Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Sep 22;100(20):11451–11456. doi: 10.1073/pnas.1635050100

Combinatorial interactions of Serpent, Lozenge, and U-shaped regulate crystal cell lineage commitment during Drosophila hematopoiesis

Nancy Fossett *,†, Kristy Hyman *, Kathleen Gajewski *, Stuart H Orkin ‡, Robert A Schulz *,§
PMCID: PMC208778  PMID: 14504400

Abstract

The GATA factor Serpent (Srp) is required for hemocyte precursor formation during Drosophila hematopoiesis. These blood cell progenitors give rise to two distinct lineages within the developing embryo. Lozenge, a Runx protein homologue, and Glial cells missing-1 and -2 are essential for crystal cell and plasmatocyte production, respectively. In contrast U-shaped, a Friend of GATA class factor, antagonizes crystal cell formation. Here we show that Srp, Lozenge, and U-shaped interact in different combinations to regulate crystal cell lineage commitment. Coexpression of Srp and Lozenge synergistically activated the crystal cell program in both embryonic and larval stages. Furthermore, expression of Lozenge and SrpNC, a Srp isoform with N- and C-terminal zinc fingers, inhibited u-shaped expression, indicating that crystal cell activation coincided with the down-regulation of this repressor-encoding gene. In contrast, whereas U-shaped and SrpNC together blocked crystal cell production, coexpression of U-shaped with noninteracting Srp proteins failed to prevent overproduction of this hemocyte population. Such results indicated that U-shaped and SrpNC must interact to block crystal cell production. Taken together, these studies show that the specialized SrpNC isoform plays a pivotal role during crystal cell lineage commitment, acting as an activator or repressor depending on the availability of specific transcriptional coregulators. These findings provide definitive proof of the combinatorial regulation of hematopoiesis in Drosophila and an in vivo demonstration of GATA and Runx functional interaction in a blood cell commitment program.

Vertebrate blood cells develop from pluripotent stem cells that are capable of both self-renewal and sequential commitment to divergent lineages of increasingly restricted developmental potential. A number of studies suggest that combinatorial interactions between hematopoietic factors regulate lineage selection. For example, Friend of GATA (FOG)-1 converts GATA-1 from an activator to a repressor of eosinophil lineage commitment (1). In addition, this complex is required for erythrocyte and megakaryocyte differentiation (2–4). FOG-1 may also act with GATA-2 to initiate megakaryopoiesis (5). In general, FOG proteins bind the N-terminal zinc-finger of their respective GATA partners. This interaction modifies GATA-regulated transcription in a variety of tissues and in taxa ranging from flies to humans (2–14). Hematopoietic coregulators of GATA factors may include members of the Runx family. Proteins from both families are expressed and function during megakaryopoiesis. Inherited mutations in either GATA-1 or Runx1 are associated with congenital thromobocytopenia (15–21). Furthermore, in vitro studies suggest that GATA-1:Runx1 complex formation regulates megakaryocyte differentiation (22). Studies in Drosophila have identified hematopoietic functions for homologues of GATA, FOG, and Runx class genes (23–26). However, combinatorial regulation of lineage commitment in the fly has not been demonstrated.

Drosophila has a rudimentary hematopoietic system with two blood cell types, crystal cells and plasmatocytes. During embryogenesis these cells develop from hemocyte precursors derived from the cephalic mesoderm (25, 27, 28). The lymph gland is the site of larval hematopoiesis, producing the two primary lineages and a third cell type, called the lamellocyte (25, 27–29). Together, these cells provide the organism with the capacity for wound healing, immune response, and removal of apoptotic cells. Crystal cells, named for their crystalline inclusion bodies, repair cuticle lacerations and encapsulate foreign invaders (28–30). Plasmatocytes, which comprise 95% of circulating hemoctyes, synthesize antimicrobial peptides and differentiate into macrophage-like cells that phagocytize microbes and apoptotic cells (28, 29). Lamellocytes mount a response against invaders that are too large to be phagocytized (29, 31, 32).

The GATA factor Serpent (Srp) is required for the production of embryonic hemocyte precursors that give rise to both primary blood cell lineages (23, 24). Expression of srp in the larval lymph glands suggests that its role is reiterated during larval development. However, this function has not been demonstrated because of the embryonic lethality of null mutations (25). Srp acts upstream of the lineage-specific transcription factors Lozenge (Lz), a Runx homologue, Glial cells missing (Gcm), and the FOG family member U-shaped (Ush) (25, 26). Gcm-1 and -2 direct plasmatocyte production (33, 34). Together with Srp, the Notch–Serrate signaling pathway up-regulates Lz in hemocyte precursors, thereby promoting crystal cell lineage commitment (25, 35, 36). In contrast, Ush antagonizes crystal cell production (26). In this report we show that Srp, Lz, and Ush interact to regulate crystal cell production. Specifically, Srp and Lz synergistically activated crystal cell lineage commitment. In addition, SrpNC, the Srp isoform with N- and C-terminal zinc fingers, and Lz functioned together to inhibit ush gene expression. Furthermore, Ush repression of crystal cell production depended on binding to its GATA partner SrpNC. These data show that SrpNC acts as an activator or repressor of both crystal cell lineage commitment and ush expression, depending on the availability of specific coregulators. Moreover, these studies provide an in vivo demonstration of opposing cross talk between regulatory pathways. Together, these findings document essential genetic interactions during Drosophila hematopoiesis and provide an in vivo demonstration of GATA:Runx synergism as required for specialized hemocyte lineage commitment.

Materials and Methods

Fly Strains. y w67c23 served as our wild-type stock. The upstream activation sequence (UAS)-ush and twi-Gal4 strains used in this study are described elsewhere (37). The following strains were generous gifts from colleagues: UAS-lz (J. Canon and U. Banerjee, University of California, Los Angeles); UAS-srpC (D. Hoshizaki, University of Nevada, Las Vegas; SrpC, Srp isoform with a single C-terminal zinc finger); e33c-Gal4 (N. Perrimon, Harvard Medical School). srpNC cDNA was generated by using RT-PCR with the following primer set: 5′ SrpI, AAAACTCGAGTCATACGCCTACCCAGCTCTACGAG; 3′ Srp2, TTTGTCTAGAGATAGATTTTAGATAGATGAGTGTTCTTTG. The PCR product was cloned into the pCRII vector (Invitrogen). The srpNCV421G mutant was produced by using the Stratagene Qwik Change kit, the srpNC cDNA template, and the following primer set: 5′ SrpVGa, GGGCGAGAATGTGGCAATTGTGGTG; 3′ SrpVGb, CACCACAATTGCCACAT TCTCGCCC. To generate UAS-srpNC and UAS-srpNCV421G transgenic fly strains, the cDNA for srpNC and srpNCV421G were cloned into the P element vector pUAST and injected into y w67c23 embryos. Multiple transgenic lines were established by using standard transformation procedures (38). Several lines were tested for each DNA and all showed a strong crystal cell phenotype.

Immunohistochemical Staining of Embryos and Larvae. Collection, fixation, and immuno-histochemical staining of embryos were performed as previously described (38). The U-shaped antibody has been described in Fossett et al. (26) and used at a 1:500 dilution for embryo staining. A polyclonal antibody, which recognizes a crystal cell-specific antigen, was produced by injecting either guinea pigs or rabbits with keyhole limpet hemocyanin conjugated to the synthetic peptide SrpNC 409 FDADYFTEGREC 420 (Alpha Diagnotics, San Antonio, TX). A 1:500 dilution was used for embryo staining. Rabbit anti-rat, goat anti-guinea pig, horse anti-mouse, and goat anti-rabbit secondary antibodies (Vector) were used at 1:1,000 dilution.

In Situ Hybridization of Embryos. Black cells (Bc) cDNA was generated by RT-PCR using the following primer set: 5′ Bc1, 5′ CACGGATCTGAAAGCCTTGG; 3′ Bc2, TTGCCCAGGATATCGATGCC. The PCR product was cloned into the pCRII vector, with clones used to produce cRNA antisense probes. The in situ hybridization was performed according to Gajewski et al. (38).

Gene Expression Analyses in Mutant and Gal4/UAS Embryos. Embryos and larvae were cultured and collected at 23°C. To visualize crystal cells in lymph glands, larvae were heat treated according to the method of Duvic et al. (35). Pan-mesodermal expression of the various transgenes was achieved by crossing UAS-srpC, UAS-srpNC, UAS-srpNCV421G, UAS-lz, or UAS-ush males to twi-Gal4 virgin females. Larval lymph gland transgene expression was achieved by crossing either UAS-srpNC or UAS-lz males to e33c-Gal4 virgin females. A two-generation cross was required to coexpress srp with either ush or lz. UAS-ush virgin females were crossed to UAS-srpC, UAS-srpNC, or UAS-srpNCV421G males. Subsequently, F1 males from each cross were crossed to twi-Gal4 virgin females to achieve coexpression of ush and srp. UAS-srpNC or UAS-srpC virgin females were crossed to UAS-lz males. Males from each cross were crossed to either twi-Gal4 or e33c-Gal4 virgin females to achieve mesodermal or larval lymph gland coexpression, respectively.

Results

Srp and Lz Synergistically Activate the Crystal Cell Program in Embryos and Larvae. Vertebrate GATA-1 and Runx1 are coexpressed and have overlapping functions during hematopoiesis, including megakaryopoiesis (15–22). In addition, in vitro studies have suggested that the GATA-1:Runx1 complex directs megakaryocyte differentiation (22). Furthermore, their respective Drosophila homologues, Srp and Lz, are coexpressed in the crystal cell lineage (25). These observations suggested that Srp and Lz might interact to promote crystal cell lineage commitment. To test this hypothesis, we assayed for crystal cell production in embryos with coexpression of UAS-srp and UAS-lz, and compared these results to those in which each transgene was expressed separately. We used the UAS/Gal4 system (39) with the twi-Gal4 driver to express the transgenes pan-mesodermally. Crystal cell production was monitored by using the Bc transcript and an antibody to a crystal cell-specific antigen (N.F. and R.A.S., unpublished data).

Expression of either UAS-srp or UAS-lz produced supernumerary crystal cells that persisted throughout embryogenesis (Fig. 1 C, G, H, L, M, Q, and R). Despite pan-mesodermal expression of these proteins, cell production was constrained in both genetic backgrounds. Forced expression of SrpNC produced patterns of Bc transcript and crystal cell antigen expression that were both temporally and spatially restricted. As a result, crystal cell markers were first detected at the same developmental stages in both twi-Gal4;UAS-srpNC and control embryos (Fig. 1 A, B, F, G, K, L, P, and Q). In addition, we observed that expanded crystal cell production was confined to the head region throughout embryogenesis by using either marker (Fig. 1 G, L, and Q, and data not shown). Similar results were obtained with twi-Gal4-driven UAS-srpC (data not shown).

Fig. 1.

Fig. 1.

Open in a new tab

Coexpression of Srp and Lz enhances crystal cell overproduction. The crystal cell population was monitored in different genetic backgrounds and in distinct developmental stages by Bc mRNA in situ hybridization or crystal cell antigen detection. (A–E) Ventral views of stage 10 embryos. (F–T) Dorsal views of stage 11–12 (F–J), stage 13 (K–O), and stage 15–16 (P–T) embryos. Bc mRNA in situ hybridization of wild type (A, F, K, and P), twi-Gal4;UAS-srpNC (B, G, L, and Q), twi-Gal4;UAS-lz (C, H, M, and R), and twi-Gal4; UAS-srpNC; UAS-lz (D, I, N, and S) embryos is shown. (E, J, O, and T) Immunohistochemical staining of twi-Gal4; UAS-srpNC; UAS-lz embryos with the crystal cell-specific antibody.

In contrast, ectopic Lz expression produced temporally and spatially altered Bc expression. In twi-Gal4;UAS-lz embryos, the transcript was detected earlier during stage 10 (Fig. 1 A and C). Furthermore, Bc was atypically expressed in the cephalic mesoderm in a pattern similar to endogenous srp expression (Fig. 1C). Interaction between ectopic Lz and endogenous Srp most likely explains the Bc up-regulation. From stage 11 through 13, Bc expression declined to near wild-type levels (Fig. 1 H and M). Expression rebounded during late embryogenesis, forming a stippled pattern throughout the mesoderm (Fig. 1R). The crystal cell-specific antigen was expressed in a similar pattern; however, it was not detected in the cephalic mesoderm in stage 10 embryos (data not shown).

Coexpression of Srp and Lz dramatically increased both Bc mRNA and crystal cell antigen levels when compared with either transgene alone (Fig. 1 B–D, G–I, L–N, and Q–S). Similar to twi-Gal4;UAS-lz embryos, the Bc transcript was first detected in the cephalic mesoderm of stage 10 embryos. However, expression was greater in embryos with coactivation of SrpNC and Lz when compared with activation of Lz alone (Fig. 1 C and D). By stage 12, coexpression of SrpNC and Lz produced panmesodermal activation of the Bc gene that persisted throughout embryogenesis (Fig. 1 I, N, and S). This result was significantly different from the restricted pattern observed in either twi-Gal4;UAS-srpNC or twi-Gal4;UAS-lz embryos. In addition, coactivated SrpC and Lz also produced pan-mesodermal Bc gene expression, indicating that either isoform could function with Lz (data not shown). Furthermore, crystal cell antigen detection was significantly greater in embryos with both active transgenes than in those with only one (Fig. 1 J, O, and T and data not shown)

The combinatorial transgene activity was clearly synergistic, and not additive, because the level of Bc and crystal cell antigen expression was considerably greater with coactivation of Srp and Lz when compared with either transgene acting alone. This was most evident during stages 11–13. During this interval, Bc was expressed throughout the mesoderm in Srp and Lz coactivated embryos. In contrast, expression was limited to the head region in embryos with single transgene expression (Fig. 1 G–I and L–N). Together, these data indicated that forced expression of Srp and Lz synergistically activated the crystal cell program.

During larval development, the lymph glands are the site of hematopoiesis. This tissue has two primary lobes and multiple pairs of secondary lobes. Genetic regulation of crystal cell production appears to be reiterated in larvae because srp is expressed throughout the cells of the lymph gland, whereas lz expression appears to be limited to a subset of srp expressing cells (25). We used the e33c-Gal4 lymph gland driver to test whether coexpression of SrpNC and Lz would synergistically activate larval crystal cell production as it had during embryogenesis. Larvae were heat treated to visualize crystal cells. Forced expression of SrpNC and Lz produced far greater numbers of crystal cells than expression of either transgene alone (Fig. 2). The findings in embryos and larvae indicate that Srp and Lz act together to promote crystal cell lineage commitment. Furthermore, the data indicate that Srp directly activates crystal cell lineage commitment in addition to its role in hemocyte precursor production.

Fig. 2.

Fig. 2.

Open in a new tab

Srp and Lz synergistically activate the crystal cell program in larval lymph glands. (A–D) Crystal cells detected in the primary lobes of third instar larval lymph glands as visualized after heat treatment. (A) Wild type. (B–D) e33c-Gal4-driven expression of SrpNC (B), Lz (C), and SrpNC and Lz transgenes (D).

SrpNC and Lz Repress ush Gene Expression. Ush is expressed with both Srp and Lz in hemocyte and crystal cell precursors (25, 26). In addition, Ush functions to repress crystal cell production (26). These data suggested that Srp and Lz activation of the crystal cell program involves repression of ush gene expression. We used a Ush peptide antibody to assay Ush expression in embryos with Srp and Lz transgenes expressed either separately or together. Pan-mesodermal expression of either SrpNC or SrpC produced uniform activation of ush throughout the embryo (Fig. 3C and data not shown). Lz alone, or in combination with SrpC, did not repress ush expression (Fig. 3 A and E and data not shown). In contrast, coactivation of SrpNC and Lz inhibited ush activity in the hemocyte precursors, during stage 10 and at the same time when the Bc gene was up-regulated in these cells (Fig. 3G). Furthermore, ush activity remained repressed in plasmatocytes throughout development (data not shown). These results suggested that the interaction of SrpNC and Lz either down-regulated the ush gene or reprogrammed hemocyte precursors to produce crystal cells in lieu of plasmatocytes. To distinguish between these possibilities, we assayed hemocyte precursor gene expression and plasmatocyte production in twi-Gal4;UAS-srpNC;UAS-lz embryos. Gcm is required for plasmatocyte production and is expressed in hemocyte precursors (33). When we used the gcmlacZ enhancer-reporter fusion as a marker, we observed gcm activity in hemocyte precursors and plasmatocytes in embryos regardless of the genetic background (Fig. 3 B, D, F, and H). These results indicated that the loss of ush activity was caused by inhibition of gene expression and not a change in cell fate. Finally, forced expression of Lz in the presence of endogenous SrpNC did not inhibit ush expression. Thus, down-regulation of ush by the combinatorial action of SrpNC and Lz appears to depend on a threshold concentration of SrpNC.

Fig. 3.

Fig. 3.

Open in a new tab

Coexpression of Srp and Lz represses Ush expression in the cephalic mesoderm. Ventral views of stage 10 embryos are presented. (A, C, E, and G) Embryos stained with an αUsh antibody. (B, C, F, and H) gcmlacZ embryos stained for β-galactosidase activity. (A and B) Wild type. (C–H) twi-Gal4-driven expression of SrpNC (C and D), Lz (E and F), and SrpNC and Lz (G and H). Solid arrows indicate normal Ush expression; open arrow highlights the absence of Ush protein.

SrpNC and Ush Interact to Block Crystal Cell Production. Repression of ush expression by the combinatorial action of SrpNC and Lz may inhibit the formation of a potential SrpNC:Ush repressor complex, thereby enhancing crystal cell production. To test whether Ush and SrpNC interact to block crystal cell production, we coexpressed Ush with SrpNC, SrpC, or a mutant version of SrpNC that does not bind Ush. The transcripts used to produce these proteins are shown in Fig. 4. SrpC is spliced to exclude the N-terminal zinc-finger, rendering it incapable of interacting with Ush. SrpNC is spliced to juxtapose the N- and C-terminal zinc-fingers, producing a protein with the characteristic GATA structure (40). The amino acid sequence of the GATA N-terminal zinc-finger is highly conserved (8). In addition, 7 of 25 amino acids required for GATA:FOG complex formation are all conserved between SrpNC and the GATA consensus sequence (8, 40).

Fig. 4.

Fig. 4.

Open in a new tab

Diagram of srp transgenes. Czf, exon encoding the C-terminal zinc finger; Nzf, exon encoding the N-terminal zinc finger. V421G indicates the SrpNC mutation changing valine 421 to glycine. Amino acids with asterisks are required for GATA–FOG interaction. The amino acid change in SrpNCV421G is indicated in red.

We produced a UAS-srpNC mutant strain in which amino acid residue valine 421 was mutated to glycine. A number of studies have shown that this residue is required for GATA:FOG complex formation. For example, an inherited valine to methione substitution in human GATA-1 precludes FOG-1 binding, resulting in familial dyserythropoietic anemia and thrombocytopenia (16). In addition, valine to glycine substituted GATA-4 knock-in mice exhibit heart defects similar to FOG-2 null mutant embryos (14). Finally, this conserved valine was among the seven residues shown to be essential for GATA:FOG binding (8). We used the twi-Gal4 driver to coexpress these Srp proteins with Ush and monitored crystal cell production by using Bc transcript and crystal cell-specific antigen expression.

We have previously demonstrated that the crystal cell population increases in ush null embryos and misexpression of Ush or FOG related proteins within these cells inhibits their production (26). Compared with wild-type controls, mesodermal expression of Ush reduced expression of both crystal cell markers (Fig. 5 A–D), consistent with our previous findings. Furthermore, coexpression of SrpNC and Ush blocked expression of both markers, indicating that crystal cell production was completely repressed (Fig. 5 E and F). Inhibition of crystal cell production was greater in embryos expressing both proteins than in those expressing Ush alone (Fig. 5 C and D). Finally, SrpNC lost the capacity to produce supernumerary crystal cells when coexpressed with Ush. To determine whether Ush and SrpNC interact to repress crystal cell production, we coexpressed Ush with the noninteracting Srp proteins, SrpC and SrpNCV421G, and assayed crystal cell formation. Coactivation of Ush with either of the nonbinding Srp proteins had no effect on the production of supernumerary crystal cells (Fig. 5 G–J), indicating that Ush must interact with a GATA-binding partner to inhibit crystal cell production.

Fig. 5.

Fig. 5.

Open in a new tab

Coexpression of SrpNC and Ush blocks crystal cell production. Dorsal views of stage 13 embryos are presented. (A, C, E, G, and I) In situ hybridization to detect Bc mRNA. (B, D, F, H, and J) Embryos stained with the crystal cell antigen-specific antibody. (A and B) Wild type. (C–J) twi-Gal4-driven expression of Ush (C and D), coexpression of SrpNC and Ush (E and F), coexpression of SrpNCV421G and Ush (G and H), and coexpression of SrpC and Ush (I and J).

It has been reported that lz-Gal4-driven Ush, SrpNC, or SrpC expression inhibited crystal cell production. This finding implies that Ush functions independently of SrpNC in this capacity (40). However, when inhibition of crystal cell production by the lz-Gal4 driver acting alone was accounted for, coexpression of Ush and SrpNC reduced cell number (data not shown). In contrast, neither SrpNC nor SrpC expressed alone inhibited production (data not shown), consistent with the results we obtained with the twi-Gal4 driver. Together, these results indicate that Ush limits crystal cell production only when bound to its GATA partner SrpNC and suggest that Ush binding converts SrpNC from an activator to a repressor of crystal cell development.

Discussion

In this report, we demonstrate that combinatorial interactions between Srp, Lz, and Ush regulate crystal cell lineage commitment. Coexpression of Srp and Lz synergistically activated the crystal cell program, elevating marker expression to a level significantly greater than that induced by either transgene alone. Our results indicate that, although Lz is necessary for crystal cell production, it is not sufficient, because lineage activation requires interaction with Srp. These findings represent an important in vivo demonstration of GATA and Runx factor synergism in hemocyte production. Moreover, the role of Srp in crystal cell production is not limited to hemocyte precursor specification and activation of the lz gene, but rather it is also directly involved in lineage activation. Coincident expression of SrpNC and Lz inhibited Ush expression, indicating that elimination of this crystal cell repressor contributes to lineage activation by these factors. In contrast, coexpression of SrpNC with Ush blocked crystal cell production. However, coexpression of Ush with either nonbinding Srp protein did not decrease supernumerary cell production, indicating that Ush repressor function depends on SrpNC binding.

Together with recent studies, our current findings have identified an integrated network of multifactor interactions that directs crystal cell production (Fig. 6). Extrinsic signals from the Notch–Serrate pathway act with Srp to up-regulate Lz expression (25, 36). Srp and Lz then interact to promote lineage commitment, which includes down-regulating the Ush repressor component. This reduces the amount of Ush available for repressor complex formation, allowing SrpNC to act with Lz to drive crystal cell lineage commitment (Fig. 6A). Ush is activated by Srp (26) and then binds SrpNC, converting it from an activator to a repressor of crystal cell production (Fig. 6B). As an end result, a precise number of crystal cells emerge from the larger hemocyte precursor population based on the overall positive or negative combination of these vital hematopoietic factors.

Fig. 6.

Fig. 6.

Open in a new tab

Model of the combinatorial control of crystal cell lineage commitment. (A) Activation of the crystal cell program. Srp acts with the Notch–Serrate signaling pathway to up-regulate lz gene expression. Lz and SrpNC then interact to promote crystal cell lineage commitment and down-regulate the repressor, Ush. (B) Repression of crystal cell lineage commitment. Srp activates ush gene expression. Ush then binds SrpNC, converting it from an activator to a repressor of crystal cell production. HP, hemocyte precursors; CC, crystal cells.

SrpNC thus plays a pivotal role in crystal cell lineage commitment functioning as an activator in combination with Lz or a repressor when bound to Ush. In a broader context, these results show that the function of hematopoietic factors can be altered by changes in cofactor association. Moreover, they show that a single factor, in this case SrpNC, can act as an intermediary between competing regulatory pathways. Finally, these results provide an in vivo demonstration of the dynamic interactions between hematopoietic factors and the influence these interactions have on lineage selection.

These findings also highlight striking similarities between the genetic regulation of hemocyte precursor maturation in the Drosophila embryo and the development of erythrocytes and megakaryocytes from the erythroid progenitor in vertebrates. In both systems, GATA factors are expressed in bipotential precursor cells. As hematopoiesis progresses, GATA expression is maintained and its function is required in both descendent lineages: GATA-1 in erythrocytes and megakaryocytes, and Srp in plasmatocytes and crystal cells (7, 15–24, 41). In contrast, Runx function is restricted to one lineage: Runx1 is restricted to the megakaryocyte lineage and Lz is restricted to the crystal cell lineage (22, 25). In both systems, GATA and Runx synergistically activate gene expression in the Runx-restricted lineage, and cooperation between these proteins is thought to promote lineage commitment. GATA-1 and Runx1 interact physically to activate αIIb integrin gene expression. The conserved GATA zinc fingers are not required for GATA-1:Runx1 binding, whereas the poorly conserved N- and C-terminal domains of GATA-1 are essential for interacting with the Runt domain. GATA-1:Runx1 binding and gene activation is enhanced by core binding factor β (22). Whether Lz and Srp form a complex with or without the aid of cofactors must still be determined. However, this would not be surprising, because it is likely that additional factors act with Srp and Lz to program crystal cells because synergistic cell production was delayed until sometime after the initial coactivation of Srp and Lz. Considering the overlap of GATA-2 and Runx1 function during definitive hematopoiesis, the identification of GATA and Runx combinatorial hematopoietic gene activation in vertebrates and Drosophila may have broader implications for the regulation of blood development.

Another class of GATA cofactors is the FOG family of multitype zinc-finger proteins. A number of studies indicate FOG must physically interact with the N-terminal zinc finger of GATA partners to function during development. For example, a FOG-1 mutant incapable of binding GATA-1 failed to rescue erythroid or megakaryocytic maturation in a FOG null cell line (8). In addition, recent studies have suggested that FOG-1 can interact with either GATA-1 or GATA-2 to promote megakaryopoiesis, arguing against the previous notion that FOG-1 functions independently in this process (5). Furthermore, in Drosophila, Ush repression of heart and sensory bristle development apparently requires the GATA binding partner Pannier (6, 37). Unequivocal examples of FOG-independent functions have not been found. Although forced expression of Srp proteins produced supernumerary crystal cells, coexpression of Ush and SrpNC completely blocked lineage commitment. However, Ush did not inhibit supernumerary cell production by the nonbinding Srp proteins, SrpC and SrpNCV421G, consistent with the interdependency between FOG function and FOG: GATA complex formation.

Multipotent precursors express a number of lineage-specific factors before lineage commitment, allowing them to remain poised for a variety of cell fate choices. A change in the concentration of these factors can alter the composition and abundance of specialized regulatory complexes, resulting in alternative lineage specification. This creates a dynamic system that responds to a variety of external inputs by varying the populations of cell types as needed to maintain the vitality of the organism. This report is a clear demonstration of the combinatorial regulation of Drosophila hematopoiesis and supports these models of blood cell development. Furthermore, it provides a framework for future investigation of multifactor complex control of hemocyte lineage commitment.

Acknowledgments

We gratefully acknowledge J. Canon and U. Banerjee for sending the UAS-lz transgenic line before publication. We also thank D. Hoshizaki for UAS-srpC and N. Perrimon for e33c-Gal4 fly strains, and L. McCord for help with figures. In addition, we are grateful to M. Russell and D. Stegemeier for technical assistance. DNA sequences were determined by the M. D. Anderson Core Sequencing Facility supported by National Cancer Institute Grant CA16672. This work was supported by grants from the National Institutes of Health (to R.A.S.), the American Heart Association (to N.F. and R.A.S.), and the M. D. Anderson Cancer Center (to N.F.).

Abbreviations: Bc, Black cells; FOG, Friend of GATA; Lz, Lozenge; Srp, Serpent; SrpC, Srp isoform with a single C-terminal zinc finger; SrpNC, Srp isoform with N- and C-terminal zinc fingers; UAS, upstream activation sequence; Ush, U-shaped; Gcm, glial cells missing.

References

  • 1.Querfurth, E., Schuster, M., Kulessa, H., Crispino, J. D., Doderlein, G., Orkin, S. H., Graf, T. & Nerlov, C. (2000) Genes Dev. 14, 2515–2525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Tsang, A. P., Visvader, J. E., Turner, C. A., Fujiwara, Y., Yu, C., Weiss, M. J., Crossley, M. & Orkin, S. H. (1997) Cell 90, 109–119. [DOI] [PubMed] [Google Scholar]
  • 3.Tsang, A. P., Fujiwara, Y., Hom, D. B. & Orkin, S. H. (1998) Genes Dev. 2, 1176–1188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Cantor, A. B., Katz, S. G. & Orkin, S. H. (2002) Mol. Cell. Biol. 22, 4268–4279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Chang, A. N., Cantor, A. B., Fujiwara, Y., Lodish, M. B., Droho, S., Crispino, J. D. & Orkin, S. H. (2002) Proc. Natl. Acad. Sci. USA 99, 9237–9242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Haenlin, M., Cubadda, Y., Blondeau, F., Heitzler, P., Lutz, Y., Simpson, P. & Ramain, P. (1997) Genes Dev. 11, 3096–3108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Orkin, S. H. (1998) Int. J. Dev. Biol. 42, 927–934. [PubMed] [Google Scholar]
  • 8.Crispino, J. D., Lodish, M. B., MacKay, J. P. & Orkin, S. H. (1999) Mol. Cell. 3, 219–228. [DOI] [PubMed] [Google Scholar]
  • 9.Fox, A. H., Liew, C., Holmes, M., Kowalski, K., Mackay, J. & Crossley, M. (1999) EMBO J. 18, 2812–2822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Lu, J. R., McKinsey, T. A., Xu, H., Wang, D. Z., Richardson, J. A. & Olson, E. N. (1999) Mol. Cell. Biol. 19, 4495–4502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Parmacek, M. S. & Leiden, J. M. (1999) Heart Development, eds. Harvey, R. P. & Rosenthal, N. (Academic, San Diego), pp. 291–306.
  • 12.Svensson, E. C., Tufts, R. L., Polk, C. E. & Leiden, J. M. (1999) Proc. Natl. Acad. Sci. USA 96, 956–961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Tevosian, S. G., Deconinck, A. E., Cantor, A. B., Rieff, H. I., Fujiwara, Y., Corfas, G. & Orkin, S. H. (1999) Proc. Natl. Acad. Sci. USA 96, 950–955. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Crispino, J. D., Lodish, M. B., Thurberg, B. L., Litovsky, S. H., Collins, T., Molkentin, J. D. & Orkin, S. H. (2001) Genes Dev. 15, 839–844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Song, W. J., Sullivan, M. G., Legare, R. D., Hutchings, S., Tan, X., Kufrin, D., Ratajczak, J., Resende, I. C., Haworth, C., Hock, R., et al. (1999) Nat. Genet. 23, 166–175. [DOI] [PubMed] [Google Scholar]
  • 16.Nichols, K. E., Crispino, J. D., Poncz, M., White, J. G., Orkin, S. H., Maris, J. M. & Weiss, M. J. (2000) Nat. Genet. 24, 266–270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Freson, K., Devriendt, K., Matthijs, G., Van Hoof, A., De Vos, R., Thys, C., Minner, K., Hoylaerts, M. F., Vermylen, J. & Van Geet, C. (2001) Blood 98, 85–92. [DOI] [PubMed] [Google Scholar]
  • 18.Mehaffey, M. G., Newton, A. L, Gandhi, M. J., Crossley, M. & Drachman, J. G. (2001) Blood 98, 2681–2688. [DOI] [PubMed] [Google Scholar]
  • 19.Yu, C., Niakan, K. K., Matsushita, M., Stamatoyannopoulos, G., Orkin, S. H. & Raskind, W. H. (2002) Blood 100, 2040–2045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Michaud, J., Wu, F., Osato, M., Cottles, G. M., Yanagida, M., Asou, N., Shigesada, K., Ito, Y., Benson, K. F., Raskind, W. H., et al. (2002) Blood 99, 1364–1372. [DOI] [PubMed] [Google Scholar]
  • 21.Walker, L. C., Stevens, J., Campbell, H., Corbett, R., Spearing, R., Heaton, D., Macdonald, D. H., Morris, C. M. & Ganly, P. (2002) Br. J. Haematol. 117, 878–881. [DOI] [PubMed] [Google Scholar]
  • 22.Elagib, K. E., Racke, F. K., Mogass, M., Khetawat, R., Delehanty, L. L. & Goldfarb, A. N. (2003) Blood 101, 4333–4341. [DOI] [PubMed] [Google Scholar]
  • 23.Rehorn, K. P., Thelen, H., Michelson, A. M. & Reuter, R. (1996) Development (Cambridge, U.K.) 122, 4023–4031. [DOI] [PubMed] [Google Scholar]
  • 24.Sam, S., Leise, W. & Hoshizaki, D. K. (1996) Mech. Dev. 60, 197–205. [DOI] [PubMed] [Google Scholar]
  • 25.Lebestky, T., Chang, T., Hartenstein, V. & Banerjee, U. (2000) Science 288, 146–149. [DOI] [PubMed] [Google Scholar]
  • 26.Fossett, N., Tevosian, S. G., Gajewski, K., Zhang, Q., Orkin, S. H. & Schulz, R. A. (2001) Proc. Natl. Acad. Sci. USA 98, 7342–7347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Lanot, R., Zachary, D., Holder, F. & Meister, M. (2001) Dev. Biol. 230, 243–257. [DOI] [PubMed] [Google Scholar]
  • 28.Rizki, T. M. (1978) The Genetics and Biology of Drosophila, eds. Ashburner, M. & Wright, T. R. F. (Academic, New York), Vol. 2b, pp. 397–452. [Google Scholar]
  • 29.Dearolf, C. R. (1998) Biochim. Biophys. Acta 1377, M13–M23. [DOI] [PubMed] [Google Scholar]
  • 30.Rizki, R. M. & Rizki, T. M. (1990) J. Insect Physiol. 36, 523–529. [Google Scholar]
  • 31.Rizki, T. M. & Rizki, R. M. (1984) Insect Ultrastructure, eds. King, R. C. & Akai, H. (Plenum, New York), Vol. 2, pp. 579–604. [Google Scholar]
  • 32.Sorrentino, R. P., Carton, Y. & Govind, S. (2002) Dev. Biol. 243, 65–80. [DOI] [PubMed] [Google Scholar]
  • 33.Bernardoni, R., Vivancos, V. & Giangrande, A. (1997) Dev. Biol. 191, 118–130. [DOI] [PubMed] [Google Scholar]
  • 34.Alfonso, T. B. & Jones, B. W. (2002) Dev. Biol. 248, 369–383. [DOI] [PubMed] [Google Scholar]
  • 35.Duvic, B., Hoffmann, J. A., Meister, M. & Royet, J. (2002) Curr. Biol. 12, 1923–1927. [DOI] [PubMed] [Google Scholar]
  • 36.Lebestky, T., Jung, S. H. & Banerjee, U. (2003) Genes Dev. 17, 348–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Fossett, N., Zhang, Q., Gajewski, K., Choi, C. Y., Kim, Y. & Schulz, R. A. (2000) Proc. Natl. Acad. Sci. USA 97, 7348–7353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Gajewski, K., Fossett, N., Molkentin, J. D. & Schulz, R. A. (1999) Development (Cambridge, U.K.) 126, 5679–5688. [DOI] [PubMed] [Google Scholar]
  • 39.Brand, A. H. & Perrimon, N. (1993) Development (Cambridge, U.K.) 118, 401–415. [DOI] [PubMed] [Google Scholar]
  • 40.Waltzer, L., Bataille, L., Peyrefitte, S. & Haenlin, M. (2002) EMBO J. 21, 5477–5486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Vyas, P., Ault, K., Jackson, C. W., Orkin, S. H. & Shivdasani, R. A. (1999) Blood 93, 2867–2875. [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES