CbfA, the C-Module DNA-Binding Factor, Plays an Essential Role in the Initiation of Dictyostelium...

EUKARYOTIC CELL, Oct. 2004, p. 1349–1358 Vol. 3, No. 51535-9778/04/$08.00�0 DOI: 10.1128/EC.3.5.1349–1358.2004Copyright © 2004, American Society for Microbiology. All Rights Reserved.

CbfA, the C-Module DNA-Binding Factor, Plays an Essential Role inthe Initiation of Dictyostelium discoideum Development

Thomas Winckler,1* Negin Iranfar,2 Peter Beck,1 Ingo Jennes,1 Oliver Siol,1 Unha Baik,1William F. Loomis,2 and Theodor Dingermann1

Institut fur Pharmazeutische Biologie, Universitat Frankfurt (Biozentrum), Frankfurt am Main, Germany,1 and Cell andDevelopmental Biology, Division of Biology, University of California—San Diego, La Jolla, California2

Received 22 April 2004/Accepted 23 June 2004

We recently isolated from Dictyostelium discoideum cells a DNA-binding protein, CbfA, that interacts in vitrowith a regulatory element in retrotransposon TRE5-A. We have generated a mutant strain that expresses CbfAat <5% of the wild-type level to characterize the consequences for D. discoideum cell physiology. We found thatthe multicellular development program leading to fruiting body formation is highly compromised in themutant. The cells cannot aggregate and stay as a monolayer almost indefinitely. The cells respond properly toprestarvation conditions by expressing discoidin in a cell density-dependent manner. A genomewide microar-ray-assisted expression analysis combined with Northern blot analyses revealed a failure of CbfA-depleted cellsto induce the gene encoding aggregation-specific adenylyl cyclase ACA and other genes required for cyclic AMP(cAMP) signal relay, which is necessary for aggregation and subsequent multicellular development. However,the cbfA mutant aggregated efficiently when mixed with as few as 5% wild-type cells. Moreover, pulsing cbfAmutant cells developing in suspension with nanomolar levels of cAMP resulted in induction of acaA and otherearly developmental genes. Although the response was less efficient and slower than in wild-type cells, it showedthat cells depleted of CbfA are able to initiate development if given exogenous cAMP signals. Ectopic expres-sion of the gene encoding the catalytic subunit of protein kinase A restored multicellular development of themutant. We conclude that sensing of cell density and starvation are independent of CbfA, whereas CbfA isessential for the pattern of gene expression which establishes the genetic network leading to aggregation andmulticellular development of D. discoideum.

Dictyostelium discoideum is a social amoeba that lives in soiland feeds on bacteria. When the cells sense environmentalconditions unfavorable for vegetative growth, they start tocommunicate with each other by means of extracellular cyclicAMP (cAMP). Individual cells move toward higher concentra-tions of cAMP and produce cAMP themselves, which is se-creted and serves to relay the signal in the aggregation field.Oscillatory secretion and perception of extracellular cAMPguide the cells into aggregation centers, in which they cooper-ate to form fruiting bodies consisting of stalk cells that supporta mass of dormant spores (2, 23).

As the density of growing cells increases, aggregation com-petence is acquired in response to secreted quorum-sensingproteins referred to as prestarvation factor (PSF) and condi-tioned-medium factor (CMF) (7). The PSF protein induces theexpression of a subset of early developmental genes, such asthe extracellular matrix discoidin proteins and the cAMP re-ceptor CAR1 (8, 32). Extracellular accumulation of PSF alsoinduces the expression of the protein kinase YakA, whichinterrupts the cell cycle and releases the translation block fromthe mRNA of the catalytic subunit of protein kinase A(PKA-C) (35, 36). CMF is not secreted by growing cells but israpidly released when the cells starve (21). CMF is an 80-kDaprotein that modulates signal transduction from CAR1 to itsdownstream effector, the aggregation-specific adenylyl cyclase

ACA. When CMF concentrations are low, ligand-induced ac-tivation of CAR1 cannot be transmitted to ACA even in thepresence of high cAMP concentrations (3, 4). Thus, the cellssense starvation of neighboring cells as a function of the CMFlevel and can proceed to aggregate (17).

CAR1 is a typical serpentine receptor that couples via trim-eric G proteins to several intracellular effectors, including ad-enylyl cyclase, guanylyl cyclase, and phospholipase C (23, 29).The gene encoding the activatable adenylyl cyclase ACA,acaA, is rapidly induced when the cells start to respond toextracellular cAMP (30). Activation of ACA is complex andrequires at least the G�� complex of heterotrimeric G pro-teins, the cytosolic regulator of adenylyl cyclase (CRAC), andcomponents of the Ras pathway (2, 23, 29). ACA activity is themajor source of cAMP produced in early stages of D. discoi-deum development (30). Most cAMP is secreted in order torelay the cAMP signal to nearby aggregation-competent cells.

PKA is an essential regulator of all stages of D. discoideumdevelopment (14, 25, 31, 33). PKA is a heterodimeric proteinconsisting of a catalytic and a regulatory subunit. CytoplasmiccAMP binds to the regulatory subunit of PKA, which releasesthe catalytic subunit to phosphorylate downstream substratesthat lead to alteration in gene expression (40). ExtracellularcAMP is degraded by the product of pdsA, a cAMP-specificphosphodiesterase, whose expression increases dramaticallyfollowing the initiation of development (13). This extracellularcAMP-specific phosphodiesterase is controlled by a specificinhibitor (PDI) whose expression is regulated in response tocAMP (15). Intracellular cAMP is broken down by the phos-phodiesterase RegA (34). The concerted action of many gene

* Corresponding author. Mailing address: Institut fur Pharmazeu-tische Biologie, Universitat Frankfurt (Biozentrum), Marie-Curie-Strasse 9, D-60439 Frankfurt, Germany. Phone: 49-69-798-29648. Fax:49-69-798-29662. E-mail: [email protected].

1349

on March 12, 2015 by guest

http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

products involved in the generation and perception of periodicnanomolar cAMP waves, including ACA, CAR1, PkaC, PdsA,RegA, and others, establishes a positive feedback loop thatleads to the activation of genes required for aggregation andsubsequent multicellular development. Inactivation of anygene involved in this regulatory network will result in mutantsthat cannot aggregate, and suppression of the aggregationblock with genes involved in the autofeedback loop can be usedto dissect the components of these regulatory pathways.

Genomewide DNA microarray analysis of expression pro-files is a powerful tool that extends our knowledge of thecomplex gene networks that regulate D. discoideum develop-ment (19, 20, 24, 39). Particularly helpful are genomewidecomparative studies of wild-type cells versus mutant cells inwhich a single known developmental gene is disrupted (19, 24).

Retrotransposons are autonomous genetic entities that canbe amplified and inserted at different chromosomal locationsin their hosts (11). Active retrotransposons are permanentsources of insertion mutagenesis. Thus, in compact genomeslike that of D. discoideum, retrotransposons must follow strat-egies to avoid disruption of host genes upon retrotransposition(45). One such strategy is the specific recognition of regionsflanking tRNA genes as integration sites (45). Studies of thetRNA gene-targeting retrotransposon TRE5-A has led to theisolation of a host-encoded protein that binds specifically to aDNA sequence at the 3� end of TRE5-A, the C module (16,18). This protein was named C-module-binding factor(CMBF). We have now changed the name of CMBF to CbfAin order to avoid confusion with CMF (conditioned-mediumfactor) and to conform to accepted Dictyostelium nomencla-ture. The DNA-binding properties of CbfA have been ana-lyzed in some detail, and the CbfA-encoding gene, cbfA, hasbeen isolated (18). The CbfA protein, as deduced from a re-cently revised version of the cbfA gene (GenBank accession no.AF052006), consists of 1,000 amino acids and contains a con-served “carboxy-terminal jumonji” domain implicated in chro-matin remodeling (10), three unusual zinc fingers, and an AThook that is required to bind to the C module in vitro (18).

The C module of retrotransposon TRE5-A serves as a pro-moter for the transcription of TRE5-A antisense RNAs, butthe role of these antisense RNAs in TRE5-A retrotransposi-tion is still obscure. We therefore attempted to generate CbfA-depleted mutants in order to better understand the role ofCbfA. We have not been able to isolate strains in which cbfAis deleted by conventional gene replacement approaches, per-haps because it is essential for growth, but we have been ableto generate mutants that express CbfA at levels �5% that ofwild-type cells by introducing a partially suppressed ambertranslational stop codon (44). Preliminary analyses of the phe-notype of cbfAam mutant cells revealed defects in growth anddevelopment (44) and also in TRE5-A retrotransposition (un-published data). We have now characterized the consequencesof CbfA depletion for D. discoideum development by using acombination of genomewide DNA microarray analyses andNorthern hybridizations. Although cbfAam cells show a prestar-vation response by inducing discoidin expression in a cell den-sity-dependent manner, they do not induce acaA when starvedon filters, which can account for their failure to aggregate orform multicellular structures. However, exposure of the mu-tant cells to artificial extracellular cAMP pulses induced the

early genes, including acaA. Mixing cbfAam cells with a fewwild-type cells also rescued development of the mutant cells,indicating that the block in morphogenesis occurs after theprestarvation response but before the cAMP positive feedbackloop is established.

MATERIALS AND METHODS

D. discoideum cell culture and development. D. discoideum AX2, AX4, andcbfAam mutant cells (strain JH.D) (44) were grown in shaken cultures either inliquid HL5 medium or in association with Klebsiella planticola bacteria. Foodbacteria were prepared as proposed by Clarke et al. (9). Multicellular develop-ment of D. discoideum cells was monitored on phosphate-buffered agar plates oron black filters (37). To study gene expression during development, 5 � 107 D.discoideum cells were spread on black nitrocellulose filters (AABP04700; 47-mmdiameter; Millipore) and allowed to develop for the time periods indicated in thefigures. The cells were then washed from the filters using phosphate buffer andcollected by centrifugation. Aliquots corresponding to 2 � 107 cells were storedas pellets at �80°C. Frozen cells were used to extract total RNA as describedbelow.

Expression vector construction and complementation studies. Expressionplasmids were transformed into cbfAam cells, and stable transformants wereselected in HL5 medium containing 5 �g of G418/ml. A nearly full-length cbfAcDNA was reconstructed using a combination of cloned and PCR-amplifiedgenomic DNA and cDNA fragments. The cDNA was inserted into the KpnIrestriction site of vector pDXA-3H (26) to create the expression vector pDXA-rCbfA. The protein product derived from this vector, named recombinant CbfA(rCbfA), was identical to authentic CbfA protein except that it lacked the twoC-terminal isoleucine residues. Control cells were transformed with emptypDXA-3H vector. The pkaC cDNA was either expressed from its endogenouspromoter (plasmid p188) (1) or under the control of the constitutively active act6promoter (plasmid p332) (12).

Induction of cAMP-induced gene expression by cAMP pulses. D. discoideumAX2, AX4, and cbfAam cells were grown to densities of 3 � 106/ml in axeniccultures. The cells were washed and adjusted to 2 � 107/ml in phosphate buffer.The cultures were shaken at 150 rpm at 22°C. cAMP was added at a 30 nM finalconcentration at 6-min intervals, starting 2 h after the cells were washed. After6 h, a single pulse of 300 �M cAMP followed. Aliquots of 2 � 107 cells werecollected and frozen as cell pellets at �80°C for further use.

Northern blots. DNA probes specific for several developmental genes weregenerated by PCR using DNA sequence information from GenBank entries. Thefollowing genes were used for hybridization experiments: acaA (L05499), carA(M21824), csaA (X66483), pkaC (M38703), lagC (U09478), and cotB (M26238).The D. discoideum histone H3 gene (hstC) (6) was used as a loading control onNorthern blots. Information on genes SLJ247 and SSD449 was obtained from theDictyostelium genome project (http://dicty.sdsc.edu). PCR fragments used forhybridization experiments were purified from agarose gels using the QIAquickGel Extraction kit (QIAGEN). DNA fragments were verified by DNA sequenc-ing using the PCR primers. 32P-labeled probes were generated with the PCRprimer complementary to the respective mRNA sequence by using the TaqCyclist DNA Sequencing kit from Stratagene. A 32P-labeled probe specific fordiscoidin I was prepared by nick translation (42).

Total RNA was prepared from 2 � 107 frozen cells using the RNeasy Mini kitfrom QIAGEN; 40 �g of total RNA was separated in denaturing gels, and RNAswere blotted onto Hybond-N membranes. Radioactive probes were hybridized toimmobilized RNA at 42°C overnight in 50% formamide–5� SSC (1� SSC is 0.15M NaCl plus 0.015 M sodium citrate)–1% sodium dodecyl sulfate–1� Den-hardt’s solution. The blots were washed for 30 min at ambient temperature in 2�SSC and for 30 min at 65°C in 2� SSC–0.1% sodium dodecyl sulfate.

The primers used to generate PCR probes were as follows: acaA-01, GTTCATCCTATGGTATGAAATTGG; aca-02, GTAGTGAATGAGCCAATTTCACCC; arA-01, CCAGCCAATGAAACATCATTGG; carA-02, GATGATGATAAAGAAGATGAAGATG; csaA-01, CATTACAGGTACTGGATTTACAG;csaA-02, CCATTGTGAGGTGCTTGAGTGAC; pkaC-01, CCACCACCAGTCAATGCAAGAGAAAG; pkaC-02, CATAGAAAGGTGGATAACCTGCCAAC; cotB-01, GGAGTAGTTGTACACCAAGTAGTGGTTTC; cotB-02, CTTATGGGAACCAATCCAGCCACCTTTTGG; lagC-01, CTCAAGATTATGGGGTAGTATAGAC; lagC-02, CCAGTACCGTATGGTGTTGGACATG;SLJ247-01, GTCAGTGGTGAATGTGCCATTGATTTCTC; SLJ247-02, GGATGATCCACTTCCTGTACCACTTCC; SSD449-01, ATGGCCAGAATTGATTATGCTGTAAGC; and SSD449-02, CCAGTTTCAATGGTTTTGAAATAGCCC.

1350 WINCKLER ET AL. EUKARYOT. CELL


http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

Microarray analyses. Corning slides were microarrayed with 6,345 cDNA andgenomic targets as previously described (19). Inserts from 5,655 cDNAs weregenerously provided by the Japanese EST Project (27, 38). A list of genes isavailable at http://www.biology.ucsd.edu/loomis-cgi/microarray/cbfA_array.html.All genes in this study were sequence verified. Probes were prepared from totalRNAs collected at 2-h intervals during wild-type and mutant development, aswell as from time-averaged reference RNA. Superscript II DNA polymerase(Invitrogen, Carlsbad, Calif.) was used to incorporate either Cy5- or Cy3-conju-gated dCTP (Amersham, St. Louis, Mo.) into DNA. Labeled probes from thetime points and the time-averaged RNA were mixed and hybridized at 42°C tomicroarrays for 6 to 12 h before being analyzed with a Genepix 4000B scanner(Axon Instruments, Foster City, Calif.). The total Cy3 signal was normalized tothe total Cy5 signal after background subtraction to allow independent slides tobe compared. The Cy3/Cy5 ratios for individual genes were then calculated ateach time point for subsequent analyses. The expression patterns of mutant cellsdepleted in CbfA were compared to those of strain AX4, which are indistin-guishable from those of strain AX2 (39).

RESULTS

CbfA-depleted cells are unable to aggregate. CbfA was ini-tially identified and purified by its specific binding to the Cmodule of the retrotransposon TRE5-A (16). Using the highlyspecific monoclonal antibody 7F3 directed against the carboxy-terminal domain of CbfA (44), we showed that the CbfA pro-tein level in AX2 cells is not sensitive to growth conditions(axenic culture or growth on bacteria) and stays constantthroughout development. Using this same antibody, we foundthat the level of CbfA protein in cbfAam cells was �5% of thatin wild-type cells during growth or following the initiation ofdevelopment (reference 44 and data not shown).

We examined the effect of CbfA depletion in cbfAam cells onthe multicellular development of D. discoideum cells. WhencbfAam cells grew on lawns of bacteria, they formed very few, ifany, fruiting bodies, reflecting poor aggregation of the mutantcells, leading to multicellular structures generated by only afew of the starving cells (44). When mutant cells were platedon phosphate-buffered agar at 1.6 � 106/cm2, the cbfAam cellsdid not show any sign of aggregation for at least 36 h and onlyvery few, mostly crippled, fruiting bodies were observed after72 h (data not shown). We could recover a few spores fromthese plates, and they gave rise to populations of mutant cellswhen spread on bacterial plates. cbfAam cells showed the ag-gregation defect at all cell densities tested (5 � 104 to 9 � 106

cells/cm2).The failure of cbfAam cells to aggregate might be the con-

sequence of their inability to sense their own density and foodstatus. Therefore, we determined the cell density-dependentinduction of discoidin expression as a measure of the prestar-vation response (8, 9). Cells were grown in shaken cultures inthe presence of bacteria and collected at various titers forNorthern analyses of accumulation of discoidin I mRNA (Fig.1). The cbfAam cells showed a pronounced lag in growth onbacteria as a consequence of impaired phagocytosis (44). Nev-ertheless, cbfAam cells showed almost normal, cell density-dependent expression of discoidin mRNA (Fig. 1) and protein(not shown), indicating that the prestarvation response ofcbfAam cells is not affected by CbfA deficiency. These resultswere confirmed with axenically growing cbfAam cells, in whichdiscoidin mRNA levels increased to 106 cells/ml and showedthe characteristic repression at cell densities of 3 � 106

cells/ml also observed with AX2 cells (43) (data not shown).

cbfAam cells fail to activate cAMP pulse-induced genes inearly development. Although cbfAam cells reacted to prestar-vation conditions, they did not acquire aggregation compe-tence and thus stayed as a monolayer. The ability of a cellpopulation to respond to and relay cAMP signals can be ana-lyzed at the transcription level by measuring the induction ofcAMP pulse-induced genes during aggregation. AX2 andcbfAam cells were spread on phosphate buffer-supported filters,and total RNA prepared from samples was collected over 14 h.Northern blots were hybridized with DNA probes specific forthe cAMP pulse-induced genes carA, acaA, csaA, and pkaC(Fig. 2). Expression of the acaA gene was completely absent incbfAam cells. The genes carA and csaA were induced to verylow levels 2 h after starvation of the mutant cells. On Westernblots, cbfAam cells did not express any detectable CsaA proteinby 11 h after starvation on filters (data not shown). The mRNAencoding the catalytic subunit of PKA (pkaC) was fairly wellexpressed in growing and developing cbfAam cells, although thetypical cAMP pulse-induced increase in pkaC expression seenat 8 h was not observed in the mutant cells (Fig. 2). We alsotested for the expression of the postaggregative, cell-type-spe-cific genes cotB and lagC. Both markers were well expressed inAX2 cells 8 to 14 h after starvation but were absent in thecbfAam cells under these conditions (Fig. 2).

The Northern blot data suggested that the gene regulatorynetworks that establish and maintain cAMP relay may not beactive in cbfAam cells. We used genomewide DNA microarrayscarrying 6,345 targets (19) to characterize the expression ki-netics of developmentally regulated genes in cbfAam cells. Wefocused on the expression patterns of 118 genes that can be

FIG. 1. Cell density-dependent expression of discoidin I. AX2 andcbfAam cells were grown in shaken cultures in association with bacteria.At time points 1 to 4, indicated by the arrows, D. discoideum cells werecollected, washed, and analyzed for discoidin I mRNA expression. Theindicated samples 1 to 4 corresponded to cell densities of 1.0 � 106, 2.0� 106, 4.4 � 106, and 7.9 � 106/ml for AX2 and 1.0 � 106, 2.4 � 106,3.8 � 106, and 7.7 � 106/ml for strain JH.D. Total RNA was preparedfrom the cell samples and analyzed on a Northern blot with a discoidinI�-specific probe (inset).

VOL. 3, 2004 CbfA REGULATES DICTYOSTELIUM DEVELOPMENT 1351


http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

grouped into clusters according to their successive induction atdifferent stages of multicellular development (19) (Fig. 3A).When cbfAam cells were developed on filters, they did not showsignificant induction of any of the developmentally regulatedgene clusters (Fig. 3B). Close inspection of the individual ex-pression profiles of the early developmental genes tested onNorthern blots revealed that expression of carA was inducedonly 3-fold in cbfAam cells (compared to 38-fold in wild-typecells) and then stayed constant throughout development atonly 5% of the wild-type level (not shown). In wild-type cells,csaA expression increased eightfold 8 h after starvation but wasnot significantly induced in cbfAam cells. The gene acaA wasnot expressed at all in cbfAam cells. The only gene from the set

of developmentally regulated genes that was expressed in amanner similar to that in the wild type was pdiA, the phospho-diesterase inhibitor that is repressed by cAMP pulses.

Developmental rescue of cbfAam cells. The aggregation de-fect of cbfAam cells most likely results from their inability tosynthesize cAMP at a high rate as a consequence of failing toexpress acaA. However, we speculated that they might be ableto respond to cells that are making cAMP. We mixed a smallnumber of AX2 cells with cbfAam cells and found that thecbfAam cells developed significantly better in the presence ofwild-type cells (Fig. 4). Spores isolated from the fruiting bodieswere plated clonally on bacterial lawns, and most gave rise toclones with the cbfAam phenotype, thus demonstrating that the

FIG. 2. Expression of developmentally regulated genes in the cbfAam mutant JH.D developed on solid support. AX2 and cbfAam cells of strainJH.D were washed free of medium and spread on black filters at 5 � 107 per filter. Cell samples were collected every 2 h. RNA was extracted fromthe samples, and 40 �g of total RNA was loaded in each slot. The origins and preparation of 32P-labeled probes are described in Materials andMethods.

FIG. 3. Microarray analysis of developmentally regulated genes in AX4 and cbfAam mutant JH.D developed on solid support. Samples werecollected every 2 h for 18 h and analyzed on chips carrying 6,345 targets. Expression profiles of 118 developmentally regulated genes were extractedand clustered into 9 groups. Each cluster contains 10 genes and is color coded. A) wild-type AX4. B) cbfAam mutant JH.D.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

cbfAam cells had participated in forming the observed fruitingbodies (data not shown).

To determine if cbfAam cells developed in response to ex-tracellular cAMP signals provided by the wild-type cells, weincubated AX2 and mutant cells in buffer and gave them arti-ficial cAMP pulses at 6-min intervals for 4 h, starting 2 h afterthey were washed, and a final boost of 300 �M cAMP after 6 h.RNA was collected at 2-h intervals and analyzed on microar-rays. Unlike cbfAam cells developed on filters, those developedin suspension and given cAMP pulses expressed the earlygenes, including acaA (Fig. 5). Accumulation of many of theseearly gene products was delayed 4 h and was somewhatattenuated compared to that of wild-type cells (Fig. 5). Thus,CbfA appears to play a role in early gene expression but can bepartially bypassed if the cells are given exogenous cAMPpulses.

The expression of some critical cAMP pulse-induced geneswas analyzed on Northern blots to verify the microarray data.In the wild-type AX2 cells, the genes carA, acaA, and csaAwere strongly induced by exogenous cAMP pulses and reachedmaximum expression 4 to 6 h after the beginning of starvation(Fig. 6). These genes were also induced in the mutant cells,albeit more slowly (Fig. 6). Expression of pkaC was almostnormal in the cbfAam cells under these conditions. Thus, the

time course of gene expression seen on Northern blots con-firmed the DNA microarray data for these genes.

In the DNA microarray experiments, we noticed two genesthat were induced in cbfAam cells but not in AX4 cells after theonset of starvation. The gene SLJ247, which encodes CF50, asubunit of counting factor (5), was induced 6-fold (SLJ247),whereas expression of SSD449, an orphan gene, was induced2-fold (data not shown). We analyzed the expression profilesof these two genes on Northern blots of cAMP-pulsed cells tosee if they were in fact overexpressed in cbfAam cells comparedto wild-type cells (Fig. 6). It turned out that both genes were infact underexpressed in growing cbfAam cells compared to AX2and were strongly induced soon after the beginning of starva-tion, thus giving an apparently stronger induction in the mu-tant than in the wild type. Both genes reached about the sameexpression levels in both wild-type and cbfAam cells by 2 h afterstarvation (Fig. 6). Notably, cAMP-induced repression ofSLJ247 was delayed in the mutant strain. Apparently, under-expression of the gene encoding CF50 in growing cells is un-likely to contribute to the phenotype of cbfAam cells, since cf50null cells form giant streams and fruiting bodies (5). Neverthe-less, the fact that SSD449 and SLJ247 were induced in acAMP-independent manner within the first 2 h following theremoval of nutrients, well before the cells were treated with

FIG. 4. Development of the cbfAam mutant in synergy with wild-type cells. Axenically cultured D. discoideum cells were washed and plated onphosphate-buffered agar. The 100% value corresponds to 108 plated cells per 9-cm-diameter petri dish. The pictures were taken 60 h after the cellswere plated. Scale bar: 1 mm.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

FIG. 5. Microarray analysis of developmentally regulated genes in AX4 and cbfAam mutant JH.D developed in shaken culture with cAMPpulses. Samples were collected every 2 h for 18 h and analyzed on chips carrying 6,345 targets. Expression profiles of 118 developmentally regulatedgenes were extracted and clustered into 9 groups. Each cluster contains 10 genes and is color coded. A) wild-type AX4. B) cbfAam mutant JH.D.

FIG. 6. Northern analysis of developmentally regulated genes in the cbfAam mutant JH.D in shaken culture treated with cAMP. Axenicallygrown AX2 and cbfAam cells were harvested at 3.0 � 106/ml, washed free of medium, and suspended at 2 � 107/ml in phosphate buffer. The cellswere shaken at 150 rpm at 22°C. Beginning 2 h after the cells were washed, cAMP pulses were applied every 6 min at a 30 nM final concentration,followed by a final boost of 300 �M cAMP after 6 h. RNA was extracted from cell samples, and 40 �g of total RNA was loaded in each slot. Theorigins and preparation of 32P-labeled probes are described in Materials and Methods.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

cAMP pulse, clearly indicates that cbfAam cells are able tosense starvation.

We also tested two postaggregation genes, lagC and cotB, forexpression in cbfAam cells after cAMP pulsing. As shown inFig. 6, cotB expression was absent under the experimentalconditions used and lagC expression was only slightly inducedand barely detectable.

When cAMP-treated cbfAam cells were plated after 8 h onphosphate-buffered agar plates to allow multicellular develop-ment, they were able to complete development significantlyfaster and more efficiently than mutant cells that had not beenexposed to exogenous cAMP (data not shown). However, thedevelopmental timing of cbfAam cells was still aberrant, sincefruiting bodies were not formed until 31 h after plating, 6 hlater than the formation of wild-type fruiting bodies. More-over, there were fewer fruiting bodies due to inefficient entry ofmutant cells into aggregates (data not shown).

The developmental phenotype of cbfAam is complemented byrecombinant CbfA. To formally prove that the observed devel-opmental defects of cbfAam cells were exclusively due to CbfAdeficiency, we expressed a nearly full-length cbfA cDNA in the

mutant cells under the control of the strong, constitutivelyactive actin15 promoter. Using the CbfA-specific antibody 7F3,we found several cbfAam transformants that accumulated sub-stantial levels of rCbfA (data not shown). These clones alldeveloped normally and gave rise to normal-size fruiting bod-ies (Fig. 7A). Moreover, expression of early developmentalgenes, such as csaA, was returned to normal by expression ofrecombinant CbfA (Fig. 7B).

Constitutive expression of PKA-C restores development ofcbfAam cells. The aggregation defect of the cbfAam mutantresembled in several respects the phenotype of acaA mutantcells. Wang and Kuspa (41) have shown that acaA mutant cellscan aggregate when they overexpress pkaC, the gene encodingthe catalytic subunit of protein kinase A (41). To determinewhether the development of cbfAam mutant cells can be re-stored by expression of pkaC, we transformed cbfAam cells withplasmids that allowed the expression of the pkaC cDNA eitherunder the control of the authentic pkaC promoter or under theconstitutively active actin6 promoter. Expression of PKA-Cfrom both plasmids reproducibly overcame the aberrant devel-opment of cbfAam cells (shown in Fig. 8 for act6/pkaC), sug-

FIG. 7. Complementation of the cbfAam phenotype by expression of recombinant CbfA. (A) D. discoideum cells were plated for multicellulardevelopment on phosphate-buffered agar plates. cbfAam cells were transformed with plasmid pDXA-rCbfA (cbfAam/rCbfA). Control AX2 andmutant cells were transformed with the empty vector pDXA-3H. The pictures were taken 40 h after the cells were plated. Scale bar: 1 mm.(B) Transformants were spread on black filters and allowed to aggregate as described in the legend to Fig. 2. Cell samples were collected every2 h. RNA was extracted from the samples, and 40 �g of total RNA was loaded in each slot. Blotted RNAs were hybridized to a radiolabeled csaAprobe.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

gesting that accumulation of constitutive PKA-C can bypassthe consequences of the lack of ACA caused by depletion ofCbfA.

DISCUSSION

CbfA is a sequence-specific DNA-binding protein that isessential for correct expression of acaA and other genes re-quired to aggregate and develop following starvation and dep-osition of Dictyostelium cells on a moist support. CbfA containsprotein motifs that likely act in DNA binding and transactiva-tion of gene expression. The zinc fingers of CbfA may contrib-ute to DNA binding and/or stabilization of the CbfA structure.The carboxy-terminal domain of CbfA contains an AT hookthat is required for binding to the C module of retrotransposonTRE5-A in vitro (18) and that may also mediate binding toAT-rich binding sites in CbfA target gene promoters. Theamino-terminal one-third of CbfA contains a jumonji domainthat is essential for the regulation of early developmental genes(unpublished data). Hence, we speculate that CbfA is a tran-scription regulator that binds to AT-rich target sites in the D.

discoideum genome and regulates the transcription of CbfA-dependent genes.

CbfA seems to be particularly important for the expressionof the aggregation-specific adenylyl cyclase ACA. At present,we cannot resolve whether CbfA regulates the acaA gene di-rectly or indirectly. CbfA recognizes two unusual binding sitesin the C module of the retrotransposon TRE5-A. These consistof 24-bp oligoadenine sequences surrounded by noncon-served nucleotides. There are similar oligoadenine runs in the500 bp preceding the acaA gene where CbfA might bind. How-ever, given the generally high A�T content of the D. discoi-deum genome, it can be tested experimentally only if CbfAbinds to the acaA promoter in vivo. In the absence of acaAgene expression, the cells cannot generate the extracellularpulses of cAMP that are necessary for chemotactic aggregationand stimulation of adjacent cells. As a result, CbfA-depletedcells fail to aggregate or to accumulate normal levels of earlydevelopmental mRNAs. However, if CbfA-depleted cells aredeveloped in mixtures with wild-type cells or are given artificialpulses of cAMP while suspended in buffer, they can expressacaA and other early developmental genes and develop intofairly normal fruiting bodies. However, both gene expressionand fruiting body formation were delayed in cbfAam cells underthese conditions, indicating that the role of CbfA cannot becompletely bypassed by artificial cAMP pulses.

CbfA does not appear to be necessary for the prestarvationresponse, since CbfA-depleted cells express discoidin I in amanner determined by the cell density, as in wild-type cells. Inwild-type cells, PSF induces the serine-threonine protein ki-nase YakA, which arrests growth and releases a translationblock on pkaC mRNA (35, 36). Our data suggest that theYakA pathway functions normally in cbfAam cells, since yakAmutant cells show enhanced growth rates (whereas cbfAam cellsgrow slowly), and cAMP-pulsed yakA mutant cells express nei-ther carA nor acaA (both of which are induced in cAMP-pulsed cbfAam cells) (Fig. 6). CbfA probably acts downstreamof the YakA pathway and, either directly or indirectly, controlstranscription of acaA and other early genes. When the cells areable to generate and respond to cAMP pulses, a CbfA-inde-pendent transcriptional pathway can take over.

CbfA does not appear to be necessary for cells to sensestarvation, since expression of some very early genes increasesimmediately after mutant cells are washed free of nutrients.The major defects of cbfAam cells can be attributed to the lackof ACA, and like those of acaA null mutant cells, they can bebypassed by the constitutive PKA activity that results fromoverexpression of the catalytic subunit encoded by pkaC. Itappears that all of the early internal signaling by cAMP ismediated by PKA. When cbfAam cells are treated with pulsesof cAMP, the MAP kinase ERK2 is transiently activated andinhibits the internal phosphodiesterase so that cAMP gener-ated by the other adenylyl cyclase ACR can accumulate (19).This stimulates PKA activity, which induces acaA and otherearly developmental genes, allowing the cbfAam cells to escapetheir trap.

A Myb-related transcription regulator (Myb2) that is alsorequired for induction of acaA (28) has been described. LikecbfAam cells, those lacking mybB fail to aggregate but can berescued by artificial pulses of cAMP. It is possible that in theabsence of high levels of PKA activity, both CbfA and Myb2

FIG. 8. Complementation of the cbfAam phenotype by expressionof recombinant PKA-C. cbfAam cells were transformed with plasmidp332 (12), so that expression of pkaC was under the control of theactin6 promoter (cbfAam/PKA-C cells). The cells were plated on phos-phate-buffered agar plates. The pictures were taken 40 h after the cellswere plated. Scale bar: 1 mm.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

are necessary to induce their target genes. Another gene,amiB, is known to be necessary for aggregation but can bebypassed by cAMP pulsing or ectopic expression of Myb2 (22).It appears to act upstream of CbfA and Myb2, since amiB nullmutant cells fail to repress the vegetative gene cprD whenstarved, suggesting that they cannot sense starvation (22). TheAmiB protein contains a domain that is significantly related toMAP kinase phosphatases and may be involved in the regula-tion of YakA, which acts upstream of PKA.

Although pulsing with cAMP partially bypassed the block todevelopmental gene expression in cbfAam cells, expression ofthese genes was generally reduced and delayed by 2 to 3 h (Fig.5 and 6). Since cbfAam cells developed quite efficiently whenmixed with wild-type cells, it may be that wild-type cells pro-vided a secreted factor missing in the cbfAam mutant in addi-tion to cAMP. This factor is unlikely to be CMF, since wefound that cmfA is normally expressed in cbfAam cells (data notpresented). However, we cannot presently exclude the possi-bility that a component of the CMF-controlled signaling path-ways is dependent on CbfA, so that the cbfAam cells are unableto respond properly to CMF.

ACKNOWLEDGMENTS

We are grateful to C. Utner and D. Fuller for expert technicalassistance. We thank B. Wetterauer, G. Gerisch, and C. Reymond forgifts of plasmids and antibodies.

This work was supported by grants from the Deutsche Forschungs-gemeinschaft to T.D. and T. W. (WI 1142/2-1 and WI 1142/2-2) andgrants from the NIH (GM60447 and GM62350) to W.F.L.

REFERENCES

1. Anjard, C., S. Pinaud, R. R. Kay, and C. D. Reymond. 1992. Overexpressionof DdPK2 protein kinase causes rapid development and affects the intracel-lular cAMP pathway of Dictyostelium discoideum. Development 115:785–790.

2. Aubry, L., and R. Firtel. 1999. Integration of signaling networks that regulateDictyostelium differentiation. Annu. Rev. Cell Dev. Biol. 15:469–517.

3. Brazill, D. T., D. F. Lindsey, J. D. Bishop, and R. H. Gomer. 1998. Celldensity sensing mediated by a G protein-coupled receptor activating phos-pholipase C. J. Biol. Chem. 273:8161–8168.

4. Brazill, D. T., R. Gundersen, and R. H. Gomer. 1997. A cell-density sensingfactor regulates the lifetime of a chemoattractant-induced G�-GTP confor-mation. FEBS Lett. 404:100–104.

5. Brock, D. A., K. Ehrenman, R. Ammann, Y. Tang, and R. H. Gomer. 2003.Two components of a secreted cell number-counting factor bind to cells andhave opposing effects on cAMP signal transduction in Dictyostelium. J. Biol.Chem. 278:52262–52272.

6. Bukenberger, M., J. Horn, T. Dingermann, R. P. Dottin, and T. Winckler.1997. Molecular cloning of a cDNA encoding the nucleosome core histoneH3 from Dictyostelium discoideum by genetic screening in yeast. Biochim.Biophys. Acta 1352:85–90.

7. Clarke, M., and R. H. Gomer. 1995. PSF and CMF, autocrine factors thatregulate gene expression during growth and early development of Dictyoste-lium. Experientia 51:1124–1134.

8. Clarke, M., S. C. Kayman, and K. Riley. 1987. Density-dependent inductionof discoidin-I synthesis in exponentially growing cells of Dictyostelium dis-coideum. Differentiation 34:79–87.

9. Clarke, M., J. Yang, and S. Kayman. 1988. Analysis of the prestarvationresponse in growing cells of Dictyostelium discoideum. Dev. Genet. 9:315–326.

10. Clissold, P. M., and C. P. Ponting. 2001. JmjC: cupin metalloenzyme-likedomains in jumonji, hairless and phospholipase A2�. Trends Biochem. Sci.26:7–9.

11. Craig, N. L., R. Craigie, M. Gellert, and A. M. Lambowitz. 2002. MobileDNA II. ASM Press, Washington, D.C.

12. Dammann, H., F. Traincard, C. Anjard, M. X. P. van Bemmelen, C. Rey-mond, and M. Veron. 1998. Functional analysis of the catalytic subunit ofDictyostelium PKA in vivo. Mech. Dev. 72:149–157.

13. Faure, M., J. Franke, A. L. Hall, G. J. Podgorski, and R. H. Kessin. 1990.The cyclic nucleotide phosphodiesterase gene of Dictyostelium discoideumcontains 3 promoters specific for growth, aggregation, and late development.Mol. Cell. Biol. 10:1921–1930.

14. Firtel, R. A., and A. L. Chapman. 1990. Role for cAMP-dependent proteinkinase-A in early Dictyostelium development. Genes Dev. 4:18–28.

15. Franke, J., M. Faure, L. Wu, A. L. Hall, G. J. Podgorski, and R. H. Kessin.1991. Cyclic nucleotide phosphodiesterase of Dictyostelium discoideum andits glycoprotein inhibitor—structure and expression of their genes. Dev.Genet. 12:104–112.

16. Geier, A., J. Horn, T. Dingermann, and T. Winckler. 1996. A nuclear proteinfactor binds specifically to the 3�-regulatory module of the long-interspersed-nuclear-element-like Dictyostelium repetitive element. Eur. J. Biochem. 241:70–76.

17. Gomer, R. H. 1999. Cell density sensing in a eukaryote. ASM News 65:23–29.18. Horn, J., A. Dietz-Schmidt, I. Zundorf, J. Garin, T. Dingermann, and T.

Winckler. 1999. A Dictyostelium protein binds to distinctoligo(dA) � oligo(dT) DNA sequences in the C-module of the retrotranspos-able element DRE. Eur. J. Biochem. 265:441–448.

19. Iranfar, N., D. Fuller, and W. F. Loomis. 2003. Genome-wide expressionanalyses of gene regulation during early development of Dictyostelium dis-coideum. Eukaryot. Cell 2:664–670.

20. Iranfar, N., D. Fuller, R. Sasik, T. Hwa, M. Laub, and W. F. Loomis. 2001.Expression patterns of cell-type-specific genes in Dictyostelium. Mol. Biol.Cell 12:2590–2600.

21. Jain, R., I. S. Yuen, C. R. Taphouse, and R. H. Gomer. 1992. A density-sensing factor controls development in Dictyostelium. Genes Dev. 6:390–400.

22. Kon, T., H. Adachi, and K. Sutoh. 2000. amiB, a novel gene required for thegrowth/differentiation transition in Dictyostelium. Genes Cells 5:43–55.

23. Loomis, W. F. 1996. Genetic networks that regulate development in Dictyo-stelium cells. Microbiol. Rev. 60:135.

24. Maeda, M., H. Sakamoto, N. Iranfar, D. Fuller, T. Maruo, S. Ogihara, T.Morio, H. Urushihara, Y. Tanaka, and W. F. Loomis. 2003. Changing pat-terns of gene expression in Dictyostelium prestalk cell subtypes recognized byin situ hybridization with genes from microarray analyses. Eukaryot. Cell2:627–637.

25. Mann, S. K. O., J. M. Brown, C. Briscoe, C. Parent, G. Pitt, P. N. Devreotes,and R. A. Firtel. 1997. Role of cAMP-dependent protein kinase in control-ling aggregation and postaggregative development in Dictyostelium. Dev.Biol. 183:208–221.

26. Manstein, D. J., H. P. Schuster, P. Morandini, and D. M. Hunt. 1995.Cloning vectors for the production of proteins in Dictyostelium discoideum.Gene 162:129–134.

27. Morio, T., H. Urushihara, T. Saito, Y. Ugawa, H. Mizuno, M. Yoshida, R.Yoshino, B. N. Mitra, M. Pi, T. Sato, K. Takemoto, H. Yasukawa, J. Wil-liams, M. Maeda, I. Takeuchi, H. Ochiai, and Y. Tanaka. 1998. The Dictyo-stelium developmental cDNA project: generation and analysis of expressedsequence tags from the first-finger stage of development. DNA Res. 5:335–340.

28. Otsuka, H., and P. J. M. van Haastert. 1998. A novel Myb homolog initiatesDictyostelium development by induction of adenylyl cyclase expression.Genes Dev. 12:1738–1748.

29. Parent, C. A., and P. N. Devreotes. 1996. Molecular genetics of signal trans-duction in Dictyostelium. Annu. Rev. Biochem. 65:411–440.

30. Pitt, G. S., N. Milona, J. Borleis, K. C. Lin, R. R. Reed, and P. N. Devreotes.1992. Structurally distinct and stage-specific adenylyl cyclase genes play dif-ferent roles in Dictyostelium development. Cell 69:305–315.

31. Primpke, G., V. Iassonidou, W. Nellen, and B. Wetterauer. 2000. Role ofcAMP-dependent protein kinase during growth and early development ofDictyostelium discoideum. Dev. Biol. 221:101–111.

32. Rathi, A., and M. Clarke. 1992. Expression of early developmental genes inDictyostelium discoideum is initiated during exponential growth by an auto-crine-dependent mechanism. Mech. Dev. 36:173–182.

33. Schulkes, C., and P. Schaap. 1995. cAMP-dependent protein kinase activityis essential for preaggregative gene expression in Dictyostelium. FEBS Lett.368:381–384.

34. Shaulsky, G., D. Fuller, and W. F. Loomis. 1998. A cAMP-phosphodiester-ase controls PKA-dependent differentiation. Development 125:691–699.

35. Souza, G. M., A. M. da Silva, and A. Kuspa. 1999. Starvation promotesDictyostelium development by relieving PufA inhibition of PKA translationthrough the YakA kinase pathway. Development 126:3263–3274.

36. Souza, G. M., S. Lu, and A. Kuspa. 1998. YakA, a protein kinase required forthe transition from growth to development in Dictyostelium. Development125:2291–2302.

37. Sussman, M. 1987. Cultivation and synchronous morphogenesis of Dictyo-stelium under controlled experimental conditions. Methods Cell Biol. 28:9–29.

38. Urushihara, H., T. Morio, T. Saito, Y. Kohara, E. Koriki, H. Ochiai, M.Maeda, J. G. Williams, I. Takeuchi, and Y. Tanaka. 2004. Analyses ofcDNAs from growth and slug stages of Dictyostelium discoideum. NucleicAcids Res. 32:1647–1653.

39. VanDriessche, N., C. Shaw, M. Katoh, T. Morio, R. Sucgang, M. Ibarra, H.Kuwayama, T. Saito, H. Urushihara, M. Maeda, I. Takeuchi, H. Ochiai, W.Eaton, J. Tollett, J. Halter, A. Kuspa, Y. Tanaka, and G. Shaulsky. 2002. Atranscriptional profile of multicellular development in Dictyostelium discoi-deum. Development 129:1543–1552.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

40. Veron, M., R. Mutzel, M.-L. Lacombe, M.-N. Simon, and V. Wallet. 1988.cAMP-dependent protein kinase from Dictyostelium discoideum. Dev.Genet. 9:247–258.

41. Wang, B., and A. Kuspa. 1997. Dictyostelium development in the absence ofcAMP. Science 277:251–254.

42. Wetterauer, B., G. Jacobsen, P. Morandini, and H. MacWilliams. 1993.Mutants of Dictyostelium discoideum with defects in the regulation of discoi-din I expression. Dev. Biol. 159:184–195.

43. Wetterauer, B. W., K. Salger, C. Carballo-Metzner, and H. K. MacWilliams.

1995. Cell-density-dependent repression of discoidin in Dictyostelium discoi-deum. Differentiation 59:289–297.

44. Winckler, T., C. Trautwein, C. Tschepke, C. Neuhauser, I. Zundorf, P. Beck,G. Vogel, and T. Dingermann. 2001. Gene function analysis by amber stopcodon suppression: CMBF is a nuclear protein that supports growth anddevelopment of Dictyostelium amoebae. J. Mol. Biol. 305:703–714.

45. Winckler, T., T. Dingermann, and G. Glockner. 2002. Dictyostelium mobileelements: strategies to amplify in a compact genome. Cell. Mol. Life Sci.59:2097–2111.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

CbfA, the C-Module DNA-Binding Factor, Plays an Essential Role in the Initiation of Dictyostelium...

Documents

Transcript of CbfA, the C-Module DNA-Binding Factor, Plays an Essential Role in the Initiation of Dictyostelium...