Kinesin-5 Is Dispensable for Bipolar Spindle Formation and Elongation in Candida albicans, but Simultaneous Loss of Kinesin-14 Activity Is Lethal

Candida albicans is one of the most prevalent fungal pathogens of humans and can infect a broad range of niches within its host. This organism frequently acquires resistance to antifungal agents through rapid generation of genetic diversity, with aneuploidy serving as a particularly important adaptive mechanism. This paper describes an investigation of the sole kinesin-5 in C. albicans, which is a major regulator of chromosome segregation. Contrary to other eukaryotes studied thus far, C. albicans does not require kinesin-5 function for bipolar spindle assembly or spindle elongation. Rather, this motor protein associates with the spindle throughout mitosis to maintain spindle integrity. Furthermore, kinesin-5 loss is synthetically lethal with loss of kinesin-14—canonically an opposing force producer to kinesin-5 in spindle assembly and anaphase. These results suggest a significant evolutionary rewiring of microtubule motor functions in the C. albicans mitotic spindle, which may have implications in the genetic instability of this pathogen.

T he mitotic spindle is a highly dynamic microtubule (MT)-based structure that undergoes a distinct set of morphological changes in order to correctly attach, orient, and then separate sister chromatids in the dividing cell. Kinesin motor proteins play major roles in shaping and organizing MTs within the spindle over the course of cell division. Early in mitosis, evolutionarily conserved kinesin-5 proteins cross-link the overlapping plus ends of interpolar MTs from newly duplicated centrosomes (spindle pole bodies in yeast) and then slide them apart via plus-enddirected motility to establish spindle bipolarity (1)(2)(3)(4)(5)(6)(7)(8). Genetic or chemical inhibition of kinesin-5 activity produces monopolar spindles or inward collapse of preanaphase spindles, usually leading to cell death (9)(10)(11). This spindle defect arises from loss of outward forces needed to counterbalance the inward forces supplied by MT minus-end-directed kinesin-14 motors, which pull spindle poles together (10,12,13). In many organisms, a nearly normal spindle phenotype can be restored by inactivating or depleting cells of kinesin-5 and kinesin-14 simultaneously because this force imbalance is eliminated (10,(13)(14)(15)(16). In this experimental scenario, pushing forces generated by MT growth are sufficient to promote spindle pole separation and bipolar spindle assembly (17)(18)(19)(20).
This interplay of motor and MT forces has been studied extensively in the model yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe (10,17,18,(20)(21)(22). S. cerevisiae encodes two kinesin-5 homologs, Kip1 and Cin8, that have overlapping, but nonequivalent functions during mitosis (1,2), while S. pombe encodes a single kinesin-5, named Cut7 (23). All three of these proteins form homotetramers that exhibit bidirectional motility, and all of them function in bipolar spindle assembly and crosslink parallel MTs to help focus kinetochore clusters (24)(25)(26)(27)(28)(29). They are also important for stabilizing the overlapping array of MTs at the anaphase spindle midzone and for promoting and regulating timely anaphase spindle elongation (17,22,(30)(31)(32)(33)(34). In both yeast species, loss or inhibition of kinesin-5 function is lethal. However, simultaneous inactivation of their kinesin-14 motors (Kar3Cik1 and Kar3Vik1 in S. cerevisiae or Pkl1 and Klp2 in S. pombe) neutralizes kinesin-5 deficiency (35)(36)(37)(38), highlighting the importance of keeping inward and outward forces acting on the spindle in balance. In contrast to the lethality of kinesin-5 loss, bipolar spindles are able to form in the absence of kinesin-14 activity, but are either short and disorganized, or their MT minus ends are unfocused and extend past the opposite spindle pole (39,40). Our studies of the homologous motors in the opportunistic fungus Candida albicans indicate that these phenotypes, and the opposing relationship of kinesin-5 and kinesin-14 proteins in spindle regulation, are not as highly conserved among eukaryotes as previously thought.
C. albicans is a close relative of S. cerevisiae and S. pombe, but it encodes only one kinesin-5 and one kinesin-14 motor, named CaKip1 and CaKar3, respectively, the latter of which forms a heterodimer with a noncatalytic kinesin-like protein, CaCik1 (41). C. albicans is viable without CaKip1 (42), and cells lacking CaKar3Cik1 activity often arrest with a monopolar spindle or two dissociated half-spindles (41). Through further investigation of these unconventional phenotypes, we found that CaKip1 is not needed for bipolar spindle assembly or nuclear division, even though it exhibits the same cellcycle-dependent localization as its homologs in budding yeast. However, kip1Δ/Δ spindles are shorter and intermittently disassemble prior to cell division. When spindle disassembly occurs, two or more independent bipolar spindles emerge that either segregate between the mother and daughter cells or elongate across the bud neck. Each bring portions of the nucleus with them, which are further subdivided when the spindles undergo anaphase. Rather than neutralizing these kinesin-5 deficiencies, we found that simultaneous loss of kinesin-14 activity is lethal. These results imply that C. albicans Kip1 and Kar3Cik1 have mostly overlapping rather than antagonistic functions in bipolar spindle assembly and that their combined loss cannot be compensated for by MT polymerization forces or other spindle-associated factors.
( Fig. 2B and C). Upon further visual inspection and quantification of the kip1Δ/Δ strain by microscopy, we observed this hyperpolarized morphology in approximately 30% of the cells (Fig. 2C). These elongated cells indicate a delay in cell cycle progression or a cell cycle arrest and could mask a slower proliferation rate on solid growth medium by giving kip1Δ/Δ dilution spots a similar appearance to the wild type. We also found that loss of CaKip1 affected filamentous growth under hypha-inducing conditions. CaKip1depleted colonies formed a smaller halo of invasive growth on Spider medium (Fig. 2D), and cells grown in serum produced shorter germ tubes and fewer septa ( Fig. 2E and F). When we added a wild-type copy of CaKIP1 back into the kip1Δ/Δ strain at the native locus, normal cell growth rate and cell morphology were restored (Fig. 2B to D), confirming that these defects were a direct consequence of CaKIP1 loss.
Expecting that the slow-growth phenotype of CaKip1-depleted cells was caused by errors in mitotic spindle assembly, we imaged fields of unsynchronized wild-type and kip1Δ/Δ blastoconidia expressing Tub2-mCherry and Spc98-GFP (a component of the spindle pole body) and examined their spindle structures. We found that most of the spindles in budded kip1Δ/Δ cells (ϳ92%) formed a stereotypical bipolar spindle structure (Fig. 3A). However, nearly twice as many kip1Δ/Δ cells had metaphase spindles (53.5%) compared to the wild type (28.9%) (Fig. 3B), and the mean length of kip1⌬/⌬ metaphase spindles was significantly shorter (kip1⌬/⌬, 0.68 Ϯ 0.01 m; wild type, 0.93 Ϯ 0.01 m) (Fig. 3C). When we tracked progression of the mitotic phases by time-lapse microscopy, we observed that kip1⌬/⌬ cells took an average of 117.1 Ϯ 7.8 min to initiate anaphase after a spindle had formed, whereas wild-type cells required only 80.0 Ϯ 5.2 min on average ( Fig. 3D and E). In contrast, when kip1⌬/⌬ spindles did eventually elongate, there was no difference in the duration of anaphase compared to the wild type (Fig. 3F). These data show that C. albicans is not solely dependent on kinesin-5 activity for bipolar spindle assembly or late anaphase spindle  tetO-KIP1/Δ, strain CF436. Cells were serially diluted to the specified concentrations, and 5-l droplets were plated on YPD with or without DOX (10 g/ml). Plates were incubated for 2 days at the indicated temperatures. (B) Cell growth assay of the independent CaKIP1 strains, including a KIP1 add-back strain (CF354). Cells in SDC medium were diluted to 2.5 ϫ 10 6 cells per ml, incubated at 30°C, and counted every hour with a hemocytometer. Data points represent the average from three independent experiments Ϯ standard error of the mean (SEM). (C) Cells were grown in SDC, and bright-field images were collected. The graph shows the proportion of normal-looking blastoconidia and arrested cells observed in these bright-field images. Data represent the average from three independent experiments Ϯ SEM. n Ͼ 3,000 cells per strain. (D) Assessment of hyphal growth on various CaKIP1-null strains. The wild type and each KIP1 mutant were plated onto Spider medium and incubated for 5 days at 30°C before imaging. (E) Bright-field (BF) and fluorescence images of wild-type and kip1Δ/Δ cell cultures. Cells were diluted 1:50 into fresh YPD medium supplemented with 10% FBS and incubated at 37°C to induce hyphae. Cells were removed from the cultures at the indicated time points and then fixed and stained with DAPI (4=,6-diamidino-2-phenylindole) before imaging. (F) Cells were induced to form hyphae under the same conditions in panel E and then fixed and stained with calcofluor white before imaging. The number of septa per hyphae was quantified and graphed. elongation, but CaKip1 is important for timely separation of spindle poles after spindle assembly.
Interestingly, kip1⌬/⌬ spindles had longer and more numerous astral MTs than the wild type ( Fig. 4A to C). In many eukaryotes, including C. albicans, the plus ends of astral MTs strike the cell cortex, where they are captured by the minus-end-directed MT motor protein dynein (54)(55)(56). When this happens, dynein can draw the MT, and the attached SPB, toward the cortical contact site to facilitate proper spindle positioning, elongation, and/or migration (57,58). In S. cerevisiae, this activity of dynein assists Cin8 and Kip1 in anaphase spindle elongation, and simultaneous loss of dynein and Cin8 activity is lethal (22,31). We found the dyn1⌬/⌬ strain to be nonviable in the presence of the CaKip1-specific inhibitor aminobenzothiazole (ABT), suggesting that CaKip1 and dynein also have overlapping functions in C. albicans (Fig. 4D). Perhaps the longer and more numerous astral MTs in the kip1⌬/⌬ strain are an adaptation to CaKip1 loss that provides more opportunities for MT capture and pulling events by dynein, which could promote both anaphase spindle elongation and SPB separation during spindle assembly.
A subpopulation of kip1⌬/⌬ cells have multiple spindles and show atypical cell cycle dynamics. Similar to previous findings by Chua et al. (42), we observed that a significant proportion (ϳ12%) of kip1Δ/Δ blastoconidia and all kip1Δ/Δ cells with a hyperelongated morphology, contained multiple spindles ( Fig. 5A and B). In some cases, monopolar and bipolar spindles were simultaneously visible within the same budding cell (Fig. 5A, row 3). To determine how these extra spindles formed, we collected time-lapse images of kip1Δ/Δ blastoconidia expressing Tub2-GFP. All of the multispindle blastoconidia that we tracked (n ϭ 20) formed two short "bars" of tubulin fluorescence in the mother cell as the new bud began to emerge. Once a new bud formed, we observed two different multispindle configurations. In 65% of the cells we imaged, one of the tubulin structures traversed the bud neck, while the other remained in the mother compartment ( Fig. 5C, row 2, t ϭ 6 min). Each short fluorescent bar then elongated simultaneously, suggesting that they had formed distinct bipolar spindles. However, both spindles broke apart or disintegrated once the cell divided ( Fig. 5C, row 3, t ϭ 36 min; see Movie S1 in the supplemental material). This phenotype suggests that CaKip1 activity is important for anaphase spindle stability. In a smaller cohort of cells (45%), one or both of the spindles elongated across the bud neck and appeared to complete anaphase (not shown). Both of these multispindle phenotypes were recapitulated in wild-type cells treated with ABT ( Fig. 5D). We were intrigued by this spindle defect because a subpopulation of wild-type C. albicans cells exposed to the antifungal agent fluconazole (FLC) display abnormal numbers of spindles as well. In the presence of FLC, DNA replication and nuclear division proceed ahead of bud emergence and completion of cytokinesis, respectively. Harrison et al. (59) showed that when this happens, some nuclei re-fuse or fail to separate due to mitotic collapse, forming tetraploid progeny with extra spindle components. Therefore, we next used time-lapse microscopy to track nucleolar segregation (using Nop1-mScarlet) in ABT-treated wild-type cells that formed multiple spindles. While the "no-drug" condition showed stereotypical nuclear and spindle dynamics that were well coordinated with bud emergence and growth ( . This indicates that a bipolar spindle had already formed and broken apart before bud evagination. As bud growth continued, the Nop1 patch divided and migrated with each spindle. When these spindles were segregated to the mother and daughter cell (cell 1, 65% of the cells  imaged), two smaller nuclear fragments were visible in each compartment after spindle elongation (four Nop1-mScarlet patches in total, cell 1, t ϭ 95 min). However, within 20 min, each pair of Nop1 patches coalesced as a result of mitotic collapse. When the spindles elongated across the neck (cell 2, 45% of the cells imaged), each spindle divided a Nop1-mScarlet patch into two different pieces, again resulting in four separate nuclear fragments (cell 2, t ϭ 45 min). Here, the two Nop1 foci in each compartment were produced from different anaphase spindles. As the time-lapse continued, these Nop1 foci appeared to merge into one (Fig. 6, cell 2, t ϭ 75 min). These results suggest that the extra spindles seen in kip1Δ/Δ cells and ABT-treated wild-type cells could have formed in nuclei that experienced mitotic collapse or after merging of nuclear fragments from two distinct spindles that completed anaphase. Simultaneous loss of CaKip1 and CaKar3/Cik1 function is lethal. In many of the eukaryotic systems, inactivation of kinesin-14 rescues the lethal spindle defects arising from inhibition or loss of kinesin-5 activity and spindles recover the ability to complete a relatively normal mitotic cycle (12-16, 37, 38, 60-65). This has been rationalized as a restoration of force balance in the spindle, where compensatory spindle forces are provided by MT polymerization and cross-linking proteins (17,34,66). In spite of repeated attempts, we were unable to obtain a kip1Δ/Δ kar3Δ/Δ strain by traditional methods (data not shown), suggesting they are synthetically lethal. To confirm this, we constructed a KIP1 knockout strain containing only one functional copy of KAR3 that is under the control of the maltose-inducible Mal2 promoter. Indeed, when we deactivated the Mal2 promoter by culturing this strain on glucose (YPD [yeast extractpeptone-dextrose] medium), cell growth was arrested, demonstrating that simultaneous loss of CaKar3 and CaKip1 is lethal (Fig. 7A). We also observed that kar3⌬/⌬ and cik1Δ/Δ cells were not viable in the presence of ABT (Fig. 7B). To visualize events leading up cell death by loss of CaKip1 and CaKar3 function, we imaged Tub2-GFP fluorescence in kar3Δ/Δ cells treated with 100 M ABT every 15 min by time-lapse microscopy to avoid photobleaching. After 165 min of imaging, we observed a short anaphase spindle that did not elongate further. Within 3 h, spindle structures disappeared and cells showed no tubulin fluorescence (Fig. 7C). These results demonstrate that kinesin-5-and kinesin-14 have more functional overlap in C. albicans than in other organisms.

DISCUSSION
In nearly every type of eukaryotic system studied, kinesin-5 activity is needed to push newly duplicated centrosomes or SPBs apart to establish spindle bipolarity (  the major providers for outward forces during anaphase spindle elongation (1-4, 7, 8, 23, 67). Without them, most cells exhibit mono-astral spindles and are nonviable (2,(9)(10)(11). In contrast, our genetic data demonstrate that the sole kinesin-5 gene in C. albicans is not essential in diploid cells. Perhaps the recent finding that CaKIP1 is an essential gene in a haploid isolate of C. albicans (51) is an indication that CaKIP1 mutants exhibit a form of ploidy-specific lethality, which is shared by other yeast genes involved in mitotic spindle stability (68). An alternative explanation for this discrepancy is that there are differences in protein expression between the haploid and diploid proteomes (69). Our results further show that kinesin-5 is dispensable for spindle assembly and anaphase spindle elongation in C. albicans. The only other organisms reported to complete mitosis without kinesin-5 activity are Caenorhabditis elegans and Dictyostelium discoideum (70)(71)(72). Although their mechanisms for kinesin-5-independent bipolar spindle assembly and elongation are not yet known, it has been suggested that cytoplasmic dynein-mediated astral MT pulling forces are involved. We propose that dynein may also fulfill these roles in C. albicans in the absence of kinesin-5 activity for several reasons. In earlier studies, C. albicans cells lacking the heavy chain of cytoplasmic dynein or the p150 Glued subunit of dynactin exhibited spindle position, orientation, and elongation defects, and dramatically slowed nuclear dynamics (73,74). In the filament-forming fungus Ustilago maydis, it was shown that stationary dynein motors capture and pull on the plus ends of astral MTs that emanate from SPBs, drawing the attached SPB toward the cortical contact site (75). In our studies, astral MTs were much longer and more numerous in kip1⌬/⌬ cells, which could increase the frequency of these MT capture and dynein-mediated pulling events for SPB separation and spindle elongation. We also found dyn1⌬/⌬ mutants to be nonviable in the presence of the inhibitor ABT, suggesting that CaKip1 and dynein have overlapping functions. Our future studies aim to better understand this putative intersection of dynein and CaKip1 functions. We will also investigate the alternative possibility that CaKip1 has a direct role in limiting the number and length of astral MTs, based on recent evidence that kinesin-5s can act as length-dependent MT depolymerases at kinetochores (46,76,77).
Although kip1Δ/Δ cells readily assembled metaphase spindles, these spindles were shorter and delayed in transitioning to anaphase relative to wild-type cells. This suggests a role for CaKip1 in maintenance of the bipolar spindle prior to anaphase, which is an important kinesin-5 function in other fungi and in Xenopus and Drosophila (3,4,45,78). The redistribution of CaKip1-GFP fluorescence along the spindle in early anaphase supports such a role. The metaphase-to-anaphase delay in kip1Δ/Δ cells may also explain their lower growth rate in liquid cultures. Our observation that kip1Δ/Δ spindles sometimes broke apart and then reassembled two new bipolar spindles also supports this function and implies that CaKip1 is acting as a MT cross-linker within the spindle. An alternative explanation for these short, unstable spindles is that CaKip1 regulates kMT dynamics, which is important for congression of bioriented sister chromosomes in metaphase. Indeed, S. cerevisiae's Cin8 is important for kinetochore clustering/positioning near the SPBs by cross-linking kMTs and promoting the disassembly of long kMTs (28,46). In S. pombe, Cut7 is recruited to the kinetochores by a spindle assembly checkpoint (SAC) protein, Mad1, to promote chromosome gliding toward the spindle equator (29). Recent EM reconstructions of C. albicans KIP1/kip1⌬ spindles show disorganized kMTs (76), suggesting that chromosomes are not properly congressed at the spindle equator during metaphase. Without proper chromosome congression, mitotic errors are more likely to occur. Perhaps the short bipolar spindles we observed in kip1Δ/Δ cells are indicative of attempts to correct erroneous kMT attachments (76), and spindle disassembly occurs when they are not corrected. If these cells initiate DNA replication and attempt mitosis again, this could explain the extra spindles and nucleoli observed in a subpopulation of C. albicans kip1Δ/Δ cells. Further work will be needed to uncover whether or not these defects in nuclear dynamics lead to an increase in the prevalence of aneuploid cells.
Surprisingly, CaKip1 loss did not extend the duration of anaphase relative to wild type, even though CaKip1-GFP accumulated at the midzone of late anaphase spindles; a site where it could exert outward MT sliding forces for spindle elongation. This is unique from other fungi and Drosophila embryos, which rely on kinesin-5 to cross-link overlapping antiparallel MTs in the spindle midzone and drive anaphase B spindle elongation via plus-end-directed motility (6,17,22,(30)(31)(32)(33)(34)63). While we suspect that dynein provides pulling forces on the spindle to assist in anaphase spindle elongation in the absence of CaKip1, it is also possible that other kinesins or MT cross-linking components within the spindle are involved. In fission yeast, kinesin-6 provides additional MT-sliding forces to kinesin-5 at the spindle midzone for anaphase spindle elongation and dynein is not involved (17,66). Although C. albicans has no kinesin-6 homolog in its genome, it encodes five other kinesin-like proteins in addition to CaKip1. Therefore, we have begun to generate strains lacking different combinations of these proteins in order to identify new collaborative roles of kinesins in mitosis. By simultaneously disrupting kinesin-5 and kinesin-14 activities, we found that C. albicans displays a puzzling exception to the widely regarded spindle force-balance model (15,37,62,79). Rather than providing antagonistic spindle forces, CaKip1 and CaKar3Cik1 may cooperate to focus and stabilize parallel and antiparallel interactions in certain areas of the spindle. In this regard, loss of both kinesins may reduce the number of MT cross-linking factors to an intolerable level that cannot support cell viability. Our previous finding that CaKar3Cik1-depleted cells often arrest with two monopolar half-spindles that become pulled apart before assembling a bipolar spindle, supports this idea (41). Combined with CaKip1 loss, MTs may not be well tethered at SPBs or fail to focus kinetochores, resulting in disorganized spindle structures that quickly break down.
C. albicans is a close relative of the model yeasts S. pombe and S. cerevisiae but is also an opportunistic fungal pathogen. An assortment of fitness attributes promote its pathogenicity (80), most of which arise by rapid genetic diversification within a population in response to stressful growth conditions as a means of adaptation (81,82). Research has shown that aneuploidy accounts for much of this diversity (83)(84)(85), and recent findings suggest that aneuploidies could be induced or enabled by altered activity of mitotic kinesin motors under stress (39,49,(86)(87)(88). As specific aneuploidies can confer resistance to antifungal drugs through altered gene copy numbers, it could be advantageous for C. albicans cells to regulate mitotic kinesins as a way to control aneuploidy occurrence. We are currently conducting studies to delineate the putative contributions of C. albicans kinesins to mitotic defects of cells under stress and to identify stress-specific regulatory factors that change kinesin activity to promote aneuploidy. Table 1. The oligonucleotides used in strain construction are listed in Table 2. Gene disruption of the C. albicans KIP1 open reading frame (Candida Genome Database, orf19.8331; NCBI Gene ID, 3645256) was conducted by PCR-based gene targeting and CRISPR-Cas9 methods (89). PCR amplification was used to generate disruption cassettes where a selectable marker was flanked by approximately 50 bp of C. albicans genomic sequence immediately 5= and 3= of the KIP1 coding region. Disruption of KIP1 in a wild-type strain (CF027) was conducted sequentially. First a kip1::LEU2 ϩ cassette was amplified from pSN40 (90) using primers P118 and P119 and transformed into strain CF027. Correct kip1::LEU2 ϩ cassette integration was confirmed using primer pairs P120/P13 and P121/P14 for the upstream and downstream junctions, respectively. To disrupt the second KIP1 allele, a kip1::HIS1 ϩ cassette was amplified from pSN52 (90) using primer pair P118/P119 and transformed to create strain CF311. Integration of the disruption cassette at the correct location was confirmed by PCR amplification across the junctions of integration using primers P120/P11 and P121/P12 for the upstream and downstream regions, respectively. CRISPR-Cas9-mediated kip1 deletion was conducted as previously described (89) using the custom guide RNA (gRNA) primer P247 and double-stranded donor DNA formed using primers P248/P249 to create the strain CF429 (89). To regulate the expression of KIP1, the tetracycline-repressible transactivator, the tetO promoter, and the NAT flipper cassette were PCR amplified from pLC605 (kindly provided by Leah Cowen) using primers P240/P241. The PCR-amplified product was transformed into the heterozygous KIP1 strain to create strain CF436. Correct integration at the KIP1 locus was verified using primer pairs P120 and P242.

Genetic manipulations. A list of C. albicans strains used in this study is presented in
To demonstrate that mutant phenotypes are solely a result of loss of KIP1, add-back strains were created to reintroduce a wild-type copy of each gene. The KIP1 gene (Ϯ1,000 bp upstream/downstream) was cloned into pCIp10-based integration plasmids bearing the ARG4 ϩ selectable marker using primers P128/P129 (41,91). The integration plasmid was digested at a unique restriction site (PmlI) to add back to the endogenous KIP1 region into CF311 to create CF354. Confirmation of integration of the pCIp10-ARG4 ϩ vector was done using P121/P170.
A strain lacking both CaKip1 and CaKar3Cik1 function was created by deleting both copies of KIP1 and one copy of KAR3 and by placing the remaining functional KAR3 copy under the control of a maltose promoter as follows: one copy of the KAR3 ORF was disrupted using kar3::HIS1 ϩ knockout cassette amplified using primers P199/P200, transformed into CF027, and confirmed using primers P201/P11 and P202/P12 for the upstream and downstream junctions, respectively. The ARG::MAL::KAR3 cassette was amplified using primer pair P212/P213 from pFA-ARG4-MALp (92) and transformed in kar3Δ::HIS1 ϩ to create CF411 (not shown). Integration of the cassette was confirmed using primers P201 and P16. KIP1 was disrupted using LEU2 (described above) and the SAT1 nourseothricin resistance marker to create the strain CF396. Correct kip1::SAT1 integration was confirmed with primer pairs P128/P17 and P129/P18. Fluorescent tagging of KIP1 ϩ in wild-type cells was accomplished using the method described by Gerami-Nejad et al. (93) and using long-tailed primers P137 and P187 and the plasmid pGFP-SAT1 as a template to create an integration cassette bearing approximately 50 bp of KIP1 ϩ ORF immediately before the stop codon and of sequence 3= to the ORF. This cassette was transformed into the wild type (CF027) to create the KIP1-GFP-SAT1 R strain. Correct integration was confirmed by PCR using the primer pair P69 and P169. The same integration cassette was also transformed in cik1Δ/Δ to create strain CF308 (not shown). KIP1-mScarlet was amplified using pScarlet plasmid pRB897 (kindly provided by Richard Bennett) using primers P284 and P285 and transformed into CF421 (pGAL1-Tub2-mNeon) to create CF443. Integration was confirmed using primers P254 and P121.
Strains expressing fluorescently labeled ␤-tubulin were constructed using the plasmids pGAL1-Tub2-GFP-SatR::NEUT5L, pGAL1-Tub2-mCherry-Arg4::NEUT5L, or pGAL1-Tub2-mNeon-Arg4::NEUT5L using the previously described method (41) and adapted to further include sequence of the neutral NEUT5L locus, which was linearized using the restriction enzyme KpnI. pGAL1-Tub2-GFP was transformed into CF027 and CF311 to create CF289 and CF226, respectively. pGAL1-Tub2-mCherry was transformed into the wild-type, kip1Δ/Δ, and cik1Δ/Δ strains to create CF363, CF368, and CF340, respectively. pGAL1-Tub2-mCherry was transformed into KIP1-GFP to create strain CF338. Correct integration for pGAL1 vectors was confirmed by PCR using primers P16 (or P17, depending on SAT/ARG markers) and P107. Induction of the GAL1 promoter, leading to expression of fluorescently tagged tubulin, was done by growing the cells in SDC-sucrose medium supplemented with 1% galactose. To visualize the nucleus, the nucleolar protein Nop1 was fluorescently labeled using pScarlet. The integration cassette was PCR amplified using primers P243 and P244 and transformed into CF289 to create CF417. Correct integration was confirmed by PCR using primers P254 and P246. To visualize spindle pole body structures, strains expressing SPC98-GFP were constructed as previously described (41).  C. albicans transformation. Disruption cassettes, fluorescent tags, and complementation plasmids were transformed into C. albicans using the lithium acetate-polyethylene glycol (PEG) heat shock method as previously described with minor modifications (94). Incubation of cells with transforming DNA in lithium acetate-PEG solution was carried out for 2 h at 30°C with rotation. Heat shock was conducted at 43°C for 30 min. Transformations involving selection using the SAT1 gene were accompanied by a 4-h incubation in YPD (1% yeast extract, 2% peptone and 2% glucose) at 30°C to allow expression of the ClonNAT resistance gene before plating on selection medium.
C. albicans cell culture and growth assays. Strains were maintained on YPD plates. YPD was supplemented with 200 g/ml nourseothricin (clonNat; Werner BioAgents) for selection of positive SAT1 gene integration. Selection for auxotrophic markers was conducted using synthetic dropout (SD) medium containing 0.66% yeast nitrogen base, 0.2% yeast dropout mix lacking uracil, arginine, leucine, and histidine, 2% glucose, and 200 mg/liter uridine and supplemented with 200 mg/liter histidine, leucine, and/or arginine where required. Experimental cultures were grown to mid-logarithmic phase in completely supplemented dropout medium (SDC) unless otherwise indicated. In order to assess the generation time, logarithmically growing cells were diluted to 2.5 ϫ 10 6 cells/ml in fresh medium, and the density was measured hourly using a hemocytometer. To create dilutions for spot assays, logarithmically growing cells were diluted to 1.0 ϫ 10 6 cells/ml in phosphate-buffered saline (PBS). Serial dilutions of 10 5 , 10 4 , and 10 3 cells/ml were made. Five microliters of cell culture dilutions was pipetted for each spot, and plates were incubated at 30°C for 2 days, unless otherwise indicated. To assess hyphal growth, cells were either plated onto Spider medium and incubated for 5 days at 30°C before imaging or were diluted 1:50 into fresh YPD medium supplemented with 10% fetal bovine serum (FBS) and incubated at 37°C to induce hyphae.
Light microscopy. Microscopy for static images was conducted using a Zeiss Axio Observer epifluorescence microscope with a 100ϫ (1.40 NA) oil objective AxioCam hRM camera controlled by Axiovision software. Time-lapse imaging was conducted using the Olympus IX83 with a 100ϫ oil objective (1.4 NA), Andor Zyla 4.2 Plus camera controlled by the cellSens software. For time-lapse and static imaging, logarithmically growing cells were immobilized between an agarose pad and a glass coverslip, as previously described (41). For time-lapses, images were captured in five z-slices 0.8 m apart. Image stacks and pole-to-pole distances were analyzed with ImageJ (NIH). Graphs were calculated and displayed using GraphPad Software; figures were compiled in Adobe Photoshop and Adobe Illustrator.
RNA sequencing. Logarithmically growing yeast cells were harvested and centrifuged at 4,000 ϫ g for 10 min at 4°C. Cell pellets were flash frozen in liquid nitrogen, and genomic DNA-free total RNA was extracted from each pellet by grinding the fungal mass to a fine powder and resuspending it in 1 ml TRIzol (Ambion) solution and using the RNeasy mini-spin columns (Qiagen) following the manufacturer's protocol. RNA quantification was carried out spectrophotometrically at 260 nm and 280 nm, and RNA integrity was evaluated by NanoDrop2000 (Thermo Scientific). Total RNA (1 g/sample) was shipped to the National Research Council of Canada, DNA Sequencing Technologies Facility (Saskatoon, Canada), where further quality check was performed using a BioAnalyzer followed by short cDNA fragment synthesis using the TruSeq Stranded RNALT kit, and finally sequenced on an Illumina HiSeq 2500 platform according to the manufacturer's guidelines (Illumina, USA). The DESeq2-based SARTools (v1.5.1) pipeline as previously described (95) was adopted for differential analysis of mapped C. albicans Assembly 22 RNA-seq count data. A BH P value adjustment was performed (96,97), and the falsediscovery rate was set at P Ͻ 0.05. Data availability. The RNA-seq data that support the findings of this study are provided in Table S1 and are available at the Sequence Read Archive (SRA) under BioProject accession no. PRJNA579546.

ACKNOWLEDGMENTS
We are grateful to the efforts at the Candida Genome Database (http://www .candidagenome.org/) for archiving and annotating Candida sequence information. We thank Shetuan Zhang, Bruce Banfield, and Peter Davies for the use of microscopy facilities. We thank Richard Bennett for the mNeon-and mScarlet-encoding genes. We thank Michèle Loewen and Simon Foote for RNA collection and analysis. We thank Leah Cowen for providing the plasmid used to generate the tetracycline-regulatable (TR) promoter system.