Escherichia coli DH5α was used as a host for plasmids propagation. To grow DH5α, LB broth medium was used and when necessary, supplemented with ampicillin (0.1 mg/ml). Cultures of E. coli were incubated at 37°C under constant agitation (220 rpm) for 16–20 h. For plate cultures, 0.7% BactoTM Agar (Brunschwig, Switzerland) was added.

C. albicans were grown in complete YEPD medium (1% Bacto peptone [Difco Laboratories, Basel, Switzerland], 0.5% yeast extract [Difco] and 2% glucose [Fluka, Buchs, Switzerland]) at 30°C under constant agitation (220 rpm). For growth on plates, 2% BactoTM Agar was added to the medium. Yeasts were transformed by a lithium-acetate procedure with slight modifications as previously described (Sanglard et al., 1996).

For the construction of the red-shifted luciferase, two rounds of mutagenesis were performed by PCR. First, two mutations (S284T and L295F) were introduced to allow peak emission red-shifting. For this purpose, two overlapping PCR were performed on pACT1-CaLucOpt (Jacobsen et al., 2014) using the Phusion® high-fidelity Taq polymerase (NEB, Switzerland). The first PCR was carried out with the primers Hind_LUCopt (GCAAGCTTAAAATGGAAGATGCTAAGAACA) and LUCopt_MUT1_R (AAAGTAGATTTAGCCAAGAAAGAGAACAAAGTTGGAACCAACAAAGCAGTTTGAATCTTG). The second PCR was performed with primers LUCopt_MUT1_F (CAAGATTCAAACTGCTTTGTTGGTTCCAACTTTGTTCTCTTTCTTGGCTAAATCTACTTT) and Nhe-LUCopt (GCAAGCTAGCCTATACAGCAATTTTACCAC). The primers LUCopt_MUT1_F and LUCopt_MUT1_R are overlapping and contained the mutations. Next, a third PCR was performed to fuse the two previous PCR products using 10 ng of each PCR product as template and the two external primers (Hind_LUCopt and Nhe-LUCopt). The final product was cloned after HindIII and NheI digestion in pACT1-GFP (Barelle et al., 2004), thus yielding pDS1901. A second round of mutagenesis was performed to introduce two mutations (T214A and A215L). The two PCR were performed on pDS1901. The first and second PCR used, respectively, the primer pair Hind_LUCopt and LUCopt_MUT2_R (GTGAGAGAATCTAACACACAAAGCTCTGTGTGGCAAAGCAAC) and the primer pair LUCopt_MUT2_F (GTTGCTTTGCCACACAGAGCTTTGTGTGTTAGATTCTCTCAC) and Nhe-LUCopt. The primers LUCopt_MUT1_F and LUCopt_MUT1_R are overlapping and contained the mutations. A last PCR was performed as above with primers Hind_LUCopt and Nhe-LUCopt and the final product was cloned in pACT1-GFP to yield pDS1902.

The initial plasmids pACT1-CaLucOpt and pDS1902 were digested by StuI and transformed as described above in CAF4-2 (Fonzi and Irwin, 1993) to yield DSY4824 (expressing Mut0) and DSY4823 (expressing Mut2), respectively. To express Mut2 in another C. albicans strain (strain 101, a gift from S. Leibundgut, University of Zürich), a plasmid (pDS1928) containing the SAT1 dominant marker was used (Reuss et al., 2004). pDS1928 was constructed from pDS1904, which contained the SAT1 resistance marker in CIp10 (Murad et al., 2000). A 3.5 kb XhoI-BamHI fragment from pDS1902 containing Mut2 was inserted in pDS1904 to yield pDS1928. This plasmid was linearized by StuI for transformation into strain 101, yielding DSY4976. Five clones of each transformation were checked for luminescence signal and those with best signals were selected (data not shown).

For in vitro cells suspension assays, C. albicans strains were grown overnight in YEPD. Each strain was subsequently diluted 100-fold in 3 ml of YEPD medium and grown to a concentration of 1.5 × 10⁷ cells/ml. Cells were then washed twice in phosphate-buffered saline (PBS: 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄) and resuspended in 3 ml of PBS. Cells concentration was measured using a spectrophotometer. Cells were next centrifuged and resuspended in 2 × RPMI (Sigma, Switzerland) buffered at pH 7 or 4.6 with MOPS at 10⁸ cells/ml. Fifty microliters of the cells suspension were then loaded on a 96-well plate (half-area, black, COSTAR). Sixty microliter of D-luciferin (Biosynth, Switzerland) resuspended in H₂O at a concentration of 1.6 mg/ml was injected automatically into the wells, followed by an orbital shacking of 3 s before recording luciferase activity (RLU: Relative Luminescent Unit) using a 96-well plates luminometer (FLUOstar Omega, BMG Labtech). Filters available for spectral analysis were at 460, 520, 530, 550, 580, 620, and 675 nm.

All animal experiments were performed at the University Hospital Center of Lausanne with approval through the Institutional Animal Use Committee, Affaires Vétérinaires du Canton de Vaud, Switzerland (authorization n° 1734.4), according to decree 18 of the federal law on animal protection. For all mice experiments, female BALB/c mice (7 weeks old; Charles River France) were housed in ventilated cages with free access to food and water. They were controlled and scored daily for health status. Score of health was based on ability to heat and drink, to move, on fur aspect, on body weight and on body temperature (1 to 3 points for each criteria, where 1 corresponds to “healthy” and 3 to “strongly altered”). Healthy mice had a score of 5 points. Mice with a score above 10 or with a body weight loss equal or above 25% or body temperature below 35.5°C were immediately euthanized. The C. albicans strains were grown in individual tubes for 16 h under agitation at 30°C in YEPD medium. Each strain was subsequently diluted 100-fold in YEPD medium and grown overnight under agitation at 30°C. Overnight cultures were washed twice with PBS and resuspended in 5 ml PBS.

The concentration of each C. albicans culture was measured through optical density, and each strain was diluted to 2 × 10⁶ cells/ml. Mice were injected through the lateral tail vein with 250 μl of the cell suspension. For drug treatments, mice were injected intraperitoneally daily starting 24 h post-infection (p.i.) with 100 μl of fluconazole (FLC: 10 mg/kg; Toronto Research Chemicals) or 100 μl of caspofungin (CAS: 1 mg/kg; Cancidas® Merck, Switzerland). For determination of fungal burden, the kidneys of euthanized animals were removed, homogenized in sterile 0.05% NP40 in water and serial dilutions were plated on YEPD agar.

Immunocompetent mice were infected sublingually with 2.5 × 10⁶ CFU C. albicans yeast cells using a cotton ball as described (Schönherr et al., 2017). For determination of fungal burden, the tongue of euthanized animals was removed, homogenized in sterile 0.05% NP40 in water and serial dilutions were plated on YEPD agar. For drug treatment, FLC was diluted in 0.5% hydroxyethyl-cellulose (Sigma) and 30 μl of the preparation was administered under isoflurane anesthesia with 200 μl pipette tips under the tongue daily starting 24 h p.i. (10 mg/kg/day).

Taking advantages of a C. albicans codon-adapted firefly luciferase (pACT1-CaLUCopt) reporter gene (Jacobsen et al., 2014), an optimization of this luciferase was undertaken to improve its properties. We were especially interested to improve in vivo light emission and tissue penetrance. We first introduce two mutations in CaLUCopt (S284T and L295F). S284T is known to result in a red-shifting in light emission, while L295F contributes to thermal stability (Branchini et al., 2005, 2007; Figure 1A). The rationale behind this idea was to reduce light absorption by tissues and hemoglobin, as reported (Papon et al., 2014). In a second round of mutations, two other substitutions were introduced (T214A and A215L) that were known to confer enhancement of luciferin-dependent reactions and a better resistance to body temperature as described (Branchini et al., 2007).

C. albicans transformants with the original and the optimized luciferase were obtained (designated as Mut0 and Mut2, respectively) and their light emission spectra analyzed first in vitro with whole cells on a standard laboratory luminometer with seven emission filters covering a 460–675 nm range. Firefly luciferase was detected in cell extracts from the C. albicans transformants in similar amounts (Figure S1). As shown in Figure 1B (upper panel), Mut0 showed a dual peak emission profile with a major peak at around 500 nm. In contrast, Mut2 exhibited a clear shift toward the red spectrum with a peak at about 625 nm. Similar in vitro measurements were conducted using the bruker in-vivo Xtreme II imaging platform, containing five filters spanning from 520 to 850 nm. The measured emission spectra profiles were similar to those obtained with a luminometer, except that the major peak of Mut0 emission could not be recorded. This instrument is dedicated to in vivo imaging and since wavelengths shorter than 550 nm are absorbed by host tissues, it does not record light emitted below 520 nm. Thus, it can be expected that Mut2 light emission could be better captured by the instrument camera, while an important portion of Mut0-emitted light may remain entrapped by mammalian tissues and is not recorded by the instrument dedicated to in vivo red-near infrared (NIR) imaging. In order to assess this assumption, emission spectra were next measured in vivo in mice infected with either C. albicans cells harboring Mut0 or Mut2 (inoculum 5 × 10⁵ CFU/mouse) at 48 h p.i.. This infection time allows the establishment of the C. albicans infection at the site of the kidneys. After spectral analysis, we could observe light emission peaks at higher wavelength for Mut2 as compared to Mut0 (Figure 1C). In addition, both luciferases displayed a peak emission shifted to higher wavelengths as compared to in vitro emission (Figure 1B, lower panel and Figure 1C), confirming that light in the green-blue region is entrapped by mammalian tissues and not recorded, as expected.

We next addressed the performance of Mut2 in parallel to Mut0 in the invasive model of infection in mice. After acquisition of total emitted bioluminescence in mice in the region of the kidneys, fungal burdens in these organs were obtained and were compared to the signals emitted by both luciferase types (Figure 2). Signal intensities were about 3.7-fold higher for Mut2 as compared to Mut0 (median of 828.9 and 3,089 photons/s/mm², respectively; p < 0.005, Mann-Whitney test). In addition, the Mut2 luminescent signal correlated better with CFU than Mut0 (R² 0.92 vs. 0.58, Figure 2). These data clearly establish Mut2 firefly luciferase as a more sensitive in vivo luminescence reporter for C. albicans.

In this part of the study we aimed to establish the limit of detection of Mut2 in vivo. For this purpose, mice were infected with three different inoculums (10⁴, 5 × 10⁴, and 10⁵ cells/mice) of C. albicans strain expressing Mut2. Infection was monitored daily for three consecutive days recording kidney luminescence emission (Figure 3A). Globally, kidney signals were stable between 24- and 72 h p.i. A dose response was visible with a statistical significant difference between mice infected with 10⁴ cells as compared to mice infected with 10⁵ cells at the three time-points. The infection dose of 10⁴ cells gave more variable outcomes of infection (Figure 3A).

To establish the limit of detection, mice were sacrificed at 72 h p.i.. Fungal burdens were measured and correlated with the corresponding previously recorded luminescent signals. Due to technical limitations, it is not possible to detect <50 CFU per pair of kidney. Luminescence signals above the background detection levels were considered as bioluminescent signals produced by C. albicans. Both thresholds are indicated in Figure 3B by dashed regions. The detection limit should be set above the two thresholds. Such limit was estimated at about 400 CFU/pair of kidney (dotted line of Figure 3B), which corresponds to about 200 CFU/kidney.

Kinetics of bioluminescence emission was addressed in infected mice (n = 8) at early (6-, 18-, and 24 h) and later time points (48- and 72 h). Light emission was originating mainly from the region of the kidneys. While a constant progression of emitted light was observed up to 24 h p.i. in almost all animals (Figure 4), signals tended to stabilize at the later time points as observed previously in Figure 3A, and dropped in one individual (Figure 4, animal No 6). This observation suggests that fungal loads increase dramatically in infected kidneys in the first 24 h p.i., and stabilized thereafter. This is consistent with observations made by Jacobsen et al. (2014), who suggested that fungal loads were principally increasing over 24 h p.i.

BLI was next used to evaluate the efficacy of antifungal treatments on C. albicans (Mut2) infections initiated by intravenous challenge. We used fluconazole (FLC) and caspofungin (CAS) as therapeutic agents in these experiments. Infected animal treatments were started intraperitoneally 24 h after inoculation, and then administrated daily. Animals were individually followed by BLI for a period of 72–96 h p.i..

In the fluconazole experiment, the mean luminescence of untreated control animals increased ~2-fold over a 72 h period of treatment (Figure 5A). In fluconazole-treated animals, bioluminescence increased up to 48 h and then declined. These data suggest a delay in fluconazole efficacy. Since fluconazole was injected 24 h p.i., this delay may reflect limitations in the pharmacokinetics of this antifungal agent. The fluconazole effect observed with luminescent signals could be confirmed measuring CFU at 72 h p.i. toward end of the experiment. While untreated animals had a fungal burden around 10⁴ CFU/g of kidney, treated animals displayed 10- to 100 times lower kidney fungal burden (Figure 5A, right panel).

The caspofungin treatment experiment was performed separately and included an independent untreated control group. Figure 5B indicates that bioluminescence signals of the untreated group were first increasing over 72 h and then declined at 96 h p.i. In contrast, caspofungin-treated mice showed a decreasing bioluminescence all along the entire treatment time course, particularly from 48 to 96 h p.i. Efficacy of caspofungin treatment could be confirmed with fungal burden of kidneys in infected animals at 96 h: while CFUs were slightly above the limit of detection in treated animals (10²/g kidneys), fungal burdens were ranging between 10⁵ and 10⁷ CFUs/g kidney in untreated animals. These experiments clearly highlight differences in the pharmacodynamics of fluconazole and caspofungin treatments.

The OPC model was also tested with bioluminescent C. albicans cells expressing Mut2. To assess this model of infection with a bioluminescence read-out, we used a persistent C. albicans clinical isolate (strain 101). This strain is reported as persistent in the OPC model of infection as opposed to the strain SC5314, which is known to be cleared rapidly in this infection model (Schönherr et al., 2017). In our hands, strain 101 can persist for at least 96 h in the OPC model (Figure 6B), while SC5314 is cleared within 3 days p.i. (data not shown). Infecting mice with the strain 101 expressing Mut2 in the oropharyngeal area resulted as expected in luminescence signals located in the head region (Figures 6A,B). We used several experiments in which CFUs were obtained from animal's tongues at the end of experimentation (96 h) to correlate oral cavity luminescent signal with tongue fungal burden. Figure 6A shows that bioluminescence signals were correlated to recovered CFUs. The correlation was significant, even though the correlation coefficient (R = 0.52) was not excellent. We next addressed the efficacy of FLC in this model by applying the substance sublingual at three different time points post infection (24-, 48-, and 72 h). While a significant decrease in light emission was detected starting at 48 h post-treatment as compared to placebo, no light was detected in FLC-treated at later time points (up to 96 h; Figure 6B). In agreement, no C. albicans cells were recovered from tongues of treated animals at 96 h post-treatment (data not shown). These data clearly demonstrate the utility of BLI in this type of infection model.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.