P. wickerhamii S1 (mucous) was isolated from an 85-year-old female patient with a skin infection in Shanghai, China (Guo et al., 2022). P. wickerhamii HN01 (rough) was isolated from a 45-year-old male patient in Hunan, China, who had suffered from multiple skin infections and a bloodstream infection. P. wickerhamii ATCC 16529 was purchased from Thermo Fisher Scientific (Waltham, MA, USA). The strains were streak-inoculated onto Sabouraud dextrose agar (SDA) (Thermo Fisher Scientific) and incubated at 35°C for 3–5 days. Then, the colonies were harvested and washed with sterilised water (Jagielski et al., 2017). Primers used for PCR amplification and sequencing of the partial CYTB gene were cytb-F: GYGTWGAACAYATTATGAGAG-5’ and cytb-R: WACCCATAARAARTACCATTCWGG. Primers were used for both amplification and sequencing of degenerate nucleotides: Y, C, or T; R, A, or G; W, A, or T. The sequence of the CYTB gene of the strains in this study was aligned with the corresponding sequences retrieved from P. wickerhamii from the GenBank database.

Clinical strains S1 and HN01 of P. wickerhamii were subjected to micromorphological examination for phenotypic analysis. The morphological features of the cells were investigated on direct wet-mounted smears from SDA using an Olympus BX51 light and fluorescence microscope (Olympus Instruments Co., Tokyo, Japan). Images were taken using an Olympus U-TV0.63XC digital camera connected to a microscope. Olympus cellSens Standard 3.1 software was used for morphometric studies. P. wickerhamii cells were selected for fluorescence microscopy using the one-step fungal fluorescence staining method (Baso Diagnostic, Inc. Zhuhai, China).

P. wickerhamii S1 and HN01 were first picking into 20% BSA in cacodylate buffer (20 μL) and 0.8 μL BSA was transferred with a few samples to a 3 mm-diameter,100 μm-deep well of an aluminium carrier disk (Engineering Office of M. Wohlwend), quickly capped with a 0.16 mm × 3 mm sapphire disc (flat side coated with hexadecene), loaded into a customised 0.66 mm × 3.05 mm HPF specimen holder for rapid freezing at ~2,050 bar pressure for 340–360 ms using a Leica ICE HPF machine, and kept in liquid nitrogen until use.

The specimens were treated with a cocktail of 1% osmium tetroxide, 0.2% uranyl acetate, 3% H₂O in acetone, and 5% methanol at -90°C for 72 h. The temperature was raised to -60°C for 15 h and kept at -60°C for 24 h, then raised to -20°C for 15 h, and increased to 20°C from 2 to 17 h. This was followed by four washes with 100% acetone, each for 15 min. Subsequently, the specimens were incubated with 1% thiocarbohydrazide in 80% methanol at room temperature (RT) for 60 min, followed by six washes, each for 10 min, with 100% acetone. The specimens were stained with 2% osmium tetroxide in acetone at RT for 1 h and washed with 100% acetone four times each for 15 min, followed by incubation in 0.5% uranyl acetate in acetone in the dark overnight at 4°C. After four washes in pure acetone, each for 15 min, the samples were incubated in Epon resin with pure acetone for 3 h (at a ratio of 1:3), 3 h (at a ratio of 1:2), 3 h (at a ratio of 1:1), 3 h (at a ratio of 3:1), and 6 h each time, with four repetitions (100% Epon resin) on a rocker. The samples were embedded with freshly made resin and cured overnight at 38°C and then polymerised in a 60°C oven for 48 h. For transmission electron microscopy (TEM), samples were sliced (approximately 70 nm thick) using a Leica EM UC7 ultramicrotome and collected on copper grids. Sections of spermatids collected by FACS were stained with 2% uranyl acetate and Sato’s triple lead. All sections were imaged using a Talos L120C TEM (Thermo Fisher Scientific) at 80 kV, equipped with a 4 k × 4 k Ceta CCD camera.

The liquid cultures of P. wickerhamii S1 and HN01 were centrifuged at 10,000 rpm. The pellets were washed with 0.1 M sodium phosphate buffer (PB) pH 7.3 three times. The pellets were then fixed with 2.5% glutaraldehyde (25% glutaraldehyde ampules; Ted Pella Co., Redding, CA, USA) and 2% paraformaldehyde (16% paraformaldehyde; Ted Pella Co.) mixture in PB by incubating at RT for 30 min, then incubating at 4°C overnight. The pellet was washed three times with PB and collected by centrifugation. The samples were dehydrated using different ethanol volumes (30, 50, 70, 80, and 90%), and for each ethanol volume, incubated on ice for 10 min. Samples were incubated in 100% ethanol at RT three times (total of 1 h). The samples were dried using a critical-point dryer with a final rinse in pure ethanol. The dried samples were mounted on stubs containing carbon adhesives and a palladium coating of approximately 6 nm. For scanning electron microscopy (SEM), the samples were placed in a Zeiss Gemini Smart FE-SEM 550 (Zeiss, Germany) and imaged at 1.5 Kv and 4 k × 4 k.

Nine samples (three P. wickerhamii strains ATCC16529, S1, and HN01, each with three biological replicates) were included in this study. The RNA-seq raw reads were preprocessed using SOAPnuke v1.4.0 (Yuxin et al., 2018). The clean reads were then mapped to the reference gene sequence (Bowtie 2 v2.2.5) (Ben and Salzberg, 2012) to acquire the position information and unique read features of the nine sequenced samples. The gene expression levels of individual samples were then calculated as fragments per kilobase per million mapped reads (FPKM) using RSEM v1.2.8 (Li and Dewey, 2011). Significantly differentially expressed genes (DEGs) between the two P. wickerhamii strains were identified using DEseq 2 (fold change > 2 and P < 0.05) (Love et al., 2014).

The MSstats R package (Meena et al., 2014) was used to quantify the peptides and proteins from the mass spectrometry data of the 12 samples (three P. wickerhamii strains, each with four biological replicates). The MSstats R package was also used for intrasystem error correction and normalisation for each sample. Significantly differentially expressed proteins (DEPs) between the two types of P. wickerhamii strains were identified with a fold change > 2 and P < 0.05. For the functional annotation of a protein set, the protein sequences were aligned to Eukaryotic Orthologous Groups (KOG) using BLASTp with an E-value of 1E−5 or less (Koonin et al., 2004).

The 18 sample result files (three P. wickerhamii strains, each with six biological replicates) from Compound Discoverer were input into MetaX for data preprocessing and further analysis. After the log 2 transformation of the data, a PLS-DA model was established to screen the differential metabolites between two types of P. wickerhamii strains, with the scaling method Pareto and a 7-fold cross-validation. Significantly differential metabolites were defined as fold change > 2, P-value < 0.05, and variable importance for projection (VIP) >1. Pathway enrichment analysis of significantly differentially expressed metabolites was performed using the Kyoto Encyclopedia of Genes and Genomes pathway database (https://www.kegg.jp/kegg/pathway.html).

Exponentially growing cells of the three P. wickerhamii strains were cocultured with bone marrow-derived macrophages (BMDMs) (multiplicity of infection [MOI] = 1:1) at 37°C with 5% CO₂ for 6 and 24 h, respectively. After incubation, the cells were fixed with 4% formaldehyde, washed, stained with calcofluor white (10 min co-incubation), and observed under a fluorescence microscope. The concentration of lactose dehydrogenase (LDH) was assayed to determine the cytotoxicity of the three P. wickerhamii strains (CytoTox 96 Non-Radioactive Cytotoxicity Assay; Promega, Madison, WI, USA). Released LDH in culture supernatants was measured for 30 min, coupled with an enzymatic assay, which was used to measure the membrane integrity for cell-mediated cytotoxicity assays to measure the lysis of target cells by the three P. wickerhamii strains.

Statistical tests were performed using the Student’s t-test and Wilcoxon test. Statistical significance was set at P < 0.05.

P. wickerhamii HN01 colony morphology was consistent with that of ATCC16529; sporangia and sporangiospores were globose, and the capsule was absent. The diameter of P. wickerhamii HN01 sporangia measured 2.69–5.90 μm. After incubation on SDA at 35°C for 3 days, the colonies were creamy white, raised, and circular, with a smooth surface and a diameter of 3–5 mm ( Figure 1A ). P. wickerhamii HN01 can grow at 28 or 37°C.

P. wickerhamii S1 sporangia were globose to ellipsoidal or angular, sporangiospores were globose to angular and water drop-shaped, and the capsule was absent. The diameter size of P. wickerhamii S1 sporangia on average was 1.53-3.65 μm. After incubation on SDA at 35°C for 3 days, colonies were creamy white, glistening, circular, raised, and slimy, with a smooth surface and margins, and were up to 3–5 mm in diameter ( Figure 1A ). P. wickerhamii S1 grew well at 28 and 37°C. The cell wall thickness of the HN01 and S1 strains were 0.14 and 0.05 μm, respectively.

To better understand the differences at the macro- and micromorphological levels between the two P. wickerhamii strains S1 and HN01, we studied their transcriptome, proteome, and metabolome data for comparison. These data were highly reproducible ( Figure S1 ). The results showed that the number of DEGs between P. wickerhamii strains S1 and ATCC16529 was much higher than that between HN01 and ATCC16529. This was also true for the DEPs and metabolites ( Figure 2A ). This is consistent with previous results showing that strains HN01 and ATCC16529 evolved closely (Guo et al., 2022).

Interestingly, we found that mannan endo-1,4-β-mannosidase was significantly downregulated in S1 compared to that in HN01 ( Figure 2B ). This produced less glucan and mannan, which are the two main components of the algal cell wall. Consequently, this contributed to a thinner cell wall in the mucous strain compared to that in isolate S1 ( Figure 1D ).

Next, we performed a functional annotation of the significantly downregulated proteins in P. wickerhamii S1. Strikingly, the results showed that they were enriched in cell wall-related functions such as cell wall/membrane/envelope biogenesis, extracellular structures, collagens (type IV and XIII), and related proteins such as GDP-mannose pyrophosphorylase/mannose-1-phosphate guanylyltransferase ( Figure 2C ). These results are consistent with the transcriptome results, further suggesting that the downregulation of mannose-related enzymes may be responsible for the thin cell wall in S1.

The mucoid appearance of the S1 strain was likely due to the secretion of abnormal metabolites. Therefore, we identified differential metabolites between S1 and HN01, including significantly increased glycerol-related metabolites in S1 such as glycerol 2-phosphate, glycerol 3-phosphate, 2-linoleoyl glycerol, etc. ( Figure 2D ). This may partially explain the mucoid appearance of S1.

We further analysed the pathways in cases in which metabolites were significantly upregulated in S1 compared to HN01. The results showed that they were enriched in glycerol-related pathways, such as linoleic acid metabolism, fatty acid metabolism, and arachidonic acid metabolism ( Figure 2E ). Taken together, we speculate that the mucoid appearance of S1 may be caused by an increase in linoleic acid, glycerol, and other metabolites.

To compare the differences between the two strains with different morphologies in inducing host immune cell responses, we investigated the interaction of P. wickerhamii and macrophages co-cultured P. wickerhamii strains with macrophages. RAW264.7 macrophages were co-cultured with P. wickerhamii at a ratio of 1:1 (MOI = 1) ( Figure 3A ). Compared to the LDH of P. wickerhamii in a medium without macrophages, we found that the macrophage cytotoxicity of isolate HN01 was 14.52% at 24 h, but S1 showed lower viability (1.87%) compared to HN01. However, the cytotoxic effects of P. wickerhamii ATCC16529 were 9.91%. These results show that S1 caused much less damage to macrophages than strains HN01 and ATCC16529.

Thus, P. wickerhamii S1 showed lower cytotoxicity against macrophages, which could help P. wickerhamii escape macrophage clearance. Furthermore, the mechanisms underlying the innate immune response against P. wickerhamii should be elucidated in future studies. Taken together, S1 caused less macrophage death than HN01 and ATCC16529 ( Figure 3B ), which may be related to the abnormal thinning of the S1 cell wall.

In China, we collected 31 P. wickerhamii strains that cause human infections, mainly from 15 provinces and cities ( Figure 4 ). These cases were mainly concentrated in provinces and cities in southern China with developed waters, including domestic sewage and artificial watercourses. Shanghai, Fujian, Jiangxi, Zhejiang, and Guangxi had the highest separation rates of P. wickerhamii strains at 16.13% (5/31), 12.90% (4/31), 12.90% (4/31), 12.90% (4/31), and 9.68% (3/31), respectively. There have also been cases of P. wickerhamii bloodstream infections that have led to death, although most cases involved multiple skin infections. In China, owing to the lack of understanding of the pathogenicity and biology of P. wickerhamii, the correct diagnosis rate of P. wickerhamii infections is very low. For example, no cases of infection were found in the Hubei and Guizhou provinces.

According to the literature, P. wickerhamii infections have also been reported in Hong Kong and Taiwan. Unfortunately, there are no studies involving a large-scale investigation of P. wickerhamii contamination in both natural and artificial waterbodies in China. However, many reports have shown that P. wickerhamii is widespread in the environment and can cause infections in animals. With further research, new species and atypical strains of Prototheca may be found in the environment and in animal or human infections.