CaMad2 Promotes Multiple Aspects of Genome Stability Beyond Its Direct Function in Chromosome Segregation

Mad2 is a central component of the spindle assembly checkpoint required for accurate chromosome segregation. Additionally, in some organisms, Mad2 has roles in preventing mutations and recombination through the DNA damage response. In the fungal pathogen Candida albicans, CaMad2 has previously been shown to be required for accurate chromosome segregation, survival in high levels of hydrogen peroxide, and virulence in a mouse model of infection. In this work, we showed that CaMad2 promotes genome stability through its well-characterized role in promoting accurate chromosome segregation and through reducing smaller scale chromosome changes due to recombination and DNA damage repair. Deletion of MAD2 decreased cell growth, increased marker loss rates, increased sensitivity to microtubule-destabilizing drugs, and increased sensitivity to DNA damage inducing treatments. CaMad2-GFP localized to dots, consistent with a role in kinetochore binding, and to the nuclear periphery, consistent with an additional role in DNA damage. Furthermore, deletion of MAD2 increases growth on fluconazole, and fluconazole treatment elevates whole chromosome loss rates in the mad2∆/∆ strain, suggesting that CaMad2 may be important for preventing fluconazole resistance via aneuploidy.


Introduction
Genome instability arises from many different causes, including mistakes in DNA replication, unrepaired DNA damage, and mis-segregation of chromosomes. These processes result in genetic variation among cells in a population. Typically, most mutations and aneuploidy events are detrimental to cells. However, in certain stressful circumstances, mutations and aneuploid chromosomes can have a selective advantage. High mutation rates are observed in human cancers and in several fungal pathogens, especially during the infection process. For example, large scale chromosomal changes rapidly generate genome diversity in Candida albicans following exposure to the mouse oral cavity [1]. Genome instability is also associated with antifungal drug resistance. In C. albicans, aneuploid cells with an isochromosome-two extra copies of 5L-have fluconazole resistance, and other specific aneuploid chromosomes can provide cross-resistance to multiple chemotherapeutic and antifungal drugs [2,3]. Aneuploidy also drives fluconazole resistance in the fungal pathogen Cryptococcus neoformans [4]. Therefore, cells must balance genome maintenance with mechanisms that promote genetic diversity in order to thrive in favorable conditions and survive in stressful environments.
Genome maintenance depends on the function of the spindle assembly checkpoint (SAC). A major role of the SAC is to detect unattached kinetochores. At unattached chromosomes, the mitotic checkpoint complex (MCC) binds to the anaphase-promoting complex (APC/C) to prevent cells from dividing before attachments are made. In the complex, Mad2 binds Cdc20, and together with BUBR1 (Mad3 in yeasts), prevents APC/C activity. Once the chromosomes are correctly aligned and spindle fibers are attached to the kinetochore, the MCC is disassembled, allowing APC/C activity and progression into anaphase through degradation of securin [5]. In addition to its roles in chromosome segregation, the SAC has recently been shown to respond to DNA damage and replication stress. The DNA damage response kinases ATR and CHK1 have been functionally linked with SAC components, including Mad2. SAC components delay metaphase in the presence of DNA damage. In Caenorhabditis elegans germ cells, Mad2 relocalizes to the nuclear periphery in a DNA damage response-dependent manner [6]. Similarly, relocalization of SAC components in response to DNA damage occurs in human cells [6]. In budding yeast, Mad2 has also been shown to have additional roles outside of mitosis to control DNA synthesis and allow cell survival in response to replication stress [7].
Mad2 is a central component of the spindle assembly checkpoint and is required for inhibition of APC/C. Mad2 exists in two forms, open and closed. The closed form is the form that allows Mad2 complex formation in the MCC and occurs when Mad2 is bound to Mad1 or Cdc20 [5]. In interphase, Mad1-Mad2 complexes are present at the nuclear pore in humans [8], budding yeast [9], and the hyphal fungus Aspergillus nidulans [10]. As cells begin to divide, Mad2-GFP localizes to kinetochores in response to non-attached kinetochores/mitotic block [11], but Mad1 and Mad2 remain at the nuclear periphery through mitosis in budding yeast when kinetochores are attached properly [9,12].
Mad2 binding to the kinetochore in yeast depends on two kinases, Bub1 and Mps1 Mph1 . Mps1 Mph1 kinase activity is required for assembly of the Mad1-Mad2 core complex [5]. Bub1 is an essential SAC kinase and is required to recruit other SAC proteins [5]. In humans, the MCC is recruited by two separable pathways, KNL1-Bub3-Bub1 and Rod-Zwilch-Zw10 (RZZ). Recent evidence shows that efficient mitotic checkpoint signaling in human cells depends on the activity of both Bub1 and the RZZ complex [13]. However, yeast lack RZZ and require Bub1 for detection of unattached kinetochores. Bub1 binds to the kinetochore in connection with the KNL1 Spc105/Spc7 kinetochore protein phosphorylated by Mps1 Mph1 and Bub3. Heterodimers of Mps1 Mph1 and Bub1 together with Mad1, Mad2 and Mad3 are sufficient to trigger metaphase arrest in both budding and fission yeast [14].
Overall, the SAC is well-conserved in diverse organisms. However, differences in functions between have been characterized, and some lineages, such as species in the bipolar budding yeast genus Hanseniaspora, have lost numerous genes in the SAC, including MAD1 and MAD2 [15]. C. albicans has many conserved functions and processes in common with other well-characterized yeasts. However, there are also several important differences in the genome dynamics of C. albicans. One notable difference is the parasexual process for recombination-based genetic diversity (rather than meiosis) [16]. Another difference is the presence of an H2A variant that cannot be phosphorylated by the Bub1 checkpoint protein [17]. Additionally, CaMps1 has roles in SAC activation but also has roles in morphological transitions from yeast to hyphal forms and in the oxidative stress response [18]. In C. albicans, CaMad2 has previously been shown to be important for accurate chromosome segregation, survival in high levels of hydrogen peroxide, and virulence in a mouse model of infection [19]. However, questions about the roles of CaMad2 in preventing sub-chromosomal genome instability and in evolving drug resistance remain unanswered.
In this work, we studied the roles of CaMad2 to learn more about the evolutionary conservation of the function of the protein in function and to better characterize the role of CaMad2 in promoting genome stability in the fungal pathogen, C. albicans. We found that CaMad2 promotes genome stability through its well-characterized role in promoting accurate chromosome segregation and through reducing smaller scale chromosome changes due to recombination and DNA damage repair. Additionally, the most commonly observed localization pattern of CaMad2-GFP at the nuclear periphery is consistent with a role in DNA damage response. Together, these results indicate that the DNA damage response through Mad2 is conserved in C. albicans. Interestingly, we also found that similar to other strains with altered chromosome segregation checkpoint function, deletion of MAD2 increases growth on fluconazole, suggesting that disruption of any of these genes may promote the evolution of drug resistance in C. albicans.

Strain Construction
C. albicans strains are listed in Table S1. Lithium acetate transformation of linearized plasmids or PCR products with at least 70 bp of homology to the targeted gene were used for strain construction. Briefly, strains to be transformed were inoculated in YPA-glucose and grown at 30 • C for [16][17][18] h. Overnight cultures were then diluted 1:166 in YPA-glucose and grown at 30 • C for 3-4 h. Cells were washed with water, then TELiAc (10 mM Tris pH 7.5, 1 mM EDTA, 100 mM LiAc), and incubated in TELiAc with transformation DNA and 50 µg of sheared salmon sperm DNA for 30 min. Four volumes of PLATE mix (40% PEG, 10 mM Tris pH 7.5, 1 mM EDTA, 100 mM LiAc) were then added, and the transformation mix was incubated for 12-18 h at 20-24 • C. Transformations were heat-shocked at 42 • C for 1 h and then plated on selective media. Strains were checked by PCR of genomic DNA.
For the construction of the MAD2 heterozygous and homozygous deletion strains, we replaced the MAD2 gene with URA3, HIS1, and/or ARG4 genes. We amplified the markers using primers with flanking 5' and 3' ends containing 70 bp homology to the targeted genes of interest from plasmids pGEM-URA3, pGEM-HIS1, and pRS-ARG4∆Spe1 [20]. Correct insertions and the absence of native MAD2 gene in homozygous deletion strains were verified with PCR. The MAD2/MAD2-GFP strain was tagged at the C-terminus with PCR-mediated transformation. Primers containing at least 70 bp homology to the targeted gene were used to amplify the GFP and marker from the C-terminal GFP tagging plasmids with HIS1 markers [21]. Fusion of GFP to the MAD2 gene was confirmed by PCR. To construct the PCK1p-CSE4 strains, the URA3-PCK1p-CSE4 plasmid [22] was linearized with EcoRV. Integration into the CSE4 upstream region was confirmed by PCR.

Growth Analysis
Strains were inoculated into 5 mL liquid YPA-glucose media and growth overnight at 30 • C with shaking. Overnight cultures were diluted into liquid media: YPA-glucose, YPA-succinate, YPA-glucose-nocodazole (50 µM), YPA-glucose-thiabendazole (222 µg/mL), YPA-glucose-hydrogen peroxide (1 mM), or YPA-glucose-fluconazole (1 µg/mL). Then, 150 µL of the 10 −3 dilution was plated into a clear 96-well plate. Growth was determined by measuring the optical density of cultures (four cultures per condition) at 600 nm. Data was collected every 15 min using a Tecan Sunrise at 30 • C with shaking between cycles for 24 h. Data were analyzed using a custom Excel Macros program (a template version is included as File S1).

Serial Dilution and Plate Spotting
Strains were inoculated into 5 mL liquid YPA-glucose media and grown overnight at 30 • C with shaking. These strains were serially diluted using a deep 96 well plate from 10 −1 to 10 −6 in sterilized water. Dilutions were then spotted onto YPAD (yeast peptone adenine dextrose agar), YPAD-fluconazole (1 µg/mL), and YPAD-nocodazole (50 µM) agar plates. For UV light treatment, strains were exposed to 100 mJ/cm 2 or 200 mJ/cm 2 using a Stratalinker crosslinker (Stratagene, La Jolla, CA, USA). Plates were incubated at 30 • C for 24 h. Plates shown are representative of at least three biological replicates.

Minimum Inhibitory Concentration Assay
Strains were streaked for single colonies and grown on YPAD at 30 • C for 24 h. The inoculum was prepared by diluting twenty single colonies from each plate (~0.5 mm each) into 5 mL of sterile water, adjusting the sample to match a 0.5 McFarland standard, and then further diluting to 1:10 in sterile water. Fluconazole was diluted in two-fold from 128 µg/mL to 0.25 µg/mL 2× RPMI-MOPS pH 7, 2% glucose [23]. Uridine was added at a final concentration of 0.1 mg/mL for strains lacking URA3. 100 µL 2× media with fluconazole was combined with 100 µL of the inoculum for final fluconazole concentrations from 64 µg/mL to 0.125 µg/mL, and the plate was incubated at 37 • C for 24 h. Optical density at 600 nm was read with a Tecan Sunrise plate reader (Tecan, Männedorf, Switzerland). At least six biological replicates were tested for each strain.

Fluctuation Analyses of Chromosome Loss Rates
Fluctuation analysis of loss rates was performed similar to prior studies [24] using the method of the median [25]. Briefly, strains were streaked for single colonies and grown on YPAD for 2 days at 30 • C. Per strain, 8 independent colonies were inoculated into 500 µL liquid media (YPA-glucose, YPA-succinate, or YPA-glucose-fluconazole (1 µg/mL)), and grown overnight at 30 • C with shaking. Cultures were diluted with sterile water and plated onto nonselective YPAD for total cell counts and selective media (5-FOA for URA3 loss). Plates were incubated at 30 • C for 1-3 days, and colony counts were used to calculate the rate of marker URA3 loss per cell division.

SNP-RFLP
At least 8 individual colonies that lost URA3 were isolated from 5-FOA plates after incubation in the fluctuation analysis. Colonies were streaked on 5-FOA plates and incubated at 30 • C for 48 h. Following genomic DNA extraction, PCR was performed on the right (5R) and left (5L) ends of Chr5 using primers 10080A and 2340/2493; the right (3R) and left (3L) ends of Chr3 using primers 2091/2447 and 1765/2519; and the right (1R) and left (1L) ends of Chr1 using primers 2106/2441 and 1799/2450. Restriction digests were performed on resulting PCR products with AluI and Taq1 for Chr5, DdeI and Hpy188I for Chr3, and AluI and PstI for Chr1. Restriction digests were incubated at 37 • C (AluI, DdeI, Hyp188I, and PstI) or 60 • C (Taq1) for 8 h. Restriction digests were run on 3% agarose gels to check for SNP homozygosity or heterozygosity (digested or non-digested alleles based on SNP present within PCR) [26].

Reverse Transcriptase qPCR
Strains were inoculated into YPA-glucose and grown at 30 • C for 16-18 h. Cultures were then diluted 1:100 into YPA-glucose and grown at 30 • C for 4 h. RNA was prepared using the MasterPure yeast RNA purification kit (Epicentre, Madison, WI, USA) according to the manufacturer's instructions. RNA was treated with DNase (Epicentre) to remove contaminating genomic DNA. cDNA was prepared using the ProtoScript M-MuLV First Strand cDNA Synthesis Kit (New England Biolabs, Ipswich, MA, USA) according to the manufacturer's instructions with oligo dT primers. cDNA was measured by qPCR using the Luna Universal qPCR Master Mix (New England Biolabs) with a Rotor-Gene cycler (Qiagen, Germantown, MD, USA) according to the manufacturer's instructions. Expression was calculated as the amount of cDNA from the gene of interest relative to the amount of TEF1 cDNA in the same sample using the second-derivative maximum to determine C T values.

Microscopy
For the localization study conducted, strains were inoculated into SDC-glucose and grown at 30 • C for [16][17][18] h. The overnight cultures were then diluted 1:10 into SDC-glucose, SDC-glucose-nocodazole (50 µM), and SDC-glucose-H 2 O 2 (1.2 mM), and then incubated for 3 h at 30 • C with shaking. DNA was stained with 5 µg/mL Hoechst 15 min and resuspended in 1× PBS prior to imaging. GFP filtered sets as appropriate using constant exposure times and scaling of the strain. Cells were imaged with a Zeiss LSM confocal microscope. Images were acquired using Zeiss ZEN software.

Deletion of CaMAD2 Reduces Growth
We independently constructed two different CaMAD2 deletion strains in derivatives of C. albicans SC5314. In one strain, CaMAD2 alleles were replaced by HIS1 and ARG4. In the other strain, CaMAD2 alleles were replaced by HIS1 and URA3. First, we asked whether the deletion of CaMAD2 reduced total growth and maximal growth rate in C. albicans by measuring optical density. Total growth was quantified by area under the curve (AUC), which varies based on lag time (when growth starts), growth rate throughout the experiment, and the carrying capacity (maximum density of the culture). Throughout, we saw a correlation between AUC and maximal growth rate, with AUC being somewhat more sensitive, as it combines several quantitative measures of growth. We found reduced growth relative to the wild-type MAD2 control in both mad2∆/∆ strains and the bub1∆U/∆L strain, as indicated by AUC ( Figure 1A) and by maximal growth rates ( Figure S1). Heterozygous deletion of MAD2 has shown haploinsufficient phenotypes in other organisms [27,28]. Therefore, we also tested the growth of two independently constructed CaMAD2 heterozygous strains with different markers used to replace one allele of CaMAD2. We saw moderately reduced growth as measured by AUC ( Figure 1A) and maximal growth rate ( Figure S1) in a MAD2/∆H strain in which one copy of MAD2 was replaced by HIS1 compared to the wild-type MAD2 control. However, we did not see significantly reduced growth in a MAD2/∆U strain in which one copy of MAD2 was replaced by URA3. One possible explanation for the difference in the two heterozygous strains is allele variation. In earlier versions of the C. albicans genome annotation, one allele of CaMAD2 was thought to have a six base pair insertion and was labeled as the long-allele [19]. More recent sequencing and assembly of the C. albicans SC5314 genome indicated that both alleles were the same length, but had differences in the coding sequence. Therefore, although the long-allele seems to be an artifact from a sequencing error, the alleles contain non-synonymous variation in the amino acid composition and might have functional differences [29]. CaMAD2 alleles were replaced by HIS1 and URA3. First, we asked whether the deletion of CaMAD2 reduced total growth and maximal growth rate in C. albicans by measuring optical density. Total growth was quantified by area under the curve (AUC), which varies based on lag time (when growth starts), growth rate throughout the experiment, and the carrying capacity (maximum density of the culture). Throughout, we saw a correlation between AUC and maximal growth rate, with AUC being somewhat more sensitive, as it combines several quantitative measures of growth. We found reduced growth relative to the wild-type MAD2 control in both mad2∆/∆ strains and the bub1∆U/∆L strain, as indicated by AUC ( Figure 1A) and by maximal growth rates ( Figure S1). Heterozygous deletion of MAD2 has shown haploinsufficient phenotypes in other organisms [27,28]. Therefore, we also tested the growth of two independently constructed CaMAD2 heterozygous strains with different markers used to replace one allele of CaMAD2. We saw moderately reduced growth as measured by AUC ( Figure 1A) and maximal growth rate ( Figure S1) in a MAD2/∆H strain in which one copy of MAD2 was replaced by HIS1 compared to the wild-type MAD2 control. However, we did not see significantly reduced growth in a MAD2/∆U strain in which one copy of MAD2 was replaced by URA3. One possible explanation for the difference in the two heterozygous strains is allele variation. In earlier versions of the C. albicans genome annotation, one allele of CaMAD2 was thought to have a six base pair insertion and was labeled as the long-allele [19]. More recent sequencing and assembly of the C. albicans SC5314 genome indicated that both alleles were the same length, but had differences in the coding sequence. Therefore, although the long-allele seems to be an artifact from a sequencing error, the alleles contain non-synonymous variation in the amino acid composition and might have functional differences [29].  Optical density (600 nm) was collected every 15 min using a Tecan incubator/plate reader. Results were calculated from the area underneath the curve of growth. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least four biological replicates, each with four cultures per condition. Significant differences are indicated with letters (ordinary one-way ANOVA, p < 0.0001; Tukey multiple comparison, p < 0.01 for letter difference). (B). C. albicans strains with/or without the deletion of MAD2 or BUB1 were grown overnight in YPA-glucose media, diluted with water, and plated onto YPAD for total cell counts and 5-FOA plates to select for URA3 loss. Colony counts were used to calculate the rate of loss per cell division. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least three biological replicates, each with eight cultures per condition. Significant differences are indicated with letters (ordinary one-way ANOVA, p < 0.0001; Tukey multiple comparison, p < 0.01 for letter difference). (C). Alleles on distal chromosome positions in 5-FOA resistant colonies were characterized by SNP-RFLP. Each 5-FOA colony tested was characterized as homozygous for both alleles or heterozygous.

Deletion of CaMAD2 Increases Genome Instability
Next, we characterized the mechanisms reducing cellular fitness by measuring genome instability. Deletion of CaMAD2 was previously shown to result in increased growth on sorbose and in homozygosis of the mating-type locus on chromosome 5, relative to strains with intact MAD2 [19]. The sorbose selection method primarily detects chromosome 5 monosomy rather than all types of loss of heterozygosity, as sorbose resistance occurs due to reduced copy number of several genes on chromosome 5. Given the potential role of Mad2 in DNA damage responses in addition to chromosome segregation, we examined the effect of CaMAD2 deletion using 5-FOA counter-selection against the URA3 marker which allows detection of multiple types of loss of heterozygosity as the URA3 gene is confined to a small region on the chromosome where it is located. Using the mad2∆H/∆U strain heterozygous for URA3 on chromosome 1, we observed elevated marker loss rates per cell division ( Figure 1B). We then used SNP-RFLP to determine whether 5-FOA resistance was due to smaller-scale loss of heterozygosity events, such as mutation or recombination, or whole chromosome loss resulting in whole chromosome homozygosis. For the wild-type MAD2-containing strain, we observed marker loss due to a mix of smaller scale events and whole chromosome loss. For the mad2∆/∆ strain, we detected only smaller scale events as all of 5-FOA R colonies tested were heterozygous at one or both chromosome ends ( Figure 1C and Table S2). Therefore, the elevated URA3 marker loss rate observed in strains lacking MAD2 was due to elevated mutation or recombination rather than whole chromosome loss. The MAD2/∆U heterozygous strain did not show increased marker loss ( Figure 1B), consistent with the absence of growth-related phenotypes. The URA3 marker loss rate on chromosome 4 for bub1∆U/∆L was similar to the homozygous mad2 deletion strain ( Figure 1B).

More Kinetochore-Microtubule Attachments Per Centromere Does Not Affect the Requirement for MAD2
The canonical role of Mad2 is to detect unattached kinetochores as part of the spindle assembly checkpoint. In C. albicans, each chromosome typically has one kinetochore-microtubule attachment per centromere. Overexpression of CSE4, encoding the centromeric histone CENP-A, results in an increased number of kinetochore-microtubule attachments per centromere. Approximately 60% of centromeres have two kinetochore-microtubule attachments when CSE4 is overexpressed [30]. We reasoned that higher numbers of kinetochore-microtubule attachments per centromere might alter the requirement for the spindle assembly checkpoint as there are more potentially unattached kinetochores. Therefore, we next tested whether deletion of MAD2 has synergistic phenotypes with increased numbers of kinetochore-microtubule attachments per centromere. Growth of the cells in succinate activated the PCK1 promoter driving CSE4, resulting in high levels of CSE4 expression relative to growth in glucose (repressing the PCK1 promoter upstream of one copy of CSE4) ( Figure S2A). We did not observe any synergistic effects of MAD2 deletion and CSE4 overexpression as measured by total growth ( Figure S2B) or maximal growth rate ( Figure S2C).
Importantly, marker loss rates of a heterozygous URA3 marker on chromosome 3 were similar in mad2∆/∆ cells overexpressing CSE4, and mad2∆/∆ cells with normal levels of Cse4 ( Figure 2). Consistent with the results seen with chromosome 1, URA3 marker loss rates on chromosome 3 were elevated in the mad2∆/∆ strain compared to the strain with MAD2 ( Figure 2). For the strains with MAD2, we observed marker loss due to a mix of smaller scale events and whole chromosome loss events with and without overexpression of CSE4. For the mad2∆/∆ strains with and without CSE4 overexpression, we detected only smaller scale events, as all of the 5-FOA R colonies tested were heterozygous at one or both chromosome ends ( Figure 2B and Table S2). Regardless of whether CSE4 was overexpressed, the elevated URA3 marker loss rate in strains lacking MAD2 was due to increased mutation or recombination rates rather than whole chromosome loss. Consistent with the results seen with chromosome 1, URA3 marker loss rates on chromosome 3 were elevated in the mad2∆/∆ strain compared to the strain with MAD2 ( Figure 2). For the strains with MAD2, we observed marker loss due to a mix of smaller scale events and whole chromosome loss events with and without overexpression of CSE4. For the mad2∆/∆ strains with and without CSE4 overexpression, we detected only smaller scale events, as all of the 5-FOA R colonies tested were heterozygous at one or both chromosome ends ( Figure 2B and Table S2). Regardless of whether CSE4 was overexpressed, the elevated URA3 marker loss rate in strains lacking MAD2 was due to increased mutation or recombination rates rather than whole chromosome loss.

Loss of MAD2 Increases Sensitivity to Microtubule-Destabilizing Compounds
Next, we exposed cells with and without Mad2 to treatments that induce microtubuledestabilization and DNA damage effects to further test the hypothesis that CaMad2 functions in direct spindle assembly checkpoint roles and has roles in preventing recombination and other types of sub-chromosomal genome protection mechanisms. We first confirmed the role of CaMad2 in the spindle assembly checkpoint by using nocodazole and thiabendazole. Nocodazole treatment destabilizes microtubules and increases unattached kinetochores. In several organisms, including C. albicans, deletion of MAD2 increases sensitivity to nocodazole [19,31,32]. We examined growth by dilutions and spot plating on media containing nocodazole and quantifying growth in the presence of nocodazole with optical density. We observed moderately reduced growth of the strain containing MAD2 by both measures. We also saw increased sensitivity, defined as a reduced percentage growth compared to the same strain in no drug conditions, to nocodazole for the heterozygous MAD2/mad2∆ and mad2∆/∆ strains with both methods of measuring growth (Figure 3). We also observed sensitivity of the bub1∆/∆ strain on plates containing nocodazole ( Figure 3A). For the AUC growth curve, the mean growth of the MAD2 strain in nocodazole was 57% of the mean growth with no drug, while C. albicans strains with the conditional PCK1 promoter regulating one copy of CSE4 with/or without the deletion of MAD2 were cultured in YPA-glucose (N) or YPA-succinate (OE) media. Cells were grown in each condition overnight, diluted with water, and plated onto YPAD for total cell counts and 5-FOA plates to select for URA3 loss. Colony counts were used to calculate the rate of loss per cell division. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least three biological replicates, each with four cultures per condition. Significant differences are indicated by letters (ordinary one-way ANOVA, p < 0.01; Tukey's multiple comparison test, p < 0.05 for letter difference). (B). Alleles on distal chromosome positions in 5-FOA resistant colonies were characterized by SNP-RFLP. Each 5-FOA colony tested was characterized as homozygous for both alleles or heterozygous.

Loss of MAD2 Increases Sensitivity to Microtubule-Destabilizing Compounds
Next, we exposed cells with and without Mad2 to treatments that induce microtubule-destabilization and DNA damage effects to further test the hypothesis that CaMad2 functions in direct spindle assembly checkpoint roles and has roles in preventing recombination and other types of sub-chromosomal genome protection mechanisms. We first confirmed the role of CaMad2 in the spindle assembly checkpoint by using nocodazole and thiabendazole. Nocodazole treatment destabilizes microtubules and increases unattached kinetochores. In several organisms, including C. albicans, deletion of MAD2 increases sensitivity to nocodazole [19,31,32]. We examined growth by dilutions and spot plating on media containing nocodazole and quantifying growth in the presence of nocodazole with optical density. We observed moderately reduced growth of the strain containing MAD2 by both measures. We also saw increased sensitivity, defined as a reduced percentage growth compared to the same strain in no drug conditions, to nocodazole for the heterozygous MAD2/mad2∆ and mad2∆/∆ strains with both methods of measuring growth (Figure 3). We also observed sensitivity of the bub1∆/∆ strain on plates containing nocodazole ( Figure 3A). For the AUC growth curve, the mean growth of the MAD2 strain in nocodazole was 57% of the mean growth with no drug, while the percentages of growth in nocodazole relative to the no drug treatment were 25% and 23% for the MAD2 heterozygous strains and 29% for the mad2∆/∆ strains ( Figure 3B . C. albicans strains with/or without the deletion of MAD2 were grown in YPA-glucose or YPA-glucose + 50 µM nocodazole at 30 • C for 24 h. Optical density (600 nm) was collected every 15 min using a Tecan incubator/plate reader. Results were calculated from the area underneath the curve of growth. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least three biological replicates, each with four cultures per condition. Significant differences are indicated with letters (two-way ANOVA: strain difference p < 0.0001, nocodazole difference p < 0.0001, interaction p < 0.0001; Tukey multiple comparison, p < 0.05 for letter difference).
Thiabendazole is another drug with microtubule destabilizing mechanisms of action [33]. Thiabendazole reduced growth of the strain containing MAD2, and MAD2/mad2∆ and mad2∆/∆ strains, with the strains lacking at least one copy of MAD2 showing increased sensitivity to thiabendazole using the total growth measurement with optical density ( Figure S3). For the AUC growth curve, the mean growth of the MAD2 strain in thiabendazole was 37% of the mean growth with no drug, while the percentages of growth in thiabendazole relative to the no drug treatment were 9% and 24% for the mad2∆/∆ strains.

Deletion of MAD2 Increases Sensitivity to DNA-Damage Inducing Treatments
We grew strains with and without CaMad2 on media containing hydrogen peroxide. Hydrogen peroxide causes oxidative damage and increases DNA recombination [26]. Hydrogen peroxide (1 mM) reduced growth for the strain containing MAD2 and for the MAD2/mad2∆ and mad2∆/∆ strains. The strains lacking MAD2 showed slightly increased sensitivity to hydrogen peroxide using the total growth measurement with optical density ( Figure 4A). When comparing total growth, the mean MAD2 growth in 1mM hydrogen peroxide was 13% of the mean growth of the control, while the percentages of growth in the same concentration of hydrogen peroxide relative to the no drug treatment were 7% and 9% for the mad2∆/∆ strains. Similarly to our results with total growth, the mad2/∆U strain was not sensitive to hydrogen peroxide (13%), while the mad2/∆H strain was sensitive (6%) ( Figure 4A). We also tested the effect of UV light exposure, another condition causing DNA damage to cells. We examined growth by dilutions and spot plating followed by exposure to 100 mJ/cm 2 or 200 mJ/cm 2 UV light. We observed moderately reduced growth of the strain containing MAD2 and an increased sensitivity to UV light for both mad2∆/∆ strains ( Figure 4B). MAD2 growth in 1mM hydrogen peroxide was 13% of the mean growth of the control, while the percentages of growth in the same concentration of hydrogen peroxide relative to the no drug treatment were 7% and 9% for the mad2∆/∆ strains. Similarly to our results with total growth, the mad2/∆U strain was not sensitive to hydrogen peroxide (13%), while the mad2/∆H strain was sensitive (6%) ( Figure 4A). We also tested the effect of UV light exposure, another condition causing DNA damage to cells. We examined growth by dilutions and spot plating followed by exposure to 100 mJ/cm 2 or 200 mJ/cm 2 UV light. We observed moderately reduced growth of the strain containing MAD2 and an increased sensitivity to UV light for both mad2∆/∆ strains ( Figure 4B).

CaMad2 Localizes to the Nuclear Periphery in C. albicans
To determine the localization of CaMad2, we tagged the C-terminus of CaMad2 with GFP, stained the nuclei with Hoescht DNA stain, and visualized live cells by confocal microscopy. We found localization to several different patterns: one dot colocalized with the nucleus, one dot not colocalized with the nucleus, more than one dot, a complete ring surrounding the nuclear periphery, a partial ring surrounding the nuclear periphery, and a diffuse distribution of GFP in the cell ( Figure  5A). A majority of the cells exhibited a ring surrounding the nuclear periphery ( Figure 5B). The ring morphology was observed in both budding and non-budding cells. The most common alternative localization patterns were the formations of a Mad2-GFP cluster at a single point within the cell (dot) Figure 4. Loss of MAD2 increases sensitivity to the DNA damage inducing agents (A). C. albicans strains with/or without the deletion of MAD2 were grown in YPA-glucose or YPA-glucose + 1 mM hydrogen peroxide at 30 • C for 24 h. Optical density (600 nm) was collected every 15 min using a Tecan incubator/plate reader. Results were calculated from the area underneath the curve of growth. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least two biological replicates, each with four cultures per condition. Significant differences are indicated with letters (two-way ANOVA: strain difference p < 0.0001; hydrogen peroxide difference p < 0.0001; interaction difference p < 0.001; Tukey multiple comparison, p < 0.05 for letter difference). (B). C. albicans strains were spotted in 10-fold serial dilutions on YPAD. For UV treatment, plates were exposed to 100 mJ/cm 2 or 200 mJ/cm 2 UV light. Relative growth was assayed after 24 h of culture at 30 • C.

CaMad2 Localizes to the Nuclear Periphery in C. albicans
To determine the localization of CaMad2, we tagged the C-terminus of CaMad2 with GFP, stained the nuclei with Hoescht DNA stain, and visualized live cells by confocal microscopy. We found localization to several different patterns: one dot colocalized with the nucleus, one dot not colocalized with the nucleus, more than one dot, a complete ring surrounding the nuclear periphery, a partial ring surrounding the nuclear periphery, and a diffuse distribution of GFP in the cell ( Figure 5A). A majority of the cells exhibited a ring surrounding the nuclear periphery ( Figure 5B). The ring morphology was observed in both budding and non-budding cells. The most common alternative localization patterns were the formations of a Mad2-GFP cluster at a single point within the cell (dot) and diffuse localizations throughout the cell ( Figure 5B). We next tested whether the localization pattern of CaMad2-GFP changed during treatment with microtubule-destabilizing or DNA damage-inducing treatments. We observed moderate (nocodazole) and slight (hydrogen peroxide) increases in the percentages of cells with dot localization patterns compared to untreated cells ( Figure 5B).
Genes 2019, 10, x FOR PEER REVIEW 10 of 16 and diffuse localizations throughout the cell ( Figure 5B). We next tested whether the localization pattern of CaMad2-GFP changed during treatment with microtubule-destabilizing or DNA damageinducing treatments. We observed moderate (nocodazole) and slight (hydrogen peroxide) increases in the percentages of cells with dot localization patterns compared to untreated cells ( Figure 5B).

Increased Growth on Fluconazole Is a Shared Feature of Mutants with Altered Chromosome Segregation
Checkpoint Function in C. albicans C. albicans strains with mutations that alter chromosome segregation checkpoint function, such as bub1∆/∆, sgo∆/∆, and h2a.2∆/∆ strains, have increased growth on the antifungal drug fluconazole compared to wild-type cells [17]. We hypothesized that deletion of CaMAD2 would also alter growth on fluconazole. We examined growth by dilutions and spot plating on YPAD media containing 1 µg/mL fluconazole. We observed increased growth relative to the MAD2 cells for both mad2∆/∆ strains and for the bub1∆/∆ strain, as colonies in these strains were visible in spots with higher dilutions than the wild-type strains. The increased growth was more substantial when considering the growth defects seen in these strains without drug treatment ( Figure 6A). When comparing total growth in fluconazole to total growth in no drug media, fluconazole had less of an effect on growth for both mad2∆/∆ strains and for the bub1∆/∆ strain than for the strain with intact spindle assembly checkpoint components ( Figure 6B). The heterozygous MAD2/mad2∆ strains had an intermediate

Increased Growth on Fluconazole Is a Shared Feature of Mutants with Altered Chromosome Segregation
Checkpoint Function in C. albicans C. albicans strains with mutations that alter chromosome segregation checkpoint function, such as bub1∆/∆, sgo∆/∆, and h2a.2∆/∆ strains, have increased growth on the antifungal drug fluconazole compared to wild-type cells [17]. We hypothesized that deletion of CaMAD2 would also alter growth on fluconazole. We examined growth by dilutions and spot plating on YPAD media containing 1 µg/mL fluconazole. We observed increased growth relative to the MAD2 cells for both mad2∆/∆ strains and for the bub1∆/∆ strain, as colonies in these strains were visible in spots with higher dilutions than the wild-type strains. The increased growth was more substantial when considering the growth defects seen in these strains without drug treatment ( Figure 6A). When comparing total growth in fluconazole to total growth in no drug media, fluconazole had less of an effect on growth for both mad2∆/∆ strains and for the bub1∆/∆ strain than for the strain with intact spindle assembly checkpoint components ( Figure 6B). The heterozygous MAD2/mad2∆ strains had an intermediate phenotype with fluconazole consistent with the haplo-insufficient phenotypes observed with other treatments as well.
Treatment with 1 µg/mL fluconazole significantly increased URA3 marker loss rate in the mad2∆H/∆U deletion strain ( Figure 6C). This increase in marker loss seems to be due to an increase in whole chromosome loss, as all strains tested were homozygous in SNP-RFLP assays (Table S2). The bub1∆/∆ strain had similar URA3 marker loss rates to the mad2∆/∆ strain in the absence of fluconazole, and showed an elevated loss rate with fluconazole treatment, although the magnitude of increase was less than for the mad2∆/∆ strain ( Figure 6C). We also measured the minimum inhibitory concentration (MIC) for MAD2, mad2∆/∆, and bub1∆/∆ strains using a broth dilution assay in RPMI. The MIC values were ≤0.125 µg/mL for the MAD2 strains. We observed MIC values of 0.125 and 0.25 µg/mL strains for the mad2∆/∆ and bub1∆/∆ strains. All of these values fall below the cutoff for fluconazole susceptibility of <8 µg/mL [34], but there was a trend towards increased MIC values in the mad2∆/∆ and bub1∆/∆ strains. Optical density (600 nm) was collected every 15 min using a Tecan incubator/plate reader. Results were calculated from the area underneath the curve of growth. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least four biological replicates, each with four cultures per condition. Significant differences are indicated with letters (two-way ANOVA: strain difference p < 0.0001; fluconazole difference p < 0.0001; interaction p < 0.0001; Tukey multiple comparison, p < 0.05 for letter difference). (C). C. albicans strains with/or without the deletion of MAD2 or BUB1 were grown overnight in YPA-glucose media; then, diluted with water and plated onto YPAD for total cell counts and 5-FOA plates to select for URA3 loss. Colony counts were used to calculate the rate of loss per cell division. Data are maximum/minimum box plots with means (indicated by the bars) calculated from at least six biological replicates, each with eight cultures per conditions. Significant differences are indicated with letters (two-way ANOVA: strain difference p < 0.0001; fluconazole difference p < 0.0001; interaction p < 0.0001; Tukey multiple comparison, p < 0.05 for letter difference).

Discussion
In this work, we found that deletion of CaMAD2 decreased cell growth, increased marker loss rates, increased sensitivity to microtubule-destabilizing drugs, and increased sensitivity to DNA damage-inducing treatments. We did not see an increased requirement for CaMad2 when CSE4 was overexpressed resulting in additional kinetochore-microtubule attachments per centromere. Together, these results show that CaMad2 promotes genome stability at the chromosomal level through its canonical spindle assembly checkpoint roles and at the sub-chromosomal level through roles in preventing DNA damage and recombination. Additionally, CaMad2-GFP localized to dots, consistent with a role in kinetochore binding, and to the nuclear periphery, consistent with an additional role in DNA damage. We also found that, similarly to other strains with altered chromosome segregation checkpoint function, deletion of MAD2 increases growth on fluconazole, and that fluconazole treatment elevates whole chromosome loss rates in the mad2∆/∆ strain.
A prior characterization of CaMad2 found that it is important for accurate chromosome segregation, survival in high levels of hydrogen peroxide, and virulence in a mouse model of infection [19]. We found that deletion of MAD2 in C. albicans results in a slow growth phenotype. This is similar to previous observations that mad2∆ strains of the Fusarium head blight fungus Fusarium graminearum and the multinucleate fungus Ashbya gossypii had reduced growth and viability, likely due to high levels of aneuploidy [35,36]. Interestingly, haploinsufficiency was also observed in our experiments with C. albicans and in A. gossypii [35]. MAD2 haploinsufficiency also causes chromosome instability in human cancer cells and mouse embryonic fibroblasts [37], while overexpression of the MAD2 gene is also associated with tumor formation and poor cancer prognosis [38,39]. The phenotypes observed with both decreased and increased levels of MAD2 suggest that MAD2's function relies on a particular level of expression in the cell. This may be related to the importance of the protein's stoichiometry. Mad2 protein is found in both open and closed conformations, which dimerize to each other. This dimerization helps convert the open form into the closed form that then binds to Cdc20 to form the MCC [40]. Thus, up or down regulation of the protein may affect the stoichiometry and the amount of bound closed Mad2, which could disrupt the spindle assembly checkpoint.
Previously, Bai et al. showed elevated sorbose resistance rates and homozygosis of the mating type locus in strains lacking CaMAD2 suggesting that CaMad2 has roles in chromosome segregation accuracy in C. albicans [19]. The sorbose resistance assay measures monosomy of chromosome 5 due to the presence of multiple sorbose utilization genes on chromosome 5. We used selection of 5-FOA resistant colonies to measure loss of the URA3 marker gene. Since this assay requires loss of only a single gene, it can detect multiple types of loss of heterozygosity events. Interestingly, in strains lacking CaMAD2, we saw elevated URA3 marker loss rates. These elevated loss rates in rich media were primarily due to smaller-scale events, such as mutation and recombination, as all colonies were still heterozygous. Overall, these results show that deletion of CaMAD2 causes both chromosome loss phenotypes and elevated small-scale DNA changes.
Our treatment sensitivity assays also support a model where CaMad2 has roles in chromosome segregation and preventing DNA damage. We confirmed the sensitivity of strains lacking CaMad2 to 50 µM nocodazole, a microtubule destabilizing drug, in our experiments. Bai et al. showed reduced viability of strains lacking CaMad2 following short-term exposure to high concentrations (147, 88, and 33 mM) of hydrogen peroxide [19]. We tested longer-term exposures (24 h) to a lower concentration (1 mM) and also found sensitivity of strains lacking CaMad2 to this lower concentration, as demonstrated by reduced growth. The hydrogen peroxide results, together with the ultraviolet light sensitivity and increased URA3 marker loss due to recombination or other smaller scale events observed with the Camad2∆/∆ strain, demonstrate that CaMad2 also has a role in preventing DNA damage. Throughout our work, we saw similar phenotypes with Camad2 deletion and Cabub1 deletion strains, and in a few experiments, even saw stronger phenotypes with the mad2 deletion strain. Prior work in budding yeast and fission yeast found that bub1 mutants had more severe chromosome segregation defects than deletions of mitotic checkpoint components, such as Mad1, Mad2, and Mad3 [41]. There are several possibilities to explain why the phenotypes were more similar in C. albicans, including the presence of an H2A form that cannot be phosphorylated by Bub1 [17] and/or the important role of CaMad2 in DNA damage responses.
Additionally, prior work has shown that the nuclear periphery localization pattern of Mad2 is associated with the response to DNA damage [6]. The nuclear periphery localization pattern we observed with CaMad2 is consistent with the hypersensitivity to DNA damage and the high proportion of sub-chromosomal marker loss events. However, CaMad2 localization to the nuclear periphery appears to be the default location rather than a specific response to DNA damage, as the treatment with hydrogen peroxide did not increase the proportion of cells with a complete ring of CaMad2-GFP at the nuclear periphery. These results are also consistent with the result in Saccharomyces cerevisiae that Mad1 and Mad2 only bind to kinetochores when there is a problem with the spindle or kinetochore [42]. Together, these results indicate that the DNA damage response through Mad2 is conserved in C. albicans.
There are a few different proposed mechanisms in S. cerevisiae for the role of Mad2 in preventing DNA damage. Mad2 prolongs cell cycle arrest in cells with unrepaired double-strand breaks, suggesting that it may be important for allowing cells time to repair the DNA damage [43]. In S. cerevisiae, Mad2 also potentially has roles in nucleotide excision repair pathways, as mad2∆ shows synthetic lethality with the DNA damage response genes RAD10 and RAD23, which are subunits of nucleotide excision repair factor 1 and nucleotide excision repair factor 2, respectively [44]. Additionally, recent evidence shows that Mad2 promotes translation of the S-phase cyclins Clb5/6 and contributes to cell survival in response to replication stress when Rad53 is not functional [7]. In the future, it will be interesting to further characterize the mechanisms by which CaMad2 contributes to DNA damage responses in C. albicans.
It will also be critical to further explore the interactions between spindle assembly checkpoints and the development of drug resistance. Deletion of MAD2 increases sensitivity to antifungals that target microtubules, such as thiabendazole in C. albicans and F. graminearum [36], but deletion of CaMAD2 increases growth on the antifungal fluconazole. This observation is similar to other strains with altered chromosome segregation checkpoint function, including bub1∆/∆, sgo∆/∆, and h2a.2∆/∆ strains [17], suggesting that disruption of any of these genes may promote the evolution of drug resistance in C. albicans. Cells lacking CaMAD2 had slightly elevated MIC values relative to cells without disruption of CaMAD2, but both strains were still below the MIC cutoff for resistance. It would be interesting to test whether cells lacking CaMAD2 evolve fluconazole resistance differently than cells with functional CaMAD2 following repeated exposure to fluconazole.
Fluconazole resistance is associated with elevated levels of aneuploidy. In one set of clinical and laboratory strains, approximately 50% of fluconazole resistant strains contained a whole chromosome or segmental aneuploidy, while 10% of fluconazole sensitive strains were aneuploid [45]. We found that the elevated marker loss rate seen in rich media conditions with the mad2∆/∆ strain increases further upon exposure to treatment. Importantly, the type of marker loss events also changed from sub-chromosomal events to whole chromosome homozygosis, indicative of whole chromosome aneuploidy. Exposure to fluconazole induces aneuploidy through the separation of the nuclear membrane segregation and cytokinesis processes [46]. Our results show that the absence of CaMAD2 may potentiate fluconazole-induced aneuploidy, which leads to better growth on low levels of fluconazole for mad2∆/∆ strains.

Conclusions
In this work, we found that CaMad2 promotes genome stability through its direct spindle assembly checkpoint role in promoting accurate chromosome segregation and through reducing smaller scale chromosome changes due to recombination and DNA damage repair. Furthermore, the frequent nuclear periphery localization pattern of CaMad2-GFP is consistent with a role in DNA damage response as similar patterns have been seen in other organisms. Together, our results demonstrate that a Mad2-dependent DNA damage response is conserved in C. albicans. We also found that deletion of MAD2 increases growth on fluconazole similar to other strains with altered chromosome segregation checkpoint function. These results indicate that disruption of spindle assembly checkpoint genes, including MAD2, may promote the evolution of azole drug resistance in C. albicans.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4425/10/12/1013/s1. Figure S1: Deletion of MAD2 decreases growth rate, Figure S2: Growth of strains with and without MAD2 are not affected by increased kinetochore-microtubule attachments per centromere, Figure S3: Loss of MAD2 increases sensitivity to the microtubule-destabilizing compound thiabendazole, Table S1: C. albicans strains used in this study, Table S2: Number of heterozygous and homozygous URA3 marker loss events measured by SNP-RFLP, File S1: Growth Curve Macros Excel Sheet Template.