Strengthening the genomic surveillance of Francisella tularensis by using culture-free whole-genome sequencing from biological samples

Abstract

Introduction: Francisella tularensis is a highly infectious bacterium that causes the zoonotic disease tularemia. The development of genotyping methods, especially those based on whole-genome sequencing (WGS), has recently increased the knowledge on the epidemiology of this disease. However, due to the difficulties associated with the growth and isolation of this fastidious pathogen in culture, the availability of strains and subsequently WGS data is still limited.

Methods: To surpass these constraints, we aimed to implement a culture-free approach to capture and sequence F. tularensis genomes directly from complex samples. Biological samples obtained from 50 common voles and 13 Iberian hares collected in Spain were confirmed as positive for F. tularensis subsp. holarctica and subjected to a WGS target capture and enrichment protocol, using RNA oligonucleotide baits designed to cover F. tularensis genomic diversity.

Results: We obtained full genome sequences of F. tularensis from 13 animals (20.6%), two of which had mixed infections with distinct genotypes, and achieved a higher success rate when compared with culture-dependent WGS (only successful for two animals). The new genomes belonged to different clades commonly identified in Europe (B.49, B.51 and B.262) and subclades. Despite being phylogenetically closely related to other genomes from Spain, the detected clusters were often found in other countries. A comprehensive phylogenetic analysis, integrating 599 F. tularensis subsp. holarctica genomes, showed that most (sub)clades are found in both humans and animals and that closely related strains are found in different, and often geographically distant, countries.

Discussion: Overall, we show that the implemented culture-free WGS methodology yields timely, complete and high-quality genomic data of F. tularensis, being a highly valuable approach to promote and potentiate the genomic surveillance of F. tularensis and ultimately increase the knowledge on the genomics, ecology and epidemiology of this highly infectious pathogen.

Keywords: Francisella tularensis; Microtus arvalis; RNA oligonucleotide baits; SureSelect; WGS.

Figure 1

Overview of the success of the target and enrichment sequencing approach and its relation to multiple sample features. (A) Relation between genome coverage and percentage of reads on-target (three bottom panels), and multiple factors: initial DNA input, absolute input of genome equivalents (GE), NGS library molarity and number of reads after quality control (QC). Samples are ordered from 1 to 49 (on the xx-axis) from the sample with the highest to the lowest percentage of horizontal genome coverage (by at least 1-fold). (B) Association between absolute GE input [available for 52 samples (45 sequenced and 7 excluded)] and: (i) the percentage of reads on-target (upper panel) and, (ii) the percentage of genome coverage (lower panel). Gray triangles in the upper panel identifies the samples that were excluded before sequencing due to low library concentration but for which GE counts were available (n = 7). The excluded sample with high GE counts (1×10⁸) corresponds to sample FT-MA1953. (C) Distribution of successful, failed and excluded samples among the samples from common voles (M. arvalis, n = 50) and Iberian hares (L. granatensis, n = 13).

Figure 2

Maximum likelihood phylogenetic tree of the newly sequenced genomes from clade B.10 (D1; n = 17) and all available Francisella tularensis subsp. holarctica genome assemblies from Spain (n = 58; Supplementary Table 2). The phylogenetic tree was generated using MEGA-CC v10.0.5 with 100 bootstraps based on 113 core single nucleotide variant positions (SNVs) extracted from a multiple genome alignment with 1,558,560 bp, generated with parsnp using the genome of strain FTNF002-00 (acc. no. NC_009749.1) as reference. Tree nodes are colored by clade (at level D5) and metadata blocks show the respective year of collection, host and canSNP clades and subclades (from level D4 to D8). The strain IDs from samples generated in this study using the capture and enrichment approach (TCE) (n = 11) are shown in red. Other genomes obtained from strains isolated in culture are shown in black and bold (n = 4). The panel on the right shows the color codes for each clade and subclade identified among the analyzed genomes and the ancestry relationships between them. The clades identified among the TCE genomes are highlighted in red. The ancestral clades (SNP path) of major clade B.45 identified by CanSNPer2 are B.1-B.2-B.3-B.5-B.6-B.10-B.11-B.44-B.45 (not shown). The lower panel on the right shows the geographical distribution of the samples, with the same sample color scheme of the nodes on the phylogenetic tree (clade level D5). A smaller map shows the geographical location of the study area within the Iberian Peninsula.