The CHK1 gene deletion mutant CHK21 (chk1Δ/chk1Δ) and reconstituted strain CHK23 (CHK1/chk1Δ) were previously constructed in Dr. Calderone’s group (Table 1) in the Department of Microbiology/Immunology, Georgetown University (Calera and Calderone, 1999c). Disruption of the CHK1 gene domain was performed according to the URA-blaster protocol using the lithium acetate transformation method (Fonzi and Irwin, 1993; Calera and Calderone, 1999c), and the genotypes of each strain are shown in Table 1. Strain C. albicans SC5314 as wild type (WT) and together with CHK23 (CHK1/chk1Δ) was used as controls for all experiments.

Construction of CHK25 (^ΔS_TkcCHK1/Δchk1) was accomplished by the transformation of plasmid pChk2 into CHK22 (chk1Δ/chk1Δ; ura3Δ/ura3Δ) (Table 1). Construction of pChk2 was accomplished by ligation of two DNA fragments excised from pChk1R and cloning into a SpeI/BamHI-digested pBSIIsk⁺ vector. The template plasmid pChk1R containing entire 2,470 amino acids of CHK1 open reading frame (ORF) and 870 bp promoter on 5′ end and 24 bp after TAA code. The KpnI/SalI fragment of pMB7 containing his1-URA3-his1 cassette (3.9 kb) was then inserted into KpnI/Xhol-linearized pChk2 at 3′ flanking side of CHK1. As shown in Figure 1, two fragments of pChk1R were individually excised with SpeI/NcoI (promoter plus first 371 aa) and NcoI/BglII (after 1,083 aa) endonucleases and nicely fused together without frameshift, which resulted in the deletion of the Ser/Thr domain (at a position of 387 aa–634 aa). The recombination in CHK22 was allowed by ∼750 bp upstream Hid III site at the 5′ end and 91 bp at 3′ end that has been induced within the CHK1 gene deletion cassette in a previous study.

A similar strategy was used for the construction of CHK26 (^ΔS_TkcΔgafCHK1/Δchk1) with KpnI/SpeI-linearized pChk3 plasmid transformation. The pChk3 was generated by a fusion of the same SpeI/NcoI DNA fragment (1/1,979) and PciI/BglII fragment of pChk1R (6,882/8,307) with the SpeI/BamHI sites of the pBSIIsk + vector. This pChk3 contains 1,115 bp CHK1 promoter, the 371 amino acids upstream of s/t MAPK domain, and amino acids sequence after amino acid at 2,004 position, resulting in deletion of the Ser/Thr domain and GAF domain (6,603 aa–7,060 aa). The same his1-URA3-his1 cassette (3.9 kb) was then subcloned into KpnI/Xhol sites of the pChk3 at 3′ flanking side of CHK1 and allowed to transform in CHK22.

The 5 strains (Table 1), namely, SC5314(WT), CHK21 (chk1Δ/chk1Δ), CHK23 (CHK1/chk1Δ), CHK25 (^ΔS_TkcCHK1/Δchk1), and CHK26 (^ΔS_TkcΔgafCHK1/Δchk1), were used to evaluate the fungal growth in vitro. First, all the 5 strains were grown in yeast extract peptone dextrose medium (YPD) medium [1% (wt/vol) yeast extract, 2% (wt/vol) peptone, 2% (wt/vol) dextrose] for 200 rpm overnight at 30°C and then harvested. Then, cell suspension was diluted into a 20 ml YPD medium at an initiation concentration of 1 × 10⁶/ml and cultured at 30°C (Zhang et al., 2018). The fungal growth in each 1 ml aliquot was measured by a spectrophotometer (OD₆₀₀) at 20, 40, 60, 80, 100, 120, 240, 480, and 960 min.

First, we used Spider medium containing 10% fetal bovine serum at 37°C for 3 days in order to restore the ability of hyphal formation (Bedell and Soll, 1979). Second, the abovementioned strains were suspended into 5 ml of YPD broth and incubated overnight at 30°C with shaking at 200 rpm. After washing three times with phosphate-buffered saline (PBS), the collected yeast cells (1 × 10⁶/ml) were inoculated into 2 ml of Spider-serum medium at 37°C with shaking at 200 rpm. Third, 5 μl cell suspension was collected at 20, 40, 60, 80, 100, 120, 240, 480, and 960 min, and then mixed with 5 μl of calcofluor white (CFW) fluorescent dye to observe under fluorescent microscopy; randomly 10 fields were chosen to calculate the percentage of germ tubes or hyphal cells (germ tubes number/total cell number × 100%). The positive germ tube is defined by three times or greater length of hyphal formation than yeast cells, and the experiment was repeated 3 times.

We use a murine model of oral mucosal candidiasis under immunocompromised and immunocompetent conditions to assess fungal invasiveness. We follow an earlier report (Samaranayake and Samaranayake, 2001) but with some modifications. For the immunocompromised model, 6-week-old female C57BL/6 mice were injected at days 2 and 1 with cyclophosphamide (100 mg/kg) intraperitoneally to simulate an immunodeficiency state. At day 2 after the last injection, the mice were intraperitoneally injected with 10% chloral hydrate at 10 ml/kg to induce a coma status for about 90 min. From a total of 150 mice, 25 mice were used for each testing group as follows: SC5314 (WT), CHK21 (chk1Δ/chk1Δ), CHK23 (CHK1/chk1Δ), CHK25 (^ΔS_TkcCHK1/Δchk1), and CHK26 (^ΔS_TkcΔgafCHK1/Δchk1) infection group and PBS control group. Each mouse was infected on the tongue with a cotton swab bearing 5 × 10⁷clonal formation unit (CFU) fungal cells. Five mice from each testing group were sacrificed at days 1–5 post-infection (dpi), and infected tongues were removed for evaluation. After photography, each infected tongue was longitudinally divided into two parts. Half of the tongue was weighed and then homogenized in 1 ml of aseptic PBS, and an aspirate of 50 μl suspension was then inoculated on Sabouraud dextrose agar (SDA) medium containing chloramphenicol for 48 h at 30°C to evaluate fungal load in tongue tissues. The remaining portion of each tongue was preserved for histopathological examination as below.

Mouse tongue tissues were fixed immediately with 10% neutral formaldehyde fixative. The fixed tissues were dehydrated and embedded in paraffin to make 0.4-μm-thick paraffin sections, which were stained with periodic acid-Schiff staining (PAS) (Samaranayake and Samaranayake, 2001). The fungal biofilm thickness formed by different Candida species was estimated by measuring the deepest invasive distance from the mucosal surface. There were 12 measurements in PAS staining slides from each mouse, and the average value was taken.

The SPSS version 23.0 software was used for statistical analysis. Differences between groups were tested and compared using the independent sample t-test and Tukey’s test in one-way analysis of variance (ANOVA). The value of p < 0.05 is considered statistically significant, and the symbol of “^***” denotes for p < 0.001, “^**” for p < 0.01, and “*” for p < 0.05 for all the graphs in this study.

The structure and domain architecture of the CHK1 gene are shown in Figure 1, in which the cassettes pChk2 (Figure 1B) used for CHK25 (^ΔS_TkcCHK1/Δchk1) construction and pChk3 (Figure 1C) for CHK26 (^ΔS_TkcΔgafCHK1/Δchk1) construction, and the upstream and downstream primers for Southern blot and PCR verification are all included. The genomic DNA of the mutant strain was extracted with a fungal genome extraction kit (Omega Fungal DNA kit). In the Southern blot analysis, a radiolabeling of an [ɑ-32P] dCTP probe of 915 bp obtained with primer set of JA13-JA6 (Table 2) confirms the presence of transformants of CHK25 and CHK26 with BglII-digested genomic DNA in the final product, as shown in Figure 1D. The primers for PCR confirmation for pChk2 and pChk3, and CHK25 and CHK26 are also listed in Table 2. The size of the amplified products of ∼300 bp verifies the correct construction of CHK25 and CHK26 (Figure 1E), and negative control for PCR confirmation of CHK25 and CHK26 was performed by primer set of JA8 and JA5.

The proliferation cycle of microbes including C. albicans can be divided into three phases, namely, the lag phase, the exponential phase, and the stationary phase. To understand the biological function of S_Tkc and GAF domains in Chk1p, the growth curves of CHK25 (^ΔS_TkcCHK1/Δchk1) and CHK26 (^ΔS_TkcΔgafCHK1/Δchk1) were plotted against CHK21 (Δchk1/Δchk1), WT, and reconstituted strain CHK23 (CHK1/Δchk1) based on OD₆₀₀ absorbencies. Growth curve in nutrient-rich YPD broth at the lag phase (0–100 min) is shown in Figure 2A, and the exponential phase and the beginning of the stationary phase (100–960 min) are shown in Figure 2B. At the lag phase, we see a quick jump of OD₆₀₀ absorbance from 0.03 to 0.15 within the first 20 min, and slight drops at 40 and 100 min are typically shown in WT and CHK23. The growth pattern shown in WT or CHK23 is disturbed in domain mutants CHK25 and CHK26 and null mutant CHK21 (Figure 2A). All three mutants retain OD₆₀₀ absorbances around 0.05 throughout the lag phase, which is significantly different from SC5314 and CHK23 (p < 0.001). In the beginning of the exponential phase, the growth defects of the three mutants seem to be less apparent. As shown in Figure 2B, the growth patterns among all five strains are more similar to each other between 100 and 960 min than between 0 and 100 min. However, when compared with WT and CHK23, the growth defects still remain in CHK21 or CHK25 (p < 0.001) and are less affected in CHK26 (p < 0.05). These data suggest that both S_Tkc and GAF domains likely mediate the replication of C. albicans, especially during the initiation phase of growth.

We followed the germ tube and hyphal formation in vitro for 960 min to monitor the impact of the CHK1 gene on hyphal growth. As shown in Figures 3A–E, the germ tubes formed at 80 min are fewer in number, shorter, and only weakly staining by CFW in CHK21 (chk1Δ/chk1Δ), CHK25 (^ΔS_TkcCHK1/Δchk1), and CHK26 (^ΔS_TkcΔgafCHK1/Δchk1) cultures. This weak staining on the cell wall of CHK25 (Figure 3D) is also observed in CHK21 (Figure 3B), and can be due to the cell wall defect. In our earlier study, the large and intermediate-sized mannoproteins are absent in null mutant CHK21 (Li et al., 2009). The germ tube formation rates within the first 100 min remain low for all strains with a slight peak at 80 min. The rate is ∼30% in WT and CHK23 (CHK1/chk1Δ) and immediately drops at 100 min to less than 20% for every strain in Figure 3F. There is no difference in germ tube formation rates between WT and CHK23 at any time (p > 0.05). However, pairwise comparisons show that the rates of germ tube formation within 960 min are universally decreased in the full-gene mutant CHK21 and in the domain mutants CHK25 and CHK26. When compared with WT and CHK23, the decrease in the hyphal formation rate in CHK21 and CHK25 is more obvious (p < 0.001) than CHK26 (p < 0.01) (Figure 3G).

The tongues of five mice from each group were excised at 1–5 days post-infection (dpi 1–5) for evaluation (Samaranayake and Samaranayake, 2001). As shown for dpi 2 (D2 in Figure 4A), the infected areas were swollen or often covered with white pseudomembrane materials in the immunocompromised mice. When measuring the weight change during 5 days of infection, the average body weight of mice infected by WT and CHK23 significantly declines (Figure 4B) when compared with the cyclophosphamide (CTX)-treated mice group (p < 0.01). Weight loss is also seen in the CHK26 (^ΔS_TkcΔgafCHK1/Δchk1)-infected group (p < 0.05), but no weight loss is seen in mice infected by CHK21 (chk1Δ/chk1Δ) or CHK25 (^ΔS_TkcCHK1/Δchk1) versus the CTX control group.

To better quantify fungal load from day 1 to day 5 post-infection and to estimate the extent of hyphal invasiveness in each infected tongue, ground-up tissues were cultured on SDA for fungal recovery (Figures 4C,D). For WT, the CFU value presents highest at day 1 in the tongue tissue and remains high at day 2, but significantly drops after day 3. At day 5, a small number of fungal cells still survive. The CFU values in CHK23 are similar to WT during the 5 days of infection (Figure 4C). For CHK21, the CFU is about eightfold less than WT at dpi 1, and no fungal cells are seen after dpi 2 (Figure 4C). The CFU value in CHK25 infection is only one-third of WT value at dpi 1, which quickly declines at dpi 2 and dpi 3, and is barely detectable through dpi 4–5. While the overall CFU values are significantly reduced in CHK 21- and CHK 25-infected mice from dpi 1 to 5 (p < 0.001), the CFU counts in CHK26-infected mice in dpi 2 and 3 are similar to the CHK23 group but less abundant at dpi 1 and diminished at dpi 5 (Figure 4D). Nevertheless, fungal persistence is shorter for the null mutant or domain mutants in mouse tongues than for WT and reconstituted strain CHK23. Since domain mutants are easily shed from mucosa, our data suggest that both S_Tkc and GAF domains in CHK1 are required for maintenance of C. albicans viability in mucosal tissue.

All our gross findings were not obvious in immunocompetent mice infected with any strain. Instead, slight swelling was only seen in WT- or CHK23-infected mouse tongues with a few fungal cells recovery in SDA counts (data not shown).

The PAS staining was used to examine and compare the mucosal damage and the fungal invasion between control strain and each mutant-infected mouse group from dpi 1 to 5. Since CFU counts at dpi 3 seem to be a turning point for fungal clearance in mouse tissues (Figure 4C), we present pathological changes at dpi 3 in Figure 5 and dpi 1–5 listed in Supplementary Figure 1. Under low-power microscopy, PAS staining shows that the epithelial layer of mucosa in WT- and CHK23 (CHK1/chk1Δ)-infected tongues was more degraded than in any mutant-infected mouse group from dpi 1 to 5 (Figure 5 and Supplementary Figure 1), as indicated by the amorphous structure of the epithelia and the absence of the basal membrane in CHK23 and the missing keratin layer on the surface until dpi 5. The immune cell infiltrate is more often found in WT- and CHK23-infected epithelium (Supplementary Figure 1). In contrast, except for a slightly greater swelling in CHK25 (^ΔS_TkcCHK1/Δchk1)-infected tongues than in CHK21 (chk1Δ/chk1Δ)-infected tongues, the mucosal barrier structures, such as keratin layer, remain intact in both CHK21-infected tongue, and better condition remains in CHK25-infected tongues (Figure 5). As with infections observed in CHK26 (^ΔS_TkcΔgafCHK1/Δchk1), intermediate mucosal damages are found from dpi 1 to 5, and microabscess is more commonly in CHK26 than CHK25 in Supplementary Figure 1.

Under PAS staining, in agreement with the significant amount of swelling and pseudomembrane materials shown in WT- and CHK23-infected mice (Figure 4A), hyphal invasion in vivo is more severe on the tongues infected by WT and CHK23 than any time points by other three mutants (Figure 5 and Supplementary Figure 1). Massive hyphae penetrate through the epithelial, connective tissue, or even muscular layers in WT- or CHK23-infected mouse tongues at dpi 1 (Supplementary Figure 1). In contrast, no hyphal invasion is seen in CHK21- or CHK25-infected tongues, and the basement membranes are intact (Supplementary Figure 1). Most fungal cells of CHK21 and CHK25 are restricted to the keratin layer in yeast form at dpi 1. The hyphal form is detected at dpi 2 for CHK25-infected tongues, but not at any time point for CHK21-infected tongues. Apparently, the additional deletion of GAF domain in CHK26 leads to the partial recovery of hyphal formation function that shows in CHK26-infected tongues. We find a mixture of yeast and hyphae in CHK26-infected tongues at dpi 1 and then hyphal form of CHK26 dominates from dpi 2 to 4. At dpi 5, a few yeasts of CHK26 are restricted to the mucus layer in CHK26-infected tongues (Supplementary Figure 1). When compared with the massively long hyphae shown in WT- or CHK23-infected tongues, hyphal structures seem to be shorter and mixed with more yeast cells in CHK25 and CHK26. Furthermore, the basement membrane zone of the tongue tissue in the CHK26-infected area did not appear to be interrupted from dpi 1 to 5 (Figure 5 and Supplementary Figure 1).

The extent of fungal invasiveness in WT, CHK23, and CHK26 was then quantified by measurement of invasive depth in PAS staining slides. Twelve sites were chosen in each PAS staining slide at random for the measurement of distance from the mucus layer interface or (if the mucus layer is no longer present) top surface to the greatest depth of fungal penetration in the epithelial and connective tissues, and averaged results are shown in Figures 6A–E for dpi 1 to 5. As shown in Figure 6, the invasive depths during dpi 1 to 5 are similar between WT- and CHK23-infected tongues. In agreement with the lower CFU count, the invasive ability was most impaired for the null mutant CHK21, and then for the CHK25 and CHK26 mutants versus WT- or CHK23-infected tongues (p < 0.001).