The Magnitude of Candida albicans Stress-Induced Genome Instability Results from an Interaction Between Ploidy and Antifungal Drugs

Organismal ploidy and environmental stress impact the rates and types of mutational events. The opportunistic fungal pathogen Candida albicans, serves as a clinically relevant model for studying the interaction between eukaryotic ploidy and drug-induced mutagenesis. In this study, we compared the rates and types of genome perturbations in diploid and tetraploid C. albicans following exposure to two different classes of antifungal drugs; azoles and echinocandins. We measured mutations at three different scales: point mutation, loss-of-heterozygosity (LOH), and total DNA content for cells exposed to fluconazole and caspofungin. We found that caspofungin induced higher mutation rates than fluconazole, although this is likely an indirect consequence of stress-associated cell wall perturbations, rather than an inherent genotoxicity. Surprisingly, we found that antifungal drugs disproportionately elevated genome and ploidy instability in tetraploid C. albicans compared to diploids. Taken together, our results suggest that the magnitude of stress-induced mutagenesis results from an interaction between ploidy and antifungal drugs. These findings have both clinical and evolutionary implications for how fungal pathogens generate mutations in response to antifungal drug stress and how these mutations may facilitate the emergence of drug resistance.

As an opportunistic pathogen, Candida albicans is isolated in 40-60% of fungal infections in humans (Pfaller andDiekema 2007, 2010;Pfaller et al. 2008;Pfaller 2012;Perfect 2017) and encounters diverse stressors within its human host. In vitro stress-induced mutagenesis studies show that stressors such as febrile temperature (39°) elevate the rate of LOH and increase the frequency of whole-chromosome missegregation by $fivefold (Forche et al. 2011). C. albicans also encounters antifungal drugs, which are used to treat infection and include the azoles, which target the cell membrane, and echinocandins, which target the inner cell wall (Pappas et al. 2016;Perfect 2017). Antifungal drugs are potentially mutagenic, as fluconazole increases LOH rates (Forche et al. 2011) and generates aneuploidy (Harrison et al. 2014;Robbins et al. 2017;Stone et al. 2019), which is commonly associated with azole-resistance (Selmecki et al. 2006(Selmecki et al. , 2009Robbins et al. 2017;Stone et al. 2019). With only a limited number of drugs that effectively treat fungal infections (Pappas et al. 2016), a deeper understanding of drug-induced mutagenesis is necessary to understand the emergence of resistance.
While fluconazole-induced mutagenesis has been studied in diploid C. albicans, no study to date has examined how other types of antifungal drugs impact C. albicans genome stability, or how ploidy impacts this phenomena. In this study, we investigated the rates and types of mutations generated in diploid and tetraploid C. albicans exposed to two classes of antifungal drugs, echinocandins and azoles. We found that caspofungin elicited very high mutation rates regardless of ploidy, although this is likely an indirect consequence of stress-associated cell wall perturbations, rather than an inherent genotoxicity. Surprisingly, we found that antifungal drugs disproportionately elevated genome and ploidy instability in tetraploid compared to diploid C. albicans. These findings indicate that there is an interaction between ploidy and drug exposure that determines the magnitude of stress-induced mutagenesis. These results have evolutionary and clinically relevant implications for how fungal pathogens generate mutations that potentially drive the emergence of antifungal drug resistance.

Drug susceptibility assay
The minimum inhibitory concentration (MIC) for diploid and tetraploid C. albicans was performed as previous described (Rosenberg et al. 2018) with slight modification. Briefly, single colonies were inoculated in YPD and incubated with shaking for $16hrs at 30°. Yeast cultures were subsequently normalized to 0.01 OD and 200mL was spread onto casitone plates containing 1% agar and left to dry for 10 min. Standardized E-test strips (Biomeureix) of fluconazole (0.016 mg/mL -256 mg/mL) and caspofungin (0.002 mg/mL -32 mg/mL) were placed upon the plates and incubated for 24 hr at 30°and then photographed.

HIS4 reversion assay
The reversion rate of the his4-G929T allele was performed as previously described (Forche et al. 2011) with slight modification. Briefly, 10 single colonies of MH296 and MH297 were inoculated into 6 mL YPD and incubated with shaking for 24 hr at 30°. 1 mL of overnight cultures were aliquoted to 4 mL of the following media: no-drug, 1 mg/mL fluconazole, 10 mg/mL fluconazole, 0.25 mg/mL caspofungin, 2.5 mg/mL caspofungin, and 100 mg/mL calcofluor white. All treatments were incubated with shaking for 5 days at 30°. To determine the number of viable colonies for each treatment, 500 mL was taken from each culture for serial dilution and 100 mL of the 10 24 or 10 25 dilutions were plated onto SDC. To determine the number of revertants, the remaining 4.5 mL the cultures were harvested by centrifugation, washed with H 2 O, resuspended with 300 mL ddH 2 0 and plated onto large, 150mm · 15mm plates containing media lacking histidine. Plates were incubated for 48 hr at 30°before counting. The rate of reversion was determined by fluctuation analysis (Luria and Delbrück 1943). All experiments were performed in triplicate.
All experiments were performed in triplicate. At least 48 single colonies from each treatment of MH84 and MH128 were picked from SDC and 2-DOG plates and stored as glycerol stocks at -80°for subsequent flow cytometry analysis.

Flow cytometry assay
Single colony isolates obtained from SDC and 2-DOG following 24 hr drug treatment were analyzed for total DNA content using flow cytometry (Hickman et al. 2015).
Briefly, strains were inoculated into 900 mL YPD and incubated with shaking for $16 hr at 30°. 50 mL of the overnight cultures were re-inoculated into 450 mL fresh YPD and incubated for $6 hr. 200 mL of cells were harvested, washed with ddH 2 0, and resuspended in 20 mL 50:50 TE (50 mM Tris, pH 8; 50 mM EDTA). Cells were fixed with 95% ethanol and incubated overnight at 4°. The following day, cells were washed twice with 50:50 TE, resuspended with 50 mL RNAse A (1 mg/ml) and incubated for 1 hr at 37°. Cells were then collected via centrifugation and resuspended with 50mL proteinase K (5 mg/ml), and incubated for 30 min at 37°. Cells were subsequently washed with 50:50 TE, resuspended with 50 mL SybrGreen (1:100 dilution in 50:50 TE; Lonza, CAT#12001-798, 10,000X) and incubated overnight at room temperature, shielded from light. Cells were collected via centrifugation and resuspended with 150 mL 50:50 TE, briefly sonicated, and analyzed using an LSRII flow cytometer, with 10,000 events collected for each sample. To calibrate the LSRII, laboratory diploid (SC5314; (Gillum et al. 1984)) and tetraploid (mating product of RBY16 and CHY477; (Bennett and Johnson 2003)) strains were used and also served as internal controls. G1 peak values were determined by FITC-A intensity fit with the multi-Gaussian cell cycle model (FloJoV10). To assess the frequency of deviations in total DNA content, G1 peaks of experimental samples were compared to the mean G1 peak for either MH84 (n = 45) or MH128 (n = 48). Deviations to total DNA content were defined as any experimental sample with a G1 peak greater than 2 standard deviations from the mean G1 of its respective parental strain (MH84 or MH128).

Statistical analysis
Statistical analysis was performed using GraphPad Prism 8 software. Data sets were tested for normality using the D'Agostino & Pearson omnibus normality test. To test for difference between the no-drug treatment and the drug treatments, we used the non-parametric, unpaired, Mann-Whitney U-test.

Data availability
All strains are available upon request. The authors ensure that the required data for reproducing these findings and confirming the conclusions are available in the article and corresponding figures and tables. Supplemental material available at figshare: https://doi.org/10.25387/ g3.8872604.

RESULTS
C. albicans cell viability depends on ploidy and exposure to antifungal drugs To asses if antifungal drug susceptibility differs between diploid and tetraploid C. albicans, we determined the minimum inhibitory concentration (MIC) of fluconazole (FLU) and caspofungin (CAS) for an isogenic diploid and tetraploid pair. We observed no differences in drug susceptibility between the two ploidies and determined the MIC for fluconazole to be $1 mg/mL ( Figure S1A) and for caspofungin to be less than 0.25 mg/mL ( Figure S1B). To determine the cell viability after antifungal drug exposure, we measured diploid and tetraploid colony forming units (CFUs) following 24-hour exposure to antifungal drug treatments representing 1x and 10x the MIC ( Figure 1A & B). While we detected minor, but statistically significant, CFU reductions following exposure to fluconazole compared to the no-drug treatment, we observed that growth rates in fluconazole were substantially slower (S1C & S1D). We found that exposure to caspofungin reduced CFUs more severely than fluconazole, in both diploid and tetraploid C. albicans ( Figure 1A & B). Interestingly, the low concentration of caspofungin inhibited cell viablility to a greater degree than the high concentration, a phenomenon coined the "the paradoxical effect of echinocandins" (Wagener and Loiko 2017). Importantly, there were differences in total CFUs for diploid and tetraploid C. albicans in the no-drug treatment, where there were 50% fewer CFUs for tetraploids compared to diploids. We saw a similar reduction for fluconazole, where there were $50% fewer tetraploid CFUs than diploid. In contrast, caspofungin impacted tetraploid CFUs more severely, with a 90% reduction compared to diploid ( Figure 1C). This demonstrates that antifungal drugs impact diploid and tetraploid C. albicans differently, despite having similar MICs and suggests there is a drug x ploidy interaction impacting cell viability.

Drug x ploidy interactions impact the rates of genome instability
Given that short-term antifungal drug exposure impacted cell viability of diploid and tetraploid C. albicans, we next investigated if antifungal drug exposure increased mutation rates. Since mutations range from small-scale, (i.e., single nucleotide mutations) to large-scale (i.e., chromosome events including recombination and aneuploidy), we measured both point mutation and LOH rates in diploid and tetraploid C. albicans following exposure to fluconazole and caspofungin. To assess the point mutation rate, we used a his4-G929T (located on Chr4) reversion assay (Forche et al. 2011) and found that the reversion rate in the no-drug treatment was extremely low for both diploid ($4 · 10 210 events/cell division) and tetraploid ($12 · 10 210 events/cell division) C. albicans. Exposure to antifungal drugs had minor impacts on point mutation rates (Table S2). For example, exposure to high concentrations of fluconazole and caspofungin increased the tetraploid mutation rate only by twofold. Due to the rarity of revertants, and the fungicidal nature of caspofungin, it was technically challenging to capture enough reversion events to determine statistical significance.
While point mutations are rare, large-scale genomic rearrangements occur frequently in C. albicans and are easily detected through loss-ofheterozygosity assays (Lea and Coulson 1949;Forche et al. 2011;Hickman et al. 2015). We found that diploid C. albicans exposed to caspofungin, but not fluconazole, showed significant increases in LOH rates compared to the no-drug treatment (Figure 2A). In contrast, for tetraploid C. albicans, exposure to either drug resulted in significant increases in LOH rates compared to the no-drug treatment ( Figure 2B). However, the degree to which the specific antifungal drug elevated LOH rates differed: fluconazole increased the LOH rate by $threefold, whereas caspofungin increased the LOH rate by $100-fold.
We also observed significant differences in LOH rates between diploid and tetraploid C. albicans ( Figure 2C). In the no-drug treatment, the rate of LOH was 1.0x10 25 events/cell division in the diploid state (Figure 2A, dashed line), whereas the tetraploid LOH rate was 2.9 x10 24 events/cell division ( Figure 2B), a difference of $30-fold ( Figure 2C), and consistent with previous reports (Hickman et al. 2015). If antifungal drugs elevate LOH rates to the same extent in both diploid and tetraploid C. albicans, then we hypothesized that there would also be a 30-fold difference across all drug treatments. Indeed, this is what we observed in caspofungin, where the tetraploid LOH rate was $40-fold higher than that of diploid ( Figure 2C). In contrast, the fluconazole-induced LOH rate was 60-fold higher in tetraploids compared to diploids ( Figure 2C). In addition to the individual contributions from ploidy (P = 0.0024, two-way ANOVA) and drug treatment (P = 0.0011, two-way ANOVA), we detected a significant interaction between these two variables ('interaction' P = 0.0019, two-way ANOVA; Figure S2) on LOH rates.
Regardless of ploidy, we observed that caspofungin significantly increased LOH rates, a surprising result since there is no evidence for its genotoxicity, yet some evidence that it induces apoptosis (Hao et al. 2013). To test whether the caspofungin-induced increase in LOH rates was due to genotoxicity of this drug, or simply an indirect consequence of cell wall stress, we exposed diploid and tetraploid C. albicans to calcofluor white, a cell-wall damaging agent (Walker et al. 2008). Importantly, both calcofluor white and caspofungin activate the MKC1 pathway to elicit a downstream stress response (Munro et al. 2007;Walker et al. 2008;Heilmann et al. 2013). We found that at concentrations of 100 mg/mL, calcofluor white inhibited cell viability to a similar degree as 0.25 mg/mL caspofungin (Fig. S3) and elevated LOH rates in diploid ( Figure 2A) and tetraploid ( Figure 2B) C. albicans. The calcofluor white-induced LOH rates were comparable to what we observed in caspofungin, and suggests that caspofungin-induced LOH is an indirect consequence of cell wall stress rather than an inherent genotoxicity. Surprisingly, we saw that calcofluor white elicited a ploidy-specific impact on LOH rates, in which diploids were more severely impacted than tetraploids ( Figure 2C). This was unexpected since caspofungin did not have any ploidy-specific responses.
Antifungal exposure results in modest gains of total DNA content in diploids Since exposure to antifungal drugs increased LOH rates, we examined whether there were corresponding changes in total DNA content. Following selection for LOH, we isolated single colonies and measured their total DNA content by flow cytometry (Figure 3A, 2-DOG). We found some changes to total DNA content following LOH selection ( Figure S4). To quantify if these changes were gains or losses, we compared the G1 peak size of LOH derivatives to the mean G1 peak size of diploid controls. The no-drug, high caspofungin, and calcofluor white treatments did not elicit a meaningful frequency of LOH derivatives displaying changes to total DNA content ( Figure 3B). However, high fluconazole and low caspofungin resulted in 35% and 20% of LOH derivatives displaying changes to total DNA content, respectively. Furthermore, when we detected changes, all were modest gains to DNA content ( Figure 3B). These results suggest that chromosome mis-segregation is associated with fluconazole-induced LOH, consistent with previous findings (Forche et.al, 2011). Additionally, our results with caspofungin-induced LOH imply that most of these LOH events did not occur through chromosome mis-segregation mechanisms.
It is possible that the changes to total DNA content are a general response to antifungal drug stress, and not specifically associated with LOH events. Therefore, we measured total DNA content of single colony isolates obtained in the absence of LOH selection ( Figures 3A, SDC and S4). The no-drug, low fluconazole, and high caspofungin treatments did not elicit a meaningful frequency of isolates displaying changes to total DNA content ( Figure 3C). However, high fluconazole and low caspofungin resulted in 8% and 20% of isolates displaying gains in total DNA content, respectively. Calcofluor white resulted in $14% of isolates with changes to total DNA content, in which half were gains and half were losses ( Figure 3C).
Next, we investigated whether there was a relationship between the frequency of changes to total DNA content (with and without LOH selection), and the mutagenicity of the specific drug treatment. To do this, we plotted the total fraction of isolates displaying changes to total DNA content (gains + losses) with LOH selection ( Figure 3D, filled circles) and without LOH selection ( Figure 3D, open circles) by the rate LOH for each drug treatment. If there is no relationship and these two variables are independent of each other, we hypothesize that there would not be a difference in the fraction of isolates with changes to total DNA content between selection regimes. Indeed, this is the case for the no-drug, caspofungin, and low fluconazole treatments ( Figure  3D), despite the 100-fold increased LOH rate in caspofungin. While high fluconazole did not significantly alter the rate of LOH, there were high frequencies of isolates displaying total DNA content changes in the presence and absence of LOH selection. However, LOH derivatives had a substantially higher fraction of isolates displaying changes in total DNA content ( Figure 3D, filled purple circle), suggesting that LOH events and changes to total DNA content were not independent in this treatment. Figure 1 Diploid and Tetraploid C.albicans are sensitive to antifungals. A) Diploid colony forming units (CFUs) following 24-hr exposure to 1mg/mL or 10mg/mL fluconazole ('FLU', light and dark purple bars) and 0.25mg/mL or 2.5mg/mL caspofungin ('CAS', light and dark green bars) treatments. Bars represent the mean of 3 independent experiments (black symbols), and the error bars indicate +/2 SEM. The dashed line and shaded box represent the mean and +/2 SEM of the no-drug treatment. Asterisks indicate drug treatments that differ significantly from the no-drug treatment ( Ã P , 0.05; ÃÃ P , 0.01 ; unpaired Mann-Whitney U-test). B) Tetraploid colony forming units. Data are displayed and analyzed the same as (A). C) The relative reduction of tetraploid CFUs compared to diploid CFUs. Data are displayed and analyzed similarly to (A).
Antifungal exposure results in drastic losses of total DNA content in tetraploids While we observed a relationship between changes in total DNA content and LOH in high fluconazole for diploids, we wanted to determine whether antifungals affected this relationship in tetraploids. Following LOH selection, we isolated single colonies and measured total DNA content by flow cytometry. For tetraploids, we found substantial deviations in total DNA content occurring across all treatments (Fig. S4).
We quantified these deviations similarly to our diploid analysis, and compared the G1 peak size of the LOH derivatives to tetraploid and diploid controls to categorize changes to total DNA content into two groups: 1) LOH derivatives with G1 peaks between 2C and 4C, or above 4C (indicating aneuploidy), and 2) LOH derivatives with $2C content, (indicating a halving of total tetraploid DNA content). LOH selection led to reductions in total DNA content, with 52-65% of LOH derivatives showing aneuploid genome contents for all treatments, except for Figure 3 Diploids undergo small change in total genome size regardless of LOH selection. A) Experimental schematic. Cultures were incubated in the presence or absence of drugs for 24hrs and plated on 2-DOG to select LOH events or on SDC to determine determine total viable growth. At least 48 single colonies were picked from each plate type, prepared, and analyzed for flow cytometry. B) Frequency of diploid LOH derivatives (2-DOG) with deviations to total DNA content after 24-hr drug treatments. Isolates with G1 peaks greater than the 2C control were considered 'gains' in DNA content, and those with G1 peaks less than the 2C control were considered 'losses'. C) Frequency of diploid isolates with deviations to total DNA in the absence of LOH selection (SDC) after 24-hr drug treatments. Data are represented similarly to (B). D) The frequency of deviations to total DNA content associated with the absence (open symbols) or presence (filled symbols) of selection for LOH events is plotted on the x-axis. The y-axis indicates the LOH rate for each drug treatment (represented by their respective colors).

Figure 2
Ploidy and antifungal drug-specific impacts on LOH in C. albicans. A) Diploid GAL1 LOH rates following 24-hr exposure to 1mg/mL or 10mg/mL fluconazole ('FLU', light and dark purple bars); 0.25mg/mL or 2.5mg/mL caspofungin ('CAS', light and dark green bars); and 100 mg/mL calcofluor white ('CW', gray bar) treatments. Bars represent the mean of 3 independent experiments (black symbols), and the error bars indicate +/2 SEM. The dashed line and shaded box represent the mean and +/2 SEM of the no-drug treatment. Asterisks indicate drug treatments that differ significantly from the no-drug treatment ( Ã P , 0.05; ÃÃ P , 0.01 ; unpaired Mann-Whitney U-test). B) Tetraploid GAL1 LOH. Analysis was performed the same as in (A). C) The fold-increase in GAL1 LOH rate in tetraploid relative to diploid C. albicans. Data are displayed and analyzed similarly to (A).
high caspofungin, where only 20% are in this category ( Figure 4A). Furthermore, a small number of LOH derivatives had DNA content at $2C, though this varied depending on drug treatment. For example, 2% of LOH derivatives were $2C in the high fluconazole compared to $18% in calcofluor white. These results support the model that tetraploids selected for LOH events are accompanied by reductions in DNA content (Hickman et al., 2015).
While it is likely that the changes to total tetraploid DNA content are specifically associated with LOH events, these changes may be a general response to antifungal drugs. Therefore, we measured total DNA content of single colony isolates obtained in the absence of LOH selection ( Figure S4). In the no-drug treatment, there were only a few minor deviations from tetraploidy, and a similar fraction detected in caspofungin, although these were mostly losses ( Figure 4B). For fluconazole, 20-30% of isolates had aneuploid genome content and a small fraction had converged to diploidy. This supports the premise that fluconazole promotes ploidy instability, given that we also observed changes to DNA content in diploids. Interestingly, in calcofluor white, there were $40% of isolates with $2C DNA content. This rapid convergence to diploidy was notable given the short exposure time and absence of LOH selection. Furthermore, this was specific to calcofluor white and not a general response to cell wall perturbations, as we did not observe similar changes in caspofungin.
For each drug treatment, we plotted the total fraction of isolates displaying changes to total DNA content (gains + losses) with LOH selection ( Figure 4C, filled circles) and without LOH selection ( Figure  4C, open circles) by the rate of LOH. Deviations to total tetraploid DNA content were frequently (10-62%) detected in the absence of LOH selection across all treatments ( Figure 4C), and is in contrast to diploids ( Figure 3D). Furthermore, tetraploid instability was substantially elevated in isolates selected for LOH events. For no-drug and low caspofungin treatments, we saw a strong relationship between devations to total DNA content and selection for LOH; for fluconazole and high caspofungin treatments there was a moderate relationship; and for calcofluor white there was no relationship. This was due, in part, to the high ploidy instability observed in the absence of LOH selection and may also reflect LOH events that are not due to chromosome mis-segregation.

DISCUSSION
Both stress-induced mutagenesis and organismal ploidy impact mutation rates and spectra, but these two phenomena are often studied in isolation. Here, we investigated how antifungal drug stress impacts mutation rates of small-and large-scale genomic events in diploid and tetraploid C. albicans. We found a significant interaction between ploidy and antifungal drug stress on LOH rates ('interaction' P = 0.0160, two-way ANOVA, Figure S2). This was surprising, since we expected that the relative increase in drug-induced LOH rate would be similar between diploid and tetraploid C. albicans, despite inherent differences in their baseline rates. In fact, we observed this for caspofungin which substantially elevated genome instability for both ploidy states. However, fluconazole treatments impacted tetraploid C. albcians more severely than diploid C. albicans. Our findings support a model in which the magnitude of stress-induced mutagenesis is the result of a drug x ploidy interaction.
While C. albicans genomic responses to fluconazole has been wellstudied (Forche et al. 2011;Harrison et al. 2014;Hickman et al. 2015;Popp et al. 2017), the genomic responses to caspofungin are largely uninvestigated, as most studies regarding caspofungin focus on mechanisms of resistance (Munro et al. 2007;Pfaller et al. 2008;Walker et al. 2008;Lee et al. 2012;Wiederhold 2016;Revie et al. 2018). Here, we found that caspofungin increased the rate of LOH ( Figure 2) and point mutations (Table S2) for both diploid and tetraploid C. albicans compared to the no-drug and fluconazole treatments. Furthermore, exposure to caspofungin increased the frequency at which diploids and tetraploids changed total DNA content (Figure 3, 4 and S4). While an increase in caspofungin-induced genome instability was expected, given that it induces apoptosis (Hao et al. 2013), the degree to which this occurred was unexpected. These findings prompted us to test which is whether caspofungin is genotoxic or if the mutagenicity is an indirect result from stress to the cell wall. To test this, we exposed C. albicans to calcofluor white, a drug that interferes with the proper construction of the fungal cell wall (Walker et al. 2008) and found that genome instability was induced to a similar degree as caspofungin (Figure 2, 3 and 4) (Walker et al. 2008). This result suggests that caspofungin-induced mutagenesis is not due to its direct genotoxicity, but likely is an indirect consequence of perturbations to the cell wall. Intriguingly, caspofungin has a paradoxical effect on fungal cells, a Figure 4 Tetraploids frequently exhibit substantial changes in total DNA content regardless of selection for LOH events. A) Frequency of tetraploid LOH derivatives (2-DOG) with deviations to total DNA content after 24-hr drug treatments. Isolates with G1 peaks greater than the 4C control were considered 'gains' in DNA content, and those with G1 peaks less than the 4C controls were considered 'losses'. A small subset of isolates had G1 peaks that were within +/2 2SD of the 2C control and are indicated by the filled bars. B) Frequency of tetraploid isolates with deviations to total DNA in the absence of LOH selection (SDC) after 24-hr drug treatments. Data are represented similarly to (A). C) The frequency of deviations to total DNA content associated with the absence (open symbols) or presence (filled symbols) of selection for LOH events is plotted on the x-axis. The y-axis indicates the LOH rate for each drug treatment (represented by their respective colors). phenomenon in which supra-MIC concentrations are permissible for proliferation, despite being susceptible at the MIC (Wagener and Loiko 2017). C. albicans restructures its cell wall at supra-MICs of caspofungin to better mediate this stress (Walker et al. 2008;Wagener and Loiko 2017). In our study we also saw the paradoxical effect of caspofungin in both diploid and tetraploid C. albicans viability and LOH rates (Figures 1 and 2). Our findings suggest that the magnitude of caspofungin-induced mutagenesis is dependent on its concentration, in which low concentrations signal a stronger stress response and elicit more genomic perturbations.
Given that there is a drug x ploidy interaction for LOH rates ( Figure S2), we expected to see a similar relationship for how frequently changes to total DNA content occurred. Indeed, this is the pattern that we observed. Caspofungin increased the frequency of changes to total DNA content in both diploid and tetraploid C. albicans (Figures 3 and  4), consistent with its impact on LOH rates ( Figure 2C). However, the directionality of genome size changes depended on ploidy. In diploids these changes were exclusively gains in DNA content, but in tetraploids they were predominantly losses. Fluconazole also elicited gains in DNA content in diploids ( Figure 3) and losses in tetraploids (Figure 4). However, the frequency of these changes was substantially greater in tetraploids, indicative of a ploidy-specific response, consistent with the fluconazole-induced increase in tetraploid LOH rates ( Figure 2C).
There is a relationship between LOH events and changes to total DNA content for diploids exposed to fluconazole (Forche et.al, 2011), in which LOH events are predominantly associated with chromosome mis-segregation, and our findings also support this. Interestingly, this was not the case for caspofungin, where changes to total DNA content were not necessarily associated with selection for LOH, since high ploidy instability was observed even in the absence of LOH selection. For tetraploids, the co-occurrence of LOH and deviations in total DNA content is well established (Hickman et al., 2015;Gerstein et al., 2017), and our findings also support this. A majority of tetraploid LOH derivatives showed reductions in DNA content, regardless of drug treatment and in fact, the absence of drug resulted in the highest frequency of isolates with changes to DNA content. Therefore, the relationship between LOH and changes to DNA content is strongest in the no-drug treatment and weaker in the drug treatments ( Figure 4C), and suggests that chromosome mis-segregation is not the major mechanism by which LOH occurs in tetraploids exposed to drugs. Taken together, we show that tetraploid C. albicans genomes are more unstable than diploids, and this instability is exaggerated upon antifungal drug exposure.
Tetraploid genome instability has been shown, both theoretically and experimentally, to facilitate rapid adaptation in yeast (Berman and Hadany 2012;Selmecki et al., 2015), and this premise has important implications for the emergence of antifungal drug resistance. Resistance to antifungal drugs can result from point mutations, though the target genes differ between fluconazole and caspofungin (Revie et al. 2018). Furthermore, homozygosis of azole-resistance alleles (i.e., LOH) substantially increases fluconazole MICs compared to heterozygous genotypes (Flowers et al. 2012;Ford et al. 2015). Likewise, aneuploidy is commonly associated with azole resistance (Selmecki et al. 2006(Selmecki et al. , 2009Robbins et al. 2017;Revie et al. 2018;Stone et al. 2019) and recent studies suggest that aneuploidy is beneficial for growth in caspofungin (Yang et al. 2019). Our study explicitly demonstrates that antifungalinduced genome instability impacts tetraploids more profoundly than diploids, and results in extremely high LOH rates and frequency of genome size changes, that are likely to contain aneuploid chromosomes. Thus, tetraploids have the capacity to provide an abundance of genetic variation for natural selection to act upon.