesiculations.” Scale bar represents 100 nm. (B) In higher magnification, the arrowheads point to large clusters in the Rac1^–/–;Rac2^–/– cytoskeleton that appear to have the same “globular” consistency with the junctions in the WT cytoskeleton. Scale bar represents 100 nm.

Figure 6.

Erythrocyte protein and phosphorylation profile in Rac1^–/–;Rac2^–/– RBCs evaluated in SDS-polyacrylamide gels. (A) RBC ghost electrophoretic pattern in Rac1^–/–;Rac2^–/– versus WT erythrocytes. Molecular weight markers and major RBC cytoskeleton proteins are indicated. Densitometry scans of the 2 lanes, respectively, are shown on the right with spectrin and actin bands noted between the dotted lines. The actin-to-spectrin ratio was reproducibly 2 to 3 times higher in the Rac1^–/–;Rac2^–/– ghosts versus the WT ghosts. (B) Profile of the major actin-interacting proteins and serine phosphorylation of adducin in WT and Rac1^–/–;Rac2^–/– erythrocytes (2 μL packed RBCs loaded/lane). Two different samples from each genotype are shown. (C) Triton-soluble (S) and pellet (P) fractions of erythrocyte ghosts evaluated by immunoblotting for the presence of spectrin, adducin, actin, and phospho-adducin (Ser724). All images are representative of at least 6 different blood samples from each genotype.