A basic helix–loop–helix protein with similarity to the fungal morphological regulators, Phd1p, Efg1p and StuA, controls conidiation but not dimorphic growth in Penicillium marneffei

Authors


Summary

Members of the APSES protein group are basic helix–loop–helix (bHLH) proteins that regulate processes such as mating, asexual sporulation and dimorphic growth in fungi. Penicillium marneffei is a human pathogen and is the only member of its genus to display a dimorphic growth transition. At 25°C, P. marneffei grows with a filamentous morphology and produces asexual spores from multicellular con-idiophores. At 37°C, the filamentous morphology is replaced by yeast cells that reproduce by fission. We have cloned and characterized an APSES protein-encoding gene from P. marneffei that has a high degree of similarity to Aspergillus nidulans stuA. Deletion of stuA in P. marneffei showed that it is required for metula and phialide formation during conidiation but is not required for dimorphic growth. This suggests that APSES proteins may control processes that require budding (formation of the metulae and phialides, pseudohyphal growth in Saccharomyces cerevisiae and dimorphic growth in Candida albicans) but not those that require fission (dimorphic growth in P. marneffei). The A. nidulansΔstuA mutant has defects in both conidiation and mating. The P. marneffei stuA gene was capable of complementing the conidiation defect but could only inefficiently complement the sexual defects of the A. nidulans mutant. This suggests that the P. marneffei gene, which comes from an asexual species, has diverged significantly from the A. nidulans gene with respect to sexual but not asexual development.

Introduction

Two of the most extensively characterized developmental processes in fungi are the formation of asexual spores (conidia) by filamentous fungi and dimorphic growth. The process of conidiation is best understood in the ascomycete Aspergillus nidulans, in which asexual spores are formed on large multicellular structures termed conidiophores through an ordered differentiation of several specialized cell types from the vegetative mycelium. Conidiation begins with the growth of an aerial stalk from a specialized foot cell in the basal mycelium. The stalk grows to a genetically predetermined height and then swells at its tip to form a globose, multinucleate vesicle. A layer of uninucleate cells termed metulae then bud simultaneously from the vesicle followed by a second tier of cells (phialides) that bud from the metulae. Conidia are then formed through interstitial budding of the phialides, generating chains of asexual spores while leaving the phialides ready for further spore production (for a review, see Adams et al., 1998). At least four gene products interact to form the core regulatory network of this process (BrlA, AbaA, StuA and MedA) (Boylan et al., 1987; Mirabito et al., 1989; Miller et al., 1991). BrlA and AbaA are sequence-specific DNA-binding transcription factors that form a conidiation-specific regulatory cascade required for the formation of all the various cell types of the conidiophore (Clutterbuck, 1969; Boylan et al., 1987; Adams et al., 1988; Chang and Timberlake, 1993; Andrianopoulos and Timberlake, 1994). StuA and MedA are developmental modifiers of the BrlA/AbaA pathway that affect conidiophore morphogenesis through the spatial and temporal regulation of brlA and abaA expression (Clutterbuck, 1969; Miller et al., 1991; 1992).

The A. nidulans stuA gene encodes a basic helix–loop–helix (bHLH) transcriptional regulator (Miller et al., 1992; Dutton et al., 1997). stuA mutant conidiophores have shortened stalks, reduced vesicles and lack sterigmata (metulae and phialides) (Clutterbuck, 1969). Unlike brlA and abaA mutant strains, however, stuA mutants are not aconidial. Spores are produced at low frequency in these strains through direct budding from the vesicle. Also, in contrast to both brlA and abaA, the effects of stuA mutations are not restricted to the asexual pathway, as stuA mutants are also self-sterile and fail to produce the major cell types and structures associated with sexual reproduction in A. nidulans (Hülle cells, cleistothecia and ascospores; Clutterbuck, 1969). The complex phenotypes shown by the stuA mutants are matched by the complexity of the stuA gene itself. The stuA gene produces two, differentially regulated transcripts (stuAα and stuAβ) that are regulated by an extremely large promoter (relative to other fungal genes), with over 3 kb of promoter sequences required for the correct regulation of stuA (Miller et al., 1992; Wu and Miller, 1997).

There have been four proteins identified in other species of fungi that have high similarity to the bHLH region of StuA. These proteins have been termed APSES proteins (Asm-1, Phd1p, StuA, Efg1p and Sok2p) and have been shown to regulate development in each of their respective species (Miller et al., 1992; Gimeno and Fink, 1994; Ward et al., 1995; Aramayo et al., 1996; Stoldt et al., 1997). In Neurospora crassa, Asm-1 is involved in a number of processes including spore germination, mycelial growth and both asexual and sexual sporulation (Aramayo et al., 1996). This protein therefore appears to have a function analogous to StuA in A. nidulans and indicates conservation of function between the two bHLH proteins across the two species.

In Saccharomyces cerevisiae and Candida albicans, the APSES proteins have been shown to regulate dimorphic growth. In S. cerevisiae, the two APSES members, Phd1p and Sok2p, regulate pseudohyphal growth, a process in which diploid cells of S. cerevisiae change from a unicellular mode of growth with ellipsoidal cells to a filamentous-like mode with elongated, conjoined cells that penetrate the surface of the agar (Gimeno et al., 1992). Expression of Phd1p enhances pseudohyphal growth, whereas expression of Sok2p has the opposite effect, repressing pseudohyphal development (Gimeno and Fink, 1994; Ward et al., 1995). C. albicans, a dimorphic human pathogen, is capable of growth as budding yeast cells, pseudohyphal cells and true filamentous hyphae depending on environmental stimuli (Merson-Davies and Odds, 1989). The C. albicans APSES protein, Efg1p, is required for dimorphic growth, virulence and chlamydospore formation. efg1/efg1 mutants fail to produce true hyphae under the majority of inducing conditions (including the presence of serum) and are partially attenuated for virulence in mouse models (Lo et al., 1997; Stoldt et al., 1997; Braun and Johnson, 2000). In addition, efg1/efg1 strains are unable to form chlamydospores. These ellipsoidal cells are formed under certain starvation and embedded conditions by budding from hyphal filaments and represent another example of a dimorphic transition in C. albicans (Sonneborn et al., 1999).

Although it appears that the APSES protein of A. nidulans functions in a developmental pathway that is distinct from the APSES proteins of S. cerevisiae and C. albicans, they do share a conserved role in regulating yeast–hyphal morphological transitions. In A. nidulans, this transition is from the filamentous growth state of the vegetative hyphae and conidiophore stalk to the uninucleate, yeast-like morphology of the sterigmata and conidia. In the yeasts, the opposite occurs with the transition being from a unicellular mode of growth to either a pseudohyphal or true hyphal form (or from hyphal to unicellular during chlamydospore formation).

Here, we present the cloning and characterization of a gene predicted to encode an APSES member from the asexual, dimorphic human pathogen Penicillium marneffei. Unlike all the fungi from which APSES proteins have been isolated to date, P. marneffei is capable of both asexual sporulation and dimorphic switching (Segretain, 1959). At 25°C P. marneffei grows in a filamentous form and produces asexual spores by a process similar to that in A. nidulans. At 37°C or in vivo in the host organism, P. marneffei grows as yeast cells that reproduce by fission (Garrison and Boyd, 1973). We have shown that the predicted protein product of P. marneffei, stuA, is highly conserved with StuA of A. nidulans. The creation of a stuA null mutant reveals a defect in conidiation, with the mutant strain lacking both metulae and phialides in the mature conidiophore, while still retaining the ability to produce spores. In contrast to S. cerevisiae and C. albicans, loss of the APSES protein from P. marneffei resulted in no obvious effect on dimorphic growth. Interestingly, in addition to the P. marneffei homologue being able to complement the conidiation defect of an A. nidulans stuA mutant, the P. marneffei protein was not able to complement the sexual defect of the A. nidulans mutant except under exceptional circumstances.

Results

The P. marneffei APSES protein most closely resembles A. nidulans stuA

A P. marneffei stuA genomic clone was generated from two partial genomic clones obtained from a combination of both a size-selected plasmid library and a λ-GEM11 genomic library using a probe derived by degenerate polymerase chain reaction (PCR; see Experimental procedures). The map of this full-length clone, pAB4746, is shown in Fig. 1. The predicted coding region of P. marneffei stuA was identified through homology to other APSES genes and encodes four exons located in conserved positions when compared with the exon/intron structure of the A. nidulans stuA gene (Fig. 1A). The predicted protein product of P. marneffei stuA shares 66% identity and 69% similarity to the A. nidulans StuA protein. The highest degree of homology between the two proteins was observed across the bHLH region (Fig. 1B), with a significant but lower level of conservation over the remaining regions of the two proteins. Comparisons across the entire APSES group showed a similar high degree of simi-larity over the bHLH region but with far less conservation outside this region (Fig. 1B).

Figure 1.

The P. marneffei stuA locus encodes a member of the APSES protein family.

A. A partial restriction map of the stuA locus. The stuA gene comprises four exons (black boxes) interrupted by three introns. The region that encodes the bHLH DNA-binding domain is found in exons 2, 3 and 4 (white box). The position of two short open reading frames upstream of the stuA coding region are also shown (grey boxes).

B. A schematic representation of the APSES protein members. The percentage identity between each protein and its closest homologue(s) are indicated for N-terminal region, bHLH DNA-binding domain (white box) and C-terminal region between each pair of proteins. S. cerevisiae Phd1p (ScP), S. cerevisiae Sok2p (ScS), C. albicans Efg1p (Ca), N. crassa Asm1 (Nc) A. nidulans StuA (An) and P. marneffei StuA (Pm).

C. An amino acid alignment of the bHLH motif from the APSES members. The positions of each of the helices are indicated by the thin lines above the alignment. The protein designations are the same as in (B). The amino acid co-ordinates of the region of each protein are indicated to the right of the alignment.

The P. marneffei stuA promoter also contains many features that are conserved in the A. nidulans stuA gene. It contains two short open reading frames (ORFs) upstream of the stuA coding region, of 145 and 49 amino acids respectively (Fig. 1A). The A. nidulans stuA promoter also contains small ORFs, termed μORFs, and at least one of these has been shown to regulate the production of the StuA protein post-transcriptionally (Wu and Miller, 1997). Examination of the P. marneffei promoter for the presence of known transcription factor recognition sequences identified several BrlA consensus binding sites just upstream of the μORF regions. BrlA sites are also found in the A. nidulans promoter region which contribute to BrlA-mediated, developmental regulation of the stuAα transcript (Wu and Miller, 1997).

Expression of P. marneffei stuA suggests a function during conidiation

The expression of the stuA gene was examined during the three major growth states of P. marneffei– vegetative hyphal cells, conidiating hyphal cells and yeast cells. The stuA transcript was only expressed at detectable levels in conidiating cells (Fig. 2). In contrast, no stuA transcript could be detected in vegetative hyphae or yeast cells of P. marneffei, suggesting that stuA may not be involved in regulating these growth states.

Figure 2.

Expression of stuA in P. marneffei is limited to the filamentous growth state. Total RNA was isolated from P. marneffei tissue grown in BHI medium at 37°C to induce yeast-like growth, at 25°C in liquid SD medium to produce vegetative mycelia or at 25°C on ANM + GABA plates to allow for conidiation to occur (25°C dev). Northern blot analysis of stuA expression was performed using 10 μg of RNA from each of the three developmental stages (37°C, 25°C and 25°C dev). RNA was isolated from FRR2161 (wild type) and screened with probes specific to either stuA or histone H3 (H3) (Ehinger et al., 1990).

Generation of a stuA mutant strain in P. marneffei

Deletion of the stuA locus was performed using a construct in which the majority of the stuA coding region had been deleted from the genomic clone and replaced with the A. nidulans pyrG gene (Fig. 3A). The A. nidulans pyrG gene is used as a selectable marker for P. marneffei transformation in combination with a P. marneffei pyrG mutant strain (SPM4) (Borneman et al., 2001). Unlike previous gene deletions reported in P. marneffei, the stuA deletion construct also contained direct repeats of the CAT gene that flank the pyrG marker. This type of construct allows for the selection of the in vivo deletion of the pyrG marker gene through recombination of the flanking sequences, a process commonly known as ‘URA blasting’ (Alani et al., 1987).

Figure 3.

Creation of a stuA null mutant strain in P. marneffei.

A. A partial restriction map of the P. marneffei stuA locus (top) is shown for comparison with the predicted restriction map for the stuA locus after integration of the deletion construct pAB5001 by homologous recombination (middle). The predicted map of the deleted stuA locus after loss of the pyrG marker gene as a result of 5-FOA counterselection is also shown (bottom). The position of the four exons of stuA (black boxes), the pyrG selectable marker (white box) and the flanking repeats (black box, white arrow) are all shown. The fragment of DNA used as a probe in (B) is also indicated (thin black line).

B. P. marneffei strain SPM4 was transformed with a purified 5.9 kb NotI fragment of pAB5001 containing the stuA deletion construct. TAB29019 was isolated as a poorly conidiating, pyrG+ transformant from this pAB5001 treatment. Genomic DNA from TAB29019 (Δ) and FRR2161 (WT) was subjected to restriction digestion with enzymes as indicated and Southern blotting to confirm that TAB29019 was a stuA deletion strain. TAB29019 was subsequently plated on medium containing 5-FOA and uridine to select for loss of the pyrG marker through homologous recombination between the flanking repeats. Analysis of the stuA locus of several independent sectors by Southern blotting confirmed the loss of the pyrG marker consistent with a specific recombination event. The results obtained for one of these clones, TAB29019b2 (Δb2), are shown. Numbers either side of the autoradiograph represent the DNA size standards measured in kb.

Thirty-two transformants were obtained using a linearized, gel-purified sample of the deletion construct (see Experimental procedures) and, of these, three had a flat, almost aconidial phenotype. The stuA locus of each of these transformants was analysed by Southern blotting, and the genomic arrangement of one of these, TAB29109, was consistent with replacement of the wild-type stuA locus with that of the deletion construct by homologous recombination (Fig. 3B). To create a strain in which pyrG could still be used as a selectable marker for the complementation experiments (see below), TAB29019 was plated onto 5-fluoroorotic acid (5-FOA) medium containing uridine and uracil to select for loss of the pyrG marker. Three 5-FOA-resistant sectors, shown to be uridine/uracil auxotrophs, had a restriction map consistent with loss of the pyrG marker at the stuA locus through recombination between the flanking repeats (Fig. 3A). The genomic Southern blot of one of these (TAB29019b2) is shown (Fig. 3B), and this strain was used for the complementation tests.

The P. marneffei stuA mutant produces abnormal conidiophores

TAB29109 was examined microscopically during several stages of growth. During vegetative growth, there were no abnormalities in the morphology of the hyphae that appeared to be associated with loss of stuA function (data not shown). Not only was the morphology of the hyphae normal, the nuclear distribution (as determined by DAPI staining) and the formation of cell wall material and septa (via calcofluor staining) were also normal (data not shown). Under conditions that normally induce conidiation, however, TAB29109 conidiophores lacked both metulae and phialides and consisted of a stalk from which chains of conidia budded directly (Fig. 4A). This indicates that stuA is required for accurate spatial organization of the developing conidiophore and is primarily required for the formation of the uninucleate sterigmata, which arise by budding in this multicellular structure.

Figure 4.

The stuA mutant produces abnormal conidiophores but is unaffected for dimorphic growth.

A. Colonies of both TAB29019 (ΔstuA) and FRR2161 (wild type) were grown for 4 days at 25°C on microscope slides coated in thin layers of ANM + GABA medium to induce conidiation. Wild-type strains of P. marneffei produce numerous conidiophores in which multiple metulae (M) and phialides (P) bud from a common stalk (S). The phialides produce chains of conidia (C). TAB29109 (ΔstuA) produces stalks (S) from which short chains of conidia are produced (C).

B. Colonies of TAB29109 and FRR2161 were grown for 4 days at 37°C on microscope slides coated with BHI medium to induce yeast-like growth. DIC optics were used to examine the overall morphology of the yeast cells (DIC), whereas DAPI staining and epifluorescence were used to investigate the arrangement of the nuclei within the individual yeast cells (DAPI). The scale bar in each figure represents 20 μM.

The P. marneffei stuA mutant shows normal yeast-like growth

In addition to the effects that members of the APSES group have on asexual development, APSES proteins have also been implicated in regulating dimorphic switching. The ability of the ΔstuA strain to grow in its yeast form was therefore examined microscopically. When grown at 37°C, TAB29109 had a wild-type morphology, producing yeast cells whose numbers, size and nuclear ratio (via DAPI staining) were all completely normal (Fig. 4B). Switching from yeast growth to hyphal growth by transfer of cultures from 37°C to 25°C also proceeded normally (data not shown). The effects of StuA in P. marneffei therefore do not extend to the dimorphic pathway.

Complementation of stuA mutants reveals differences between the A. nidulans and P. marneffei stuA genes

To confirm that the conidiation defects observed in TAB29109 resulted from loss of the stuA locus and not from a secondary mutation, complementation analysis was performed using the cloned P. marneffei stuA gene. As mentioned previously, ≈ 3.5 kb of sequence upstream of the start codon is required for full complementation of both asexual and sexual defects observed in A. nidulans stuA mutants. The asexual phenotype is less sensitive to the length of the stuA promoter, and shorter promoter lengths (1.7 kb relative to the start codon) have been shown to complement the asexual phenotype while failing to complement the sexual (Miller et al., 1991; 1992). TAB29109b2 was therefore co-transformed with pAB4342 (pyrG) and either pAB4530 (1.6 kb promoter) or pAB5000 (3.6 kb promoter), and the uracil prototrophic transformants were analysed for complementation of the conidiation defect. Unlike the situation observed in A. nidulans, in which stuA under the control of either short or long promoters can complement conidiation, only the longer P. marneffei promoter construct present in pAB5000 could completely complement the phenotype of TAB29109b2 (Fig. 5).

Figure 5.

Complementation of the P. marneffei stuA mutant.TAB29109b2 was co-transformed with pAB4342 (pyrG) and either pAB4530 or pAB5000. pAB4530 has 1.6 kb of stuA promoter sequence present upstream of the predicted stuA start codon, whereas pAB5000 has 3.6 kb of promoter sequence upstream. Four independent co-transformants using both pAB4530 (P. m 1.6 kb) and pAB5000 (P. M 3.6 kb) and a vector-only control (using only pAB4342) are shown for comparison with both TAB29019 (ΔstuA) and FRR2161 (wild type, WT).

As the effects of stuA promoter deletions on A. nidulans development have been analysed in some detail, complementation of the A. nidulansΔstuA strain using the P. marneffei constructs was also performed. The A. nidulansΔstuA strain (UI70) was transformed with both short and long P. marneffei stuA plasmids as well as with the short (pK5) and long (pk6.8) promoter versions of the A. nidulans stuA gene (Miller et al., 1992). Each construct was capable of complementing the conidiation defect of the ΔstuA strain; however, only those transformants carrying the longer version of the A. nidulans promoter (plasmid pK6.8, 3.5 kb of promoter) were capable of complementing the sexual defect in this strain, as shown by their ability to form large, black, spherical, sexual structures (cleistothecia) upon induction of the A. nidulans sexual cycle (Fig. 6).

Figure 6.

Figure 6.

Complementation of the A. nidulans stuA mutant strain.

A. A. nidulans strain UI70 (ΔstuA) was co-transformed with the bleomycin resistance plasmid pAmPh520 (Austin et al., 1990) and either the pAB4530 and pAB5000 P. marneffei stuA plasmids or the pK5 and pK6.8 A nidulans stuA plasmids (Miller et al., 1991; 1992). pK5 is analogous to the pAB4530 plasmid, containing 1.7 kb of promoter, whereas pK6.8 is analogous to the pAB5000 plasmid, containing 3.5 kb of promoter upstream of the stuA start codon. Magnified images of three independent transformants for which complementation of the conidiation defect of UI70 was observed are shown for each plasmid in addition to a vector-only control (using only the selectable marker plasmid, pAmPh520, of the co-transformation) for each after 2 days growth at 37°C. pAB4530 (P. m 1.6 kb), pAB5000 (P. m 3.6 kb), pK5 (A. n 1.7 kb) and pK6.8 (A. n 3.5 kb).

B. The colonies shown in Fig. 3A were transferred to ANM medium containing NO3 as a nitrogen source and grown for 2 days at 37°C. Plates were then sealed to induce the A. nidulans sexual cycle and incubated for 10 days at 37°C. The wild-type A. nidulans (biA1) control produces numerous cleistothecia (large, black spheres indicated by black arrows), which are found interspersed throughout the colony. In addition to cleistothecia, conidiophores are also present (green background) as the sexual cycle occurs after the asexual cycle has taken place. The UI70 strain (ΔstuA) does not produce either cleistothecia or normal conidiophores and so lacks both the black spheres and the green background. Transformants containing the pK6.8 plasmid (A. n 3.5 kb) produce conidiophores and cleistothecia (black arrow). Transformants containing the pAB5000 plasmid (P. m 3.6 kb) or the pK5 and pAB4530 plasmids (not shown) produce conidiophores but do not produce cleistothecia. The scale bar in each part equals 0.2 mm.

Figure 6.

Figure 6.

Complementation of the A. nidulans stuA mutant strain.

A. A. nidulans strain UI70 (ΔstuA) was co-transformed with the bleomycin resistance plasmid pAmPh520 (Austin et al., 1990) and either the pAB4530 and pAB5000 P. marneffei stuA plasmids or the pK5 and pK6.8 A nidulans stuA plasmids (Miller et al., 1991; 1992). pK5 is analogous to the pAB4530 plasmid, containing 1.7 kb of promoter, whereas pK6.8 is analogous to the pAB5000 plasmid, containing 3.5 kb of promoter upstream of the stuA start codon. Magnified images of three independent transformants for which complementation of the conidiation defect of UI70 was observed are shown for each plasmid in addition to a vector-only control (using only the selectable marker plasmid, pAmPh520, of the co-transformation) for each after 2 days growth at 37°C. pAB4530 (P. m 1.6 kb), pAB5000 (P. m 3.6 kb), pK5 (A. n 1.7 kb) and pK6.8 (A. n 3.5 kb).

B. The colonies shown in Fig. 3A were transferred to ANM medium containing NO3 as a nitrogen source and grown for 2 days at 37°C. Plates were then sealed to induce the A. nidulans sexual cycle and incubated for 10 days at 37°C. The wild-type A. nidulans (biA1) control produces numerous cleistothecia (large, black spheres indicated by black arrows), which are found interspersed throughout the colony. In addition to cleistothecia, conidiophores are also present (green background) as the sexual cycle occurs after the asexual cycle has taken place. The UI70 strain (ΔstuA) does not produce either cleistothecia or normal conidiophores and so lacks both the black spheres and the green background. Transformants containing the pK6.8 plasmid (A. n 3.5 kb) produce conidiophores and cleistothecia (black arrow). Transformants containing the pAB5000 plasmid (P. m 3.6 kb) or the pK5 and pAB4530 plasmids (not shown) produce conidiophores but do not produce cleistothecia. The scale bar in each part equals 0.2 mm.

The lack of complementation of the sexual cycle defects of the A. nidulansΔstuA strain by the P. marneffei stuA constructs may have resulted from the loss of this function in P. marneffei or the differences in regulation between the two species resulting in misexpression of a fully functional P. marneffei protein. To distinguish between these possibilities, the A. nidulans stuA coding region of the long promoter construct (pK6.8) was replaced with the P. marneffei stuA coding region. Several independent constructs were tested for their ability to complement the A. nidulansΔstuA strain. The hybrid stuA genes were all capable of complementing the conidiation phenotype of the A. nidulansΔstuA strain; however, only two of the 15 transformants containing the hybrid genes were capable of complementing the sexual defect by forming cleistothecia (Table 1). This is in contrast to transformants containing the long A. nidulans construct (pK6.8), in which the majority of the transformants were complemented for both asexual and sexual phenotypes. Although the frequency of sexual complementation by the hybrid constructs is low, the presence of the two complementing transformants is significant as neither the short P. marneffei (pAB4530) or A. nidulans (pK5) promoter constructs nor the long P. marneffei construct (pK5000) were capable of producing any sexually complementing transformants (Table 1).

Table 1. Complementation of the A. nidulansΔstuA strain.
 PromoterComplementation
PlasmidSpeciesSizeaCoding
region
AsexualSexual
  • a.

    Promoter length relative to +1ATG.

  • b.

    Results were pooled from the analysis of three transformants with each of five independent hybrid constructs.

pK6.8A.n3.5 kbA.n3/32/3
pK5A.n1.7 kbA.n3/30/3
pAB5000P.m3.6 kbP.m3/30/3
pAB4530P.m1.6 kbP.m3/30/3
HybridSizebA.n3.5 kbP.m15/152/15

Given the nature of the stuA promoter and the effect that truncations have on the complementation pattern of the gene, Southern blotting was used to confirm the presence of the intact hybrid construct in each of these transformants. Most of the transformants had a single, intact copy of the plasmid inserted into the genome. For those transformants that had more than one integrated copy of the plasmid, there was no correlation between the presence of multiple plasmid insertions and the ability to complement the sexual phenotype.

Discussion

The control of morphological transitions by P. marneffei StuA is restricted to those that occur during the formation of the conidiophore

APSES proteins have been implicated in regulating development in several species of fungi, primarily associated with controlling alterations to cellular morphology. P. marneffei is unique among the fungi from which an APSES homologue has been characterized as it is capable of both conidiation and dimorphic growth. In P. marneffei, only the conidiation pathway is dependent on StuA, where it is required for correct cell patterning to ensure the proper formation of the metulae and phialides of the conidiophore. The morphology of these cell types is similar to those of pseudohyphal yeast cells, as they are uninucleate and are formed by budding (Gimeno et al., 1992). It has been suggested previously that the formation of these cells in filamentous fungi may involve similar processes to those needed during pseudohyphal growth in yeast (which is analogous to dimorphic growth) (Busby et al., 1996). As P. marneffei conidiates but is also truly dimorphic, the hypothesis that StuA would control transitions between cells with a yeast-like morphology and those with a filamentous morphology during both developmental pathways could be formally tested. Unlike the effect of loss of stuA on the development of the yeast-like cells of the conidiophore, the formation of yeast cells during dimorphic growth is unaffected in P. marneffei. This distinguishes P. marneffei, which forms yeast cells by fission, from the budding yeasts S. cerevisiae and C. albicans, in which the APSES proteins Phd1p, Sok2p and Efg1p regulate the yeast–hyphal transition (Gimeno and Fink, 1994; Ward et al., 1995; Stoldt et al., 1997). This suggests that APSES proteins may not be general regulators of all yeast–hyphal transitions, but may instead be limited to controlling those pathways in which a budding mode of growth is required.

The finding that stuA controls the conidiation pathway but not the dimorphic pathway also differs from the situation observed for the AbaA protein in P. marneffei, which appears to function during both conidiation and dimorphic growth pathways (Borneman et al., 2000). This suggested that regulators of conidiation in P. marneffei may also regulate dimorphic growth. The analysis of P. marneffei StuA therefore not only shows that APSES proteins may not be general regulators of all yeast–hyphal transitions, but also indicates that the function of ‘conidiation-specific’ genes during dimorphism in P. marneffei may not be a general rule.

StuA and sexual development in P. marneffei

Although the P. marneffei stuA gene was capable of complementing the asexual sporulation defect of an A. nidulansΔstuA mutant, the function of the P. marneffei protein during sexual development relied on the A. nidulans promoter driving the P. marneffei coding region. Furthermore, complementation was only observed in a small percentage of transformants and was probably caused by position effects resulting from the random integration of the transforming plasmids affecting stuA expression in the complementing strains. The P. marneffei protein may therefore only be functional when expressed at levels that are higher than those produced by either P. marneffei or A. nidulans promoters alone. This indicates that not only are there differences in the function of the P. marneffei promoter during the A. nidulans sexual cycle, but that the two proteins are also significantly different. Although this is not unexpected given that P. marneffei has no recognized sexual cycle, it is of interest for two reasons. First, it has been shown that Efg1p of C. albicans (also an asexual species) is functionally interchangeable with Sok2p during the regulation of IME1 expression during sporulation in S. cerevisiae (Shenhar and Kassir, 2001). Secondly, it has been shown that the stlA gene of P. marneffei (a homologue of the S. cerevisiae STE12 gene) is able to function interchangeably with its A. nidulans homologue, steA, during sexual development (Vallim et al., 2000; Borneman et al., 2001). The divergence of the P. marneffei stuA gene may be a direct result of the asexual nature of the fungus. It has been shown previously that the A. nidulans stuA promoter is complex and the sexual cycle is very sensitive to alterations in this region, whereas the asexual cycle is less sensitive (Wu and Miller, 1997). Loss of the sexual cycle in P. marneffei may therefore have been a recent event, with the differences observed in the complementation of the A. nidulans sexual cycle by stuA and stlA resulting from a lower level of tolerance to new mutations in both the promoter and the coding region of stuA, resulting in a non-functional gene.

Alternatively, as suggested by the stlA data, P. marneffei may indeed have an unidentified sexual cycle. In this situation, the limited function of the P. marneffei protein in A. nidulans may reflect differences in the factor(s) that interact with either the stuA promoter or protein during sexual development in the two species. The function of stuA during a cryptic P. marneffei sexual cycle, although being supported by the presence of stlA, is also complicated by it. The conservation of stlA sequence and function with the steA gene of A. nidulans suggests that, if a sexual cycle is present in P. marneffei, it would be dependent on the StlA protein. The overlapping pattern of expression observed for stuA and steA in A. nidulans is consistent with both genes being involved in regulating sexual development (Vallim et al., 2000). In P. marneffei, however, the expression pattern of stlA (high levels in vegetative tissues, lowered expression during conidiation) is completely different from that of stuA. This may represent a significant difference in P. marneffei, which, in conjunction with the failure of the stuA homologue to complement the A. nidulansΔstuA mutation fully, suggests that, if P. marneffei has a sexual cycle, then StuA would not play role in this process.

APSES proteins – general regulators of budding growth

Of the developmental processes in which APSES proteins are involved and for which detailed morphological data have been obtained (sexual reproduction in N. crassa, formation of the metulae and phialides in A. nidulans, dimorphic switching and chlamydospore formation in C. albicans and pseudohyphal growth in S. cerevisiae), it appears that the most generalized function for APSES members is in controlling transitions either to or from budding modes of growth or between different modes of budding (Miller et al., 1992; Gimeno and Fink, 1994; Ward et al., 1995; Aramayo et al., 1996; Stoldt et al., 1997; Sonneborn et al., 1999). This regulation of budding appears to be quite specific as, even though P. marneffei produces yeast cells, these cells are formed by fission. It is interesting to note that Schizosaccharo-myces pombe, a yeast that divides by fission rather than budding, appears to lack an APSES homologue (based on the S. pombe genome sequence; http://www.sanger.ac.uk/Projects/S_pombe/).

This budding specificity may also explain why conidia are still produced in P. marneffei and A. nidulans stuA strains as the production of spores requires a specialized, basipetal mode of division, whereas the metulae and phialides are produced by acropetal divisions (Mims et al., 1988; Sewall et al., 1990). StuA may only control the more general acropetal division pattern, resulting in loss of the metulae and phialides, while leaving the specialized spore-producing basipetal mode of division able to produce spores from the conidiophore vesicle. This hypothesis also predicts that sexual reproduction in A. nidulans involves some aspect of budding growth as it is dependent on StuA. This may well be the case as some of the cell types that are formed during sexual development in this species have morphological characteristics consistent with a budding mode of division (e.g. Hülle cells) (Champe and Simon, 1992). If P. marneffei does have a cryptic sexual cycle, it is of considerable interest to examine the role that stuA plays in this programme.

Experimental procedures

Fungal strains and growth conditions

Fungal strains used in this study include the A. nidulans strain UI70 (biA1, methG1, argB2, ΔstuA::argB; provided by B. Miller, University of Idaho, USA) and the P. marneffei strains SPM4 (pyrG1, niaD1; Borneman et al., 2001), TAB291019 (pyrG1, niaD1, ΔstuA::pyrG) and TAB29109b2 (pyrG1, niaD1, ΔstuA). A. nidulans strains were grown on Aspergillus standard media (ANM) (Cove, 1966). Carbon sources were used at 1% (w/v) unless otherwise stated, nitrogen sources at 10 mM and other supplements as required. P. marneffei strains were grown on either standard A. nidulans media with γ-amino butyric acid as the sole nitrogen source (ANM + GABA), brain–heart infusion (BHI) medium (Oxoid) or yeast synthetic dextrose (SD) medium (Ausubel et al., 1994), with supplements as required. Strain SPM4 was supplemented with 10 mm uridine or uracil.

Molecular techniques

DNA-mediated transformation of P. marneffei was performed by the polyethylene glycol-mediated protoplast fusion method (Borneman et al., 2001), whereas A. nidulans transformations were performed as described by Andrianopoulos and Hynes (1988). Genomic DNA and RNA were isolated as described previously (Borneman et al., 2000). Southern and Northern blotting are also described elsewhere (Sambrook et al., 1989). Dideoxy sequencing was performed at the AGRF sequencing facility (University of Queensland, Australia) using ABI prism Big Dye terminator chemistry and the M13-20 and M13 reverse primers on nested deletions and the gene-specific primers stuA566 (TGTGCTTTCGCT GTTCTG) and stuA2584 (CAGAACAGCGAAAGCACAAC). Sequence alignments were performed using PILEUP (GCG software package) (Devereux et al., 1984) and MACBOXSHADE sequence analysis tools. Identical amino acids are shaded black, similar are shaded grey. The genomic sequence of the stuA gene has been deposited in GenBank as accession number AF436076.

Cloning and disruption of the stuA locus

The stuA homologue was isolated by degenerate PCR using the primers STUPHD1 (ACNHTGTGGGARGAYGAR) and STUPHD2 (CANACNCCYTTNARRTGC), designed using the conserved bHLH regions of A. nidulans StuA (accession number AAA33325) and S. cerevisiae Phd1p (AAA18358). PCR conditions consisted of an initial denaturation step of 94°C for 2 min, after which Taq DNA polymerase (Promega) was added, and 30 cycles of 94°C for 30 s, 50°C for 30 s, 72°C for 30 s were performed. Three predominant products were obtained, one of which was of the expected size of ≈ 250 bp. Sequence analysis of individual clones revealed that this band was composed of several species, one of which had significant homology to APSES homologues. This PCR product was used to probe a 4.5–6 kb EcoRI–BamHI size-selected FRR2161 genomic library in pBluescript-II SK+ (Stratagene). One positive clone was isolated (pAB4417), but subsequent sequencing analysis revealed that this clone lacked the 3′ portion of the gene. A 2.3 kb SacI fragment, which overlapped the 3′ end of pAB4417, was therefore obtained from a λ-GEM11 FRR2161 genomic library and cloned into pBluescript-II SK+ to give pAB4747. The BglII–XbaI fragment of pAB4747 was then cloned into the BglII–XbaI sites of pAB4417 to give the full-length clone (pAB4746). The 6.6 kb EcoRI fragment of pAB4746 was cloned into the EcoRI sites of pGEM-T easy (Stratagene) to give pAB5000. pAB4530 was created by digesting pAB4417 with HindIII and religating the vector to delete a 2 kb fragment of the stuA 5′ region to give pAB4420. The 3′BglII–XbaI fragment of the stuA gene from pAB4747 was then inserted into the BglII–XbaI sites of pAB4420 to give an intact stuA coding region in pAB4530.

To disrupt stuA, a 2.2 kb XhoI–BamHI fragment containing the pyrG blaster cassette from pAB4626 was cloned into the XhoI–BamHI sites of pAB5000 to give pAB5001. pAB4626 consists of a 1.8 kb fragment containing the A. nidulans pyrG gene flanked by 400 bp direct repeats of the Escherichia coli chloramphenicol acetyltransferase (CAT) gene cloned into the pBluescript-II SK+ vector at the PstI and SmaI sites. pAB5001 was digested with NotI, subjected to gel electrophoresis, and the 5.9 kb insert was gel purified using the Bresaclean gel purification kit (Geneworks). A gel-purified fragment (500 ng) was used to transform SPM4 as described by Borneman et al. (2001), and transformants were selected for complementation of the uridine/uracil auxotrophy. To select for loss of the pyrG marker by homologous recom-bination, transformant spores were point inoculated onto ANM + GABA medium containing 1 mg ml−1 5-FOA, 5 mM uracil and 5 mM uridine and incubated at 25°C until resistant sectors were observed. These sectors were then maintained on ANM + GABA medium supplemented with 10 mM uracil or uridine.

Complementation of the A. nidulans ΔstuA strain and swapping of the stuA coding region

The A. nidulansΔstuA strain (UI70) was transformed as described previously (Andrianopoulos and Hynes, 1988). Plasmids containing various stuA constructs were co-transformed in conjunction with the bleomycin resistance plasmid pAmPh520 (Austin et al., 1990), and bleomycin-resistant colonies were selected by plating on regeneration medium containing 1 mg ml−1 bleomycin. The A. nidulans stuA plasmids pK6.8 and pK5 were kindly provided by B. Miller (University of Idaho, USA). To exchange the P. marneffei stuA coding region for the A. nidulans sequence in pK6.8, phosphorylated PCR primers PmstuAF (ATGAATCAAACACAAT CATACATGG) and PmstuAR (TCATCGGCGTCGCATTAC TCCT) were used to amplify the P. marneffei stuA coding region, whereas the primers AnstuA5′ATG (CTGGGTCTG GTTGGGGTAGTAG) and AnstuA3′stop (AGGCGGGTG GCAGTCTGATAAG) were used to amplify the AnstuA 5′ untranslated region (UTR), promoter, 3′ UTR and the pBR322 vector sequences of pK6.8 by inverse PCR. The PCR was performed using Pfu Turbo polymerase (Stratagene) with the cycling conditions consisting of an initial denaturation step of 94°C for 2 min, followed by 20 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 8 min with a final extension step of 72°C for 15 min. The two PCR products were ligated together, and hybrid plasmids were screened for those clones with the correct orientation of the P. marneffei stuA insert in the pK6.8 backbone by restriction digest. A. nidulans transformants were screened for both copy number and intact stuA constructs by Southern blotting as indicated.

Microscopy

Penicillium marneffei cultures were treated and examined as described by Borneman et al. (2000).

Acknowledgements

We thank Bruce Miller (University of Idaho, USA) for the kind gift of strains and plasmids, and Bruce Miller and Mogens Trier Hansen for useful discussion on this work. This work was supported by a grant from the Australian Research Council. A.R.B. was supported by an Australian Postgraduate Award (APA).

Ancillary