Phosphoregulation of Nap1 Plays a Role in Septin Ring Dynamics and Morphogenesis in Candida albicans

We previously reported that Gin4 and Cdks regulate hyphal growth through phosphoregulation of the septins in C. albicans (19, 20). To further unravel the underlying molecular mechanisms, we continue to study proteins that have been implicated in the regulation of Gin4 and Cdks. Previous findings of Nap1’s association with Clb2 and Gin4 in S. cerevisiae (18, 23) prompted us to investigate its C. albicans homologue. Using the S. cerevisiae Nap1 (ScNap1) sequence to search the C. albicans protein database, we found a 49% identity with an uncharacterized protein encoded by orf19.7501, which was thus named C. albicans Nap1 (CaNap1). To determine its function, we first constructed nap1Δ/Δ mutants. We noted that unlike wild-type (WT) cells that grew colonies with a smooth surface, the nap1Δ/Δ cells formed wrinkled colonies with filaments emanating from the edge (Fig. 1A), indicating filamentous growth. nap1Δ/Δ cells also invaded the agar. The colonies could not be washed off the agar surface as easily as could WT cells (Fig. 1B), and microscopic examination of the agar sections underneath the mutant colonies revealed deep penetration by fungal filaments (Fig. 1C). Consistently, in liquid cultures, many nap1Δ/Δ cells exhibited elongation and swelling at 30°C (Fig. 1A). This phenotype was enhanced at 37°C with the formation of many pseudohyphae, whereas WT cells grew as yeast cells under the same conditions (Fig. 1A). However, true hyphal growth was unaffected in nap1Δ/Δ cells under all conditions tested (data not shown). Reintegrating a WT copy of NAP1 at its native locus in the nap1Δ/Δ mutant fully rescued the defects (Fig. 1).

CaNap1 was previously identified in affinity-purified septin complexes (19), suggesting colocalization with the septins. To determine if and when Nap1 colocalizes with the septins, we constructed a strain expressing a functional Nap1-green fluorescent protein (GFP) fusion as the sole source of this protein under the control of its native promoter. In yeast cells, Nap1-GFP first localized at the presumptive budding site and then at the tip of the emerging bud. When the small bud was taking shape, Nap1-GFP formed a ring at the bud neck, which persisted until cytokinesis when the ring split into two parallel rings (Fig. 2A). During hyphal growth, Nap1-GFP localized first to the tip of the germ tubes and later formed a band marking the site of the septum, and the band split into two bands at cytokinesis which did not disassemble, at least through the next cell cycle (Fig. 2A). The dynamic localization of Nap1 closely resembles that of the septins (33, 34). Consistently, Western blot (WB) analysis showed that hemagglutinin (HA)-tagged Nap1 coimmunoprecipitated (co-IP) with Myc-tagged Cdc3 and Gin4 (Fig. 2B).

The septin ring recruits proteins to the bud neck (14, 35). To ask if Nap1’s cellular localization depends on the septins, Nap1-GFP was expressed in the cdc10Δ/Δ and cdc11Δ/Δ mutants, in which the septin ring can form but with abnormalities (7). Interestingly, we found Nap1-GFP entirely in the cytoplasm in both mutants (Fig. 2C). WB analysis showed that the cellular levels of Nap1-GFP in the septin mutants were comparable to that in WT cells (see Fig. S1 in the supplemental material). Thus, the data indicate that proper septin ring organization is required for recruiting Nap1 to the bud neck.

We also examined if Nap1 is required for septin ring assembly by expressing Cdc3-GFP in the nap1Δ/Δ mutant. We found that although septin rings formed at the bud neck in most yeast cells with a normal morphology, abnormalities in both septin localization and organization were often observed in the elongated cells (Fig. 2D): the septin rings frequently appeared broken, and random dots were present in many cells as well. The defects were stronger at 37°C than at 30°C, consistent with the presence of more pseudohyphae at 37°C. Previously, we reported that the septin ring assembly depends on the kinase-independent functions of Gin4 (19). It is thus possible that Gin4’s localization to the bud neck and its function were also impaired in the nap1Δ/Δ mutant. Consistent with this idea, Gin4-GFP in the nap1Δ/Δ cells also mislocalized in ways similar to Cdc3-GFP (Fig. 2E), suggesting that the abnormalities in Gin4 localization might be responsible for the septin defects in the nap1Δ/Δ mutant. However, it is also possible that Nap1 has both direct and indirect roles in regulating septin ring assembly.

Cell cycle-dependent phosphorylation plays a key role in regulating the structure and function of the septins and their associated proteins (19, 20). To ask if Nap1 is also phosphoregulated, we performed Phos-tag gel electrophoresis (36) and WB analysis of protein extracts and detected Nap1 as two bands. The slower-migrating one disappeared after pretreatment of the sample with λ-phosphatase (λPP), indicating a phosphoisoform (Fig. 3A). We next examined if Nap1 phosphorylation is cell cycle dependent. We used centrifugal elutriation to collect early G₁ cells, >95% of which had not formed the new septin ring, and then released them into glucose minimal medium (GMM) for growth at 30°C. Aliquots of cells were collected at timed intervals to determine the budding index as well as the phosphorylation status of Nap1. Phosphorylation at each time point was quantified by measuring the band intensity and calculating the intensity ratio of phosphorylated to unphosphorylated Nap1 (Phos/Unphos). Figure 3B shows that the Phos/Unphos ratio was at the highest (~1.0) in newly collected G₁ cells, declined when the cells were traversing through the G₁ phase, and reached the lowest level of ~0.51 around the time of bud emergence. The Phos/Unphos ratio increased and remained high through the rest of the cell cycle until the beginning of the next one, when it decreased again. The data demonstrated that Nap1 phosphorylation is cell cycle dependent. To map the phosphorylation sites, Nap1-GFP was isolated from yeast cells using GFP-Trap for MS analysis. Twelve phosphorylated serine (S) and threonine (T) residues were detected. Interestingly, 10 of the 12 residues resided from S16 to T52 near the N terminus. Phosphorylation at similar sites was also observed in ScNap1 (27), suggesting that the phosphorylation has a conserved but yet undefined function. The other two phosphoresidues are S403 and S411, which are not conserved in ScNap1.

We next studied the physiological significance of Nap1 phosphorylation in the N-terminal region. All 10 phosphorylated residues, S16, T20, T24, S27, T29, S31, S35, T42, S46, and T52, were mutated to either the nonphosphorylatable alanine (nap1-10A) or the phosphomimetic glutamic acid (nap1-10E). To evaluate the function of Nap1-10A and Nap1-10E, they were tagged at the C terminus with GFP, HA, or Myc and expressed in the nap1Δ/Δ mutant from the native promoter. We first examined the abilities of the two alleles to rescue the growth defects of the nap1Δ/Δ mutant. We found that while nap1-10E could largely rescue the defects, nap1-10A failed to do so (Fig. 4A), indicating that the N-terminal phosphorylation of Nap1 is important for yeast morphogenesis. Next, we examined the cellular localization of Nap1-10A-GFP and Nap1-10E-GFP. Strikingly, Nap1-10A was found primarily in the cytoplasm with only a faint signal at the bud neck, whereas Nap1-10E localized normally to the bud neck (Fig. 4A). The observation suggested that Nap1 phosphorylation is important for its bud neck localization. We thought that the mislocalization of Nap1-10A might reflect weakened interaction with the septins. In support of this idea, repeated co-IP experiments consistently showed an ~20% decrease in the interaction between Nap1-10A and Cdc3 and an ~11% increase between Nap1-10E and Cdc3 compared with that between WT Nap1 and Cdc3 (Fig. 4B and C). The co-IP experiments were performed on cells grown at both 30°C and 37°C, which produced similar results, and WB analysis showed that the cellular levels of both mutant Nap1 proteins were comparable to that of the WT protein (data not shown). The moderate changes detected by co-IP in the interaction between Nap1-10A and Cdc3 are apparently in contrast to the strong reduction of Nap1-10A localization at the bud neck. A plausible explanation is that the mutations might have mainly affected Nap1’s interaction with the septin polymers at the bud neck without significantly reducing its affinity for the septin molecules in the cytoplasm. Taken together, the data suggest that phosphorylation in the N terminus has a role in regulating Nap1’s localization to the bud neck.

We next examined the phosphorylation status of Nap1-10A and Nap1-10E. Unexpectedly, both proteins exhibited a Phos/Unphos ratio greater than 2 (Fig. 5A). This was in contrast to the WT Nap1, which had a Phos/Unphos ratio of <1.0. This may suggest that mutating the 10 phosphorylation sites led to enhanced phosphorylation at other sites, which might occur only transiently in WT cells and/or at sites which are normally not phosphorylated. It is also possible that the mutations may influence Nap1 dephosphorylation due to mislocalization of the protein or change of the structural context required for dephosphorylation. To begin to test these hypotheses, we examined if Nap1 phosphorylation depends on its localization to the bud neck. To this end, we analyzed Nap1 phosphorylation in cdc10Δ/Δ and cdc11Δ/Δ mutants in which Nap1 localized to the cytoplasm (Fig. 2C). Strikingly, Nap1 in both mutants showed not only large amounts of phosphoisoforms but also an additional band with much lower mobility compared with Nap1 in WT cells (Fig. 5B). The data strongly suggest that the regulation of Nap1 phosphorylation requires its localization to the bud neck. We next examined the phosphorylation status of Nap1-10A and Nap1-10E in cdc10Δ/Δ cells. Surprisingly, although Nap1-10A in cdc10Δ/Δ cells showed a high Phos/Unphos ratio, as it did in WT cells, it did not show the slowest-migrating phosphoisoform observed for WT Nap1 in the same septin mutants (Fig. 5C). This lack of further phosphorylation of Nap1-10A in cdc10Δ/Δ cells suggested that the phosphorylation sites in Nap1-10A might overlap the additional phosphorylation sites in WT Nap1 when expressed in the septin mutant. This is consistent with the fact that both Nap1-10A in WT cells and WT Nap1 in the septin mutants fail to localize to the bud neck. We were unable to detect Nap1-10E in the nap1Δ/Δ cdc10Δ/Δ and nap1Δ/Δ cdc11Δ/Δ mutants (see Fig. S2 in the supplemental material), although it was expressed in exactly the same way as Nap1-10A, indicating that Nap1-10E is highly unstable in the septin mutants.

The coincidence of greater Nap1 phosphorylation with its failure to localize to the bud neck suggested that Nap1 might be dephosphorylated at the bud neck in WT cells. Next, we examined Nap1 phosphorylation in rts1Δ/Δ and cdc14Δ/Δ mutants. Rts1 is a regulatory subunit of protein phosphatase 2A (PP2A), and Cdc14 is a phosphatase that regulates the mitotic exit. Both phosphatases transiently localize to the bud neck (37, 38). In this experiment, WT, rts1Δ/Δ, and cdc14Δ/Δ cells with tagged Nap1 were grown to the log phase. Phos-tag gel analysis revealed a significant increase in the Phos/Unphos ratio of Nap1 in both mutants, 0.97 in rts1Δ/Δ and 2.49 in cdc14Δ/Δ cells, in comparison with 0.64 in WT cells (Fig. 5D). Thus, the data support the idea that Nap1 is dephosphorylated on the septin ring and is dependent on Cdc14 and Rts1.

We next asked if the septin-associated kinases Cla4 and Gin4 are involved in Nap1 phosphorylation. We used a cla4Δ/Δ mutant and a gin4-KD (kinase-dead) mutant (19). While the role of Gin4 in septin organization is well established (19, 20, 39), the effect of CLA4 deletion on the septins had not been reported in C. albicans. Thus, we first examined septin behavior in cla4Δ/Δ cells. Figure 5E shows that although Cdc3-GFP localized to the presumptive budding site, it failed to form a septin ring at the bud neck. Instead, the septins remained at the bud tip, leading to an elongated bud. Only at a later stage was a septin ring formed close to the tip. cla4Δ/Δ cells were also defective in cytokinesis/cell separation, forming clusters of cells with mislocalized septins. The results confirmed a role for Cla4 in regulating the septin ring. Next, we examined Nap1 phosphorylation in cla4Δ/Δ and gin4-KD mutants. Strikingly, we found that Nap1 phosphorylation was abolished in both mutants (Fig. 5D), showing dependence of Nap1 phosphorylation on Cla4 and Gin4.

To gain more insight into the functional relationship between Nap1 and septins, we investigated the genetic interaction between nap1Δ/Δ and cdc10Δ/Δ mutants. We compared the morphological phenotypes of the single mutants and the double mutant. Although the morphology of cdc10Δ/Δ yeast cells was rather normal, deletion of CDC10 together with NAP1 caused a much stronger filamentous phenotype than did deleting NAP1 alone (Fig. 6A). The pseudohyphal phenotype of the nap1Δ/Δ cdc10Δ/Δ mutant was corrected when a WT copy of NAP1 was introduced. The results demonstrated a genetic interaction between NAP1 and CDC10, indicating that they contribute to the same cellular process at least partially through parallel pathways or work in the same protein complex.

In S. cerevisiae, septin defects often result in filamentous growth through a Swe1-mediated morphogenesis checkpoint (40). In C. albicans, the hsl1Δ/Δ mutant undergoes Swe1-mediated pseudohyphal growth (41). To determine if Swe1 has a role in the filamentous growth in the nap1Δ/Δ mutant, we constructed a nap1Δ/Δ swe1Δ/Δ mutant and found that deleting SWE1 had no significant effect (Fig. 6B), suggesting that the filamentous growth of nap1Δ/Δ is independent of Swe1. Wightman et al. (39) also reported a weak dependence of pseudohyphal growth on Swe1 in the gin4Δ/Δ mutant.

Septin ring dynamics plays an important role in cell division and morphogenesis (37, 42). Next, we conducted fluorescence recovery after photobleaching (FRAP) analysis to determine if the deletion of NAP1 affects septin ring dynamics. To visualize the septin ring, we tagged Cdc3 with GFP. FRAP analysis was performed at three different budding stages, namely, in small-budded cells, large-budded cells with a single ring, and large-budded cells with split rings. The full ring was bleached, and fluorescence recovery was recorded at 1-min intervals (see Fig. S3 in the supplemental material). In WT cells, the fluorescence intensity of Cdc3-GFP recovered to a maximum of ~60% in small-budded cells and ~40% in large-budded cells with either a single ring or split rings (Fig. 7). Interestingly, in nap1Δ/Δ cells, the recovery was consistently higher in all three budding stages, suggesting that the septin rings assembled in the absence of Nap1 are more dynamic.

To determine if the change of septin dynamics in the NAP1 mutants affects the timing of septin ring disassembly, we used time-lapse imaging to record the time elapsed between the time of septin ring split and that of ring disappearance at the end of the cell cycle in WT, nap1-10A, and nap1-10E cells. However, we did not detect a significant difference between the WT and the mutant cells (data not shown). A possible explanation is that the moderate changes in septin ring dynamics observed in the NAP1 mutants are not sufficient to cause a significant shift in the timing of septin disassembly. Alternatively, septin ring disassembly and dynamics may not always be coupled.

Next, we asked if Nap1 is important for the virulence of C. albicans. We first determined the growth rates of the mutants in GMM at 37°C. We found that the doubling times of the NAP1 mutants were not significantly different from that of the WT (Fig. 8A). To test virulence, we conducted intravenous infections of BALB/c mice. To assess the abilities of the mutants to colonize the kidney, we counted CFU per kidney 2 days postinoculation (Fig. 8B, top). The mice infected with the WT strain had ~3.3 × 10⁵ CFU/kidney. In sharp contrast, the CFU/kidney from the animals infected with the nap1Δ/Δ mutant was ~1,000 times less (6.1 × 10²). The reintegrant strain produced 4.2 × 10⁵ CFU/kidney, comparable to the WT strain. Interestingly, despite the fact that Nap1-10E, but not Nap1-10A, can significantly alleviate the growth defects of the nap1Δ/Δ mutant, both nap1-10A and nap1-10E mutants showed poor colonization of the kidney, as they produced 8.3 × 10³ and 6.6 × 10² CFU/kidney, respectively. All mice infected with the WT or the reintegrant strain died within 5 days. Among the mice infected with the mutants, only two mice infected with the nap1-10A mutant died later, on the 8th and 15th days (Fig. 8C). As the nap1Δ/Δ mutant forms pseudohyphae at 37°C, we thought that the filamentous shape might prevent the cells from reaching the kidney effectively. To test this idea, we sacrificed mice 5 h after inoculation to count the CFU/kidney but found no significant difference between mice infected with the WT strain and those infected with the nap1Δ/Δ mutants (Fig. 8B, bottom). The results indicate that the attenuation of virulence in the NAP1 mutants is not a consequence of pseudohyphal growth. Most likely, the capacity for proliferation or survival of the mutants in the host is compromised. Nevertheless, the tests demonstrate that Nap1 is crucially required for the virulence of C. albicans.

In S. cerevisiae, Nap1 has been implicated in the regulation of gene expression. However, we do not observe CaNap1 localizing to the nucleus even after deleting a putative NES, deletion of which in ScNap1 causes a nuclear localization (27). Furthermore, the CK2 phosphorylation sites required for nuclear localization in ScNap1 (26) are not conserved in CaNap1. To confirm that deletion of NAP1 acts primarily through effects on the septin ring rather than through gene regulation, we performed microarray analysis comparing gene expression in WT cells to that in nap1Δ/Δ mutants. In total, 142 genes changed in expression by at least 2-fold at a 5% false discovery rate (FDR) threshold, 53 up and 89 down (see Table S2 in the supplemental material). KEGG and GO enrichment analysis revealed that many of these genes were related to the metabolism of glucose (KEGG pathways such as “glycolysis/gluconeogenesis” and “carbon metabolism” and GO terms such as “carbohydrate metabolic process” were identified); in particular, genes in the glycolytic pathway were downregulated (see Table S3).

In this study, we found that CaNap1 is phosphorylated in a cell cycle dependent manner and identified a cluster of phospho-S/T residues near the N terminus. Unlike WT Nap1 that localized to the bud neck, the nonphosphorylatable mutant Nap1-10A localized mainly in the cytoplasm. Also, the Nap1-10A mutant failed to rescue the morphological defects of nap1Δ/Δ cells. In contrast, the phosphomimetic Nap1-10E exhibited normal localization and could significantly correct the morphological defects of the nap1Δ/Δ mutant. Therefore, the N-terminal phosphorylation of Nap1 is important for its function. Our data further suggest that the phosphorylation may regulate Nap1’s interaction with the septins, because co-IP experiments detected reduced interaction of Nap1-10A and increased interaction of Nap1-10E with Cdc3. However, the high instability of Nap1-10E observed in the nap1Δ/Δ cdc10Δ/Δ background and its negative impact on virulence in mice clearly indicate that Nap1-10E is not fully functional.

Other than the phosphorylation at the 10 N-terminal sites identified by MS, it is likely that Nap1 is also phosphorylated at other sites that MS failed to detect, as both Nap1-10A and Nap1-10E showed a significantly higher phosphorylation level than the WT protein. As Nap1-10A localized primarily to the cytoplasm, we hypothesized that some residues of Nap1 may be phosphorylated in the cytoplasm and then dephosphorylated once Nap1 localizes to the septin ring. This hypothesis is supported by the observation of Nap1 hyperphosphorylation in the cdc10Δ/Δ and cdc11Δ/Δ mutants where Nap1 also fails to localize to the septin ring. Furthermore, the fact that Nap1-10A did not show a further increase in its phosphorylation level when expressed in the septin mutants suggests that the additional phosphorylation caused by the 10A mutation is likely to have occurred at the same residues as that caused by the deletion of CDC10 or CDC11. Although the exact mechanism of Nap1 dephosphorylation remains unclear, we discovered that two septin-associated protein phosphatases, PP2A and Cdc14, are involved in this process. We realize that the above argument is unable to explain the hyperphosphorylation of Nap1-10E, which apparently can localize to the septin ring. One possible explanation is that the phosphomimetic mutations caused phosphorylation at sites that are poor substrates of phosphatases or the phosphorylation caused structural changes that render the sites inaccessible to phosphatases. The extreme instability of Nap1-10E observed in nap1Δ/Δ cdc10Δ/Δ cells is in agreement with this explanation.

Calvert et al. (27) reported that Nap1 is phosphoregulated in S. cerevisiae. They identified about a dozen phosphoresidues on Nap1 and demonstrated that 3 residues near the C terminus are targets of CK2. Notably, the rest of the phosphorylated residues were found in a cluster near the N terminus, although the role of phosphorylation was not determined. Not all the phosphorylated residues found in ScNap1 are conserved in CaNap1. For example, the three CK2 sites are not found in CaNap1 by sequence alignment. As these CK2 sites are important for the nuclear localization of Nap1 in S. cerevisiae (27), their absence may imply that Nap1 has reduced nuclear function in C. albicans. Consistently, we have never detected Nap1 in the nucleus in C. albicans, even when we disrupted the putative NES (data not shown), mutation of which in ScNap1 caused Nap1 nuclear localization (27). We found that Nap1 phosphorylation was completely abolished in cla4Δ/Δ and gin4-KD mutants, suggesting that the two septin-associated kinases play an important role in the regulation of Nap1 phosphorylation. The role could be direct or indirect or both, and future investigation is required to distinguish these possibilities. As Nap1, Cla4, and Gin4 also regulate septin organization in S. cerevisiae (6, 17), the phosphoregulation of Nap1 elucidated in this study may have revealed evolutionarily conserved mechanisms for septin regulation in fungi. The septin and morphological defects observed in the NAP1 mutants together with the dependence of its phosphorylation on septin-associated kinases and phosphatases strongly support the idea that Nap1’s role in C. albicans is primarily at the bud neck, but not through the regulation of gene expression. This view is supported by the microarray data.

The nap1Δ/Δ and the nap1-10A and nap1-10E mutants exhibited greatly reduced virulence in the mouse model of candidemia. All mice infected with the nap1Δ/Δ or nap-10E mutant and 75% of the mice infected with the nap1-10A mutant survived to the end of the experiment, whereas all those infected with the WT or the NAP1 reintegrant strain died within 5 days. Warenda et al. (31) reported that both cdc10Δ/Δ and cdc11Δ/Δ mutants are defective in virulence. As these septin mutants also destabilize the septin ring (7), the impaired virulence of the NAP1 mutants could also be a result of septin ring defects. When looked at closely, however, there are distinct features in the defects between NAP1 and septin mutants. The nap1Δ/Δ and nap1-10A cells showed enhanced invasive growth on agar, whereas septin mutants are defective in invasive growth. The septin mutants showed normal colonization of the kidney, and the reduced virulence was likely due to their defects in tissue invasion. As the average CFU/kidney of mice infected with nap1Δ/Δ and nap1-10A cells at day 2 was ~1,000-fold less than that for the WT strain, we thought that the pseudohyphal growth of the mutant cells might have reduced their ability to travel through the blood vessels to reach the kidney for colonization. Consistent with this view, Umeyama et al. (41) demonstrated that hsl1Δ/Δ mutant cells, which also exhibit filamentous growth and invasion of agar, have reduced virulence due to defective organ colonization. However, when we sacrificed the mice at 5 h postinfection, we did not observe a significant difference in CFU/kidney between mice infected with the WT strain and those infected with the nap1Δ/Δ mutant. Thus, the attenuated virulence of the NAP1 mutants in mice may be a result of impairment of other abilities required for pathogenesis, such as the ability to survive and proliferate in the host. A surprising result was the loss of virulence in the nap1-10E mutant, which seemed indistinguishable from WT cells in growth and morphology. However, the extreme instability of Nap1-10E in the nap1Δ/Δ cdc10Δ/Δ mutant indicates that Nap1-10E is not fully functional, which could underlie the loss of virulence. Regardless of the exact mechanisms by which Nap1 contributes to infection, the critical dependence of C. albicans virulence on Nap1 in mouse suggests that new antifungal drugs may be developed by targeting Nap1.

Medium preparation, phosphorylation site mapping, protein extraction, immunoprecipitation, Western blotting, and microscopy were performed as described previously (13, 20).

Strain BWP17 (43) was used to create all the gene deletion mutants in this study. A gene deletion cassette contained a selection marker gene, URA3, HIS1, or ARG4, flanked with DNA fragments corresponding to the 5′ and 3′ untranslated region of the gene. The two copies of NAP1 were deleted with the recyclable hisG-URA3-hisG cassette, so that all three selection markers (HIS1, ARG4, and URA3) can be used for subsequent genetic manipulations. All gene deletion mutants were verified by PCR. To reintegrate the WT or mutated NAP1 gene into the nap1Δ/Δ mutant, the full-length NAP1 open reading frame (ORF) together with the 971-bp upstream promoter region sequence was PCR amplified with KpnI and XhoI sites added to the 5′ and 3′ ends, respectively. The amplified DNA fragment was cloned in frame with a C-terminal Myc, HA, or GFP tag in the CIP10 vector (44). A PmeI site was generated in the promoter region for linearization by site-directed mutagenesis using the QuikChange multi-site-directed mutagenesis kit (Agilent Technologies). The same mutagenesis procedure was used to generate all NAP1 mutants in this study. Mutated constructs were integrated into the HZX01 strain at the endogenous NAP1 promoter region.

FRAP analysis was performed using an inverted confocal laser LSM700 microscope (Carl Zeiss). The septin ring was bleached with a 488-nm laser for 25 iterations at 50% intensity. After the bleaching, images were collected every 1 min. Fluorescence intensity was analyzed using ImageJ. For each strain at each cell cycle stage examined, five individual recovery curves were determined. Average values of the bleached region of the septin ring were calculated. A reference cell in the imaging field was used to correct for general bleaching. The statistical analysis and the graphs were done using Prism software.

C. albicans cells were grown overnight at 30°C in yeast extract-peptone-dextrose (YPD). Cells were washed twice in phosphate-buffered saline (PBS) and diluted with PBS to 5 × 10⁶ cells/ml. Two hundred microliters of the cell suspension was injected into each BALB/c mouse. For each strain, 12 mice were injected. Two mice were sacrificed 5 h and 2 days postinjection for quantifying CFU/kidney. Kidneys were homogenized individually with a mortar and pestle for 1 min, and then serial dilutions of the homogenate were plated onto YPD plates for growth at 30°C for 2 days before the colonies were counted. The remaining 8 mice were monitored for 20 days, and a mouse was considered to be dead when it had lost 20% of its body weight or become moribund. The statistical analysis and the graphs were done using Prism software. Animal experiments were carried out according to National Advisory Committee for Laboratory Animal research guidelines, and all procedures were approved by the IACUC of the Agency for Science, Technology and Research of Singapore.

Total RNA was isolated from log-phase yeast cells in YPD using the RNeasy minikit from Qiagen. Per extraction, ~2 × 10⁷ cells were used, and cells were disrupted with a MicroSmash MS-100 beater (Tomy Medico, Tokyo, Japan).

Arrays were purchased from Microarrays Inc. (Huntsville, AL) and processed according to the manufacturer’s instructions and published protocols (45, 46). Arrays were scanned using the GenePix 4000B Axon Instruments scanner (Molecular Devices, CA). Images were analyzed using the GenePix Pro 6.0 software (Molecular Devices). Microarray data were analyzed using the limma package (version 3.14.4) from the R/Bioconductor collection (43). Briefly, background intensity was corrected using the “normexp” method. Probes with a background-corrected intensity of >50 in both channels were removed from further analysis. Arrays were then normalized using print tip loess normalization, and differential expression was assessed using an empirical Bayes moderated t test. Tests were controlled for multiple testing using a Benjamini and Hochberg false discovery rate (45). Where multiple probes for a single gene were significant at a 5% FDR, fold change was calculated as the mean of significant probes.