The Genome of Banana Leaf Blight Pathogen Fusarium sacchari str. FS66 Harbors Widespread Gene Transfer From Fusarium oxysporum

Abstract

Fusarium species have been identified as pathogens causing many different plant diseases, and here we report an emerging banana leaf blight (BLB) caused by F. sacchari (Fs) discovered in Guangdong, China. From the symptomatic tissues collected in the field, a fungal isolate was obtained, which induced similar symptoms on healthy banana seedlings after inoculation. Koch's postulates were fulfilled after the re-isolation of the pathogen. Phylogenetic analysis on two gene segments and the whole genome sequence identified the pathogen belonging to Fs and named as Fs str. FS66. A 45.74 Mb genome of FS66 was acquired through de novo assembly using long-read sequencing data, and its contig N50 (1.97 Mb) is more than 10-fold larger than the previously available genome in the species. Based on transcriptome sequencing and ab initio gene annotation, a total of 14,486 protein-encoding genes and 418 non-coding RNAs were predicted. A total of 48 metabolite biosynthetic gene clusters including the fusaric acid biosynthesis gene cluster were predicted in silico in the FS66 genome. Comparison between FS66 and other 11 Fusarium genomes identified tens to hundreds of genes specifically gained and lost in FS66, including some previously correlated with Fusarium pathogenicity. The FS66 genome also harbors widespread gene transfer on the core chromosomes putatively from F. oxysporum species complex (FOSC), including 30 involved in Fusarium pathogenicity/virulence. This study not only reports the BLB caused by Fs, but also provides important information and clues for further understanding of the genome evolution among pathogenic Fusarium species.

Keywords: Fusarium sacchari; banana leaf blight; comparative genomics; de novo assembly; gene transfer.

FIGURE 1

Banana leaf blight disease and pathogen isolation. (A) Symptoms of the BLB disease in the field. (B,C) The appearances of the fungal colony at 7 days after single-spore inoculation on PDA plates from the front (B) and the back (C). (D) Morphology of the fungal conidia and spores under the optical microscope. Blue, black, and red arrows denote large conidia, small conidia, and ascospores, respectively. (E) Symptoms on leaves inoculated with both the candidate pathogen and sterile water after the puncture. The white and black arrows pointed to the area punctured with needles and inoculated with water and FS66, respectively. (F) Symptoms on a banana seedling treated with leaf non-wound inoculation. Pictures (E,F) were taken at 30 days after inoculation.

FIGURE 2

Phylogenetic tree based on the partial coding sequence of RPB2. The strain names and NCBI accession IDs (in brackets) are shown at the tips. The phylogenetic trees constructed by neighbor-joining (NJ) and maximum likelihood (MX) have the same topology, and the bootstrap support rates (percentage) from NJ and MX are shown before and after the slash next to the bipartition nodes, respectively. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates have been collapsed.

FIGURE 3

Distribution of genes, RNA-seq reads, and genetic variations across the FS66 genome. The data shown in this graph were calculated in 10 kbp continuous non-overlapping windows across the genome. From outer to inner: the GC content (GC%) panel (black); the ideogram (sky blue) of the 47 non-organelle contigs arranged by contig size; the gene count panel (red) showing the number of annotated genes; the RNA-seq panel (light blue) in which the log₂ (read count mapped in each window) was shown; the unaligned region panel (yellow) which depicted the unaligned length between FS66 and Fs str. NRRL 66326 in each window; the SNP count panel showing the number of SNPs (including small indels) between FS66 and Fs str. NRRL 66326.

FIGURE 4

The best supported phylogenetic tree based on whole-genome proteins of FS66 and several adjacent Fusarium species. The species tree has been constructed based on the 9050 orthogroups, which include members from all the analyzed genomes. The lengths of the branches are proportional to genetic distance. The number on top of the branches denote the bipartition’s support rate at their right ends, which is the proportion of orthogroups supporting the bipartition. The number below each branch indicates the number of identified gene duplication events.

FIGURE 5

Orthogroup sharing relationships among six Fusarium genomes. The first two characters in the isolate names denote the species and the rest characters were the strain names listed in Supplementary Table 1. For some isolates, alphabet characters such as ‘NRRL’ were omitted for convenience. FM, F. mangiferae; FP, F. proliferatum; FF, F. fujikuroi; FV, F. verticillioides; FO, F. oxysporum.

FIGURE 6

Secondary metabolite gene cluster distribution in 12 Fusarium genomes. The isolates were named in the same way as in Figure 5 except for FVA3/5 in which ‘FV’ denotes F. venenatum. (A) Distribution of gene cluster sharing patterns among the 12 genomes in a total of 67 gene clusters. In the bottom panel, gray dots indicate absence, while colored dots (yellow, blue, red, and green) show the presence of the gene clusters in the corresponding genomes. The patterns in green color denote specific gene cluster loss in FS66, red denotes gene clusters putatively transferred from FOSC in FS66, yellow patterns include gene clusters present in more than 50% genomes, and blue pattern gene clusters were present in less than 50% genomes. (B) Availability of previously annotated gene clusters in the genomes. Green rectangles indicate the presence of the biosynthetic gene clusters in the corresponding genomes.

FIGURE 7

Putative genomic regions transferred from FOSC in the FS66 genome identified by phylogenetic incongruence analysis. (A) Species tree (left) and the incongruent molecular tree (right) suggesting gene transfer from FOSC in FS66. (B) Genomic regions putatively transferred from FOSC across the FS66 genome. Red links were inferred by phylogenetic incongruence analysis and green links were inferred based on that they were shared between FS66 and FOSC but not found in other tested FFSC genomes.