The Pho80-like Cyclin of Aspergillus nidulans Regulates Development Independently of Its Role in Phosphate Acquisition*[boxs]

In Saccharomyces cerevisiae, phosphate acquisition enzymes are regulated by a cyclin-dependent kinase (Pho85), a cyclin (Pho80), the cyclin-dependent kinase inhibitor Pho81, and the helix-loop-helix transcription factor Pho4 (the PHO system). Previous studies in Aspergillus nidulans indicate that a Pho85-like kinase, PHOA, does not regulate the classic PHO system but regulates development in a phosphate-dependent manner. A Pho80-like cyclin has now been isolated through its interaction with PHOA. Surprisingly, unlike PHOA, An-PHO80 does play a negative role in the PHO system. Similarly, an ortholog of Pho4 previously identified genetically as palcA also regulates the PHO system. However, An-PHO81, a putative cyclin-dependent kinase inhibitor, does not regulate the PHO system. Therefore, there are significant differences between the classic PHO system conserved between S. cerevisiae and Neurospora crassa compared with that which has evolved in A. nidulans. Most interestingly, under low phosphate conditions, the An-PHO80 cyclin also promotes sexual development while having a negative effect on asexual development. These effects are independent of the role An-PHO80 has in the classic PHO system. However, in high phosphate medium, An-PHO80 affects development because of deregulation of the PHO system as loss of palcAPho4 function negates the developmental defects caused by lack of An-pho80. Therefore, under low phosphate conditions the An-PHO80 cyclin regulates development independently of the PHO system, whereas in high phosphate it affects development through the PHO system. The data indicate that a single cyclin can control various aspects of growth and development in a multicellular organism.

In the budding yeast Saccharomyces cerevisiae, there are six CDKs: Cdc28, Pho85, Kin28, Srb10, Ctk1 (4), and Bur1 (12) of which Cdc28 and Pho85 bind to distinct classes of cyclins to perform different functions. Cdc28 controls cell cycle transitions by associating with different cyclin partners (9). In a similar manner Pho85 regulates diverse cellular processes depending on its cyclin partner (13). Deletion of Pho85 results in a pleiotropic phenotype including constitutive synthesis of repressible acid phosphatase (14), slow growth in nutrient-rich medium, and inability to utilize nonfermentable carbon sources (15), accumulation of glycogen (16,17), G 1 delay (18), and morphogenetic defects (15,19). The multiple functions of Pho85 may be ascribed to individual cyclins or cyclin groups.
In the filamentous fungus Aspergillus nidulans, three CDKs have been identified: NIMX cdc2 , PHOA, and PHOB. NIMX cdc2 , the homolog of S. cerevisiae Cdc28, functions in cell cycle progression when bound by its cyclin partner NIME cyclinB (32). PHOA was identified as a nonessential homolog of Pho85, which controls the developmental program of A. nidulans in an environment-dependent manner (6). The P i concentration, pH, and inoculation density influence development through PHOA. The absence of PHOA results in a switch from asexual to sexual development (at pH 8.0) or the absence of development (at pH 6.0) under the condition of low P i and high inoculum level (6).
A. nidulans contains a second Pho85 homolog, PHOB, which is 77% identical to PHOA (33). Deletion of phoB causes no phenotype, even in phosphorous-limited growth conditions. However, inactivation of phoA in a ⌬phoB background is lethal and leads to defects in nuclear division and polarity, indicating that the Pho85-like kinase pair PHOA and PHOB have an essential function and may be involve in cell cycle control and morphogenesis (33).
Identification of cyclin partners and inhibitors that regulate the PHOA CDK family member in A. nidulans should enhance our understanding of the cyclin-dependent class of kinase and their role in differentiation within multicellular eukaryotes. Here we have used the two-hybrid system to isolate a PHOAinteracting protein with homology to the S. cerevisiae Pho80 cyclin and named it An-PHO80 (A. nidulans PHO80). Although the An-pho80 gene is nonessential, it regulates P i acquisition genes through a mechanism not involving the PHOA or PHOB CDKs. The data further indicate that in addition to regulating the PHO system, An-PHO80 plays a role in the developmental program of A. nidulans. Thus in A. nidulans the PHO system cyclin plays two independent roles during growth and development.

EXPERIMENTAL PROCEDURES
A. nidulans Strains and Medium-The A. nidulans strains used in this study are listed in Table I, and all contain the veA1 marker. Sexual crosses were performed as described previously (34). Minimal medium with different concentrations of P i and MAG medium with supplements were prepared as described previously (6,34). Phenotypic characterization and conidia counting were also performed as described previously (6).
Yeast Two-hybrid Screening-A bait plasmid (pH1) was constructed by cloning the phoA cDNA into the EcoRI and BamHI site of the vector pGBKT7 (with c-Myc epitope tag, Clontech). A. nidulans cDNA library preparation and two-hybrid screening were performed by following the instruction of HybriZAP Two-hybrid cDNA Gigapack Cloning Kit (Stratagene) as described previously (35).
In Vitro Interaction Assay of Proteins-The An-pho80 cDNA isolated from a yeast two-hybrid screen was subcloned into EcoRI and XhoI sites of vector pGADT7 (with HA epitope tag, from Clontech) resulting in plasmid pGAD36. This plasmid and the bait plasmid pH1 were used to express PHOA and An-PHO80 protein in vitro using the TNT Quick Coupled Transcription/Translation system (Promega). Coimmunoprecipitation between the in vitro expressed An-PHO80 and PHOA proteins was performed as described in the manual of Matchmaker Co-IP Kit (Clontech). Immunoprecipitated protein using anti-c-Myc antibody (pull-down PHOA) was separated by SDS-PAGE (10% gels) and electroblotted onto Hybond ECL nitrocellulose membrane. Blots were blocked for 1 h in 5% nonfat dried milk in TBST (10 mM Tris-HCl at pH 7.5, 150 mM NaCl, 0.05% Tween 20) and incubated for 1 h with anti-HA antibody 3F10 (Roche Applied Science) (detecting An-PHO80) diluted 1:4,000 in blocking solution. Blots were washed in TBST and incubated for 1 h with a 1:10,000 dilution of rabbit anti-rat secondary antibody conjugated with horseradish peroxidase (Sigma) in TBST with 5% nonfat dry milk. The blots were washed and detected using an ECL Western blotting detection system (Amersham Biosciences) according to the manufacturer's recommendations. Prestained rainbow molecular mass markers (Amersham Biosciences) were used.
Rapid Amplification of cDNA Ends (RACE)-The full-length cDNA sequence of An-pho80 was obtained by the RACE method as described previously (6). For 5Ј-RACE, primers GSP36 (Table II) and AP1 were used to amplify the upstream sequence of An-pho80. RACE PCR products were gel purified and cloned in pCR4-TOPO vector using a TOPO TA Cloning Kit (Invitrogen) prior to sequencing.
Deletion of the An-pho80 and An-pho81 Genes-An-pho80 was deleted by replacing its coding region with the nutritional marker pyrG gene (37) using a two-step fusion PCR method (38). The first step involves PCR synthesis of three separate fragments: An-pho80 upstream and downstream sequences along with the pyrG gene cassette. An-pho80 upstream (1721 bp) and downstream (1983 bp) sequences were amplified with primers W36UP5, W36UP3 (for upstream fragment), and W36DN5, W36DN3 (for downstream fragment) using genomic DNA as template. The nutritional marker pyrG gene was amplified with primers PG36-5 and PG36-3, containing the 24-bp 5Јextensions complementary to the primer W36UP3 and W36DN5, respectively (Table II), using plasmid pCDA21 (containing the pyrG gene) as template. The three synthesized fragments were used in a fusion PCR to amplify the whole deletion cassette with primers W36UP5 and W36DN3. The fusion PCR product was transformed into A. nidulans GR5 strain selecting for growth without uridine and uracil. Successful replacement of An-pho80 by the pyrG cassette was confirmed by PCR amplification using primers W36-5 and W36-3 (Table II) anchored out of the An-pho80 gene deletion area. Putative null mutants were verified by Southern blot analysis. Southern blots were probed by an An-pho80specific DNA fragment or an An-pho80 flanking sequence using ECL random primer labeling and detection system (Amersham Biosciences). Deletion of An-pho81 was performed using the same method but a different nutritional marker gene, pyroA (39).
Enzyme Assays-Colony staining for acid phosphatase, alkaline phosphatase, and phosphodiesterase was performed as described previously (40). Activity assays of acid phosphatase and alkaline phosphatase followed the method of Caddick and Arst (40) using 1 mM pnitrophenyl phosphate (pNPP) as substrate. Phosphodiesterase activity was assayed as described previously (41) using 1 mM p-nitrophenyl phenylphosphonic acid (pNPP-P) as substrate.
Construction of the An-pho80-overexpressing Plasmid-The fulllength coding region of An-pho80 was amplified from A. nidulans cDNA by PCR using primers DKp36 and DBg36. The PCR product was digested with KpnI and BglII and then ligated with plasmid pAL6 digested with KpnI and BamHI. The constructed plasmid was named pAL36 with An-pho80 C-terminally tagged with two tandem repeats of the HA epitope under control of the inducible alcA promoter.

RESULTS
Isolation of An-pho80 Using the Yeast Two-hybrid System-Upon screening of an A. nidulans cDNA library using the yeast two-hybrid system and the PHOA CDK (6) as bait, three PHOA-interacting genes were isolated. One encodes a protein similar to NOT2 (CDC36). The other two genes encode a Pcl7like and a Pho80-like cyclin (Table III). The PHO80-like cyclin named An-pho80 is the focus of this work. Based on comparisons between the genetic and physical maps of A. nidulans (www.broad.mit.edu/annotation/fungi/aspergillus/maps.html) An-pho80 lies in the telomere-proximal region of the left arm of linkage group V. RACE analysis showed An-pho80 to lack introns, although the gene encoding protein AN5156.2 (An-PHO80) is predicted to have a single intron by automated annotation. There is therefore a discrepancy at the N-terminal 49 amino acids between AN5156.2 and our predicted protein sequence derived from cDNA sequence (GenBank accession number AY590766). An-pho80 encodes 390 amino acids with an estimated molecular mass of 42.4 kDa, which shares homology with the Pho80 subfamily of Pho85interacting cyclins of S. cerevisiae (13).
To test the physical interaction between PHOA and An-PHO80 proteins, the An-pho80 cDNA was subcloned into a vector to incorporate an N-terminal HA epitope tag. phoA was also cloned into a vector to incorporate an epitope tag, which in this case was c-Myc. The tagged genes were expressed in vitro using a coupled transcription/translation system. Immunoprecipitates using anti-c-Myc antibody to pull down PHOA also coprecipitated An-PHO80 protein (Fig. 1C), indicating direct association with PHOA in vitro. Control immunoprecipitation with anti-c-Myc antibody to pull down in vitro expressed control c-Myc peptide instead of c-Myc-tagged PHOA failed to precipitate An-PHO80 protein (Fig. 1C).
To verify the interaction between the two proteins in vivo, a strain mildly expressing C-terminal HA-tagged An-PHO80 was constructed. Western blotting specifically detected AN-PHO80-HA as a doublet both before (Fig. 1D, Loading control) and after immunoprecipitation (HA IP). Both PHOA and PHOB were found to coprecipitate with An-PHO80-HA using antibodies raised against PHOA, which also detects PHOB (6, 33) (Fig. 1D). The interaction between An-PHO80 and PHOB was confirmed using the two-hybrid system (Table III). To determine whether the interaction between An-PHO80 and PHOA and PHOB is specific, the An-PHO80-HA immunoprecipitates were also probed using antibodies that detect another CDK from A. nidulans, ␣-NIMX cdc2 . No interaction was detected between An-PHO80 and NIMX cdc2 (Fig. 1D). The data indicate that the An-PHO80 cyclin specifically interacts with its partner CDKs PHOA and PHOB but not with the related NIMX cdc2 CDK.
Deletion of An-pho80 Causes Constitutive Induction of Phosphate-repressible Enzymes in P i -rich Medium-In budding yeast, deletion of either the PHO85 CDK or its interacting cyclin partner PHO80 leads to constitutive expression of P i acquisition genes including PHO5 (an acid phosphatase gene) in P i -sufficient medium (43)(44)(45)(46)(47). In contrast, phoA, the counterpart of PHO85 in A. nidulans, has no apparent role in regulation of acid phosphatase expression (6). To determine whether the An-pho80 cyclin has a role in regulation of P i acquisition, An-pho80 was deleted by replacing its coding sequence with the pyrG nutritional marker (37). Four identical An-pho80 null mutant strains were isolated and verified by PCR and Southern blot analysis (data not shown). Although deletion of An-pho80 did not cause lethality it generated several interesting phenotypes.
It has been shown previously that normal P i levels (11 mM) are repressing, and low P i (0.1 mM) is inducing for the expression of acid phosphatase, alkaline phosphatase, and phosphodiesterase activities in A. nidulans (40,41). We therefore measured these activities in a ⌬An-pho80 strain along with wild type and ⌬phoA control strains in normal and low P i conditions.
AATCTAGTCCTCCACTCTCG a Sequences complementary to W36UP3 and W36DN5, and W81UP3 and W81DN5 are underlined. High phosphatase and phosphodiesterase activities were measured for each strain when grown in low P i conditions (Table IV). Less than 2% of these activities were measured in both the wild type and ⌬phoA strains when grown in high P i medium. In contrast, repression of phosphatase and phosphodiesterase activities did not occur in the ⌬An-pho80 strain under high P i conditions (Table IV). Similarly, secreted phosphodiesterase activity was not repressed in the ⌬An-pho80 strain under high P i conditions but was repressed in the wild type and ⌬phoA strains (data not shown). Thus, deletion of An-pho80 leads to constitutive induction of P i acquisition enzymes in a manner identical to that seen after deletion of S. cerevisiae PHO80. These two related cyclins therefore play similar negative roles in the classic PHO system. Interestingly, previous studies have indicated that overexpression of Pho80 is able to suppress acid phosphatase activity in S. cerevisiae even when PHO85 has been deleted (47). We a Phosphatase activity is expressed as nmol of p-nitrophenol liberated from pNPP/mg/min. b Phosphodiesterase activity is expressed as nmol of p-nitrophenol liberated from pNPP-P/mg/min. 2) and S. cerevisiae Pho85 cyclins. The phylogenetic tree was generated using ClustalW and full-length protein sequences (13). The scale bar indicates the relative distance on the tree. Predicted A. nidulans cyclins are marked by asterisks. B, An-PHO80 protein has similarity to Pho80-like cyclins from different fungi. The cyclin box regions of An-PHO80 (390 amino acids) and the Pho80-like cyclins from N. crassa (PREG, 483 amino acids), A. fumigatus (Af-PHO80, 396 amino acids), M. grisea (Mg-PHO80, 442 amino acids), and S. cerevisiae (Pho80, 293 amino acids) were aligned using ClustalW. Asterisks indicate identical residues between An-PHO80 and the other proteins, colons indicate conserved residues, and periods mark semiconserved residues. C, in vitro transcribed and translated HA-An-PHO80 and c-Myc-PHOA were incubated together prior to immunoprecipitation with an anti c-Myc antibody. In vitro expressed c-Myc peptide was used as a control. After Western blotting proteins were visualized using an anti-HA antibody. The input level of HA-An-PHO80 is shown in the left lane (L). D, protein extracts from a strain expressing C-terminal HA tagged An-PHO80 (An-PHO80-HA) and a strain with empty vector (An-PHO80) were immunoprecipitated with an anti-HA antibody, and Western blots were performed with different antibodies as indicated. The loading controls represent 1/20 of the protein used for the immunoprecipitations. therefore placed expression of An-pho80 under the control of the regulatable alcA promoter (48) and determined whether ectopic An-pho80 expression could suppress secreted acid phosphatase activity ( Fig. 2A) or phosphodiesterase activity (Fig.  2B) under low P i growth conditions. As expected, the control strains secreted acid phosphatase under low P i conditions as detected using a plate assay ( Fig. 2A). However, overexpression of An-PHO80 severely compromised this activity to a level lower than that observed in a palcA1 strain deficient in phosphatase induction (Fig. 2A). We also tested for phosphodiesterase activity under low P i with and without ectopic expression of An-PHO80. As expected, a strain containing a control vector induced phosphodiesterase activity under low P i conditions (Fig. 2B). In contrast, overexpression of An-PHO80 greatly reduced phosphodiesterase activity to a level similar to that observed in a palcA1 mutant (Fig. 2B). These data indicate that ectopic expression of An-pho80 prevents synthesis of enzymes normally made under limiting P i conditions, further demonstrating a negative role for An-pho80 in the regulation of P irepressible enzymes.
The fact that An-pho80 plays a role in phosphatase synthesis is surprising because its CDK partner phoA is not involved in the regulation of P i acquisition enzymes (Table IV and Ref. 6).
One potential explanation for why An-pho80 is important for regulation of P i acquisition enzymes, but its CDK partner is not, could be that there are two highly related PHO85-like CDKs in A. nidulans, PHOA and PHOB (33). Recent work has established that phoA and phoB have an essential overlapping function(s) because a double phoA/phoB deletion causes lethality. We therefore tested whether phoA and phoB have a redundant role in P i regulation of P i acquisition enzymes. Acid phosphatase assays were completed using a strain with deleted phoB and a temperature-sensitive allele of phoA (33). At near restrictive temperature (40°C) for growth of the phoA ts ⌬phoB strain, no acid phosphatase activities could be detected in high P i medium using plate assays (Fig. 3A) or protein extracts assays (Fig. 3B), similar to the wild type, ⌬phoA, and ⌬phoB controls. Conversely, under the same high P i concentration and temperature the ⌬An-pho80 strain had elevated acid phosphatase activity (Fig. 3), consistent with its proposed negative role in P i acquisition. These data strongly suggest that phoA and phoB do not play redundant roles in P i acquisition enzyme synthesis.
In S. cerevisiae the Pho81 CDK inhibitor regulates P i acquisition by inhibiting Pho85. Mutation of Pho81 prevents expression of phosphatase genes because of constitutive activity of Pho85 under low P i conditions. As our data suggest that the kinase activity of PHOA/PHOB is not required in the PHO signal transduction pathway, we tested whether the Pho81 CDK inhibitor ortholog in A. nidulans was required for regulation of this pathway. We generated a null allele of the Pho81 ortholog of A. nidulans (AN4310.2, which shares 28% identity with yeast Pho81 and 52% with N. crassa NUC-2; see supplemental Fig. S1) and tested whether the absence of this putative CDK inhibitor would prevent expression of P i acquisition genes under low P i conditions. Using both plate assays and protein extracts assays we found no evidence that An-PHO81 plays a role in regulating P i acquisition enzymes in A. nidulans (Fig. 4, A and B). This finding further supports the regulation of P i acquisition enzymes being independent of PHOA/PHOB activity.
An-pho80 Deletion Affects Development and Growth-In addition to playing a role in regulating expression of phosphatase genes we found that deletion of An-pho80 caused some specific developmental defects. We used previously described pour plate growth conditions (6) to investigate the developmental response of the wild type, ⌬phoA, and ⌬An-pho80 strains to different growth conditions. Under limiting P i concentrations and low inoculum levels, deletion of An-pho80 caused increased asexual development at either pH 6.0 or pH 8.0 compared with the control wild type and ⌬phoA strains, leading to increased conidia production (Fig. 5A). An-pho80 deletion also decreased sexual development relative to the two controls (Fig. 5B). Sexual development of nascent cleistothecia surrounded by Hulle cells is apparent for the wild type and ⌬phoA strains but is strikingly absent in the ⌬An-pho80 strain (Fig. 5B).
Because loss of An-PHO80 function can promote asexual development but repress sexual development (Fig. 5, A and B), it suggests that An-pho80 may play a positive role in sexual development and a negative role in asexual development. Consistent with this, ectopic expression of An-pho80 in a wild type background promoted sexual development and suppressed asexual development (Fig. 5C, ϩAn-pho80). Under the same conditions the wild type (Fig. 5C) and ⌬phoA strains (not shown) underwent vigorous asexual development, generating many conidia.
We have shown previously that deletion of phoA stops asexual or sexual development in plate cultures grown with limiting P i and high inoculum levels at pH 6.0 (Fig. 8 of Ref. 6). Under these conditions, a ⌬An-pho80 strain did not have such

FIG. 2. Overexpression of An-pho80 stops phosphatase and phosphodiesterase induction under low P i condition.
A, colony staining for acid phosphatase. Wild type strain (GR5), palcA1 mutant strain (HB31), duplicate An-pho80 overexpressing strains (ϩAn-pho80), and a strain transformed with empty vector were inoculated on minimal medium containing 40 mM threonine and 0.1 mM P i at pH 6.0 and grown at 37°C for 22 h. Colony staining was performed for acid phosphatase activity indicated by red staining around the colonies. B, phosphodiesterase activity assay. A strain with empty vector (Vector), an An-pho80 over-expressing strain (ϩAn-pho80), and a palcA1 mutant strain were grown in liquid minimal medium containing 40 mM threonine and 0.1 mM P i at pH 6.0 at 37°C for 20 h. Protein extracts were analyzed for phosphodiesterase activity, which is expressed in nmol of p-nitrophenol liberated from pNPP-P/mg of soluble protein extract/min. defects but underwent normal asexual development in a manner identical to the control wild type strains (data not shown). Thus, under several growth conditions, we failed to observe any defects caused by deletion of An-pho80 which are similar to the defects caused by deletion of phoA.
Two additional obvious phenotypes were seen in ⌬An-pho80 strains but in this case under conditions of normal P i rather than low P i . Notably, An-pho80 nulls grow slower than wild FIG. 3. phoA and phoB do not play redundant roles in regulation of P i acquisition enzymes. A, wild type (WT), ⌬phoA, ⌬phoB, phoA ts ϩ ⌬phoB, and ⌬An-pho80 strains were grown on minimal medium plates (pH 6.0) with high (11 mM) or low (0.1 mM) P i at 40°C for 22 h, and the duplicate colonies were stained for acid phosphatase activity indicated by red staining. B, the indicated strains were grown in minimal liquid medium (pH 6.0) containing high (11 mM) P i or low (0.1 mM) P i at 40°C for 18 h. Protein extracts were analyzed for acid phosphatase activity as described under "Experimental Procedures." The acid phosphatase activity was expressed as nmol of p-nitrophenol released from pNPP/mg of soluble protein/min.

FIG. 4. Deletion of An-pho81 does not lead to inhibition of phosphodiesterase activity.
A, the wild type (WT), a ⌬An-pho80, a palcA1 mutant, and three ⌬An-pho81 strains were grown on minimal medium plates (pH 6.0) with high (11 mM) or low (0.1 mM) P i at 37°C for 22 h, and the colonies were stained for acid phosphatase indicated by red staining. B, the indicated strains were grown in minimal liquid medium (pH 6.0) containing high (11 mM) or low (0.1 mM) P i at 37°C for 18 h and analyzed for phosphodiesterase activity. The phosphodiesterase activity is expressed as nmol of p-nitrophenol released from pNPP-P/mg of soluble protein/min. type and ⌬phoA strains in non-limiting P i medium (Table V). ⌬An-pho80 strains consistently displayed one-third reduction of the growth rate in liquid culture compared with both wild type and ⌬phoA strains under normal P i conditions. This effect was not observed in P i -limited medium in which ⌬An-pho80 strains grew at the same rate as wild type and ⌬phoA strains (Table V). We also observed that under normal P i conditions conidia production was greatly reduced in the ⌬An-pho80 mutant compared with a wild type strain (Fig. 6A). At its most extreme, when high inoculum levels were used, this phenotype revealed a complete lack of conidiation, although vegetative growth was apparent within the agar substrate (Fig. 6, A and  B, and data not shown). Very few aerial hyphae were generated, and this phenotype is therefore not the same phenotype seen for some A. nidulans developmental mutants known as fluffy mutants.
palcA Encodes a Pho4-like Transcription Factor Necessary for the Developmental Defects Caused by ⌬An-pho80 in P isufficient Medium-Given that ⌬An-pho80 strains increase the activity of P i acquisition enzymes (Fig. 3), it is possible that elevated levels of intracellular P i may lead to the developmental defects described above. If this is the case, mutation of a regulator required for synthesis of P i acquisition enzymes may suppress the defects caused by An-pho80 deletion by preventing accumulation of elevated intracellular P i . One such regulator is A. nidulans palcA.  5. An-pho80 plays a positive role in sexual development but a negative role in asexual development. A, An-pho80 null mutants display elevated conidia (asexual spores) formation under conditions of low P i and low inoculum. Conidia of strains were plated in top agar on minimal medium as indicated. The numbers of conidia formed after 4 days are expressed relative to the highest value observed (100% conidiation ϭ 2.99 ϫ10 11 conidia/m 2 ). B, the plates were photographed under a dissecting microscope. Asexual development results in formation of conidiophores carrying chains of conidia that cause the green color, whereas sexual development is indicated by the yellow color of hulle cells occurring in clusters that surround nascent sexual fruiting bodies called cleistothecia. C, overexpression of An-pho80 in a wild type background promotes sexual development and suppresses asexual sporulation. A wild type strain, a palcA1 mutant strain (HB31) and a strain containing the An-pho80-overexpressing plasmid (all with wA3 marker, showing white conidiophores) were grown on minimal medium plates with 0.1 mM P i and 40 mM threonine to induce overexpression of An-PHO80 protein at 37°C for 7 days. Plates were photographed under a dissecting microscope. Sexual development is shown after An-PHO80 induction by the yellow color of hulle cells occurring in clusters that surround nascent cleistothecia. The scale bar applies to B and C.
In both the S. cerevisiae and N. crassa PHO systems, a regulatory transcription factor (Pho4/NUC-1) is responsible for regulated expression of genes for P i acquisition. Genetically the palcA gene of A. nidulans has characteristics suggesting that it could be the functional homolog of Pho4/NUC-1. It acts in a positive manner to control the synthesis of several P i -repressible enzymes and probably also a P i permease (41). Although palcA has not been cloned, it has been mapped to the centromere-proximal region of the left arm of linkage group II (41) and linked (8% recombination) to phoA (6). We searched the A. nidulans genome using BLAST analysis for a potential pho4 or nuc-1 ϩ -like gene, which both encode HLH-type transcription factors. A putative ortholog was uncovered and its cDNA sequence defined (GenBank accession number AY590767). The predicted amino acid sequence (AN8271.2) shares 19% identity with PHO4 and 27% with NUC-1 and, like these proteins, contains a HLH domain (supplemental Fig. S2). This gene is located 27 kb from phoA and, given that palcA and phoA are genetically linked, indicates that this PHO4-like gene could be palcA. To confirm this we sequenced the palcA1 temperaturesensitive allele (49,50) and compared it with the wild type sequence. A single point mutation changing asparagine codon AAC (586) to a UAC tyrosine codon in the HLH domain of palcA1 was defined. We also sequenced the palcA40 allele and identified a mutation within the coding region. At codon 472 (S) an insertion of 22 bp was observed encoding 10 additional amino acids followed by a stop codon (TAG). The net result is a truncation of palcA Pho4 before its helix loop helix domain likely generating a null allele. Identification of these mutations strongly suggests that palcA Pho4 encodes the Pho4/Nuc-1-like HLH transcription factor of A. nidulans. Consistent with this, as reported previously (50), the palcA1 mutation failed to induce either acid phosphatase or phosphodiesterase activities in low P i medium (Fig. 2). FIG. 7. The function of palcA Pho4 is needed for constitutive induction of phosphatase in a An-pho80 null mutant strain. A wild type strain, ⌬An-pho80 strain, palcA1 mutant strain, and the ⌬An-pho80 ϩ palcA1 double mutants were grown on minimal medium plates containing 0.1 mM P i at pH 6.0 at 37°C for 22 h and then stained for acid phosphatase activity.
FIG. 6. Deletion of An-pho80 leads to reduced conidia production under normal P i conditions. A, conidia of the wild type strain ⌯〉38 and ⌬An-pho80 strain DWD2 were plated in top agar on minimal medium (11 mM P i , pH 8.0) at the densities (conidia/plate) indicated. The number of conidia formed after 4 days is expressed relative to the highest value observed (100% conidiation ϭ 2.4 ϫ 10 11 conidia/m 2 ). B, conidia of the wild type strain ⌯〉38, ⌬An-pho80 strain DWD2 and ⌬An-pho80 ϩ palcA1 double mutant strain were plated in top agar on minimal medium (11 mM P i , pH 8.0) at the densities (conidia/plate) indicated. The plates were photographed under a dissecting microscope after 4 days of growth at 37°C. Normal asexual development results in formation of conidiophores carrying chains of asexual conidia that cause the green color (left and right panels), whereas inhibition of development caused reduced (middle panel of top row) or very few (middle panel of bottom row) small conidiophores to develop on the surface of the plates. The scale bar is 1 mm.
The main reason to identify a mutation in a gene required for synthesis of P i acquisition enzymes was to test whether overproduction of P i acquisition enzymes in the ⌬An-pho80 mutant strain caused high levels of intracellular P i leading to developmental defects. Supporting this concept, loss of palcA Pho4 function suppressed the developmental defects of the An-pho80 null mutants grown on normal P i concentrations (Fig. 6B, compare  ⌬An-pho80 with palcA1ϩ⌬An-pho80). This was a very dramatic effect with mutation of palcA in an An-pho80 null strain increasing the conidia production more than 500-fold. This indicates that the developmental defects of ⌬An-pho80 strains on normal P i concentration may result from high levels of intracellular P i caused by inappropriate increase in P i acquisition enzymes under control of the palcA Pho4 transcription factor.
To confirm that palcA Pho4 is required for induction of An-PHO80-regulated P i acquisition enzymes, the double palcA1 ϩ ⌬An-pho80 mutant was assayed for acid phosphatase activity in low P i growth conditions (Fig. 7). The double palcA1 ϩ ⌬An-pho80 failed to induce phosphatase activity, whereas under identical conditions the ⌬An-pho80 strain did induce acid phosphatase activity as expected. This demonstrates that pal-cA Pho4 is needed for constitutive induction of phosphatases in an An-pho80 null strain, and places palcA Pho4 downstream from An-pho80 in this pathway. DISCUSSION In S. cerevisiae the Pho85/Pho80 CDK/cyclin pair is essential for regulation of the PHO system by phosphorylating and controlling the nuclear localization of the HLH transcription factor Pho4 (51). In turn, Pho4 induces expression of P i acquisition genes under limiting P i conditions. In this work we have defined An-pho80 as the Pho80 cyclin ortholog and palcA Pho4 as the Pho4 ortholog of A. nidulans. Both An-pho80 and palcA Pho4 play the expected roles in the PHO system of A. nidulans. Thus An-pho80 deletion deregulates P i acquisition enzymes, whereas its overexpression suppresses P i acquisition enzymes. An-PHO80 therefore plays a negative role in the A. nidulans PHO system like Pho80 in S. cerevisiae. Similarly, the Pho4-like HLH PALCA Pho4 transcription factor is required for expression of P i acquisition genes, even when An-PHO80 is deleted. There are therefore clear parallels between the roles of An-PHO80 and PALCA Pho4 with regard to their orthologs in both S. cerevisiae and the filamentous fungus N. crassa. However, there are clearly distinct differences in the PHO system of S. cerevisiae and N. crassa compared with the PHO system in A. nidulans. Additionally, An-pho80 plays roles during development which are independent of its role in the P i acquisition system.
The Classic PHO System Is highly Conserved between S. cerevisiae and N. crassa but Is Not Conserved in A. nidulans-Both An-pho80 and palcA Pho4 play their expected roles in P i acquisition control. However, in marked contrast, the Pho85related cyclin-dependent kinase PHOA does not play a role in the regulation of P i acquisition enzymes (Ref. 6 and current work). This conclusion is perhaps compromised by the fact that PHOA plays a redundant essential function in A. nidulans with a second Pho85-like kinase termed PHOB (33). Although we have shown that a temperature-sensitive allele of phoA in a ⌬phoB background fails to express phosphatase enzymes at near restrictive temperatures with sufficient P i , we cannot exclude the possibility that the level of PHOA activity required for PHO regulation is lower than that required for growth. Nevertheless, further supporting the contention that PHOA and PHOB do not play a role in P i acquisition enzyme control, deletion of the putative Pho81 CDK inhibitor of A. nidulans also failed to affect P i acquisition enzyme control. We are therefore left with the conclusion that although the genes involved in the regulation of P i acquisition enzymes are conserved among A. nidulans, N. crassa, and S. cerevisiae, only two of the four PHO regulatory genes have a conserved function among all three organisms; the Pho80 cyclin orthologs and the Pho4 HLH transcription factor orthologs.
This situation clearly poses some interesting issues with regard to the regulation of the PHO system in A. nidulans. How does the An-PHO80 cyclin regulate the PALCA Pho4 transcription factor without PHOA or PHOB? In S. cerevisiae ectopic expression of the Pho80 cyclin is able to inhibit Pho4-regulated genes in a strain in which the Pho85 CDK is deleted (47). Additionally, in human cells cyclin D1 can repress STAT3 transcription factors and activate estrogen receptor without the apparent participation of CDKs (52,53). It is therefore in principle possible that the An-PHO80 cyclin could directly influence the PALCA Pho4 transcription factor independently of a CDK partner. Alternatively, and perhaps more likely, another kinase partner may play a role in the PHO system with An-PHO80. Apart from having two Pho85-like CDKs, the genome of A. nidulans encodes a repertoire of CDK catalytic subunits similar to S. cerevisiae (Fig. S3), any of which could in theory partner with An-PHO80 to generate a regulatory CDK-cyclin complex involved in P i acquisition enzyme regulation. Future experiments to identify An-PHO80 interacting proteins may therefore identify other potential partners for An-PHO80 which could be involved in P i acquisition.
Another interesting question is what role An-PHO80 and its CDK partners PHOA and PHOB play as their interaction appears to be specific but not required for regulation of the PHO system of A. nidulans. Because both PHOA and An-PHO80 play roles during development, they may have common functions during development by interacting with each other. However, although PHOA represses sexual development under P ilimited conditions, An-PHO80 promotes sexual development under the same conditions. This may suggest that An-PHO80 could inhibit PHOA rather than activate its function during development. Clearly, further experimentation will be required to clarify the roles of An-PHO80 when interacting with either PHOA or PHOB.
Caution Should Be Exercised When Considering Function Based on Sequence Similarities-The PHO system would appear to be highly conserved between S. cerevisiae and N. crassa. They both have a CDK inhibitor (Pho81 in S. cerevisiae, NUC-2 in N. crassa) which binds to a CDKcyclin complex (Pho85-Pho80 in the former, presumably PGOV-PREG in the later) under low P i growth condition to inhibit its kinase activity. Then, the underphosphorylated transcription factor (Pho4 in S. cerevisiae, NUC-1 in N. crassa) activates phosphatase and P i permease genes to acquire P i from the environment (42,54,55). Given this degree of conservation and that A. nidulans also contains these four regulatory genes, it would seem extremely likely that the PHO regulatory system would be conserved between these fungi. Consistent with this, both Pho80 and Pho4 from A. nidulans play their expected negative and positive roles in the PHO system. However, at least for An-PHO80, additional functions can be ascribed to this cyclin during development, and these functions are apparently independent of its conserved role in P i acquisition. Even more remarkable, deletion analysis of the counterparts of Pho81 and Pho85 in A. nidulans clearly indicates that they play no role in controlling the P i acquisition genes regulated by the Pho4 HLH transcription factor. Our findings therefore suggest that during evolution the same regulatory module can be put to different uses. The finding that conserved genes with conserved functions between S. cerevisiae and N. crassa can play significantly different roles in A. nidulans suggest that extreme caution should be exercised when considering gene function based solely on sequence similarities, even between different filamentous fungal species.
The An-PHO80 Cyclin Influences Development under Both High and Low P i Levels-Under low P i conditions, deletion analysis and ectopic expression experiments indicate that An-PHO80 promotes sexual development while having a negative effect on asexual development. We conclude this because ectopic expression of An-PHO80 promotes sexual development and also reduces the level of asexual development. Additionally, deletion of An-PHO80 completely suppresses sexual development but markedly promotes asexual development.
During normal development there is a switch from asexual to sexual development as A. nidulans colonies grow and mature (for review, see Ref. 56). It is therefore possible that An-PHO80 plays an important role in this developmental switch, particularly under limiting P i concentrations. In the absence of An-PHO80 function this switch is unable to occur resulting in continued conidia production when normally a switch to sexual spore development would occur. Conversely, after artificially increasing the level of An-PHO80, the switch to sexual development is triggered to occur prematurely at the expense of conidia production. It will be interesting to follow the developmental pattern of An-PHO80 expression and to see whether its interacting partners change during asexual and sexual development. In addition, the effects of An-PHO80 on the production of the hormone-like substance psi (57) will be important to investigate. psi plays a role in the balance between asexual and sexual development in A. nidulans (58 -61), which is exactly the transition that is affected by the An-PHO80 cyclin.
Given that An-PHO80 appears to have at least two roles in A. nidulans, one involving control of P i acquisition enzymes and the other in promoting the switch from asexual to sexual development, it is important to determine whether these two functions are related. For example, perhaps the developmental defects seen after manipulation of An-PHO80 are caused by defects in its role in P i acquisition enzyme control. This would appear to be unlikely because under low P i conditions, the regulation of P i acquisition enzymes is not compromised by deletion of An-PHO80, but lack of An-PHO80 does have marked effects on development under these conditions. Similarly, overexpression of An-PHO80 in low P i conditions represses P i acquisition enzymes and also promotes sexual development. This may suggest that repression of P i acquisition enzymes could subsequently cause developmental defects. However, deletion of palcA Pho4 , which prevents induction of P i acquisition genes in low P i , does not cause developmental defects (Fig. 5C). We therefore conclude that An-PHO80 is likely to have developmental functions in addition to its role in regulating P i acquisition enzymes.
Under conditions of sufficient P i , lack of An-PHO80 function completely prevents production of conidia at high inoculation density. This effect was diminished somewhat at lower inoculation levels where some conidia were formed. The lack of asexual development caused by deletion of An-PHO80 was reversed dramatically by inactivation of PALCA Pho4 . As palcA-Pho4 encodes the Pho4-like HLH transcription factor required for expression of P i acquisition genes, this indicates that the developmental defects caused by deletion of An-PHO80 may be caused by increased cellular P i concentrations. However, if this is the mechanism by which deletion of An-PHO80 causes developmental defects, it is hard to explain two observations. Why would P i "toxicity" have a preferential effect on asexual development? And second, why would the effect be increased by high inoculation levels? With high inoculation levels it could be argued that less P i would be available per cell and so decrease P i toxicity. We have therefore not completely excluded the possibility that the effect of An-pho80 deletion, and the reversal of the resulting developmental defects caused by inactivation of PALCA Pho4 , may reflect a specific role for these genes in asexual development. It should be pointed out that asexual development in A. nidulans is likely under the influence of extracellular signaling molecules (56), and it will be of interest to determine whether either An-pho80 or palcA Pho4 has any genetic interaction with developmental regulators.
Finally, a second Pcl-like cyclin gene has been identified in A. nidulans which also plays a role during development termed pclA (62). This cyclin has been shown to interact with the central cell cycle nimX cdc2 gene of A. nidulans (63). Unlike An-pho80, induction of pclA does not affect development. However, deletion of pclA affects the rate of conidia formation, but plays no role in the switch from asexual to sexual development as described here for An-pho80. It would therefore appear that these two related cyclins play quite distinct roles in the developmental program of A. nidulans.