C. albicans Gene Expression during Growth within an Immunosuppressed Mammalian Host
To initiate studies on C. albicans
genes required for infection or colonization, preliminary microarray analysis of gene expression in C. albicans
cells recovered from sites of oral infection was performed. WT C. albicans
cells were orally inoculated into germ-free newborn piglets as described [16
]. Following immunosuppression with cyclosporine A and methylprednisolone for 7–10 d, the animals developed dramatic thrush and esosphagitis, and C. albicans
plaques were visible in the esophagus at necropsy. C. albicans
cells were scraped from the tongues of two immunosuppressed, gnotobiotic (IGB) piglets and from the esophagus of one of them and were used for microarray analysis as described in Materials and Methods
. Because the majority of spots exhibited poor hybridization with samples from the oral infections, a comprehensive analysis of differential gene expression could not be performed. However, there were several candidate genes more highly expressed in samples taken from infected piglets than in reference laboratory-grown cells.
To confirm that candidate genes from the microarray studies exhibited increased expression in mammalian hosts compared to log phase laboratory-grown cells, quantitative real-time reverse transcriptase PCR (qRT-PCR) was performed. For the analysis, RNA was prepared from C. albicans
cells scraped from the tongue, esophagus, or roof of the mouth of two IGB piglets or recovered from the contents of the large intestine of one IGB piglet. cDNA prepared from the RNA was used as the template for qRT-PCR amplification as described in Materials and Methods
. Results were normalized using C. albicans ACT1
(encoding actin) and are shown relative to the reference sample of log phase, laboratory-grown cells that were used for the microarrays.
Seven genes that exhibited relatively high expression in C. albicans
cells recovered from the oral cavity or intestinal tract were identified, including EFH1
, and ECE1
(), which became the subjects of this study. EFH1
encodes a paralog of the well-studied transcription factor Efg1p [17
]. Efg1p regulates several morphological transitions as well as the expression of hyphal and virulence genes [14
]. However, the absence of Efh1p does not lead to a pronounced phenotype during laboratory growth [14
], and thus, its function is not understood. YHB5
(orf19.3710) encodes a flavohemoglobin; its paralog, YHB1
, is induced by growth in the presence of nitric oxide and is important for resistance to nitrosative stress [22
, by contrast, is not induced during laboratory growth in nitric oxide [22
, a gene that influences adhesion [24
], but whose function is poorly understood, is highly expressed by cells that are growing in the hyphal form (filamentous cells lacking constrictions at their septa) [21
is also highly expressed during invasion of host tissue [13
other genes expressed under laboratory conditions that resulted in increased ECE1
expression (hyphal growth and post-exponential phase, respectively) were analyzed. RBT1
, genes of poorly understood function that are expressed in hyphal cells [15
], and SOD3
(orf19.7111.1), encoding a manganese-containing superoxide dismutase that is expressed in post-exponential phase [26
], were studied. Finally, a control housekeeping gene, TEF1
(encoding translation elongation factor 1-alpha), was analyzed in some experiments.
The qRT-PCR results for the genes of interest are shown in . In C. albicans cells recovered from oral infection, five of the genes were highly expressed relative to their expression in log phase laboratory-grown cells (A). The sixth gene, EFH1, was not convincingly upregulated in cells recovered from oral lesions. In cells recovered from the piglet intestine, EFH1 and the five other genes showed relatively high expression (B, closed circles). Since only one piglet intestinal tract sample was studied, these initial data on gene expression in the intestinal tract were not definitive. Further studies in mice corroborated the results, as described below.
To compare expression during growth in the host with expression during growth in laboratory conditions that might mimic certain host parameters, cells were grown to a post-exponential phase (rich medium 37 °C, 3 d) or under conditions that resulted in formation of hyphae (serum-containing medium, 37 °C), and compared to log phase cells (rich medium 37 °C). Expression was measured by qRT-PCR as above. C shows that most of the genes of interest were more highly expressed in post-exponential phase than in log phase and all genes except for EFH1 were relatively highly expressed in hyphae. In contrast, expression of the housekeeping gene TEF1 was not markedly induced under these conditions. Thus, several genes were relatively highly expressed during growth within multiple tissues in the host and in either post-exponential phase or hyphal growth in the laboratory.
Gene Expression during Intestinal Colonization Is Not Dependent on Host Immunosuppression
For further studies of the interaction between C. albicans
and a mammalian host, we analyzed gene expression during colonization of immunocompetent Swiss Webster mice. Following oral inoculation of antibiotic-treated Swiss Webster mice, commensal colonization persisted for several weeks as previously described, e.g., [27
]. No symptoms of disease were observed in the colonized mice.
To determine whether the genes of interest were expressed during commensal intestinal colonization, gene expression was measured by qRT-PCR using RNA prepared from C. albicans cells recovered from the cecum or ileum of colonized mice. Expression normalized to ACT1 is expressed relative to the same reference sample as above. As shown in B (closed triangles), all six genes of interest were expressed at relatively high levels in C. albicans cells recovered from the mouse cecum in contrast to the housekeeping gene TEF1. Expression was very similar to the expression observed in C. albicans cells recovered from the IGB piglet intestinal tract. Therefore, the expression of these genes was not dependent on immunosuppression of the host. In addition, to compare expression in a different part of the intestinal tract, expression of three genes was characterized in cells recovered from the ileum. ECE1 and RBT4 were relatively highly expressed and RBT1 was slightly increased. In summary, five of the six genes were relatively highly expressed during growth within multiple tissues of immunosuppressed and immunocompetent hosts. EFH1 exhibited a distinct pattern of expression as it was relatively highly expressed in C. albicans cells recovered from the intestinal tract but not from sites of oral infection (esophagus or tongue lesions).
Uncharacterized Regulatory Pathways Control Gene Expression during Growth in the Intestinal Tract
Because most of the genes of interest were highly expressed in laboratory-grown hyphae (C), it was of interest to determine the morphology of the colonizing cells. C. albicans
cells expressing yeast-enhanced green fluorescent protein (yEGFP) [30
] were orally inoculated into mice by gavage. Three, seven, or 17 days post-inoculation, the contents of the ileum were recovered, and the C. albicans
cells were observed by fluorescence microscopy. In all samples, the vast majority of the fluorescent cells (>90% +/− 3% standard deviation) exhibited yeast cell morphology (). Because the cecum is anaerobic, it was not possible to visualize GFP in this organ, so phase contrast microscopy was used to visualize organisms. As with the ileum, the majority of fungal cells detected in cecum contents exhibited the yeast-cell morphology (71% +/− 10% standard error of the mean).
C. albicans Cells Colonizing the Murine Intestinal Tract Are Predominantly in the Yeast Form
Interestingly, cells recovered from the ileum expressed higher levels of ECE1 and RBT4 than laboratory-grown yeast cells and these levels were comparable to laboratory-grown hyphal cells. This observation suggested that the yeast-form cells recovered from the ileum, rather than the rare hyphal-form cells, were expressing high levels of ECE1 and RBT4. Therefore, gene expression during growth within the host differed from the characterized expression of ECE1 and RBT4 under laboratory conditions.
To compare gene expression during laboratory growth and host growth further, the role of the morphogenesis-regulating transcription factor Efg1p was studied. In the laboratory, expression of ECE1
, and RBT4
in hyphae is dependent on Efg1p (D and [31
]). Expression is similarly reduced in a double mutant lacking a second partially redundant transcription factor, Cph1p [33
] and Efg1p (unpublished data and [32
]). Therefore, the effect of deletion of both EFG1
on expression of these three genes during growth in the intestinal tract was studied. Strain CKY138 (efg1 cph1
double null mutant) was orally inoculated into mice by gavage. RNA was prepared from cells recovered from the intestinal tract and expression of ECE1
, and RBT4
was measured by qRT-PCR. Unexpectedly, double null mutant C. albicans
cells recovered from the intestinal tract expressed ECE1
, and RBT4
(B). Thus, there was substantial Efg1-, Cph1p-independent expression of ECE1
, and RBT4
during growth within the intestinal tract. These observations demonstrate that an uncharacterized regulatory pathway(s) controls the expression of these genes during growth in the intestinal tract, underscoring the importance of studying C. albicans
physiology during growth within a host.
Expression of EFH1 Results in Reduced Colonization of the Intestinal Tract
Five of the six genes of interest were expressed at elevated levels in all samples recovered from the host, while EFH1 expression was elevated in cells recovered from the intestinal tract but not from the tongue or esophagus. To determine whether EFH1 performs a function during colonization of the intestinal tract, efh1 null mutants and EFH1 reconstituted strains were constructed.
As previously reported [14
], deletion of EFH1
did not result in defects in growth under laboratory conditions. Hyphal formation by the mutant strain was also normal and the efh1
null mutant was indistinguishable from WT in post-exponential phase stress resistance (heat shock at 55 °C or menadione treatment; unpublished data). The mutant strain grew well anaerobically in defined medium [34
] (I. Soltero and C. A. Kumamoto, unpublished data). Therefore, as noted previously [14
], the efh1
deletion appeared to have minimal effects on the physiology of laboratory-grown cells.
When WT C. albicans, efh1 null mutant, or EFH1 reconstituted null mutant cells were orally inoculated by gavage into immunocompetent mice, intestinal tract colonization was established. As shown in , WT C. albicans colonized the intestinal tract and was detectable in the contents of the cecum up to 3 wk post-inoculation (A, closed circles). Colonization was also detectable in the ileum, stomach, and fecal pellets (B and C); analysis of these samples from three individual mice revealed a correlation in colonization levels in different organs (C).
Deletion of EFH1 Alters Murine Intestinal Colonization
To determine the effect of coprophagy on colonization, an uninoculated mouse was introduced into a cage with three to four inoculated mice, 24 h post-inoculation. Fecal pellets from the uninoculated mouse were initially positive for C. albicans
, but the levels of C. albicans
declined very rapidly, and 8 d post-introduction into the cage were at least 50-fold below the geometric mean for the inoculated mice. Therefore, consistent with previous studies [29
], the contribution to C. albicans
titers due to coprophagy was minimal at later time points.
Surprisingly, two independently isolated efh1 null mutants colonized the murine intestinal tract at higher levels (A, blue triangles) than WT C. albicans (A, red circles). In the cecum at day 21 post-inoculation, the geometric mean for the efh1 null mutant was 100-fold higher than that for WT C. albicans (p < 0.000005 by t test). For both the efh1 null mutant and WT strains, C. albicans associated with the cecum wall represented a small fraction of the total cecum-associated C. albicans. Analysis of C. albicans in fecal pellets demonstrated that, initially, colonization by the mutant was similar to colonization by WT C. albicans (B) but at later time points, e.g., day 21 post-inoculation, colonization by the efh1 null mutant persisted at higher levels than the WT strain (B, p < 0.0003 by t test). Enhanced colonization was also observed in the ileum and stomach (C and unpublished data).
As a control to show that the differences in colony forming units (CFUs) recovered from WT or efh1 mutant–containing ceca truly reflected differences in the numbers of fungal cells rather than preferential recovery of mutant cells, an alternative method of quantification was used. DNA was extracted from cecum contents and the amount of C. albicans genomic DNA was determined by quantitative PCR (qPCR). The results showed that the amounts of C. albicans DNA recovered from cecum contents containing WT or efh1 null mutant cells (in arbitrary units) correlated with the CFUs determined by plating (D), and the average ratio of the amount of DNA/CFU was the same for the WT and efh1 null mutant strains. Therefore, differences in CFUs reflected differences in the numbers of colonizing C. albicans.
To demonstrate that the enhanced colonization reflected the absence of Efh1p, EFH1
was added back to the deletion mutant under control of a strong promoter to ensure expression of the reintroduced gene. Introduction of the ectopically controlled EFH1
to the mutant strain resulted in reduced colonization (A, black diamonds), indicating that the level of EFH1
in the strain determined colonization levels. For unknown reasons, introduction of EFH1
at its native locus did not result in full complementation (unpublished data), as has been observed by others for other genes, e.g., [15
]. These studies demonstrate that in the absence of Efh1p, colonization of the intestinal tract was enhanced. Thus, paradoxically, expression of C. albicans EFH1
during commensal colonization of the intestinal tract resulted in reduced colonization.
To observe the morphology of efh1 null mutant cells, GFP-expressing efh1 null mutants were orally inoculated into mice by gavage. On day 3 or day 13 post-inoculation, ileum contents were collected and organisms were visualized by observing green fluorescence. As observed for WT C. albicans, the vast majority (>91% +/− 1% standard deviation) of cells exhibited yeast-form morphology ().
To detect dissemination from the intestinal tract in mice, the kidneys, liver, and spleen were homogenized and cultured. With both efh1 null mutant and WT, all samples were either negative or contained very few organisms (). Therefore, there was no evidence of high-level colonization of deep organs, indicating that C. albicans was not escaping from the intestinal tract.
Colonization of the tongue of mice was analyzed to determine whether a significant level of oral candidiasis was occurring. Colonization was generally not detectable, although an occasional mouse exhibited below 103 CFU/gm tongue tissue (). No consistent differences in colonization by WT and mutant organisms were observed. Therefore, there was no evidence of either mucosal or systemic disease in these immunocompetent mice.
To determine the levels of residual bacteria remaining after antibiotic treatment, ileum and cecum homogenates were cultured under aerobic or anaerobic conditions on rich media. The ranges of bacteria levels were very similar for uninoculated mice and for mice inoculated with all of the C. albicans strains described above (unpublished data).
EFH1 Overexpression Reduces Intestinal Colonization
Since reduction of Efh1p by deletion resulted in enhanced colonization, the effect of increased expression of Efh1p was tested using an EFH1+
strain carrying a third copy of EFH1
expressed from the strong ADH1
promoter. The growth rate of the EFH1
overexpressing strain was close to WT during laboratory growth in CM medium at 37 °C (WT doubling time 78 min; overexpressing strain, 82 min), and high levels of EFH1
transcript were produced under these conditions (unpublished data). In rich medium, the overexpressing strain produced yeast-form cells while in certain minimal media the mutant produced pseudohyphae, consistent with previous results [14
When the overexpressing mutant was orally inoculated into mice by gavage, colonization was initially similar to that of WT C. albicans (). Subsequently, however, colonization by the overexpressing strain declined more rapidly than colonization by the WT strain, and at days 18 and 21 post-inoculation, the geometric mean for the overexpressing strain was more than 100-fold lower than that of the WT strain (p < 0.002 using t test). This result is consistent with that of the EFH1 reconstituted null mutant (). Therefore, high expression of EFH1 attenuated colonization, while deletion of EFH1 resulted in enhanced colonization, demonstrating that Efh1p is a regulator of the level of colonization during growth of C. albicans in the murine intestinal tract.
Attenuated Colonization of Mice by an EFH1 Overexpressing Strain
Effects of EFH1 on Disease
In piglet oral lesions, expression of EFH1 was low relative to its expression in cells growing within the intestinal tract. These findings suggest that when infection occurs, the negative effects of EFH1 on host colonization are relieved by lowering EFH1 expression. Therefore, ectopic EFH1 overexpression might reduce the ability of C. albicans to cause disease. To test this model, the behavior of an EFH1 overexpressing strain was analyzed in two different animal models.
In immunocompromised patients, C. albicans
causes mucosal infections, such as OPC. Paralleling the susceptibility of AIDS patients to OPC, mice lacking T cells show an enhanced susceptibility to oral colonization by C. albicans
]. To determine whether EFH1
overexpression would influence colonization in the oral cavities of immunocompromised mice, competition experiments were performed, in which mixtures of EFH1
overexpressing and genetically marked WT C. albicans
cells were inoculated directly into the oral cavities of athymic mice by swabbing. The marked WT strain carried the SAT1
nourseothricin resistance gene [36
] under control of the maltase promoter. To monitor the ability of the two strains to persist in the oral cavity relative to one another, the oral cavities were swabbed at various times post-inoculation, and the ratio of the two strains was determined by replica plating. The competitive index (CI) was determined by dividing the ratio of NouS
strains at a particular time point by the ratio in the inoculum.
When unmarked WT and marked WT strains were mixed and inoculated, the geometric means for the CI remained above 1 throughout the time course, indicating a slight competitive advantage for the unmarked WT strain (). For the competition between the NouS EFH1 overexpressing strain and the NouR WT strain, the CI was close to 1 on day 1 post-inoculation, but declined thereafter, demonstrating that the EFH1 overexpressing strain exhibited a competitive disadvantage relative to the WT strain (). The difference in CI for WT and EFH1 overexpressing strain was statistically significant (p < 0.01 by t test). At longer times, few colonies were obtained, precluding the accurate measurement of ratios. Following 24 h of laboratory growth in CM medium at 37 °C, the CI for the overexpressing strain relative to the marked WT was 1.1 (average of two determinations), showing a similar growth rate for the strains under these conditions. Therefore, these data show that forced high expression of EFH1 reduces colonization in the oral cavity of immunodeficient mice. In immunocompetent BALB/c mice, oral colonization was rapidly lost (unpublished data), and it was not possible to measure ratios.
EFH1 Overexpression Reduces Colonization of the Oral Cavity in Immunodeficient Mice
To determine whether EFH1 was important in other host niches, the ability of the efh1 null mutant and EFH1 overexpressing strain to cause lethal infection in a disseminated candidiasis model was tested. When inoculated intravenously in mice, both the efh1 null mutant and the EFH1 overexpressing strain retained the ability to cause lethal infections. The mean survival time for mice inoculated with WT C. albicans was 5 ± 3 d (23 mice). For the efh1 null mutant, the mean survival time was 5 ± 3 d (16 mice) and for the EFH1 overexpressing strain, the mean survival time was 7 ± 3 d (seven mice) (composite results from at least two experiments). Therefore, both the efh1 null mutant and the EFH1 overexpressing strain were virulent in this model.
rbt1 and rbt4 Mutants Colonize the Murine Intestinal Tract at WT Levels while the ece1 Mutant Is Attenuated
The ability of mutants lacking some of the other genes of interest to colonize the intestinal tract of mice was analyzed. The rbt1 null mutant (A), rbt4 null mutant (A), and yhb1 yhb5 double null mutant (B) colonized the intestinal tract at close to WT levels. In contrast, the ece1 mutant exhibited an attenuated colonization phenotype, which was reversed when the WT ECE1 gene was added back to the mutant strain (C). Therefore, two of the six genes tested in this study that were relatively highly expressed during growth in the intestinal tract, EFH1 and ECE1, influenced the ability of C. albicans to colonize.
rbt1 and rbt4 Mutants Colonize the Murine Intestinal Tract at WT Levels