Control of Filament Formation in Candida albicans by the Transcriptional Repressor TUP1

See allHide authors and affiliations

Science  04 Jul 1997:
Vol. 277, Issue 5322, pp. 105-109
DOI: 10.1126/science.277.5322.105


The pathogenic yeast Candida albicans regulates its cellular morphology in response to environmental conditions. Ellipsoidal, single cells (blastospores) predominate in rich media, whereas filaments composed of elongated cells that are attached end-to-end form in response to starvation, serum, and other conditions. The TUP1 gene, which encodes a general transcriptional repressor in Saccharomyces cerevisiae, was isolated fromC. albicans and disrupted. The resulting tup1mutant strain of C. albicans grew exclusively as filaments under all conditions tested. TUP1 was epistatic to the transcriptional activator CPH1, previously found to promote filamentous growth. The results suggest a model where TUP1represses genes responsible for initiating filamentous growth and this repression is lifted under inducing environmental conditions.

The yeast Candida albicans is an opportunistic pathogen of humans, causing common superficial infections as well as life-threatening disseminated and organ infections. Fungal pathogens such as C. albicans are of increasing concern because of the rising incidence of immunosuppression brought about by AIDS, diabetes, cancer therapies, organ transplantation, and other conditions (1).

Typically, C. albicans grows as single ellipsoidal cells called blastospores (also called blastoconidia). In the presence of inducing environmental signals, C. albicans can assume filamentous forms in which cells remain attached to each other after dividing and thereby form long branched strings of connected cells. These filamentous forms range from pseudohyphae (where cells that form filaments are elongated, but still ellipsoidal) to true hyphae (where highly elongated cells that form the filaments are cylindrical and are separated by perpendicular septal walls). The ability of C. albicans to adopt these different morphologies is thought to contribute to colonization and dissemination within host tissues, and thereby to promote infection (2, 3). All morphological forms can be found within infected tissues. In the laboratory, environmental conditions influence the morphological state of C. albicans. Serum causes blastospores to sprout true hyphae (termed germ tubes at their initial appearance). High temperature (37°C), high ratio of CO2 to O2, neutral pH, and nutrient-poor media also stimulate hyphal growth. Conversely, low temperatures, air, acidic pH (4 to 6), and enriched media promote blastospore growth (2, 4). Intermediate conditions can induce various pseudohyphal forms as well as true hyphae (We use “filamentous” to refer to both pseudohyphae and hyphae).

One pathway that regulates cell morphology in C. albicans has been discovered. The gene products ofCPH1, HST7, and CST20 are the C. albicans homologs of the S. cerevisiae STE12,STE7, and STE20 products, respectively.Candida albicans strains mutant in any of these genes show retarded filamentous growth but no impairment of serum-induced germ tube and hyphae formation (5, 6). These results suggest that a kinase signaling cascade, similar to that leading to STE12activation in Saccharomyces cerevisiae, plays a part in stimulating the morphological transition between blastospore and filamentous forms in C. albicans.

We now describe another regulator of filamentous growth, theTUP1 gene, whose function has been studied in S. cerevisiae, where it represses transcription of many different genes (7-9). Targets of TUP1 regulation include glucose-repressed genes, oxygen-repressed genes, DNA damage– induced genes, a-specific mating genes, haploid-specific genes, and flocculation genes. These sets of genes are each regulated by a distinct upstream DNA-binding protein, and each DNA-binding protein recruits to the promoter a complex containing theTUP1 gene product. Several lines of evidence indicate that the TUP1 gene product plays the principal role in bringing about transcriptional repression by mechanisms still not well understood (10).

In our search for homologs of S. cerevisiae TUP1, we isolated a gene from the closely related yeast Kluyveromyces lactis, which has the ability to complement a tup1deletion mutation in S. cerevisiae cells. The K. lactis TUP1 gene was similar to S. cerevisiae TUP1, and we used the shared sequence information to design degenerate PCR (polymerase chain reaction) primers to amplify conserved regions in the COOH-terminus of TUP1 from other organisms includingC. albicans. The principal PCR product generated fromC. albicans genomic DNA was cloned, sequenced, and used as a probe to isolate a full-length gene from a C. albicansgenomic library (11). Sequencing and conceptual translation revealed an open reading frame similar to that of TUP1 fromS. cerevisiae (67% identity over the entire amino acid sequence) (Fig. 1A). Major conserved features were the seven WD40 repeats at the COOH-terminus ofTUP1 (which anchor TUP1 to DNA-binding proteins) and the NH2-terminus, including a proximal glutamine-rich segment (Fig. 1B). WD40 amino acid sequence repeats are found in many other proteins, including β subunits of heterotrimeric G proteins (12).

Figure 1

Sequence of C. albicans TUP1. (A) TUP1 gene products from C. albicans and S. cerevisiae, as conceptually translated from their respective DNA sequences, are compared. Alignment was performed by the program pileup (GCG, Inc.), and identities are highlighted. (B) The twoTUP1 gene products and the relative arrangement of their NH2-terminal conserved domains (hatched) and their COOH-terminal WD40 repeated motifs (filled). Abbreviations for the amino acid residues are: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K. Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

To determine whether the C. albicans TUP1 gene had functional as well as structural similarity to S. cerevisiae TUP1, we obtained expression of C. albicans TUP1 under galactose control in tup1 S. cerevisiae cells. The overexpressed C. albicans gene restored repression of a genomic a-specific gene reporter,Mfa2:lacZ to that in wild-type (13). In addition, tup1 S. cerevisiae cells overexpressing theC. albicans TUP1 were not flocculent, were not temperature sensitive, exhibited wild-type cell shape, and grew rapidly, indicating that several other phenotypes characteristic of tup1 cells had also been corrected by the C. albicans gene. Thus in these two species, TUP1 apparently has the same molecular function: It is recruited to DNA by various DNA binding proteins, and it represses transcription.

To determine which pathways are controlled by the TUP1repressor in C. albicans, which is diploid, we disrupted both copies of the gene in two rounds (14). The disruption consisted of a large deletion that excised most of the TUP1gene as well as 330 bp of DNA upstream of the open reading frame. To ensure that the phenotypes described below resulted from loss ofTUP1 function rather than loss of the upstream DNA or other features of the locus separate from the TUP1 open reading frame, we performed a second round of disruption with a DNA fragment that carried tup1 C. albicans with an NH2-terminal frameshift mutation instead of a large deletion (Fig. 2A). The resulting strains were, in all respects, phenotypically identical to the homozygous mutant strains carrying the large deletions of TUP1, which are described below. Wild-type C. albicans phenotypes were fully restored by insertion of a wild-type copy of the TUP1gene linked to an adjacent URA3 marker (Fig. 2A) back into the disrupted locus (Fig. 2B, lane 4). Furthermore, insertion of a wild-type copy of the gene under the control of a maltase promoter into the genome also rescued the tup1 deletion mutant phenotypes in a maltose-dependent manner (13).

Figure 2

Disruption of C. albicans TUP1. (A) The open reading frame of the C. albicans TUP1 locus shows as a box containing conserved sequence elements (as in Fig. 1B). The top line represents the original genomic clone, the insert of plasmid p371. The second line represents the disruption fragment contained on p383C. The third line represents the rescuing fragment carried on p405, and the last line corresponds to the frameshift mutant (p418), created by filling in the indicated Eco RI site of p405. (B) A DNA blot of C. albicans genomic DNA (cut with Nhe I–Spe I) was probed with the Hind III–Spe I fragment from the TUP1 genomic locus (A, top line). Lanes 1 and 2, DNA fromTUP1/TUP1 strains (length equals 3 kbp); lane 3, DNA from a heterozygoustup1/TUP1 strain (3 kbp and 2.3 kbp); lane 6, DNA from a homozygous tup1/tup1 mutant strain. Lanes 4 and 5 show integration of the p405 rescuing fragment (third line from top in A) into the TUP1 locus. Integration of the subportion of the fragment with URA3 but without TUP1 resulted in the slightly smaller band shown in lane 5 (approximately 9 kbp) and did not restore TUP1 function, whereas integration of the entire fragment, shown in lane 4 (approximately 9.7 kbp), did restore TUP1 function.

Differences were observed when tup1 and wild-type C. albicans were compared under the microscope (Fig.3) (15). For A to D, both strains were grown under conditions (YEPD) (16) that favor the blastospore form of growth, and, as expected, the wild-type strain exhibited the blastospore form under these conditions (Fig. 3, A and C). In contrast, the homozygous tup1/tup1 mutant strain was completely filamentous (Fig. 3, B and D). The mutant strain formed only filaments on all media tested, including common and specialized media, namely, YEPD, YD, Saboraud, corn meal with or without Tween 80, Spider, 20% calf serum, Lee's defined, and minimal S medium with a variety of fermentable and nonfermentable carbon sources (16). On most media, mutant cells grew as pseudohyphae rather than as true hyphae; but under certain hyphal-inducing conditions, they attained elongated and straight-walled shapes indistinguishable from those of true hyphae (Fig. 3, B and D; and Fig. 4). Some of these conditions included growth on nutrient-poor media such as corn meal agar, and micro-aerobic growth under glass coverslips. The distinction between true hyphae and pseudohyphae is based on cell shape and cell division timing, and a spectrum of intermediate morphologies is observed in wild-type C. albicans cells (2, 4, 6).

Figure 3

Morphological characteristics of tup1 C. albicans. (A and C) Wild-type cells (SC5314) and (B and D) tup1 cells (BCa2-10) were grown in YEPD at 30°C to late log phase and stained with DAPI (15) to highlight the DNA before being photographed at 40× through differential interference contrast (A and B) and fluorescence optics (C and D). (E) Wild-type cells (SC5314) and (F) tup1 cells (BCa2-10) were grown in Lee's medium, pH 6.7, at 37°C, conditions that promote germ tube formation and hyphal growth, and then stained with calcofluor and DAPI to highlight both the cell walls and DNA before being photographed at 100× through fluorescence optics. Scale bar, 50 μM.

Figure 4

Colony growth of homozygous and heterozygoustup1 strains. (A) Wild-type (SC5314), (B) heterozygous (BCa2-3′), and (C) homozygous (BCa2-10) cells were placed on a cornmeal agar plus Tween 80 plate under a coverslip and grown for 25 hours at 25°C before being photographed at 40× with phase optics. Scale bar equals 50 μM.

Closer examination of homozygous tup1 mutant cells revealed that, apart from their overall altered morphology, they resembled filamentous wild-type cells in most respects (Fig. 3, E and F). In particular, DNA was centrally located in non-mitotic cells, filaments branched several septal compartments behind the growing hyphal tip, and branches were situated near the apical septa, as is normally seen in wild-type C. albicans. One minor difference was that the mutant cells often had slightly misshapen cell walls (Fig. 3F).

Heterozygous TUP1/tup1 strains showed a morphological phenotype intermediate between the wild-type and homozygous strains. Although their cells resembled wild-type cells in morphology, on most media heterozygous colonies developed a higher proportion of filaments compared to wild-type colonies (Fig. 4B), confirming the filament-repressing role of TUP1 and suggesting that its gene product is present in limiting amounts.

Whereas deletion of the TUP1 gene caused constitutive filamentous growth in C. albicans, there was a surprising lack of response of tup1 cells to some strong germ tube and filamentous growth inducers such as mammalian serum and Lee's medium. Germ tube formation from the blastospore state is a special property ofC. albicans and as such is used for clinical identification. Wild-type and TUP1/tup1 heterozygous blastospores exhibited rapid germ tube formation progressing to true hyphae on YEPD or minimal media containing 10 to 20% calf serum (2, 3). However, in these same media the homozygous tup1 mutant cells showed no detectable change in filamentous morphology; in particular, they showed no sign of germ tubes or of increased transformation toward true hyphae. The blastospore to hypha transition can also be experimentally manipulated with the defined medium developed by Lee et al. (17) which, depending on the pH and temperature of incubation, promotes blastospore growth or germ tube formation and filamentous growth. As on serum,tup1 mutant cells were unaffected by Lee's medium and grew with the same filamentous morphology regardless of pH and temperature (Fig. 4F). One hypothesis to explain these observations is that initiation of the pathway blastospore to germ tube to hyphae requires the blastospore cell type. Since this cell type is absent in thetup1 homozygote, the pathway would, according to the hypothesis, fail to initiate. Another hypothesis is that serum induction normally operates through TUP1.

tup1 mutants of S. cerevisiaeshow various phenotypes including sensitivity to 37°C, slow growth, lack of glucose repression, poor growth on glycerol, inability of the α cell type to mate, inability to sporulate, flocculence, and irregular cell shape (7, 9). We therefore examined thetup1 strains of C. albicans for additional phenotypes. Differences in the growth rate between wild-type andtup1 C. albicans were examined under numerous growth conditions. After a slightly longer lag time, growth of the homozygoustup1 mutant strain (BCa2-10) was virtually as rapid as the wild-type cells in rich YEPD media (doubling times of 64 and 58 min in log phase, respectively, as assayed by optical density at 600 nm). Growth of the mutant cells was arrested at 42°C but was normal at 37°C, whereas wild-type cells grew at both temperatures. No auxotrophies were detected, and growth on most carbon sources was similar. Growth of the strains on sucrose, glucose, galactose, and acetate was comparable.

One of the few metabolic phenotypes identified in tup1 C. albicans was a faster growth rate and accumulation to higher density on glycerol when compared to wild-type. During growth on glycerol and acetate, tup1 mutant cells exhibited the shortest cell length of all conditions tested. Short chains of stubby cells were typical, with poor cell-to-cell attachment marked by occasional single elongated cells. A formal notation, termed morphological index (Mi) has been developed to describe C. albicans cell shape (18). According to this system, where blastospores rank at 1 and true hyphae rank near 4, thetup1 mutant has values of 3.0 to 3.5 when grown on YEPD, and values of 1.5 to 2.5 when grown in minimal medium with glycerol.

The foregoing observations indicate that tup1 C. albicanshas several mutant phenotypes; some (temperature-sensitive growth, for example) are similar to those of tup1 mutants of S. cerevisiae. In most respects, however, the effects of aTUP1 deletion appeared different in the two species. SinceS. cerevisiae is capable of filamentous growth we also determined the effects of a TUP1 deletion on filamentous growth in S. cerevisiae.

Saccharomyces cerevisiae exhibits filamentous growth (exclusively in the form of pseudohyphae) in response to nitrogen starvation in diploid cells and in response to unknown inducers in haploid cells (19, 20). Saccharomyces cerevisiaestrains that do exhibit filamentous growth (21), anda/α diploid homozygous tup1 derivatives were constructed (22). The resulting strains exhibited typicaltup1 phenotypes, such as flocculence, temperature-sensitive growth, and an inability to sporulate; however, when grown on pseudohyphal growth-inducing media (SLAHD), they showed a marked reduction of pseudohyphal growth. Haploid cell types of S. cerevisiae show a different type of filamentous growth, termed invasive growth (20). Invasive growth was reduced in haploidtup1 S. cerevisiae strains derived from the diploids described above. The interpretation of these observations is complicated by the multiple defects of tup1 mutant strains, especially since TUP1 is required to maintain thea/α and α cell types of S. cerevisiae. However, TUP1 does not repress filamentous growth inS. cerevisiae as it does in C. albicans.

CPH1 is a transcriptional activator that positively regulates filamentous growth in C. albicans. Filamentous growth is reduced in cph1 cells under certain conditions (5). We constructed a double mutant to investigate the interaction of TUP1 and CPH1. Thistup1/tup1 cph1/cph1 strain was indistinguishable in its morphological characteristics from strains containing the tup1 mutations alone. The cph1mutant strains that were heterozygous for TUP1, however, resembled wild-type cells in their tendency to form filaments in some media, such as pH 7.0 Spider plates. That is, they were intermediate between cph1 mutants (which induce more poorly than wild type) and double tup1 cph1 mutants (which are constitutively induced for filamentous growth), suggesting that both genes participate in the same pathway.

If we assume that the C. albicans blastospore is the default state, the finding that deletion of TUP1 activates a filamentous morphology regardless of external conditions indicates thatTUP1 is a repressor of filamentous development. In the yeastS. cerevisiae, TUP1 is a transcriptional repressor and two lines of evidence indicate that TUP1 has the same molecular function in C. albicans: (i) the two proteins show a high degree of amino acid sequence conservation (Fig.3A) and (ii) expression of the C. albicans TUP1 gene fully complemented S. cerevisiae cells carrying a tup1deletion. On the basis of these results, we offer a simple model for the involvement of TUP1 in the blastospore to filamentous growth transition of C. albicans (Fig.5).

Figure 5

Model for control of filamentous growth in C. albicans by TUP1. Repression byTUP1 is regulated by environmental signals through a postulated DNA-binding protein. One regulator of this DNA-binding protein may be CPH1, which is placed upstream ofTUP1 based on epistasis of TUP1to CPH1. In the absence of TUP1repression, filamentous growth is constitutive and still responds to some environmental signals, suggesting the presence of both regulated and constitutive activators at genes controlled byTUP1.

According to this model, TUP1 represses genes whose expression is required to initiate or maintain filamentous growth. TUP1 is brought to the DNA upstream of these genes by postulated DNA-binding protein (or proteins) whose synthesis or activity is down-regulated by filamentous growth-inducing environmental conditions. In the absence ofTUP1, its target genes are always expressed, leading to constitutive filamentous growth. In this model the activator of filamentous growth, CPH1, is placed upstream ofTUP1 since the phenotypes of the double mutanttup1Δcph1Δ resemble those of the tup1Δmutant. More complicated models are also consistent with our observations. For example, TUP1 might participate in one of several redundant pathways through which genes required for filamentous growth can be turned on. Another possibility is that the absence ofTUP1 might alter cell physiology (via a stress response or metabolic defect, for example) in a general way that makes blastospore formation impossible. Although we believe that the specific morphological phenotype of the tup1Δ mutant cells as well as their general vigor argues for a direct role of TUP1 in regulating filamentous growth, these alternative models are formally possible. The fact that serum and pH activation of germ tube formation and hyphal growth were absent in the tup1Δ mutant cells suggests that TUP1 may also have a role in this pathway (23).

A comparison of the phenotypes produced by a TUP1 deletion suggests that TUP1 controls genes in C. albicansthat are different from those it controls in S. cerevisiae. For example, disruption of TUP1 in C. albicansresults in constitutive filamentous growth and enhanced growth in glycerol, two phenotypes not seen in a S. cerevisiae tup1Δstrain. Likewise, flocculence and defects in glucose repression are properties of S. cerevisiae tup1Δ strains, but are not seen in C. albicans tup1Δ strains. Given that the C. albicans gene complements a S. cerevisiae tup1Δmutant, it seems likely that TUP1 serves as a transcriptional repressor in both species and that interactions between the TUP1 gene product and the DNA-binding proteins to which it binds have been conserved. Therefore, it appears that the regulation of these DNA-binding proteins and the identity of the genes to which they bind have changed since S. cerevisiae and C. albicans diverged from a common ancestor.


View Abstract

Navigate This Article