NRG1 represses yeast–hypha morphogenesis and hypha‐specific gene expression in Candida albicans

A.Munir A. Murad, Ping Leng, Melissa Straffon, Jill Wishart, Susan Macaskill, Donna MacCallum, Norbert Schnell, Driss Talibi, Daniel Marechal, Fredj Tekaia, Christophe d'Enfert, Claude Gaillardin, Frank C. Odds, Alistair J.P. Brown

Author Affiliations

  1. A.Munir A. Murad2,
  2. Ping Leng3,
  3. Melissa Straffon1,
  4. Jill Wishart1,
  5. Susan Macaskill1,
  6. Donna MacCallum1,
  7. Norbert Schnell4,
  8. Driss Talibi5,
  9. Daniel Marechal5,
  10. Fredj Tekaia6,
  11. Christophe d'Enfert7,
  12. Claude Gaillardin8,
  13. Frank C. Odds1 and
  14. Alistair J.P. Brown*,1
  1. 1 Molecular and Cell Biology, Institute of Medical Sciences, University of Aberdeen, Aberdeen, AB25 2ZD, UK
  2. 2 Present address: Pusat Pengajian Biosains Molekul dan Bioteknologi, Universiti Kebangsaan Malaysia, 43600, Bangi, Selangor, Malaysia
  3. 3 Present address: Division of Cellular and Molecular Biology, Ontario Cancer Institute, University of Toronto, 610 University Avenue, Toronto, Ontario, M5G 2M9, Canada
  4. 4 AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA, 02451, USA
  5. 5 Eurogentec, Parc Scientifique du Sart Tilman, 4102‐Seraing, Belgium
  6. 6 Unité de Génétique Moléculaire des Levures (URA‐CNRS 2171), Institut Pasteur, 25 rue du Docteur Roux, 75724, Paris, Cedex 15, France
  7. 7 Unité Microbiologie et Environnement (URA‐CNRS 2172), Institut Pasteur, 25 rue du Docteur Roux, 75724, Paris, Cedex 15, France
  8. 8 Génétique Moléculaire et Cellulaire, CNRS URA 1925 INRA UMR216, Institut National Agronomique Paris‐Grignon, 78850, Thiverval, Grignon, France
  1. *Corresponding author. E-mail: Al.Brown{at}


We have characterized CaNrg1 from Candida albicans, the major fungal pathogen in humans. CaNrg1 contains a zinc finger domain that is conserved in transcriptional regulators from fungi to humans. It is most closely related to ScNrg1, which represses transcription in a Tup1‐dependent fashion in Saccharomyces cerevisiae. Inactivation of CaNrg1 in C.albicans causes filamentous and invasive growth, derepresses hypha‐specific genes, increases sensitivity to some stresses and attenuates virulence. A tup1 mutant displays similar phenotypes. However, unlike tup1 cells, nrg1 cells can form normal hyphae, generate chlamydospores at normal rates and grow at 42°C. Transcript profiling of 2002 C.albicans genes reveals that CaNrg1 represses a subset of CaTup1‐regulated genes, which includes known hypha‐specific genes and other virulence factors. Most of these genes contain an Nrg1 response element (NRE) in their promoter. CaNrg1 interacts specifically with an NRE in vitro. Also, deletion of two NREs from the ALS8 promoter releases it from Nrg1‐mediated repression. Hence, CaNrg1 is a transcriptional repressor that appears to target CaTup1 to a distinct set of virulence‐related functions, including yeast–hypha morphogenesis.


Morphological transitions between unicellular and filamentous growth forms contribute significantly to the virulence of fungal pathogens. The major plant pathogens Ustilago maydis and Magnaporthe grisae form hyphae that penetrate their hosts in the early stages of infection (Hartmann et al., 1996; Hamer and Talbot, 1998). Transitions from avirulent filamentous forms to virulent unicellular forms are essential for the establishment of Histoplasma capsulatum and Paracoccidioides brasiliensis infections in humans (da Silva et al., 1999; Maresca and Kobayashi, 2000). The most prevalent fungal pathogen of humans, Candida albicans, undergoes reversible morphological transitions between unicellular yeast‐like, pseudohyphal and hyphal growth forms, which contribute to disease establishment and progression (Odds, 1994; Brown and Gow, 1999). The yeast form is probably disseminated more effectively than filamentous forms, whereas the hyphal form seems better adapted to penetrate tissue and is thought to help the fungus evade host defences (Sherwood et al., 1992; Lo et al., 1997).

Numerous parameters influence the yeast–hypha transition in C.albicans (Odds, 1988), and these presumably reflect the variety of signals detected by the fungus in the different microenvironments it encounters within its human host (Brown and Gow, 1999). Hyphal development is subject to both positive and negative regulation. Positive regulation is mediated by multiple signalling pathways including mitogen‐activated protein kinase (MAPK), cAMP and pH signalling pathways (Brown and Gow, 1999; Ernst, 2000). Negative regulation is dependent upon the factor CaTup1 (Braun and Johnson, 1997). Inactivation of CaTup1 leads to constitutive filamentous growth and the derepression of hypha‐specific genes (Braun and Johnson, 1997, 2000; Braun et al., 2000; Brown et al., 2000). CaTup1 appears to act independently of the MAPK and cAMP pathways to regulate morphogenesis (Braun and Johnson, 2000), but the mechanisms by which CaTup1 operates remain obscure. In Saccharomyces cerevisiae, ScTup1 represses the transcription of sets of functionally related genes, interacting with their promoters indirectly through specific DNA‐binding proteins (Keleher et al., 1992; Treitel and Carlson, 1995; Park et al., 1999; Smith and Johnson, 2000). CaTup1 might operate in a similar way to repress the expression of hypha‐specific genes and control morphogenesis in C.albicans.

In this study, we report the identification of CaNrg1, which represses hyphal development and hypha‐specific gene expression in C.albicans. Our data suggest that CaNrg1 achieves this by targeting CaTup1 to hypha‐specific genes.


Identification of C.albicans Nrg1

The HYR1 and ALS8 genes are activated specifically during hyphal development in C.albicans (Bailey et al., 1996; Leng et al., 2001). Both promoters contain the sequence C4T (Leng, 1999), which corresponds to the stress response element (STRE) in S.cerevisiae (Marchler et al., 1993). Since many conditions that promote yeast– hypha morphogenesis impose a stress upon the C.albicans cell (Odds, 1988; Brown and Gow, 1999), we reasoned that C4T elements might contribute to the regulation of hypha‐specific genes. Hence, we sought C.albicans factors that interact specifically with C4T.

A C.albicans cDNA library was screened for proteins that interact specifically with a 32P‐labelled C4T‐containing oligonucleotide. The screen yielded a cDNA that reacted reproducibly with the C4T‐containing oligonucleotide, but not with three different control oligonucleotides lacking C4T. This cDNA was used to isolate the corresponding gene.

The 310 amino acid sequence encoded by this gene is most similar to S.cerevisiae Nrg1 (Figure 1), and hence it was named CaNrg1. The overall level of sequence identity between CaNrg1 and ScNrg1 is low (22%), but this rises to 67% within a (C2H2)2 zinc finger region from residues 230 to 280. This region is closely related to the zinc fingers of several fungal transcription factors that are important regulators of carbon assimilation and stress responses (Nehlin and Ronne, 1990; Lundin et al., 1994; Martinez‐Pastor et al., 1996; Park et al., 1999). CaNrg1 is also related to early growth response transcription factor, colon Kruppel‐like protein and Wilms' tumour protein, which regulate important developmental processes in humans (Gessler et al., 1990; Blok et al., 1995; Shi et al., 1999).

Figure 1.

Candida albicans Nrg1 is related to C2H2 zinc finger proteins from other species. Alignment of the CaNrg1 C2H2 domain (residues 226–286) with related fungal and mammalian proteins: ScNrg1, S.cerevisiae Nrg1; ScNrg2, S.cerevisiae Nrg2; ScMsn2, S.cerevisiae Msn2p; ScMig1, S.cerevisiae Mig1; HuEGRTF, human early growth response transcription factor; HuCKLP, human colon Kruppel‐like protein; KlMIG1, Kluyveromyces lactis MIG1 protein; HuWTP, human Wilms' tumour protein; AnCREA, Aspergillus niger CREA protein. Sequence identities to CaNrg1 are provided (%).

Gel shift assays were performed to confirm the DNA binding specificity of CaNrg1. Synthetic CaNrg1, made by in vitro transcription–translation, was close to its predicted mass of 34 kDa (Figure 2A). CaNrg1 extracts formed two complexes with the C4T‐containing oligonucleotide that were not formed by control extracts lacking CaNrg1 (Figure 2B). These complexes were competed out by unlabelled C4T‐containing oligonucleotide, but not by an oligonucleotide lacking C4T (Figure 2B). Hence, CaNrg1 can interact specifically with a C4T‐containing sequence in vitro, like ScNrg1 (Park et al., 1999).

Figure 2.

Nrg1 is a sequence‐specific DNA‐binding protein. (A) Synthesis of in vitro transcribed–translated CaNrg1. SDS–polyacrylamide gel of 35S‐labelled products with a rabbit in vitro transcription–translation mix: control reaction with luciferase plasmid, L; reaction with NRG1 plasmid, N; reaction with no plasmid, –. (B) Gel shift assays of complexes formed with a 32P‐labelled double‐stranded C4T‐containing oligonucleotide: Nrg1 extract, N; control in vitro transcription–translation luciferase extract, L; competition with the same unlabelled C4T oligonucleotide, S; competition with a control oligonucleotide lacking C4T, C. Numbers represent fold excess of competitor. A putative rabbit STRE‐specific complex is highlighted by the open arrow, and Nrg1‐dependent, sequence‐specific complexes are highlighted by the closed arrows.

Morphology of C.albicans nrg1 mutants

Both CaNRG1 alleles in this diploid fungus were inactivated by mutating the CaNRG1 start codon (ATG→GTG) and deleting 438 bp of the coding region. Independent nrg1 mutations were introduced into CAI4 (ura3/ura3) and CAI8 (ura3/ura3, ade2/ade2) backgrounds (Table I). NRG1 mRNA was undetectable in homozygous nrg1 mutants (not shown).

View this table:
Table 1. Candida albicans strains

Homozygous C.albicans nrg1 mutants grew in filamentous forms under conditions that normally promote the growth of yeast cells (Figure 3A). The mutants formed chains of relatively short pseudohyphal cells in YPD at 25°C, and more extended pseudohyphae in synthetic complete medium. They also formed characteristically wrinkly colonies on YPD agar.

Figure 3.

Inactivation or depletion of NRG1 derepresses filamentous growth in C.albicans. (A) Cells were grown in YPD at 25°C (upper panels) or YPD containing 20% serum at 37°C (lower panels): wild‐type, CAF2‐1; nrg1, MMC3; tup1, BCa2‐10; mig1, LOZ124. (B) MMC5 cells (nrg1/nrg1, MET3‐NRG1) growing in SC lacking or containing 10 mM methionine (Met) and cysteine (Cys).

To confirm that these phenotypes correlated with the loss of CaNrg1, the CaNRG1 gene was placed under the control of the CaMET3 promoter, which is repressed by methionine and cysteine (Care et al., 1999). The nrg1/nrg1/MET3‐NRG1 strain (Table I) grew normally in the yeast form in the absence of methionine and cysteine, but formed pseudohyphae following methionine and cysteine addition (Figure 3B). Therefore, CaNrg1 promotes growth of C. albicans in the yeast form.

Phenotypes of C.albicans nrg1, mig1 and tup1 mutants

By analogy with their homologues in S.cerevisiae, CaNrg1 and CaMig1 are probably CaTup1 targeting proteins. Therefore, we compared the phenotypes of homozygous C.albicans nrg1, tup1 and mig1 mutants. Wild‐type and mig1 cells grew normally in the yeast form in YPD at 25°C, and formed normal germ tubes (the progenitors of hyphae) following serum stimulation (Figure 3A). In contrast, nrg1 and tup1 cells formed pseudohyphae in YPD at 25°C. This was consistent with previously reported phenotypes for tup1 and mig1 cells (Braun and Johnson, 1997; Zaragoza et al., 2000). The C.albicans tup1 and nrg1 mutants invaded YPD agar constitutively, unlike wild‐type and mig1 cells (Figure 4A). In addition, tup1 and nrg1 colonies formed filaments at an enhanced rate on low ammonia SLAD medium compared with wild‐type and mig1 colonies (Figure 4B). Hence, nrg1 and tup1 mutants displayed related morphological and invasive phenotypes, suggesting that they might play similar roles in these processes.

Figure 4.

Candida albicans nrg1 and tup1 mutants display subtly different phenotypes. (A) Agar invasion (revealed after washing cells from the surface of agar plates) and temperature effects upon growth: wild‐type, CAF2‐1; NRG1/nrg1, MMC1; nrg1/nrg1, MMC3; tup1/tup1, BCa2‐10; mig1/mig1, LOZ124. (B) Colony fringes of mutants grown on different solid media: 3 days at 37°C on YPD agar containing 20% serum; 6 days at 37°C on Spider agar; 4 days at 30°C on SLAD agar; chlamydospore formation after 7 days at 25°C on cornmeal–Tween agar under coverslips. Homozygous strains are as described in (A).

Further analyses revealed subtle morphological differences between the nrg1 and tup1 mutants. The tup1 mutant remained fixed in a pseudohyphal morphology, whereas nrg1 cells formed true hyphae following serum stimulation (Figure 3A) or pH induction (not shown). Similarly, wild‐type and nrg1 colonies generated extended filaments on serum and Spider plates, whereas tup1 colonies formed truncated extensions (Figure 4B). Therefore, unlike tup1 cells, nrg1 cells are able to form normal hyphal cells under a range of experimental conditions.

The tup1 and nrg1 mutants displayed other phenotypic differences. Unlike the tup1 strain, the nrg1 mutant grew at 42°C (Figure 4A), grew on low glucose media and formed true hyphae on glycerol (not shown). Also, chlamydospore formation was delayed in the tup1 mutant, unlike the nrg1 strain (Figure 4B). Therefore, compared with CaNrg1, CaTup1 appears to play additional roles in the growth and development of C.albicans.

Expression of hypha‐specific genes in nrg1, mig1 and tup1 mutants

In addition to controlling yeast–hypha morphogenesis, CaTup1 represses the expression of hypha‐specific genes (Sharkey et al., 1999; Braun and Johnson, 2000; Brown et al., 2000). Therefore, we compared the expression of hypha‐specific genes in nrg1, mig1 and tup1 strains (Figure 5A). As expected, the hypha‐specific HYR1, ALS8, HWP1 and ECE1 mRNAs were not expressed in wild‐type and mig1 cells under conditions that promote growth of yeast cells. Significantly, these mRNAs were derepressed in nrg1 and tup1 cells. Therefore, both CaNrg1 and CaTup1 repress these hypha‐specific genes under yeast growth conditions.

Figure 5.

Nrg1 represses hypha‐specific genes. (A) Northern analysis of hypha‐specific mRNAs (HWP1, ECE1, HYR1 and ALS8) and a gluconeogenic mRNA (PCK1): WT, CAF2‐1 (wild‐type); nrg1, MMC3; tup1, BCa2‐10; mig1, LOZ124 (Table I). (B) Northern analysis of NRG1 mRNA levels during hyphal development in CAF2‐1: Y, yeast form; H, hyphal form.

We also analysed the glucose‐repressed PCK1 gene (Leuker et al., 1997). The PCK1 mRNA remained at low levels in wild‐type and nrg1 cells, but was derepressed in tup1 and mig1 cells. Hence, both CaTup1 and CaMig1 repress the C.albicans PCK1 gene. Clearly, CaNrg1 only represses a subset of CaTup1‐regulated genes in C.albicans.

CaNRG1 expression during yeast–hypha morphogenesis

We examined CaNRG1 expression during the yeast–hypha transition in wild‐type cells. CaNRG1 mRNA levels were reduced 10‐fold when hyphal growth was stimulated with serum at 37°C for 2 h (Figure 5B). Hence, CaNRG1 expression appears to decline in response to serum stimulation.

Identification of CaNrg1‐, CaTup1‐ and CaMig1‐regulated genes by transcript profiling

CaNrg1 and CaTup1 co‐regulate the HYR1, ALS8, HWP1 and ECE1 genes, and CaMig1 and CaTup1 co‐regulate the PCK1 gene (Figure 5A). Hence, CaNrg1 and CaMig1 appear to regulate distinct subsets of CaTup1‐regulated genes, in a similar fashion to their S.cerevisiae counterparts (Treitel and Carlson, 1995; Smith and Johnson, 2000). To examine this further, we performed transcript profiling using arrays carrying 2002 C.albicans genes representing about one‐quarter of the estimated protein‐coding genes in this fungus.

Replicate membranes were hybridized with cDNA from wild‐type, nrg1, mig1 and tup1 strains grown under conditions that favour the yeast form (CAF2‐1, MMC3, LOZ124 and BCA2‐10, respectively; Figure 6A). Expression was detected for 1851 genes (91.9%). A significant proportion of genes were derepressed >3‐fold in the nrg1 (10.7%), mig1 (10.9%) or tup1 (9.9%) mutants. As expected, distinct subsets of CaNrg1–CaTup1‐co‐regulated genes and CaMig1–CaTup1‐co‐regulated genes were observed (Figure 6B). Furthermore, the hypha‐specific genes ALS3, ALS8, HWP1 and ECE1 all appeared in the list of the top 25 CaNrg1–CaTup1‐co‐regulated genes (Table II). Hence, most of the previously defined hypha‐specific genes fell into the subset of CaNrg1– CaTup1‐co‐regulated genes defined by transcript profiling. This suggests that these transcription factors play a major role in the establishment of morphogenetic expression patterns in C.albicans, and that some of the unknown genes on this list might execute hypha‐specific functions.

Figure 6.

The use of transcript profiling to identify targets of Nrg1 regulation. (A) 33P signals from a small portion of membrane‐based arrays carrying 2002 C.albicans ORFs, showing 28‐fold derepression of one gene of unknown function in the nrg1 strain (MMC3) compared with the wild‐type control (CaF2). (B) Proportions of those genes that show detectable levels of expression, which are regulated by Nrg1, Tup1 and/or Mig1. (C) Percentage of regulated genes that are predicted to belong to the functional categories cell rescue, energy and transcription: Nrg1‐regulated genes, N; Tup1‐regulated genes, T; Mig1‐regulated genes, M; all genes with a predicted function, A.

View this table:
Table 2. Top 25 Nrg1–Tup1‐co‐regulated C.albicans genes revealed by transcript profiling

Not all genes that were regulated by CaNrg1 were also regulated by CaTup1 (Figure 6B). Hence, CaNrg1 might execute additional regulatory functions in C.albicans. The same was true for CaMig1. Hence, we sought functional categories that were enriched in the subsets of C.albicans genes that are regulated by CaNrg1, CaTup1 or CaMig1. This task was complicated by the fact that the functions of only a small proportion of C.albicans genes have been defined experimentally. Hence, for the purposes of this experiment, we assumed that sequence homologues execute similar functions in C.albicans and S.cerevisiae (Figure 6C). Certain functional categories were under‐represented in CaNrg1‐ and CaMig1‐regulated genes, compared with the gene set as a whole. These included transcription, cell growth and division, and protein destination (not shown). Other functional categories, such as metabolism and transport facilitation, were over‐represented in both CaNrg1‐ and CaMig1‐regulated subsets (not shown). CaMig1‐, but not CaNrg1‐regulated genes, were enriched for the functional category energy. In contrast, cell rescue functions were enriched in CaNrg1‐, but not CaMig1‐regulated genes (Figure 6C), suggesting that CaNrg1 might play a role in the regulation of stress responses as well as yeast–hypha morphogenesis.

Inactivation of CaNrg1 attenuates virulence and resistance to some stresses

To test whether CaNrg1 influences stress responses in C.albicans, we measured the doubling times of nrg1 cells following exposure to certain stresses. The nrg1 strain grew more slowly in YPD at 30°C (MMC3 Td = 190 ± 9 min) than its isogenic parent (CAF2‐1 Td = 108 ± 4 min), but was as resistant to 0.7 M NaCl and amino acid starvation as the wild‐type strain. However, the nrg1 cells showed increased sensitivity to H2O2 (Td = 285 ± 11 min), 5% ethanol (Td = 447 ± 62 min) or carbon limitation (0.2% glucose: Td = 530 ± 24 min). Therefore, as suggested by transcript profiling, CaNrg1 influences some stress responses.

CaNrg1–CaTup1‐responsive genes included hypha‐ specific genes and other virulence‐related functions such as a secreted aspartyl proteinase, adhesins and proteins involved in iron assimilation (Table II). Therefore, we compared the lethality of isogenic C.albicans nrg1, tup1 and wild‐type strains in a mouse model of systemic candidosis (Figure 7). The virulence of both mutants was attenuated significantly. The fungal load in the kidneys of infected animals was significantly lower for the nrg1 (2.5 × 105 c.f.u./g) and tup1 strains (<1 × 103 c.f.u./g) compared with the parental strain (1.1 × 106 c.f.u./g). Hence, reduced fungal loads correlated with attenuated virulence. These data are consistent with the idea that CaNrg1 and CaTup1 regulate virulence traits.

Figure 7.

Inactivation of CaNrg1 or CaTup1 attenuates the virulence of C.albicans in systemic infections. Survival of mice was monitored following injection of C.albicans cells into the tail vein: wild‐type (CAF2‐1, circles); nrg1/nrg1 (MMC3, triangles); tup1/tup1 (BCa2‐10, squares).

Nrg1 response elements in C.albicans

CaNrg1 binds a C4T‐containing sequence in vitro (Figure 2), and ScNrg1 interacts with C4T or C3TC (Park et al., 1999). Hence, we searched for these elements in the top 25 CaNrg1–CaTup1‐co‐regulated promoters. Some CaNrg1‐regulated genes, e.g. ECE1 (Figure 5), lacked these elements. Therefore, the sequences C4T and C3TC are not sufficient to define an Nrg1 response element (NRE) in C.albicans.

Further in silico analyses of CaNrg1–CaTup1‐responsive promoters using regulatory sequence analysis tools (van Helden et al., 2000) revealed that most carry an element that is closely related to the core sequence C3T: (A/C)(A/C/G)C3T (Table II). The ALS8 promoter, which is regulated by CaNrg1 (Figure 5), carries two such sequences (Figure 8A). Hence, we tested whether ALS8 repression is dependent upon these putative NREs using the Renilla reniformis luciferase reporter (Srikantha et al., 1996). As expected, an ALS8‐RrLUC reporter was derepressed by inactivation of CaNrg1 (Figure 8B). However, mutation of the C4TC region alone at −80 did not derepress the ALS8 promoter in yeast cells (not shown), confirming that C4T and C3TC are not sufficient to define the C.albicans NRE. Inactivation of both putative NREs did relieve ALS8‐RrLUC repression in wild‐type CaNRG1 cells growing in the yeast form (Figure 8B). The level of derepression of this mutant ALS8 promoter was not increased further in nrg1 cells, indicating that these effects were not additive. Furthermore, the levels of derepression were equivalent if not greater than those observed for the ALS8 promoter in wild‐type hyphae (Figure 8B). Therefore, these (A/C)(A/C/G)C3T sequences mediate the repression of ALS8 by CaNrg1 during growth in the yeast form, and this repression is released completely by inactivation of CaNrg1 alone.

Figure 8.

Nrg1 response elements are required for Nrg1p‐mediated repression. (A) Cartoon of the ALS8 promoter showing the positions of NREs (N), an E‐box (E; Leng et al., 2001) and the TATA‐box (T) relative to the major transcriptional start site. NRE sequences are highlighted (bold, underlined), and the mutations are show below the wild‐type sequences. (B) Activities of wild‐type and mutated ALS8 promoter fusions in yeast (Y; closed bars; YPD at 25°C) and hyphal cells (H; open bars; YPD containing 10% serum at 37°C). Luciferase activities in CAI8 (NRG1/NRG1) and MMC9 (nrg1/nrg1) cells were measured for the empty RrLUC vector (pCRW3; V), the wild‐type ALS8 promoter (WT) and the mutated ALS8 promoter (M).


Mode of action of CaNrg1

We identified the C.albicans NRG1 gene on the basis that CaNrg1 can bind specifically to C4T‐containing oligonucleotides. This sequence‐specific DNA binding activity was confirmed using gel shift assays (Figure 2). DNA sequencing revealed that S.cerevisiae Nrg1 is the closest known homologue of CaNrg1 (Figure 1). Although ScNrg1 has been reported to interact with C4T or C3TC (Park et al., 1999), two observations suggest that these elements are not sufficient to account for CaNrg1 binding at C.albicans promoters. First, the ECE1 promoter lacks C4T or C3TC, but is strongly repressed by CaNrg1 (Figure 5). Secondly, mutation of the single C4TC sequence in the ALS8 promoter does not relieve its repression by CaNrg1. Detailed in silico analyses revealed a related sequence, (A/C)(A/C/G)C3T, which is present in most of the promoters of CaNrg1–CaTup1‐regulated genes (Table II). The genes in Table II that do not carry this element may be regulated indirectly by CaNrg1. Significantly, two of these elements exist in the ALS8 promoter, and CaNrg1‐mediated repression of this promoter is relieved when both of these elements are mutated (Figure 8). Therefore, CaNrg1 is a sequence‐specific DNA‐binding protein that represses transcription via (A/C)(A/C/G)C3T. We have called this sequence the NRE.

ScNrg1 interacts physically with ScTup1 complexes (Park et al., 1999). By analogy with S.cerevisiae, CaNrg1 might interact with a CaTup1 complex to regulate gene expression in C.albicans. The fact that C.albicans nrg1 and tup1 mutants display related phenotypes is consistent with this idea. Both nrg1 and tup1 mutants show constitutive filamentous morphologies, wrinkly colonies and invasive growth (Figures 3 and 4; Braun and Johnson, 1997). Also, hypha‐specific genes are derepressed in both nrg1 and tup1 mutants (Figure 5). Furthermore, transcript profiling confirms that a subset of C.albicans genes is co‐regulated by CaNrg1 and CaTup1 (Figure 6). Hence, CaNrg1 might target CaTup1 to a subset of C.albicans promoters. However, our data suggest that CaNrg1 does more than target CaTup1 to hypha‐specific promoters, because some CaNrg1‐regulated genes are not derepressed in the tup1 mutant (Figure 6).

CaNrg1, CaTup1 and CaMig1 regulate distinct but overlapping subsets of genes

In S.cerevisiae, ScTup1 regulates the expression of several distinct sets of genes, only some of which are controlled by ScNrg1 or ScMig1 (e.g. Keleher et al., 1992; Treitel and Carlson, 1995; Park et al., 1999; Smith and Johnson, 2000). By analogy, CaNrg1 and CaMig1 might control the expression of different subsets of CaTup1‐regulated genes in C.albicans. As predicted, the hypha‐specific genes ALS8, ECE1, HWP1 and HYR1 were co‐regulated by CaNrg1 and CaTup1, but not by CaMig1 (Figure 5). Also, PCK1 was co‐regulated by CaMig1 and CaTup1, but not by CaNrg1. The more global view provided by transcript profiling showed clearly that CaNrg1, CaMig1 and CaTup1 repress distinct, but overlapping gene sets in C.albicans (Figure 6). However, some genes were co‐regulated by all three factors, and some genes were only controlled by one of these three factors.

These data indicate that CaNrg1 and CaTup1 play overlapping but distinct roles in C.albicans, and hence that nrg1 and tup1 mutants might display phenotypic differences. As predicted, such differences were observed. The nrg1 cells form hyphae following serum induction, whereas tup1 cells appear to be fixed in a pseudohyphal state (Figure 3). This is reflected in the different behaviour of nrg1 and tup1 colonies on a range of solid media (Figure 4). Further subtle differences between the mutants were observed. Unlike tup1 cells, nrg1 cells are able to grow following carbon limitation or at 42°C (Figure 4). Also, chlamydospore formation is delayed in the tup1 strain, but not in the nrg1 mutant (Figure 4).

These subtle differences between nrg1 and tup1 cells may be due to the pleiotropic nature of the tup1 mutation. In other words, the inability of tup1 cells to form hyphae might be due to secondary effects of the mutation caused by changes in metabolism, for example. Pleiotropic effects of the tup1 mutation might also contribute to the difficulties in defining the hierarchical relationships between CaTup1 and other morphological regulators (Braun and Johnson, 2000).

Biological roles of CaNrg1

Several observations indicate that CaNrg1 plays a central role in the negative regulation of yeast–hypha morphogenesis (Figure 9). Most significantly, the inactivation of CaNrg1 derepressed filamentous and invasive growth (Figures 3 and 4), and derepressed hypha‐specific genes (Figure 5; Table II). The repression of hypha‐specific genes by CaNrg1 appears to be direct, because these genes contain at least one NRE in their promoter (Figure 8; Table II). However, CaNrg1 might also play an indirect role in the control of hyphal development, because CEK1 was present in the CaNrg1–CaTup1‐co‐regulated gene set revealed by transcript profiling (Table II). CEK1 encodes the MAPK on one of the morphogenetic signalling pathways that activates hyphal development (Csank et al., 1998; Brown and Gow, 1999; Ernst, 2000). Therefore, CaNrg1 might regulate hyphal development at two levels: directly by repressing hypha‐specific functions, and indirectly by down‐regulating the MAPK signalling pathway. This repression appears to be relieved, at least in part, by the down‐regulation of CaNRG1 in response to a morphogenetic signal (Figure 9).

Figure 9.

Model for Nrg1 function in C.albicans. According to this model, CaNrg1 binds to the NRE and brings the CaTup1 repressor to the promoters of genes involved in hyphal development, some stress responses and virulence attributes. CaNRG1 mRNA levels are down‐regulated by serum, suggesting that morphogenetic signals can relieve CaNrg1‐mediated repression by reducing CaNrg1 levels. CaNrg1 also appears to execute CaTup1‐independent roles in C.albicans (see text).

A second putative CaTup1‐targeting protein regulates morphogenesis in C.albicans. Inactivation of Rfg1, a homologue of S.cerevisiae Rox1, derepresses hyphal growth and hypha‐specific genes (Kadosh and Johnson, 2001). Hence, some rfg1 and nrg1 phenotypes are similar. However, nrg1 cells seem to display stronger hyphal phenotypes than rfg1 cells (Braun et al., 2001). Also, CaNrg1‐mediated repression is sufficient to account for the repression of at least one hypha‐specific gene (Figure 8). Nevertheless, in the absence of a morphogenetic stimulus, nrg1 cells grow as pseudohyphae (Figure 4), suggesting that inactivation of CaNrg1 does not fully derepress hyphal development. Hence, CaNrg1 and Rfg1 both repress morphogenesis, apparently via similar mechanisms (Figure 9; Braun et al., 2001; Kadosh and Johnson, 2001), although CaNrg1 might play the predominant role.

CaNrg1 plays an additional role in some stress responses (Figure 6). The NRE (A/C)(A/C/G)C3T includes the sequence C4T, which acts as a stress response element (STRE) in S.cerevisiae (Marchler et al., 1993; Mager and De Kruijff, 1995). The sequence C4T is present in the promoters of stress response genes in C.albicans (not shown), and putative cell rescue functions are over‐represented in the subset of CaNrg1‐regulated genes revealed by transcript profiling (Figure 6). However, not all (A/C)(A/C/G)C3T sequences contain C4T. Hence, there is only partial overlap between NREs and STREs. Not surprisingly, therefore, C.albicans nrg1 mutants only displayed increased sensitivity to a subset of stresses, including an oxidative stress, ethanol treatment and carbon limitation. Nevertheless, CaNrg1 provides a mechanistic link between stress responses and cellular morphogenesis in C.albicans (Brown and Gow, 1999).

Our data also provide clues about the role of CaMig1 in C.albicans. Consistent with the findings of Zaragoza et al. (2000), the C.albicans mig1 mutant did not display a morphological phenotype, except on Spider medium (Figure 4). Also, CaMig1 does not influence the expression of hypha‐specific genes (Figure 5). However, CaMig1 does regulate the gluconeogenic gene PCK1 (Figure 5), and putative energy functions were over‐represented in the subset of CaMig1‐regulated genes (Figure 6), suggesting that CaMig1 might control aspects of C.albicans metabolism, like its homologue in S.cerevisiae (Gancedo, 1998).

Co‐regulation of virulence factors by CaNrg1 and CaTup1

Our data also provide evidence for the co‐regulation of virulence factors in C.albicans. Several reports have suggested that the expression of some virulence factors, such as adhesion and morphogenesis, might be linked in C.albicans (Hube et al., 1994; Staab et al., 1996, 1999; Leng, 1999). Our data reinforce these observations, since the set of CaNrg1–CaTup1‐co‐regulated genes includes adhesins (HWP1, ALS3 and ALS8), a secreted aspartyl proteinase (SAP5) and two genes involved in iron assimilation (CFL1 and FTR1; Yamada‐Okabe et al., 1996; Ramanan and Wang, 2000) (Table II). Hence, CaNrg1 might play a central role in the coordination of key virulence attributes in C.albicans (Figure 9). Not surprisingly, therefore, inactivation of CaNrg1 significantly attenuates the virulence of C.albicans (Figure 8).

Materials and methods

Strains and growth conditions

Candida albicans strains (Table I) were grown in: YPD (Sherman, 1991); YPD containing 10% fetal calf serum; Soll's medium at pH 4.5 or 6.5 (Swoboda et al., 1994); SC, synthetic complete medium (Kaiser et al., 1994); YNB medium containing low glucose (0.67% yeast nitrogen base without amino acids, 0.2% glucose); SLAD agar (Gimeno et al., 1992); Spider agar (Liu et al., 1994); and cornmeal–Tween agar (Braun and Johnson, 1997). The C.albicans cell morphology was analysed using an Olympus BX50 microscope mounted with a 35 mm Olympus camera, and quantified using an improved Neubauer haemocytometer. Colonies were analysed using a Zeiss stereo microscope stemi 2000‐C.

CaNRG1 isolation

A C.albicans cDNA expression library in λ‐ZAPII (Swoboda et al., 1993) was screened using published procedures (Singh, 1993) with a C4T‐containing oligonucleotide (top strand, 5′‐CGCGTGTATAAACCCCCTTTTCTTGGGGCCCCTTTTCTTGGGGGA). Clones that gave positive and reproducible signals with this oligonucleotide, but no significant signals with oligonucleotides lacking C4T, were selected. A cDNA clone was converted into its pBlueScript form (Stratagene, Cambridge, UK) to create pBS‐NRG1, and was then used to isolate the complete NRG1 locus from a C.albicans genomic library (Smith et al., 1992) by colony hybridization (Sambrook et al., 1989).

CaNRG1 disruption

The NRG1 locus (−1262 to +1243) was cloned into pGEM‐T® (Promega, Southampton, UK). Reverse PCR was used to mutate the initiation codon to GTG, delete codons 92–237 and introduce a BglII site at the deletion to generate pGEM‐Δnrg1. The 3 kb hisG‐URA3‐hisG sequence (Fonzi and Irwin, 1993) was then inserted into pGEM‐Δnrg1 to create the nrg1::hisG‐URA3‐hisG cassette. This cassette was released from the pGEM‐T backbone using NotI and transformed into C.albicans (Gietz and Woods, 1998). Both CaNRG1 alleles in the strains CAI4 and CAI8 (Table I) were disrupted using two rounds of ura‐blasting (Fonzi and Irwin, 1993). Disruptions were confirmed by Southern blotting and PCR diagnosis.

To create MET3‐NRG1, the NRG1 open reading frame (ORF) was PCR amplified using the primers 5′‐CATTAAGATCTAAACAATCATTATGC and 5′‐GCAATTAACCCTCGAGATTTAACCCG (BglII and XhoI sites underlined), cloned into pGEM®‐T Easy to make pGEM‐NRG1, and resequenced. The NRG1 ORF was then released from pGEM‐NRG1 using PstI and SphI, and ligated into pCaEXPa (Care et al., 1999) to make pMET3‐NRG1. This plasmid was linearized with StuI, transformed into C.albicans, and single copy integration at the RP10 locus (Murad et al., 2000) was confirmed by Southern blotting.

RNA and DNA analyses

Published methods were used for RNA and DNA preparation, Southern blotting and northern analysis (Hoffman and Winston, 1987; Brown, 1994; Wicksteed et al., 1994; Planta et al., 1999). cDNA and genomic clones were sequenced using Big Dye Terminator Cycle Sequencing kits (Perkin Elmer, Workington, UK) and run on an ABI 377 automated DNA sequencer. The NRG1 sequence has DDBJ/EMBL/GenBank accession No. AF321521, and is available in the Stanford University C.albicans genome sequence, in Contig6‐2506 ( DNA sequences were analysed at the Stanford Genome Database (, and promoter sequences were analysed using regulatory sequence analysis tools (; van Helden et al., 2000).

Transcript profiling

Details of the transcript profiling methods will be published elsewhere (A.M.A.Murad, C.Gaillardin, C.d'Enfert, F.Tekaia, D.Marechal, D.Talibi and A.J.P.Brown, submitted). Briefly, ORFs were identified using Release 3 of the C.albicans genome sequence obtained from the Stanford DNA Sequencing and Technology Center website at (April 1999). Replicate arrays of PCR products corresponding to 2002 C.albicans ORFs were hybridized with [33P]cDNA prepared using RNA from mid‐exponential C.albicans cells (Hauser et al., 1998). Cells were grown in YPD at 30°C to an OD600 of 0.8, subcultured into fresh YPD at 30°C, grown to an OD600 of 0.8 and then harvested for analysis. Signals were detected using a Fuji FLA‐3000 phosphoimager and analysed using Array Vision software (Amersham, Buckinghamshire, UK).

Gel shift assays

Synthetic Nrg1 was made in vitro using pBS‐NRG1 and the TNT® reticulocyte lysate coupled transcription/translation system (Promega, Southampton, UK). Control reactions were performed using the control luciferase DNA provided by the manufacturers. 35S‐labelled products were analysed by SDS–PAGE. Unlabelled reaction mixes were used in gel shift assays with 32P‐end‐labelled double‐stranded oligonucleotides (Sambrook et al., 1989): C4T‐containing oligonucleotide, 5′‐GTCCTGTATAAACCCCCTTTTCTTGGGGCCCCTTTTCTTGGGGAG (top strand); control oligonucleotide, 5′‐GTCCTGGCTTCCAAAAGACCTTGAATTGAGGTCTGATAGTAG (top strand). Assays were performed on 20 fmol of oligonucleotide with 20 ng of iv product and 0.5 μg/ml poly(dI–dC)·poly(dI–dC), as described (Carey, 1991), and electrophoresed on 5% polyacrylamide gels.

Reporter assays

RrLUC promoter fusions were made using pCRW3 (Srikantha et al., 1996). The wild‐type ALS8 promoter, site‐directed from −696 to +4, was fused in‐frame with the RrLUC ORF. Putative NREs were altered by mutagenesis and resequenced (Sambrook et al., 1989) as shown in Figure 8A. Wild‐type and mutant ALS8‐RrLUC plasmids and the pCRW3 control were linearized with HindIII, and transformed into C.albicans CAI8 (Gietz and Woods, 1998). Single copy integration at the ade2 locus was confirmed by PCR diagnosis. To measure luciferase activities, protein extracts were prepared from C.albicans transformants after 3 h growth in YPD at 25°C, or YPD containing 10% serum at 37°C. Quadruplicate luciferase assays (in RLU/20 μg protein/20 s) were performed on fresh extracts with 0.5 μM co‐elentrazine (Molecular Probes, Leiden, The Netherlands) in a Lumat LB9507 luminometer (EG&G Berthold) (Srikantha et al., 1996). Similar data were obtained in three experiments using independent transformants.

Virulence assays

The virulence of C.albicans strains was assessed in a mouse model of systemic candidosis. Cells were grown for 24 h at 30°C in 4% glucose, 1% peptone, 0.1% yeast extract, 1% glycerol, harvested by centrifugation, and washed twice with water. Fungal biomass was compared on the basis of ATP concentration (Odds and Abbott, 1984), and an equivalent of 5000 c.f.u. per mouse was injected into the lateral tail vein of DBA2 mice (average weight 19 g; Harlan UK).


We are grateful to Phil Carter for help with DNA sequencing, and to Nicole Hauser and Joerg Hoheisel for providing access to Array Vision software. We thank Burk Braun, Alexander Johnson, Oscar Zaragoza and Carlos Gancedo for providing C.albicans strains, and Susan Budge for excellent technical assistance. We are particularly grateful to Alexander Johnson and David Kadosh for communication of results prior to publication. We thank the Stanford DNA Sequencing and Technology Center for access to their C.albicans genome sequence data (, which were generated with the support of the NIDR and the Burroughs Wellcome Fund. A.M.A.M. was supported by the Malaysian Government [UKM(PER)7371], P.L., M.S., J.W. and S.M. by the UK Biotechnology and Biological Sciences Research Council (CEL04563, 1/P11585, 97/B1/P/03008, 98/A1/P/04001), and A.J.P.B. by The Wellcome Trust (055015, 063204). C.E. and C.G. were supported by the French Ministère de la Recherche (Réseau Infections Fongiques, P.R.F.M.M.I.P.). A.J.P.B., C.E. and C.G. were also supported by the European Commission (QLRT‐1999‐30795).