Penicillium marneffei is a thermally dimorphic opportunistic human pathogen with a saprophytic filamentous hyphal form at 25°C and a pathogenic unicellular yeast form at 37°C. During infection. P. marneffei yeast cells exist intracellularly in macrophages. To cope with nutrient deprivation during the infection process, a number of pathogens employ the glyoxylate cycle to utilize fatty acids as carbon sources. The genes which constitute this pathway have been implicated in pathogenesis. To investigate acetate and fatty acid utilization, the acuD gene encoding a key glyoxylate cycle enzyme (isocitrate lyase) was cloned. The acuD gene is regulated by both carbon source and temperature in P. marneffei, being strongly induced at 37°C even in the presence of a repressing carbon source such as glucose. When introduced into the non-pathogenic monomorphic fungus Aspergillus nidulans, the P. marneffei acuD promoter only responds to carbon source. Similarly, when the A. nidulans acuD promoter is introduced into P. marneffei it only responds to carbon source suggesting that P. marneffei possesses both cis elements and trans-acting factors to control acuD by temperature. The Zn(II)2Cys6 DNA binding motif transcriptional activator FacB was cloned and is responsible for carbon source-, but not temperature-, dependent induction of acuD. The expression of acuD at 37°C is induced by AbaA, a key regulator of morphogenesis in P. marneffei, but deletion of abaA does not completely eliminate temperature-dependent induction, suggesting that acuD and the glyoxylate cycle are regulated by a complex network of factors in P. marneffei which may contribute to its pathogenicity.
Genes encoding enzymes of the glyoxylate cycle have been implicated in both fungal and bacterial pathogenesis (McKinney et al., 2000; Lorenz and Fink, 2001; Munoz-Elias and McKinney, 2005). The glyoxylate cycle involves two critical steps catalysed by the enzymes isocitrate lyase and malate synthase, which bypass the two decarboxylation steps of the TCA cycle. Isocitrate lyase hydrolyses isocitrate (C6) to succinate (C4) and glyoxylate (C2) and subsequent condensation of glyoxylate and acetyl-CoA (C2) by malate synthase produces malate (C4), a TCA cycle intermediate. Malate is further oxidized to oxaloacetate (C4) and then converted into citrate (C6) by the condensation of another molecule of acetyl-CoA. C4 intermediates are required for anabolic processes. For example, oxaloacetate is employed during gluconeogenesis to generate glucose, which is necessary for the incorporation of carbon into the cellular macromolecules. Malate, oxaloacetate, citrate and isocitrate are intermediates in the glyoxylate and TCA cycles, so the glyoxylate cycle allows two carbon compounds to replenish the TCA cycle. Therefore, the glyoxylate cycle is required for growth on gluconeogenic carbon sources, such as acetate, and it is usually activated under conditions of nutrient deprivation. Such conditions are believed to occur inside macrophages and pose particular challenges to intracellular pathogens (Lorenz and Fink, 2002).
In Candida albicans, the isocitrate lyase gene icl1 is upregulated during growth inside macrophages. Strains deficient in this gene are less virulent than wild-type strains in a mouse pathogenicity model (Lorenz and Fink, 2001). Recently it has been reported that the glyoxylate cycle genes in C. albicans are repressed by the physiological concentrations of glucose found in the bloodstream. However, these genes are induced upon phagocytosis by macrophages or neutrophils, emphasizing the importance of carbon metabolism during pathogenesis (Barelle et al., 2006). Plant pathogens have also been shown to require the glyoxylate cycle for pathogenicity (Idnurm and Howlett, 2002; Wang et al., 2003; Solomon et al., 2004). Despite the importance of the glyoxylate cycle in pathogenesis, the molecular mechanisms which regulate the isocitrate lyase encoding genes in pathogens are unknown.
In the non-pathogenic fungus Aspergillus nidulans, the isocitrate lyase encoding acuD gene is expressed during growth on gluconeogenic compounds such as acetate or fatty acids. Acetate induction of acuD is FacB-dependent (Todd et al., 1997). FacB is a transcriptional activator with a zinc binuclear cluster (Zn(II)2Cys6) DNA binding domain. FacB binds two dissimilar DNA sequences found in the 5′ regions of acetate-regulated genes such as acuD (Todd et al., 1998). Besides acetate regulation, acuD is also regulated by fatty acids via FarA (long-chain fatty acids) and ScfA and FarB (short-chain fatty acids). All these fatty acid transcriptional regulators also contain a Zn(II)2Cys6 DNA binding domain and work independently of FacB (Hynes et al., 2006). In Saccharomyces cerevisiae the glyoxylate cycle genes are activated by the transcriptional activators Cat8 and Sip4, which also contain a Zn(II)2Cys6 DNA binding domain, in response to glucose limitation. Cat8 is the main activator of the expression of the glyoxylate genes and is also required for the expression of SIP4, which plays a minor role in the upregulation of the glyoxylate genes (Schuller, 2003). In addition to the acetate and fatty acid induction of the glyoxylate cycle genes, expression is also subject to glucose-mediated repression. In A. nidulans, this is achieved in part by the action of the carbon catabolite transcriptional repressor CreA (Bowyer et al., 1994; De Lucas et al., 1994a).
Penicillium marneffei is an opportunistic fungal pathogen of humans. It is the only dimorphic species within the Penicillium and related genera. At 25°C, P. marneffei grows as multicellular hyphal cells, which resemble other Penicillium species while at 37°C it grows as a unicellular yeast, oval in shape, and divides by fission. Yeast cells are the pathogenic form. In addition, at 25°C in response to specific environmental cues it can undergo asexual development (conidiation) to produce conidia (spores), which are presumed to be the infective form (Andrianopoulos, 2002). In the monomorphic fungus A. nidulans, asexual development is controlled by a cascade of transcriptional regulators, encompassing brlA, which encodes a C2H2 zinc finger protein (Adams et al., 1988), abaA, which encodes an ATTS/TEA protein (Andrianopoulos and Timberlake, 1991; Burglin, 1991) and wetA, which encodes a spore-specific protein (Marshall and Timberlake, 1991). BrlA is activated by conidiation inducing conditions, and then in turn activates abaA expression. brlA mutants strains fail to produce most of the conidiophore cell types and are only able to differentiate conidiophore stalks (Clutterbuck, 1969). AbaA activates the expression of wetA and also feedback regulates the expression of itself and brlA (for review see Adams et al., 1998). abaA mutant strains produce aberrant conidiophores lacking conidia, instead producing reiterated phialide-like cells which bud by acropetal division (Sewall et al., 1990). In P. marneffei, asexual development is similarly regulated but AbaA is also required to couple nuclear and cellular division during dimorphic switching (Borneman et al., 2000; A.R. Borneman, M.J. Hynes and A. Andrianopoulos, unpublished results). The expression of abaA is upregulated 10-fold during dimorphic switching and 30-fold during conidiation (Borneman et al., 2000).
In the present work, we show that the P. marneffei acuD gene is required for growth on gluconeogenic carbon sources such as acetate and fatty acids, is strongly induced by acetate and is dependent on the FacB transcriptional activator for acetate induction. More importantly, P. marneffei acuD is also independently regulated by the dimorphic switching developmental program and part of this control is through the AbaA transcriptional activator. The developmental regulation of P. marneffei acuD has both cis- and trans-acting elements which are not present in the A. nidulans acuD gene or in A. nidulans, showing a unique evolutionary path for acetate and fatty acid regulation in this dimorphic pathogen.
The P. marneffei acuD homologue
P. marneffei acuD was originally identified in a differential display expression screen as a gene highly expressed in yeast cells (C.R. Cooper Jr, Youngston University, pers. comm.; accession number AF373018). The open reading frame spans 1860 bp and consists of five exons and four introns (Fig. 1A). The predicted protein is 540 aa long and highly similar to other fungal isocitrate lyases. It shows 85% identity to the Aspergillus fumigatus predicted isocitrate lyase (Q6T267), 82% to AcuD of A. nidulans (P28298), 81% to Icl1 of Coccidioides immitis (Q96TP5), and 64% to Icl1 of C. albicans (Q9P8Q7). Although predicted to be localized to glyoxysomes (peroxisomes) these isocitrate lyases, included P. marneffei AcuD, lack obvious peroxisomal targeting sequences including the carboxyl terminal PTS1 sequence.
Disruption of acuD function
An acuD deletion strain of P. marneffei was generated by transformation of SPM4 (niaD1 pyrG1) with the deletion construct in pDAP4, in which the acuD open reading frame was replaced with the A. nidulans pyrG selectable marker (Fig. 1A). Transformants were selected for uracil prototrophy and then screened for the absence of growth on minimal media containing acetate as the sole carbon source. From 22 pyrG+ transformants, 10 were unable to grow on acetate but showed wild-type growth on glucose as a carbon source. Examination of the deletion arrangement in these transformants by Southern blot hybridization showed that in one transformant the acuD gene was deleted and there were no ectopic integration events of the deletion construct. The remaining transformants either had the wild-type copy of the gene and/or ectopic integrations of the deletion construct (data not shown).
Additional growth tests revealed that the deletion mutant was also unable to grow on butyrate (4C) and Tween 80 (18C) as sole carbon sources at 25°C (Fig. 1B). When the ΔacuD strain was grown on minimal media containing both glucose and acetate at 25°C, growth was partially impaired (data not shown). At 37°C, the deletion mutant failed to grow on SA (YNB medium containing acetate as the sole carbon source) or SD (YNB medium containing glucose as the sole carbon source) (Fig. 2B). The addition of uracil to the culture medium restored growth of the acuD deletion strain on SD but not on SA at 37°C (Fig. 3). Addition of uracil also restored growth of the acuD deletion strain on minimal medium containing glucose and acetate at 25°C (data not shown). One hypothesis for these observations is that conditions which induce expression from the acuD promoter lead to interference with expression of the pyrG selectable marker in the acuD deletion construct and therefore a failure to complement the uracil auxotrophy. To test this hypothesis, an ΔacuD::pyrG– strain was isolated by selection on 5′-fluoroorotic acid. The resultant strain was transformed with pALX223 (pyrG+) to isolate transformants with randomly integrated copies of pyrG in the genome which would be unlinked to the acuD promoter and not be subject to transcriptional interference from the expression arising from this promoter. The resultant transformants [ΔacuD::pyrG, pyrG+ (pALX223)] were able to grow on SD (containing glucose) but not on SA (containing acetate) at 37°C (Fig. 2B), showing that a uracil auxotrophy occurring in the ΔacuD::pyrG+ was responsible for the absence of growth on glucose at 37°C. Introduction of the acuD+ gene (pDAP70) complemented the growth defect of the deletion mutant on acetate at 25°C and 37°C (Fig. 2).
A number of additional experiments point to a mechanism. First, the acuD deletion mutant also showed poorer growth on medium containing glucose and acetate than on glucose alone at 25°C and this was remediated by uracil. Second, a number of transformants containing ectopic copies of the acuD deletion construct failed to grow on medium containing acetate as the sole carbon source but lacking uracil despite having an intact acuD locus. Coupled with the experiment where additional ectopic copies of pyrG complemented the uracil auxotrophy, these observations are consistent with the idea that these strains are conditional uracil (pyrG–) auxotrophs where the restrictive condition is dependent on acuD promoter activity, which in turn interferes with pyrG expression. One possible mechanism given the inverted orientation of the acuD and pyrG promoters in the deletion construct is that acuD promoter activity produces a mRNA complementary to the pyrG transcript and that this complementary mRNA acts in trans by RNAi or antisense mechanisms to interfere with the pyrG selectable marker (Agrawal et al., 2003).
The acuD deletion construct removes, in addition to the entire coding region, the first 543 bp of acuD promoter (relative to the ATG) which encompasses four of the five predicted FacB binding sites and the transcription initiation site (Figs 1 and 6A). Therefore, acuD promoter sequences upstream of −543 are sufficient for both acetate induction and cell type (temperature) induction. Exactly which sequences are responsible for this regulation remains to be determined.
acuD expression is induced by acetate and temperature
Expression of acuD is regulated by carbon source in A. nidulans and other fungi like S. cerevisiae (Todd et al., 1997; Schuller, 2003). To analyse acuD expression in P. marneffei, we performed Northern blot analysis using RNA isolated from cells grown on SYBD (glucose) or SYBA (acetate) at both 25°C and 37°C (Fig. 3). As for A. nidulans acuD at 25°C, expression of P. marneffei acuD is dependent on the carbon source, being induced on acetate while having no detectable expression on glucose at 25°C. At 37°C acuD was expressed in the presence of either glucose or acetate and the level of expression was substantially higher than at 25°C. The addition of acetate to the culture medium at 37°C slightly increased the expression of acuD over the levels on glucose. This clearly showed a different pattern of acuD regulation between A. nidulans and P. marneffei. Northern blot analysis also revealed a second but minor transcript (Fig. 3).
Comparison of acuD regulation in A. nidulans and P. marneffei
To analyse whether the differing regulation of acuD in A. nidulans and P. marneffei is regulator dependent (trans effect) or promoter dependent (cis effect), we placed the red fluorescent protein (monomeric) mrfp reporter gene under the control of a 1.4 kb DNA fragment containing the acuD promoter from each organism, denoted PmacuD(p) and AnacuD(p). These constructs were used for the transformation of A. nidulans (targeted to the pyrG locus for PmacuD(p)::mrfp and to the pyrG or acuD locus for AnacuD(p)::mrfp) and P. marneffei (targeted to the areA locus). Of the pyrG+ transformants, two to three independently isolated strains were grown on glass slides covered with a thin layer of SD (containing glucose) or SA (containing acetate) media, and expression arising from the acuD promoter was detected by fluorescence microscopy (Figs 4 and 5). In A. nidulans, induction via either acuD promoter occurred exclusively in the presence of acetate at both 25°C and 37°C (Fig. 4). Neither of the promoters was subject to temperature regulation in A. nidulans. In addition, a faint signal was detected in the spores of the strains containing the PmacuD(p)::mrfp construct but not the AnacuD(p)::mrfp construct, suggesting that P. marneffei acuD may be expressed during the early stages of germination.
In P. marneffei at 25°C, acuD expression is carbon source dependent and both PmacuD(p)::mrfp and AnacuD(p)::mrfp constructs were induced in the presence of acetate while no expression could be detected on glucose (Fig. 5A). As for PmacuD(p)::mrfp in A. nidulans, a faint signal could be detected in the spores of strains bearing this promoter construct. In contrast to the expression pattern of these constructs in A. nidulans, the PmacuD(p)::mrfp construct was strongly expressed at 37°C in the presence of glucose while the AnacuD(p)::mrfp construct was barely detectable (Fig. 5A). Expression of both constructs was higher on acetate than on glucose, corroborating the results obtained by Northern blot analysis for P. marneffei acuD (Fig. 3) and expression at 37°C was higher than at 25°C, regardless of the carbon source.
Comparison of the acuD promoter of P. marneffei and that in the two closest relatives A. nidulans and A. fumigatus
From the expression studies, it was hypothesized that there are two levels of regulation of acuD in P. marneffei which are likely to operate independently. We looked for predicted transcription factor binding sites in the promoter of P. marneffei acuD and then compared these sites with those found in the acuD promoters of A. nidulans and A. fumigatus (Fig. 6 and Table 1). The acuD promoter of A. nidulans has eight potential FacB binding sites, of which three sites are bound by FacB in vitro (Todd et al., 1998). In P. marneffei, there are five potential FacB binding sites, three of which are overlapping and contiguous. Within the three overlapping FacB binding sites there is also a predicted binding site for the glucose-dependent repressor CreA (Kulmburg et al., 1993) (Fig. 6). In addition, the A. nidulans acuD promoter has five CreA binding sites, one BrlA binding site (Chang and Timberlake, 1993), three AbaA binding sites (Andrianopoulos and Timberlake, 1994) and three CCGAGG motifs (Hynes et al., 2006) while the P. marneffei acuD promoter has two CreA, two AbaA binding sites and one CCGAGG motif but no BrlA binding sites. The acuD promoter of A. fumigatus was more similar to the A. nidulans than P. marneffei containing three CreA binding sites, four AbaA binding sites, nine FacB binding sites, three CCGAGG motifs and no BrlA binding sites (Table 1). In addition, promoter analysis using match software (http://www.gene-regulation.com/pub/programs.html#match) with a cut-off set to minimize false positives revealed two CCAAT boxes inverted with respect to each other, separated by 100 bp and surrounding the three overlapping FacB binding sites and two overlapping CCAAT displacement protein (CDP) binding sites in the P. marneffei acuD promoter but not in the Aspergillus acuD promoters. In higher eukaryotes, CDP is believed to act as a repressor of gene transcription (Barberis et al., 1987). However, no homologues of CDP have been found in fungi to date (Gillingham et al., 2002).
Table 1. Summary of predicted transcriptional regulator binding sites in acuD promoters.
acuD expression is regulated by the acetate-dependent regulator FacB in P. marneffei
FacB is a transcriptional activator involved in the acetate-dependent induction of acuD in A. nidulans (Katz and Hynes, 1989). To test whether acuD expression in P. marneffei was dependent on FacB, Northern blot analysis was performed on a facB deletion mutant strain (D. Cánovas and A. Andrianopoulos, unpublished) grown on SD (glucose) or SA (acetate) media (Fig. 7A). The level of acuD expression on both YNB-based media (SD and SA) at 25°C was very poor (data not shown). Induction by acetate was reduced, but not abolished, in the facB deletion mutant at 37°C. Therefore, facB mediates acetate-inducible expression of acuD, but other factor(s) also control the expression of acuD in response to acetate.
acuD expression is regulated by the asexual development transcriptional cascade in P. marneffei
acuD expression is regulated by both carbon source and temperature. FacB is partly responsible for controlling acuD expression in response to acetate but not temperature. The abaA gene is a key regulator of morphogenesis in response to the temperature shift to 37°C (Borneman et al., 2000). AbaA also regulates the asexual development program at 25°C where it is activated by, and feedback activates, the brlA regulatory gene. Given that there are two predicted AbaA binding sites in the promoter of P. marneffei acuD (Fig. 6A), the role of AbaA in acuD expression was examined. Wild-type, ΔabaA and ΔbrlA mutant strains of P. marneffei were grown under the standard growth conditions which promote hyphal growth and asexual development [A. nidulans medium (ANM) at 25°C] or dimorphic switching [brain–heart infusion (BHI) at 37°C], and acuD expression was analysed by Northern blot (Fig. 7B). The expression of acuD at 37°C was severely reduced in the abaA deletion mutant but unaffected in the brlA deletion mutant. At 25°C during asexual development, acuD was expressed at low levels in the wild-type strain but absent in both the ΔabaA or ΔbrlA mutants. Given that AbaA strongly induces the expression of acuD at 37°C and directs the expression of genes during conidiation at 25°C, it is not clear why acuD expression is low during conidiation. One possibility is that acuD is repressed during asexual development, possibly via brlA, thereby overriding AbaA activation. To test this hypothesis, a strain of P. marneffei bearing A. nidulans brlA under the control of the inducible xylP promoter was grown at 37°C on BHI under repressing (glucose) and inducing (xylose) conditions. Northern blot analysis revealed that overexpression of brlA dramatically reduced the expression of acuD (Fig. 7C). There are no BrlA binding sites in the promoter of P. marneffei acuD (Fig. 6A), suggesting that BrlA-mediated repression is likely to be indirect.
AbaA is required for controlling acuD expression in a temperature-dependent manner. To examine whether AbaA is also involved in the acetate-dependent induction of acuD, the abaA mutant strain was grown in SD and SA. Northern blot analysis showed that the level of acuD expression on acetate was significantly lower in the ΔabaA mutant compared with the wild type suggesting that acetate induction is partially mediated by AbaA at 37°C (Fig. 7A). In addition, the level of acuD expression on glucose was higher than the wild type.
Disruption of abaA does not prevent growth on acetate
The expression of acuD is significantly reduced at 37°C in the abaA deletion mutant. Growth tests revealed that the abaA mutant can grow on acetate as sole carbon source both at 25°C and 37°C (Fig. 8). This demonstrated that temperature and carbon source are two independent levels of regulation of acuD expression and that acuD is functional in an abaA deletion mutant even during the yeast growth phase.
The glyoxylate cycle is necessary for the incorporation of C2 carbon compounds into cellular anabolic processes by bypassing the two decarboxylation steps in the TCA cycle, thereby permitting the generation of glucose by gluconeogenesis. In P. marneffei, the glyoxylate cycle is required for growth on acetate and fatty acids, as demonstrated by deletion of the acuD gene which encodes isocitrate lyase, the first enzyme in the bypass. In contrast to other characterized systems, regulation of P. marneffei acuD expression occurs in response to both carbon and cell type-specific signals, and heterologous expression experiments show that the latter requires P. marneffei-specific cis and trans elements.
Expression of acuD is subject to carbon regulation in P. marneffei, as in most other fungi where it has been studied (Todd et al., 1998; Maeting et al., 1999; Rude et al., 2002; Wang et al., 2003; Barelle et al., 2006). In contrast to other fungal systems, the P. marneffei acuD gene also shows strong cell type regulation which is independent of the carbon regulation, such that the presence of glucose does not affect the expression levels, suggesting that cell type regulation overrides glucose repression. Heterologous expression of the acuD promoters from A. nidulans and P. marneffei in P. marneffei and A. nidulans, respectively, shows that the cell type (temperature) regulation is restricted to P. marneffei harbouring its native acuD promoter. This suggests the existence of unique elements in the promoter of P. marneffei acuD, which are controlled by specific regulator(s) present in P. marneffei but not in A. nidulans. This regulatory system could also control the expression of other genes including those involved in pathogenesis. Comparison of the promoter sequence in both organisms revealed a number of putative binding sites for transcriptional regulators common to both organisms. Besides positional and number differences for the various transcription factor binding sites, the other difference is the presence of two CCAAT boxes which are binding sites for a wide domain, multimeric transcriptional activator complex found in eukaryotes (Brakhage et al., 1999). This CCAAT binding complex has been shown to contribute to setting high basal levels of expression as well as in facilitating activation, and in one case repression, by other transcriptional regulators (Aramayo and Timberlake, 1993; Brakhage et al., 1999; Zhang et al., 1999; Schuller, 2003; McNabb and Pinto, 2005). The fact that this complex has been implicated in carbon metabolism, nitrogen metabolism, development and virulence in fungi suggests that the CCAAT binding factor may contribute to the expression of gluconeogenic genes (i.e. acuD) that may be upregulated during infection in P. marneffei.
acuD is independently regulated by acetate and temperature in P. marneffei and a combination of both inducing conditions resulted in the highest level of expression, suggesting additive relationship. A minor transcript was also evident but it is currently unclear whether this is a result of differential splicing or differential transcriptional initiation and/or termination. Both A. nidulans acuD and S. cerevisiae ICL1 genes are strictly regulated by carbon source but not temperature (De Lucas et al., 1994b; Ordiz et al., 1995). At the transcriptional level, A. nidulans acuD and S. cerevisiae ICL1 expression is upregulated in the presence of acetate and repressed by glucose (Fernandez et al., 1993; Bowyer et al., 1994; De Lucas et al., 1994a). Glucose repression in S. cerevisiae is mediated by Mig1, a C2H2 zinc finger regulator; however, glucose repression of ICL1 is indirect even though the promoter contains a Mig1 binding site. Glucose derepression is mediated by Snf1, which inactivates Mig1 and this, in turn, allows the expression of CAT8, which activates the expression of SIP4. Both Cat8 and Sip4 act in combination to induce the expression of gluconeogenic genes, including ICL1, by binding to a carbon source responsive element, YCCRTTNRNCCG (Schuller, 2003). In A. nidulans, glucose repression is mediated by CreA, which binds to the consensus sequence SYGGRG (Kulmburg et al., 1993; Cubero and Scazzocchio, 1994), preventing the transcription of target genes (Mathieu and Felenbok, 1994). However, in P. marneffei, acuD is also expressed in the presence of glucose at 37°C, therefore temperature induction overrides any putative glucose repression control of acuD expression. It is noteworthy that two potential CreA binding sites are present in the acuD promoter, one of which is located directly on the DNA binding region containing three overlapping FacB sites (Fig. 6).
In A. nidulans acetate induction of acuD expression is dependent on the transcriptional regulator FacB and three FacB binding sites have been identified, consisting of two dissimilar DNA sequences (De Lucas et al., 1994a; Todd et al., 1998). Examination of the P. marneffei acuD promoter sequence and comparison with the A. nidulans and A. fumigatus promoters revealed a number of conserved binding sites for FacB. There are five additional potential sites in A. nidulans, nine in A. fumigatus and five in P. marneffei. The position of the sites with respect to the ATG are not conserved in the three species. Interestingly, the two pathogenic species contain overlapping FacB binding sites (two in A. fumigatus and three in P. marneffei); however, the biological significance of this is unclear. In the P. marneffei facB deletion strain, acetate induction of acuD is substantially reduced at 37°C. Given that acuD expression at 37°C is not completely abolished, this shows that FacB is not required for temperature-dependent induction. Furthermore, the higher levels of acuD mRNA at 37°C on acetate compared with glucose indicates that this is either the result of glucose derepression leading to high basal levels of acuD expression or there is an additional weak acetate induction mechanism involving an unidentified transcriptional activator.
The P. marneffei abaA gene controls both differentiation of the sporogenic phialide cell type during asexual development at 25°C and the coupling of nuclear division with cell division during yeast cell morphogenesis at 37°C (Borneman et al., 2000). The downstream structural genes which are regulated by AbaA are poorly understood. There are two potential AbaA binding sites in the P. marneffei acuD promoter and four potential binding sites in the promoters of the A. nidulans and A. fumigatus acuD genes. Deletion of the P. marneffei abaA gene leads to an almost complete loss of acuD expression at 37°C showing that this is the prime morphological regulator of acuD. In addition, loss of abaA also leads to a defect in acetate induction of acuD at 37°C suggesting that either AbaA activates the facB gene which in turn activates acuD expression or that FacB and AbaA are co-dependent for acetate induction of acuD at 37°C. Studies in other systems have shown that pathogens reprogram their metabolic patterns when growing inside the host (McKinney et al., 2000; Lorenz and Fink, 2001) and this is also likely to be true for P. marneffei, possibly via the developmental regulator AbaA. Interestingly, there are three binding sites for the AbaA homologue Tec1 (Andrianopoulos and Timberlake, 1994; Madhani and Fink, 1997) in the ICL1 promoter of C. albicans but none in the gene from the non-pathogenic S. cerevisiae. Tec1 is involved in filamentation and pathogenesis in C. albicans (Schweizer et al., 2000) but only in pseudohyphal growth in S. cerevisiae (Lengeler et al., 2000). Whether this is a general mechanism occurring in other dimorphic fungi remains to be discovered. A screen for phase-specific expressed sequence tags in the thermally dimorphic fungal pathogen Paracoccidioides braziliensis also identified an acuD orthologue and it was shown to be differentially expressed during the transition from yeast to mycelium (Goldman et al., 2003). The fact that AbaA regulates a gene involved in carbon metabolism at 37°C, and implicated in pathogenicity in some other fungal pathogens, implies that it may be important for the regulation of other pathogenic determinants.
The acuD gene is expressed at low levels during asexual development and this expression is absent in the abaA deletion mutant. During asexual development but not yeast cell morphogenesis, the abaA gene is activated by the product of the brlA gene (A.R. Borneman et al., unpublished). Consistent with these observations, loss of brlA has no effect on acuD expression at 37°C but does lead to a loss of expression during asexual development at 25°C. The high level of AbaA-dependent acuD expression in yeast cells is in stark contrast to the low levels during asexual development. One possible explanation for these results is that BrlA is acting as both an activator and repressor of gene transcription. In support of this, ectopic overexpression of BrlA at 37°C results in a dramatic downregulation of acuD expression. Such a mechanism may be one way in which the activity of regulatory factors such as AbaA, which participates in two distinct developmental programs, are cell type limited in their targets. It is unknown whether BrlA-mediated repression occurs by direct binding to the promoter (no BrlA binding site was found in P. marneffei acuD promoter), by activating the transcription of some other repressing regulator(s) or by some other unknown mechanism. Alternatively, these results may be explained by the loss of the cell types in which acuD is expressed.
The second step in the glyoxylate bypass is catalysed by malate synthase and it would be expected that the gene encoding this enzyme would be co-ordinately regulated with acuD, The P. marneffei acuE gene encodes malate synthase and this gene is regulated in an identical manner to acuD responding to both temperature and acetate induction, showing AbaA-dependent expression and showing reduced expression when brlA is overexpressed (Fig. S1). Therefore, the described regulatory pattern is not unique to acuD, as the acuE gene exhibits identical regulation, showing that the whole glyoxylate bypass is co-ordinately regulated by this mechanism.
In A. nidulans a suppressive subtraction hybridization screen for genes expressed in conidia identified a number of genes including acuD and it was shown that expression of acuD is significantly reduced in strains carrying mutations in the brlA and abaA genes (Osherov et al., 2002). The acuD gene was also previously identified in a screen for genes expressed during asexual development (termed CAN genes for conidiation in A.nidulans) (Boylan et al., 1987; Stringer, 1994). The role of acuD expression during asexual development is not clear but it can be hypothesized that the glyoxylate cycle is used to derive energy in developing conidiophores which grow away from the growth substrate and thus may be deprived of nutrients. In P. marneffei, the acuD deletion strain does not show any signs of delayed germination at either 25°C or 37°C suggesting that it is not required for conidia to exit from dormancy and initiate growth (data not shown), further implicating its role in conidiogenesis rather than germination.
The pathology of P. marneffei is poorly understood. It is assumed that infection occurs via inhalation of conidia, which are subsequently phagocytosed by macrophages (Vanittanakom et al., 2006). Macrophages are considered an environment poor in nutrients. Consequently, genes involved in the glyoxylate cycle in other fungi, such as C. albicans, are upregulated when the fungal cells are inside the macrophages (Lorenz and Fink, 2001; Barelle et al., 2006). In C. albicans, concentrations of glucose as low as those found in the bloodstream are enough to downregulate glyoxylate cycle genes. Therefore, C. albicans reprograms its carbon metabolism from an activated glyoxylate cycle during early stages of infection inside the macrophage to the activation of glycolysis during the progression of systemic disease (Barelle et al., 2006). Although P. marneffei primarily produces pulmonary infections, it can also be disseminated to other organs, so that yeast cells are found both in intracellular and extracellular environments within the host (Vanittanakom et al., 2006). As the acuD gene is tightly regulated by both temperature and carbon source, it would be expected that in contrast to C. albicans, the glyoxylate cycle remains upregulated during the entire P. marneffei infection cycle. These regulatory differences may be a consequence of the commensal versus opportunistic life styles of these two pathogens.
Induction of the glyoxylate cycle in pathogenic microorganisms is presumed to be a response to nutrient deprivation inside the macrophages (Lorenz and Fink, 2002); however, the regulatory mechanisms controlling the expression of the glyoxylate cycle genes in pathogens had not been addressed previously. This study shows that the genes belonging to the glyoxylate bypass are highly expressed during the pathogenic yeast form in P. marneffei. The expression of these genes is tightly controlled by the acetate-responsive regulator FacB and the developmental regulator AbaA (Fig. 9). These results provide evidence for a link between carbon metabolism, development and pathogenicity in the dimorphic fungal pathogen P. marneffei.
Strains, media and growth conditions
Fungal strains used in this study are listed in Table 2. A. nidulans strains were regularly grown on Aspergillus defined medium (ANM) (Cove, 1966). P. marneffei strains were grown on either ANM or yeast synthetic medium (YNB, Yeast Nitrogen Base) (Ausubel et al., 1989) containing the appropriate carbon and nitrogen sources, or the complex BHI (Oxoid) or SYB (2% yeast extract, 1% protease peptone) medium. (NH4)2SO4 at a final concentration of 10 mM was used as the nitrogen source in the defined media. Glucose (1–2%), acetate (50 mM), butyrate (20 mM) or Tween 80 (0.2%) were used as carbon sources in the media. Auxotrophic supplements were added as required. Escherichia coli TOP10 (Invitrogen Life Technologies) and JM110 dam– were used for the propagation of plasmid DNA.
Genomic DNA was isolated as previously described (Borneman et al., 2001). RNA was prepared by using the FastRNA red kit (BIO101). RNA samples were separated on 1.2% agarose formaldehyde denaturing gels, blotted onto Amersham Hybond N + membranes and analysed by Northern blot hybridization (Sambrook et al., 1989). The acuD and acuE probes for Northern blotting were generated by PCR using the primers pm-icl1upper1 (5′-GTGGTACACTCAAGATTGAGTATC-3′) and pm-icl1lower1 (5′-GTTCTAGAATAACGAACATTAGAG-3′), and pmacuEnort43 (5′-GAACTGTCAACAACGAACATC-3′) and pmacuEnort592 (5′-GACTGTGAAATGCTTCTCCTC-3′) respectively. DNA-mediated transformation of P. marneffei and A. nidulans has been previously described (Borneman et al., 2001; Nayak et al., 2005).
Cloning and plasmid construction
The acuD (ICL1) DNA sequence of P. marneffei is deposited in the GenBank database (accession number AF373018). acuD was amplified from genomic DNA by using primers icl1for1 (5′-GATCGTAGAGACTAGATGACC-3′) and icl1rev1 (5′-ACCTGTATCGATTGACAATTC-3′) and cloned into pGEM®-T Easy vector (Promega). The ΔacuD deletion construct plasmid (pDAP4) was generated by replacing an XbaI-EcoRV fragment spanning the entire acuD coding region with the A. nidulans pyrG selectable marker (Fig. 1B).
PmacuD(p)::mrfp and AnacuD(p)::mrfp were constructed by assembling two PCR products. Both contain 1.4 kb of the promoter region immediately adjacent to the ATG of each gene. The promoter of P. marneffei acuD was amplified using the primers acudfor5′xho1 (5′-ACTCGAGCACCCCATACCAACAATC-3′) and acudrev5′mrfp (5′-GTCCTCGGAGGAGGCCATTGTGAATTTT-3′). The A. nidulans acuD promoter was amplified by using primers AnacuDpup (5′-ATCTTCTTCCCATGCAACGG-3′) and AnacuDplowRFP (5′-GTCCTCGGAGGAGGCCATGATGGCAGTATTCAGC-3′). The mrfp gene was amplified from plasmid pMT-mRFP1 (Toews et al., 2004) using the primers mRFPforw (5′-CATATGGCCTCCTCCGAGGACG-3′) and mRFPrev (5′-GGCGCCGGTGGAGTGGCG-3′). The PCR fragment containing the P. marneffei promoter fused to the mrfp reporter gene was cloned into pGEM®-T Easy to produce the PmacuD(p)::mrfp fusion vector (pDAP28). pDAP75 was generated by inserting the 3′ region of the P. marneffei areA gene into pALX223 to allow targeting into the areA locus of P. marneffei (H. Smith and A. Andrianopoulos, unpubl. results). pDAP28 was then digested with NotI and the PmacuD(p)::mrfp fragment cloned into pDAP75 to give the final plasmid pDAP77 (pyrG+, areA*, PmacuD(p)::mrfp). The PCR fragment containing the A. nidulans acuD promoter fused to the mrfp reporter gene was cloned into pALX223 digested with EcoRV (pDAP71). areA* was then cloned into the EcoRI site giving the final plasmid pDAP76 (pyrG+, areA*, AnacuD(p)::mrfp).
For complementation of the ΔacuD deletion mutant, the acuD gene was generated by PCR using primers icl1for1 (5′-GATCGTAGAGACTAGATGACC-3′) and icl1rev1 (5′-ACCTGTATCGATTGACAATTC-3′) and cloned into pALX223 digested with EcoRV to yield pDAP70.
P. marneffei and A. nidulans strains were grown on slides covered with a thin layer of solid medium, inoculated with conidia from the appropriate strains and incubated at the indicated temperature, except for P. marneffei cultures at 37°C where conidia were allowed to germinate on liquid medium and then visualized under the microscope. All strains were grown on yeast nitrogen base containing 10 mM (NH4)2SO4 as a nitrogen source and either 2% glucose (SD) or 50 mM sodium acetate (SA) as a sole carbon source at the indicated temperature. Slides were visualized on a Reichart Jung Polyvar II microscope using differential interference contrast (DIC) or epifluorescence optics for RFP. Images were captured using a SPOT CCD camera (Diagnostic Instruments) and processed in Adobe Photoshop™ 7.0.
D.C. thanks all members of the Andrianopoulos, Hynes and Davis groups for helpful discussions. Dr Richard Todd, Dr Reinhard Fischer and Dr Berl Oakley kindly provided a partial clone of the P. marneffei facB gene, the pMT-mRFP1 plasmid and the A. nidulans TNO2A3 strain respectively. D.C. was supported by a Marie Curie OIF. This work was supported by grants from the Australian Research Council, National Health and Medical Research Council and Howard Hughes Medical Institute.