Experimental design and methodology.
Following the complete sequence determination of the S. cerevisiae
genome, Affymetrix DNA microarrays have emerged as a powerful tool for examining the simultaneous expression pattern of more than 6,000 yeast genes (10
). Because of the scope of each data set obtained from a microarray hybridization, treatments used in this study were limited to a single concentration at a single time point. Rich medium (YPD) was used so as not to limit the growth rate, and one doubling time (90 min) was selected for the duration of exposure to agent at 0.5 times the MIC. Cells were exposed to agents and harvested during the mid-logarithmic-growth phase.
RNA was prepared following exposure of the parent strain, BY4743, to one of five previously characterized azoles (clotrimazole, fluconazole, itraconazole, ketoconazole, and voriconazole), a novel imidazole, PNU-144248E (shown in Fig. ), an allylamine (terbinafine), or a morpholine (amorolfine) as described in Materials and Methods. In addition, RNA was prepared from three strains each bearing a homozygous deletion of ERG2, ERG5, or ERG6 and from two untreated control cultures. Figure shows the genes involved in the biosynthesis of ergosterol from farnesyl pyrophosphate in S. cerevisiae.
FIG. 1 Treatments used in the study and their relationships to ergosterol biosynthesis. Gene names are as listed in reference 9). Genes in deletion strains are boldfaced and boxed. Arrows point to the sites of action of the antifungal agents. The structure of (more ...) Identification of gene expression patterns.
Each Affymetrix yeast genome set represents 6,593 ORFs, including 172 control genes. Thus, the 11 data sets from the treatments comprising this study represent 72,523 data points. Hybridization intensity values are expressed in ADI units, as described in Materials and Methods. Following normalization to account for variations in chip intensity (described in Materials and Methods), a filter was applied requiring the experimental ADI value to be above 50 ADI units, thereby limiting the data to 29,428 points. Use of a second filter requiring the baseline ADI value to be greater than 50 ADI units further limited the data to 20,697 points. The ADI values used in the filters were chosen after examination of the background ADI value calculated by the GeneChip software and the ADI values for selected unexpressed genes (e.g., haploid-specific genes in diploid cells [data not shown]); with both approaches, a value of up to ~25 ADI units corresponded to no signal.
Genes for which ADI ratios changed beyond 1 standard deviation were considered to be responsive to the treatment. By this criterion, 1,154 genes responded with increased mRNA levels in at least one treatment; 1,358 genes responded with decreased mRNA levels relative to the baseline. To distinguish gene responses to perturbations in ergosterol biosynthesis from other transcriptional changes, genes responding in at least 5 of the 11 experimental treatments were considered in the subsequent analysis. The intention of selecting by these criteria is to identify genes that have a convergent pattern of expression across many individual treatments, which may be indicative of a common response.
A total of 156 genes showed significant increases in transcript levels in five or more treatments, and 78 showed significantly decreased transcript levels in five or more treatments. These were annotated using the biological roles assigned by the Yeast Protein Database (YPD) (18
). The number and characteristics of the responsive genes grouped according to biological role are shown in Table . The category with the largest number of responses (hits) is “unknown” group of 53 genes. Next most abundant are the 36 responsive mitochondrial genes, followed by 22 genes involved in biosynthesis of lipids, fatty acids, and sterols. This group includes nine genes in the ergosterol pathway. The category “other related genes” refers to genes that are related to ergosterol perturbation by other experimental results; these genes are described in Tables and and Fig. and . The category “other genes” includes responsive genes each of which was the sole representative of a particular biological pathway, and whose relationship to ergosterol perturbation could not be discerned.
TABLE 1 Genes responding in five or moretreatments
TABLE 2 Genes encoding products associated with cell envelope structure orfunction
TABLE 3 Genes encoding mitochondrialproteins
FIG. 2 Genes involved in the biosynthesis of ergosterol and membrane components. Boldfaced genes were responsive in the study; boldface italics indicate genes with decreased transcript levels. Superscripts indicate the number of treatments to which the gene (more ...)
FIG. 3 Genes involved in the biosynthesis and utilization of heme. Boldface, italics, and superscripts are as described for Fig. . Lists of genes were obtained from the YPD (18).
Genes were also categorized by the relative number of responses due to chemical versus genetic perturbation. The percentage of the response that was due to chemical perturbation ranged from highs of 87 to 88%, in the case of membrane-associated proteins, to a low of 62% for the class of lipid-, fatty acid-, or sterol-related genes. Analysis using a k-means algorithm did not reveal transcriptional patterns associated with particular treatment classes (data not shown). Interestingly, the ergosterol genes were highly responsive to genetic disruption of ERG2, ERG5, and ERG6. The 10 responsive ergosterol genes had increased transcript levels in all three mutant strains (Fig. ) with the exception of the particular gene disrupted in each strain.
As a means to assess the specificity of the ergosterol response identified in this study, the behavior of these responsive genes was examined following the transition from logarithmic growth to saturation. Table shows the number of responsive genes in each category whose transcript levels change as the cells enter saturated growth and undergo the diauxic shift (10
; G. F. Bammert and J. M. Fostel, unpublished data). The proportion of responsive genes which are altered in response to saturated growth varies from 0% of genes in several pathways to highs of 78 and 83% of ergosterol-responsive genes related to amino acid and carbohydrate metabolism, respectively. This suggests that some groups of transcripts may be responding to alterations in the metabolic rate rather than to ergosterol perturbation directly.
Responsive genes in the ergosterol pathway.
Figure shows the genes involved in the biosynthesis of ergosterol from acyl coenzyme A (acyl-CoA), highlighting some of the responsive genes identified above. Most notably, nine genes in the ergosterol pathway responded with increased transcript levels to the conditions used in this study. The ergosterol pathway represented the highest proportion of responsive genes identified, in agreement with previous studies showing that this pathway is the target of azoles and is responsive to modulations of the ergosterol level.
The biosynthesis of ergosterol involves the coordination of many factors by which the cell regulates the synthesis of this essential component. ERG19
is reported to contribute to the regulation of flux through the mevalonate pathway (3
) and is increased in response to perturbations here. ERG3
expression increases following treatment with antifungal agents (43
) and in strains with mutations in ERG2
, and ERG6
). These data are supported by the observations reported here: ERG3
expression increased following drug treatment and in the three ERG
deletion strains. Additionally, expression of NCP1
, which encodes NADP-cytochrome P450 reductase and the electron donor for squalene epoxidase, lanosterol 14α-demethylase, and sterol C-22 desaturase (45
), increases fivefold in a strain constitutively overexpressing ERG11
), consistent with coordinate increases in the transcript levels of these two genes in this study.
Measurements using promoter fusions in various genetic backgrounds found a number of genes affecting ERG9
). Two treatments overlapped those reported here: exposure to ketoconazole and expression in an ERG2
disruptant. In both cases ERG9
expression increased (23
). While ERG9
did not meet the criterion of change in five or more treatments used here, levels of the ERG9
transcript did increase in four treatments: ERG2
knockouts and exposure to fluconazole or PNU-144248E. The difference in the ketoconazole treatments used in the two studies (18 h the study reported in reference 23
versus 90 min here) may account for the difference, as observed in Fig. for ERG25
FIG. 5 Expression time course following exposure to ketoconazole. (A) Culture OD at the times sampled (solid bars, untreated culture; heavily shaded bars, 4 μM ketoconazole; open bars, 8 μM ketoconazole; light shaded bars, 16 μM ketoconazole) (more ...)
Another gene in the pathway, ERG1
, did not change significantly in any of the treatments used here, including terbinafine, although Erg1p enzyme activity does increase following terbinafine exposure (28
). Erg1p activity responds to oxygen and sterol limitation (34
) and appears to be regulated by protein localization or other factors (28
). Taken together, these observations suggest that changes in Erg1p activity arise from posttranscriptional regulation.
In a study using promoter fusions as a readout of transcriptional changes, Dimster-Denk et al. (11
) observed 19 genes with a change in expression greater than 1.5-fold following 21
h of exposure to fluconazole relative to expression levels in untreated cells. Of these 19 responsive genes, 7 were found to be responsive in this study (ERG2
, and ERG19
) and ERG9
expression was changed in four treatments. The other 11 genes had transcript levels too low to be measured reliably, including 5 mating-response genes not expressed in the diploid strain used here (ERG8
[also referred to as CAT5
, and STE2
). Responses by ERG24
were seen on the microarray but were not reported by Dimster-Denk et al. Thus, the two methods are in agreement on the identification of responsive genes among genes detected by both. The change in transcript level was not always in the same direction in the two studies, however. Decreases in ERG3
, and ERG11
transcript levels at 21 h were reported by Dimster-Denk et al. (11
), while these genes showed increased transcript levels following 90 min of exposure as measured by microarrays. It is more likely that this reflects differences between the biological responses of these genes at the different exposure times than a discrepancy between the methods used to measure expression.
Other responsive genes related to lipid, fatty acid, and sterol biosynthesis.
In addition to participating in cell membranes, esterified ergosterol is found in lipid particles, which may serve as storage reservoirs or as intermediates in intracellular transport (53
is one of the two genes responsible for the esterification of ergosterol, the final step in the pathway leading to ergosteryl, and was responsive in this study. Another responsive gene, CYB5
(encoding cytochrome b5
), was identified by an ability to overcome ketoconazole hypersensitivity when overexpressed in an erg11 background (47
) and may help protect the organism from azole exposure. ACH1
(encoding acetyl-CoA hydrolase) may be responding to an overproduction of sterol intermediates formed during inhibition of the pathway. FAS1
(encoding fatty-acyl-CoA synthase) deletion strains have reduced levels of ergosterol esters and sphingolipids, indicating a possible role in lipid biosynthesis and metabolism (9
encodes serine C
-palmitoyltransferase, the first enzyme involved with the biosynthesis of the long-chain base component of sphingolipids. An increase in the level of transcripts of this transferase following perturbation of the ergosterol pathway suggests an interaction between the ergosterol and sphingolipid biosynthetic pathways in yeast.
Transcript levels of several other genes involving lipid, fatty acid, and sterol metabolism decreased in this study. The SUR2
product hydroxylates the sphingoid C-4 of ceramide (15
), and the decrease observed in the level of this transcript further suggests an interaction between the sphingolipid and sterol pathways. ELO1
encodes an enzyme responsible for the elongation of fatty acids. A decline in the level of this transcript under the conditions of this study may indicate a compensatory response in cellular fatty acid content to limited sterols following perturbation of the ergosterol pathway. OLE1
, encoding δ-9 desaturase, is needed for the formation of unsaturated fatty acids and also shows a decreased transcript level here. OLE1
is repressed by the presence of saturated fatty acids (12
); thus, the decline in OLE1
transcript levels may indicate an increase in saturated fatty acid levels, possibly another compensatory response to altered ergosterol. Interestingly, the fatty-acid-responsive repression of OLE1
is mediated through the FAA1
), and the level of FAA4
transcripts decreased here. A connection of fatty acids to perturbation of the ergosterol pathway may suggest a restructuring of the cell membrane in response to reduced ergosterol levels.
Other responsive pathways.
While ergosterol is found throughout the cell membranes, it is most abundant in the plasma membrane and secretory vesicles and is important for mitochondrial respiration (9
). Depletion of ergosterol with concomitant accumulation of sterol intermediates can result in alterations in membrane functions, synthesis and activity of membrane-bound enzymes, and mitochondrial activities, as well as in uncoordinated behavior of the yeast cell (36
). The changes in transcript pattern reported in Table may reflect these stresses.
It can be predicted that perturbations of ergosterol levels within the cell may affect the functioning of mitochondrial enzymes. This pattern does indeed emerge with the increase in transcript levels of several components of the mitochondrial electron transport system, as shown in Table . Transcript levels of four members of the cytochrome c oxidase complex, COX4, COX5A, COX8, and COX13, were increased. Similarly, transcript levels increased for four members of the cytochrome c reductase complex, RIP1, CYT1, QCR6, and COR1, and five genes encoding subunits of ATP synthase, ATP1, ATP14, ATP15, ATP16, and INH1. Other genes involved in energy generation that showed increased transcript levels are listed in Table , as are a number of other responsive genes encoding mitochondrial proteins.
Levels of transcripts from four of the five members of the hypoxic gene family (ANB1
, and HEM13
) are reduced in response to treatments used in this study (Table ). Two of these genes, CYC7
, encode the hypoxic isoforms of cytochrome c
and cytochrome c
oxidase. Since their expression levels depend on the level of available oxygen (6
), the decrease in transcript levels of these anaerobiosis-induced genes may be occurring in response to increased levels of intracellular oxygen. Another line of evidence to support this hypothesis is the observation that transcripts from three genes involved with oxidative-stress response (GRE2
, and SOD2
) are increased (Table ). In the course of normal catalysis, cytochrome c
oxidase transfers two electrons to molecular oxygen, and incomplete reactions can result in the formation of reactive oxygen species. Machida et al. (32
) have shown that exposure to farnesol results in the generation of reactive oxygen species through an indirect effect on mitochondrial electron transport. Farnesol can be generated from farnesyl pyrophosphate, a precursor to sterols which may accumulate if late stages of ergosterol biosynthesis are inhibited. All these observations are consistent with increased oxidative stress following perturbation of ergosterol biosynthesis.
TABLE 4 Genes responding to oxygenstress
Perturbation of mitochondrial electron transport could arise from a decrease in ergosterol levels in the inner membrane due to interference with ergosterol biosynthesis or from a direct interaction between the chemical agents used and the mitochondrial enzyme complexes. For example, metal-chelating drugs block reduction at the ubiquinol oxidation site of the cytochrome c
), and azoles inhibit their target, lanosterol demethylase, through an interaction with the heme iron component of the complex (54
). Since more than half of the responsive mitochondrial genes are found to respond both to genetic alterations and to drug treatment, it is likely that the effect is mediated through ergosterol biosynthesis and does not arise solely as a direct consequence of drug action.
Interestingly, one drug-specific pattern was observed. Transcript levels of HEM1, the first gene involved in the biosynthesis of heme, increased in the eight drug treatments but not in the three deletions. This suggests a compensatory response to drug treatment not represented in the mutant strains. Transcript levels of HEM13
, the rate-limiting step of heme biosynthesis (57
), were reduced in response to seven drug treatments and, again, in none of the deletion strains. HEM13
is repressed by heme and oxygen (1
), raising the possibility that levels of one or both are elevated following treatment. This is consistent with the hypothesis of increased intracellular oxygen suggested by the responses of anaerobiosis-induced and oxygen stress genes described above. Heme also plays a role in sensing intracellular oxygen levels (57
Heme plays a central role in sterol synthesis and regulates the transcription of several genes involved with this process (37
). The accumulation of 5-aminolevulenic acid, the product of Hem1p, derepresses 3-hydroxy-3-methyl-glytaryl CoA reductase, leading to increased levels of 2,3-oxidosqualene (31
). Heme is required for the enzymatic activities of Erg3p (C-5 sterol desaturase) and Erg5p (C-22,23 desaturase) (36
). Erg11p also contains heme and shows heme-regulated expression (48
). Other genes regulated by heme are listed in Fig. . In the present study, levels of transcripts from four heme-induced genes are increased and levels of transcripts from three heme-repressed genes are decreased, consistent with induction of heme-regulated expression, and suggesting elevated intracellular heme levels following treatment, particularly treatment with chemical agents.
Correlation of the expression pattern of a novel azole to other conditions.
Included in this analysis was the expression profile in response to treatment with a compound containing an azole moiety. PNU-144248E contains an imidazole ring, yet is structurally distinct from the other azoles tested. The rationale for including it in this study was to determine if exposure to this compound would result in a pattern of expression similar to that seen in response to treatment using azoles with known biochemical functions, thereby suggesting a similar mode of action. This information could then be used to assess the predictive ability of expression profiles observed in response to an agent with an unknown mode of action. Of the 156 genes with increased transcript levels following ergosterol perturbation, 144, or 92%, also had increased transcript levels in response to treatment with PNU-144248E. All of the ergosterol, lipid, fatty-acid, and sterol metabolism genes and 17 of the 19 genes involved with energy generation were included (the exceptions were QCR6 and COX13). Twenty-nine of the 34 unknowns were also included. From the set of 78 genes with decreased transcript levels following ergosterol perturbations, 40, or 51%, also had decreased transcript levels following treatment with PNU-144248E. These data suggest that PNU-144248E does indeed behave as an azole as measured by cellular responses at the level of gene transcription.
transcript levels increased following treatment with PNU-144248E. This gene encodes a member of the multidrug resistance pump family homologous to genes which are upregulated in azole-resistant isolates of C. albicans
). It might be anticipated that PDR5
transcript levels increase in all of the drug treatments; however, they were unaltered in other treatments used in this study. It is possible that PDR5
requires a longer exposure time for full induction and that PNU-144248E was the only treatment of sufficient potency to elicit a response in the 90 min of exposure used here. PDR5
transcript levels in cells treated with mucidin did not increase until 2 h of exposure (35
), and the response to ergosterol perturbation may follow similar kinetics.
Comparison of microarray measurements with PCR measures of transcript levels.
Microarrays represent a new technology for measuring transcript levels. Control experiments indicate that measures using this technique are quantitative; however, these experiments were carried out for a limited subset of the genes on the array (52
). It was of interest to compare microarray ADI measurements for responsive genes to measurements of the same RNA preparation using the Taqman quantitative PCR system. This was performed for two genes of interest, ERG25
(Fig. ), one showing an increase in response to treatment and the other a decrease.
FIG. 4 Comparison of expression measurements by RT-PCR and microarrays. Each data point represents the measurement of a given gene by each method under a specific treatment, as indicated by the label. Squares, ERG25; circles, YER067W; solid symbols, untreated (more ...)
Data from the experiments shown in Fig. were normalized as described in Materials and Methods. Microarray ADI measures are proportional to transcript abundance, while PCR measures of Ct are inversely proportional, i.e., more-abundant transcripts require fewer cycles to achieve detectable levels, and hence the Ct is lower than for less-abundant material. For this reason the correlation seen in Fig. suggests good agreement between measures of transcript levels by these two methods.
Changes in transcript levels over time.
A single doubling time was selected for the duration of treatment with chemical agents in this study. It is likely that different transcripts may be responsive at earlier or later times and that the responsive transcripts detected herein may peak at times other than 90 min. In order to assess the response kinetics for genes of interest, measures of transcripts from responsive genes were made following different times of exposure to several of the agents used in this study. ERG25
is a late-stage transcript in the ergosterol pathway and has previously been shown to be responsive to ergosterol levels (29
are uncharacterized ORFs which respond to ergosterol perturbation but do not change in many other treatments tested (data not shown). YER067W
is a ploidy-regulated gene (13
is predicted to be a glycosylphosphatidylinositol (GPI)-anchored protein (16
) with weak homology to the potential cell wall stress sensor Mid2p (18
). Microarray measurements show that both ERG25
transcripts decrease under conditions where the growth rate is slowed, for example, entry into stationary phase, while YER067W
transcript levels increase under these conditions (data not shown). This is opposite to the direction of response to ergosterol perturbation, where microarray data show that YER067W
levels increase and ERG25
Figure shows the response kinetics for these three genes during a 24-h exposure to ketoconazole. Panel A shows the growth of the cultures used in this experiment both by OD and by the concomitant change in the TEF1 level as the cultures become saturated between 5 and 24 h. Panels B, C, and D show the PCR measures of ERG25, YER067W, and YNL300W RNA, respectively, from cells exposed to ketoconazole at three different doses. All show the TEF1 measure and the experimental transcript. In all cases, TEF1 is relatively unchanged by treatment and serves to normalize slight differences in the total RNA in each reaction tube during logarithmic growth (0, 60, 90, 180, and 300 min). In contrast, exposure to differing concentrations of ketoconazole results in an almost dose-dependent response in the three transcripts measured, and the direction of change is opposite that seen during growth saturation.
One interesting exception is the ERG25 response. The initial microarray measure found a 1.8-fold increase following a 90-min treatment with ketoconazole, and this was confirmed by PCR (Fig. ). In the time course experiment shown in Fig. , however, ERG25 levels were not observed to increase until the 3-h point, when they were 2.5 cycles or approximately 5.6-fold increased relative to levels in untreated cells. This most likely reflects a difference in sensitivity (a 1.8-fold change is within one PCR cycle) or small differences between the cultures used in the two experiments. This experiment illustrates the different strengths of the two methods used. Microarrays reveal the transcript profile at a particular time and state and can be used to identify all transcripts responding under the conditions of the study. Quantitative PCR probes can then be generated to follow the kinetics of expression of transcripts of interest, allowing easy measurement of their response to a variety of treatments.
Genome-wide transcriptional changes in S. cerevisiae observed in response to treatment of the cells with chemical agents correlate with responses to genetic alterations in the same biosynthetic pathway. A number of responsive genes which encode products with functions impinging on ergosterol biosynthesis or products related to membrane structure and function were identified. This supports the interpretation of expression profiles to define the mode of action of a drug. In addition to changes in transcript level relating directly to ergosterol biosynthesis, expression changes suggestive of interruption of heme biosynthesis and increased intracellular oxygen tension were also observed, indicating additional effects perturbing the ergosterol pathway. The approach used to identify genes responsive to the treatments studied does not rely on prior understanding of the biological effects of the treatments. Thus, it is possible to contemplate applying this method to predict the mode of action of novel agents with antifungal activity. The novel azole PNU-144248E has provided a validation of this method, causing transcriptional changes with a high degree of correlation to those seen following treatment with other azoles with known biological effects. This study has outlined a method to identify genes of interest common to a particular cellular response that will be of utility in the future study of novel antifungal compounds.