The homeobox-containing gene Wariai regulates anterior-posterior patterning and cell-type homeostasis in Dictyostelium

In Dictyostelium, the multicellular organism is formed by the chemotactic aggregation of up to 10⁵ individual cells (Firtel, 1995; Van Haastert, 1995; Chen et al., 1996). This is mediated by cAMP that interacts with cell surface, G-protein-coupled receptors, activating a series of downstream signaling pathways that control directed cell movement, relay of the cAMP signal and gene expression. Upon mound formation, rising levels of extracellular cAMP trigger a transcriptional cascade regulated through the transcription factor GBF and a cAMP receptor-mediated signaling pathway to initiate morphogenesis and cell-type differentiation (Schnitzler et al., 1994, 1995; Firtel, 1995).

In wild-type organisms, cell-type differentiation and morphogenesis are tightly coupled. Shortly after mound formation, a tip emerges and the organism elongates to form a cylindrical first finger that falls over onto the substratum to form a migrating slug or pseudoplasmodium (Loomis, 1975). Under appropriate conditions, culmination initiates and the organism differentiates into a mature fruiting body containing a spore mass on top of an elongated vacuolated stalk sitting on a basal disc. Spatial patterning within the migrating slug is well-understood. For the most part, spatial patterning in the slug is organized along an anterior-posterior axis. The anterior ∼20% of the slug is composed of three distinct classes of prestalk cells (Williams and Morrison, 1994; Early et al., 1993; Jermyn et al., 1989). The very anterior contains an internal core of prestalk AB (pstAB) cells. This lies in an anterior domain constituting ∼10% of the organism composed of prestalk A (pstA) cells followed posteriorly by a domain of ∼10% of the slug composed of prestalk O (pstO) cells. The posterior ∼80% of the organism represents the prespore domain (Takeuchi and Sato, 1965; Krefft et al., 1984; Gomer et al., 1986; Haberstroh and Firtel, 1990). Within this domain and also scattered throughout the entire organism is another population designated anterior-like cells (ALCs) that have many properties of prestalk cells (see below; Devine and Loomis, 1985; Sternfeld and David, 1982). The very posterior of the slug (rearguard zone) contains an enrichment of ALCs that are lost as the slug migrates (Esch and Firtel, 1991; Sternfeld, 1992). Recent analysis by the Williams laboratory suggests that the prestalk A and pstO cells arise independently as the mound is formed and differentially sort during the process of tip formation (Early et al., 1995). The induction of pstA and pstO cells and subsequent induction of the pstAB cells requires the morphogen DIF in addition to a prior cAMP signal with pstA and pstO cells having a differential sensitivity to DIF (Early et al., 1995; Morris et al., 1987; Town et al., 1976; Williams et al., 1987; Jermyn et al., 1987; Berks and Kay, 1990). The expression of prestalk-specific genes is mediated by a STAT transcription factor (Kawata et al., 1997); however, the role of DIF in potentially regulating STAT function is not known. Prespore cells arise independently within the mass of the mound (Haberstroh and Firtel, 1990; Krefft et al., 1984; Williams et al., 1989). Differential cell sorting of the various prestalk cell populations gives rise to the anterior-posterior patterning observed within the slug (Esch and Firtel, 1991; Traynor et al., 1992; Siegert and Weijer, 1995).

Cell-type patterning is independent of the size of the organism, which can range over >3 orders of magnitude from <100 cells to ∼10⁵ cells (Loomis, 1975; Raper, 1940; Bonner and Slifkin, 1949; Bonner, 1944). In contrast to that in metazoans, cell-type differentiation in Dictyostelium is plastic up to the time of culmination. Classic experiments by Raper showed that if slugs are cut, both the anterior and posterior regions, which preferentially contain prestalk and prespore cells, respectively, give rise to complete organisms with normal spatial patterning if slugs are given time to migrate (Raper, 1940). This has been generally called transdifferentiation. Additional studies in which individual cell types were ‘tagged’ using reporter constructs have shown that, even in wild-type slugs, there is transdifferentiation of prespore to anterior-like cells, and anterior-like cells (ALCs) are in equilibrium with the pstO population, which in turn is in equilibrium with the pstA population [psp<=>ALC<=>pstO<=>pstA] (Abe et al., 1994). These observations, taken in combination with the knowledge that spatial patterning is independent of the size of the organism, suggest the existence of highly regulated homeostatic mechanisms that control both the initial induction and maintenance of cell-type proportioning. Both of these processes are thought to be regulated in part by cAMP emitted from an oscillator in the anterior part of the organism combined with molecular mechanisms to differentially regulate the level of DIF in different parts of the organism (Firtel, 1995; Siegert and Weijer, 1995; Brookman et al., 1987; Kay, 1992). While some mutants have been identified that affect the proportions of different prestalk cell-types or the ratio of prestalk to prespore cells (MacWilliams et al., 1985; MacWilliams and David, 1984; Wood et al., 1996), little is understood about the actual molecular mechanisms controlling these processes.

Homeobox-containing (Hbx) transcription factors have been identified in most if not all phyla of eukaryotes (McGinnis et al., 1984; Scott and Weiner, 1984; Burglin, 1995). In metazoans, Hbx-containing proteins are known to be involved in regulating the anterior-posterior body plan and in the specification of certain cell types and anterior-posterior compartments (McGinnis and Krumlauf, 1992). In organisms from fly to man, many of these genes control the formation of the anterior-posterior axis. We have identified two homeobox-containing genes in Dictyostelium, DdHbx-1 (named Wariai) and DdHbx-2, both of which are developmentally regulated and preferentially expressed starting at the mound stage in response to cAMP and require the transcription factor GBF. In addition to the Hbx domain, the encoded Wariai ORF contains seven ankyrin repeats (similar to IκB) in the C-terminal half of the protein. Our analysis suggests that Wariai is involved in regulating the size of the pstO compartment. wariai null cells show at least a two-fold increase in the pstO compartment by regulating the posterior boundary of this compartment. There is a concomitant decrease in the prespore domain and no observable change in the pstA or pstAB domains. wariai null cells also have a morphological defect consistent with an increase in the prestalk population. We show that Wariai is preferentially expressed in the pstA cells and ALCs with little, if any, expression in pstO cells. Our results suggest that Wariai may be a negative regulator of pstO differentiation and may control the equilibrium between pstO, anterior-like cells (ALCs) and pstA cells, and thus regulate the homeostatic mechanism controlling spatial proportioning. Both the homeobox and ankyrin repeats of Wariai are necessary for Wariai function. ddhbx-2 null cells have no overt phenotype but the gene appears to accentuate the wariai null phenotype in cells in which both genes have been disrupted. These results suggest that, in Dictyostelium, homeobox genes may also control cell-type differentiation and spatial proportioning along the anterior-posterior axis, possibly by regulating equilibrium between various prestalk populations and prespore cells. Wri may represent the earliest evolutionary example of a homeobox-containing gene regulating anterior-posterior axis patterning, suggesting that this mechanism to control spatial patterning in multicellular organisms may have evolved prior to the evolution of metazoans.

We utilized a single degenerate primer PCR method to clone potential homeobox-containing genes. Two degenerate oligonucleotides were designed by reference to conserved region of amino acids 48-54 in the homeodomain as described in the text and used with the T3 sequencing primer to amplify potential homeobox-containing sequences from a λZap cDNA library made from RNA isolated from cells between 8 and 16 hours of development (Schnitzler et al., 1994). Oligonucleotide 1, 5’-C(A,G)(T,A)C(T,G)(A,G)TT(T,C)TG(A,G)A-ACCA; oligonucleotide 2, 5’-(T,A)C(T,G)(T,A)C(T,G)(T,A)T-T(A,T)(C,G)(T,A)(A,G)AACCA. DdHbx-1 (Wri) was identified in the PCR products from oligonucleotide 1, while DdHbx-2 was identified in the PCR products from oligonucleotide 2. The T3 primer, which hybridizes to the N-terminal end of the polylinker region of the library, initiates from the 5’ end of all cDNA clones. Taq DNA polymerase from Gibco BRL was used. The reactions were performed with 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 3 mM MgCl, 4 mM dATP, 3 mM dCTP, 3 mM dGTP, 4 mM dTTP, 0.5 µg of each primer, and 0.1 µl of cDNA library with a titer of ∼10¹⁰/ml. The PCR reaction steps included denaturation at 94°C for 2 minutes; then 40 rounds of denaturation at 94°C for 0.5 minutes, annealing at 37°C for 1 minute, and extension at 72°C for 1.5 minutes. The final extension step was at 72°C for 7 minutes. The PCR fragments generated were subcloned into Bluescript SK and sequenced. Those fragments that encode potential homeobox were used as probes for screening full-length cDNA from the same λZap library. Approximately 10⁶ plaques were screened.

Cell culturing, transformation of Dictyostelium cells and subsequent analysis, and RNA and Southern blot analyses are all standard and were performed as previously described (Mann and Firtel, 1987; Howard et al., 1988; Datta and Firtel, 1988; Dynes et al., 1994). For all experiments, clonal isolates were obtained and more than one clone was examined.

For the Wri knockout construct, the Wri cDNA clone was digested with NdeI to linearize the DNA at the NdeI site located within the homeobox domain. The DNA was blunt-ended with Klenow fragment and a BglII linker was ligated into the site. Either the 1.3 kb BamHI fragment containing the blasticidin-resistant (Bsr) gene cassette (Sutoh, 1993) or the 3.5 kb BamHI fragment containing the auxotrophic marker Thy1 cassette (Dynes and Firtel, 1989) was cloned into this site. The Bsr construct was digested with PstI and HindIII, then electroporated into wild-type KAx-3 cells. The Thy1 construct was digested with SmaI and StuI, then electroporated into JH10 cells (thy1 null cells) (Mann and Firtel, 1991). Homologous recombination events were identified by genomic Southern blot analysis on randomly selected clones.

For the DdHbx-2 knockout construct, a BamHI site was created inside the homeobox by PCR. The original cDNA was subcloned into the EcoRI and XbaI sites of Bluescript SK, placing the N-terminal of the cDNA near the T7 primer. Two PCR reactions were performed to create an internal deletion and a BamHI site. One used the T7 sequencing primer and a DdHbx-2-specific primer: 5’-GTTTGGATCCAATAAGACATACCTAAACG. The product was digested with EcoRI and BamHI. The other fragment was amplified using the T3 sequencing primer and the DdHbx-2-specific primer: 5’-GTTTTGGATCCGTTCTCAAATAAACGTCAAG-OH. This product was digested with BamHI and XbaI. The two fragments were simultaneously ligated into the EcoRI and XbaI sites of Bluescript. This procedure deleted amino acids 43 to 48 in the homeodomain and created a BamHI site in their place. Either the Bsr gene cassette or the Thy1 gene cassette was inserted into the new BamHI site. The Bsr construct was digested with EcoRI and XbaI and electroporated into wild-type KAx-3 cells. The Thy1 construct was digested with EcoRI and XbaI and electroporated into JH10 cells. Homologous recombination events were identified by genomic Southern blot analysis. See Fig. 1.

The WriΔAnk construct carrying an in-frame deletion in the ankyrin repeats was created by digestion of the cDNA with StyI followed by religation of the DNA. The construct has a deletion of amino acids 403 to 628, which includes 6 of the 7 ankyrin repeats. See Fig. 1.

The WriΔHbx construct carrying an in-frame deletion of amino acids 12 to 38 in the homeobox was created using PCR. Specifically, the EcoRI fragment of the Wri cDNA was subcloned into Bluescript, which served as the PCR template. The two primers used for this PCR were 5’-GTTTACTAGTATTATTTTCTAAAAGATGGCATCA-3’ and 5’-GTTTCATATGGTCTGGTGATGTTCTCTTTCT-3’. The PCR product was digested with SpeI and NdeI, then subcloned back into Wri cDNA to create pWriΔHbx. This procedure also deleted 220 bp of 5’ untranslated sequence immediately upstream of the ATG initiation codon.

To express the two deletions constructs and the complete Wariai, the ORF-containing inserts were cloned into the SpeI and XhoI sites of a Dictyostelium expression vector downstream from the Act15 promoter (Dynes et al., 1994). DdHbx-2 expression vectors were made using similar constructs.

Accession number for Wariai is AF036171 and AF036170 for DbHbx-2.

In Dictyostelium, promoters reside within 1 kb of the coding region. To clone the promoter, a pUC vector was inserted into the Wri gene by homologous recombination and then excised with 2.5 kb of upstream sequences. Specifically, pUCBsrΔBamHI, which consists of the 1.3 kb Bsr cassette and pUC118, was introduced into Wri in the genome by homologous recombination. To do this, pUCBsrΔBamHI was inserted into the NdeI and EcoRI sites of WriΔHbx. This created a deletion of amino acids 173 to 244 in the Wariai ORF, including amino acids 12 to 60 of the 60 amino acids homeodomain. The deleted sequences were replaced by the Bsr/pUC118. The resulting DNA was digested with SpeI and HindIII, then electroporated into wild-type KAx3 cells. Homologous recombination events were identified as outlined above. Then genomic DNA were purified, digested with HincII, self-ligated and transformed into supercompetent Epicurian E. coli cells (Stratagene). The rescued plasmid contains 2.5 kb upstream of the Wri coding region. See Fig. 1 for maps.

To make the Wri promoter-lacZ construct, PCR was used to create an in-frame fusion between the promoter and the lacZ gene. The HindIII-XhoI fragment from the rescued plasmid was subcloned into Bluescript SK HindIII and XhoI sites and served as a template for PCR. The primers used in this PCR were the T3 sequencing primer and 5’-GTTTAGATCTCATAACAATTGATGCCATCTT. The PCR product was digested with HindIII and BglII, and ligated into pDdGal-17 combined with the remainder of the promoter region (the HincII to HindIII part of the promoter region). This process created an in-frame fusion in which the complete 2.5 kb promoter region plus sequences coding for the first six amino acids joined to the multiple linker region are in-frame with the lacZ coding sequences.

Histochemical staining experiments for β-gal and β-gluc activity were performed as previously described (Esch and Firtel, 1991; Jermyn and Williams, 1991; Mann et al., 1994). Cells were plated for development on 25 mm Millipore filters (Type HA, 45 µm from Millipore). At a specific developmental stage, the organisms were fixed briefly in Z buffer with 0.5% glutaraldehyde and 0.05% Triton X-100, then stained in Z buffer containing 2.5 mM K₃Fe(CN)₆, 2.5 mM K₄Fe(CN)₆ and 1 mM X-gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside). For β-galactosidase and β-glucuronidase (β-Gus) double staining, X-gal was replaced by 2 mM X-GlcA (5-bromo-4-chloro-3-indolyl-β-D-Glucuronide, Sigma) and 1 mM salmon-β-D-gal (magenta-gal) (Early et al., 1993; Mann et al., 1994). X-GlcA gives blue color and salmon-β-D-gal gives red color.

For neutral red staining, log phase cells were washed several times with 12 mM Na/K buffer (pH 6.2) and incubated with 4 µg/ml neutral red in 40 mM phosphate buffer (pH 7.2), 10 minutes on ice followed by 5 minutes at room temperature. Then cells were washed three times in 12 mM Na/K buffer (pH 6.2) and plated on Na/K agar plates for development. Pictures were taken of migrating slugs.

In order to identify potential homeobox-containing genes, we implemented a ‘single primer’ PCR approach in which we used one degenerate primer from the highly conserved region near the C terminus of the 60 amino acid homeobox domain [amino acids 48-54; WFQNRR(M/A); WFSNRR(R/S)] (Burglin, 1995) combined with a PCR primer complementary to the polylinker used in construction of a λZapII library to identify Dictyostelium homeobox-containing genes. Two Hbx-directed degenerate oligonucleotides were designed, taking into consideration the sequence of various subclasses of metazoan homeoboxes as well as those identified in plants and yeast (Burglin, 1995; see Materials and Methods). PCR amplification of a λZapII library made from RNA isolated from cells between 8 and 16 hours in development (Schnitzler et al., 1994) identified two potential homeobox-containing clones in the initial sequencing screen of 25 randomly selected clones. These cloned PCR products were then used to screen the λZap library for full-length cDNA clones. The derived amino acid sequence of both of these ORFs is presented in Fig. 1. Visual sequence comparison of the domain identified in the PCR clones and BLAST (Altschul et al., 1990) searches identified a homeobox domain in both DdHbx-1 and DdHbx-2. Because DdHbx-1 regulates cell-type proportioning (see below), we have named this gene Wariai (Wri), the Japanese word for proportioning or ratio. Comparison to a consensus homeobox (Hbx) sequence (Burglin, 1995) is shown in Fig. 1A. In addition, Wariai also contains seven ankyrin repeats in the carboxy-terminus that are most closely homologous to those found in mouse ankyrin (Birkenmeier et al., 1993; Fig. 1A). Like many other Dictyostelium proteins (Burki et al., 1991; Mann and Firtel, 1991; Pitt et al., 1992), Wariai and DdHbx-2 contain runs of repeated amino acids, including a long, very asparagine-rich region near the N terminus of Wariai and a more unusual glycine-rich-containing region followed by a region containing alternating histidine and asparagine residues. The DdHbx-2 sequence shows Methods). Multiple disrupted strains made using both selections were identified by Southern blot analysis (data not shown) of randomly selected, independent clones. All showed the same phenotype and no expression of Wri transcripts (data not shown). wri null cells aggregated and formed mounds normally when plated for development on non-nutrient agar. Approximately 30% of the organisms arrest at this stage, while the remainder proceed through development and form fruiting bodies with an enlarged ‘basal disc’ and lower portion of the stalk, suggesting there is an increase in the number of basal disc and stalk cells (Fig. 3D,F), compared to wild-type cells (Fig. 3B). In addition, predominantly long homopolymer regions and stretches of hydrophobic amino acids in addition to the homeobox domain. The Wri homeobox shows strongest homology to the Hbx of the tapeworm gene EgHbx4 (accession no. JC1389). Neither of the two Dictyostelium genes are members of any known homeobox gene family (Burglin, 1995).

Developmental RNA blot hybridization (Fig. 2) shows both genes are developmentally regulated: expression of DdHbx-2 is first detected and that of Wariai rises significantly at 8 hours, the time of mound formation, when there is a developmental switch from aggregation to multicellular differentiation and just prior to cell-type differentiation (Firtel, 1995). No expression of DdHbx-2 or increased expression of Wri was observed in gbf null cells in which the transcription factor GBF, which is essential for differentiation past the mound stage, is knocked out (Schnitzler et al., 1994), indicating Wri and DdHbx-2 are dependent on GBF function for a high level of expression during the multicellular stages. Both genes are induced in suspension in response to cAMP (data not shown), as would be expected for genes preferentially induced at the onset of mound formation.

Wariai was disrupted by homologous recombination in two independent strains: (1) wild-type KAx-3 cells using the Bsr (blasticidin) dominant drug selectable marker, and (2) in the thy1 null strain JH10, which exhibits a wild-type pattern of development, using the Thy1 (thymidine) auxotrophic marker (Dynes and Firtel, 1989; see Materials and the anterior of wri null slugs was not as cylindrical as that of wild-type slugs (Figs 4, 5). When Dictyostelium cells are grown on nutrient plates in association with a bacterial food source, clones originating from a single cell grow vegetatively, feeding on the bacteria, and form a ‘plaque’ in the bacterial lawn as the bacteria are eaten. As the bacteria are depleted, the cells starve and initiate multicellular development. The morphological phenotype of the wri null cells is more severe under these developmental conditions than when grown axenically and plated on non-nutrient buffered agar (Fig. 3C,E). Wild-type cells show a similar developmental morphology whether allowed to develop on nutrient or non-nutrient plates (Fig. 3A,B).

ddhbx-2 null cells, made and examined using the same approaches as for disruption of Wri, are morphologically indistinguishable from wild-type cells regardless of which genetic background and marker was used (data not shown). Double knockouts (wri/ddhbx-2 null cells) were made in which the Bsr Wri knockout construct was used to disrupt the Wariai gene in ddhbx-2 null cells created using Thy1 selection. The DdHbx-2 Bsr knockout construct was also used to disrupt the DdHbx-2 gene in wri null cells made using the Thy1 marker. In this way, both reciprocal double knockout strains were obtained. Morphologically, the double knockout strains appeared similar to wri null cells, although they may have a slightly enhanced morphological phenotype (Fig. 3G-I). In addition, the phenotypes are slightly stronger than those of wri null cells (Fig. 2), with a greater fraction of the aggregates arresting at the mound stage and larger increase in the lower stalk region. All sets of experiments described below were done with both wri null strains and both wri/ddhbx-2 double knockout strains. No differences were identified between the wri null strain made using either marker or between the reciprocal double knockout strains.

Spatial patterning of the cell types was examined in wild-type, wri null, ddhbx-2 null and double knockout strains transformed with lacZ reporter constructs in which β-gal is expressed from cell-type-specific promoters [SP60 (prespore-specific), ecmA (pstA-specific), ecmO (pstO-specific), ecmB (pstAB and ALC) and ecmAO (expressed in pstAB, A and O cells and ALCs, with a stronger expression in pstA than pstO cells) (Early et al., 1993; Jermyn et al., 1989; Haberstroh and Firtel, 1990)]. In addition, we examined the spatial pattern of expression of the rasD/lacZ reporter, which shows a significantly stronger level of expression in the anterior pstA/O/AB region and ALC population than in prespore cells (Esch and Firtel, 1991). A cartoon is presented in Table 1 that outlines the pattern of expression in the slug of the reporters that were used. Spatial patterning was also analyzed statistically by Developmental RNA blots were performed to examine the timing and level of expression of ecmAO, ecmB and SP60 in wri null and wild-type cells. No reproducible, significant differences were observed between the mutant and wild-type quantifying the anterior-posterior boundaries of the staining of individual reporters where 0.0 is the very anterior of the slug and 1.0 is the very posterior. Because cell-type proportioning is independent of the size of the slug (Firtel, 1995; Loomis, 1975), such measurements provide an accurate mechanism of quantitating the relative size of various cell-type compartments. The size independence of spatial patterning, which has been observed previously for wild-type organisms, was also observed in our mutant studies in which slugs of varying sizes showed the same spatial patterning.

Analysis of the prestalk- and prespore-specific ecmAO and SP60 reporter staining shows that the anterior prestalk domain comprising the pstAB, A and O domains is significantly enlarged in the wri null and double knockout strains compared to that seen in wild-type cells and there is a concomitant decrease in the size of the prespore domain in the mutant strains (Fig. 4A; see Table 1 for quantitation). To determine which prestalk domains were affected in wri null cells, we used the pstA-(ecmA/lacZ) and pstO-specific (ecmO/lacZ) reporters (Fig. 4B). Examination of stained slugs and a quantitative analysis of the data (Table 1) show that there is an approximate doubling (see below) in size of the pstO compartment that is found at the coordinates ∼0.1-0.22 units along the anterior-posterior axis in wild-type KAx-3 cells. This increases to coordinates ∼0.1-0.35 in wri null cells, with a decrease in the size of the posterior prespore/rearguard domains. In contrast, the pstA (ecmA-expressing) and pstAB (anterior ecmB-expressing) compartments are unchanged (Fig. 4B; data not shown for the ecmB marker that is specific for the anterior pstAB domain in the slug). We also notice that, in wri⁻/SP60/lacZ slugs, there are some prespore cells (cell expressing β-gal) in the enlarged pstO domain that are not observed in wild-type slugs. The results obtained from direct measurements of the various domains and indirect subtractive measurements of the stained slugs are all consistent (Table 1). No difference was observed between measurements made in which the Wariai gene was disrupted by the Thy1 marker and using the Bsr selection (data not shown). cells, presumably because northern blots are not sufficiently quantitative and pstO cells only weakly express the entire ecmAO promoter compared to pstA cells (Early et al., 1993), which are unaffected in wri null cells (data not shown). Similar studies performed with ddhbx-2 null strains showed no difference from the results observed in wild-type strains (data not shown). When these studies were done in the wri/ddhbx-2 double knockout cells, the results were similar to those in wri null cells except that the increase in the prespore O domain was slightly larger (Fig. 4; Table 1). We also observe a greater number of SP60/lacZ-expressing cells in the pstO domain than are observed for the wri null cells (Fig. 4, part I, SP60/lacZ panels).

The increase in the number of the ecmO/lacZ-staining cells was quantified counting the number of ecmO/lacZ-staining cells in wri null and wild-type slugs after they were dissociated into single cells and the cells histochemically stained. This analysis showed 11± 1% of the wild-type and 26 ± 2% of the wri null cells stained, indicating there was an ∼2.5-fold increase in the number of ecmO-expressing cells in the mutant. The ecmO/lacZ-staining cells are expected to be a combination of pstO cells and ALCs.

As a further method to examine the relative size of prestalk and prespore domains, we employed double-label reporter analysis using ecmAO/lacZ and SP60/β-glucuronidase (GUS) reporters. Clonal isolates of strains expressing both markers were obtained and histochemically stained for β-galactosidase and β-glucuronidase activity (Haberstroh and Firtel, 1990; Jermyn and Williams, 1991; Early et al., 1993; Mann et al., 1994). The results presented in Fig. 5 clearly show a significant increase in the prestalk (magenta-stained) domain (ecmAO/lacZ) and a decrease in the prespore (blue) domain (SP60/GUS) in wri and wri/ddhbx-2 null strains when compared to wild-type cells. The increase in the pstAO domain in these experiments using the ecmAO/lacZ reporter and magenta-gal substrate is greater than that observed using X-gal, although the constructs used are identical (Jermyn and Williams, 1991). This may be due to the staining properties and/or differences in the hydrolysis rates of the two substrates by β-gal. Nonetheless, using either marker, there is a significant increase in the pstAO domain in wri null cells. The double-staining results also suggest little overlap, if any, between the prestalk and prespore domains.

The spatial pattern of expression of ecmAO, ecmO and SP60 was also examined in mature fruiting bodies. With either the ecmAO/lacZ or ecmO/lacZ reporter, wri null cells showed a more intense staining in the basal disk and upper and lower cups, regions that are derived from the ALC populations (data not shown). Prespore marker (SP60/lacZ) staining of the spore mass showed a narrower band of staining in the null cells compared to wild-type cells (data not shown). These data are consistent with an increase in the pstO/ALC population in the mutant strains when compared to wild-type cells.

Neutral red is a vital dye that preferentially stains an acidic vacuole found in prestalk cells and anterior-like cells (ALCs) but is absent in prespore cells in the slug (Loomis, 1982; Sternfeld and David, 1982). Neutral red has been used as a prestalk and anterior-like cell marker and a mechanism to quantitate sizes of prestalk compartments and localize and monitor movement of ALCs within the prespore region (Abe et al., 1994; Dormann et al., 1996; Hadwiger et al., 1994; Jermyn and Williams, 1991; Siegert and Weijer, 1992, 1995). In confirmation of the lacZ reporter analysis, neutral red staining shows that the anterior, neutral red staining region is significantly enlarged in wri null cells compared to wild-type cells (Fig. 6). Transformation of the wri null cells with the Wariai expression vector restores the wild-type neutral red staining pattern (Fig. 6) and complements the wri null morphological defects (data not shown). To determine whether both the homeobox and the ankyrin repeats are required, constructs carrying an in-frame deletion of either part of the Hbx domain (Act15-WriΔHbx) or the ankyrin repeats (Act15-WriΔAnk) were transformed into wri null cells. As shown, neither deletion construct was able to restore neutral red staining (Fig. 6) nor morphological defects observed in wri null cells (data not shown). These results indicate that both the homeobox domain and the ankyrin repeats are required for Wariai. Cells overexpressing the Wariai coding region showed no morphological defect or change in neutral red staining (data not shown).

We used a novel approach to clone the Wariai promoter. To accomplish this, we integrated, by homologous recombination, a pUC-based E. coli plasmid into the Wri gene (see Materials and Methods). Homologous recombination into the Wri locus was confirmed by Southern blot analysis and the strains exhibited the same morphological defect as the other wri knockouts (data not shown). The 5’ flanking region of Wri was isolated by linearizing genomic DNA 2.5 kb upstream from the start codon and at a designated site within the inserted vector, ligating the DNA and cloning the recircularized vector carrying the 5’ flanking region into E. coli. This cloned 5’ region was fused in-frame to the β-gal reporter and transformed into wild-type and wri null cells and stably transformed clones were isolated. As shown in Fig. 7C,D, the lacZ expression pattern in wild-type slugs is prestalk-specific with very strong staining in the pstA and some scattered throughout the posterior, which probably represents expression in anterior-like cells. No definitive staining of the pstO compartment of migrating slugs is observed. In the developing mound, staining is first seen as a ring of cells (Fig. 7B) that then moves toward the developing tip and is observed as a centrally localized group of cells near the top of the mound (Fig. 7A). This staining pattern is remarkably similar to that of pstA cells, as previously described (Early et al., 1995). In the mature fruiting body, expression is seen in the stalk and upper and lower cups of the sorus, consistent with a pstA/O-like expression pattern (data not shown). When the staining pattern of Wri/lacZ was examined in wri null cells, the staining pattern was unchanged, remaining highly localized to the anterior pstA compartment (Fig. 7E). No expansion of the staining pattern was observed as is seen with ecmAO/lacZ or ecmO/lacZ. This is consistent with Wri expression in the anterior of the slug being restricted to the pstA region.

Expression of the Wri construct from the cloned Wri or the ecmA, pstA-specific, promoter complemented the wri null phenotype as determined by quantitation of the size of the prestalk zone by neutral red staining as described above (data not shown). However, expression of Wri from either the pstO-specific ecmO or the prespore-specific SP60 promoter did not complement the wri null phenotype (data not shown).

From the Wri/lacZ staining, Wri expression appears localized to pstA and ALCs even though the wri null mutation alters the size of the pstO compartment, suggesting that Wri may function non-autonomously. To further examine a possible non-autonomous function for Wri, we created chimeras composed of wri null and wild-type cells, with one of the strains expressing the ecmO/lacZ reporter to mark pstO cells. Similar experiments were performed with the prestalk marker SP60/lacZ. In these experiments, the cells carrying the reporter were mixed with three parts of unlabeled cells. In some experiments, a fraction of the cells carrying the lacZ reporter were also labeled with the vital fluorescent stain Cell-Tracker Green (CMFDA, Molecular Probes) as a ‘genetic’ tag, which allows us to determine the general distribution of this strain within the slug. As shown in Fig. 8B, mutant and wild-type cells distributed uniformly throughout the slug. As expected, homologous mixtures (one part ecmO/lacZ-wri null cells mixed with three parts unlabeled wri null cells or one part ecmO/lacZ wild-type cells mixed with three parts unlabeled wild-type cells) showed the same spatial distribution of ecmO/lacZ-expressing cells as described above when the wri null and wild-type strains were examined individually, except that the density of the label is reduced due to the presence of fewer ecmO/lacZ-expressing cells (compare Fig. 8D,E to Fig. 4B). When one part ecmO/lacZ-wri null cells was mixed with three parts unlabeled wild-type cells, the size of the pstO domain was slightly larger than that of a wild-type:wild-type homologous chimera (compare Fig. 8A-D). However, when one part ecmO/lacZ-wild-type cells was mixed with three parts unlabeled wri null, the labeled wild-type cells (Fig. 8B,C) showed a significantly broader distribution than that observed with the control homologous wild-type:wild-type mixture in the reciprocal mix, which was similar to the homologous wri⁻/wri⁻ mixture. Thus, in a slug where 75% of the cells are unlabeled wri null cells, there is a significant expansion of the region occupied by the wild-type pstO cells, which is consistent with a significantly enlarged pstO compartment.

Experiments using the SP60/lacZ reporter show a similar pattern, with the majority strain in the chimera dictating the size of the prestalk and prespore compartments (Fig. 9). In the chimera containing one part wri⁻/SP60/lacZ and three parts wild-type cells, the size of the prestalk domain was only slightly larger than that observed for wild-type strains and the prespore domain was slightly smaller (Fig. 9B). In the reciprocal chimera (75% of the cells are unlabeled wri null cells), there is a significant reduction of the region occupied by wild-type prespore cells (SP60/lacZ-expressing cells; Fig. 9D), which is consistent with the above model.

Wariai and DdHbx-2 are the first homeobox-containing genes that have been identified in Dictyostelium. Disruption of Wri leads to a more than doubling of ecmO/lacZ-expressing cells and the pstO compartment and no detectable change in the relative size of the more anterior pstA and pstAB compartments. Furthermore, we show that both the homeobox domain and the ankyrin repeats are required for Wri function. As there is recent evidence that there is no cell division during morphogenesis (Shaulsky and Loomis, 1995) in Dictyostelium and spatial patterning is independent of the size of the organism, this increase in the size of the pstO compartment is not simply due to a mitotic doubling of the pstO cells. Our data suggest that there is a concomitant decrease in the number of prespore cells, it is probable that the increase in the pstO cells is at the expense of the prespore cells. As we discuss below, we propose that this might be the result of a conversion of prespore cells or cells destined to become prespore cells to pstO cells. While it is difficult to determine quantitatively if there is an increase in the ALC population, staining with an ALC marker (Gα4/lacZ reporter, Hadwiger et al., 1994) shows no clear differences between wri null and wild-type cells (ZH and RAF, unpub. obser.).

Our analysis suggests that disruption of Wri does not affect the ability of any known cell type to differentiate but leads to a shift in the equilibrium of prespore and pstO populations. It is possible that Wri regulates an interconversion of pstO and prespore cells via transdifferentiation, which may play an important role in regulating the maintenance of cell-type proportioning. Two alternate models are possible. The first is that Wri does not affect the initial patterning but affects cell-type homeostasis, which leads to an increase in pstO cells and a decrease in prespore cells. In this model, Wri would not necessarily alter the initial spatial patterning of the prespore, pstA and pstO cell types, which are thought to derive from distinct subpopulations of cells within the developing mound (Early et al., 1995). The second is that Wri functions to control the sensitivity of cells to a morphogen such as DIF and effects the relative number of cells that become pstO cells. In this scenario, pstO cells would be induced at lower DIF concentrations and/or prespore cells would be more sensitive to DIF, resulting in a decrease in prespore cell differentiation. While we observed no differences in the initial number of ecmO/lacZ-staining (pstO) cells in the forming mound of wri null and wild-type strains (data not shown) in support of the first model, prestalk cells are still differentiating as the mound is forming, making it difficult to accurately quantitate the number of cells that will initially become pstO cells. We also identified a second homeobox gene, Ddhbx-2, that may potentiate the function of Wri. Both genes are developmentally regulated and induced by cAMP as the mound forms. While transdifferentiation is known to occur in migrating slugs and if slugs are dissected, it is possible that other mechanisms may also control cell-type interconversion and Wri may regulate these processes.

(ecmO-staining) cells. These data are also consistent with an enlarged pstO compartment and a cell non-autonomous function for Wri in regulating patterning. Such a possible non-autonomous function of regulating spatial patterning by a homeobox-containing gene is quite distinct from that seen in metazoans where the pathways regulated by Hbx-containing genes are predominantly cell-autonomous (McGinnis and Krumlauf, 1992).

Approximately 30% of the wri aggregates arrest at the mound stage and do not form tips, a percentage that is higher in wri⁻/ddhbx-2 null cells. Tips are formed through the chemotactic migration of prestalk cells to the apex of the mound (see Introduction). As Wri is expressed in pstA, it is possible that Wariai may have a cell autonomous function to control tip formation, with null cells being less efficient in tip formation. It is possible that Wri’s regulation of tip formation and the pstO compartment are distinct functions of the gene and that Wri controls two very different pathways.

It is interesting to speculate on the possible evolutionary significance of the role of Wariai. In metazoans, it is known that some homeobox-containing genes regulate cell-type patterning and are involved in controlling anterior-posterior axis formation and determination. The Dictyostelium slug is a relatively simple and ancient multicellular organism with only a few cell types but a clear anterior-posterior axis. This forms after the cells sort Wariai shows a pstA- and ALC-specific spatial pattern of expression during multicellular development, with expression being highest in pstA and little, if any, expression in pstO cells. Moreover, expression of Wri protein from a pstO-specific promoter construct does not complement the wri null phenotype, suggesting the regulation by Wri on patterning is cell non-autonomous. The doubling of the size of the pstO domain in wri null cells suggests that Wri functions genetically to negatively regulate a signaling pathway controlling the size of the pstO compartment and is required for properly specifying the posterior boundary of this domain. We expect that Wri encodes a transcription factor that may function genetically as a repressor for a signaling pathway that is involved in negatively regulating prestalk cell specification. This is consistent with our observation of wild-type pstO cell distribution in chimeras containing a predominance of unlabeled wri null cells in which there is a significant increase in the size of the domain occupied by wild-type pst during tip formation and this anterior-posterior pattern can first be distinguished in the first finger stage. It is possible that the role of homeobox-containing genes in the control of the anterior-posterior axis is sufficiently ancient to include cell-type proportioning in Dictyostelium. In Dictyostelium, this may be regulated by changing the proportioning mechanism after differentiation of the individual cell types, possibly via the regulation of transdifferentiation or another process, in contrast to mechanisms that may be used in metazoans, which, in general, have more determinative cell fate decisions. The identification of genes directly regulated by Wariai should help elucidate the mechanism by which this gene controls spatial patterning.

We would like to thank J. Posakony, W. McGinnis, and T. Burglin for helpful comments, suggestions, and consultations. This work was supported by USPHS grants to R. A. F.