The C-Module DNA-Binding Factor Mediates Expression of the Dictyostelium Aggregation-Specific Adenylyl Cyclase ACA

Abstract

Aggregation of Dictyostelium discoideum amoebae into multicellular structures is organized by cyclic AMP (cAMP), which acts as a chemoattractant, as a second messenger, and as a morphogen. Aggregation of D. discoideum cells depends on the expression of adenylyl cyclase ACA, which provides extracellular cAMP for signal relay and intracellular cAMP for the induction of genes required at multicellular stages. We have identified a DNA-binding activity specific for a highly A+T-enriched motif in the upstream region of the ACA-encoding gene, acaA. The factor shows DNA-binding characteristics very similar to those of C-module-binding factor (CbfA). Although CbfA was originally identified as a putative regulator of the activity of D. discoideum retrotransposon TRE5-A, it also was found to be essential for aggregation of D. discoideum cells. The identified DNA-binding activity was absent in mutant cells depleted of CbfA, and CbfA could be precipitated using an acaA promoter fragment. We propose that CbfA binds to the acaA promoter to provide a basal transcription activity that is required for induction of ACA expression after the onset of D. discoideum development.

Dictyostelium discoideum is a social amoeba that lives in soil and feeds on bacteria. When the environmental conditions limit vegetative growth, the amoebae collect into aggregates and form multicellular organisms in which the cells cooperate to form fruiting bodies consisting of a stalk that supports a mass of dormant spores (5). All stages of D. discoideum development are coordinated by cyclic AMP (cAMP) (9). Starving D. discoideum cells spontaneously start to produce and secrete pulses of cAMP. Surrounding cells can sense cAMP by means of a cAMP-specific receptor, CAR1, which belongs to the family of G protein-coupled receptors. Signaling by CAR1 leads rapidly to the activation of the cAMP-dependent protein kinase A (PKA) and the mitogen-activated protein kinase extracellular signal-related kinase 2 (ERK2) (6, 8), which phosphorylate downstream substrates, influence the expression of genes required for aggregation and later development, and regulate intracellular cAMP levels by modulation of both adenylyl cyclase ACA and phosphodiesterase RegA (15, 16). The coordinated activation of ACA, PKA, and ERK2 by CAR1 and the subsequent suppression of ACA and ERK2 activity by PKA lead to a cAMP-induced feedback loop that ensures oscillatory activation of ACA in about 6-min intervals and pulsatile production of cAMP that is required for full induction of aggregation-specific genes (7).

Several genes have been shown to be required for aggregation of D. discoideum cells. In support of the central role of ACA during early development, mutants defective in the gene for ACA (acaA) fail to aggregate (12). It was surprising to find that acaA null cells can develop into multicellular stages if given artificial extracellular pulses of cAMP (13). This result was interpreted by the presence of another adenylyl cyclase, ACR, whose expression in growing and aggregating cells is low and unable to support relay of the extracellular cAMP signal (17). Pulsing of ACA-deficient cells is thought to activate ERK2, which then represses RegA, thus leading to elevated levels of ACR-produced intracellular cAMP in the absence of ACA; these cAMP levels are high enough to activate PKA.

The present data point to the central role of ACA during aggregation, yet little is known about the gene regulatory networks that lead to the induction of the acaA gene during early development. It has been reported that D. discoideum mutants defective in Myb2, a factor that contains DNA-binding motifs found in Myb-related transcription factors, are unable to aggregate (11). Myb2 null cells complete multicellular development after exposure to extracellular cAMP pulses, and ectopic expression of ACA in the mybB mutant rescued its aggregation-minus phenotype (11). Thus, it has been proposed that Myb2 may induce ACA during starvation, but is has not been shown whether Myb2 mediates acaA induction by direct interaction with the acaA promoter.

CbfA, the C-module-binding factor, was discovered as a DNA-binding activity specific for a regulatory DNA element (the C module) in the D. discoideum retrotransposon TRE5-A (2). There are two binding sites for CbfA in the C module as determined by footprinting analyses, and these binding sites are unusual in that they are about 30 bp long and are characterized by central oligothymidine stretches of at least 20 bp in length (4). Biochemical analyses of the interaction of CbfA with the C module have suggested that CbfA may recognize unusual local DNA conformations enforced by the high A+T content and homopolymeric thymidine runs. Specific high-affinity binding of CbfA to the C module requires interaction with the minor groove of the recognized DNA segments (4). Nevertheless, it is somewhat puzzling how CbfA can discriminate specific target sequences in the background of the A+T-biased D. discoideum genome, and analysis of other promoters directly regulated by CbfA in D. discoideum cells will be required to address this issue.

We have recently used the translation stop codon suppression technology to create a D. discoideum mutant that expresses less than 5% of wild-type CbfA levels (cbfA^am mutant JH.D) (18). This mutant is unable to aggregate and fails to express most early and late genes required in the multicellular phase (19). Similar to acaA and mybB knockout strains, the cbfA^am mutant is able to enter multicellular development and produce live spores after exposure to extracellular cAMP pulses. Northern blots have revealed the complete absence of acaA expression in the cbfA^am mutant (19). Here we show that the aggregation deficiency of the cbfA^am mutant can be overcome by ectopic expression of ACA and that CbfA binds to the acaA upstream region in a highly A+T-rich region characterized by long homothymidine stretches. Since deletion of this binding site eliminates acaA promoter activity, we suggest that CbfA is a crucial factor in D. discoideum development that is required for proper expression of ACA during aggregation.

MATERIALS AND METHODS

D. discoideum cell culture, transformation, and development.

D. discoideum AX2 and cbfA^am (strain JH.D) cells (19) were grown in liquid HL5 medium. ACA was expressed by transforming plasmid CP43 (a gift of C. Parent, NCI, Bethesda, MD) into JH.D cells. G418-resistant transformants were selected in the presence of 10 μg/ml G418 and analyzed for their ability to enter multicellular development on phosphate-buffered agar plates or on black nitrocellulose filters (19).

Cloning of acaA promoter fragments.

The acaA upstream region (acaAP) was cloned using information from Dictybase entry DDB0214814 (www.dictybase.org). A 773-bp fragment was generated using primers acaAP-01 (GGAATTCTGGGTAGGTTAAAATTTGAAACTGAATAGAATTGTG) and acaAP-02 (GGAATTCATGATCATTAAACATTGGTGAGCTAGATGCC). The promoter fragment acaAP(−739/+34) was inserted into pGEM-T (Promega). To create smaller promoter fragments suitable for mobility shift assays, two EcoRI sites were introduced into the acaAP(−739/+34) fragment using primers acaAP-mut01 (5′-TTTATTTTACTAGTAGAATTCATTTGTTGTACC-3′) and acaAP-mut02 (5′-GTTAATTTGAATTCAAAATATATTGTATTGATAGG-3′). The three resulting acaAP subfragments were separately cloned into the EcoRI site of pGEM7Zf(-). Oligo(dT) stretches in the acaAP(−739/−448) promoter fragment were disrupted with oligo(dA) by site-directed mutagenesis using oligonucleotides acaAP-mut03 (5′-CCCTTTTTTTTTAAAAAAAAAAATTTTTTTTTCTTTTTTTTTTTTTTTTTTTTT TG-3′) and acaAP-mut04 (5′-TTTTTTTTTTTTCTTTTTTAAAAAAAAAATTTTTTGTTAGTAATATTATTAC-3′), yielding promoter fragment acaAP(−739/−448*) (underlining indicates introduced adenine bases). A 98-bp region of the A+T-rich CbfA binding site in acaAP(−739/−448) was deleted by inserting flanking BglII sites by site-directed mutagenesis. Digestion of pGEM-acaAP(−739/−448^Bgl) with BglII and religation of the vector yielded acaAP derivative acaAP(−739/−448Δ). With the same strategy, the deletion of the 98-bp fragment was reproduced in vector pGEM-acaAP(−739/+34) to generate acaAP derivative acaAP(−739/+34D). A 90-bp DNA fragment containing part of the C module of TRE5-A.1 was isolated with BamHI and BglII from pUC9#17 (2) and inserted into the BglII site of pGEM-acaAP(−739/+34Δ).

Reporter assay with β-galactosidase.

The complete acaA upstream region acaAP(−739/+34) and derivative acaAP(−739/+34D) were cloned into the promoterless D. discoideum lacZ expression vector pDdGal-17 (3) to generate in-frame fusions of the first 11 ACA amino acids and β-galactosidase. The pDdGal-acaAP plasmids were transformed into D. discoideum AX2 cells. Transformants were allowed to develop on black filters (19) for the time periods indicated in the figures. Cells were harvested and stored as frozen pellets. β-Galactosidase activity was assayed as described previously (14). Total RNA was prepared from frozen cells as described previously (19). We then prepared mRNA from total RNA using a QIAGEN Oligotex kit in order to remove trace amounts of genomic DNA. The absence of genomic DNA was confirmed by the size of the acaA reverse transcription (RT-)PCR fragment spanning intron 3 of the acaA gene (data not shown). First-strand cDNA was synthesized with an oligo(dT) primer. Expressions of acaA, acaAP-lacZ, and cbfA were analyzed in separate PCRs to account for optimal template concentrations required for each gene. The primers were as follows: acaA-05, 5′-CAAGATCTTCTTTAACTCGTGTTTGTGCAAC-3′; acaA-06, 5′-GGTTCATCATATTCTTGGAAACCTGCAATTTG-3′; lacZ-02, 5′-GGAACTCCGCCGATACTGACGGGCTC-3′; lacZ-03, 5′-GGTAGTGGTCAAATGGCGATTACCGTTG-3′; cbfA-122, 5′-GGCGGCCGCAATGGAATGGCAAGAATATCTTTC-3′; and cbfA-123, 5′-GGCGGCCGCTATGTATGAGGTAATGATGAAAAGTAAGAG-3′.

EMSAs.

Nuclear proteins were prepared by extracting isolated nuclei with 400 mM KCl (NE400 fraction), and electrophoretic mobility shift assays (EMSAs) were performed essentially as described previously (4, 18). Unless stated otherwise, assay mixtures contained 1 μg of each of the nonspecific competitor DNAs poly(dAdT) · poly(dAdT) (Sigma #0883) and plasmid pGEM7Zf(-) (Promega). Radiolabeled C module and acaAP probes were generated as detailed previously (18). The A+T-rich CbfA binding site of acaAP [acaAP(−630/−526)] was generated with two synthetic oligonucleotides that were annealed and filled in with Klenow polymerase in the presence of desoxynucleotides and [α-³²P]dATP. Assays with competitor DNA fragments or DNA-specific drugs were performed as described previously (4). In some experiments, amounts of protein-DNA complexes were quantified by cutting off the bands from the dried gels and quantifying the enclosed radioactivity in a liquid scintillation counter.

DNA pull-down assays and Western blots.

Biotinylated DNA probes were generated by PCR using the subcloned C module, acaAP(−739/−448), and acaAP(−739/−448*) as templates, one biotin-labeled pGEM-specific primer, and an unlabeled reverse pGEM primer. Biotin-labeled DNAs were gel purified and adjusted to similar DNA concentrations. For pull-down experiments, 500 μl NE400 from AX2 cells was preincubated at room temperature for 15 min with 25 μg of each of the nonspecific competitor DNAs poly(dAdT) · poly(dAdT) and plasmid pGEM7Zf(-). Five μg of biotinylated DNAs was added, and the mixtures were incubated for an additional 60 min. Then, 50 μl of streptavidin-loaded Dynabeads (Dynal) was added, and incubation was continued for 30 min with shaking. Magnetic beads were separated from the mixtures and washed five times with GP50 buffer (2). The beads were then boiled in sample buffer, and eluted proteins were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Protein samples were blotted onto nitrocellulose BA85 membranes and stained for the presence of CbfA using monoclonal CbfA antibody 7F3 as described previously (18).

RESULTS

Expression of recombinant ACA rescues development of the CbfA mutant.

CbfA-depleted cells developing on a solid surface are unable to express ACA (19). Hence, it was tempting to speculate that loss of ACA activity is the major cause of the cbfA^am mutant phenotype (i.e., the aggregation defect). To test this assumption, we expressed ACA in the cbfA^am mutant under the control of the constitutively active actin15 promoter. As shown for a representative clone in Fig. 1, the obtained transformants reproducibly aggregated and entered multicellular development with a slight temporal delay compared to the wild type. Nevertheless the results suggested an almost complete reversal of the cbfA^am mutant phenotype by expression of ACA.

FIG. 1.

Expression of recombinant ACA in cbfA^am mutant JH.D. Plasmid CP43 was transformed into JH.D cells and G418-resistant clones were plated on phosphate-buffered agar plates. Multicellular development of the transformants was observed and compared with wild-type AX2 and untransformed JH.D cells. One representative cbfA^am/rACA transformant is shown. Pictures were taken 28 h (left panels) and 44 h (right panels) after plating.

Characterization of the acaA upstream region.

Since ectopic ACA expression rescued the developmental phenotype of the CbfA mutant, we wanted to explore whether CbfA directly modulates ACA expression by interacting with the acaA promoter. We decided to clone positions −739 through +34 of the acaA upstream region based on Dictybase data (www.dictybase.org). We noticed two copies of the sequence 5′-TTGTGTGTGTGT-3′ within the cloned acaAP (acaA promoter) region located at positions −708 through −684 of acaAP (Fig. 2). This sequence is similar to known G-rich motifs in developmentally regulated D. discoideum promoters and may represent a cyclic AMP response element (10), but we did not focus on this element in this study, since this motif is not a potential CbfA binding site.

FIG. 2.

Analysis of the acaA upstream region. Shown is the DNA sequence of the cloned acaA promoter. The +1 nucleotide denotes the translation start of ACA (uppercase letters; ATG start codon in bold). Two GT-rich motifs are indicated in bold lowercase letters. EcoRI sites used to clone the promoter are underlined.

We first showed that the activity of the cloned acaA promoter region was induced during early development. We inserted acaAP(−739/+34) into a promoterless lacZ expression vector. RT-PCRs were performed on mRNA preparations of D. discoideum transformants to compare the developmentally regulated induction of the lacZ transcript expressed from acaAP(−739/+34) with that of the endogenous acaA mRNA (Fig. 3A). The acaA message was induced by 4 hours of starvation and declined after about 12 h, which corresponded to the tight aggregate stage observed under a microscope. The lacZ mRNA was first detectable after 2 hours of development, and peak expression was observed at time points similar to those for endogenous acaA (Fig. 3A). Significant expression of lacZ was also detected between 12 and 16 h of development in this experiment. We confirmed the induction of lacZ expression at the protein level. Activity of β-galactosidase was strongly induced by 12 h of starvation, i.e., with a delay of about 2 hours after mRNA induction (Fig. 3B). It is worth mentioning that β-galactosidase activity stayed high until slug stage (probably owing to the relative stability of the enzyme in D. discoideum cells), suggesting that the induction of lacZ expression under the control of acaAP(−739/+34) can be measured by comparing vegetative and slug-stage cells (see Fig. 7).

FIG. 3.

The acaA promoter fragment acaAP(−739/+34) was cloned as an EcoRI fragment into pDdGal-17 to produce an in-frame fusion of 11 amino acids from ACA with β-galactosidase. Transformed cells were developed on filters, and cell pellets were collected at the indicated time points. Slug stage was reached after 16 h. (A) Semiquantitative RT-PCR on the endogenous acaA mRNA, the lacZ mRNA expressed from acaAP(−739/+34), and cbfA (used to show the presence of intact mRNA in all preparations). (B) Activity of β-galactosidase in cell extracts prepared from 10⁷ cells.

FIG. 7.

Influence of homothymidine runs on acaA promoter activity. (A) DNA sequence of the acaA promoter after deletion of a 98-bp segment that spans the homothymidine runs. The deleted part is indicated by dashes; the BglII site that remained after deletion of the 98-bp segment is underlined. Underlining at the beginning and end of the sequence indicates EcoRI sites used to clone the promoter. (B) Assay for β-galactosidase activity in total cell extracts prepared from slug-stage AX2 cells transformed with wild-type acaAP(−739/+34) (n = 7; column 1) or acaAP(−739/+34Δ) (n = 7; column 2) and untransformed AX2 cells (n = 3; column 3). OD₅₅₀, optical density at 550 nm.

CbfA binds to the acaA upstream region.

We performed a survey of DNA-binding proteins with specificity for the subcloned acaAP promoter region. We separated the acaA upstream region into three parts to produce DNA fragments suitable for EMSAs. The DNA fragments were incubated with D. discoideum nuclear extracts prepared after 2 hours of starvation. Only the most distal fragment, acaAP(−739/−448) (Fig. 4A), produced reproducible gel shifts. The EMSAs were performed under the same experimental conditions used to study CbfA, and the mobility shifts produced on the acaAP(−739/−448) probe resembled complexes formed by CbfA on the C-module probe (compare Fig. 4B and C). Both the C-module-binding activity and the activity specific for acaAP(−739/−448) were strongly reduced in nuclear extracts prepared from CbfA-depleted mutant cells (Fig. 4B and C).

FIG. 4.

Identification of a DNA-binding activity specific for the acaA promoter. (A) DNA sequence of the acaA(−739/−448) promoter fragment. (B) EMSA with acaAP(−739/−448) as the probe. Increasing amounts of nuclear extract (NE400) prepared after 2 hours of development from AX2 cells (lanes 1 to 5) and cbfA^am cells (lanes 6 to 10) were added to the acaAP(−739/−448) probe. (C) EMSA with the C module as the radioactive probe. Amounts of nuclear extracts and lanes are as described above.

The results presented in Fig. 4 could be interpreted in two ways. The first interpretation is that the identified DNA-binding activity could be due to a protein whose expression depends on CbfA. The second is that the DNA-binding activity could be CbfA itself. Binding of the unknown nuclear protein to acaAP(−739/−448) showed several characteristics of CbfA binding to the C module, and we will show several lines of evidence to suggest that the protein that binds to acaAP(−739/−448) is in fact CbfA.

At high concentrations of nuclear extracts in the EMSAs, an additional low-mobility CbfA-DNA complex that most likely represents two molecules of CbfA bound simultaneously to the DNA fragment is observed with the C module (Fig. 4C and references 2 and 4). This was also observed with the acaAP(−739/−448) fragment, suggesting that two adjacent protein molecules bind to the same DNA (Fig. 4B). In EMSAs, the binding of CbfA to acaAP(−739/−448) was effectively competed with excess C module but not with a typical D. discoideum expression vector that contains several A+T-rich promoter and terminator sequences (pDneo2) (Fig. 5A). Thus, CbfA showed a remarkable binding preference for the acaA promoter and the C module while ignoring other model promoter sequences of the D. discoideum genome. In most experiments, we observed an affinity of CbfA to acaAP even higher than that of CbfA to the C module (e.g., the competition experiment shown in Fig. 5A), suggesting that longer thymidine runs may favor stronger binding of the factor to the target DNA. We have previously shown that CbfA binds to the C module by interacting with the narrow minor groove of the A+T-rich kernel of the target sequence (4). Similarly, we found that binding of CbfA to acaAP(−739/−448) was effectively competed by distamycin, which binds into the minor groove of A+T-rich DNA (Fig. 5B). By contrast, mithramycin, which prefers to bind into the minor groove of G+C-rich DNA, had no effect on CbfA-DNA complex formation on acaAP(−739/−448) (data not shown).

FIG. 5.

DNA-binding specificity of the factor that binds to acaAP(−739/−448). (A) Competition experiments with plasmid DNAs added to the acaAP(−739/−448) probe. All incubation mixes contained constant amounts of NE400 prepared from AX2 cells after 2 hours of development and 1 μg of poly(dAdT) · poly(dAdT) as nonspecific competitors. In addition, control mixes also contained 1 μg pGEM plasmid (lanes 1, 5, and 9). Competitor plasmids were 0.1 μg, 0.5 μg, and 1 μg of pGEM-C (C module) (lanes 2 to 4); pGEM-acaAP(−739/−448) (lanes 6 to 8); and pDneo2 (lanes 10 to 12). The total amount of pGEM DNA was kept constant at 1 μg in each lane. (B) Reaction mixtures contained 1 μg of poly(dAdT) · poly(dAdT), 1 μg pGEM plasmid, and the acaAP(−739/−448) probe. This DNA mixture was preincubated for 10 min with distamycin A at 0, 0.01, 0.05, 0.1, 0.5, 1, 5, and 10 μM final concentrations (lanes 13 to 20) before the protein fraction was added.

In footprinting analyses, CbfA protected two ∼30-bp regions with nearly homopolymeric 22- and 24-bp thymidine runs in the middle of each binding site (2). In the C module, the thymidines are located on the upper DNA strand with respect to the transcription direction. Several homopolymeric thymidine stretches are present in the acaA promoter (Fig. 2). Most pronounced are two oligo(dT) runs of 29 and 22 dT nucleotides at positions −612/−561 separated only by a single cytidine. Reasoning that the best candidate binding site for CbfA was the homothymidine motif located at positions −612/−561 of the acaA upstream region, we performed site-directed mutagenesis to place two runs of oligo(dA) in the middle of each of the two homothymidine motifs (Fig. 6A). This manipulation of the acaAP fragment should alter the DNA conformation of the CbfA target sequences, which is nicely illustrated by the EMSA shown in Fig. 6B, in which the free mutant DNA acaAP(−739/−448*) clearly migrated differently from the wild-type DNA fragment. After adding nuclear extract to the promoter fragments, we observed a >95% reduction of CbfA binding to the acaAP(−739/−448*) fragment compared to the wild type (Fig. 6B), supporting the central role of the homothymidine stretches in mediation of a characteristic local DNA conformation required for recognition by CbfA.

FIG. 6.

Involvement of homothymidine stretches in protein binding. (A) DNA sequence of the acaAP(−739/−448*) promoter fragment, in which the homothymidine stretches are interrupted by two 10-mer adenine runs (bold uppercase letters). EcoRI sites used to clone the promoter are underlined. (B) EMSA with increasing amounts of nuclear extract (NE400) prepared from AX2 cells after 2 hours of development. Radioactive probes were wild-type promoter fragment acaAP(−739/−448) (lanes 1 to 5) and mutant acaAP(−739/−448*) (lanes 6 to 10).

CbfA is the only protein that binds to the C module under our experimental conditions (2, 4). Therefore, we wanted to determine whether the putative CbfA binding site(s) in the acaA promoter could be replaced by the part of the C module that contains the CbfA binding sites. For this purpose, we first deleted a 98-bp region including the homothymidines from the acaAP(−739/+34) promoter (Fig. 7A) and linked the mutant promoter acaAP(−739/+34Δ) to the lacZ reporter gene. As shown in Fig. 7B, deletion of the CbfA binding site(s) reduced reporter gene expression in slug-stage cells to the background level, suggesting that binding of CbfA to this region of the acaA promoter is of functional significance for ACA expression. By use of acaAP(−739/−448Δ), it was confirmed that deletion of the CbfA binding sites in fact abolished binding of CbfA to this part of the acaA promoter (Fig. 8B). We then inserted the C module including the two CbfA binding sites into the mutant acaAP(−739/−448Δ) promoter fragment (Fig. 8A). This restored about 63% of the original CbfA binding level to the wild-type acaAP(−739/−448) promoter fragment (Fig. 8B), suggesting that CbfA binding sites can be shuffled between promoters and restore local DNA conformations being recognized by CbfA.

FIG. 8.

Exchange of the CbfA binding site of the acaA promoter for the C module. (A) DNA sequence of the EMSA probe acaAP(−739/−448^CM), which contains a part of the C module (uppercase bold letters) including the CbfA binding sites (underlined). Underlining at the beginning and end of the sequence indicates EcoRI sites used to clone the promoter. (B) EMSAs with increasing amounts of nuclear extract (NE400) prepared from AX2 cells prepared after 2 hours of development. Radioactive probes were wild-type promoter acaAP(−739/−448) (lanes 1 to 4), promoter acaAP(−739/−448Δ) carrying the 98-bp deletion (lanes 5 to 8), and acaAP derivative acaAP(−739/−448^CM) containing the C module instead of the deleted 98-bp segment (lanes 9 to 12).

To further demonstrate that the homothymidine stretches and surrounding sequences in region −612/−561 of the acaA promoter provide autonomous CbfA binding sites, we generated the 105-bp double-stranded DNA fragment shown in Fig. 9A. We observed that in EMSAs CbfA bound to this DNA fragment with high affinity (Fig. 9B) and with binding properties similar to those seen with the complete acaAP(−739/−448) fragment. We again observed an additional low-mobility shift representing two binding sites in this highly A+T-rich DNA sequence (Fig. 9B, lane 5). As expected, the described DNA-protein complexes were greatly reduced in nuclear extracts from cbfA^am cells (Fig. 9B, lanes 6 to 10).

FIG. 9.

Protein binding on isolated binding sites. (A) DNA sequence of the synthetic acaA promoter fragment acaAP(−612/−561). (B) The radiolabeled acaAP(−612/−561) DNA fragment was incubated under standard conditions with increasing amounts of NE400 fraction prepared from AX2 cells (lanes 1 to 5) and cbfA^am cells (lanes 6 to 10) after 2 hours of development.

To provide additional direct evidence to support the theory that CbfA binds to the acaA promoter, we performed a pull-down experiment with biotinylated DNA fragments. A nuclear extract fraction prepared from wild-type AX2 cells was used as the source for CbfA. As shown in Fig. 10, it was possible to pull down CbfA from nuclear extracts with biotinylated C module and acaAP(−739/−448) but not with acaAP(−739/−448*) that had a strongly reduced affinity to CbfA in EMSAs.

FIG. 10.

Pull-down experiment with biotinylated DNA probes. NE400 fraction prepared from AX2 cells after 2 hours of development (input) was incubated with biotinylated DNAs as described in Materials and Methods. Washed beads were boiled in Laemmli sample buffer, and proteins were separated in a 10% sodium dodecyl sulfate-polyacrylamide gel. After transfer of the proteins to a nitrocellulose membrane, CbfA was stained with monoclonal antibody 7F3. DNA probes used to pull down CbfA are indicated at the top of the figure.

DISCUSSION

CbfA regulates the acaA promoter.

In this report, we provide direct evidence that CbfA binds to the acaA upstream region and that this factor is required for the expression of ACA during aggregation of D. discoideum cells. Although considerable evidence has been collected in recent years pertaining to the genes involved in the transition from growth to development of D. discoideum cells and the genes required to achieve aggregation competence, little is known about the transcription factors that facilitate the reorganization of genetic networks in early development. The complete absence of acaA transcripts in CbfA-depleted cells starved for several hours has prompted us to investigate whether CbfA binds directly to the acaA upstream region. We found that a DNA-binding activity bound to the distal acaAP(−739/−448) fragment. This DNA-binding activity was absent in CbfA-depleted cells, suggesting that the factor is CbfA. The very unusual binding characteristics of the factor on the acaA promoter together with DNA-protein pull-down experiments provide compelling evidence that CbfA binds to the acaA promoter. In detailed biochemical studies of retrotransposon TRE5-A.1, we have never been able to show that proteins other than CbfA bind the C module. In this work, we show that CbfA binding sites in the C module can replace corresponding sites in the acaA promoter. On the other hand, we have isolated a highly A+T-rich DNA element from the acaA promoter that is bound by CbfA with high affinity (Fig. 9) and likely acts as a principal autonomous binding site of the factor similar to the C module. Deleting the part of the acaA promoter that is recognized by CbfA prevented induction of the mutant promoter in multicellular stages, suggesting a functional role for CbfA in ACA expression during development.

How does CbfA discriminate certain A+T-rich sequences in the background of the 78% A+T-rich D. discoideum genome? Our current view is that CbfA detects certain DNA conformations rather than strict DNA sequence motifs, and these critical DNA conformations may be provided only by homothymidine runs embedded in particular genomic sequences. This view is directly supported by the experiment in which the homothymidine stretches in the acaA promoter were disrupted by homoadenine, which led to a DNA fragment with altered migration in EMSAs and almost complete loss of affinity for CbfA (Fig. 6), and is also reflected by inhibition of CbfA binding to acaAP by distamycin, which is known to alter DNA conformations at high concentrations (compare to Fig. 5B). CbfA is a member of the “jumonji-type” transcription factors, which are thought to act as chromatin-remodeling enzymes (1). We speculate that CbfA is a general transcription activator for certain genes in D. discoideum cells. Although the recognition sites of CbfA in the genome are highly A+T rich, CbfA binds only to a limited set of genes in the D. discoideum genome, since preliminary proteome analyses suggest that fewer than 30 proteins are differentially expressed 2 hours after the starvation of cbfA^am cells (manuscript in preparation). Future work will focus on the isolation of other CbfA-regulated promoters and CbfA binding sites therein, thus providing us with more data to understand how the remarkable specificity of the factor in the background A+T-biased genome is facilitated.

CbfA and Myb2 may act synergistically to regulate ACA expression.

ACA-, Myb2-, and CbfA-depleted cells all show intriguingly similar developmental phenotypes. They are unable to aggregate on their own at moderate cell densities but can be helped to enter the multicellular stages by small numbers of wild-type cells. The mutants express cAMP pulse-induced genes only after exposure to artificial nanomolar cAMP pulses in shaken cultures and are rescued by ectopic expression of ACA. This suggests that Myb2 and CbfA are the principal regulators of ACA expression, but it is unclear whether the two factors act in a dependent or a cooperative manner. It is unlikely that CbfA expression depends on MybB, as no growth defects characteristic for CbfA-depleted cells (18) have been described for the mybB cells (11). On the other hand, previous DNA microarray analyses of developing cbfA^am cells (19) have clearly revealed wild-type mybB transcript levels. Thus, the best explanation is that Myb2 and CbfA act in independent pathways but act synergistically to mediate induction of ACA after starvation. We propose that CbfA provides a basal transcription competence of the acaA promoter that is required for full induction by a second factor, probably Myb2. If Myb2 is in fact the inducer of ACA expression, it would fail to enhance acaA transcription in the absence of the basal promoter activity mediated by CbfA. Although we have not detected binding to the cloned acaAP(−739/+34) promoter fragment of proteins other than CbfA, we cannot exclude the possibility that Myb2 binds in an even more distal region of the promoter. Our model of ACA induction does not necessarily require direct binding of Myb2 to the acaA promoter if induction was mediated by an as-yet-unknown factor whose expression depends on Myb2. Further work is required to identify Myb2 binding sites in the D. discoideum genome, particularly in the acaA upstream region.