Background

The giant panda (Ailuropoda melanoleuca) is one of the rarest endangered animals[1, 2]. Dermatomycosis has become the second major disease of giant pandas, which seriously affect the survival of giant pandas [3]. Our previous studies have shown that Trichosporon spp. is the dominant genera of the body surface [4] and genitals of giant pandas [2]. These species may lead to trichosporidiosis in humans and animals.

Trichosporon is a genus of fungi that belongs to the order Tremellales in the class Tremellomycetes (division Basidiomycota) and is widely distributed in nature [5]. Trichosporon spp. can cause superficial fungal infections such as tinea pedis, onychomycosis, and dermoid infections [6]. With the increasing prevalence of immunocompromised patients, the incidence of invasive fungal diseases has increased, and Trichosporon has become the second commonest genus of yeast in deep fungal infections in patients with hematologic malignancies, granulocytic deficiency, and bone marrow transplants [7]. Meanwhile, Trichosporon spp. infections of animals have increased, such as disseminated trichosporosis in cats [8], canine meningitis [9], and tortoise cutaneous infection [10]. Owing to the difficulties in classification and identification of Trichosporon spp., studies on them have substantially lagged behind other species in many areas such as clinical characteristics, antifungal susceptibilities, and the selection of therapeutic drugs [11]. To accurately identify Trichosporon spp., a number of molecular methods have been developed, of which DNA sequencing of the internal transcribed spacer (ITS) region, the D1/D2 domain of the 26S subunit of the rRNA gene region, and the intergenic spacer 1 (IGS1) region are the most frequently used. The IGS1 gene region is particularly useful in phylogenetic studies and the description of intraspecies variation [12, 13]. Ribeiro et al. identified 21 clinical isolates as belonging to six species on the basis of the ITS and IGS1 regions [14]. In 2009, Chagas Neto et al. identified 22 isolates from human blood by analyzing the IGS1 region [15]. In 2011, Guo identified 29 clinical isolates of Trichosporon by analyzing 3 loci, and eight Trichosporon spp. were found, of which Trichosporon asahii was the commonest [13].

Morphology study also used to identify Trichosporon spp., but the result was not revealing. Li performed slide culture of six clinically common Trichosporon spp. and found no significant differences in colony morphology [16].

Trichosporon spp. are the dominant fungal species on giant panda skin and genitals [4]. Meanwhile, there are no reports on systemic identification of Trichosporon spp. isolated from animals. So identification Trichosporon spp. at the species level is important for preventing and treating dermatomycoses in giant pandas.

We collected 29 isolates of Trichosporon spp. from the skin of giant pandas breeding at the China Conservation and Research Center for Giant Pandas, Ya’an. Owing to the shortage of sequence data for the IGS1 region in individual species, we had to analyze the ITS region, D1/D2 domain, and IGS1 region of the isolates to obtain accurate classification information. Afterward, the morphological development process was observed by the slide culture method. Mice were artificially infected, and their livers and skin were taken for pathological analysis.

Methods

Sampling procedure

Samples were collected from clinically healthy giant pandas (22 females and 22 males) at the China Conservation and Research Center for Giant Pandas (Ya’an, China) in 2015–2016. Pandas that had been treated with antifungal drugs during the previous 6 months or with a recent history of disease were excluded from this study. The pandas lived in a semi-captive semi-enclosed breeding environment, were fed a diet of about 10% steamed cornbread and fruits and 90% bamboo shoots, and were allowed to drink water ad libitum [2].

All personnel involved in sampling wore sterile protective clothing, hats, masks, and latex gloves. Use sterilized shears to remove most of the hair when a panda ate fruits, 70% alcohol was used to sterilize the surface of the upper back of forearm wrist (5.0 cm × 5.0 cm, approximately), use the edge of the sterilized scalpel to scrape the surface and then a suitable amount of dander were collected. All samples were quickly placed in sterilized plastic sample bags, transported to the laboratory on ice within 2 h, and then immediately processed in a BSL-2 safety cabinet. No repeat sampling was performed on the same panda, and all 44 samples were processed for isolating Trichosporon spp.[2].

Fungal culture

Samples were streak-inoculated under aerobic conditions onto Sabouraud dextrose agar (SDA) (MOLTOX, Inc., Boone, NC) containing 4% (m/v) glucose, 1% (m/v) peptone, and 1.5% (m/v) agar. When samples were first inoculated, media were supplemented with chloramphenicol (0.005%, m/v). The chloramphenicol are not added to the medium in subsequent culture.

Fungal culture was carried out in a BSL-2 safety cabinet in a bioclean room. Sterilized sealing film was used to cover each plate. Each sample was plated onto three culture plates with three control plates. All culture dishes were inoculated and stored at 25 °C for 7–30 days before being considered negative [2].

Molecular identification

Fungal DNA was extracted from pure culture as described previously [17]. Amplification of the ITS region, D1/D2 domain, and IGS1 region was performed as described with the primer pairs ITS1/ITS4 (ITS1: 5′-TCCGTAGGTGAACCTGCGG-3′; ITS4: 5′-TCCTCCGCTTATTGATATGC-3′), F63/R635 (F63: 5′-GCATATCAATAAGCGGAGCAAAAG-3′; R635: 5′-GGTCCGTGTTTCAAGACG-3′), and 26SF/5SR (26SF: 5′-ATCCTTTGCAGACGACTTGA-3′; 5SR: 5′-AGCTTGACTTCGCAGATCGG-3′), respectively [12, 18]. PCR amplification was performed in a 50 μl reaction mixture containing 19 μl 2 × Taq Master Mix (Tsingke Biotech Co., Ltd., Chengdu, China), 2 μl primers, 25 μl double-distilled water, and 2 μl fungal genomic DNA. The thermocycling conditions were as follows: 5 min at 98 °C (initial denaturation), 35 cycles of 10 s at 98 °C, 10 s at 58 °C, and extension at 72 °C for 10 s, and final extension for 4 min at 72 °C. A total of 8 μl of the amplified PCR products were visualized on 2% agarose gel after staining with GreenView (Solarbio, Beijing). The PCR products were then sequenced by Tsingke Biotech Co., Ltd. (Chengdu, China).

All the chromatograms of DNA sequences were examined to ensure high-quality sequences. For species identification, the sequences of the ITS region, D1/D2 domain, and IGS1 region were queried against the NCBI database (https://www.ncbi.nlm.nih.gov/genbank). The sequence of each locus and concatenated sequences were then aligned using the NCBI BLAST and formed the consensus sequences for all 29 isolates. Phylogenetic trees were computed with MEGA version 6 (Molecular Evolutionary Genetic Analysis software version 6.0.2; http://www.megasoftware.net) using the neighbor-joining method, in which all positions containing gaps and missing data were eliminated from the dataset. The ITS region plus the D1/D2 domain and IGS1 region were used to produce two separate phylogenetic trees. All sequences of the three genes from the 29 isolates were deposited in the GenBank database (https://www.ncbi.nlm.nih.gov/genbank/) and were assigned ID numbers (Table 1).

Table 1 Nucleotide sequence accession numbers
Full size table

Morphological development process

Microscopic observations of the 29 isolates were made after slide culture on SDA [5]. A 0.5 ml sample of melted medium was injected into a closed glass Petri dish, which comprised a slide glass, a cover glass, and a copper ring with a hole in the wall, and was inoculated via the hole [19, 20]. All isolates were incubated at 25 °C and were observed after 24, 48, 72 and 96 h. The cover glass was stained with 5 ml Lactophenol cotton blue (Hopebio, Qingdao, China) and was observed with a microscope (BX51, Olympus).

Pathogenicity experiment

Animal experiment

In total of 216 sex-matched SPF Kunming mice (Dashuo Science and Technology Co., Ltd., China), which belonged to 11 experiment groups and one control group, with ages of 6–8 weeks were used. Each Trichosporon sp. was inoculated by skin inunction (We abraded skin with emery paper until a slight bleeding and cut the hair (2 cm × 2 cm) on the back of mouse after anesthesia by diethyl ether), subcutaneous injection, and intraperitoneal injection into immunosuppressed and non-immunosuppressed mice. In total, six groups were used, each of which comprised three mice. Groups A (intraperitoneal injection), B (subcutaneous injection), and C (skin inunction) were immunosuppressed (Mice were given intraperitoneal injection with 50 mg/kg cyclophosphamide at intervals of 48 h, three times in total, and each mouse was given 15 mg of penicillin sodium under the skin.); groups D (intraperitoneal injection), E (subcutaneous injection), and F (skin inunction) were non-immunosuppressed.

Preparation of fungal suspension and inoculation

Before being inoculated, mice in the immunosuppressed groups were intraperitoneally injected with Trichosporon spp. that had been cultured in SDA at 25 °C for 5 d. The mycelium and spores were scraped, washed with physiological saline, and mixed well. A hemocytometer was used to adjust the concentration of the spore suspension to 1 × 107 CFU/ml. Except for control group, each mouse was inoculated with 0.1 ml fungal suspension. Mice in the control group received 0.1 ml physiological saline instead. The backs of the mice treated by skin inunction were shaved, sterilized with 75% (w/w) alcohol, and lightly abraded with a 25G needle, and then 0.1 ml fungal suspension was gently rubbed onto the skin with a sterile injector.

Tissue sample processing

The ingestion and clinical symptoms of the mice were observed daily. The mice were anesthetized and euthanized with 5 ml diethyl ether (Chengdu Kelong Chemical Reagent Factory, Chengdu, China) using anesthetic gas box (AC-100, Yuyan Instruments, Shanghai), decapitated, and dissected to observe lesions on the seventh day after infection. The livers of mice in groups A and D and skin lesions from mice in the other groups were taken for fungal culture and pathological evaluation. The livers and skin were placed in 100 ml 4% formalin (w/w) (Chengdu Kelong Chemical Reagent Factory; Chengdu, China) for histopathological study via staining with hematoxylin/eosin (HE) and periodic acid/Schiff stain (PAS)[9].

Results

Molecular identification and genotyping

The interspecies identification of 22 isolates of Trichosporon spp. was performed from the data in Fig. 1. Seven strains of Trichosporon (JYZ1221, JYZ1224, JYZ915, JYZ1223, JYZ1261, JYZ1252, and JYZ12922) could not be identified because of the lack of sequence information for IGS1 in the GenBank database. However, it was determined that these seven Trichosporon strains belonged to three species. The interspecies identification of all 29 isolates of Trichosporon spp. was performed using the data in Fig. 2. The structures of the two phylogenetic trees were basically identical, and the method could be used to determine the accuracy of the identification of the 29 isolates. The 29 isolates of Trichosporon spp. were identified as belonging to 11 species, namely, T. laibachii (4 strains), T. gracile (1), T. brassicae (1), T. domesticum (1), T. guehoae (3), T. asteroides (5), T. jirovecii (5), T. cutaneum (1), T. shinodae (1), T. middelhovenii (3), and T. moniliiforme (4). T. jirovecii and T. asteroides were the commonest species (17%, 5/29). Moreover, T. middelhovenii and T. shinodae were isolated from the surfaces of animals for the first time.

Fig. 1
figure 1

Phylogenetic tree based on IGS1 sequences

Full size image
Fig. 2
figure 2

Phylogenetic tree based on ITS sequences plus D1/D2 sequences

Full size image

In this study, preliminary genotyping was performed on Trichosporon spp. T. asteroides (JYZ1251, JYZ1281, JYZ371, JYZ1255, and JYZ951) had the same genotype as its reference strain, which was isolated from the blood of immunocompromised patients [15], whereas T. laibachii (JYZ3252, JYZ921, JYZ321, and JYZ912) had different genotypes. The reference strain was isolated from humans. The five isolates of T. jirovecii were identified as having genotype 1[13]. T. brassicae (JYZ1253) had the same genotype as the reference strain, which was isolated from cabbage [12]. T. gracile (JYZ1291) had the same genotype as the reference strain, which was isolated from spoiled milk [12]. T. domesticum (JYZ983) and the reference strain had the same genotype; the reference strain was isolated from human sputum [12]. Three strains of T. moniliiforme (JYZ932, JYZ331, and JYZ323) had the same genotype as the reference strain, which was isolated from curdling milk [12], but the strain JYZ372 did not have this genotype. It was difficult to determine whether T. cutaneum had the same genotype as the reference strain. According to the phylogenetic tree, the genetic relationship was distant and it may not have had the same genotype.

Morphological development process

All 29 isolates were identified as Trichosporon spp. after molecular identification. The morphological development processes of the Trichosporon spp. were clearly different, and the difference was significant for the processes of single-spore development. Spores of T. moniliiforme and T. shinodae tended to be self-replicating at the beginning of their development and reproduced mainly by budding. Spores of the other species tended to become mycelia and reproduced mainly by the differentiation of mycelia to form spores or produce arthrospores. Morphological development could be used as an important basis for the identification of Trichosporon spp. For example, T. moniliiforme produced a large number of spores, and the spores were transformed into an oval shape after development was completed. The growth of T. shinodae was the slowest among the Trichosporon spp.; the shape of mycelia was specific and served as a basis for identification. T. laibachii produced a large number of arthrospores and had the unique feature that the mycelia were folded together. The main method of growth of T. guehoae comprised the reproduction of arthrospores by budding, and it had the specific feature that new mycelia grew from gaps in segmented mycelia. The morphological appearances of T. brassicae, T. domesticum and T. gracile were analogous in some ways. All these Trichosporon spp. tended to reproduce via the differentiation of mycelia into spores, and few arthrospores were seen during the process of development. The shapes of spores that differentiated from mycelia were different: those of T. gracile were quadrilateral, those of T. domesticum were round, and those of T. brassicae were disciform. The arthrospores of T. middelhovenii were fusiform and were always located in a bifurcation of the mycelium. This characteristic was distinct from other Trichosporon spp. T. asteroides had the feature that the spores overlapped each other, and the mycelia were thin and short. The mycelia could differentiate into spores, and the pigmentation of the spores was uneven, as shown by dyeing with cotton blue. The morphologies of T. jirovecii and T. cutaneum were similar, but the mycelia of T. cutaneum were more curved and tended to differentiate into spores, in contrast to T. jirovecii. Trichosporon spp. have individual morphological characteristics and hence could be distinguished by means of a comparison of their morphological development processes.

After culture for 24 h, spores of T. moniliiforme divided independently (Fig. 3a), no hyphae differentiated into spores (Fig. 3b, c, and d), and the mycelium and spores were evenly stained (Fig. 3b, c, and d). The mycelium produced a large number of arthrospores (Fig. 3c), and the shape of the spores changed from circular (Fig. 3c) to elliptical (Fig. 3d).

Fig. 3
figure 3

Morphological development process of Trichosporon moniliiforme from day 1 to day 4. a Most spores divided and a small number of spores expanded; b: Scattered hyphae appeared and began to produce conidia on day 2; c: Basic hyphae formed; d: Spores increased in number and their shapes were transformed from circular to elliptical

Full size image

The spores of T. laibachii grew into mycelium after cultivation for 24 h, and no spores divided independently (Fig. 4a). The mycelium spread radially (Fig. 4a) and was folded together (Fig. 4c), and differentiated into spores. Arthrospores were abundant, whereas spores were round and few in number (Fig. 4b, c, and d). The spores and mycelium were unevenly colored: the spores were darker, whereas the mycelium was lighter (Fig. 4b, c, and d). Hyphal folding was typical in structure.

Fig. 4
figure 4

Morphological development process of Trichosporon laibachii from day 1 to day 4. a Some spores germinated and mycelium began to divide; b: Mycelium produced arthrospores; c: Mycelium was folded; d: Mature mycelium and arthrospores

Full size image

The structural development of T. guehoae was completed after cultivation for 24 h (Fig. 5a). The mycelium spread radially but was scattered (Fig. 5a, b, c, and d). A large number of round spores were produced as grape-like clusters (Fig. 5a, b, and c). Hyphae and spores were evenly stained (Fig. 5a, b, c, and d). The mycelium grew from a segmented section to form a new mycelium (Fig. 5b), and the new spores were generated by arthrospores (Fig. 5d); this was a typical structure.

Fig. 5
figure 5

Morphological development process of Trichosporon guehoae from day 1 to day 4. a Mycelial buds formed new mycelium; b: New mycelium appeared at the point at which hyphae were segmented; c: A large number of botryoidal spores of T. guehoae were observed; d: New spores developed from arthrospores

Full size image

After T. gracile was cultured for 24 h, a large number of hyphae appeared and became segments, and no spores could be seen (Fig. 6a). It could be seen that the mycelium differentiated into spores (Fig. 6b, c and d). A large number of spores were produced, which were mostly square (Fig. 6c) and became oval after reaching maturity (Fig. 6d). The mycelia and spores were evenly stained (Fig. 6c). Some hyphae did not divide into sections and differentiated into spores at intervals; this was a typical structure (Fig. 6d).

Fig. 6
figure 6

Morphological development process of Trichosporon gracile from day 1 to day 4. a and b: Hyphae were segmented; c: A large number of hyphae differentiated into spores, some of which were square; d: Mycelium divided into oval spores

Full size image

After culture for 24 h, the spores of T. domesticum swelled to form mycelia and no spores divided independently (Fig. 7a). Hyphae were abundant and parallel to each other and had spindle-type buds (Fig. 7b). A large number of hyphae differentiated into spores (Fig. 7c and d). The number of spores was small with no arthrospores, and the spores were round (Fig. 7b, c, and d). The mycelia and spores were evenly stained (Fig. 7a, b, c, and d).

Fig. 7
figure 7

Morphological development process of Trichosporon domesticum from day 1 to day 4. a Spores expanded and sprouted; b: Mycelium had multiple branches and spindle-shaped buds; c: Mycelium began to differentiate into oval spores; d: Mycelium had completely differentiated into oval spores

Full size image

After culture for 24 h, the spores of T. brassicae exhibited no significant changes (Fig. 8a), whereas after culture for 48 h the spores swelled and formed hyphae (Fig. 8b). No spores were found to divide independently (Fig. 8a and b). The mycelia became segmented and were distributed parallel to each other (Fig. 8c). It could be seen that the mycelium differentiated into spindle-shaped spores (Fig. 8d). The mycelium and spores were evenly stained (Fig. 8d), and there were few arthrospores (Fig. 8c).

Fig. 8
figure 8

Morphological development process of Trichosporon brassicae from day 1 to day 4. a Spindle-shaped spores; b: Spores sprouted; c: Hyphae were segmented and spores were less oval; in the segmented hyphae, the arthrospores were fewer in number and spindle-shaped, and some spores developed into mycelium; d: Hyphae differentiated into spindle-shaped spores

Full size image

After cultivation for 24 h, some spores of T. shinodae swelled (Fig. 9a), whereas other spores divided independently (Fig. 9b). The hyphae were short and sparse and grew very slowly (Fig. 9c). A large number of round arthrospores were produced (Fig. 9d). The spores and mycelium were evenly colored (Fig. 9c and d). Crude short mycelium was produced after culture for 72 h (Fig. 9c); this was a typical structure.

Fig. 9
figure 9

Morphological development process of Trichosporon shinodae from day 1 to day 4. a Spores swelled and budded; b: Spores divided; c: Spores formed thick and short hyphae; d: Mycelium produced a large number of arthrospores

Full size image

After culture for 24 h, T. asteroides formed slender hyphae, and no spores divided independently (Fig. 10a). The mycelium was elongated (Fig. 10c) and could differentiate into spores (Fig. 10b and d). Spores on bifurcated mycelium aggregated into spheres (Fig. 10c); a large number of spores were produced, and the spores were round, oval, or spindle-shaped (Fig. 10b and d). The spores were not evenly pigmented and some were dark in color (Fig. 10d). The bifurcation of the mycelia and the aggregation of spores into spheres were typical structures (Fig. 10c).

Fig. 10
figure 10

Morphological development process of Trichosporon asteroides from day 1 to day 4. a Tenuous mycelium, some of which began to become segmented; b: Mycelium differentiated into circular, oval, and spindle-shaped spores; c: The tail ends of mycelium produced a large number of small, aggregating spores and bifurcated; d: Mycelium differentiated into darker spores

Full size image

After T. middelhovenii was cultivated for 24 h, hyphae were generated and no spores were found to divide independently (Fig. 11a). The hyphae were elongated (Fig. 11a and d), and their segments were small and indistinct (Fig. 11b and d). No hyphae differentiated into spores. Spores at bifurcations were spindle-shaped (Fig. 11a, b, c, and d). The spores were few in number and darker (Fig. 11c and d). Shuttle-type articular spores were characteristic structures (Fig. 11a, b, c, and d).

Fig. 11
figure 11

Morphological development process of Trichosporon middelhovenii from day 1 to day 4. a Hyphae were slender and did not divide; b: Hyphae produced fusiform arthrospores; c: Arthrospores budded to form new hyphae; d: Mature mycelium produced a large number of fusiform spores

Full size image

After T. jirovecii was cultivated for 24 h, hyphae appeared and no spores were seen to divide independently (Fig. 12a). The hyphae were segmented and radial (Fig. 12a) and differentiated into spores (Fig. 12d). Arthrospores were round and large in number (Fig. 12b); mycelia and spores were uniformly colored (Fig. 12d).

Fig. 12
figure 12

Morphological development process of Trichosporon jirovecii from day 1 to day 4. a Mycelium was slightly segmented and formed arthrospores; b and c: Mycelium formed a large number of circular arthrospores; d: Mycelium differentiated into spores

Full size image

After cultivation for 24 h, spores of T. cutaneum swelled and budded, and some spores divided independently (Fig. 13a). The mycelium was curved and segmented (Fig. 13b and c) and differentiated into spores (Fig. 13d). A large number of round arthrospores (Fig. 13c) were produced. The mycelium and spores were evenly stained (Fig. 13d), and curved mycelium was a typical structure (Fig. 13b and d).

Fig. 13
figure 13

Morphological development process of Trichosporon cutaneum from day 1 to day 4. a Some spores expanded and sprouted, and some spores divided; b: Spores formed curved mycelium; c: Most mycelium was segmented, and the mycelium produced round arthrospores; d: Mycelium began to fold and differentiated into spores

Full size image

Pathogenicity

The results of this study indicated that Trichosporon spp. mostly caused necrosis or swelling of hepatocytes and enlargement of the inter-hepatocyte space, and necrosis of hepatocytes mostly occurred near liver vessels. Subcutaneous injection of Trichosporon spp. caused lymphocyte infiltration into the skin, abscesses, and thickening of the stratum corneum. Mice that were inoculated via skin inunction had no obvious lesions, and most of them exhibited changes in the thickness of the stratum corneum, which in some cases resulted in subcutaneous abscesses. Tissues infected with Trichosporon spp. exhibited congestion easily both in skin and liver because of bleeding. Most of the Trichosporon spp. caused significant damage to the liver and skin, for example, T. laibachii, T. brassicae, T. guehoae, T. asteroides, T. jirovecii, T. cutaneum, T. shinodae, and T. middelhovenii. T. asteroides, T. laibachii, T. brassicae, T. guehoae, T. cutaneum, T. shinodae, and T. middelhovenii all produced spores in the skin infection model (Additional file 1: Figure S1, Additional file 2: Figure S2, Additional file 3: Figure S3, Additional file 4: Figure S4, Additional file 5: Figure S5, Additional file 6: Figure S6, Additional file 7: Figure S7, Additional file 8: Figure S8, Additional file 9: Figure S9, Additional file 10: Figure S10). In particular, T. asteroides gave rise to disseminated infections in the reticular layer of the skin (Fig. 14)G1 and budding in the dermis (Fig. 14G2). T. gracile, T. moniliiforme, and T. domesticum caused inconspicuous pathological changes, and hence their pathogenicity was weak.

Fig. 14
figure 14

Pathological sections of tissue damaged by Trichosporon asteroides (JYZ1255) infection. a: Hepatocyte necrosis, hepatic sinusoidal congestion, and unclear hepatic cord structure; b1: Abscess of the dermis, necrosis of muscle tissue, and blood capillary congestion; b2: Epidermis with a large area of blood stasis and abscess; c1: Thickening of granular layer of skin and cuticle; c2: Infiltration of neutrophilic granulocytes into the epidermis; d: Hepatocyte necrosis and hepatic sinusoidal congestion; E1: A large number of neutrophils infiltrated into the reticular layer, muscle tissue, and dermis, and some cells are necrotic; E2: Osteonecrosis of the dermis and hyperplasia of connective tissue; F: Thickening of granular layer of skin and cuticle; G1: Spores stained with periodic acid/Schiff stain (PAS) in the reticular layer of the skin of a mouse in group B; G2: Spores stained with PAS in the dermis and spore germination in a mouse in group B; H: Spores stained with PAS in the dermis of a mouse in group E

Full size image

Discussion

Interspecies identification of Trichosporon spp.

There have been reports on the isolation and identification of fungi from the body surface of the giant panda [4]. It was concluded that Trichosporon spp. were the dominant genus among skin flora on the giant panda [4]. Recently, there have been many reports on infections by T. asahii [21,22,23,24,25,26], but few mentions of other Trichosporon spp.[9, 27]. However, there have been reports that some animals are susceptible to rare Trichosporon spp.[27,28,29]. Because the phylogenetic relationship between Trichosporon spp. was very close, it was impossible to distinguish the different species of Trichosporon spp. according to the ITS region or D1/D2 domain in every case [30]. The sequence similarity between the ITS regions of T. asahii and T. asteroides, in which only two or three bases are different, is 99–99.3%, and T. montevideense and T. domesticum have identical ITS regions [31]. Scorzetti et al. found that the differences between the 28 s rDNA D1/D2 domains of different Trichosporon spp. are greater than those between the corresponding ITS regions. The ITS regions of T. laibachii and T. multisporum are identical, and seven bases are different in the D1/D2 domains. Two bases are different in the D1/D2 domains of T. montevideense and T. domesticum [32]. Guo amplified all three loci (ITS, D1/D2, and IGS1) and constructed a phylogenetic tree for the ITS region and D1/D2 domain and a separate phylogenetic tree for the IGS1 region. Both trees could completely distinguish the Trichosporon spp.[13]. In this study, seven strains could not be identified by their IGS1 regions because of the lack of sequence information for IGS1 in the NCBI database. Hence, we used the joint contribution of the ITS region and D1/D2 domain, which we compared with the phylogenetic tree for IGS1. It was found that the clades of the phylogenetic trees were basically identical and authenticated each other, so that all 29 strains could be identified completely.

Pathogenicity of dominant Trichosporon spp. isolated from pandas

T. asteroides and T. jirovecii were the dominant Trichosporon spp. that were isolated from the giant panda samples, and these species are widely present in giant pandas [4]. Their pathogenicity has a great influence on the health of giant pandas [4]. Especially, T. asteroides showed high pathogenicity because it caused disseminated infections in the reticular layer of the skin. This is consistent with the results of Chagas-Neto’s report [15]. T. jirovecii genotype 1 has been isolated from the human body [13], but its pathogenicity was unknown. Thus far, there have been few reports on T. jirovecii: Malgorzata et al. reported one case of mixed respiratory infection in a dog caused by T. jirovecii and Rhodotorula [33, 34], and Nardoni reported one case of back infection in a tortoise caused by T. jirovecii [10]. In the present study, four strains of T. laibachii (JYZ3252, JYZ921, JYZ321, and JYZ912) and one strain of T. moniliiforme (JYZ372) were identified as having new genotypes. Their pathogenicity remains to be confirmed by future studies.

Genotyping of Trichosporon spp.

At present, IGS1 sequence analysis is generally used for genotyping Trichosporon spp. For example, Chagas-Neto et al. completed the genotyping of 14 strains of T. asahii by IGS sequence analysis [15], whereas Guo completed the genotyping of 39 strains of T. asahii [12]. The main target in the genotyping of Trichosporon spp. has been T. asahii, and the genotyping of other Trichosporon spp. has been rare. In this study, among all the isolates only T. jirovecii had been assigned to two genotypes, and the other species had not been studied [13]. The main reason was that the identification of Trichosporon spp. is difficult and sequence information for IGS1 is scarce. In this study, only preliminary genotyping was performed for Trichosporon spp., and further research will rely on improvements in sequence information for IGS1in Trichosporon spp.

Morphological development process of Trichosporon spp.

There have been few studies on the morphology of Trichosporon spp. In 2005, Li et al. performed ITS-PCR detection and morphological and susceptibility testing on six Trichosporon spp. [16]. The colonies of different Trichosporon spp. were similar, but the morphologies of their mycelia and spores were significantly different. The structure of the mycelium was not destroyed, and the test results were credible. The morphology of the mycelia of T. domesticum was very similar to that in this study. The morphological development process of Trichosporon spp. was significantly different, and the majority of Trichosporon spp. had a typical structure: for example, septal differentiation of the mycelium in T. gracile (Fig. 8d); short thick mycelium during the development of T. shinodae (Fig. 11c); elongation and bifurcation of the mycelium and the aggregation of spores into spheres in T. asteroides (as shown in Fig. 10c); and a spindle-type articular spore structure in T. middelhovenii (Fig. 13a, b, c, and d). The above results proved that the morphological development process and typical structure have great significance as references for morphological identification.

From the point of view of the development and sporulation of mycelia, Trichosporon is an intermediate genus between molds and yeasts. Its mycelia can differentiate into a large number of spores like yeasts and also produce conidia like molds. Spores in the early stages of development can either bud like hyphae or divide like those of yeasts. Colonies of some Trichosporon spp. resemble yeasts in being milky, oily, and reflective, whereas colonies of some Trichosporon spp. have a radiate texture similar to that of molds [16]. Trichosporon might represent an intermediate genus in the evolution of yeasts into molds. In the study of the morphology of Trichosporon spp., they should be regarded as molds in order to observe their sporulation and mycelial structure.

Pathological changes in Trichosporon spp. infections

Different Trichosporon spp. cause similar pathological lesions on the skin and liver. T. asahii caused hepatic sinusoidal dilatation, mild to moderate dilatation of small blood vessels, hyperemia, neutrophil-based focal infiltration of inflammatory cells, and proliferation or degeneration of hepatocytes [35]. T. dermatis caused hepatic sinusoidal dilatation and congestion, swelling, degeneration, or necrosis of hepatocytes, and hyperplasia of Kupffer cells [36]. These lesions were similar to the pathological changes in the liver observed in this study. In the literature there are few mentions of skin lesions, subcutaneous abscesses, and bruises that were caused by T. dermatis.

However, most of the Trichosporon spp. identified in this study could cause skin lymphocyte infiltration, abscesses, and thickening of the stratum corneum. The pathological changes were significantly different between the groups treated by subcutaneous injection and skin inunction. These conclusions were similar to those of a study that was reported in China for the first time in 2010[37]. The skin damage caused in the group treated by skin inunction was lighter, and only T. laibachii and T. asteroides caused obvious pathological changes, which might be related to the uncontrollable amount of the spore coating and the pathogenicity of the Trichosporon spp. themselves. The spores developed a strong tendency to form mycelium, and the process of formation of mycelium could cause mechanical damage, which might be a reason for these observations.

Pathogenicity of the Trichosporon spp.

Except for T. moniliiforme, T. domesticum, and T. gracile, all the Trichosporon spp. in this study caused significant damage to the liver and skin in healthy mice (Additional file 1 Figure S1, Additional file 2 Figure S2, Additional file 3 Figure S3, Additional file 4 Figure S4, Additional file 5 Figure S5, Additional file 6 Figure S6, Additional file 7 Figure S7, Additional file 8 Figure S8, Additional file 9 Figure S9, Additional file 10 Figure S10). In most cases spores stained with PAS could be observed in skin sections. T. brassicae, T. guehoae, T. middelhovenii, and T. shinodae were found for the first time to be pathogenic forms of Trichosporon that could provoke obvious lesions in immunosuppressed and non-immunosuppressed groups. Hitherto, reports of the pathogenicity of these four Trichosporon spp. had not been found. This might be related to difficulties in the identification of Trichosporon spp. and differences in pathogenicity caused by the differences between strains.

In this study, T. asteroides (JYZ1255) exhibited strong pathogenicity. The infected tissue was extensively congested, and there was a large area of ​​abscess. Two mice in group A died 2 days after being inoculated with a suspension of T. asteroides. There have been many reports on T. asteroides, which is one of the main pathogens involved in trichosporosis in humans and is a dominant strain among fungi on the body surface of the giant panda. T. asteroides caused purulent keratitis [38] and was also isolated from the blood of patients with disseminated trichosporosis [13, 15]. T. asteroides gave rise to obvious disseminated infections in the reticular layer of the skin (Fig. 14G1) and budding in the dermis (Fig. 14G2). The infections might cause damage or even be life-threatening to immunocompromised giant pandas. T. jirovecii (JYZA10) was identified as having genotype 1 in previous studies (Fig. 1) and was significantly more pathogenic in immunosuppressed mice than in non-immunosuppressed mice and tissue. The degree of damage was significantly lower than that caused by T. asteroides, and it was inferred that genotype 1 of T. jirovecii was opportunistically pathogenic. Four strains of T. laibachii were identified as having new genotypes by phylogenetic analysis of the IGS1 sequence (Fig. 2). Although there have currently been no reported cases of infection involving T. laibachii, T. laibachii (JYZ3252) caused skin ulceration in mice (Fig. 4). This lesion demonstrated that the new genotype of T. laibachii is more pathogenic.

The pathogenicity test only studied the effects of Trichosporon spp. on the skin and liver, and most of the isolated Trichosporon spp. caused more severe damage to skin than to the liver. For example, T. brassicae (JYZ1253) and the reference strain, which was isolated from rancid milk, had the same genotype, but the isolated strain caused severe inflammatory reactions in skin tissue and skin necrosis. T. asteroides (JYZ1255) and the reference strain, which was isolated from human blood, had the same genotype, but the isolated strain caused disseminated infections in the reticular layer of the skin. It was assumed that the pathogenicity of Trichosporon spp. is related to their parasitic environment and that Trichosporon spp. that are isolated from the skin surface cause more pronounced damage to the skin.

Conclusions

We can concluded that combination of ITS, D1/D2, and IGS1 loci can effectively identify the genotype of Trichosporon spp. The morphological development process and typical structure of Trichosporon molds type have great significance as references for morphological identification.