RESOURCE ANNOUNCEMENTOpen Access icon OPENOpen Access license

Draft Genomes of Five Fusarium oxysporum f. sp. niveum Strains Isolated from Infected Watermelon from Texas with Temporal and Spatial Differences

    Authors and Affiliations
    • Vanessa E. Thomas
    • Thomas Isakeit
    • Sanjay Antony-Babu
    1. Department of Plant Pathology and Microbiology, Texas A&M University, College Station, TX 77843


    Fusarium oxysporum f. sp. niveum (FON) is a soilborne fungal pathogen of watermelon that has significantly impacted production worldwide. In the United States, watermelon growers can experience 30 to 100% yield loss from this soil pathogen. In this study, we present five draft genome sequences of FON-race 2 isolates collected from four different counties within Texas, spanning more than 23 years. The estimated genome sizes of the de novo assembled genomes were W2 = 49 Mb, COM = 48 Mb, HAR = 49 Mb, STORM = 49 Mb, and PRU = 48 Mb. These sizes were on par with the previously sequenced FON races 0 and 2 from Texas, which ranged from 49 to 54 Mb. Gene prediction program AUGUSTUS predicted 13,554 genes in W2, 13,658 genes in COM, 13,678 genes in HAR, 13,821 genes in STORM, and 13,347 genes in PRU. The information from our five genome drafts contributes to the southeast genomic dataset of FON for comparative genomic analysis. Furthermore, the sequences are beneficial in the understanding of changes within the FON genome over time.

    The author(s) have dedicated the work to the public domain under the Creative Commons CC0 “No Rights Reserved” license by waiving all of his or her rights to the work worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law, 2023.

    Genome Announcement

    Fusarium oxysporum f. sp. niveum (FON), the causative agent of Fusarium wilt of watermelon, is a soilborne pathogen that has significantly impacted watermelon production throughout the world (Keinath et al. 2017). Four races (0, 1, 2, and 3) of this soilborne pathogen have been recognized, and all the variants have been demonstrated in the United States. Whereas race 2 was first reported in Texas in 1985, reports of race 3 are more recent, in 2010 and 2018, from Maryland and Florida, respectively (Amaradasa et al. 2018; Martyn 1985, 1987; Zhou et al. 2010). FON can endure in soils for a long time owing to its ability to prevail as dormant chlamydospores in infected soils, persist on the roots of nonhosts without causing disease, and survive as soil saprobes. Studies have documented the pathogen's persistence for up to 10 to 20 years (Ma et al. 2021; Wu et al. 2019; Xu et al. 2015). The long-term endurance can potentially lead the fungi to evolve its pathogenesis strategies as it inhabits soils and hosts over time. Hence, it is important to monitor FON genomes for evolutionary changes that can potentially enhance pathogen success. Furthermore, comparative population genomics between FON genomes over time will advance the information required to recognize genomic characteristics relevant to develop diagnostic methods for race differentiations. Additionally, the genome data will provide information on the evolution of virulence genes in FON as new strains emerge over time (Petkar et al. 2019). Thus, these draft genomes are a major resource to counter the economic losses incurred due to Fusarium wilt of watermelon. Here, we describe draft genomes of five isolates of FON race 2 from different regions in Texas that were initially isolated over a span of 23 years (1997 to 2020) and thus were both spatially and temporally separated (Fig. 1). Only one other FON race 2 genome from Texas is available in the literature (van Dam et al. 2016). Our study expands this resource.

    Fig. 1.

    Fig. 1. Map of Texas counties with our sample locations. COM was sampled in 2020, W2 was sampled in 2018, HAR was sampled in 2002, PRU was sampled in 1998, and STO was sampled in 1997. HAR and W2 were sampled in the same county; data on acres of watermelon grown by county in Texas are from U.S. Department of Agriculture, National Agricultural Statistics Service (USDA-NASS 2017).

    Download as PowerPoint

    The five FON isolates were isolated by one of the authors (T. Isakeit) from symptomatic watermelon stems growing in Texas fields. The tissues were surfaced disinfected with 10% bleach prior to plating onto water agar supplemented with streptomycin (100 mg/ml), with subsequent transfers to half-strength potato dextrose agar (0.5× PDA) medium. Identification was based on morphology, and putative verification of FON race 2 isolates was determined by inoculating race 1-resistant watermelon cultivars. The isolates COM, W2, HAR, PRU, and STO represent strains collected in the years 2020, 2018, 2002, 1998, and 1997, respectively. The isolates were enriched in tryptic soy broth (TSB) and placed in a shaker for 7 days at 125 RPM and 21°C. The biomass was transferred to fresh TSB supplemented with 100 mg/liter of streptomycin antibiotic to inhibit any co-isolated bacteria. The fungal cultures were incubated in a shaker for a minimum of 7 days at 125 RPM and 21°C. Once there was substantial fungal growth, the isolates were centrifuged at 12,000 RCF for 60 min, and pellets were washed three times with phosphate-buffered saline buffer. DNA from the enriched isolates was extracted using Quick-DNA Fungal/Bacterial Kits from Zymo Research Group (CA, U.S.A.) according to the manufacturer's standard protocol. The purity of the extracted DNA was assessed in a SpectraMax QuickDrop Micro-volume spectrophotometer (Molecular Devices, CA, U.S.A.) and quantified with a Qubit 2.0 Fluorometer (Invitrogen, Life Technologies, CA, U.S.A.). Absence of bacterial DNA was confirmed by lack of PCR amplification with primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) (Lane 1991) using the KAPA HIFI Hotstart Readymix PCR kit protocol (KAPA Biosystems, MA, U.S.A.). The same PCR protocol was applied to confirm fungal identity using the primers ITS1 (5′-TCCGTAGGTGAACCTTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) (White et al. 1990). PCR products were sent to Eton Bioscience (Triangle Park, NC, U.S.A.) for Sanger sequencing, and the resulting sequences confirmed that the samples were Fusarium oxysporum through nBLAST using the NCBI database.

    DNA samples that were confirmed to be void of bacterial DNA and belonged to Fusarium oxysporum were sent to Novogene Corporation (Sacramento, CA, U.S.A.) for microbial whole genome sequencing. A whole genome paired-end library was prepared for Illumina Novaseq PE150 sequencing to generate a range of 10−8 million raw reads per sample. Reads were analyzed on Texas A&M University's High Performance Research Computing (HPRC) Terra cluster ( with a de novo assembler, MEGAHIT (Li et al. 2015). Initial filtering was performed to remove adapters through the Trim-galore (Version 0.6.7) program's automatic detection setting. Potential mitochondrial genes and other possible contaminants from the genomes were filtered out using the Barrnap (Version 0.9) program (Krueger 2012; Seemann 2018). The remaining contigs were filtered at 1,000 bp, and the quality of the assembly was visualized with QUAST (Version 5.0.2) (Gurevich et al. 2013) within the GALAXY platform (Galaxy Version (Afgan et al. 2018). Individual Quast reports of the assemblies with different kmer parameters of the five isolates were used to compare the assembled quality of each genome. The contigs obtained through kmer 141 from MEGAHIT were selected based on superior N50 and L50 values as denoted by Quast analysis (Table 1). The resultant draft genomes (COM, HAR, STORM, PRU, W2) were all >48 Mb in size. GC content was similar across all five genomes at 48%. However, the largest contig sizes varied with isolates. Isolate STO (from 1997) had the highest overall genome length at 49.5 Mb. The isolate W2 (from 2018) had the largest contig at 834 kb (Table 1).

    Table 1. Summary of QUAST statistics of the five FON2 genomes

    The de novo assembled FASTA output were mapped to a reference genome of FON BioSample: SAMN15791673 from Hudson et al. (2020) using BWA-MEM (Li 2013). Outputs from BWA-MEM were further analyzed with antiSMASH fungal version 6.0 (Blin et al. 2021) to predict biosynthetic gene clusters (BGCs) and secondary metabolite biosynthetic pathways (Table 2). Twelve of the putative BCGs from the antiSMASH results showed significant similarities to known metabolites. Eight of these recognized BCGs were present in all genomes (Equisetin, gibberellin, AbT1, squalestatin S1, beauvericin, oxyjavanicin, fusaric acid, and bikaverin; Table 2). Enniatin was not detected in isolates STO and COM. ACT-Toxin II was not found in PRU and HAR, whereas gibepyrone-A was found in four of the five genomes, with W2 being the exception. The Cyclosporin C gene cluster was only found in the COM genome, and this secondary metabolite derived from the cluster can antagonize other fungi and utilizes nonribosomal peptide synthases (NRPSs) (Moussaïf et al. 1997). Other secondary metabolites that utilize the NRPS pathway include enniatin and beauvericin. Other pathways from recognized gene clusters were from type-one polyketide synthases (PKSIs) (ACT-Toxin II, gibepyrone-A, and oxyjavanicin) (Janevska et al. 2016; Miyamoto et al. 2010; Studt et al. 2012). The terpene pathway is utilized by gibberellin and squalestatin S1 (Bonsch et al. 2016; Hoffmeister and Keller 2007). The hybrid pathway of PKSI-NRPS is utilized by equisetin and fusaric acid (Boettger and Hertweck 2013; Studt et al. 2016). The five genome de novo assemblies underwent gene prediction and annotation through the Funannotate v1.8.11 pipeline (Palmer and Stajich 2017), with parameters for the closely related organism Fusarium. Funannotate training files were generated utilizing two genome predictor programs, GeneMark-ES (Besemer et al. 2001) and AUGUSTUS (Stanke et al. 2006). Further analysis included the application of BUSCO v4.1.2 (Simão et al. 2015) to evaluate genome completeness and to compare with other genomes based on single-copy benchmarking universal single-copy orthologs (BUSCO). Overall, we found completeness scores of 97.9% in COM, W2, and HAR and 98.0% in STO and PRU. Annotated protein files underwent orthologous gene cluster analysis to find unique differences among the five isolates through OrthoVenn2 parameters e-value 1e-2 and inflation value 1.5 (Xu et al. 2019). The five isolates shared 13,711 orthologous gene clusters (Fig. 2). In isolate COM, there were 14,576 orthologous gene clusters and 334 singletons, whereas W2 had 14,704 clusters and 186 singletons. HAR had 14,852 clusters with 568 singletons, PRU had 15,035 clusters with 181 singletons, and STO had 14,819 clusters with 199 singletons.

    Table 2. The antiSMASH summary of secondary metabolites from our five genomes

    Fig. 2.

    Fig. 2. Venn diagram of the shared and unique protein-coding genes or pseudogenes from the five Fusarium oxysporum f. sp. niveum (FON) isolates. Of the five isolates, COM and HAR are the only ones to have unique genes. The figure was generated from the OrthoVenn2 online software.

    Download as PowerPoint

    The five draft genomes described here are unique and represent different spatial and temporal representative genomes of a major watermelon pathogen. The assembled genome and annotated gene resources of our five drafts will be helpful to study FON's evolution, environmental adaptation, and overall pathogenicity. In addition, the information provided is beneficial to elucidate the genetic differences that may correlate with virulence variations of FON isolates. Finally, the comparative analysis of the five isolates identified genes that were uniquely present or absent among genomes that will provide insights into the increasing pathogenicity of FON.


    Bio sample draft genome accessions: SAMN30676175, SAMN30676176, SAMN30676177, SAMN30676178, and SAMN30676179. Sanger sequencing accession: PRJNA878855.


    The authors thank Texas A&M AgriLife Excellence fellowships for student aid for V. E. Thomas. V. E. Thomas thanks Sabin Khanal for his technical guidance in this pipeline.

    The author(s) declare no conflict of interest.

    Literature Cited

    Funding: This work was supported by the United States Department of Agriculture HATCH 1021870 project number TEX09714.

    The author(s) declare no conflict of interest.