Ploidy influences cellular responses to gross chromosomal rearrangements in saccharomyces cerevisiae

Collection of the studied strains

The URA2 gene of S. cerevisiae, which is located on the chromosome X, encodes a multifunctional protein occurring in the pyrimidine biosynthesis pathway. This protein is composed of 3 active domains: glutamine amidotransferase (GATase), carbamoylphosphate synthetase (CPSase) and aspartate transcarbamoylase (ATCase). Three point mutations located in the proximal region of the ura2 _15-30-72 allele (Figure 1) lead to a loss of functions in the Ura2 protein and to auxotrophy towards uracil [18]. This auxotrophy is due only to the lack of ATCase activity, since the missing GATase and CPSase functions are compensated for by two isoenzymes present in the arginine biosynthesis pathway. Only ATCase reactivation is therefore required for prototrophic reversion to occur and it has been previously established that this reactivation is possible via complex chromosomal rearrangements [18–23].

To investigate the effects of the same GCRs on cellular behaviour in haploid and diploid strains, revertants were constructed in both contexts. Unfortunately, haploid strains could not be generated from diploids carrying segmental deletions and nonreciprocal translocations. In fact, these rearrangements involve essential genes, which result in non-viable haploid strains [20, 21]. Meiosis was therefore induced in the diploid mutant strains carrying an aneuploidy or a segmental duplication. After sporulation, segregants were grown in minimal medium and spores still carrying the GCR were selected. The ura⁺ spores were conserved for further analysis. On the other hand, haploid revertants were crossed with a strain possessing the ura2 _15-30-72 allele in order to obtain diploid strains. This yielded a final set of 21 revertants, 16 of which carried the same GCRs in two different genetic backgrounds (Table 1).

Ploidy content generates different colony morphologies

The colony morphology of laboratory strains of S. cerevisiae is generally smooth whereas those isolated from nature and vineyards show a structured fluffy pattern [24]. However, the same strain can also show various morphological features, depending on the carbon source used [25]. Mutant and parental strains were plated onto a solid YPD medium containing 2% glucose. In the haploid context, the colony morphology was found to be of two kinds in the revertant A5-n strain: one with a smooth appearance and one with a granular appearance (see Additional file 1, Figure S1). When these two colonies were spread on minimum medium containing no uracil in order to maintain the selection pressure, only granular colonies developed. This suggests that the additional chromosome essential to uracil prototrophy is unstable in this mutant. This chromosome is therefore conserved only in cells generating granular colonies. By contrast, colony morphology of other revertants showed only one homogeneous type of phenotype differing from the parental strains. For example, the haploid parental strain has an irregular shape whereas the mutants can be either circular and smooth (D1-n, SD1-n) or circular and rough (A2-n, A3-n, A4-n) (Figure 2 and Figure S 2).

The use of other carbon sources such as glycerol (3%), galactose (2%), sorbitol (2%), ethanol (2%) and acetic acid (2%) brought to light the existence of some more striking discrepancies between the mutant and parental strains. Differences were observed, for example, between the NRT1-2n revertant and the parental diploid strain in the presence of glycerol. The parental strain has peripheral swellings with a central invagination, whereas NRT1-2n strain is circular with a fluffy appearance (Figure 2 and see Additional file 2, Figure S2).

These comparisons also showed the existence of morphological differences between revertants carrying the same GCRs, depending on the ploidy context. As shown in Figure 2B, colonies of the revertant D1-n were found to be smaller than the D1-2n colonies, which suggests that the ploidy has decisive effects on the cellular responses. In addition to these morphological findings, the patterns of growth on non-fermentable substrates such as glycerol and ethanol showed that the revertants' respiration capacity was preserved (see Additional file 2, Figure S2).

GCRs generate abnormal cellular morphologies

It was established a decade ago that the increase in ploidy level from haploidy to tetraploidy results in an increase in the intracellular volume, which induces cell shape modifications [26]. However, the presence of GCRs can also give rise to abnormal cellular morphological features such as pseudohyphal growth, multibudded cells and cells with elongated buds [12, 27]. With a view to comparing the cellular morphology of revertants and parental strains, cells were grown in minimum liquid medium with 2% glucose up to mid-log phase and examined under light microscopy at a magnification of × 1000. Haploid and diploid parental strains both showed a classical ellipsoidal shape, but not the mutants (Figure 2). As shown in Figure 2C, revertants A2-n and D1-n formed aggregates, whereas A3-2n and A4-n were more elongated than the parental strains, which suggests that an increase in the cellular volume had occurred. This increase is consistent with the delayed G1 phase observed in haploid disomic strains [10]. This time lag may be due to the regulation of the 17 genes thought to be involved in the cellular volume increase in S. cerevisiae such as the G1 cyclins encoded by the CLN1 and PCL1 genes or FLO11[26, 28]. The presence of an additional chromosome or GCRs may impair the molecular pathway, leading to the deregulation of several genes including cyclins and FLO11.

Growth defects are more obvious in diploid revertants

To analyze the effects of GCRs on the doubling time, the specific growth rates of cells grown in minimum medium were determined during the exponential growth phase. All experiments were performed in triplicate. Student's T-tests were performed on the results obtained. Revertants giving a p-value < 0.05 were taken to have a different doubling time from that of the parental strains. The diploid parental strain has a doubling time equal to one division every 1.5 hour, whereas all the revertants except A1-2n, A2-2n and D2-2n have longer doubling times. In the main diploid mutant strains studied, the specific growth rate was even found to be twice as fast as that of euploid strains (Figure 3). In the haploid context, parental strain buds every almost 2.5 hours, and the doubling time increased by a factor of less than two in the case of only three revertants (SD1-n, SD3-n and D1-n) (Figure 3). To compare the effects of GCRs depending on the ploidy, the ratios between the doubling times of revertant and parental strains (g_revertant/g_wt) were determined. Comparisons between these ratios showed that with the same GCRs, this ratio was almost equal to 2 in case of diploid revertants but only about 1 in that of haploids, which suggests that GCRs generate two distinct responses, depending on the ploidy.

Metabolic properties and ATP production rates are impaired

To determine the metabolic impairments involved, various properties, including the specific glucose consumption rate (Q_Glu) and the specific ethanol production rate (π_EtOH), were studied. These parameters were defined as the amount of glucose consumed and the amount of ethanol produced per hour and per gram of biomass generated. As previously, all theses experiments were again performed in triplicate. Student's T-tests were performed on the data obtained. The values of these parameters were found to differ between the parental and mutant strains at both ploidy levels. In diploid revertants, only A1-2n consumed glucose faster than the diploid parental strain, and the differences did not differ significantly in the case of A2-2n, A3-2n, D1-2n and D2-2n (Figure 4). Lower glucose consumption rates were observed in the last seven diploid revertants studied (Figure 4). In haploid strains, three revertants showed a decrease in Q_Glu, whereas this rate did not differ significantly in the other haploid mutants in comparison with the parental strain (Figure 4). Likewise, π_EtOH was found to decrease in all the mutants except for the diploid A1-2n and the haploids A2-n, SD1-n and SD3-n (Figure 4).

The Q_Glu/π_EtOH ratio can be used to assess the amount of ethanol produced per mmole of glucose consumed. A high Q_Glu/π_EtOH ratio means that a low amount of ethanol is produced. Classically, the alcoholic fermentation of 1 molecule of glucose results in the formation of 2 molecules of ethanol, as in the case of the haploid parental strain showing a ratio of 0.49. Higher ratios (of up to 0.80 in the case of D1-n), suggesting the occurrence of a decrease in the ethanol production rates, were obtained in most of the haploid revertants (Figure 4). In the diploid strains, Q_Glu/π_EtOH also increased in the five mutant strains A2-2n, NRT2-2n, SD1-2n, D2-2n and D3-2n, ranging from 0.74 to 0.81 (Figure 4). This decrease in the ethanol production rate per molecule of glucose uptake suggests that this sugar may be used differently in the mutants to achieve the ATP production rates required for cell survival and proliferation.

Cellular response depends on the ploidy context

To determine the effects of GCRs on the patterns of gene expression, each pair of mutants and parental strains were grown in minimum liquid medium up to mid-log phase and a comparative transcriptomic analysis based on microarray hybridizations was conducted. Each experiment was performed in duplicate. Results obtained on 14 revertants were used for further analyses depending on the reproducibility of the replicates (Table 1). A global analysis of the transcriptomic data was performed on all the genes to define the proportion of genes that are differently regulated in mutants in comparison with the parental strains. Results showed that up to 2/3 of all the genes are misregulated in haploid mutants in comparison with the parental strain, whereas this proportion is less than 1/3 in diploids.

We clustered misregulated genes in different groups depending on the ploidy and the type of GCRs. Each group was then divided into 2 subgroups depending on whether these genes were up- or down-regulated. To identify the functional categories that were enriched in the various subgroups, the GO Term Finder program was used [29]. In all the diploid and haploid revertants studied, a similar pattern of up-regulation was observed among the genes involved in energy metabolism, such as those encoding subunits of the succinate dehydrogenase complex and those encoding proteins involved in the tricarboxylic acid cycle. This finding is consistent with the fact that higher ATP production rates were observed in revertants than in the parental strains. However, some differences were also found to exist between haploid and diploid revertants showing exactly the same GCRs. Genes involved in the composition of the ribosome and in RNA processing were found to be down-regulated in diploids but up-regulated in haploids (Figure 5). In addition, only haploids showed over-expression of genes involved in the regulation of the proteasome, which suggests the occurrence of hyperactivity in this complex. All in all, these data suggest not only that the cellular responses depend on the ploidy level, but also that both haploid and diploid strains require to be provided with additional energy supplies to be able to proliferate.

GCRs generate a stress response in both ploidy contexts

Almost 850 genes in S. cerevisiae have been described as environmental stress response elements (ESR) and are known to be involved in carbohydrate metabolism or protein metabolism [30]. A t-test of these stress response genes in all the revertants studied here showed that 457 of these genes are misregulated. The most noteworthy result obtained was that ESR genes are regulated differentially, depending on their ploidy context (Figure 5). As mentioned above, many genes involved in translation and ribosome biogenesis are differentially expressed in revertants. These genes are down-regulated in diploid strains and up-regulated in haploids (Figure 5). These results suggest that all GCRs induce a general stress response in these mutants, regardless of their ploidy.

One haploid revertant (D1-n) is especially interesting because only 223 of its genes are differentially expressed, whereas the average number of regulated ESR genes in haploid mutants is 408. This strain contains a deletion event at the locus where the ATCase region of URA2 is fused to the PBS2 gene. The protein encoded by this gene is a MAP kinase kinase involved in the HOG signalling pathway [31]. The phosphorylation of the Hog protein by Pbs2p results in the activation of transcription factors Msn2p and Msn4p, which regulate many ESR genes [32, 33]. When the PBS2 gene is mutated, cells are more sensitive to the polymyxin B antibiotic [34]. The fact that D1-n is more sensitive to this antibiotic than the haploid parental strain shows that the kinase is not functional in this revertant. This finding helps to explain the existence of this small set of misregulated ESR genes. It also suggests that all the revertants respond to GCRs acting as environmental stresses. This response is partly independent of the HOG signalling pathway.

Strains and growth media

All the strains used in this study are isogenic to the laboratory strain FL100 and were obtained using the positive selection screening method described previously [19]. Yeast cells were grown at 30°C in liquid or solid (2% agar) yeast peptone (YP) with various sources of carbon and supplemented yeast nitrogen base (YNB).

Strain construction

Diploid revertants were sporulated to isolate the ura⁺ haploid strains. Haploid revertants were crossed with a strain isogenic to FL100 that possesses the ura2 _15-30-72 allele in order to obtain diploid cells with the chromosomal rearrangements required for prototrophy towards uracil.

Characterization of colony morphology

Cell cultures were grown in YNB medium in mid-log phase (0.35 < OD₆₀₀ < 0.45), placed on slides, stained with neutral red and examined by light microscopy (magnification × 1000). The colony morphology was determined after incubating the cells for 6 days on solid YP media with glucose (2%), galactose (2%), ethanol (2%), sorbitol (2%), glycerol (3%) or acetic acid (2%). Pictures were obtained using a Leica Z16 APO A Macroscope (magnification × 10).

Fitness, glucose uptake and ethanol production

Cell cultures were grown in 50 ml of YNB medium, starting with a low OD₆₀₀ = 0.002. After 12 hours of growth, samples were collected every 2 hours and the OD₆₀₀ was measured to determine the growth rates. All samples were also used to determine the glucose uptake and ethanol production rates. Residual glucose rates and ethanol concentrations were determined using the D-Glucose kit and the Ethanol kit from R-BIOPHARM, respectively. Batch cultures were performed concomitantly in 200 ml of YNB medium to determine the dry weight of each strain. Cell cultures were collected in the mid-log phase and dried at 65°C for 2 days.

Specific glucose consumption rates (Q_Glu) were determined during the exponential growth phase and expressed as the amount (in millimoles) of glucose consumed divided by the amount of biomass produced (in grams of dry weight) and multiplied by the corresponding specific growth rates (in terms of the increase in the biomass per hour). The specific ethanol production rates (π_EtOH) were also determined and expressed as the amount of alcohol produced per hour per gram of biomass.

RNA extraction and gene expression arrays

10⁸ cells in the mid-log phase (0.33 < OD₆₀₀ < 0.36) grown in YNB medium were harvested on filters, frozen in liquid nitrogen and conserved at -80°C until RNA preparation was performed. Filters were incubated at 65°C for 1 hour in lysis buffer (10 mM EDTA, 0.5% SDS and 10 mM Tris, pH 7.5) and acid phenol and then placed on ice for 10 min. The aqueous phase was further extracted with an equal volume of chloroform. Total RNAs were then precipitated overnight in the presence absolute ethanol and sodium acetate 4M. The RNA was then purified using RNeasy columns (Qiagen).

Quantity and integrity of the RNA were measured with the RNA 6000 nano assay kit using an Agilent 2100 bioanalyser (Agilent Technologies, Germany). 1.5 μg of total RNA was used for each reverse transcription reaction and indirectly labeled with Cy3 and Cy5, and subsequently hybridized on microarrays. Agilent DNA microarray slides (GE 8 × 15K n AMADID 015761) containing probes for most of the yeast open reading frames were used. Two hybridizations were performed for each comparison using the dye-switch procedure. The arrays were scanned using a Genepix 4000 scanner.

The full dataset can be downloaded from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) [40] under the accession series number GSE28285.

Data analysis

The microarray data were normalized without performing background substraction, using the global Lowess method performed with the Goulphar software program [41]. Gene profile analyses were performed using the Java Treeview software program [42]. Hierchical clustering and Student's t-tests were performed using Pearson correlation with the MeV software program [43]. Gene ontology was analyzed using GO Term Finder http://go.princeton.edu/cgi-bin/GOTermFinder[29].