TupA, the Penicillium marneffei Tup1p homologue, represses both yeast and spore development

Authors


Summary

Fungal pathogenesis is frequently associated with dimorphism – morphological changes between yeast and filamentous forms. Penicillium marneffei, an opportunistic human pathogen, exhibits temperature-dependent dimorphism, with growth at 25°C as filamentous multinucleate hyphae switching at 37°C to uninucleate yeast cells associated with intracellular pathogenesis. The filamentous hyphae also undergo asexual development generating uninucleate spores, the infectious propagules. Both processes require a switch to coupled nuclear and cell division. Homologous regulators, including Tup1p/GROUCHO-related WD40 repeat transcription factors, control dimorphism in Candida albicans and asexual development in Aspergillus nidulans. Unlike these fungi, P. marneffei has both developmental programmes allowing examination of common and programme-specific controls. We show that deletion of tupA, the P. marneffei TUP1 homologue, confers reduced filamentation and inappropriate yeast morphogenesis at 25°C, in stark contrast to constitutive filamentation observed when C. albicans TUP1 is deleted. Deletion of tupA also confers premature brlA-dependent asexual development, unlike reduced asexual development in the corresponding A. nidulans rcoA deletion mutant. Furthermore, the A. nidulans rcoA deletion mutant is self-sterile, and we show that tupA from P. marneffei, which lacks an apparent sexual cycle, complements both the asexual and sexual development phenotypes. Therefore, TupA coordinates cell fate by promoting filamentation and repressing both spore and yeast morphogenetic programmes.

Introduction

Penicillium marneffei is emerging as an important opportunistic human fungal pathogen endemic to South-East Asia, where it is AIDS-defining, and it is being found increasingly around the world (Wong et al., 1999; Cooper and Haycocks, 2000). Penicillium marneffei infections may occur as superficial cutaneous lesions or disseminated infections that result in death when untreated. Infection by fungal pathogens such as P. marneffei, Histoplasma capsulatum, Candida albicans, Cryptococcus neoformans, Coccidioides immitis and Aspergillus fumigatus has shown an alarming increase over the last decade posing a substantial health threat to immunocompromised individuals (Walsh and Groll, 1999). Fungal pathogenesis is frequently associated with dimorphism – morphological changes between yeast and filamentous forms. Penicillium marneffei exhibits temperature-dependent dimorphism and is the only Penicillium species known to be dimorphic (Segretain, 1959; Garrison and Boyd, 1973; Pitt, 1979; Chan and Chow, 1990). At 25°C growth is filamentous with multinucleate hyphae that can differentiate to form asexual reproductive structures (conidiophores) bearing uninucleate spores (conidia). Growth at 37°C is as uninucleate yeast cells, which divide by fission, and this form is associated with disease as an intracellular pathogen similar to H. capsulatum. Switching from filamentous to yeast morphology upon shift from 25°C to 37°C occurs by arthroconidiation, as in C. immitis, where the hyphal cells form double cell walls and separate from each other. The switching process is reversible in that, when transferred to 25°C, yeast cells grow by apical extension to form hyphal filaments.

In addition to the dimorphic phase transition, P. marneffei also undergoes asexual development resulting in formation of conidia – a process well studied in Aspergillus nidulans (Adams et al., 1988; Adams et al., 1998). For both processes a switch from filamentous, uncoupled nuclear and cell division to yeast-like, coupled nuclear and cell division is required and, therefore, the presence of both programmes in the same organism is of considerable interest with respect to common and specific controls. It has been found that P. marneffei abaA, an homologue of the A. nidulans abaA asexual regulator (Andrianopoulos and Timberlake, 1994), is required for correct conidiophore development at 25°C and the coupling of nuclear and cell division during yeast cell growth at 37°C (Borneman et al., 2000). TEC1, the abaA homologue from the dimorphic fungi C. albicans and Saccharomyces cerevisiae, is required for hyphal and pseudohyphal growth, respectively, and expression of A. nidulans abaA in S. cerevisiae results in pseudohyphal growth (Gavrias et al., 1996; Madhani and Fink, 1997; Schweizer et al., 2000). AbaA and TEC1 are members of the ATTS transcriptional regulator family. Other regulators, such as members of the APSES DNA-binding domain class of transcription factors –A. nidulans stuA, C. albicans EFG1, and S. cerevisiae PHD1 and SOK2– also control asexual sporulation and dimorphic development (Gimeno and Fink, 1994; Ward et al., 1995; Dutton et al., 1997; Lo et al., 1997; Stoldt et al., 1997). Tup1p/GROUCHO-related WD40 repeat transcription factors are also involved in regulation of both developmental programmes. In A. nidulans and Neurospora crassa the TUP1 homologues rcoA and rco-1, respectively, encode pleiotropic repressors involved in asexual development (Yamashiro et al., 1996; Hicks et al., 2001). In C. albicans homozygous tup1 deletion mutants show constitutive filamentous growth and reduced virulence (Braun and Johnson, 1997; Braun et al., 2000) and in S. cerevisiae tup1 loss-of-function mutants show reduced diploid pseudohyphal growth and reduced haploid invasive growth (Braun and Johnson, 1997). Saccharomyces cerevisiae Tup1p also acts as a pleiotropic repressor of glucose repressible, oxygen repressible, DNA-damage inducible, haploid specific, a-mating type specific and flocculation genes (Williams and Trumbly, 1990; Komachi et al., 1994; Treitel and Carlson, 1995;Braun and Johnson, 1997; DeRisi et al., 1997). Tup1p is recruited to specific sets of promoters, as a complex with Ssn6p, through interaction with specific DNA-binding proteins for each functionally related set of genes (Komachi et al., 1994; Treitel and Carlson, 1995; Tzamarias and Struhl, 1995). Tup1p mediates repression by interaction with RNA polymerase II holoenzyme components and alteration of chromatin structure through interactions with histone deacetylases and histones H3 and H4 (Cooper et al., 1994; Edmondson et al., 1996; Kuchin and Carlson, 1998; Watson et al., 2000).

Here we show that P. marneffei tupA, the homologue of TUP1, plays a fundamental role in repressing both developmental processes: asexual development and yeast cell morphogenesis. Therefore tupA represents a focal point for regulation of the switches between undifferentiated filamentous growth and development and an important factor in determining virulence due to its role in regulation of production of both the infectious propagules (spores) and the pathogenic yeast form.

Results

TupA, the highly conserved P. marneffei TUP1 homologue, controls growth and morphogenesis

The P. marneffei tupA gene was cloned by homology to the A. nidulans rcoA gene (Hicks et al., 2001) (see Experimental procedures). The TupA protein contains a predicted N-terminal coiled coil (residues 13–40), a proline-rich region (residues 127–223), seven WD40 repeats (residues 284–577) (Fig. 1) and is highly conserved with A. nidulans RcoA, N. crassa RCO1, C. albicans TUP1, and S. cerevisiae Tup1p (Fig. 1). The tupA gene was inactivated by homologous gene replacement (see Experimental procedures). The tupA deletion mutant showed a reduced growth rate at both 25°C and 37°C on a range of sole carbon and nitrogen sources and severe morphological defects at 25°C (Fig. 2A–C). These phenotypes were fully complemented by introduction of the tupA gene into the tupAΔ mutant (Fig. 2A–C and data not shown). Production of a diffusible red pigment at 25°C, a defining characteristic of P. marneffei (Segretain, 1959), was increased in the tupAΔ mutant (Fig. 2D).

Figure 1.

Penicillium marneffei TupA is highly conserved. The P. marneffei TupA (GenBank Accession AY082798), A. nidulans RcoA (GenBank Accession AF197225), C. albicans TUP1 (GenBank Accession AF005741), N. crassa RCO1 (GenBank Accession U57061) and S. cerevisiae Tup1p (GenBank Accession U57061) protein sequences were aligned using the clustalw1.5 program (Thompson et al., 1994) and displayed as standard single letter code using the Boxshade public domain software available by anonymous ftp (ftp.isrec.isb-sib.ch). Conserved residues are shaded black (identity) or grey (similar) according to Boxshade. P. marneffei TupA is highly conserved with A. nidulans RcoA (91% similarity/86% identity), N. crassa RCO1 (76% similarity/63% identity), C. albicans TUP1 (70% similarity/53% identity), and S. cerevisiae Tup1p (61% similarity/44% identity). The predicted N-terminal coiled-coil (residues 13–40) (dashed line), proline-rich region (residues 127–223) (grey line) and seven WD40 repeats (residues 284–577) (black lines) in the P. marneffei TupA protein are shown.

Figure 2.

P. marneffei tupA is required for growth and morphogenesis.
A–C. Deletion of tupA results in reduced growth rate and morphological defects. Wild type, tupAΔ mutant and two complementing tupAΔ transformants, containing 1–2 copies (bottom left) and 10–20 copies (bottom right) of the wild-type tupA gene, were streaked (B) and point inoculated (C) as indicated (A) on SD ammonium medium and grown for 4 days at 37°C (B) and 3 days at 25°C (C).
D. Deletion of tupA results in inappropriate pigment production. The wild type and tupAΔ mutant were grown in liquid SD ammonium for 4 days at 25°C and examined for red pigment production.

TupA represses asexual development at 25°C

Microscopic examination of tupAΔ colonies indicated that they prematurely initiated the asexual development programme. In the presence of 2.0% glucose the tupAΔ mutant produced conidiophores and conidia prolifically after 3 days (Fig. 3A). In contrast, no asexual reproductive structures (conidiophores) were evident at 3 days in the wild type (Fig. 3A) and became abundant only after 7 days (data not shown). Growth on limiting carbon (0.1% glucose) or poor carbon sources (e.g. 2.0% ethanol or 2.0% glycerol) resulted in conidiation in both strains after 3 days (Fig. 3A and data not shown). This effect is due to carbon limitation rather than carbon catabolite repression (CCR) as asexual development is observed in both strains on a range of repressible carbon sources in the presence of 2-deoxy-d-glucose (2-DOG), a glucose analogue which triggers CCR (data not shown). Furthermore, asexual development appears to be a specific response to carbon limitation rather than a general stress response, as osmotic stress (0.8 M NaCl) does not accelerate asexual development in the wild type (data not shown). Therefore tupA is necessary for the repression of asexual development in the presence of high concentrations of glucose.

Figure 3.

Penicillium marneffei tupA is required for repression of asexual development.
A. In the presence of 2.0% glucose asexual development is repressed in the wild type and derepressed in the tupAΔ mutant at 25°C. The wild type and tupAΔ mutant were grown on synthetic media containing 2.0% and 0.1% glucose as the sole carbon source on microscope slides for 3 days (d) at 25°C. Fixed samples were viewed using Differential Interference Contrast (DIC) optics. Conidiophore cell types are indicated: metulae (M), phialides (P) and conidia (C) Scale bar represents 20 µm.
B. The brlA gene is prematurely expressed at 25°C in the tupAΔ mutant. Total RNA was isolated from wild-type and tupAΔ 25°C cultures synchronized as described in (Borneman et al., 2000) and grown on solid SD medium for the indicated number of days (d), subjected to electrophoresis and Northern blotting and probed with the P. marneffei brlA gene (A. R. Borneman, M. J. Hynes and A. Andrianopoulos., unpublished data).

The P. marneffei brlA gene is necessary and sufficient for asexual development as deletion of brlA results in stalk formation but no further development of asexual structures and overexpression of brlA confers asexual development (A. R. Borneman, M. J. Hynes and A. Andrianopoulos, unpublished data). These data indicate that, as is well-established in A. nidulans (Adams et al. (1988), brlA encodes the primary transcriptional regulator of the asexual development pathway. brlA is expressed during asexual development and not during vegetative filamentous or yeast growth (A. R. Borneman, M. J. Hynes and A. Andrianopoulos, unpublished data). Northern analysis showed that the brlA transcript was prematurely expressed after 1 day in the tupAΔ mutant compared with the wild type where the brlA transcript was not observed until 3 days (Fig. 3B). This is consistent with premature conidiation in the tupAΔ and indicates that TupA represses asexual development temporally by acting on the brlA regulatory pathway. This is in striking contrast with the phenotype of the corresponding rcoAΔ mutant in A. nidulans which shows reduced conidiation and decreased brlA transcript levels (Hicks et al., 2001).

TupA represses yeast morphogenesis at 25°C

Although early in incubation (3–7 days) at 25°C the tupAΔ colonies appeared filamentous, upon extended incubation (10–12 days) they changed to a yeast-like appearance (Figs 2C and 4A). The inappropriate formation of uninucleate yeast-like cells after 12 days at 25°C on solid media (Fig. 4A and B), suggests that tupA is required for repression of yeast morphogenesis during filamentous growth. The brlA asexual development regulator is expressed in the tupAΔ mutant after 12 days at 25°C (Fig. 4C). However, neither asexual structures nor spores were observed in the tupAΔ mutant in 12 day, 25°C cultures, indicating that, unlike in early culture, brlA expression is unable to drive asexual development in this context. Simultaneous derepression for asexual development and yeast morphogenesis late in 25°C cultures may result in yeast-like cells predominating over the multicellular structure of conidiophores. The absence of yeast-like cells in early 25°C developmental culture is probably caused by a lack of developmental competence in the early culture. In support of this, switching of wild type from 25°C to 37°C requires 3–4 days of filamentous growth prior to yeast morphogenesis.

Figure 4.

tupA is required for repression of yeast morphogenesis during filamentous growth.
A. The tupAΔ mutant changes in morphology over time at 25°C. The wild type and tupAΔ mutant were grown on SD (2.0% glucose) medium for 7, 10 and 12 days (d) at 25°C. Scale bar represents 2 mm. Note the difference in magnification for wild type and the tupAΔ mutant.
B. Uninucleate yeast-like cells are inappropriately formed in the tupAΔ mutant at 25°C. Wild type and the tupAΔ mutant were grown as described in Fig. 4A for 12 days (d) with the microscope slides embedded in agar to prevent dessication prior to fixing. DIC and fluorescence imaging (HC) using Hoechst 33258 (nuclei; white arrowhead) and calcofluor (cell walls; grey arrowhead) are shown. Scale bar represents 20 µm.
C. The brlA gene is expressed in the tupAΔ mutant during yeast morphogenesis. Total RNA was isolated from wild type and tupAΔ 25°C synchronous cultures grown on solid SD medium for 12 days (d) and liquid SD cultures grown at 37°C for 4 days (d). Northern analysis was carried out as described in Fig. 3B. D. Uninucleate yeast-like cells are inappropriately produced at 25°C in submerged culture in the tupAΔ mutant. The wild type and tupAΔ mutant were grown at 25°C in SD liquid media for 6 days. DIC and fluorescence imaging (HC) using Hoechst 33258 (nuclei; arrowhead) and calcofluor (cell walls; grey arrowhead) are shown. Red pigment production visible as dark colouration in DIC obscures the nuclear stain in some cells. Scale bar represents 20 µm.

Asexual development is strongly inhibited by growth in submerged culture at 25°C in wild-type P. marneffei. In 25°C submerged culture the tupAΔ mutant also does not undergo asexual development (Fig. 4D). Therefore repression of asexual development in submerged culture appears to operate independently of TupA. In contrast to the filamentous growth of the wild type, the tupAΔ mutant shows little filamentous growth and uninucleate yeast-like cells, which are deformed in shape and not always detached, predominate (Fig. 4D), indicating that the tupA gene is required for the repression of yeast cell morphogenesis at 25°C.

TupA is involved in maintenance but not initiation of filamentous growth

To determine whether TupA regulates the dimorphic phase transition or the maintenance of filamentous growth, a time course analysis of the dimorphic transitions between filamentous and yeast forms was performed. When the temperature was shifted from 37°C to 25°C, the wild type and the tupAΔ mutant both switched from yeast to filamentous morphology (Fig. 5A). However, the tupAΔ mutant showed premature asexual development, little filamentous growth and initiated a return to yeast-like morphology upon extended incubation (data not shown). Thus, the tupAΔ mutant is able to initiate the switch from the yeast to filamentous form but cannot maintain the filamentous morphology. When the temperature was altered from 25°C to 37°C, the wild-type strain switched from filamentous to yeast morphology at the actively proliferating periphery of the colony while the tupAΔ mutant switched to the yeast form throughout the colony (Fig. 5B). This is consistent with the yeast-like colony morphology observed in the tupAΔ mutant upon extended incubation at 25°C and suggests that TupA is required for maintenance of the filamentous state by repression of yeast morphogenesis. This contrasts with the situation in C. albicans where tup1 mutants show constitutive filamentation and TUP1 is required for the repression of filamentous growth (Braun and Johnson, 1997). The negative control of yeast morphogenesis by TupA suggests that the yeast form may be the default in P. marneffei.

Figure 5.

Deletion of tupA does not affect initiation of dimorphic transitions.
A. The Yeast to Filamentous Switch. Wild-type and tupA deletion strains were streaked on SD ammonium medium and grown for 4 days at 37°C, then moved to 25°C to induce the yeast to filamentous transition.
B. The Filamentous to Yeast Switch. Wild-type and tupA deletion strains were streaked on SD ammonium medium and grown for 4 days at 25°C, then moved to 37°C to induce the filamentous to yeast transition. Scale bar represents 1.0 mm.

TupA is required for correct yeast cell polarity at 37°C

At 37°C with ammonium as the sole nitrogen source the wild type and the tupAΔ mutant formed smooth yeast colonies, but the tupAΔ mutant grew more slowly than the wild type (Figs 2B and 6A). Microscopic examination revealed that the tupAΔ yeast-like cells were predominantly uninucleate but deformed in shape (Fig. 6B). Many cells showed non-uniform chitin deposits suggesting a cell wall abnormality. Thus tupA is required for correct yeast cell polarity. The asexual development regulator brlA was inappropriately expressed at 37°C in the tupAΔ mutant (Fig. 4C) but asexual structures were not observed (Fig. 6A and B). This is similar to our observations that in the tupAΔ mutant brlA expression in 25°C yeast cells is unable to drive asexual development.

Figure 6.

Deletion of tupA affects morphogenesis and yeast cell polarity at 37°C.
A. Deletion of tupA promotes yeast morphogenesis at 37°C. Wild type and the tupAΔ mutant were streaked on SD media containing 10 mM ammonium (NH4), 10 mM glutamate (GLU), 10 mM alanine (ALA) and 10 mM gamma-amino butyric acid (GABA) as the sole nitrogen source and incubated for 3 days at 37°C. Scale bar represents 0.25 mm.
B. Wild type and the tupAΔ mutant were grown on SD ammonium medium on slides for 4 days at 37°C, fixed and co-stained with Hoechst 33258 (nuclei; white arrows) and calcofluor (cell walls; grey arrows). Samples were viewed using Differential Interference Contrast (DIC) or the appropriate UV fluorescence filter set (HC) for Hoechst 33258 and calcofluor staining. Inappropriate calcofluor staining chitin deposits are indicated (^). Scale bar represents 10 µm.

TupA negatively regulates yeast morphogenesis during filamentation in response to nitrogen source at 37°C

Filamentation, visible as fuzzy colony edges and as hyphal filaments at the microscopic level, is observed in wild type in the presence of some sole nitrogen sources (see Experimental procedures) other than ammonium at 37°C (Fig. 6A and data not shown). In the tupAΔ mutant the degree of filamentation was greatly reduced (Fig. 6A). Thus, similar to its role in repressing yeast morphogenesis at 25°C, TupA acts negatively on yeast morphogenesis under conditions favouring filamentation at 37°C.

P. marneffei tupA complements the growth, asexual and sexual development phenotypes of the A. nidulans rcoAΔmutant

To determine whether tupA and A. nidulans rcoA are functional homologues tupA was introduced into the A. nidulans rcoAΔ mutant (see Experimental procedures). tupA complemented the rcoAΔ reduced conidiation and growth defects to the same extent as rcoA (Fig. 7A and B). Unlike P. marneffei, which lacks a known sexual cycle, A. nidulans is homothallic (self-fertile) and undergoes sexual development to form sexual structures (cleistothecia) that harbour meiotic products (ascospores). The A. nidulans rcoAΔ mutant is self-sterile due to a defect in female sexual development resulting in a lack of cleistothecia (R. B. Todd, M. J. Hynes and A. Andrianopoulos, unpublished data). The P. marneffei tupA gene complemented the self-sterile phenotype of the A. nidulans rcoAΔ mutant allowing the production of mature cleistothecia that contained viable ascospores (Fig. 7B). Thus, with respect to growth and asexual development, tupA and rcoA may be true functional homologues and the difference in the phenotypes of the P. marneffei and A. nidulans mutants probably reflects a fundamental difference in the mechanism of regulation of asexual development in these two fungi. Furthermore, the sexual function of tupA has been conserved even though P. marneffei is considered an asexual species.

Figure 7.

The tupA gene from the asexual fungus P. marneffei complements the growth, asexual and sexual development defects of the A. nidulans rcoAΔ mutant.
A. Penicillium marneffei tupA complements the growth and asexual development defects of the A. nidulans rcoAΔ mutant. The P. marneffei tupA and A. nidulans rcoA genes were introduced into the A. nidulans rcoAΔ mutant as described in Experimental procedures. The wild type (biA1), rcoAΔ parent, and transformants containing the pyrG marker and A. nidulans rcoA, pyrG and P. marneffei tupA and pyrG alone were inoculated on Aspergillus complete media and grown at 37°C for 2 days.
B. Penicillium marneffei tupA complements the sexual and asexual development defects of the A. nidulans rcoAΔ mutant. The strains described in Fig. 7A were inoculated to appropriately supplemented Aspergillus nitrate minimal media, sealed after 2 days at 37°C to induce sexual development by oxygen deprivation and incubated for an additional 6 days. Sexual structures (cleistothecia; S) and asexual structures (conidiophore heads; A) are indicated. Scale bar represents 200 µm.

Discussion

TupA promotes filamentous growth by repression of asexual development and yeast morphogenesis

The data presented here implicate P. marneffei TupA as a central coordinator of morphogenesis and growth. Penicillium marneffei is capable of both asexual development and yeast morphogenesis from the filamentous state and must therefore repress both developmental programmes in order for vegetative filamentous growth to proceed. The primary function of TupA is to repress both of these developmental processes during filamentous growth. For asexual development, TupA acts on the brlA transcriptional regulator of this programme. The mechanism by which TupA acts to repress yeast morphogenesis awaits the isolation of regulators of yeast morphogenesis and potential downstream target genes.

P. marneffei TupA and A. nidulans RcoA represent conserved morphogenetic switches

The role of TupA bears both similarities to and differences from Tup1p homologues involved in morphogenesis in other fungi (Yamashiro et al., 1996; Braun and Johnson, 1997; Hicks et al., 2001). Penicillium marneffei TupA is required for the repression of asexual development. In contrast, RcoA and RCO1 from A. nidulans and N. crassa, respectively, are required for wild-type levels of asexual development (Yamashiro et al., 1996; Hicks et al., 2001). In P. marneffei brlA is prematurely expressed early in development in the tupAΔ mutant whereas in A. nidulans the rcoAΔ mutant shows delayed expression of brlA (Hicks et al., 2001). That tupA and rcoA are interchangeable with respect to asexual development indicates that although protein function is conserved, the molecular mechanisms of action or context are fundamentally different in these fungi. The difference in tupAΔ and rcoAΔ phenotypes may be because P. marneffei asexual development is stimulated by nutrient limitation whereas A. nidulans conidiates under nutrient sufficiency. This link with metabolism is supported by preliminary evidence that P. marneffei TupA contributes to carbon catabolite repression (CCR) of a subset of carbon metabolic activities (R. B. Todd, M. J. Hynes and A. Andrianopoulos, unpublished data). In S. cerevisiae Tup1p is required for repression of all activities subject to CCR (Williams and Trumbly, 1990; Treitel and Carlson, 1995) whereas the TupA homologues in A. nidulans and N. crassa play little or no role in CCR (Lee and Ebbole, 1998; Hicks et al., 2001). Penicillium marneffei tupA is required for repression of red pigment biosynthesis indicating a similar role to A. nidulans rcoA in regulation of biosynthesis of the secondary metabolite sterigmatocystin (Hicks et al., 2001).

Conservation of tupA sexual function

Although P. marneffei lacks a known sexual cycle the tupA gene functions in sexual development when introduced into A. nidulans. Therefore the cellular and physiological function of tupA has been conserved for sexual development. The N. crassa rco1 and S. cerevisiae tup1 mutants are female- and MATα-sterile, respectively (Williams and Trumbly, 1990; Yamashiro et al., 1996). Therefore the role of TupA homologues in sexual development appears widely conserved. The retention of TupA sexual function in the apparent absence of a sexual cycle in P. marneffei may be incidental, due to structural constraints imposed by conservation of unrelated TupA activities. Alternatively, P. marneffei may have a cryptic sexual cycle, a hypothesis supported by complementation of the sexual defects of the A. nidulans steA and stuA mutants by the P. marneffei STE12-like, stlA, and EFG1/PHD1-like, stuA, genes respectively (Borneman et al., 2001; 2002).

In contrast to C. albicans TUP1, which represses filamentation, P. marneffei TupA represses yeast morphogenesis

Penicillium marneffei TupA regulates yeast-filamentous dimorphism by repressing yeast morphogenesis during filamentous growth. Similarly, TUP1 from the dimorphic pathogen C. albicans also acts to regulate yeast-filamentous dimorphism but in this case represses filamentous growth (Braun and Johnson, 1997). Although at the morphological level the tupA and TUP1 genes in these two organisms act in opposing directions, mutations in these genes promote production of the pathogenic form. Filamentous growth is required for virulence in C. albicans (Lo et al., 1997) and the yeast form is pathogenic in P. marneffei.

How does TupA repress two distinct morphogenetic programmes?

The RFG1 HMG-domain and NRG1 zinc finger proteins have recently been implicated as DNA-binding repressors that recruit TUP1 for repression of filamentation in C. albicans (Braun et al., 2001; Khalaf and Zitomer, 2001; Murad et al., 2001). The manner in which TupA is harnessed for the repression of both asexual and yeast morphogenetic programmes during filamentous growth at 25°C is of considerable interest. A single DNA-binding transcriptional repressor may recruit TupA for repression of both developmental processes, or distinct DNA-binding proteins may recruit TupA for repression of yeast morphogenesis and asexual development.

Experimental procedures

Isolation and characterization of the Penicillium marneffei tupA gene

Standard molecular methods were as described (Sambrook et al., 1989). The tupA gene was cloned from a P. marneffei genomic library using a genomic 3.5 kbp EcoRI fragment of the A. nidulans rcoA gene (Hicks et al., 2001) as a heterologous probe in low stringency (40% formamide, 4 × SSPE (0.6 M sodium chloride, 0.04 M sodium dihydrogen phosphate, 4 mM disodium ethylenediamine tetra acetate (EDTA) pH 7.4), 1.0% SDS (sodium dodecyl sulphate), 5.0% blotto (10% skim milk powder, 2 mg ml−1 sodium azide), 100 µg ml−1 sonicated herring sperm (DNA) hybridization at 37°C. Filters were washed at 37°C twice for 30 min with 2 × SSC (0.3 M sodium chloride, 0.03 M tri-sodium citrate pH 7.0), 0.1% SDS. Sequence determination was undertaken at the Australian Genomic Research Facility (AGRF) and sequences were analysed using the Australian National Genomic Information Service (ANGIS). The tupA sequence data have been submitted to DDBJ/EMBL/GenBank under accession number AY082798. The P. marneffei tupA gene showed significant homology with the A. nidulans rcoA homologue (66%) and the six predicted introns were conserved in position.

Deletion of the P. marneffei tupA gene

A construct containing a ‘pyrG-blaster’ two-way selection cassette (Borneman et al., 2002) flanked by approximately 3.5 kbp XhoI-PstI 5′ and 0.7 kbp EcoICR1-EcoICR1 3′tupA sequence was introduced into a P. marneffei uridine/uracil auxotrophic strain (SPM4 niaD1 pyrG1) by DNA-mediated transformation (Borneman et al., 2001). Eighty PyrG+ transformants were selected and one transformant, which showed severe morphological defects at 25°C, was shown by genomic Southern hybridization (data not shown) to have an homologous gene replacement of the tupA gene. This tupA::pyrG allele lacks sequences +159 to +1011 relative to the ATG. The pyrG selectable marker used in gene replacement of tupA was deleted from the tupA::pyrG mutant by counter-selection for resistant sectors in the presence of 1 mg ml−1 5-fluoro orotic acid and 5 mM uracil and 5 mM uridine. The resulting tupAΔ pyrimidine auxotroph contained the deletion of tupA sequences as confirmed by Southern analysis (data not shown) and was phenotypically identical to the tupA::pyrG parent. The integrative pyrG selectable marker plasmid containing the P. marneffei tupA gene (pRT5313, BamHI approximately 2.2 kbp 5′ of the ATG to XhoI approximately 0.8 kbp 3′ of the stop codon) was introduced into the P. marneffei tupAΔpyrG1 mutant and selected directly for pyrimidine prototrophy (Oakley et al., 1987; Borneman et al., 2001). The reconstituted wild type (niaD1 pyrG1 tupAΔ[pyrG+tupA+]) was indistinguishable from the actual wild type (SPM3 niaD1) with respect to all phenotypes and there were no copy number or integration effects observed amongst transformants with different numbers of copies of tupA as determined by Southern analysis (data not shown).

Analysis of effects of deletion of tupA on growth

Penicillium marneffei growth testing and analysis used S. cerevisiae synthetic dextrose (SD) 2.0% glucose or synthetic (carbon-free) media with 10 mM ammonium as the sole nitrogen source, except where indicated. Wild type and the tupAΔ mutant were grown on a range of sole carbon sources (2.0% glucose, 0.1% glucose, 2.0% sucrose, 2.0% fructose, 2.0% maltose, 2.0% starch, 2.0% lactose, 2.0% glycerol, 2.0% ethanol, 2.0% raffinose and 50 mM acetate) with 10 mM ammonium as the sole nitrogen source, and on a range of sole nitrogen sources (10 mM ammonium, 10 mM glutamine, 10 mM glutamate, 10 mM alanine, 10 mM GABA, 10 mM arginine, 10 mM proline, 10 mM nitrite or 5 mM methyl-ammonium) with 2.0% glucose as the sole carbon source. The degree of filamentation at 37°C on different nitrogen sources was examined using a dissecting microscope.

Microscopy

Growth and fixation of P. marneffei cultures for microscopy and microscopic methods were as described (Borneman et al., 2000), except using SD as the growth media and co-staining with 50 µg ml−1 Hoechst 33258 and 500 µg ml−1 calcofluor white.

Complementation of the A. nidulans rcoAΔ mutation

Growth of A. nidulans strains used Aspergillus Nitrogen-free media (ANM) plus 10 mM nitrate as the sole nitrogen source or Aspergillus complete media (Cove, 1966). Complementation of the A. nidulans rcoAΔ mutation was undertaken using the protocol of Andrianopoulos and Hynes (1988) for transformation into the recipient strain MH10198 (biA1 pyrG89ΔrcoA::argB), derived from a meiotic outcross of TJH105.5 (biA1; argB2; methG1;ΔrcoA::argB) (Hicks et al., 2001). Transformants harbouring the tupA plasmid pRT5313 were directly selected for pyrimidine prototrophy. The rcoA cosmid pW03D10 (Hicks et al., 2001) was introduced by co-transformation with a plasmid containing the pyrG selectable marker (Oakley et al., 1987; Borneman et al., 2001).

Acknowledgements

We acknowledge technical assistance by S. Delimitrou, R. Croft and Q. Lang. We thank Drs N. P. Keller, J. M. Kelly and R. A. Lockington for provision of A. nidulans rcoA clones and the rcoAΔ mutant, and Drs M. A. Davis, C. S. Cobbett, D. G. Heckel and M. Trier Hansen for critical reading of the manuscript. This work was supported by the Australian Research Council (ARC) and Novozymes A/S.

Ancillary

Advertisement