De Novo Genome Assembly and Comparative Genome Analysis of the Novel Human Fungal Pathogen Trichosporon austroamericanum Type-Strain CBS 17435

TL;DR

This paper presents the genome assembly of a new human fungal pathogen, Trichosporon austroamericanum, and compares it to a related species.

Contribution

The study provides the first de novo genome assembly of Trichosporon austroamericanum type-strain CBS 17435 using nanopore long-read technology.

Findings

The genome of Trichosporon austroamericanum was assembled using nanopore sequencing.

Comparative analysis revealed genetic differences between T. austroamericanum and T. inkin.

Abstract

Trichosporon austroamericanum is a recently described species recognized for its emerging clinical significance in invasive trichosporonosis. In this study, we present the nanopore long-read-based de novo genome assembly of the type-strain CBS 17435. Additionally, we performed genomic comparative analyses with its closest relative, Trichosporon inkin.

Full text

Trichosporon austroamericanum is a recently recognized emerging pathogen, noted for its clinical relevance in causing a range of infections, including both superficial and invasive trichosporonosis [1]. The first case of T. austroamericanum was identified during an epidemiological survey in 2013, when it was isolated from a urine sample of a Brazilian kidney transplant recipient. However, sequence analyses from retrospective studies and genomic databases have documented the presence of this species in Europe, Asia, and Latin America [1, 2]. Phylogenetic analyses indicated that this species is most closely related to Trichosporon inkin. Notably, T. austroamericanum exhibits a range of physiological characteristics, including the ability to grow at 45 °C, that differentiates it from other Trichosporon species [1].

We aimed to perform long-read nanopore sequencing and genomic analysis of the T. austroamericanum type-strain CBS 17435. This strain was subcultured onto 2% glucose, 50% yeast-extract water, 0.5% bacteriological peptone, 2% technical agar #2 (GYPA), supplemented with 0.5 M sodium chloride to reduce the excessive formation of extracellular polysaccharides, which otherwise negatively impact genomic DNA purification. The previously reported detailed protocol for high-quality genomic DNA purification was followed, with the modification of using proteinase K from Roche Diagnostics (Mannheim, Germany) [3]. The quality and quantity of genomic DNA was assessed using the Qubit in combination with the High Sensitivity kit (ThermoFisher, Waltham, MA, U.S.A.), and by 0.8% agarose gel electrophoresis.

One microgram of gDNA was used as starting point for the library preparation using the multiplex native ligation kit (SQK-NBD114.24; ONT, Oxford, United Kingdom) following the manufacturer’s instructions (protocol version NBE_9169_v114_revQ_15Sep2022, last updated February 16, 2024). The library was loaded onto a MinION R10.4.1 flow cell and raw data was collected using the GridION platform (software release 24.02.16; ONT). Basecalling was performed using Dorado (basecall_model_version_id = dna_r10.4.1_e8.2_400bps_hac@v4.3.0; ONT). Reads with a length ≥ 1000 bp and a quality-score (Q) of ≥ 10 were collected into a single FASTQ file for downstream analyses.

FASTQ data was subjected to an additional quality check using chopper v0.7.0 to collect reads with a length of ≥ 1100 bp and a ≥ Q10, followed by removal of an arbitrary 50 bp from the 5’- and 3’-ends [4]. Thereafter Flye v2.9.3-b1797 was used to generate the de novo genome assembly which was subsequently manually curated [5]. The haploid nuclear genome was found to be 20,968,827 bp in size, comprising eight fragments measuring 3,928,915; 3,876,917; 3,401,721; 2,954,482; 2,383,423; 2,186,886; 1,418,410; and 818,073 bp in length. The circular mitochondrial genome was 35,357 bp in length. Coverage of the nuclear genome was 79X, while the mitochondrial genome had a coverage of 3,479X. The haploid genome size of T. austroamericanum CBS 17435 closely matches that of T. inkin JCM 9195 (= CBS 5585), which has a haploid genome size of 20.35 Mbp long [6]. Both species have a rather decreased genome size compared to other haploid species in the Trichosporonales, which have an average genome size of 23.73 Mbp (± 4.38 Mbp; range 17.23–36.62 Mbp) [6]. The mitochondrial genome of T. inkin was recently determined to be 39,466 bp in length, and that of the more distantly related Trichosporonales species Apiotrichum gamsii and Apiotrichum gracile were 38,096 and 34,648 bp in length, respectively [7, 8]. The mitochondrial genome length of 35,357 bp reported here for T. austroamericanum is within the observed range of Trichosporonales species.

To assess the quality of the de novo genome sequence, a BUSCO v5.8.0 analysis was performed, that yield with the eukaryota_odb10 database 96.5% complete (95.7% single, 0.8% duplicated), 2.7% fragmented, and 0.8% missing BUSCO’s among 255 tested, with the tremellomycetes_odb10 database this was 92.7% (92.3%, 0.4%), 1.4%, and 6.0%, respectively, of the 4,284 BUSCO’s tested [9]. As a comparison, similar BUSCO analysis was done for the T. inkin reference genome of JCM 9195 (= CBS 5585) that was retrieved from NCBI Genome (accession number GCA_040365635.1, version April 4, 2024). This yielded comparable values, for the eukaryota_odb10 database 96.9% complete (96.1% single, 0.8% duplicated), 2.4% fragmented, and 0.8% missing genes, and for the tremellomycetes_odb10 database 92.4% (92.0%, 0.4%), 1.5%, and 6.1%, respectively. Additionally, compleasm v0.2.6 was run for the tremellomycetes_odb10 database and resulted in higher completeness scores, with 94.31% complete (94.19% single, 0.12% duplicated), 1% fragmented, and 4.69% missing BUSCO’s for the genome of T. austroamericanum CBS 17435, and 94% complete (93.84% single, 0.16% duplicated), 1.07% fragmented, and 4.93% missing BUSCO’s for the genome of T. inkin JCM 9195 [10].

We used the web-based deep learning tool Helixer v0.3.4 to predict the number of genes for CBS 17435 and JCM 9195, which were found to be 8,275 and 8,395 genes, respectively [11]. The latter represents a 9–24.1% increase compared to the previously reported 6,766–7,700 predicted genes, which were obtained using the tools Augustus and GeneMark-ES based on data from the Cryptococcus neoformans reference genome [6, 12].

The GC% of the nuclear genome of T. austroamericanum CBS 17435 was calculated to be 61.38%, comparable to the 63% of T. inkin JCM 9195. The GC% of the mitochondrial genome of CBS 17435 was found to be 27.11%, nearly similar to the 27.56% of that of JCM 9195 [8]. The average nucleotide identity (ANI) between the genomes of the T. austroamericanum and T. inkin type strains was calculated by OrthoANI using the USEARCH algorithm [13]. This analysis returned an OrthoANI value of 84.6472% between the genomes of CBS 17435 and JCM 9195. The ANI between this two Trichosporon species is comparable to the ~ 82% previously reported for members of the genus Cutaneotrichosporon, as well as to the ANI of 84.73% between Cutaneotrichosporon oleaginosus and Apiotrichum akiyoshidainum [14, 15]. These ANI values fall well below the 95% threshold recently proposed for species delineation in bacteria based on genome data, which correlates with the historical species delineation criterion of 70% similarity in DNA-DNA hybridization. The ANI reported here for the closely related siblings T. austroamericanum and T. inkin further supports their distinction as separate species.