Variation in chromosome copy number influences the virulence of Cryptococcus neoformans and occurs in isolates from AIDS patients

Disomy at chromosome 13 of C. neoformans is correlated with reduced melanin formation

In comparison to the serotype A strain H99, the reference strain that is commonly used to investigate the virulence of C. neoformans, two clinical isolates, CBS7779 and WM626, were reported to have significantly diminished production of the major virulence factor melanin [18]. Since both strains were shown by comparative genome hybridization (CGH) to be disomic for chromosome 13 (chr 13), we hypothesized that their reduction in melanin formation was due to chromosomal variation such as disomy. We focused our subsequent work on CBS7779 because this strain was similar in virulence to strain H99, and strain WM626 was attenuated for virulence (see below). To examine the relationship between melanin formation and disomy, we initially screened strain CBS7779 for variants that showed a more pronounced defect in melanin production (hereafter called white variants). These were readily identified along with black variants that formed more highly melanized colonies. This screen yielded 12 black isolates (designated B1-B12), and 12 white isolates (W1-W12) for further analysis. These isolates generally maintained their melanin phenotypes upon repeated streaking on L-DOPA medium. Each of the 12 black and 12 white melanin variants were first examined by quantitative PCR using primers for genes at three positions along chr 13 (CNN00820, CNN01890 and CNN02400) to determine the copy number of the chromosome. Based on the genome sequences of serotype A and D strains [29, 30], the actin gene on chr 1 (CNA04650) and a gene on chr 4 (CND00707) were employed as a single copy controls. In addition, the SMG1 gene on chr 4 was used as a control because Fraser et al. [31] showed that this gene was duplicated in the serotype D strain JEC21, but not in strain H99. The PCR analysis revealed that each of the 12 black variants was monosomic at chr 13 and chr 4. All 12 of the white variants were monsomic at chr 4 but nine were disomic and three were trisomic at chr 13 (Table 1). These analyses were repeated several times on independent DNA samples with identical results. The PCR results were also confirmed and extended by CGH analysis on independent DNA samples that revealed monosomy for chr 13 in the three black variants (B4-6) and disomy in the three white variants (W1-W3) (Figure 1).

Strain	SMG1 (chr 4)	CNN00820 (chr 13)	CND00707 (chr 4)
H99	1	1	1
JEC21	2.33	1.07	1.22
CBS7779 B1	1.1	1.09	ND
CBS7779 B2	1.08	0.95	ND
CBS7779 B3	1.16	1.08	ND
CBS7779 B4	1.1	1.03	1.08
CBS7779 B5	1.24	0.99	1.28
CBS7779 B6	1.16	1.01	1.37
CBS7779 B7	1.27	0.68	ND
CBS7779 B8	1.38	1.34	ND
CBS7779 B9	1.43	1.01	ND
CBS7779 B10	1.38	0.74	ND
CBS7779 B11	1.26	1.15	ND
CBS7779 B12	1.36	0.99	ND
CBS7779 W1	1.3	1.84	ND
CBS7779 W2	1.16	2.01	ND
CBS7779 W3	1.1	1.97	ND
CBS7779 W4	1.23	2.23	1.36
CBS7779 W5	1.11	1.85	1.24
CBS7779 W6	0.96	1.78	1.19
CBS7779 W7	1.29	1.8	ND
CBS7779 W8	1.13	2.66	ND
CBS7779 W9	1.25	2.89	ND
CBS7779 W10	1.15	2.01	ND
CBS7779 W11	1.15	2.87	ND
CBS7779 W12	1.13	3.1	ND

Quantitative PCR was used to examine copy number (2^-ΔΔCt method).

ND, not determined

In addition to a correlation between reduced melanization and disomy at chr 13, we also found that white strains were more susceptible to the antifungal drug fluconazole, and to the trafficking inhibitors brefeldin A (BFA) and monensin (Figure 2A and data not shown). The latter inhibitors are relevant to virulence because we previously showed that they inhibit the secretory pathway needed for capsule formation [32]. In addition, several secreted enzymes contribute to virulence in C. neoformans including phospholipase B, urease and laccase [33]. The black and white variants exhibited no differences in capsule formation or their responses to oxidative and nitrosative stress (data not shown). However, the white strains did display slower growth in liquid culture (Figure 2B). Overall, these results established a correlation between disomy at chr 13 and phenotypic traits with potential relevance to virulence and clinical outcome, i.e., melanin formation, secretion and susceptibility to fluconazole.

Disomic, non-melanized strains have reduced virulence

The white and black variants with defined chromosomal complements provided the opportunity to determine the effect, if any, of variation in chromosome copy number on virulence in a mouse inhalation model of cryptococcosis. Cells from the same cultures that were employed to isolate DNA for the CGH analysis of the three white (W1-3) and black (B4-6) variants were also used as inoculum in the virulence assays. Initially, we performed a pilot virulence test by inoculating three mice with each of the three black isolates and the three white isolates. This study also included strain H99 because of its well-characterized and high level of murine virulence, as well as the original CBS7779 and WM626 strains. The results indicated that CBS7779 was similar in virulence to strain H99, WM626 was attenuated for virulence, and the white (disomic) variants were less virulent than the black (monosomic) variants (Figure 3A). These results were substantiated in a larger study that employed 10 mice for each variant. As shown in Figure 3A, the white variants again showed reduced virulence (i.e., prolonged survival) compared with the black variants.

Infections with the black and white variants produced similar fungal burdens in the lungs and the brains of infected mice, although greater variation between mice was observed for the black variants in the brain (Figure 3B). Considering that white variants grew more slowly in culture and the black variants were more lethal, their comparable fungal loads in mice were unexpected. These results suggested that other properties besides fungal proliferation were contributing to disease, and we therefore examined lung and brain tissue by histopathology to more closely compare the infections. As shown in Figure 4, lungs infected with the black strain (CBS7779-B4) exhibited more extensive airway remodeling compared with the lungs infected with the white strain (CBS7779-W2), which showed greater damage to the lung parenchymal as indicated by thickened tissue and reduce air space. These differences in pulmonary histopathology were more pronounced at days 14 and 21 than at day 7 (Figure 4 and data not shown). The infections with both strains resulted in similar levels of perivascular and peribronchial inflammation (Figure 4). Overall, the distinct virulence properties of the black and white strains are associated with differences in the type of damage observed in infected lung tissue.

Disomy at chromosome 13 influences gene expression

A microarray comparison of the transcriptomes of the black and white isolates was performed to determine whether disomy at chr 13 influences gene expression (Figure 5). For this study, CBS7779 isolates W2 and B4 were grown in liquid L-DOPA medium for six hours and RNA was prepared for hybridization. The analysis revealed that 127 genes were down regulated and 168 were up regulated by two-fold or more in the disomic W2 strain relative to the B4 strain (Additional files 1 and 2). Only two of the down-regulated genes in W2 were on chr 13, compared with 97 of the up-regulated genes. Chr 13 is 756,744 bp in length and contains ~234 annotated genes (see Materials and Methods). Thus, the array data indicated that 41% of the chr 13 genes were up regulated two-fold or more in association with disomy for this chromosome. We also examined the expression of genes on chr 13 (or other chromosomes) that might influence melanin formation and identified the CCC2 gene (CNAG_06415) as a potential candidate. Walton et al. [34] found that loss of CCC2 results in a melanin defect, and this gene encodes a transporter that may load copper onto the laccase enzyme that catalyzes the polymerization of melanin. However, the expression of CCC2 was not substantially different between the B4 and W2 strains (~1.5X higher in the B4 strain) and additional work will be needed to test whether this subtle difference influences melanization.

Copy number variation at other chromosomes also correlates with melanin formation

To further examine the association between chromosome copy number and melanization, we plated the white strains CBS7779-W2 and CBS7779-W3 on L-DOPA medium and screened for "second generation" black variants under the same conditions employed earlier for the parental strain CBS7779. These variants were readily detected and we used CGH to examine the complement of chromosomes in randomly selected colonies (Figure 6A, Table 2). All four of the black variants tested maintained the initial disomy at chr 13 that was present in the starting strains, but they were also disomic for chr 4 (variant W3-BB), chr 12 (W2-BB and W3-BC) or both chr4 and 12 (W2-BA). A subsequent comparison of the transcriptomes of strains W2-BA and B4 also revealed elevated transcription of many of the genes on chr 4, 12 and 13 (Additional file 3). The results for chr 4 and chr 13 in these strains were confirmed by qPCR (data not shown). With regard to melanin formation in the strains, we hypothesize that disomy at chr 4 and/or chr 12 may have compensated for disomy at chr 13. The examination of a random white colony from the starting strain W2 revealed that disomy at chr 13 was maintained and that additional changes occurred for segments of chr 3, 5, 12 and 14 (Table 2, Additional file 4).

Three second-generation white variants (B1-WA, B1-WB, B1-WC) from the monosomic black strain CBS7779-B1 were also examined and found to show changes in additional chromosomes (Figure 6A, Table 2). By CGH, (i) strain B1-WA had an increase in the number of copies of segments of chr 3 and chr 11, (ii) B1-WB was disomic for chr13 and showed duplications of segments of chr 3, chr 5 and chr 11, and (iii) strain B1-WC had duplications of chr 13, chr 6 and a segment of chr 5. These results and those described below indicated that changes in copy number occurred for several chromosomes and involved partial or complete chromosome duplication. In some cases, changes in melanization correlated with changes in specific chromosomes (i.e., chr 4, 12, and 13). However, the observed variation for part or all of four additional chromosomes (chr 3, 5, 11, 14) suggests that genome plasticity and/or other factors, such as background mutations, may also influence melanization. Overall, these results indicate that changes in chromosome copy number readily occur in the CBS7779 variants.

Chromosome copy number variation is detected after passage in mice

We next tested whether passage through mice influenced the transitions from the monosomic and disomic states. Initially, we examined the transition from disomy to monosomy by examining colonies recovered from the mice employed in the virulence assay with the three white strains CBS7779-W1, -W2 and -W3 (Figure 3A). To screen for melanin formation as an indicator of disomy, we determined the proportion of black variants in the starting inoculum of each strain by plating ~30,000 cells of each strain on L-DOPA medium. The W1 inoculum yielded nine black colonies (0.03%) and both W2 and W3 produced six (0.02%). We then determined the fungal burden in the lungs and brains of each mouse (Figure 3B) and screened ~50,000 colonies from each inoculated group (W1-W3) for black colonies. We found the following average percentages of black variants in brain (B) and lung (L) tissue: W1 (B: 0.08%, L: 0.32%), W2 (B: 0.06%, L: 0.22%), W3 (B: 0.05%, L: 0.30%). It appeared from these data that the percentage of black colonies recovered from the brain was lower than that recovered from the lung. This result may indicate a tissue-specific influence on the appearance of melanin variants.

From mice inoculated with the three white strains, we randomly selected 10 white and 14 black colonies and analyzed their chromosome copy numbers. Analysis by qPCR revealed that nine of the white isolates and seven of the black isolates from mice were disomic for chr 13 (data not shown). As with the in vitro analysis, this finding indicates that some strains can recover the ability to form melanin while still maintaining disomy at chr13. Subsequent CGH analysis of seven black colonies revealed that two had become monosomic at all chromosomes, three had retained the disomy at chr 13 and gained disomy for chr 4, one had lost disomy at chr 13 but gained disomy at chr 4 with a segmental change in chr 14, and one retained disomy at chr 13 with a elevated copy number of both chr 4 and 12 (Figure 6B, Table 2 and Additional file 5). These results are similar to those obtained with the second-generation black colonies from cultures (described above), where disomy at chr 4 and/or chr 12 correlated with the restoration of melanin formation in strains originally disomic for chr 13. Three white colonies recovered from mice infected with white strains maintained disomy at chr 13 (Figure 6B, Table 2, Additional file 5).

We also examined whether disomy at chr 13 arises in the host by inoculating each of six mice with ~ 1000 cells of the black strain B4. The starting inoculum did not contain white variants among ~30,000 colonies screened prior to inoculation of mice. Mice were sacrificed 24 days after infection and a screen of 2933 colonies from homogenized brain tissue did not reveal any white colonies. However, a screen of ~50,000 colonies from lung tissue yielded 27 white colonies, and subsequent qPCR analysis of 13 of these white isolates revealed that five were disomic for chr 13 (data not shown). In contrast, six randomly selected black colonies that arose from lung tissue were found to be monosomic for chr 13. Although we have not examined other chromosomes, this analysis confirmed that the transition from monosomy to disomy at chr 13 occurs during infection, as well as in vitro. However, it was also clear that other mechanisms besides disomy at chr 13, such as the accumulation of background mutations, may result in reduced melanin production.

Disomy at chromosome 13 occurs in a laboratory strain but does not correlate with melanin formation

The link between disomy at chr 13 and melanin formation was discovered in strain CBS7779 during a survey of genome variation in serotype A strains, and it is possible that strain CBS7779 is particularly prone to changes in chromosome copy number. To determine whether these results are strain-specific or a general property of C. neoformans, we investigated whether disomy readily occurs at chr 13 in a laboratory strain. We were also interested in testing whether specific chromosomes could be amplified through selection. We therefore tagged chr 13 in laboratory strain H99 by insertion of a neomycin resistance marker and then plated tagged strains on medium containing neomycin at 10X the concentration generally employed for selection of DNA transformants (2 mg/ml). Four tagged transformants (#s 32, 33, 34 and 35) were plated on L-DOPA medium containing neomycin at 10X the inhibitory concentration (10 × 10⁶ cells per plate). After five days of incubation, approximately 10 colonies were observed on each plate. White and black colonies were identified on L-DOPA medium with neomycin and the isolates were characterized for capsule formation and cell morphology (Additional file 6). Two of the white variants showed a larger cell size and two showed larger capsules, which indicated phenotypic changes beyond melanin formation. Subsequent CGH analysis of four white and two black isolates revealed that only two strains, one white (35W1) and one black (32B7), had become disomic at chr 13, and both of these strains also had changes at chr 14. One white strain (32W1) had changes at chr 11 and chr 14, and three strains (two white and one black) had no chromosome changes. The CGH results for the three strains with chromosomal changes are shown in Figure 7. Overall, these results indicate that elevated copy number at chr 13 could be achieved in a laboratory strain through tagging a specific chromosome with a selectable marker. However, this approach was only partially effective, and in one strain, an elevated chr 13 copy number did not correlate with reduced melanin formation.

Disomy is relatively rare in a collection of clinical and environmental strains

Our initial examination of four serotype A strains revealed two (CBS7779 and WM626) that were disomic at chr 13 [18]. To more widely assess the frequency of disomy, we used CGH to survey 19 additional clinical and environmental serotype A strains of different global origins (Additional files 7 and 8). Only two of the isolates showed chromosome copy number changes (Figure 8). Specifically, strain A5-35-17 appeared to have an elevated copy number for chr 1, and strain JP1086 contained an amplified region on chr 7. Both have the A5/M5 genotype as defined by multilocus sequence typing (MLST) and amplified fragment length polymorphism (AFLP) analyses [23, 24]. The CGH analysis did reveal that strains A7-35-23 and C45 had a greater overall level of sequence divergence from H99; this result is consistent with the fact that these strains have a VNII molecular type compared with the VNI subtype of strain H99 [24]. We previously demonstrated the ability of CGH to distinguish the three molecular types of serotype A strains, VNI, VNII and VNB [18]. These isolates have been stored and subcultured since the time of their isolation from patients or the environment, and this prolonged period of in vitro growth in complete medium may have selected for strains with a haploid complement of chromosomes.

Chromosome copy number variation is present in isolates from AIDS patients

We next examined whether chromosome copy number variation could be detected in isolates freshly collected from the cerebrospinal fluid of AIDS patients. Multiple colonies were obtained upon plating cerebral spinal fluid from 13 patients and individual colonies from each specimen were used to prepared DNA for CGH. No differences in chromosome copy number were observed for the isolates from 11 of the patients (RCT23, RCT31, HC2, HC3, HC5, RCT17, RCT50, RCT52, RCT55 (day 1), RCT55 (day 4) and RCT60 (data not shown, additional file 9). However, three different colonies from the patient HC2 had divergent patterns of hybridization on the CGH array containing the H99 genome, suggesting sequence heterogeneity. Subsequent testing revealed that these isolates were D serotype or molecular type VNIV (data not shown). It is known that the genomes of A and D strains exhibit sequence divergence of 5-10% [18, 29, 30]. A divergent hybridization pattern was also noted for two colonies of patient HC3 (HC-3a and HC-3b), but a third colony (HC-3c) showed a tighter hybridization pattern; this result also suggests that a mixed infection of both serotype A and D strains was present in HC3 (Additional file 9). This observation is reminiscent of the finding by Desnos-Ollivier et al. [28] that mixed infections are relatively common (see Discussion below). The colonies obtained from patient HC4 also indicated a mixed infection because one colony showed copy number elevation for part of chr 9 (indicating a segmental duplication, Figure 8), and another colony showed no copy number variation. Interestingly, a copy number change at chr 6 was present in all three of the isolates tested from patient HC6 (Figure 8). In these isolates, it appeared that the chromosome was duplicated but that one copy had a deleted region near the center of the chromosome. It is possible that duplication of the deleted region is not tolerated. FACS analysis indicated that the HC-6 isolates were diploid (Additional file 10). Overall, these results indicated that variation in chromosomal copy number was present in isolates from patients with HIV/AIDS, and provided additional evidence for mixed infections.

Strains and growth conditions

The strains for this study are listed in Additional file 7 and were maintained on yeast extract, peptone, dextrose medium (YPD; Difco). To screen for variants of strain CBS7779, single colonies were grown in liquid YPD overnight at 30°C, and plated on L-DOPA medium ((0.5 mM 3,4-hydroxy-L-phenylalanine [L-DOPA], 1 mM MgSO₄ · 7H2O, 22 mM · KH₂PO₄, 3 μM thiamine-HCl, 0.1% glucose, 0.1% L-asparagine, pH 5.6) at ~500 cells per plate. The plates were incubated at 30°C for 3 days and melanin production (darker or lighter colony pigmentation) was evaluated visually. Strains from the cerebral spinal fluid of AIDS patients were obtained by plating directly on YPD plates. To minimize the potential influence of prolonged culture on chromosome number, single colonies were picked and individually cultured in 50 ml YPD overnight prior to genomic DNA isolation. Growth in liquid yeast nitrogen base medium was monitored in a TECAN plate reader over 72 hours.

To insert a selectable marker on chr 13 of strain H99, a construct containing the neomycin resistance cassette was prepared by overlapping PCR and introduced by biolistic transformation [64]. The construct replaced a fragment of 31 base pairs in an intergenic region (chr 13: 316840-316870). The primer sequences for the construct are given in Additional file 11. Transformants were screened by colony PCR and four positive strains with the correct insertion of the neomycin resistant cassette (#s 32, 33, 34 and 35) were selected. A single colony of each tagged strain was grown in YPD overnight at 30°C and then plated on L-DOPA medium containing neomycin (2 mg/ml). After three days at 30°C, black and white colonies were collected, genomic DNA was isolated and qPCR and/or CGH analysis was employed to examine the copy number of chr 13 as described [18].

Comparative Genome Hybridization

The genomic DNA for hybridization experiments was isolated as previously described [18]. The genomic sequence of H99 was retrieved from the Cryptococcus neoformans var. grubii H99 Sequencing Project, Broad Institute of Harvard and MIT (http://www.broadinstitute.org/). The assembled sequences were used by NimbleGen Systems, Inc. (Madison, WI) to design and manufacture the oligonucleotide genomic arrays. Because the focus of this study was on chromosome copy number changes, rather than the duplication/deletion of single genes or genome fragments, the design for the arrays was modified for our earlier design [18], with oligonucleotide probes that cover all 14 chromosomes of H99 tiled at an average interval spacing of 133 bp on one strand. The average length of each probe was 50 bp (range 45 to 85 bp) and the average Tm of the probes on each array was 76°C. The number of oligonucleotide probes for this new design for H99 was 134, 953. The probes for each array were designed to uniquely match a single sequence in the genome and highly repetitive centromeric regions and the rDNA repeat cluster were not included. The hybridization procedures, data acquisition and analysis were as previously described [18].

Phenotypic assays

Melanin production was examined on solid L-DOPA medium. Capsule measurements, and stress- and drug-response assays were performed as previously described [32, 65]. Briefly, exponentially growing cultures of test strains were washed, resuspended in H₂O, and adjusted to 2 × 10⁴ cells/μl. The cell suspensions were serially diluted 10-fold, and 5 μl of each dilution was spotted on YPD and/or YNB plates supplemented with or without 20 μg/ml or 40 μg/ml brefeldin A, 0.5 mg/ml monensin, or 5 μg/ml fluconazole. Plates were incubated for two to five days at 30°C or 37°C and photographed. To estimate the frequency of spontaneous emergence of color variants in vitro, ~3 × 10⁴ cells of each strain were plated onto L-DOPA medium. After 72 hours of incubation at 30°C, dark colonies originating from a white parent (e.g., CBS7779-W1, -W2, or -W3), or white colonies from a black parent, were identified, and frequencies were estimated by dividing the number of variant colonies by the total CFU. A similar method was used to estimate the frequencies of changes in melanization in vivo; the percentages of colonies of the total fungal load in each organ with color variant to parental phenotype was assessed.

Virulence assays and histopathology

The virulence of each strain was examined using female BALB/c mice (4 to 6 weeks old) from Charles River Laboratories (Ontario, Canada). Fungal cells were grown in 5 ml of YPD at 30°C overnight, washed twice with PBS (Invitrogen, Canada), and resuspended in PBS. The BALB/c mice, in groups of 10, were anesthetized intraperitoneally with ketamine (80 mg/kg of body weight) and xylazine (5.5 mg/kg) in PBS and suspended on a silk thread by the superior incisors. A suspension of 5 × 10⁴ cells in 50 μl was slowly dripped into the nares of the anesthetized mice, and the mice were suspended for 10 minutes on the thread. The number of cells in each inoculum was confirmed by CFU determination after plating serial dilutions. The health status of the mice was monitored daily post-inoculation. Mice reaching the humane endpoint were euthanized by CO₂ anoxia. All ten mice from each infection group were used at the endpoint to assess fungal loads in organs. Following euthanasia, the brain and the lungs were aseptically removed and immersed in 1 ml sterile PBS containing a metal bead. Organs were homogenized using an automated tissue homogenizer (Retsch, PA, USA). The samples were serially (1:10) diluted and 100 μl of 10^-4 and 10^-5 dilutions were plated and spread with 4 mm glass beads on L-DOPA medium containing 100 μg/ml chloramphenicol. After 3 days of incubation at 30°C, the L-DOPA plates were photographed using a Nikon Coolpix 990 camera. The images were opened with AlphaImager 3400, and the CFUs were digitally assessed by AlphaEase FC software (Alpha Innotech, San Leandro, CA). The number of color variants was tallied manually.

Female BALB/c mice (8 weeks old) from Charles River Laboratories (Ontario, Canada) were used in histopathology experiments to examine whether CBS7779-B4 and CBS7779-W2 cells elicited differential immune responses and/or exhibited different dissemination efficiency in vivo. Groups of nine mice were inoculated as described above, and on days 7, 14, and 21 post-infection, three mice from each group were euthanized by CO2 asphyxiation, and lungs and brains were immediately fixed in 10% neutral buffered formalin. Tissue embedding and sectioning (5 μm) were prepared by Wax-IT Histology (Vancouver, Canada) and stained with hematoxylin and eosin (H&E) or Mayer's mucicarmine (MM) to visualize the cryptococcal capsule. Lung slides were examined by light microscopy (AxioSkop 2 MOT, Carl Zeiss Ltd, Canada) in a blinded fashion and scored for degree of fungal infiltration, tissue damage and inflammation. Brain sections were examined by light microscopy for the presence and distribution of cryptococcal cells.

All experiments with animals were conducted according to guidelines set by the Canadian Council on Animal Care. The protocol (A08-0586) for the virulence assays with mice was approved by the University of British Columbia's Committee on Animal Care (Animal Welfare Assurance Number A5090-01). C. neoformans isolates from patients were obtained as part of the Database and Specimen Repository for Infectious Disease-Related Studies (reference number CR3_Pro00005314), as approved the Duke University Internal Review Board.

Statistical analyses

Statistical analysis of survival differences was performed by Kaplan-Meier survival curves and log rank tests. An unpaired, two-tailed t-test was used to determine the differences in mean CFU per organ (i.e., brain and lung) between mice inoculated with different strains. A P-value of 0.05 or less was considered significant. All statistical analyses were carried out using GraphPad Prism version 4 for Windows (GraphPad Software, San Diego, CA).

Quantitative real-time PCR

Quantitative real time PCR was used to confirm the copy number of chromosomes 4 and 13 in selected colonies as previously described [18]. Each reaction mixture contained 100 ng of template DNA, 0.25 μM of forward and reverse primers, and Power SYBR Green PCR Master Mix (Applied Biosystems). An ABI 7500 instrument (Applied Biosystems) was used for signal detection and data collection. Primer sequences for genes CNA04650 (actin), SMG1 (chr 4), 00_070 (chr 4) and CNN00820 (chr 13) were used for CCNV determination, as listed in Additional file 12 and as previously described [18].

Flow cytometry

To examine the genomic profile of the original parental strains (CBS7779, CBS7779B4-6, CBS7779W1-3) and selected strains harvested post-infection, flow cytometry was performed as described [49], with slight modifications. Briefly, overnight cultures were centrifuged at 10,000 rpm for 1 minute, washed once with 1 mL of distilled water, and 1 × 10⁷ cells were harvested and fixed in 70% ethanol overnight at 4°C. Subsequently, the ethanol was removed and the cells were washed with NS buffer (10 mM Tris-HCl [pH 7.6], 250 mM sucrose, 1 mM EDTA [pH 8.0], 1 mM MgCl₂, 0.1 mM CaCl₂, 0.1 mM ZnCl₂), and stained with 10 μg/mL propidium iodide (Sigma) in 0.5 mL NS buffer containing 1 mg/mL RNase A (Roche) at 4°C overnight. The stained cells were then suspended in 50 mM Tris-HCl (pH 8.0) (without sonification). Propidium iodide fluorescence was analyzed using a BD FACSCalibur (BD Biosciences, Canada) on the FL3 channel. Graphs and histograms were generated using Flowjo v8.8.6 (Ashland, OR).

RNA labeling and microarray analysis

The strains CBS7779-W2 and CBS7779-B4 were used for microarray analysis. Three independent biological replicates for each strain were grown in 25 ml of YPD overnight at 30°C. Cells were washed twice with water, re-suspended in L-DOPA medium and then incubated at 30°C for 6 h prior to RNA extraction. RNA was purified with the RNeasy kit (Qiagen) and treated with DNase (Qiagen) following the manufacturer's recommendations. The quality of the RNA was analyzed with an Agilent 2100 Bioanalyzer. cDNA synthesis, labeling with Cy3 or Cy5, and array hybridization to 70-mer microarrays (version 2.0, http://www.mogene.com were performed by Mogene (MO, USA). Dye-swap experiments were performed for all hybridizations. Modification of the gene list of microarray probes was described previously [66]. Single 60-mer oligonucleotide probes were designed for each of the ORFs and were duplicated on the microarray to provide an estimate of intra array variance. Of the 7,302 ORFs designated by the Broad Institute, 100 ORFs were eliminated because the sequences were of low complexity. The probes covered 6,932 of the final annotated genes. Agilent Technologies fabricated the arrays.

After ANOVA normalization, genes showing log₂-transformed fold changes (P < 0.05) were selected for further analyses. To classify genes by their functions, a web-based program, CateGOrizer (http://www.animalgenome.org/bioinfo/tools/catego/), was used [67]. The heatmap was generated using the "heatmap.2" routine in the package gplots after the gene expression data were normalized using R-package, by limma with options of 'minimum' background correction, 'Lowess' within-array normalization, and 'quantile' between-array normalization as recommended [68].