Abstract

Mammalian erythroid cells undergo enucleation, an asymmetric cell division involving extrusion of a pycnotic nucleus enveloped by the plasma membrane^{1, 2, 3}. The mechanisms that power and regulate the enucleation process have remained obscure. Here, we show that deregulation of Rac GTPase during a late stage of erythropoiesis completely blocks enucleation of cultured mouse fetal erythroblasts without affecting their proliferation or differentiation. Formation of the contractile actin ring (CAR) on the plasma membrane of enucleating erythroblasts was disrupted by inhibition of Rac GTPases. Furthermore, we demonstrate that mDia2, a downstream effector of Rho GTPases and a formin protein required for nucleation of unbranched actin filaments^{4, 5, 6}, is also required for enucleation of mouse fetal erythroblasts. We show that Rac1 and Rac2 bind to mDia2 in a GTP-dependent manner and that downregulation of mDia2, but not mDia1, by small interfering RNA (siRNA) during the late stages of erythropoiesis blocked both CAR formation and erythroblast enucleation. Additionally, overexpression of a constitutively active mutant of mDia2 rescued the enucleation defects induced by the inhibition of Rac GTPases. These results reveal important roles for Rac GTPases and their effector mDia2 in enucleation of mammalian erythroblasts.

Introduction

Mammalian erythropoiesis involves the erythropoietin (Epo)- and fibronectin-dependent proliferation and differentiation of committed CFU-e erythroid progenitors to mature erythrocytes. The signalling pathways involved in the early stages of erythropoiesis, in which erythropoietin binds to its receptor and induces activation of JAK2 and several downstream signal transduction proteins, are well established. In the late stages of erythropoiesis when Epo is no longer required⁷, erythroblast proliferation, but not differentiation, requires signalling by 41 integrins that adhere to fibronectin⁸. Late erythroblasts undergo terminal cell-cycle exit, chromatin condensation and extrusion of the pycnotic nucleus via an asymmetric cell division^{9, 10, 11}, but the signalling pathways and genes involved in these last steps of erythropoiesis are not known.

As a first step in addressing these long-standing questions, enucleation was monitored in vivo during fetal liver erythropoiesis (this system was used as more than 90% of the cells are erythroid). To this end, mouse fetal liver cells simultaneously labelled for the erythroid-specific plasma membrane glycoprotein TER119 and the cell-permeable DNA staining dye Hoechst 33342 were analysed by flow cytometry at different embryonic days. The percentage of reticulocytes (defined as Hoechst^lowTER119^high cells) in fetal liver gradually increased during fetal development (Fig. 1a).

Figure 1: Enucleation of mammalian erythroid cells in vivo and during in vitro culture.

(a) Flow cytometric analysis of mouse fetal erythroblast enucleation in vivo. The total population of fetal liver cells from E12.5 to E15.5 embryos was stained with Hoechst 33342 and PE-conjugated TER119 and analysed by FACS. Debris (low forward scatter) and dead cells (propidium iodide positive) were excluded from the analysis. The percentage of cells in each quadrant is indicated; enucleated reticulocytes fractionate in the lower right quadrant. (b) Analysis of mouse fetal erythroblast enucleation during in vitro culture. E13.5 fetal liver cells were stained with a biotinylated anti-TER119 monoclonal antibody. Purified TER119-negative erythroid progenitor cells were cultured in vitro in fibronectin-coated plates in medium containing serum and erythropoietin (Epo). The differentiation and enucleation status of cultured cells at different days were analysed by flow cytometric analysis using FITC–CD71 and TER119–PE, and Hoechst 33342 and TER119–PE, respectively. (c) Visualization of extruded nuclei and reticulocytes formed during in vitro culture. Three distinct populations of cells after three days of culture were defined: R6 (blue; extruded nuclei), R7 (magenta; nucleated erythroblasts) and R8 (red; reticulocytes). R6 to R8 cells were sorted by a Moflo cell sorter and analysed by Benzidine–Giemsa staining. The scale bar represents 20 m. (d) Time course of enucleation in culture. Flow cytometric analysis of cultured TER119-negative mouse fetal erythroblasts at different times. The percentages of R8 reticulocytes are indicated. (e) Actin staining of erythroblasts in culture. TER119-negative mouse fetal erythroblasts were cultured as in b. The erythroid cells were stained with Alexa Fluor 488–phalloidin at the indicated time point. Nuclei were stained with DAPI. The arrows indicate the CAR. The scale bar represents 7 m.

Full size image (127 KB)

TER119-negative erythroid progenitor cells were purified from E13.5 mouse fetal livers and cultured for 2 days, with erythropoietin, on fibronectin-coated plates. As shown previously, during the 2-day culture most of these erythroid progenitor cells (CD71^lowTER119^low) divided approximately four times. They exhibited induction of the transferrin receptor (CD71) by day 1 (CD71^highTER119^med) and then full induction of TER119 and downregulation of CD71 by day 2 (CD71^medTER119^high; Fig. 1b), which mimics the different erythroid developmental stages in vivo¹². Enucleation occurred by day 2, as observed by staining with Hoechst 33342 and TER119. Three populations of cells were distinguished: Hoechst^highTER119^med (R6), Hoechst^medTER119^high (R7) and Hoechst^lowTER119^high (R8), which represent extruded nuclei, nucleated erythroblasts and incipient reticulocytes, respectively (Fig. 1b, c). R7 erythroblasts were slightly bigger than R6 nuclei and R8 reticulocytes when analysed by forward scatter (Fig. 1c). Benzidine–Geimsa staining confirmed these results: R6 contained highly condensed extruded nuclei, R7 populations contained nucleated late-stage erythroblasts and R8 contained enucleated reticulocytes (Fig. 1c).

To dissect enucleation from the early stages of erythropoiesis, time course experiments were performed to pinpoint the start of enucleation. Enucleation started after approximately 35 h in culture (Fig. 1d) and reached a peak at 48 h, with 25% of the cells having enucleated (data not shown).

Studies many years ago indicated that actin filaments are critical to enucleation^{13, 14}. These studies also demonstrated the presence of a CAR at the boundary between the incipient reticulocyte cytoplasm and the nucleus of an enucleating cell¹³. To examine actin filaments, cells were fixed at different times in culture and then stained with Alexa Fluor 488–phalloidin; the levels of actin decreased during in vitro culture (Fig. 1e). Initially actin filaments were observed at high density at the plasma membrane, and in the late stages of erythropoiesis were observed to gradually accumulate at specific areas of the plasma membrane. Eventually, formation of the CAR was visualized as a bright green dot on the plasma membrane of most late-stage erythroblasts, at the boundary of the reticulocyte cytoplasm and the nucleus of enucleating cells (Fig. 1e). Erythroid cell nuclei were observed undergoing gradual condensation during erythropoiesis (Fig. 1e).

As Rho GTPases are master regulators of actin polymerization, we examined the possibility that specific Rho GTPases are involved in enucleation. Purified TER119-negative mouse fetal erythroid progenitors were infected with retroviruses encoding dominant-negative mutants of several Rho family GTPases. A bicistronic retroviral vector that expressed both the gene of interest and human CD4 as a marker of infected cells was used; thus uninfected cells (hCD4 negative) served as an internal control. After a 1 h spin infection, the expression levels of both the gene of interest and hCD4 peaked after approximately 30 h in culture (data not shown), roughly coincident with the start of enucleation (Fig. 1d). After 48 h in culture, cells were analysed and enucleation was quantified by flow cytometry. The infection efficiencies were similar among various retroviruses, as determined by the ratio of hCD4-negative and hCD4-positive cells (data not shown). Cells expressing the dominant-negative Rac1^T17N or Rac2^T17N mutants exhibited a significant decrease in enucleation, as indicated by a reduction in the percentage of R8 reticulocytes (Fig. 2a). In contrast, cells infected with control virus or dominant-negative mutants of RhoA or Cdc42 showed a slightly increased percentage of enucleation. The effect of constitutively active mutants of Rho family GTPases was also examined, and Rac1^G12V and Rac2^G12V, but not RhoA^G14V or Cdc42^G12V, inhibited enucleation (Fig. 2b). Equal levels of protein expression of these mutants were confirmed by western blot analysis of the Flag tag appended to these proteins (Fig. 2c). These results suggest that the activities of Rac GTPases are in a dynamic on-and-off status during the late stages of erythropoiesis and that either inhibition or excessive activation of Rac GTPases leads to inhibition of enucleation; in fact, similar kinds of phenotypes occur with dominant-negative and constitutively active mutants of many Rho GTPase family proteins¹⁵.

Figure 2: Deregulation of Rac GTPases blocks enucleation.

(a, b) Effect of overexpression of mutants of Rho, Rac and CDC42 on formation of reticulocytes during in vitro culture. TER119-negative mouse fetal erythroblasts were infected with bicistronic retroviruses encoding CD4 and different Flag-tagged mutant forms of several Rho-family GTPases, or as control encoding human CD4 alone. Cells were analysed by flow cytometric analysis after 48 h in culture and the percentages of incipient reticulocytes were quantified. Grey bars indicate the uninfected hCD4-negative cells that provide an internal control, and the black bars indicate hCD4-positive cells expressing different Rho GTPase mutants. (c) Cells from a and b were harvested after 48 h in culture and analysed by western blot using anti-Flag antibodies. (d) TER119-negative mouse fetal erythroblasts were cultured as in Fig. 1. After 35 h, the indicated amount of NSC23766 was added. Cells were harvested after 48 h and analysed by flow cytometry and Benzidine–Giemsa staining. The percentages of R8 cells (reticulocytes) are indicated. The scale bar represents 12 m. (e) Cells from d were harvested after 48 h in culture. Total levels of Rac1/2, RhoA and Cdc42 were determined by western blot using the indicated antibodies. The activities of Rac, RhoA and Cdc42 were determined by GTPase activity assay. (f) The same cells as in c were analysed by flow cytometry for their differentiation status. The percentages of TER119-negative and positive cells are indicated. (g) The same cells as in c were counted and the total number is shown (grey bars). The number was normalized (see main text) and is shown as black bars. (h) TER119-negative mouse fetal erythroblasts were cultured as in Fig. 1. After 35 h in culture, 100 M NSC23766 was added to some cells and after 48 hours in culture the erythroblasts were stained with Alexa Fluor 488–phalloidin. The scale bar represents 8 m. (i) Quantification of results from h. All error bars represent mean s.d. (n = 3).

Full size image (102 KB)

To study enucleation specifically without disturbing the early stages of erythropoiesis, a Rac-specific inhibitor, NSC23766, which does not affect RhoA or Cdc42 activity^{16, 17}, was used. TER119-negative mouse fetal erythroid progenitors were treated after 35 h in culture when enucleation is beginning (Fig. 1d). Cells were then harvested at 48 h in culture and analysed by flow cytometry. The effect of NSC23766 on enucleation was dose dependent and treatment with 100 M completely blocked enucleation (Fig. 2d). Benzidine–Geimsa staining showed that most of the NSC23766-treated cells had a condensed nucleus located close to the plasma membrane (Fig. 2d). Inhibition of the activities of Rac GTPases, but not RhoA and Cdc42, by NSC23766 was confirmed by Rho GTPase activity assays (Fig. 2e). Flow cytometric analysis of cell surface expression levels of glycophorin A (TER119) and the transferrin receptor (CD71) showed a normal percentage of CD71^medTER119^high cells in NSC23766-treated cells (Fig. 2f), indicating that the blockage of enucleation was not due to inhibition of normal erythroblast differentiation. There was also no difference in apoptosis between treated and control cells (see Supplementary Information, Fig. S1). NSC23766 treatment did result in a slightly lower number of total cells (T = R6 + R7 + R8; Fig. 2g). However, recognizing that enucleation forms one R6 nucleus and one R8 reticulocyte, the cell number was normalized using the equation T = R7 + (R6 + R8)/2 and no significant difference in numbers of treated and control cells was observed (Fig. 2g). This demonstrated that NSC23766 treatment at 35 h had no effect on proliferation of late erythroid cells. Similarly, expression of a dominant-negative or constitutivly active Rac1 or Rac2 protein caused no inhibition of differentiation or proliferation compared with control cells (see Supplementary Information, Fig. S2 and data not shown). Cells were also treated with C3 exoenzyme, a specific inhibitor of RhoA, after 35 h in culture and no obvious defects in enucleation were observed, further excluding a role for RhoA in enucleation (see Supplementary Information, Fig. S3).

As expected, treatment with 100 M NSC23766 after 35 h in culture completely blocked CAR formation (Fig. 2h bottom panels and 2i). Taken together, these results demonstrate that Rac GTPases are required for the formation of the CAR and enucleation.

We next determined the downstream effector of Rac GTPase on mouse fetal erythroblast enucleation. In other types of mammalian cells, diaphanous-related formins (DRFs) nucleate unbranched actin filaments, including CARs^{4, 5, 6}. There are three known DRFs in mammals: DRF1, DRF2 and DRF3, which encode mDia1, mDia3 and mDia2 proteins, respectively. All mDia proteins contain a Rho GTPase binding domain (GBD) in their amino terminus that binds to the carboxy-terminal Diaphanous-autoregulatory domain (DAD) and forms an auto-inhibitory sturcture¹⁸. The mRNA and protein levels of mDia1 and mDia2 increased during the 2-day in vitro culture of erythroid progenitors, whereas the mRNA expression of mDia3 was very low; therefore mDia3 was not analysed further (Fig. 3a, b).

Figure 3: Rac1 and Rac2 bind to mDia2 in a GTP-dependent manner.

(a) Expression levels of mDia mRNAs during in vitro culture of TER119-negative mouse fetal erythroblasts, as quantified by quantitative real time PCR analysis. The relative level compared to 18S rRNA was calculated using the delta-delta Ct method. (b) Expression levels of mDia1 and mDia2 proteins during in vitro culture of TER119-negative mouse fetal erythroblasts analysed by western blot. (c) Glutathione-coated beads with bound GST or GST fusions with Rac1 or Rac2 were incubated with lysates of 293T cells that had been transfected with Flag-tagged mDia1 GBD or mDia2 GBD, together with GDP or GTPS as indicated. Beads were then subjected to western blotting using antibodies against the Flag tag. (d) Glutathione-coated beads with bound GST or GST fusions with indicated proteins were incubated with lysates of 293T cells that had been transfected with Flag-tagged mDia2 mutants as indicated, together with GTPS. Beads were then subjected to western blotting using antibodies against the Flag tag. (e) Glutathione-coated beads with bound GST or GST fusions with Rac1 or Rac2 were incubated with 1 g of purified mDia2 GBD produced in BL21 cells, together with GDP or GTPS as indicated. Beads were then subjected to western blotting using a polyclonal antibody against mDia2. The input of mDia2 GBD was shown by Coomassie staining. (f) Rac GTPases colocalize with mDia2 but not mDia1. TER119-negative mouse fetal erythroblasts were cultured as in Fig. 1 and harvested after 48 h in culture. Cells were fixed and stained with antibodies against Rac GTPases, mDia2 or mDia1. The scale bar represents 5 m. (g) Inhibition of Rac GTPases disrupts mDia2 localization at the CAR. TER119-negative mouse fetal erythroblasts were cultured as in Fig. 1. After 35 h in culture, 100 M NSC23766 was added to some cells and after 48 h in culture the erythroblasts were stained with Alexa Fluor 488–phalloidin for actin and antibodies against mDia2. The scale bar represents 7 m.

Full size image (97 KB)

An interaction between the mDia2 GBD domain and the GTP form of Rac1 and Rac2 was observed, but did not establish that this interaction is direct, or that no other proteins facilitate this interaction (Fig. 3c–e). GST-pulldown experiments using mDia2 GBD (amino acids 1–280) produced in transfected 293T cells, showed that both Rac1 and Rac2 bind to the GBD of mDia2 in a GTP-dependent manner, and that there is no interaction between Rac GTPases and the mDia1 GBD (amino acids 1–300; Fig. 3c). However, full-length mDia2 binding to Rac GTPases could not be detected using the same GST-pulldown assay. In this system, full-length mDia2 also showed no interaction with RhoA, a well-established binding protein of mDia2 (data not shown). Indeed, similar observations were also documented between Rif and mDia2 interaction¹⁹. We reasoned that this could be because of the auto-inhibitory structure of mDia2 protein; therefore we generated different mutants of mDia2 and determined their binding efficiency with Rac GTPases. As shown in Fig. 3d, both Rac1 and Rac2 bound to mDia2 GBD and DAD (amino acids 1–1030) mutants, but not a GBD (amino acids 280–1171) mutant. Their binding efficiencies were comparable to that of RhoA, but were much stronger than between Cdc42 and either mDia2 GBD or the DAD mutants. A GST-pulldown experiment using mDia2 GBD purified from bacterial cells was also performed and showed that Rac1 and Rac2 bind to mDia2 GBD in a GTP-dependent manner (Fig. 3e). However, this binding was significantly less efficient than the interactions between RhoA and the mDia2 GBD (Fig. 3e). This suggests that the interaction between Rac GTPase and mDia2 could be indirect and that other unidentified proteins may facilitate the Rac–mDia2 interaction in vivo.

The Rac–mDia2 interaction was further examined using immunofluorescence microscopy after staining the cells for endogenous Rac GTPases and mDia2. The staining showed that Rac GTPases colocalized with mDia2, but not mDia1, on the plasma membrane where the CAR was located (Fig. 3f and 3g top panel). Treatment with 100 M NSC23766 after 35 h in culture blocked CAR formation and the localization of mDia2 at the CAR (Fig. 3g). Taken together, these results indicate that mDia2 mediates the effects of Rac1 and Rac2 on enucleation.

To directly investigate the function of mDia1 and mDia2 on mouse fetal erythroblast enucleation, we infected isolated TER119-negative mouse fetal erythroid progenitors with retroviruses encoding short hairpin RNA (shRNA) targeting mDia1 or mDia2 mRNA. A scrambled shRNA sequence was used as control. A maximum decrease of mDia1 and mDia2 mRNA levels was achieved after 30 h in culture, approximately coincident with the start of enucleation (data not shown). Western blots showed a dramatic reduction in mDia1 and mDia2 protein levels after 48 h in culture (Fig. 4a). Flow cytometric analysis after 2 days of culture showed that downregulation of mDia2 by shRNA dramatically inhibited enucleation, whereas there was no obvious effect of mDia1 downregulation (Fig. 4b, c). The differentiation, proliferation and apoptotic status of mDia2 shRNA-infected cells were similar to those of control shRNA infected cells and cells expressing mDia1 shRNA (Figs 4b, d and see Supplementary Information, Fig. S4). Benzidine–Geimsa staining of mDia2 shRNA-infected cells showed late-stage erythroblasts with condensed nuclei located close to the plasma membrane, similar to cells expressing deregulated Rac GTPases (Figs 2d and 4e). In addition, many partially enucleated cells with nuclei halfway extruded from the cytoplasm were evident (Fig. 4e). Binucleated cells were also observed (Fig. 4e), which suggests that mDia2 is involved in cytokinesis of erythroblasts at earlier stages in development. As expected, Alexa Fluor 488–phalloidin staining of the mDia2 shRNA-infected cells after 2 days in culture illustrated defects in CAR formation that were not observed in mDia1 shRNA infected cells (Fig. 4f, g).

Figure 4: Depletion of mDia2 inhibits enucleation.

(a) Depletion of mDia1 and mDia2 in primary erythroid cells by shRNA. TER119-negative mouse fetal erythroblasts were infected with retrovirus vectors encoding an shRNA specific for mDia1 or mDia2, and cultured for 48 h. Lysates were subjected to western blotting for mDia1 or mDia2. Controls show GAPDH levels. (b) Flow cytometric analysis of cells infected with retroviruses encoding mDia1 or mDia2 shRNAs and cultured for 2 days, as in Fig. 1. Cells were stained with TER119–PE, FITC–CD71 and Hoechst 33342 and analysed by FACS. The percentages of R8 cells (reticulocytes) and TER119-negative and -positive cells are indicated. (c, d) Quantification of incipient reticulocytes in cells infected with indicated retrovirus (c) and the number of these cells (d). As in Fig. 2d, the total cell number is shown as grey bars and the normalized number as black bars. The error bars represent mean s.d. (n = 3). (e) Benzidine–Giemsa staining of 2 day-cultured erythroblasts infected with retrovirus expressing mDia2 shRNA. The arrowheads indicate cells blocked in the process of enucleation and the arrow indicates a binucleated cell. (f) Alexa Fluor 488–phalloidin staining of 2 day-cultured erythroblasts infected with retrovirus expressing mDia1 or mDia2 shRNA. The arrowheads indicate CAR in late-stage erythroblasts. The arrow indicates an incipient reticulocyte. The scale bar represents 6 m. (g) Quantification of cells containing a CAR from f. The error bars represent mean s.d. (n = 3).

Full size image (65 KB)

We also directly showed that mDia2 lies downstream of Rac1 and Rac2 in triggering enucleation (Fig. 5). Isolated TER119-negative mouse fetal erythroid progenitors were coinfected with retroviruses encoding a dominant-negative mutant of Rac GTPase and a constitutively active mutants of mDia2 or mDia1 with the GBD deleted (mDia2^GBD, amino acids 280–1171; mDia1^GBD, amino acids 300–1255)^{20, 21}. Overexpression of mDia2^GBD or mDia1^GBD alone had no effect on enucleation, compared with control cells. This is in contrast to the enucleation defect induced by constitutively active Rac GTPases (Fig. 2b); note that mDia proteins are different from Rac GTPases in that their activities are not directly controlled by intracellular GTP levels. Significantly, overexpression of mDia2^GBD, but not mDia1^GBD, rescued the defects of enucleation induced by dominant-negative Rac1 and Rac2. Lysates from the cells were subjected to western blot analysis to demonstrate that the levels of ectopic protein expression observed in singly infected and coinfected cells were the same (Fig. 5). Immunofluorescence microscopy demonstrated that the overexpressed mDia2^GBD partially colocalized with the CAR, but the majority of the proteins were located with the extruded nucleus (see Supplementary Information, Fig. S5).

Figure 5: Constitutively active mutant of mDia2 rescues the defects of enucleation induced by dominant-negative mutants of Rac GTPases.

TER119-negative mouse fetal erythroblasts were infected with bicistronic retroviruses encoding hCD4 and the indicated proteins (groups 2, 3, 4 and 7), or coinfected with retroviruses encoding hCD4 and mutant Rac proteins together with retroviruses encoding hCD4 and mDia2^GBD (groups 5 and 6) or mDia1^GBD (groups 8 and 9). Cells in the control group (group 1) were infected with retroviruses encoding hCD4 alone. Cells were then analysed by flow cytometric analysis after 48 h in culture and the percentages of incipient reticulocytes were quantified. The grey bars indicate the uninfected hCD4-negative cells that provide an internal control, and the black bars indicate hCD4 positive cells expressing different proteins as indicated. Lysates from the cells were subjected to western blot analysis to demonstrate that the levels of ectopic protein expression were unaltered in coinfected cells. mDia2^GBD was detected by polyclonal antibody against mDia2. mDia1^GBD was detected by monoclonal antibody against mDia1. Rac1 and Rac2 mutants were detected by anti-Flag antibody. The error bars represent mean s.d. (n = 3). P values were calculated using t-test.

Full size image (76 KB)

Our studies on the role of Rac GTPases on enucleation were made possible by a recently developed in vitro culture system in which large numbers of erythroid progenitor cells from mouse fetal livers can easily be isolated, cultured and developmentally synchronized to study their terminal proliferation and differentiation in a step-by-step manner¹². In this way, the enucleation process is properly dissected from erythropoiesis, which is difficult to achieve in vivo using gene-altered animal models. Our results demonstrate that Rac GTPases and their effector protein mDia2 play significant roles in mouse fetal erythroblast enucleation by affecting the formation of the CAR in late-stage erythroblasts. Previous studies suggested that enucleation is a type of cytokinesis in which various membrane skeleton network proteins are involved^{9, 10, 13, 14}. A recent study using Rac1–Rac2 double-knockout mice showed that Rac GTPases are important for the morphology and deformability of the erythrocyte cytoskeleton²². Therefore, our results are consistent with the role of Rac GTPases in the integrity of the erythroid cell actin cytoskeleton network.

Enucleation of erythroblasts was a crucial step in mammalian evolution, as eliminating the inactive nuclei from erythrocytes allowed an increase in haemoglobin concentration in the blood. As Rac GTPases and mDia2 affect actin polymerization in many body cells, evolution co-opted these proteins for erythroblast enucleation. We aim to identify all the proteins that participate in and regulate this asymmetric cell-division process involving extrusion of a plasma-membrane-enveloped pycnotic nucleus from the reticulocyte cytoplasm, as well as all the proteins that cause chromatin condensation and transcriptional inactivation prior to enucleation.

Acknowledgements

We thank: A. Alberts for critical reading of the manuscript; S. Lux for helpful comments; J. Zhang for help with the in vitro culture system; G. Pardis for help with flow cytometry. This study was supported by National Institutes of Health (NIH) grant (P01 HL 32262) and a research grant from Amgen, Inc. to H.F.L.

Received 17 October 2007; Accepted 23 January 2008; Published online 10 February 2008.

References

1. Richmond, T. D., Chohan, M. & Barber, D. L. Turning cells red: signal transduction mediated by erythropoietin. Trends Cell Biol. 15, 146–155 (2005). | Article | PubMed | ISI | ChemPort |
2. Koury, M. J., Sawyer, S. T. & Brandt, S. J. New insights into erythropoiesis. Curr. Opin. Hematol. 9, 93–100 (2002). | Article | PubMed | ISI |
3. Ihle, J. N. & Gilliland, D. G. Jak2: normal function and role in hematopoietic disorders. Curr. Opin. Genet. Dev. 17, 8–14 (2007). | Article | PubMed | ISI | ChemPort |
4. Faix, J. & Grosse, R. Staying in shape with formins. Dev. Cell 10, 693–706 (2006). | Article | PubMed | ChemPort |
5. Goode, B. L. & Eck, M. J. Mechanism and function of formins in the control of actin assembly. Annu. Rev. Biochem. 76, 593–627 (2007). | Article | PubMed | ChemPort |
6. Zigmond, S. H. Formin-induced nucleation of actin filaments. Curr. Opin. Cell Biol. 16, 99–105 (2004). | Article | PubMed | ISI | ChemPort |
7. Koury, M. J. & Bondurant, M. C. Maintenance by erythropoietin of viability and maturation of murine erythroid precursor cells. J. Cell Physiol. 137, 65–74 (1988). | Article | PubMed | ISI | ChemPort |
8. Eshghi, S., Vogelezang, M. G., Hynes, R. O., Griffith, L. G. & Lodish, H. F. Alpha4beta1 integrin and erythropoietin mediate temporally distinct steps in erythropoiesis: integrins in red cell development. J. Cell Biol. 177, 871–880 (2007). | Article | PubMed | ChemPort |
9. Simpson, C. F. & Kling, J. M. The mechanism of denucleation in circulating erythroblasts. J. Cell Biol. 35, 237–245 (1967). | Article | PubMed | ChemPort |
10. Skutelsky, E. & Danon, D. An electron microscopic study of nuclear elimination from the late erythroblast. J. Cell Biol. 33, 625–635 (1967). | Article | PubMed | ChemPort |
11. Repasky, E. A. & Eckert, B. S. A reevaluation of the process of enucleation in mammalian erythroid cells. Prog. Clin. Biol. Res. 55, 679–692 (1981). | PubMed | ChemPort |
12. Zhang, J., Socolovsky, M., Gross, A. W. & Lodish, H. F. Role of Ras signaling in erythroid differentiation of mouse fetal liver cells: functional analysis by a flow cytometry-based novel culture system. Blood 102, 3938–3946 (2003). | Article | PubMed | ChemPort |
13. Koury, S. T., Koury, M. J. & Bondurant, M. C. Cytoskeletal distribution and function during the maturation and enucleation of mammalian erythroblasts. J. Cell Biol. 109, 3005–3013 (1989). | Article | PubMed | ISI | ChemPort |
14. Chasis, J. A., Prenant, M., Leung, A. & Mohandas, N. Membrane assembly and remodeling during reticulocyte maturation. Blood 74, 1112–1120 (1989). | PubMed | ISI | ChemPort |
15. Luo, L. Rho GTPases in neuronal morphogenesis. Nature Rev. Neurosci. 1, 173–180 (2000). | Article |
16. Nassar, N., Cancelas, J., Zheng, J., Williams, D. A. & Zheng, Y. Structure-function based design of small molecule inhibitors targeting Rho family GTPases. Curr. Top. Med. Chem. 6, 1109–1116 (2006). | Article | PubMed | ChemPort |
17. Akbar, H., Cancelas, J., Williams, D. A., Zheng, J. & Zheng, Y. Rational design and applications of a Rac GTPase-specific small molecule inhibitor. Methods Enzymol. 406, 554–565 (2006). | Article | PubMed | ChemPort |
18. Higgs, H. N. Formin proteins: a domain-based approach. Trends Biochem. Sci. 30, 342–353 (2005). | Article | PubMed | ISI | ChemPort |
19. Pellegrin, S. & Mellor, H. The Rho family GTPase Rif induces filopodia through mDia2. Curr. Biol. 15, 129–133 (2005). | Article | PubMed | ISI | ChemPort |
20. Destaing, O. et al. A novel Rho–mDia2–HDAC6 pathway controls podosome patterning through microtubule acetylation in osteoclasts. J. Cell Sci. 118, 2901–2911 (2005). | Article | PubMed | ISI | ChemPort |
21. Palazzo, A. F., Cook, T. A., Alberts, A. S. & Gundersen, G. G. mDia mediates Rho-regulated formation and orientation of stable microtubules. Nature Cell Biol. 3, 723–729 (2001). | Article |
22. Kalfa, T. A. et al. Rac GTPases regulate the morphology and deformability of the erythrocyte cytoskeleton. Blood 108, 3637–3645 (2006). | Article | PubMed | ChemPort |
23. Ghaffari, S. et al. AKT induces erythroid-cell maturation of JAK2-deficient fetal liver progenitor cells and is required for Epo regulation of erythroid-cell differentiation. Blood 107, 1888–1891 (2006). | Article | PubMed | ISI | ChemPort |
24. Watanabe, N., Kato, T., Fujita, A., Ishizaki, T. & Narumiya, S. Cooperation between mDia1 and ROCK in Rho-induced actin reorganization. Nature Cell Biol. 1, 136–143 (1999). | Article |
25. Yuan, B., Latek, R., Hossbach, M., Tuschl, T. & Lewitter, F. siRNA Selection Server: an automated siRNA oligonucleotide prediction server. Nucleic Acids Res. 32, W130–W134 (2004). | Article | PubMed | ISI | ChemPort |
26. Luo, B., Heard, A. D. & Lodish, H. F. Small interfering RNA production by enzymatic engineering of DNA (SPEED). Proc. Natl Acad. Sci. USA 101, 5494–5499 (2004). | Article | PubMed | ChemPort |

1. Whitehead Institute for Biomedical Research, Department of Biology, Massachusetts Institute of Technology, 9 Cambridge Center, Floor 6, Cambridge, MA 02142, USA.
2. Current address: Stem Cell Group 4, Genome Institute of Singapore #08-01, Genome, 60 Biopolis Street, Singapore 138672.

Correspondence to: Harvey F. Lodish¹ e-mail: lodish@wi.mit.edu

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated.

NPG Journals

by Subject Area

• Chemistry

  • Chemistry
  • Drug discovery
  • Biotechnology
  • Materials
  • Methods & Protocols

• Clinical Practice & Research

  • Cancer
  • Cardiovascular medicine
  • Dentistry
  • Endocrinology
  • Gastroenterology & Hepatology
  • Methods & Protocols
  • Pathology & Pathobiology
  • Urology

• Earth & Environment

  • Earth sciences
  • Evolution & Ecology

• Life sciences

  • Biotechnology
  • Cancer
  • Development
  • Drug discovery
  • Evolution & Ecology
  • Genetics
  • Immunology
  • Medical research
  • Methods & Protocols
  • Microbiology
  • Molecular cell biology
  • Neuroscience
  • Pharmacology
  • Systems biology

• Physical sciences

  • Physics
  • Materials

by A - Z Index