Genomic insights into Ceratobasidium sp. associated with vascular streak dieback of woody ornamentals in the United States using a metagenomic sequencing approach

Abstract

Woody ornamentals are integral to urban landscapes and play important roles in habitat restoration and ecological conservation, yet their national and international trade facilitates the spread of plant diseases with significant ecological and economic consequences. Vascular streak dieback (VSD) recently emerged on woody ornamentals in the United States and was found to be associated with the fungal pathogen Ceratobasidium sp. (Csp), but little is known about its genomic diversity and associated microbial communities. We thus applied metagenomic sequencing to 106 symptomatic samples that had tested positive for Csp and had been collected from 34 woody ornamental species in seven states. Taxonomic profiling identified Csp as the only putative pathogen of which we recovered 17 high-quality draft genomes. Phylogenomic and pangenome analyses revealed that U.S. Csp isolates form a tight genetic cluster, distinct in gene content from C. theobromae, a pathogen of cacao, avocado, and cassava in Southeast Asia. Comparative analyses highlighted gene content differences, including candidate effectors and secondary metabolite clusters, which may underlie host interactions and offer diagnostic targets. These findings provide the first genomic insights into the U.S. Csp population, suggest the recent introduction of a single genetic lineage with a broad host range, and establish a framework for improved detection, monitoring, and management of VSD in woody ornamentals.

Importance: Identification of the pathogen that causes an emerging disease, be it of humans, animals, or plants, is a prerequisite to develop effective treatment and/or management practices and to try to control the disease outbreak to prevent further pathogen spread. Vascular streak dieback (VSD) is an emerging disease of ornamental bushes and trees in the United States. Identification of the pathogen has been hindered by the difficulty in growing the fungal pathogen found to be associated with diseased plants in pure culture. Here, we succeeded in sequencing the DNA of the likely pathogen directly from plant tissue or from the fungal mass growing out of collected plant tissue. The sequences were assembled into genomes, which allowed us to precisely identify the pathogen, compare it to related pathogens of other plants, and predict how it causes disease. These results can now be used to inform management and control of VSD.

Keywords: emerging disease; metagenomic sequencing; pangenome analysis; phylogenetics; woody ornamentals.

Fig 1

Representative VSD symptoms observed in Csp-positive woody ornamental hosts: (A) Acer rubrum “October Glory”; (B) Cercis canadensis “Flame Thrower”; (C) Cornus kousa “Greensleeves.” Symptoms include vascular discoloration, foliar chlorosis progressing to necrosis, and terminal dieback.

Fig 2

Geographic distribution of Csp sample collection sites across the United States. (A) Map of the United States with the states from which samples were obtained highlighted in purple. The numbers indicate the number of samples collected from each state. (B) A map of Virginia showing counties and the number of samples that were collected.

Fig 3

Boxplots of log₁₀-transformed normalized read counts (reads per million) for plant, fungal, Oomycota, and bacterial taxa, comparing mycelium-derived (left) and plant-derived (right) samples.

Fig 4

Boxplots of log₁₀-transformed normalized read counts of Csp comparing Kraken2 (red) and Minimap2 (blue) across mycelium-derived (left) and plant-derived (right) samples. In both sample types, Minimap2 shows a slightly higher median and tighter distribution of read counts than Kraken2.

Fig 5

Two-panel boxplots of log₁₀-transformed normalized reads for the top 10 taxa identified by Kraken2. (A) Fungal species and (B) bacterial species, each comparing mycelium-derived (left) and plant-derived (right) samples. Boxes show the interquartile range with the median (thick line); whiskers extend to 1.5 × IQR; points are per-sample values. Species marked with an asterisk (*) in the legends were not supported by read-mapping or sourmash and are therefore likely Kraken2 false positives.

Fig 6

(A) Maximum likelihood phylogenetic tree constructed using 76 single-copy orthologous genes from assembled U.S. Csp genomes, publicly available Ceratobasidium genomes, and R. solani isolates used as an outgroup to root the tree (best model: VT + F + I + I + R6). Branch tips are labeled with the sample IDs. Bootstrap support values are shown at each node. (B) Core genome phylogenetic tree of Csp genomes recovered from woody ornamental hosts in the United States (best model: VT + F + I + I + R6). A total of 1,473 core genes were used to construct the tree, which was rooted on genomes of Ct isolates from cacao and cassava in Southeast Asia. Branch tips are labeled with the sample IDs.

Fig 7

Maximum‐likelihood phylogeny constructed under a GTR + Γ model from 32 Ceratobasidium isolates using 1.32 million SNPs identified by the NanoCaller tool by aligning reads against the Ct LOA1 genome, which was chosen as the root. Ultrafast bootstrap values (≥50%) are shown at nodes.

Fig 8

Pangenome analysis. (A) Composition of core and accessory orthogroups and singleton genes in the Ceratobasidium/Rhizoctonia pan-genome. (B) Gene accumulation curve (orthogroup count vs. number of genomes): each bar represents how many gene families are present when considering a specific number of genomes. The numbers decrease as more genomes are included, with a sharp increase near the rightmost bars (particularly at 15 genomes), reflecting the core gene families shared by nearly all genomes. This curve illustrates how the pangenome grows and highlights both unique and shared gene content across genomes. (C) UpSet plot showing intersections of orthogroups across 15 Ceratobasidium and R. solani genomes. Bar heights indicate the number of orthogroups for each presence/absence pattern (only intersections with ≥50 OGs are shown), with the connected dots below identifying the genomes included in each intersection. (D) Composition of core, accessory, and singleton genes per genome.