CaGdt1 plays a compensatory role for the calcium pump CaPmr1 in the regulation of calcium signaling and cell wall integrity signaling in Candida albicans

Strains and reagents

Construction of C. albicans mutants

We constructed the double-gene deletion mutant by replacing two CaGDT1 alleles in the pmr1/pmr1 mutant in the CAI4 background [10], with the SAT1 flipper cassette [28] and the hisG-URA3-hisG cassette [29], respectively, as described previously [30]. Genotypes of the resultant mutants were confirmed by PCR (data not shown).

DNA manipulation

The 2150-bp CaGDT1 gene containing its promoter, ORF and 3′ terminator was amplified from genomic DNA with primers GDT1 (BamHl) –UP and GDT1 (BamHl) –DOWN, and cloned into the C. albicans expression vector pCR4 [13], which yielded pCR4-CaGDT1. Similarly, we amplified the CaGDT1 gene using the primers GDT1 (Kpnl) –UP and GDT1 (Sphl) –DOWN, and cloned it into the Kpnl and Sphl sites of S. cerevisiae vector pHAC181 [31], which yielded pHAC181-CaGDT1.

Growth test

Strains were cultured overnight in liquid YPD or SD-URA medium at 30 °C. Overnight cultures were serially diluted by 10-fold, and spotted onto appropriate plates. Phenotypes were recorded after plates were incubated for 2–3 days at 30 °C.

⁴⁵Ca²⁺ uptake assays

Cellular accumulation of Ca²⁺ was measured basically as described [32]. Cells were inoculated in YPD medium containing 5 μCi ml^{− 1 45}CaCl₂ with a specific activity of 9.6 × 10⁶ cpm ml^− 1 (PerkinElmer), and grown to log phase at 30 °C in 96-well filtration plates. Cultures were harvested by filtration, washed four times with ice-cold buffer (5 mM Na-HEPES, pH 6.5, 10 mM CaCl₂, 150 mM NaCl), and dried at room temperature. The filters were counted using a TopCount NXT (Packard) liquid scintillation counter.

Assays of β-galactosidase activity

The CaUTR2p-lacZ reporter in the pGP8 plasmid [13] was integrated into the CaUTR2 locus in the wild type CAI4, the gdt1/gdt1 mutant, the pmr1/pmr1 mutant and the gdt1/gdt1pmr1/pmr1 mutant. Three independent transformants of each strain were assayed for β-galactosidase activity as described previously [13].

Analysis of cell survival and cell integrity

Test strains were grown to log phase before the culture was divided into two parts, which were treated with 25 μg ml^− 1 CsA and ethanol (as control), respectively. Two parts were further incubated with shaking for 1 h, 2 h, 4 h and 8 h, respectively, before samples were taken for measuring colony-forming units (clonogenic survival) as well as co-staining with 20 μg ml^− 1 Annexin V-FITC and 50 μg ml^− 1 propidium iodide (CLONETECH Laboratories, USA). Vacuoles in these cells were labelled in vivo with N-(3-triethylammoniumpropyl)-4-(6 [4-(diethylamino) phenyl) hexatrienyl] pyridinium dibromide (FM4–64; Molecular Probes, USA) as described [33].

Protein extract and western blot analysis

Cells were grown to mid-exponential phase, and treated with 50 μg ml^− 1 Calcofluor white for 2 h before they were collected for protein extraction and Western blot analysis as described [7]. The anti-phospho-p44/42 MAPK (Thr²⁰² /Tyr²⁰⁴) antibody (Cell signaling, USA) was used to simultaneously detect phosphorylated forms of both CaMkc1 and CaCek1 [23, 24]. Anti-HA antibody (Abmart, Shanghai) was used to detect CaMkc1-HA and CaCek1-HA proteins [7].

Chromosomal integration of green fluorescent protein (GFP) and HA epitope

GFP tagging cassette was PCR-amplified with CaGDT1-GFP-UP and CaGDT1-GFP-DOWN from the pGFP-URA plasmid [34], using primers CaGDT1-GFP-UP and CaGDT1-GFP-down, and integrated into the heterozygous mutant (WJCA102). Genotype of the CaGDT1-GFP tagged strain with correct integration was confirmed by PCR (Additional file 1: Figure S2B). To examine the co-localization between CaGdt1 and CaPmr1, we first chromosomally tagged HA epitope to the C-terminus of CaPMR1 in the wild-type RM1000 strain, yielding RM1000 PMR1/PMR1::HA-HIS1 strain (WJCA110). The HA-HIS1 cassette was PCR-amplified with primers PMR1-HA-HIS-UP and PMR1-HA-HIS-DOWN from the pFA-HA-HIS1 plasmid [34, 35]. We subsequently tagged the GFP to the C-terminus of CaGDT1 in the WJCA110 strain, which generated the RM1000 PMR1/PMR1::HA-HIS1 GDT1/GDT1:: GFP-URA3 strain (WJCA111). Genotypes of WJCA110 and WJCA111 strains were confirmed by PCR (Additional file 1: Figure S2B).

Two HA-URA3 cassettes were PCR-amplified with primer pairs, MKC1-HA-UP/ MKC1-HA-DOWN and CEK1-HA-UP and CEK1-HA-DOWN, respectively, from the pFA-HA-URA3 plasmid [34], to chromosomally tag the HA epitope to the C-termini of CaMKC1 and CaCEK1 in relevant strains. Their genotypes were confirmed by PCR (Additional file 1: Figure S2C and S2D).

Co-localization of GDT1-GFP and PMR1-HA proteins

Cells of the WJCA111 strain was grown to log-phase in SD-URA-HIS medium before they were treated with 100 mM CaCl₂ or 1 mM EGTA for 2 and 6 h, respectively. Cells were treated with 3.7% formaldehyde for 10 min and incubated in fixation buffer (0.1 M KH2PO4,0.5 M MgCl2, 3.7% formaldehyde, pH 6.5) for 20 min. Cells were then resuspended in 1 ml buffer (0.1 M KH₂P₄, 1.2 M sorbitol, pH 6.5) supplemented with 4 μl of beta-mercaptoethanol and 4 μl of 5 mg ml^− 1 Zymolyase 20 T (Seikagaku Biobusiness), and incubated at 37 °C for 20 min. Spheroplasts were washed twice with 1 ml of PBS buffer, and suspended in 100 μl of PBS supplemented with 0.05% Tween 20 (PBST).

BSA of 400 μg ml^− 1 was added to the spheroplast mixture, and incubated for 20 min before mouse anti-HA antibodies was added at a dilution of 1:500. The mixture was incubated for 2 h, and washed twice with PBST before goat anti-mouse IgG conjugated to Alexa Fluor 555 (Invitrogen, USA) was added. The mixture was incubated in dark for 45 min before spheroplasts were collected, washed three times, and visualized under a Nikon 80i microscope equipped with DS-U2 CCD. The images were acquired at 1000 × and processed using the NIS – ELEMENTS F3.0 software.

Co-localization of GDT1-GFP protein and Golgi-tracker red dye

For Golgi apparatus labeling, the pmr1/pmr1 GDT1-GFP/gdt1 cells were grown to log-phase in SD-URA, harvested, washed twice with PBS, and suspended in 200 μl PBS supplemented with 4 μl beta-mercaptoethanol and 15 U lyticase enzyme (Sigma). Cells was incubated at 37 °C for 40 min to partially digest the cell wall, and were washed once with PBS and suspended in 200 μl PBS before they were mixed with 2 μl Golgi-Tracker Red dye (33.3 mg ml^− 1) (Beyotime Institute of Biochemistry, China) and incubated at 4 °C for 30 min. Cells were then washed twice and incubated in 1 ml SD-URA at 30 °C for 30 min, before they were visualized by the Nikon 80i fluorescent microscope.

RNA extraction and qRT-PCR assays

Total RNA was extracted as described previously (12). Contaminated genomic DNA in the reverse transcription products was examined by PCR of 35 cycles using a pair of primers CaACT1-g-UP and CaACT1-g/m-DOWN flanking the intron of the C. albicans ACT1 gene. Real-time qPCR was performed in the CFX96™ Real-Time System I (BIO-RAD) using SYBR Premix EX Taq™ II (TAKARA Biotechnology, China). All assays were performed in duplicate. For quantification, the abundance of each gene was determined relative to the transcript of CaACT1.

Transcript profiling and data analysis

Total RNA samples were extracted from the wild type CAI4, the gdt1/gdt1, the pmr1/pmr1, and the gdt1/gdt1 pmr1/pmr1 mutants grown to log phase in SD-URA at 30 °C. RNA integrity was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, USA). RNA-seq libraries were constructed using Illumina’s TruSeq RNA Sample Preparation Kit (Illumina Inc., USA). RNA sequencing, data analysis and sequence assembly were performed by Beijing BioMarker Technologies (Beijing, China). Preparation of the paired-end libraries and sequencing were performed following standard Illumina methods and protocols. The mRNA-seq library was sequenced on the Illumina Hiseq 2500 and Illumina Genome Analyzer system.

To obtain high-quality clean read data for de novo assembly, raw reads from mRNA sequencing were filtered by discarding the reads with adaptor contamination, masking low-quality reads with ambiguous ‘N’ bases and removing reads in which more than 10% bases had a Q-value < 20. Clean reads were assembled into full-length transcriptome from RNA-Seq data with the reference genome (http://www.candidagenome.org/).

Digital analysis of gene expression

Gene expression levels were measured as numbers of reads, and normalized with RPKM in RNA-Seq analyses as described [36]. Differentially expressed genes were identified in pair-wise comparison with the IDEG6 software [37], and all statistical test results were corrected for multiple testing with the Benjamini–Hochberg false discovery rate (FDR<0.01). Gene transcripts were considered to be significantly differentially expressed if the adjusted P value was <0.001 and there was at least a two-fold change (>1 or <-1 in log 2 ratio value) in PKM between two libraries.

Semi-quantitation of gene expression levels by RT-PCR

To confirm the RNA-sequencing data, we carried out semi-quantitative RT-PCR with reverse transcription products that were treated to remove their contaminated genomic DNA. RT-PCR was carried out as described previously [12]. PCR products were separated on agarose gel and stained with ethidium bromide. DNA contents in all lanes of the gel were quantified and analyzed using the Bio-Rad Gel DocTM XR+ System and the Image Lab software (Version 4.0.1). Relative levels of target gene transcripts in each strain were calculated using CaPGK1 as the internal control.

Statistical analysis

Significant differences were analysed by GraphPad Prism version 4.00. P values of < 0.05 were considered to be significant.

CaGDT1 complements the function of ScGDT1 in cell wall stress

As reported previously [30], expression of CaGDT1 in the S. cerevisiae vector, pHAC181, fully suppressed the sensitivity of the Scgdt1 single-gene deletion mutant to 0.6 M and 0.7 M CaCl₂ as well as the sensitivity of the Scgdt1/Scpmr1 double-gene deletion mutant to 0.4 M CaCl₂ (Additional file 1: Figure S1A). In addition, expression of CaGDT1 reverted the sensitivity level of the Scgdt1/Scpmr1 double-gene deletion mutant to that of the Scpmr1 single-gene deletion mutant under higher concentrations of CaCl₂ (0.6 M and 0.7 M) (Additional file 1: Figure S1A). Therefore, CaGDT1 is a functional homolog of ScGDT1 in calcium stress [30]. Although S. cerevisiae cells lacking ScGDT1 were not sensitive to 5 mM EGTA, 0.4 M CaCl₂, 100 μg ml^− 1 Congo red (CR) and 50 μg ml^− 1 Calcofluor white (CFW) [7], we found that deletion of ScGDT1 increased the sensitivity of cells lacking ScPMR1 to all these reagents (Additional file 1: Figure S1A). Expression of CaGDT1 reverted the sensitivity level of the Scgdt1/Scpmr1 double-gene deletion mutant to that of the Scpmr1 single-gene deletion mutant under these cell-wall stress conditions (Additional file 1: Figure S1A). Taken together, these data demonstrate that CaGdt1 complements the function of ScGdt1 in the response of S. cerevisiae cells to cell-wall stress.

CaGDT1 shows genetic interactions with CaPMR1 in the response to cell wall and ER stress

Since mutations of human TMEM165, the CaGdt1 homolog, affect the glycosylation pathway [25, 38], we tested cells lacking CaGDT1 against cell wall and ER stresses. Surprisingly, C. albicans cells lacking CaGDT1 did not show significant alteration in their sensitivity to CR, CFW, caffeine and tunicamycin (Fig. 1a).

Unlike ScGDT1, C. albicans cells lacking CaGDT1 were not sensitive to 0.7 M CaCl₂ (Fig. 1a; Additional file 1: Figure S1A; [25]). However, ScGDT1 genetically interacts with ScPMR1 in the calcium sensitivity of S. cerevisiae cells [25]. We next investigated possible interactions between CaGDT1 and CaPMR1 in cell wall stress. As reported previously [10], cells lacking CaPMR1 were hypersensitive to CR, CFW, caffeine and tunicamycin (Fig. 1a and b). Further deletion of CaGDT1 increased the sensitivity of cells lacking CaPMR1 to CR, CFW, caffeine and tunicamycin, which could be reversed by the introduction of the CaGDT1 gene (Fig. 1a and b). These results demonstrate that CaGDT1 has an additive effect on CaPMR1 in the response of C. albicans cells to cell wall and ER stresses.

Calcineurin function is required for C. albicans cells in response to cell wall stress

Further deletion of CaGDT1 led cells lacking CaPMR1 hypersensitive to cyclosporine A (CsA) and FK506, specific inhibitors of calcineurin, in YPD medium with or without cell wall stress (Fig. 1b; Additional file 1: Figure S1B). This indicates that normal growth of C. albicans cells lacking both CaGDT1 and CaPMR1 requires the calcineurin function.

The homozygous mutant for CaGDT1 became hypersensitive to CR and CFW in the presence of 50 μg ml^− 1 CsA, and its sensitive degree was comparable to that of C. albicans cells lacking both CaGDT1 and CaPMR1 on YPD plates containing these reagents without CsA (Compare Fig. 1b to a). Furthermore, addition of CsA in YPD plates made C. albicans cells more sensitive to CR and CFW independent of CaGDT1 and CaPMR1 (Compare Fig. 1b to a). This suggests that calcineurin function is additive to the overlapping function between CaGdt1 and CaPmr1 in the response of C. albicans cells to cell wall stress.

Inhibition of calcineurin promotes cell death and disrupts cell wall and vacuolar integrity of cells lacking both CaGDT1 and CaPMR1

To further investigate the growth defect of C. albicans cells lacking CaGDT1 and CaPMR1 in the presence of CsA, we determined their clonogenic survival ability by measuring their colony-forming units. In response to 25 μg ml^− 1 CsA in YPD medium, cells of the gdt1/gdt1 pmr1/pmr1 mutant lost their viability in both time-dependent and dose-dependent fashions (Fig. 2a and b). However, CsA treatment of the wild type did not affect its viability (Data not shown). This indicates that CsA promotes the death of cells lacking both CaGDT1 and CaPMR1.

Compared to untreated cells (Fig. 2c), most of cells treated with 25 μg ml^− 1 CsA for 8 h were abnormal in their internal structures (Fig. 2d), and appeared to be flakes in culture solution (Fig. 2e). To examine their cell wall integrity, we co-stained these cells with Annexin V-FITC (green fluorescence) and propidium iodide (red fluorescence), which measure externalization of phosphatidylserine and loss of membrane integrity of yeast cells, two indicators of cell death [39, 40]. Only a few of the wild-type and the mutant cells untreated with CsA were stained (Fig. 2f). Comparing to the CsA-treated wild-type with only a few of stained cells, most of mutant cells treated with CsA were stained, by these two dyes (Fig. 2g), indicating susceptible cell walls of CsA-treated mutant cells for an easy entry of these dyes.

To understand the nature of the abnormal intracellular compartments in CsA-treated cells lacking both CaGDT1 and CaPMR1 (Fig. 2d), we examined the integrity of their vacuoles with the dye FM4–64 [33]. Untreated wild-type and mutant cells showed a uniform staining pattern of vacuoles, indicating normal endocytosis and enrichment of FM4–64 in the vacuolar membrane (Fig. 2h). Wild-type cells treated with CsA showed a uniform staining pattern, while mutant cells treated with CsA exhibited various staining patterns (Fig. 2i), including cells with normal intracellular compartments and no intracellular staining, indicating their lost ability of transporting FM4–64 through endocytosis, and cells with abnormal intracellular compartments and strong whole-cell staining, indicating their disrupted vacuolar structures. Taken together, these results suggest that calcineurin function is essential for the cell survival as well as the cell wall and vacuolar integrity of cells lacking both CaGDT1 and CaPMR1.

Calcium uptake and activation of the calcium/calcineurin signaling in C. albicans cells lacking CaGDT1 and CaPMR1

Although deletion of CaGDT1 did not alter the sensitivity of C. albicans cells to calcium stress, we showed that CaGDT1 complemented the function of ScGDT1 in calcium stress ([30]; Additional file 1: Figure S1A). Both ScGdt1 and its human ortholog TMEM165 are involved in the regulation of cytosolic Ca²⁺ and pH homeostasis [7]. We next measured the uptake of ⁴⁵Ca²⁺ in mutants for CaGDT1 and/or CaPMR1. As compared to the wild type, both the gdt1/gdt1 and the pmr1/pmr1 mutants exhibited similar high levels of ⁴⁵Ca²⁺ uptake in YPD medium, while their double-gene mutant showed an even higher level of ⁴⁵Ca² uptake (Fig. 3a). Increased calcium uptake was also observed for cell lacking ScPMR1 [41]. These results suggest that CaGdt1 and CaPmr1 have additive functions in the regulation of calcium homeostasis. This is consistent with observations for ScGdt1 and ScPmr1 in S. cerevisiae [7, 42].

We next examined activation levels of the calcium/calcineurin signaling in these mutants with the lacZ reporter of CaUTR2, a downstream target of CaCrz1 [43]. There was no significant difference in the lacZ activity between the wild type and each of three mutants in YPD medium (Fig. 3b). In the presence of 0.2 M CaCl₂, both the pmr1/pmr1 and the double-gene mutants showed higher lacZ activities than, but the gdt1/gdt1 mutant showed a similar lacZ activity to, the wild type (Fig. 3b). These CaUTR2p::lacZ activity results agree with our transcriptomic profiling data on the transcript levels of CaUTR2 (CAWG_02505) in these mutants and the wild type (SuppInfo 1 and 2 [GEO Accession number: GSE100737]). Taken together, these data indicate that CaPmr1 is the major controller of calcium homeostasis in C. albicans, which in consistent with observations on ScPmr1 in S. cerevisiae [7, 42].

CaGdt1 plays a role in the CaCek1-mediated, but not CaMkc1-mediated, cell wall integrity signaling

Deletion of CaPMR1 activates the Mkc1-mediated CWI signaling [10]. Since deletion of CaGDT1 increased the sensitivity of cells lacking CaPMR1 to cell wall stresses (Fig. 1a), we examined the effect of CaGDT1 deletion on CWI signaling. We chromosomally tagged the HA epitope to the C-terminus of the CaPKC1 allele and the CaMKC1 allele, respectively, in the wild type CAI4, the WJCAG22 (gdt1/gdt1), the NGY98 (pmr1/pmr1) and the WJCAG25 (gdt1/gdt1 pmr1/pmr1) strains (Additional file 1: Figure S2A, 2C and 2D). In YPD medium, phosphorylation levels of the wild type CaMkc1 (or the HA-tagged CaMkc1-HA fusion) were similar between the wild type and the gdt1/gdt1 strains in the absence or presence of 50 μg ml^− 1 CFW (Fig. 4a). As reported previously [10], phosphorylation levels of CaMkc1 (or CaMkc1-HA) increased with similar degrees in both the pmr1/pmr1 and the gdt1/gdt1 pmr1/pmr1 mutants as compared to the wild type (Fig. 4a). Similar patterns for CaMkc1 phosphorylation levels were also observed in C-terminally HA tagged CaCEK1 strains (Fig. 4b). These results indicate that CaPmr1 does, but CaGdt1 does not, play a role in the CaMkc1-mediated CWI signaling -.

The CaCek1-mediated CWI signaling is also involved in cell wall construction [18]. Phosphorylated CaCek1 and CaCek1-HA) were not detected in both the wild type and the gdt1/gdt1 strains, while the gdt1/gdt1 pmr1/pmr1 mutant showed a higher level of phosphorylated CaCek1 and CaCek1-HA than the pmr1/pmr1 mutant in the absence or presence of CFW (Fig. 4b). Similarly, in C-terminally HA tagged CaMKC1 strains, the gdt1/gdt1 pmr1/pmr1 mutant showed a higher level of phosphorylated CaCek1 than the pmr1/pmr1 mutant in the absence or presence of CFW, while phosphorylated CaCek1 was detected at very low levels in both the wild type and the gdt1/gdt1 strains, with a slightly higher level in the latter (Fig. 4a). Taken together, these data suggest that CaGdt1 and CaPmr1 are both involved in the activation of CaCek1-mediated CWI signaling independent of cell wall stress.

CaGdt1 is localized in the Golgi apparatus but at distinct sites from CaPmr1

To determine the subcellular localization of CaGdt1, we chromosomally tagged the green florescent protein (GFP) to the C-terminus of CaGdt1 in the heterozygous mutant for CaGDT1 generating WJCA102 (CAI4 pmr1pmr1 GDT1-GFP/gdt1) (Additional file 1: Figure S2). Like the pmr1/pmr1 mutant, the WJCA102 mutant was not sensitive to CsA, while the pmr1/pmr1 gdt1/gdt1 mutant was (Fig. 5a). This indicates that the CaGDT1-GFP allele is functional. In log-phase growing cells of WJCA102, CaGDT1-GFP was colocalized with the Golgi Tracker Red dye (Fig. 5b).

Similarly, we constructed the WJCA111 (RM1000 PMR1/PMR1::HA-HIS1 GDT1/GDT1::GFP-URA3) strain to examine the colocalization between CaGdt1 and CaPmr1 (Data not shown). Through indirect immunofluorescent approach, we found that the CaGDT1-GFP protein was not colocalized with the CaPMR1-HA in log-phase growing WJCA111 cells (Fig. 5c). It should be noted that there was a discrepancy in CaGDT1-GFP localization between Fig. 5b and c, which could be due to the difference in sample processing for florescent microscopy as described in the Methods section. In response to depletion of extracellular calcium (1 mM EGTA treatment) or high levels of extracellular calcium, the CaGDT1-GFP protein was also not colocalized with the CaPMR1-HA in WJCA111 cells (Additional file 1: Figure S3). Taken together, these results indicate that CaGdt1 is present in the Golgi apparatus but at distinct sites from CaPmr1 in C. albicans.

Transcriptomic profiling of cells lacking CaGDT1 and/or CaPMR1

To help us further understand functions of CaGdt1 and CaPmr1, we carried out transcript profiling for the wild type CAI4, the gdt1/gdt1, the pmr1/pmr1, and the gdt1/gdt1 pmr1/pmr1 mutants, growing in log phase in YPD medium at 30 °C (GEO accession number: GSE100737). Transcripts of 5901 genes at various expression levels were detected in each of these four strains (SuppInfo 1 [GEO Accession number: GSE100737]). As compared to the wild type, there are 249, 446 and 440 differentially expressed genes (DEGs) in the gdt1/gdt1, the pmr1/pmr1 and the gdt1/gdt1 pmr1/pmr1 mutants, respectively (Fig. 3c; SuppInfo 2). Deletion of CaGDT1 does not affect the transcript level of CaPMR1, and vice versa (SuppInfo 2).

There are 105 shared DEGs between the gdt1/gdt1 and the pmr1/pmr1 mutants, which accounts for 42% (105/249) of total DEGs due to deletion of CaGDT1 or 24% (105/446) of total DEGs due to deletion of CaPMR1 (Fig. 3c). This indicates that CaGdt1 and CaPmr1 have shared and distinct functions. Furthermore, there are 76 shared DEGs among the gdt1/gdt1, the pmr1/pmr1 and the gdt1/gdt1 pmr1/pmr1 mutants, with 23 of them being up-regulated and 53 of them being down-regulated (Fig. 3c; Table 3). Using a semi-quantitative RT-PCR approach, we confirmed the transcript levels of selected 15 out of these 76 genes in the wild type, the gdt1/gdt1, the pmr1/pmr1 and the gdt1/gdt1 pmr1/pmr1 mutants (Additional file 1: Figure S4). The 23 up-regulated DEGs are involved in the cellular transport, including the transporters of long-chain fatty acids (PXA1 and PXA2), the calcium channel (CCH1) and the potassium transporter (TRK1) (Table 3). Similarly, a big portion of the 53 down-regulated DEGs are also involved in the cellular transport, with different substrates including amino acids (MUP1, QDR1, ALP1, GNP1, TPO3 and TPO4) [44], lipids (PDR16), nucleosides (CNT) [45], calcium ions (RCH1) [13] and sodium ions (ENA2). In addition, 10 of these down-regulated DEGs are involved in the protein synthesis (Table 3). These data suggest that core functions shared by CaGdt1 and CaPmr1 are involved in the regulation of cellular transport of metal ions and amino acids.

CaGdt1 homologs in humans and S. cerevisiae have been shown to regulate glycosylation pathway [25, 27]. In S. cerevisiae, there are two main types of glycosylation, the N-linked and the O-linked glycosylation (http://www.kegg.jp/kegg/pathway.html). The N-linked glycosylation is the attachment of the oligosaccharide known as glycan to the amide nitrogen of an asparagine residue of a protein, There are 30 genes involved in the N-linked glycosylation in S. cerevisiae, with 20 of them having homologs in C. albicans (Additional file 1: Table S2). Transcript levels of CAWG_02847 (DIE2) and CAWG_03939 (STT3) are up-regulated only in the gdt1/gdt1 pmr1/pmr1 mutant, while that of CAWG_04868 (OST3) is up-regulated in both the pmr1/pmr1 mutant and the gdt1/gdt1 pmr1/pmr1 mutant, but not in the gdt1/gdt1 mutant (Additional file 1: Table S2). The O-linked glycosylation is the attachment of a sugar molecule (such as N-acetyl-galactosamine, N-acetylglucosamine and mannose) to an oxygen atom in serine or threonine residue of a protein. There are 13 genes involved in the O-linked glycosylation n S. cerevisiae, with 7 of them showing homologs in C. albicans (Additional file 1: Table S3). Expression of CAWG_05623 (PMT1) and CAWG_04353 (PMT4) is up-regulated only in the gdt1/gdt1 pmr1/pmr1 mutant (Additional file 1: Table S3). Furthermore, from the C. albicans genome database we identified 20 genes involved in the CWI pathway based on their homologs in S. cerevisiae (Additional file 1: Table S4) [46]. Only one gene CAWG_01201 (SWI4) was transcriptionally upregulated in each of three mutants, while expression of the gene CAWG_05506 (BEM2) was down regulated in the pmr1/pmr1 mutant, but not affected in both the gdt1/gdt1 mutant and the gdt1/gdt1 pmr1/pmr1 mutant (Additional file 1: Table S4).