A novel silkworm infection model with fluorescence imaging using transgenic Trichosporon asahii expressing eGFP

Trichosporon asahii is a pathogenic fungus that causes deep mycosis in patients with neutropenia. Establishing an experimental animal model for quantitatively evaluating pathogenicity and developing a genetic recombination technology will help to elucidate the infection mechanism of T. asahii and promote the development of antifungal drugs. Here we established a silkworm infection model with a transgenic T. asahii strain expressing eGFP. Injecting T. asahii into silkworms eventually killed the silkworms. Moreover, the administration of antifungal agents, such as amphotericin B, fluconazole, and voriconazole, prolonged the survival time of silkworms infected with T. asahii. A transgenic T. asahii strain expressing eGFP was obtained using a gene recombination method with Agrobacterium tumefaciens. The T. asahii strain expressing eGFP showed hyphal formation in the silkworm hemolymph. Both hyphal growth and the inhibition of hyphal growth by the administration of antifungal agents were quantitatively estimated by monitoring fluorescence. Our findings suggest that a silkworm infection model using T. asahii expressing eGFP is useful for evaluating both the pathogenicity of T. asahii and the efficacy of antifungal drugs.

Trichosporon asahii is a basidiomycetous yeast widely distributed in the environment [1][2][3][4] , and commonly isolated from human blood, sputum, skin, feces, and urine [5][6][7][8] . T. asahii is a pathogenic fungus that causes severe deep mycosis in patients with neutropenia [8][9][10] . Whereas the mortality rate of deep mycosis caused by Candida albicans is approximately 40%, that caused by T. asahii is approximately 80%, and the prognosis is poor 11 . T. asahii is resistant to echinocandin antifungal drugs and causes severe infections in patients treated with micafungin 12 . Moreover, strains resistant to amphotericin B and azole antifungal drugs such as fluconazole have been isolated from patients 13,14 . Therefore, T. asahii has become a serious clinical problem as a pathogenic fungus that causes systemic infection in immunocompromised hosts 8 .
T. asahii forms hyphae, which are branching filamentous structures 14 . In pathogenic fungi that form hyphae such as Candida albicans, hyphal formation is crucial for the host epithelial cell damage and biofilm formation that are involved in infections 15 . As with C. albicans, T. asahii makes treatment difficult by forming a biofilm on devices such as catheters 14,16 . Hyphal growth of T. asahii in blood vessels causes necrotic thrombi, and may contribute to infection 17 . The molecular mechanisms of infection caused by T. asahii, however, remain unclear. Establishing a simple animal experimental system for systemic infection with hyphal formation of T. asahii and developing genetic recombination technology in T. asahii will contribute to elucidate the infection mechanism.
The silkworm, an invertebrate, is a useful experimental animal for evaluating the pathogenicity of pathogenic microorganisms and the therapeutic effects of antibacterial, antifungal, and antiviral drugs [18][19][20][21][22] . Silkworms have advantages for conducting infection experiments that require a large number of individuals [23][24][25] as they are much less expensive to rear and maintain than mammals and their experimental use partly avoids the ethical problems associated with that of mammalian models 23,24 . These advantages allow for the use of a large number of silkworms to calculate the dose of a pathogen required to kill half of the silkworms (LD 50 ) and the administration dose of www.nature.com/scientificreports/ an antifungal drug required to promote survival in half of the silkworms (ED 50 ) for quantitative evaluation of the pathogenicity of microorganisms and the efficacy of therapeutic agents [26][27][28] . Furthermore, silkworm infection models can be used for in vivo screening experiments, and allow for the identification of virulence-related genes of microorganisms and therapeutically effective drugs. Virulence-related genes of pathogenic microorganisms such as Staphylococcus aureus, Candida albicans, and Candida glabrata have been discovered using silkworm infection models and gene-deficient mutant libraries 19,29,30 . Avirulent mutants that exhibit lower pathogenicity against silkworms also exhibit lower pathogenicity in infection experiments in mice. Many studies using antimicrobial drugs, toxic compounds, and natural products demonstrate that the therapeutic effects, pharmacokinetic parameters, and toxicity are similar between silkworms and mammals [31][32][33][34][35] . Moreover, lysocin E 36 , nosokomycin 37 , ASP2397 38 , and GPI0363 34 were identified by exploratory research using a silkworm infection model from microbial culture broths and chemical libraries as novel antimicrobial compounds that show therapeutic effects in mouse infection experiments. Therefore, the silkworm infection model is useful for elucidating pathogenic mechanisms and evaluating the efficacy of antifungal drugs 24,39 . Fungal hyphae comprise multiple cells, making it difficult to quantitatively estimate hyphal formation by counting colony-forming units 40 . Fluorescence imaging using enhanced green fluorescent protein (eGFP) is a simple and effective method for quantitatively evaluating cell proliferation and cell death in vivo 41,42 . We previously established a silkworm infection model using transgenic dermatophytes expressing eGFP, and developed a method for quantifying hyphal formation by fluorescence imaging 43 .
In the present study, we demonstrated that T. asahii can kill silkworms and successfully established a transgenic T. asahii strain expressing eGFP by genetic recombination using Agrobacterium tumefaciens. The therapeutic effects of antifungal drugs can be evaluated using a silkworm infection model with the transgenic T. asahii strain expressing eGFP. The silkworm infection model and gene recombination technology established in this study were effective toward elucidating the infection mechanism of T. asahii.

Material and methods
Culture of T. asahii. T. asahii JCM2466 was grown on Sabouraud dextrose agar plates and incubated at 27 °C for 2 days.

Minimum inhibitory concentration determination. The minimum inhibitory concentration (MIC)
was determined using a drug sensitivity test kit, yeast-like fungi DP (Eiken Chemical, Tokyo, Japan), according to the Clinical and Laboratory Standards Institute (CLSI) M27-A3 method. Briefly, T. asahii cells (2 × 10 3 cells per well of a 96 well plate) were incubated with twofold serial dilutions of antifungal agents at 37 °C for 2 days, and the MIC values were determined. eD 50 measurement. To evaluate the therapeutic effects of antifungal agents, T. asahii cells (1-5 × 10 6 cells) were injected into the silkworm hemolymph, and various concentrations of the antifungal agents (50 μl) dissolved in saline were injected immediately afterwards into the silkworm hemolymph. The doses were created by fourfold serial dilutions. To determine the ED 50 values, five or six silkworms were injected with each dose of the antifungal agents. Survival of the silkworms at 48 h was monitored. The ED 50 values were calculated from combined data of 4-5 independent experiments by simple logistic regression model using Prism 8 (GraphPad Software, LLC, San Diego, CA, USA, https ://www.graph pad.com/scien tific -softw are/prism /).

Construction of T. asahii expressing eGFP.
The plasmid for expressing eGFP in T. asahii was constructed according to a previous report 43 . The eGFP gene was introduced into a pAg1-NAT1 vector 44  www.nature.com/scientificreports/ eGFP gene, and the termination sequence of the Aspergillus nidulans trpC gene was incorporated downstream of the eGFP gene 45 . Cloning was performed using the infusion method according to the general method (In-Fusion HD Cloning Kit, Takara, Shiga, Japan). The primers used for polymerase chain reaction (PCR) amplification of each DNA region are shown in Table 1.
The pAg1-NAT1-eGFP was introduced into the T. asahii JCM2466 strain using the A. tumefaciens-mediated transformation method described previously 46,47 . The pAg1-NAT-eGFP was introduced into A. tumefaciens EHA105 strain by electroporation and transformants were grown on 2 × YT agar containing rifampicin (50 µg/ ml), chloramphenicol (25 µg/ml), and kanamycin (50 µg/ml). The transformant was co-cultured with the T. asahii JCM2466 strain at 27 °C for 2 days. The candidate transgenic T. asahii expressing eGFP were isolated as colonies grown on Sabouraud dextrose agar containing nourseothricin (50 µg/ml) and cefotaxime (100 µg/ml). Introduction of the eGFP gene into genome of the candidate strains was confirmed by PCR using the primers described in Table 1.

Imaging of eGFP-expressing T. asahii in silkworms.
Fluorescence imaging using eGFP-expressing fungi in silkworms was performed according to a previous report 43 . eGFP-expressing T. asahii cells (2 x 10 6 cells) were injected into silkworms. The silkworms were reared at 37 °C and their hemolymph was collected after 24 h. The hemolymph was placed on glass slides and covered by a glass coverslip. The samples were examined with bright light or ultraviolet light under a microscope equipped with a fluorescence lens (DP-74; Olympus, Tokyo, Japan). The pictures were randomly obtained and the image fluorescence was determined by Image J software (ImageJ 1.47t; National Institutes of Health, Bethesda, MD, https ://image j.nih.gov/ij/).

Statistical analysis.
The significance of differences between groups was calculated using the Tukey-Kramer method. P < 0.05 was considered significant.

Killing of silkworms by injection of Trichosporon asahii cells.
To establish a silkworm infection model with fungi, the silkworm rearing temperature after infection should be determined 24,48 . In Cryptococcus neoformans (H99 strain), by injection of 10 7 cells into silkworm hemolymph, the silkworms survived for 2 days rearing at 27 °C, but all silkworms died within 2 days at 37 °C 27 . Therefore, we first determined the rearing temperature at which T. asahii would kill silkworms. When silkworms injected with 1.9 × 10 6 cells of T. asahii were reared at 27 °C, most of silkworms survived at 48 h, but when reared at 37 °C, all of the silkworms died within 48 h (Fig. 1A-C). Hyphal formation of T. asahii was observed in the silkworm hemolymph after T. asahii injection under 37 °C rearing conditions (Fig. 1D). Silkworms died in a dose-dependent manner after administering T. asahii (Fig. 2A). The LD 50 was 3.9 × 10 5 cells (Fig. 2B and Supplementary Fig. 1). On the other hand, silkworms injected with autoclave-treated T. asahii cells did not die (Fig. 3A,B). We previously reported the LD 50 value of C. neoformans after 48 h at 37 °C 27 , which are the same experimental conditions we used to determine the LD 50 value of T. asahii. The LD 50 value of the C. neoformans H99 strain was 6 × 10 6 cells per larva. The LD 50 value of the T. asahii JCM2466 strain determined in this study is lower than that of the C. neoformans H99 strain, indicating that the pathogenicity of T. asahii JCM2466 strain is higher than that of the C. neoformans H99 strain under these experimental conditions. Evaluation of the therapeutic effects of antifungal drugs using the silkworm infection model. We previously found that the therapeutic effect of antifungal drugs could be evaluated using a silkworm infection model with C. neoformans 27 . Administration of the antifungal drugs amphotericin B, fluconazole, and voriconazole prolonged the silkworm survival time after injection of T. asahii (Fig. 4A). The ED 50 value of amphotericin B, fluconazole, and voriconazole was 1.3, 3.9, and 0.4 µg g −1 of larva, respectively (Table 2 and Supplementary Fig. 2). Hyphal formation of T. asahii in the silkworm hemolymph at 24 h after injection was inhibited by administering amphotericin B, fluconazole, and voriconazole (Fig. 4B). These findings suggest that the therapeutic effect of antifungal drugs can be quantitatively evaluated using the silkworm infection model with T. asahii. Table 1. Primers used in the study.

Establishment of a transgenic T. asahii strain expressing eGFP. By gene recombination using
Agrobacterium tumefaciens, we successfully introduced a gene into dermatophytes, which cause a superficial cutaneous fungal infection 46 . Using the Agrobacterium gene transfer method, a plasmid inserted with a nourseothricin resistance gene and a gene encoding eGFP was introduced into T. asahii to obtain a strain that grew on agar medium containing nourseothricin (Fig. 5A,B). When the genome of the nourseothricin-resistant strain was used as a template, a band of the expected size of the eGFP gene was amplified by PCR, but when the wild strain genome was used as a template, no amplification was observed (Fig. 5C). Moreover, the nourseothricinresistant strain emitted green fluorescence when irradiated with excitation light (Fig. 5D). We named the strain eGFP-Tg. These results suggest that a nourseothricin-resistance gene and gene encoding eGFP could be introduced into T. asahii by gene recombination using A. tumefaciens.

Fluorescence imaging analysis of T. asahii infection. We examined whether T. asahii grew in silk-
worms by microscopic observation of the hemolymph of silkworms infected with T. asahii. Hyphal formation was observed in the silkworm hemolymph at 24 h after injection of the T. asahii eGFP-Tg strain (Fig. 6A). On the other hand, we observed no increase in the number of colony-forming units of T. asahii in the hemolymph (Fig. 6B). We quantified the hyphal formation of dermatophytes in silkworms by measuring fluorescence intensity using transgenic dermatophytes expressing eGFP 43 . The fluorescence intensity increased in hemolymph of silkworms at 24 h after injection of T. asahii eGFP-Tg strain (Fig. 6C). The increased fluorescence intensity was attenuated by administering the antifungal drugs amphotericin B, fluconazole, and voriconazole (Fig. 7A,B). We www.nature.com/scientificreports/ also confirmed that the pathogenicity of the T. asahii eGFP-Tg strain in the silkworm infection model was similar to that of the wild-type strain ( Supplementary Figs. 3 and 4). These results suggest that fluorescence imaging using the transgenic T. asahii strain expressing eGFP is effective for quantitatively evaluating hyphal formation in the host.

Discussion
We established a silkworm infection model with T. asahii and evaluated the therapeutic effects of antifungal drugs using the silkworm infection model. Moreover, we generated a transgenic T. asahii strain expressing eGFP by gene recombination using A. tumefaciens. The silkworm infection model and the T. asahii gene recombination technology will be useful for elucidating the infection mechanisms of T. asahii and developing antifungal drugs. Silkworm systemic infection models caused by pathogenic fungi, such as C. albicans, C. tropicalis, C. glabrata, C. neoformans, Aspergillus fumigatus, and dermatophytes, are reported 26,27,29,30,38,43 . T. asahii infection models using mice with neutropenia and Galleria mellonella, an insect, have been reported 4 . In these infection models, the number of viable T. asahii cells in the host organs is measured by colony-forming units and therefore quantifying the actual cell proliferation of T. asahii may not be accurate. The fluorescence quantification method using the transgenic T. asahii expressing eGFP can be used to quantify cell proliferation by considering hyphal formation in the infection models using mice and Galleria mellonella.
We demonstrated that the death of silkworms caused by T. asahii was abolished by autoclaving the fungal cells, and that antifungal drugs inhibited the hyphal formation in silkworms and effectively treated silkworm infections. T. asahii also forms arthroconidia that are segmented conidia. The arthroconidia of T. asahii were observed in multinucleated giant cells present in the blood of a patient with Job's syndrome 49 . To our knowledge, however, the molecular mechanisms involved in hyphal and arthroconidia formation in T. asahii are unknown. Genetic evidence from experiments using a mutant that cannot form hyphae or arthroconidia is needed to clarify the contribution of hyphal and arthroconidia formation to the pathogenicity of T. asahii. Silkworms are useful for identifying virulence genes of pathogenic microorganisms 24 . Novel pathogenesis-related genes of Staphylococcus aureus, Pseudomonas aeruginosa, Bacillus cereus, C. albicans, and C. glabrata have been identified by screening mutants with reduced pathogenicity in silkworms from genetic-disrupted mutant libraries 19,29,30,50,51 . We previously reported in generating gene-deficient strains of dermatophytes by gene recombination using A. tumefaciens 46 . In the present study, we revealed that gene recombination using A. tumefaciens is a useful transgenic technique in T. asahii. Future studies will establish gene-deficient strains and screening for virulence genes in T. asahii. On the other hand, in the experimental system established in this study, multiple experiments are required to determine LD 50 value by the simple logistic regression model. Moreover, there are the difference in   Fig. 1). Optimizing the experimental condition to enable more reproducible measurement of LD 50 values is a future subject. Silkworms are useful for evaluating the pharmacokinetics and toxicity of compounds [31][32][33][34][35] . In mammalian animals, organs such as the intestinal tract, liver, and kidney govern the pharmacokinetics of drugs. Recent studies revealed that silkworms have functionally similar organs that affect drug pharmacokinetics and toxicity [32][33][34][35] . Many in vitro and in vivo analyses revealed that the absorption of compounds from the silkworm intestinal tract is similar to that in mammals 26,27,31,43 .  www.nature.com/scientificreports/ The ED 50 values of antibiotics in silkworms are similar to those in mice 26 . The difference in the total clearance, volume of distribution, and half-life values of anti-microbial agents such as chloramphenicol, tetracycline, vancomycin, rifampicin, micafungin, and fluconazole is less than 10-fold between silkworms and mammals 35 . The LD 50 values for compounds are also similar between silkworms and mammals 28,32 . Lysocin E, nosokomycin, ASP2397, and GPI0363 were identified by exploratory research using a silkworm infection model from microbial culture broths and chemical libraries as novel antimicrobial compounds that show therapeutic effects in mouse Figure 6. Growth of transgenic T. asahii expressing eGFP in silkworm hemolymph. Saline or T. asahii JCM2466 eGFP-Tg strain (1 × 10 6 cells) was injected into the silkworm hemolymph. Silkworm hemolymph was collected at 0, 6, and 24 h after injection and observed with a fluorescence microscope (A). The colony forming units (B) and fluorescence (C) were calculated. BF: bright field. Flu: fluorescent, in which the field of view was irradiated with excitation light for fluorescence detection. Scale bar, 20 µm. Statistically significant differences between groups were evaluated using Tukey-Kramer method. *P < 0.05.  52 . Therefore, silkworms are useful for in vivo screening of compounds that are candidate antibacterial and antifungal agents. The ED 50 values of amphotericin B and voriconazole in the silkworm infection model with T. asahii were 1.3 mg/kg and 0.4 mg/kg, respectively. A previous study using a guinea pig infection model with T. asahii reported amphotericin B and voriconazole ED 50 values greater than 1.5 mg/kg and 5-10 mg/kg, respectively 53 . Therefore, the ED 50 value of voriconazole in the silkworm model is lower than that in the guinea pig model. In the previous report using the guinea pig model, antifungal drugs were administered 24 h after infection. On the other hand, in the silkworm infection model, antifungal drugs were administered immediately after infection. The difference in the timing of the administration may account for the difference in ED 50 values. T. asahii exhibits natural resistance to candin antifungals such as micafungin, and strains resistant to azole antifungals such as fluconazole and amphotericin B have also been clinically isolated [12][13][14] . New antifungal drugs might be obtained by in vivo screening using a silkworm infection model with a multidrug-resistant T. asahii strain.

Scientific RepoRtS
In conclusion, the silkworm T. asahii infection model is useful for quantitatively evaluating the pathogenicity of T. asahii and the therapeutic effects of antifungal drugs. We also revealed that T. asahii can be genetically modified using Agrobacterium. The transgenic T. asahii expressing eGFP might be useful for in vivo fluorescence imaging in various infection models. Figure 7. Antifungal drug-induced inhibition of growth of transgenic T. asahii expressing eGFP in silkworm hemolymph. (A) Saline or T. asahii JCM2466 eGFP-Tg strain (3 × 10 6 cells) was injected into the silkworm hemolymph, followed by injection of 50 µl of amphotericin B (AMPH-B, 100 µg/ml), fluconazole (FLCZ, 100 µg/ml), or voriconazole (VCZ, 100 µg/ml) into the hemolymph. Silkworm hemolymph was collected at 24 h after injection and observed with a fluorescence microscope (A). The fluorescence was calculated (B). Statistically significant differences between groups were evaluated using Tukey-Kramer method. *P < 0.05.