The study protocol was approved by the Medical Ethics Committee of Beijing Shijitan Hospital: sjtky11-1x-2020 (20).
Strains of C gattii in previous reports and their molecular types
We used “Cryptococcus gattii” and “China (or specific location)” to search all related publications in PubMed and China National Knowledge Infrastructure (CNKI). After screening abstracts and reviewing full texts, we found 22 studies containing 86 strains between 2006 and 2020 with molecular types.
Strain isolation and molecular types of C gattii in our study
32 clinical C gattii strains from 18 hospitals between 2011 and 2017 in China were isolated and were stored at -80°C in glycerol. After subculture, the isolates were re-identified by canavanine glycine bromothymol blue (CGB) agar or matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS, Bruker), but five [15·6%] of them could not be recovered due to improper storage. The molecular type (VGI and VGII) was determined by traditional PCR based MLST.13
Clinical data of patients with C gattii in our study
We retrospectively collected clinical data of 32 patients infected with C gattii, including gender, age, onset address, immune status, contact history, symptoms, lung CT, brain MRI presentation, intracranial pressure (ICP), biochemistry of cerebrospinal fluid (CSF), initial diagnosis, strain source, strain isolation time, admission to diagnosis time, treatment, survival prognosis, final diagnosis, and genotyping (table 1).
DNA/RNA Sequencing for C gattii strains
DNA extraction was performed similarly as previous report.14 All genomes were sequenced by Illumina HiSeq X Ten platform by 150 bp paired-end sequencing. Nanopore Sequencing was performed on four representative isolates (three VGIb and one VGIa, appendix p 6). RNA was extracted by following the protocol from commercial Kit (CWBIO, CW0581, China) and was sequenced by Illumina HiSeq X Ten platform by 150 bp paired-end sequencing.
DNA-Seq data of C gattii strains analyzed from publicly available data
Additional public C gattii sequencing data was downloaded from NCBI (appendix p 7-8), which was then converted to fastq format by fastq-dump, version 184.108.40.206 Then the adapters were removed by fastp, version 0.20.0.16
Each Illumina data was aligned to the reference genome (VGIa reference WM276 in phylogenetic analysis and VGIb assembled reference from our G8 strain in SNP calling analysis) by bwa-mem2.17 To call variants, the Picard tools (http://picard.sourceforge.net) were used for pre-processing, followed by GATK HaplotypeCaller, version v220.127.116.11, 18 with recommended filtering parameters. The final vcf file was annotated with a customized database by SnpEff, version 4.3t.19 In addition, ploidies were estimated based on the ratio of heterozygous sites given by the visualized output of ploidy NGS, version 3.1.2,20 and copy number variation (CNV) was called by CNV kit, version 0.9.6.21
The single nucleotide polymorphisms (SNPs) were retained if all samples were homozygous, and then were converted to phylip format, which was used to construct a phylogenetic tree by RAxML, version 8.2.12, with 1000 bootstraps.22 VGII strain R265 was set as the outgroup, and the final phylogenetic tree was visualized by iTOL, version 6.23
Multiple plate-based virulence assays
The strains were incubated in YPD liquid medium at 30°C for 16 h, centrifuged, and then washed twice with phosphate-buffered saline (PBS). The strains were uniformly adjusted to 5×107 cells/ml, and then gradient diluted to 107 cells/ml, 2×106 cells/ml, 4×105 cells/ml, 8×104 cells/ml, and 1.6×104 cells/ml. The cells (3 ml from each concentration) were respectively plated onto yeast extract-peptone-dextrose (YPD) agar medium to observe the growth at different temperatures (30℃, 37℃, and 39℃), plated onto YPD agar medium containing hydrogen peroxide (0·03% H2O2) and diamide (3 mM) to observe the growth at 30℃ plated onto YPD agar medium containing different concentrations of antifungal drugs to observe the growth at 30℃ (Flucytosine: 300 μg/ml; Fluconazole: 20 μg/ml; Amphotericin B: 0·75 μg/ml). 24
Induction of melanin using minimal medium containing L-3,4-dihydroxyphenylalanine (L-DOPA): Culture as described above, the strains were also uniformly adjusted to 5×107 cells/ml, and then gradient diluted to 107 cells/ml, 2×106 cells/ml, 4×105 cells/ml, 8×104 cells/ml, and 1.6×104 cells/ml. The cells (3 ml from each concentration) were respectively plated onto minimal medium containing L-3,4-dihydroxyphenylalanine (L-DOPA) and incubated at 30°C for three days. Incubated at 30°C for three days and photographed for melanin production. 25
Induction of capsule using Dulbecco's minimal essential media (DMEM) medium: The culture conditions are as described above. In brief, the strains were washed three times with PBS and then uniformly adjusted to 2×105 cells/ml. Cells were resuspended in DMEM and incubated at 37°C and 5% CO2 for 24 hours. Staining with India ink, acquire images using Carl Zeiss Microscopy, and use the cell measurement software (Zen 2011) to manually measure capsule and cell diameter. 26
In vivo virulence study in mouse
Survival test: In order to verify the difference of virulence between VGIa and VGIb, representative strains were selected for mouse experiment, a VGII strain was used as the reference strain. The cells were adjusted to 2×106 cells/ml. Female C57BL/6 mice aged 8-10 weeks were used for survival test, and each group of VGIa, VGIb, and VGII contained ten mice. After the mice were anesthetized with 10% chloral hydrate, 50 μl (2×106 cells /ml) was slowly dropped into the nasal cavity of each mouse. The state of the mice was observed every day and the death was recorded.
Lung and brain fungal burden test: The process of mouse infection was the same as survival test, five mice in each testing group. After 14 days of infection, the mice were killed, the lung and brain tissues were dissected and placed in a 15 ml centrifuge tube containing 4 ml PBS, the tissues were broken up with a homogenizer. The tissue suspension was serially diluted (ten times) , 100 ml of tissue suspension were plated onto YPD agar medium, incubated at 30℃ and counted after two days.
Hematoxylin and eosin (H&E) staining of lung tissues: H&E staining was performed as previously described.27 Briefly, the entire lung of Cryptococcus infected mice was dissected, fixed with 4% PFA and embedded with paraffin. Maximum cross section was selected and stained. The entire sections were scanned for analysis by Pannoramic midi: 3D/histech. CaseViewer, version 2.4, was utilized to visualize the stained section.
Assembly and annotation of C gattii genomes
Initial assemblies of four isolates with Nanopore sequencing reads were generated by Canu, version 2.0, with parameter genome size = 18,000,000.28 Illumina reads were aligned to genome drafts by bwa-mem217, which was used to polish for two rounds with Pilon, version 1.23.29 Assembly contigs were discarded if they represented the mitochondrial genome, the rest of fully assembled contigs were aligned against WM276 with NUCmer, version 3.1,30 to evaluate whether a chromosome-level assembly was obtained correctly and examine potential recombination events. QUAST, version 5.0.2, was also used to evaluate the assembly quality.31 After we identified soft-masked repeat regions by RepeatMasker, version 18.104.22.168 BRAKER2, version 2.1.4, was performed to predict genes in C gattii.33 RNA-seq reads were aligned with STAR, version 2.5.3a,34 which was incorporated to improve gene callings. tRNA and rRNA was separately predicted by tRNAscan-SE, version 2.0.535 and rRNAmmer, version 1.2.36 Eggnog-mapper version2 was then used to predict gene functions,37 which were filtered if over 30% coding sequences overlapped with noncoding regions.
We applied Kruskal-Wallis with Dunn’s post-hoc test when comparing in vitro experiments and time from onset of symptoms to final diagnosis of VGIa, VGIb, and VGII. We used one-way ANOVA test and a Tukey’s honestly significant difference (Tukey’s HSD) post-hoc test in organ fungal burden test between three groups. In addition, Wilcoxon rank sum test was used to compare behaviors between two groups. All statistical analyses were done with R, version 3.6.3.
Role of the funding source
The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.