Dissection of the Candida albicans Cdc4 protein reveals the involvement of domains in morphogenesis and cell flocculation

Background CDC4, which encodes an F-box protein that is a member of the Skp1-Cdc53/Cul1-F-box (SCF) ubiquitin E3 ligase, was initially identified in the budding yeast Saccharomyces cerevisiae as an essential gene for progression through G1-S transition of the cell cycle. Although Candida albicans CDC4 (CaCDC4) can release the mitotic defect caused by the loss of CDC4 in S. cerevisiae, CaCDC4 is nonessential and suppresses filamentation. Results To further elucidate the function of CaCDC4, a C. albicans strain, with one CaCDC4 allele deleted and the other under the repressible C. albicans MET3 promoter (CaMET3p) control, was made before introducing cassettes capable of doxycycline (Dox)-induced expression of various C. albicans Cdc4 (CaCdc4) domains. Cells from each strain could express a specific CaCdc4 domain under Dox-induced, but CaMET3-CaCDC4 repressed conditions. Cells expressing domains without either the F-box or WD40-repeat exhibited filamentation and flocculation similarly to those lacking CaCDC4 expression, indicating the functional essentiality of the F-box and WD40-repeat. Notably, cells expressing the N-terminal 85-amino acid truncated CaCdc4 partially reverse the filament-to-yeast and weaken the ability to flocculate compared to those expressing the full-length CaCdc4, suggesting that N-terminal 85-amino acid of CaCdc4 regulates both morphogenesis and flocculation. Conclusions The F-box and the WD40-repeat of CaCdc4 are essential in inhibiting yeast-to-filament transition and flocculation. The N-terminal region (1–85) of CaCdc4 also has a positive role for its function, lost of which impairs both the ability to flocculate and to reverse filamentous growth in C. albicans.


Background
Candida albicans is a natural diploid without a complete sexual cycle and exists as yeast, pseudohyphal, and hyphal cells [1]. It is capable of a morphological switch induced by environmental stimuli [2], essentially via cAMP-mediated and MAPK signaling pathways [3]. Importantly, its ability to alter morphology among cell types is associated with virulence to humans [4]. Many cell cycle regulators including cyclins are also known to control morphogenesis in C. albicans [5].
Although CaCdc4 is a structural homolog of S. cerevisiae Cdc4 (ScCdc4) and is capable of rescuing the mitotic defect caused by the loss of ScCDC4 in S. cerevisiae [7], the functions of CaCdc4 and ScCdc4 are dissimilar as the null Cacdc4 mutant is viable and the depletion of CaCdc4 causes the accumulation of Sol1 (Sic1 like) for hyphal development rather than initiation of cell cycle arrest [6]. This verifies that CaCDC4 is nonessential and suppresses filamentation and suggests that controlling the degradation on Sol1 in C. albicans by CaCdc4 is important for inhibition of filamentation. Therefore, while C. albicans Sol1 is likely a substrate of SCF CaCdc4 , which can be demonstrated by the reduction of Sol1 when CaCdc4 is overexpressed [6], there has not been any direct evidence to support this hypothesis. Additionally, the filamentous properties for mutants of Cacdc4 null and Cacdc4 sol1 double null were comparable. This refutes the idea that Sol1 is the sole target of CaCdc4. Indeed, with an affinity-purification approach, we have isolated at least two novel CaCdc4-associated proteins [12] that are potential substrates of CaCdc4.
To further elucidate the role of CaCDC4 and its mediation through a characteristic F-box protein of SCF ubiquitin E3 ligase in C. albicans, we have sought to dissect the CaCdc4 domains associated with filamentation. In this study, we made a C. albicans strain with one deleted CaCDC4 allele and repressed the other by CaMET3 promoter (CaMET3p) using methionine and cysteine (Met/Cys). We used this strain to introduce plasmids capable of inducing expression of various CaCdc4 domains with doxycycline (Dox). We observed the roles of F-box and WD40-repeat for CaCdc4 function and the possible role of the N-terminal 85-amino acid for morphogenesis. We also showed that C. albicans cells that lacked CaCdc4 triggered flocculation. Moreover, we found that N-terminal 85-amino acid of CaCdc4 is required for inhibition of both filamentation and flocculation.

Strains and growth conditions
E. coli strain DH5α was used for the routine manipulation of the plasmids. They were grown at 37°C in LB broth medium [13] or on plates containing 1.5% agar (Difco, BD Biosciences), with 50 μg/ml ampicillin or 30 μg/ml kanamycin. All C. albicans strains (Table 1) were derived from auxotrophic strain BWP17 (arg4/arg4 his1/his1 ura3/ura3) [14]. They were grown at 30°C in either yeast extract-peptone-dextrose (YEPD) or supplemented minimal synthetic defined (SD) medium with 2% glucose with or without 2% agar [15]. While Ura + prototrophs were selected on SD agar plates without uridine, His + prototrophs were selected on SD plates without histidine. Selection for the loss of the C. albicans URA3 (CaURA3) marker was performed on plates with 50 μg/ml uridine and 1 mg/ml 5-fluoroorotic acid (5-FOA, MD Bio). To repress the CaCDC4 expression that was controlled by CaMET3p, strains were grown on SD medium or on plates with 2.5 mM Met/Cys, which has been shown to optimally switch off the expression of the CaMET3p-driven downstream gene [16]. To induce gene expression under the Tet-on system, 40 μg/ml Dox (Sigma) was added to YEPD or SD media.

Plasmid DNA manipulation
Plasmid DNA was extracted routinely from E. coli cultures using Gene-Spin TM MiniPrep purification Kit-V 2 (PRO TECH, Taipei, Taiwan) and the instructions provided by the manufacturer. E. coli was transformed with plasmid DNA by using CaCl 2 . The DNA cassettes were introduced into C. albicans by the lithium acetate method as described previously [17].

Construction of C. albicans strains
Initially, a strain with repressed CaCDC4 expression was made. A mini-Ura-blaster cassette, flanked with 60-bp sequences homologous to CaCDC4, was PCR-amplified using a template of plasmid pDDB57 and long primers of CaCDC4-URA3-F and CaCDC4-URA3-R (Table 1). BWP17 was transformed by integration of the cassette into the CaCDC4 locus to generate Ura + strain JSCA0018. The plasmid pFA-HIS1-MET3p-CaCDC4, with a partial CaCDC4 coding sequence for N-terminal CaCdc4 (1-563), was linearized with BspEI and used to transform JSCA0018 to generate His + JSCA0021 ( Figure 1A; Table 1). Cells of JSCA0021 were plated with 5-FOA to induce recombination between two copies of dpl200 flanking the mini-Ura-blaster for a loss of CaURA3 to generate JSCA0022.
To define the function of the distinct CaCdc4 domains (Figure 2A), different CaCDC4 portions were used to replace the full length CaCDC4 coding sequence on pTET25M-CaCDC4-6HF. By using the primer sets listed in Table 2, the following constructs were made: pTET25M-ΔNCaCDC4-6HF (with primers CaCDC4 ΔN AatII and CaCDC4 ΔN XhoI), which encodes the Nterminal truncated CaCdc4; pTET25M-F-6HF (with primers CaCDC4 F-box AatII and CaCDC4 F-box XhoI), which encodes the F-box domain with flanking regions; pTET25M-WD40-6HF (with primers CaCDC4 WD40 AatII and CaCDC4 ΔN XhoI), which encodes eight copies of WD40-repeat; and pTET25M-ΔNF-6HF (with primers CaCDC4 ΔN AatII and CaCDC4 F-box XhoI), which encodes truncated N-terminal CaCdc4 and the F-box domain. All inserts of the constructs were released with AatII and XhoI to replace the full-length CaCDC4 on pTET25M-CaCDC4-6HF. Consequently, plasmids bearing those CaCDC4 segments flanked with common C. albicans ADH1 (CaADH1) sites were digested with SacII and KpnI, each of which was transformed into C. albicans for integration at the CaADH1 locus. All strains were verified by colony PCR with specific primers before subjecting to Southern blotting analysis.

Southern blotting analysis
Genomic DNA from the C. albicans strains was isolated by the MasterPure™ Yeast DNA Purification Kit (Epi-centre®, an Illumina company) according to the manufacture's instruction. Southern blotting was performed with the aid of the Rapid Downward Transfer System (TurboBlotter™, Whatman) using 10 μg of the restriction enzyme-digested genomic DNA. The DNA on the blot was hybridized with a probe amplified by the PCR DIG probe synthesis kit (Roche) with the primers CaCDC4_Probe_F and CaCDC4_Probe_R for CaCDC4 locus or CaADH1 Probe_F and CaADH1 probe_R for ADH1 locus ( Table 2) using DIG Easy Hyb (Roche). To reveal the structure of gene locus, the DIG Luminescent Detection Kit (Roche) was used after hybridization, and the luminescent images of blot were captured with the imaging analysis system (ImageQuant LAS4000 mini, GE Healthcare Life Sciences).

Protein extraction and Western blot analysis
Cultured cells were collected, and the total protein from each sample was extracted as described previously [20]. The proteins were resolved by 10% SDS-PAGE and transferred to PVDF membranes (PerkinElmer, Boston, USA). Proteins on the membranes were probed with polyclonal antibody to FLAG (Sigma) in 1:2000 dilution and detected using the SuperSignal West Pico Chemiluminescent Substrate Kit (PIERCE). These were recorded with the Luminescent Image Analyzer (FUJIFILM LAS-1000) and analyzed by ImageGauge 3.46 and L Process v 1.96 (FUJIFILM).

Ca 2+ -initiated flocculation assay
The FLO-encoded flocculins are known to be essential for flocculation in S. cerevisiae [21]. Functional homologues of FLO genes have been found in C. albicans. In particular, the important S. cerevisiae gene FLO11 responsible for flocculation has C. albicans functional counterpart ALS1 [22]. Since FLO11-associated flocculation is dependent on the presence of Ca 2+ , we adopted an alternative flocculation assay in which the rate of flocculation is initiated by Ca 2+ and the optical density was assessed within a short time-frame [23]. Briefly, to initiate flocculation, an aliquot of 800 μl deflocculated cell suspension was transferred into a 1-ml cuvette, followed by addition of 200 μl of 100 mM CaCl 2 . The cuvette was mixed robustly by pipetting and the absorbance (OD 600 ) was assessed instantly at 30-s intervals for 5 minutes using a spectrophotometer (DU800, Beckman Coulter, Inc.). All assays were conducted in triplicate.
Cacdc4::URA3-dpl200  Figure 1 Construction of a C. albicans strain for repressibly expressing CaCDC4. (A) Strain construction (detailed in the Methods). The first CaCDC4 allele on BWP17 was deleted by mini-Ura-blaster to obtain JSCA0018. Plasmid pFA-HIS1-MET3p-CaCDC4 containing partial CaCDC4 coding sequence was linearized at a unique site for introducing into strain JSCA0018 to generate JSCA0021. 5-FOA was used to counter-select CaURA3 removal to obtain JSCA0022 for re-introducing the Tet-on plasmid with a CaURA3 marker. (B) Verification of constructed strains by Southern blotting analysis. Organization of the CaCDC4 locus with respect to NdeI sites is shown. The relative positions of the probe used and the predicted NdeI-digested pattern of the CaCDC4 locus are indicated. Two NdeI-fragments of 14 kb and 8.5 kb, specific to CaCDC4, could be detected in genomic DNA from BWP17 digested with NdeI; two NdeI-fragments of 8.5 kb and 4.5 kb, specific to CaCDC4 and Cacdc4::URA3-dpl200, respectively, could be detected in JCSA0018; two NdeI-fragments of 4.5 kb and 7.4 kb specific to Cacdc4::URA3-dpl200 and Cacdc4::P MET3 -CDC4:HIS1, respectively, could be detected in JSCA0021; and two NdeI-fragments of 13.5 kb and 7.4 kb specific to Cacdc4::dpl200 and Cacdc4::P MET3 -CDC4:HIS1, respectively, could be detected in JSCA0022. A non-specific NdeI-fragment is indicated as "*" can be detected in all strains tested.

Results
Constructing a C. albicans strain capable of conditionally repressing the expression of CaCDC4 To establish C. albicans strains capable of expressing CaCDC4 and its domains solely controlled under a Tet promoter directly in C. albicans, BWP17, with both alleles of CaCDC4 deleted, was constructed to accommodate Tet-on plasmid cassettes capable of expressing assorted CaCdc4 domains induced by Dox. The first allele of CaCDC4 was deleted in BWP17 by mini-Ura-blaster [24] to generate the JSCA0018 strain ( Figure 1A; Table 1). This strain was used to delete the second CaCDC4 allele to obtain a Cacdc4 null mutant. However, Cacdc4 null mutant cells growing as filamentous form with toughened cell walls obstructed transformation.
To overcome this problem, the strain JSCA0021 ( Figure 1A; Table 1) was created that had one CaCDC4 allele deleted and the other under CaMET3 control that was Met/Cys repressible. To allow the introduction of Tet-on cassettes with the same CaURA3 selectable marker as the mini-Ura-blaster on JSCA0021, 5-FOA was used as a counter-selection agent to remove CaURA3 from JSCA0021, from which JSCA0022 was obtained ( Figure 1A; Table 1). The strains were PCR-confirmed with specific primers before subjecting to Southern blotting analysis. The CaCDC4 locus from BWP17 strain could detect two NdeI-digested fragments with size of 14 kb and 8.5 kb, respectively ( Figure 1B). The size shifting of NdeI-fragment flanking CaCDC4 from 14 kb to 4.5 kb demonstrated that one CaCDC4 allele was integrated with the mini-Urablaster cassette as in strain JSCA0018 ( Figure 1B). The size shifting of NdeI-fragment flanking CaCDC4 from 8.5 kb to 7.4 kb demonstrated that the other CaCDC4 allele integrated with the MET3-diven CaCDC4 plasmid as in strain JSCA0021 ( Figure 1B). Strain JSCA0021 could be further popped out the mini-Ura-blaster cassette to obtain strain JSCA0022 in which the size shifting of NdeI-fragment flanking CaCDC4 from 4.5 kb to 13.5 kb ( Figure 1B). These results indicate that all strains constructed have expected organizations in their genome.
Phenotypic verification of C. albicans strains capable of conditionally repressing the expression of CaCDC4 It has been shown that Uraauxotrophic mutants are avirulent [25] and other virulence-associated features can be influenced by the level of CaURA3 gene expression [26]. To assess presence of CaURA3 having effect on yeast-to-filament transition, the yeast-to-filament transitions between strain JSCA0021 and JSCA0022 were compared, cells of those strains were assessed under CaMET3p repressed or de-repressed conditions. Cells of both strains on SD plates without Met/Cys grew as circular colonies with smooth surfaces (Figure 2). By contrast, cells on plates with Met/Cys formed irregular colonies with filaments ( Figure 2). Under the microscope, these strains exhibited equivalent filamentous forms, suggesting a comparable ability to deplete CaCDC4 for expression and inability of CaURA3 interfering with yeast-to-filament transition in C. albicans. Subsequently, JSCA0022 was used as a parental strain to introduce the Tet-on cassettes (with CaURA3 marker) that encoded assorted CaCdc4 domains.
Establishment of Tet-on cassettes capable of expressing assorted CaCDC4 domains in C. albicans reveals that both the F-box and WD40-repeat are required for CaCdc4 function The filamentous development of JSCA0022 under CaMET3p-CaCDC4 repressed conditions, with Met/Cys and the Tet-on system, allows us to study the function of the CaCdc4 domains. A set of Tet-on cassettes (obtained from pTET25M-CaCDC4-6HF, pTET25M-ΔN-6HF, pTET25M-F-box-6HF, pTET25M-WD40-6HF, and pTET25M-ΔNF-6HF) that encoded each of the assorted domains of CaCdc4 ( Figure 3A) were used to transform JSCA0022 (which contained a CaMET3prepressible CaCDC4) to Ura + by integration at the CaADH1 locus ( Figure 3B). The correctness of the strains was confirmed by yeast colony PCR with specific primers before Southern blotting analysis. The CaADH1 locus from strain JSCA0022 could detect a SpeI-digested fragment with size of 3.3 kb ( Figure 3C). The CaADH1 locus from strains JSCA0023 and JSCA0024 detected an increased SpeI-digested fragment of 9.4 kb due to the integration of Tet-on cassettes of either pTET25M-CaCDC4 or pTET25M-CaCDC4-6HF ( Figure 3C). The CaADH1 locus from other strains also showed expected alteration in size according to the size of different CaCDC4 domains ( Figure 3C). These results confirmed the correctness of the strains. The JSCA0022 strain, which expressed the non-tagged and repressible CaCdc4, was used as a negative control. The sample obtained from JSCA0022 contained two prominent proteins of approximately 55 kDa and 72 kDa ( Figure 4A) which were presumably a result of crossreactivity to the anti-FLAG antibody. Those two proteins were used as an internal control. The F-box and WD40repeat proteins from strains JSCA0026 and JSCA0027 migrated to their expected positions of approximately 19 kDa and 43 kDa ( Figure 4A), respectively. However, the full-length CaCdc4 and the N-terminus truncated CaCdc4 (ΔN) from strains JSCA0024 and JSCA0025 exhibited signals at positions corresponding to 100 kDa and over 100 kDa ( Figure 4A), respectively, as opposed to 86 kDa and 77 kDa, respectively. Three distinctive signals ( Figure 4A) were observed for strain JSCA0030 expressing ΔNF of CaCdc4, but none of them matched the expected size of 34 kDa; however, the signal at the lowest position could be meaningful. These patterns of expression were similar to strains expressing each of the domains, with either BWP17 or JSCA0021 as a parental strain (Lai WC, unpublished results). Therefore, even though some of the strains expressed domains with unexpected size, they were unique from the negative control of JSCA0022. We concluded that the Tet-on system functions in JSCA0022 and that CaCdc4 might be undergoing undefined modifications.
To determine the function of the assorted CaCdc4 domains, JSCA0022-based strains capable of repressing CaCDC4 and inducing expression of assorted CaCdc4 domains were grown in SD medium with or without Met/Cys and in the presence or absence of Dox. Cells from strains in SD medium without Met/ Cys grew as yeast in the presence or absence of Dox ( Figure 4B). By contrast, cells from strains in medium with Met/Cys grew with filaments ( Figure 4B). As expected, cells of JSCA0023 and JSCA0024 growing on medium with Met/Cys and Dox and that expressed the full-length CaCdc4 with or without tag grew as yeast. Disregarding the full-length CaCdc4, cells from all strains, except JSCA0025 expressing assorted domains, still grew as filaments ( Figure 4B). Under Met/Cys and Dox conditions, cells from JSCA0025 expressing the N-terminal 85-amino acid truncated CaCdc4 seemed to have an ability to suppress filamentation but not complete back to the yeast form ( Figure 4B). This is in consistent with our previous observation in which, comparing with cells capable of expressing the fulllength CaCdc4 under the CaMET3p repressible control, those cells expressing the N-terminal 85-amino acid truncated CaCdc4 lagged behind in reaching exponential stage (Additional file 1: Figure S1) and converted to filamentous form earlier (Additional file 2: Figure S2) in the repressed condition.    Table 2) containing common AatII and XhoI sites for replacing full-length CaCDC4 on pTET25M-CaCDC4-6HF as described in the Methods. By digestion with SacII and KpnI, each cassette was used to transform C. albicans strain JSCA0022 for integration into the CaADH1 locus. (C) Verification of Tet-on cassettes being integrated into CaADH1 locus by Southern blotting analysis. Organization of the CaADH1 locus with respect to SpeI sites is shown in Figure 3B. The relative positions of the probe used and the predicted SpeI-digested pattern of the CaADHI locus are indicated in Figure 3B. One SpeI-fragment of 3.3 kb specific to CaADH1 locus could be detected in genomic DNA from strain JSCA0022 and its derivatives digested with SpeI.

Met/Cys
The correctness of integration of the cassette into the CaADH1 locus of various strains was confirmed by alteration of the SpeI-fragment from size of 3.3 kb to 9.4 kb ( Figure 3B) or various sizes as indicated in Figure 3A.
Ca 2+ to the culture medium being depleted with Ca 2+ beforehand was used [23]. Only cells of JSCA0023 and JSCA0024 remained resistance in flocculation during the time-frame of 5-minute assay compared with those of the rest of strains ( Figure 6), which were consistent with the results shown in Figure 5. However, both strains JSCA0025 and JSCA0027 exhibited greater ability to resist flocculation than that of JCSA 0026 and JSCA0030 when considering the differences in OD 600 from the initial to the end points.

Discussion
In this study, we aimed to dissect the function of CaCdc4 domains by introducing a Tet-on system with cassettes that encoded for a variety of CaCdc4 domains in a C. albicans mutant of Cacdc4 null. However, the Cacdc4 null mutant with a filamentous form could not be easily used to introduce the Tet-on cassettes; therefore, we constructed the JSCA0022 strain, where CaURA3 was released from the strain JSCA0021, and CaCDC4 expression was repressible. Under repressed conditions, the JSCA0022 strain showed similar filamentous morphology (Figure 2) to those from previous reports of cells with CaCDC4 repressed strain [6,7] and of cacdc4 null mutant [6] (Tseng TL, Hsu WH, and Shieh JC, unpublished results). We confirmed that the JSCA0022 strain under repressed conditions was equivalent to a strain that had completely lost CaCDC4 function. Hence, by introduction of the Tet-on cassettes into JCSA0022 strain, each of the strains was capable of expressing individual CaCdc4 domains in the presence of Met/Cys and Dox for functional comparisons.
To verify the ability of the Tet-on cassettes in C. albicans, each of the cassettes encoding various CaCdc4 domains was transformed into BWP17 and JSCA0021 before introducing them into JSCA0022 at the CaADH1 locus. Individual CaCdc4 domains from relevant strains were all detectable, suggesting that the Tet-on system functions in C. albicans. However, while cells expressing the F-box and the WD40 repeat could be detected as their expected sizes, those expressing the full-length CaCdc4, the N-terminus truncated CaCdc4 (ΔN), and the ΔNF of CaCdc4 could be detected at positions higher than anticipated ( Figure 4A). In particular, the sample from strain JSCA0030 expressing the ΔNF could be detected three signals ( Figure 4A), all of which were greater than the predicted sizes. These results suggest that the N-terminal CaCdc4 from residue 85 to 241 ( Figure 3A) might be undergoing post-translational modification during the Tet-on-induced expression,  Table 1.
although its functional significance is unknown. Interestingly, the region between residue 85 and 241 of CaCdc4 contains abundant serine and threonine residues, the majority of which are homologous to S. cerevisiae Cdc4 [7]. This implies possible phosphorylations or other modifications on these residues that is specific to C. albicans. However, the genuine nature of these residues remains to be determined, and their functional significance of this N-terminal CaCdc4 requires further study.
With regards to integration of CaADH1 locus by the Tet-on cassette, it is known that C. albicans adh1 homozygous null mutant gains the ability to form biofilm both in vitro and in vivo [27], suggesting a possible role of CaADH1 in flocculation. However, the heterozygous CaADH1 null mutant with which the homozygous adh1 null mutant is reintegrated a functional copy of CaADH1 to the CaADH1 locus appears to be similar in biofilm formation as its isogenic wild-type strain. In addition, disruption of CaADH1 has no consequence of morphology alteration in C. albicans [27] (Lai WC, unpublished results). Therefore, the possible effect of Tet-on cassette on flocculation and filamentation by integration, hence disruption of a copy of CaADH1 locus can be excluded.
Under the Met/Cys and Dox conditions, cells expressing F-box, WD40 repeat, and the ΔNF of CaCdc4 exhibited filamentous forms similar to those of JSCA0022, whose CaCDC4 was repressed, compared to those expressing the full-length CaCdc4 without or with tag (JSCA0023 and JSCA0024), which exhibited yeast forms Cultures of the indicated strains were grown in SD medium with histidine, arginine, uridine for 2 days before diluting into the SD medium to an initial OD 600 = 0.1 with addition of 2.5 mM Met/Cys to repress the expression of CaMET3p-driven CaCDC4 and 40 μg/ml Dox to induce the expression of CaCdc4 domains tested for 48 hrs to OD 600 ≈ 1.6. Cultures were photographed before and after centrifugation. (A). A representative of the cultures. Upper panel: two-day culture. Bottom panel: cultures being spun down with 500 rpm for 1 minute. (B). Quantitative results. Data are represented as means with standard deviation from three independent experiments, each sample was in duplication. The data from JSCA0022 were compared with those of other strains. **: P < 0.01. The designations of strains are the same as in Table 1.
( Figure 4B). These results suggest that both the WD40 repeat and F-box are essential to suppress the yeastto-filament transition. Cells from strain JSCA0025 expressing the ΔN of CaCdc4, which were grown in the presence of Met/Cys and Dox, were only partially able to reverse filamentous cells to yeast cells, suggesting that the N-terminal 85-amino acid of CaCdc4 plays a role in the yeast-to-filament transition in C. albicans. The role of the N-terminal 85-amino acid of CaCdc4 for growth was observed previously, in which cells expressing N-terminal 85-amino acid truncated CaCdc4 lagged slightly in proliferation during the exponential stage (Additional file 1: Figure S1), and repression of the expression of the N-terminal 85-amino acid truncated CaCdc4 resulted in prominently lagging behind in growth, which was presumably due to the morphological alteration of cells to filaments in advance (Additional file 2: Figure S2) that delays proliferation as compared to those of yeast cells. Since the N-terminal 85-amino acid of CaCdc4 is unique compared to that of the S. cerevisiae Cdc4 [7], our finding reveals a role of N-terminal 85-amino acid of CaCdc4 on morphogenesis, which is unknown previously.
Importantly, cells of all JSCA0022-based strains exhibited flocculation in medium with Met/Cys, but the strains JSCA0023 (CaCDC4) and JSCA0024 (CaCDC4-6HF) exhibited less flocculation by adding Dox simultaneously ( Figure 5). Unlike cells of JSCA0023 and JSCA0024, those of JSCA0025 expressing N-terminal 85-amino acid truncated CaCdc4 were unable to totally overturn filamentous-to-yeast cells, suggesting that N-terminal 85-amino acid is required for full activity of CaCDC4 function in C. albicans to inhibit filamentation. However, if flocculation is tightly associated with filamentation, we expect to see the extent of flocculation in JCSA0025 (ΔN 6HF) being greater than that of JSCA0022 but less than that of JSCA0023 and JSCA0024 in the presence of Met/Cys and Dox. This was not revealed by the low speed-centrifugation method but by the Ca 2+ -initiation assay. Importantly, both JSCA0025 and JSCA0027 expressing CaCdc4 lacking N-terminal 85-amino acid ( Figure 3A) exhibits similar extent of flocculation. Moreover, JSCA0025 that expressing CaCdc4 lacking N-terminal 85-amino acid could only partially suppress filamentation yet JSCA0027 that expressing CaCdc4 lacking N-terminal 85-amino acid and F-box with flanking regions completely lose the ability to inhibit filamentation ( Figure 3A and Figure 4B). These results imply that N-terminal 85-amino acid of CaCdc4 has a role in inhibition of cell flocculation in C. albicans and that the F-box and its flanking region in addition to the N-terminal 85-amino acid of CaCdc4 might be associated with proper control of both morphogenesis and flocculation.

Conclusions
Therefore, we conclude that F-box and WD40-repeat are important in suppressing yeast-to-filament transition and flocculation and that the N-terminal region (1-85) has a positive role in CaCDC4 function, lost of which impairs reverse of filament-to-yeast and reduces the ability to flocculate in C. albicans. Moreover, the function of CaCdc4 for suppressing flocculation that is related to cell-cell adhesion [21] implies a role of CaCDC4 in biofilm formation [28] that is under investigation.

Additional files
Additional file 1: Figure S1. N-terminal 85-amino acid of CaCdc4 is required for normal growth of C. albicans. Strains: BWP17, heterozygous null mutant CaCDC4 +/−, M3CaCDC4 +/− carrying CaMET3-full-length CaCDC4, and M3NTCaCDC4 +/− carrying CaMET3-partial CaCDC4 (capable of expressing N-terminal 85-amino acid of truncated CaCdc4). Cells of the strains were grown initially in SD medium without Met/Cys to saturation and were diluted to the same initial concentration. Cells were grown for 12 hrs in SD either with or without 2.5 mM Met/Cys (−Met/Cys or + Met/Cys) and at each 2-hr interval the cells were sampled to determine the optical density of 595 nm (O.D. 595) in which the growth curves could be plotted.
Additional file 2: Figure S2. N-terminal 85-amino acid of CaCdc4 is required for suppression of yeast-to-filament transition in C. albicans. Cells of the strains were grown initially in SD medium without Met/Cys to saturation and were diluted to the same initial concentration. Cells were grown for 8 hrs in SD either with or without 2.5 mM Met/Cys (−Met/Cys or + Met/Cys). The images were visualized and recorded with a Nikon 50i microscope at 400× magnification. Bars represent 10 μm. The designations of strains are the same as in Additional file 1: Figure S1.  Figure 6 Analysis of Ca 2+ initiated cell flocculation. The strains were grown as described in Figure 5 to saturation at OD 600 ≈ 1.4. The cultures were harvested and washed twice with deflocculation buffer, followed by initiation of flocculation as described in the Methods. The assays were conducted in triplicate. The name of each strain shows only the last two numbers.