1. DNA microarrays for gene expression profiling of C. albicans
To construct DNA microarrays of C. albicans, the QIAGEN Operon 70mer
probe sets (Array-Ready Genome Oligo Set and Candida Genome AROS Upgrade Set) will be used. This oligo set contains 7,925 optimized probes that represent the entire genome of C. albicans and 10 different controls. The oligo set has been successfully used in studies of C. albicans biology and pathogenesis (Cao et al., 2006; Chen et al., 2004; Magee et al., 2003; Zhao et al., 2005).
a. RNA isolation, sample labeling and microarray hybridization.
Recognizing that one of the most important aspects of microarray analysis is the source and quality of the RNA used in these experiments, we will use standardization of protocols for RNA isolation as previously described (Lan et al., 2002; Lan et al., 2004). In general, cells grown at different conditions will be collected throughout lag, log and stationary phases of growth. RNA isolation, sample labeling and microarray hybridization will be performed using established protocols
(http://microarrays.org). Briefly, cells are harvested by centrifugation immediately after sampling; pellets are either snap-frozen in dry ice/ethanol or extracted in the presence of 15% SDS and buffered phenol:chloroform (1:1). Total RNA is
precipitated with absolute ETOH. After centrifugation, the RNA pellet is air dried and suspended in DEPC-treated water. For labeling, cDNA containing a T7 RNA polymerase recognition sequence are prepared from total RNA amplified with T7 RNA polymerase. After processing, blunt-ended cDNAs are used as templates to produce antisense RNA using a T7 Ampliscribe kit from Epicentre Technologies (Phillips and Eberwine, 1996). A second round of double-stranded cDNA synthesis is performed in the presence of random hexamer primers, reverse transcriptase and aminoallyl-dUTP. Cy3-/Cy5- dyes are incorporated into single-stranded cDNAs with monofunctional NHS esters that bind to free amino groups. Un-incorporated dye is removed by ultrafiltration using a Centricon 30 unit (Amicon). DNA microarrays will be analyzed using GenePix 4000B scanner and GenePix Pro 6.0 software (Molecular Device).
b. Experimental design and statistical analysis for expression profiling.
To derive accurate signals for the gene expression profiling, the first issue is the proper experimental design. Since the main purpose of this proposal is to understand the signaling/regulatory network involved in C. albicans response to environmental stimuli, especially the host environments, the number of time points measured is
crucial to the characterization for the upstream and downstream components. For most studies, at least five time points will be measured. Some effort will also be made in meta-analysis of the existing data to understand the variation and to decide the number of replicates needed for the expected significance level. At least five replicates are considered at this point and the number is subjected to change. With the high cost of microarray experiments, properly arranging the samples hybridized together with loop design in oligo spotted arrays can reduce the arrays needed without loss of too much information (Churchill, 2002). If not enough samples or arrays are available, pooling the samples is considered as a choice to get more reliable signals (Kendziorski et al., 2005).
Microarray data are known to be very noisy and no single statistical method has been recognized as standard approach for the normalization. The data derived will be analyzed with several different statistically solid algorithms. Image segmentation resultsfrom GenePix™,Spot(Jain et al., 2002), and model-based approaches will be compared and the one with the most reproducibility of replicates will be chosen for the next step. For the spotted oligo arrays proposed in this proposal, we will use ANOVA model (Wolfinger et al., 2001) to adjust for systematic noises. Also, global intensity-based patterns will be corrected with loess or quantile normalization when needed.
Gene expression profiles will be compared across a variety of conditions.
Comparisons will include, for example, wild type and mutants strains, drug-treated and un-treated, and various environmental conditions. Since the experiment is designed according to the statistical significance needed, we will follow the longitudinal model decided well ahead of the experiment and to choose genes that show differential expression among the control and experimental group. Given the large number of genes compared, care will also be exercised to avoid proportionately large numbers of false positives. To properly assess differential gene expression, we will employ at least two approaches: (1) Significance Assessment for Microarrays (SAM) (Tusher et al., 2001) to control the false discovery rate (2) the one controls the family wise error rate, but avoids the conservatism of Bonferroni correlation by utilizing a step-down method (Dudoit and Speed, 2000).
Exploratory data analyses using cluster methods can allow us to find patterns embedded in the data set and provide us some clues of the major targets. There exist a number of publicly available programs that allow one to cluster genes on the basis of similarity of their expression patterns across multiple experiments. This type of analysis has proven a very useful predictor of the function of unknown genes, and determined previously unknown linkages between different signaling/regulatory
pathways. Analysis of the promoter region of members of a cluster is another criterion that has been used to strengthen a case for possible physiological connections between its members (Bussemaker et al., 2001). Datasets from our microarray analysis will be analyzed using various methods such as hierarchical cluster analysis (Eisen et al., 1998), self-organizing maps (Tamayo et al., 1999), k-means clustering (Tavazoie et al., 1999) and principal component analysis (Raychaudhuri et al., 2000).
2. Understanding the molecular involvement of Efg1-CaEno1 pathway in the pathogenesis of C. albicans
Studies of C. albicans and its potential mechanisms of pathogenesis have relied heavily on the expression of various phenotypes induced by environmental changes or by its morphogenetic transitions. The relationship of these conditions/pathways to one another is complex. Here, we propose to study the molecular mechanism of the Efg1-CaEno1 pathway in drug resistance and virulence using molecular genetic tools and to identify Efg1 target genes using gene expression profiling.
a. Construction of a Caeno1/Caeno1 mutant
Since there is no known plasmid of C. albicans, homology recombination will be the main approach to generating knockout and knock-in mutations and other genetic manipulation. And since C. albicans is a diploid organism without known sexual cycle, both alleles of a given gene will have to be replaced at the same time to generate mutations. A homozygous Caenol/Caeno1 mutant will be constructed by gene disruption method based on the homology recombination strategy described previously (Gerami-Nejad et al., 2001; Wilson et al., 2000; Wilson et al., 1999).
Both copies of CaENO1 gene will be replaced by SAT1 cassette. A DNA fragment containing the SAT1 construct flanked with homology regions of CaENO1 at two extremities will be transformed into the C. albicans strain SC5314. The
transformants will be selected for the drug resistance marker and then screened for lose of the marker for the pup-out of the SAT1. Then, PCR product containing the SAT1 sequence flanking with the CaENO1 homology regions at two extremities will be transformed again into the CaENO1/Caeno1 strain to generate the Caeno1/Caeno1 homozygous mutant via homology recombination. Since CaEno1 is required for cell growth in the presence of glucose (Yang et al., 2006), the Caeno1/Caeno1
homozygous mutant will be constructed by selecting for SAT1+ transformants on the selective medium using glycerol as the carbon source. After selection for STA1-again, a DNA fragment containing the wild-type CaENO1 will be transformed into the Caeno1/Caeno1 homozygous mutant to generate the Caeno1/Caeno1::CaENO1
strain.
b. Determining whether CaENO1 is involved in morphogenesis/virulence, drug resistance in Candida albicans and other characterization of the Caeno1/Caeno1 mutant
To determine if CaENO1 regulates the morphogenesis of C. albicans, we will compare the wild-type and the Caeno1/Caeno1 mutant about their morphology, germ tube formation, colony formation, and cellular growth under filament-inducing condition including the addition of serum, temperature, pH, and other environment cues. Since the Efg1 also involved in drug resistance, we will also test the drug susceptibility of the Caeno1/Caeno1 mutant by Etest and/or agar dilution to unveil the connection between those pathways.
c. Characterization of new interested genes obtained by DNA microarray.
DNA microarray will be also employ to find out other genes regulated by Efg1.
The interested candidates will be subjected to mutagenesis by homologous
replacement as described in (1). Heterozygous and homozygous null mutants will be functional characterized by comparing their phenotypes with that of the wild type strain. The phenotypes to be studied include cell growth, cell morphogenesis, susceptibility to antifungal drugs and sensitivity to different stress conditions.
Finally, the correlation of the genes of interest with virulence will be assessed using a mouse model of systemic infection.
d. Elucidation of the relationship between CaENO1 and other genes
Making double mutations on two genes to assess whether they have related function is an approach commonly used for studying gene functions (Lo et al., 1997).
Thus, first of all, we will construct Caeno1/Caeno1 and new target gene double mutant. A PCR product containing the URA3-dpl200 sequence with the short homology regions (70 bps) flanking at the two extremities of another interested gene (NEW) will be transformed into Caena/Caeno1 to generate
NEW/new::dpl200-URA3-dpl200 Caeno1::ARG4/Caeno1::dpl200 strain. The NEW/new::dpl200 Caeno1::ARG4/Caeno1::dpl200 cells will be selected by growing the NEW/new::dpl200-URA3-dpl200 Caeno1::ARG4/Caeno1::dpl200 cells into a medium containing 5FOA.
The presence of 5FOA will select for the recombinants that have lost URA3.
Again, the same PCR product containing the URA3-dpl200 sequence with the NEW short homology regions (70 bps) at two extremities will be transformed into
NEW/new::dpl200 Caeno1::ARG4/Caeno1::dpl200cells to generate
new::dpl200-URA3-dpl200/new::dpl200 Caeno1::ARG4/Caeno1::dpl200 double mutant by selecting for Ura+ transformants.
3. The response of C. albicans to the environmental stimuli
Since the relationship of conditions/pathways of pathogenesis to one another is complex, we intend to study the possible cross-talk between the Efg1 pathway and other pathways of our interest.
a. Cell morphogenesis.
Efg1, Rim101, Nrg1 have been indicated to control cell morphogenesis. As described above, we will identify target genes of Efg1. To reveal genes that are commonly regulated by all three transcriptional factors or subsets of genes that are specifically regulated by one of the three factors, the target genes of Rim101 and Nrg1 will be also studied. Experiments will be performed by comparing expression
profiles between wild-type and mutants lacking functions of each transcriptional factor, and that between cell growth of yeast and hyphal forms. The mutant strains will be generated using the SAT1-flipper method (as described below) or the methods described above.
b. Iron-responsive gene regulation.
In the regulation of morphogenesis, Efg1 receives its upstream signal via a cAMP/PKA (protein kinase A)-dependent pathway (Ernst, 2000). Although that has not been studied in C. albicans, components of the cAMP/PKA pathway not only affect morphogenesis, but also affect iron-acqusition gene expression in S. cerevisiae.
To explore the possibility of the cAMP/PKA and Efg1 pathway to control
iron-responsive gene expression, we will also generate null mutant of PKA. We will compare patterns of gene expression between wild-type and mutants lacking functions of PKA and Efg1, and that between cell grown in iron-limiting and iron-repletion conditions. In addition to cAMP/PKA and Efg1 pathway, other potential
transcriptional factors controlling gene expression in response to iron availability will also be examined. We are generating deletion mutation of C. albicans Orf19.2272, which encodes a protein with a high homology with S. cerevisiae Aft1p. In S.
cerevisiae, Aft1p is an activator for iron acquisition and many other iron-responsive genes. The media representing iron-limiting and iron repletion conditions are used as previously described (Lan et al., 2004).
c. Other stress responses.
In the host, the survival of C. albicans is also dependent on evasion of the host’s
immune system, including the microbial killing mechanisms of phagocytosis.
Macrophages and neutrophils are the main components of the innate immune system and use reactive oxygen and nitrogen species to protect the host (Nathan and Shiloh, 2000). Superoxide readily dismutates to hydrogen peroxide or combines with nitric oxide to form strong oxidant peroxynitrite, which is fungicidal (Vazquez-Torres and Balish, 1997). In the presence of transition metals such as iron, hydrogen peroxide can even break down to form the highly reactive hydroxyl radical. Therefore, the regulation of iron may be of great importance to C. albicans to deal with oxidative stress.
To study the cross-talk between cell responses to iron availability and to oxidative stress, C. albicans will be treated with 0.4 mM H2O2or 0.5 mM menadione (a
superoxide generating agent) that allows the organism to tolerate ordinarily lethal levels of these oxidants (Jamieson et al., 1996). A comparison of S. cerevisiae and C.
albicans indicate that the latter can adapt much higher levels of reactive oxygen species (Jamieson et al., 1996). This analysis will be the identification of genes that may functionally confer this ability. Revealing the overlap or difference between cell responses to iron and oxidative stresses will be important for our understanding to the survival/persistence of C. albicans in its host environment.
d. Functional characterization of genes of interest
It is the expectation that DNA microarray data obtained will allow us to identify significantly expressed ORFs of both known and unknown functions. From these data, we will be possible to identify a limited number of ORFs which are related to general functions of interest. To analyze functions of these genes of interest (GOI), we will construct target gene disruptions using the SAT1-flipper method (Reuss et al., 2004). This method relies on the use of a cassette that contains a dominant
nourseothricin-resistance marker (CaSAT1) for the selection of integrative
transformants and an inducible FLP recombinase system for subsequent excision of the cassette (Reuss et al., 2004). Briefly, the flanking sequences of GOI are located at the both sides of the cassette. Following integration of the marker cassette by homologous recombination of the GOI flanking sequences, transformants were grown in a medium containing 10% BSA (bovine serum albumin) to induce the recombinase for marker construct excision. Cells were plated at a low nourseothricin concentration (25g/ml) to identify SAT1-negative colonies which grew to a smaller size under these conditions than the colonies from SAT1-positive cells. Selecting the small colonies and plating on the high drug concentration (100g/ml) dish to make sure it is SAT1-negative indeed. Two rounds of integration/excision result in the disruption of both alleles of the GOI.
Heterozygous and homozygous null mutants will be functional characterized by comparing their phenotypes with that of the wild type strain. The phenotypes to be studied include cell growth, cell morphogenesis, and susceptibility to antifungal drugs and sensitivity to different stress conditions. Finally, the correlation of the genes of interest with virulence will be assessed using a mouse model of systemic infection.