Background & Summary

Apple and stone fruit trees are integral to Ontario’s agriculture economy. The primary areas for apple production in Ontario are located along the shores of Lake Ontario, Lake Erie, Lake Huron and the Georgian Bay. Recent years have seen a concerning downturn, with apple cultivation decreasing to 6,414 hectares in 2023, a 28% drop from the 8,903 hectares recorded in 20021. Similarly, the Niagara region, responsible for over 90% of Ontario’s stone fruit production, has seen a decrease in cultivation area, down 37% from 5,080 hectares in 2007 to 3,192 hectares in 20231.

Since 2016, both apple and stone fruit trees have faced alarming levels of mortality in Ontario, with reports of up to 42% mortality in apples and up to 72% in stone fruits, particularly affecting trees aged between 2 to 10 years2,3,4,5,6,7,8. Symptoms of decline vary, from leaf discoloration and canker development in apples to progressive dieback in stone fruits. These symptoms collectively define Fruit Tree Decline (FTD), with Rapid Apple Decline (RAD) specifically identified in apples9. The decline observed in apple and stone fruit trees extends beyond Ontario, echoing in other regions across North America10. While the exact cause of FTD remains elusive, the emergence of new pathogenic fungi2,3,4,5,6,7,8,11 and viruses12,13,14, as well as the re-emergence of known pathogens15,16,17, compounded by unexpected abiotic stressors such as drought stress, heat waves, floods, winter injury, and nutrient deficiency, are thought to play a significant role in its development9,18,19.

Fungal pathogenicity in a plant depends on multiple factors, including the specific traits of the pathogen, the plant’s defenses, and the conditions within their shared environment20. While weak pathogens may individually induce negligible FTD symptoms, the severity of the disease can escalate under the influence of other biotic or abiotic stresses. Understanding these interactions and the underlying genetic mechanisms is key for effective disease management.

Sequencing fungal genomes lays the foundation for exploring their ecological niche, evolutionary patterns, and disease-causing capabilities21. By deciphering the genetic makeup of these fungi, we aim to improve disease diagnosis, elucidate pathogenic mechanisms, and support the development of targeted strategies for disease prevention and management22. In the present study, we sequenced, de novo assembled and annotated the genomes of 12 fungi associated with FTD and RAD syndromes in Ontario, Canada, spanning a spectrum from pathogens to weak pathogens and non-pathogens/endophytes. By identifying genetic differences between pathogens and non-pathogens, we can gain a stronger understanding into the genetic determinants driving pathogenic evolution23. This knowledge is required for developing predictive models for fungal pathogen emergence and designing effective control measures.

This study provides the first assembled and annotated genomes for the Ascomycetes species Diplodia intermedia, Diatrype stigma, and Nothophoma quercina. These represent the only genome assemblies currently available for the Diatrype and Nothophoma genera. We also present high-quality and well-annotated reference genome for Diplodia seriata, Didymella pomorum and Neofusicoccum ribis. These new genomic resources significantly enhance our ability to study fungal pathogenicity in agricultural systems and inform the development of targeted control measures against pathogens. Moreover, the availability of these assemblies enables downstream functional analyses, including the identification of effector proteins, which are key virulence factors used by pathogens to facilitate host colonization24. Since effector characterization can inform resistance breeding and integrated disease management strategies, this work supports future discovery efforts and comparative genomic studies aimed at improving the resilience and sustainability of fruit production systems.

Methods

Sample collection, fungal isolation and identification

Between 2018 and 2021, wood samples were collected from 25 apple, 30 apricot, six peach, and six nectarine trees showing extensive symptoms of tree fruit decline across 15 commercial orchards in Ontario. Tree mortality rates reached up to 42% in apples and up to 72% in stone fruits, with the highest impact observed in trees between 2 and 10 years of age across the sampled orchards. Small sections of diseased wood (1 cm long) underwent surface sterilization with 70% ethanol for 30 seconds, followed by treatment with 1% NaClO for 20 minutes. After rinsing thrice in sterile water, the wood samples were placed on 2% potato dextrose agar (PDA, Difco™, Franklin Lakes, NJ, USA) with kanamycin (50 mg/L) and incubated at 22 °C for 5 days in darkness. Fungal colonies consistently isolated underwent hyphal-tip transfer to individual PDA plates, then incubated at 22 °C for 7 days in darkness. Purified isolates were morphologically classified into morphotypes before molecular identification.

Genomic DNA was extracted from the mycelium grown on PDA of 7-day-old cultures of one representative isolate from each morphotype group using the Plant/Fungi DNA Isolation Kit (Norgen Biotech, ON, Canada, cat. no. 26200), following the manufacturer’s instructions with slight modifications. Specifically, fungal tissue was vortexed for 15 minutes with 1 mm glass beads in 500 μL of lysis buffer supplemented with 1 μL of RNase A, followed by incubation at 65 °C for 10 minutes. Subsequently, 100 μL of Binding Buffer I was added, the mixture was thoroughly mixed, and incubated on ice for 5 minutes before centrifugation at 10,000 rpm to separate the lysate from the beads and biomass. Resin drying during the column wash was achieved by spinning for 10 minutes at 14,000 rpm. Finally, DNA was eluted at 10,000 rpm for 2 minutes.

Polymerase chain reaction (PCR) was performed in a VWR PCR Thermal Cycler XT96 Gradient (VWR, USA) under the following conditions: 98 °C for 30 sec; 35 × (98 °C, 10 sec; 60 °C, 30 sec; 72 °C, 30 sec); 72 °C, 2 min. The internal transcribed spacer (ITS), translation elongation factor 1-α (EF1-α) and β-tubulin (TUB2) gene regions were each amplified in separate PCR reactions using the primers ITS1/ITS425, EF1-728 F/EF1-986R26 and Bt2a/Bt2b27, respectively. Each 30 μL reaction contained 9 μL of nuclease-free Water, 15 μL of Q5 Hot Start High-Fidelity 2X Master Mix (New England Biolabs, ON, Canada, cat. no. M0494S), 3 μL of extracted genomic DNA and 1.5 μL (10 μM stock) of the appropriate fungal-specific primers. Reactions were performed with negative controls containing no DNA.

The quality of the PCR products was examined using electrophoresis in 1% agarose gel. Sanger sequencing was carried out at Genome Quebec’s Sequencing Facility (Montreal, Canada). Sequences were analyzed with the Basic Local Alignment Search Tool (BLAST)28 against the nucleotide database for an early approximation of species classification.

DNA isolation and genome sequencing

Genomic DNA was extracted from fungal mycelium obtained from 12 pure cultures of fungi identified as previously described. These fungi include strains such as Botryosphaeria dothidea, Cytospora paraplurivora, Diaporthe eres, Diatrype stigma, Didymella pomorum, Diplodia intermedia, Diplodia seriata, Neofusicoccum ribis, Nothophoma quercina, and Paraconiothyrium brasiliense, all isolated from symptomatic apple, apricot, and nectarine trees (Table 1). Eleven of the 12 strains were deposited at the Canadian Collection of Fungal Cultures (DAOMC), each assigned a unique identification number as detailed in Table 1. The mycelium was cultured in potato dextrose broth (PDB, Difco™, Franklin Lakes, NJ, USA) for 5 days at room temperature on an orbital shaker (3 g), followed by filtration through sterilized WhatmanTM Qualitative filter paper grade 1 (Cytiva, Marlborough, MA, USA), freeze-drying for 48 hours using a Benchtop Freeze Dry System (Labconco FreeZone® 4.5 Liter, Kansas City, MO, USA), and bead beating using 1 mm glass beads until a powdered form was obtained. Subsequently, the modified DNA extraction protocol, as previously described (Norgen Biotech, Thorold, ON, Canada), was employed.

Table 1 Genome assembly and annotation statistics.
Full size table

Sequencing strategies were designed to species abundance and data requirements. Diplodia seriata FDS-637, one of the most abundant species in the study, was sequenced using both PacBio and Illumina platforms to generate high-quality long-read data for comprehensive genome assembly. Two additional abundant species, Cytospora paraplurivora FDS-564 and Diaporthe eres M63-4 were sequenced using PacBio only, while the remaining nine isolates were sequenced using Illumina, which provided sufficient resolution.

Genomic DNA (gDNA) libraries for Cytospora paraplurivora FDS-564, Diaporthe eres M63-4, and Diplodia seriata FDS-637 were constructed and sequenced at the SickKids sequencing facility (Toronto, ON, Canada). Combinations of barcoded primers and SMRTbell adapters were used to prepare the samples, which were afterward pooled. Long reads for genome assembly were generated from a library prepared with 5 µg of unsheared gDNA using the Express Template Prep Kit (v2.0) (Pacific Biosciences, Menlo Park, CA). A post-library size-selection step targeting fragments > 14 kb was performed using the Agilent 4200 TapeStation System (Agilent, Waldbronn, Germany). The final size-selected library was sequenced on a PacBio Sequel® II sequencer using a single 8 M SMRT Cell in continuous long-read (CLR) mode with a 15-hour movie acquisition time. Raw reads were processed using PacBio’s P-filter to remove low-quality reads and adapter sequences.

Illumina gDNA libraries for Diplodia seriata FDS-637 and the remaining nine fungal isolates were prepared and sequenced at the Centre d’Expertise et de Services, Génome Québec, McGill University (Montréal, QC, Canada). Genomic DNA was quantified using the Quant-iT™ PicoGreen® dsDNA Assay Kit (Life Technologies™, Grand Island, NY, USA), and its integrity was assessed on a TapeStation 2200 (Agilent Technologies, Inc. Santa Clara, CA, USA). Libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs, Whitby, ON, Canada) according to the manufacturer’s instructions. Adapters and PCR primers were purchased from Integrated DNA Technologies (IDT, Coralville, IA, USA). Size selection to obtain the desired insert size was performed using sparQ beads (Quantabio, Beverly, MA, USA). Library quantification was carried out using the KAPA Library Quantification Universal Complete Kit (Kapa Biosystems; Wilmington, MA, USA), and average fragment size was determined using a LabChip GX II instrument (PerkinElmer, Waltham, MA, USA). Libraries were normalized, pooled, denatured with 0.02 N NaOH, and neutralized using HT1 buffer. The pooled libraries were loaded at 225 pM on an Illumina NovaSeq 6000 S4 flow cell using the Xp protocol, following the manufacturer’s guidelines. Sequencing was conducted in paired-end mode (2 × 150 bp cycles). A 1% PhiX control library was spiked into the run. Base calling was performed using Illumina Real-Time Analysis (RTA) software (v3), and demultiplexing and FASTQ file generation were done using bcl2fastq2 Conversion Software (v2.20) (Illumina, San Diego, CA, USA).

Genome de novo assembly and annotation

Illumina NovaSeq reads (shown in Table 1) were trimmed with Trimmomatic (v0.38.1)29. The ILLUMINACLIP setting was used to remove adapter sequences, and a sliding window approach (4 bp window with an average quality threshold of Q = 20) was applied to trim low quality base calls. The quality of the reads were analyzed before and after Trimmomatic by FastQC (v0.72). Trimmed reads were assembled and gaps were filled using SPAdes (v3.12.0)30 with the K-mer values 21, 33, 45, 69, 81, 93, 105, and 117.

PacBio reads for Cytospora paraplurivora FDS-564, Diaporthe eres M63-4 and Diplodia seriata FDS-637 were trimmed and assembled using Canu (v2.1.1)29 with default settings, as described by Ilyukhin et al.8. For Diplodia seriata FDS-637, Illumina paired-end reads were aligned to the Canu-assembled PacBio genome using Burrow Wheeler Alignment (BWA (v0.7.17))31 with default settings to generate a SAM file. This file was then converted to a sorted BAM file using SAMtools (v1.14)32, which served as input for Pilon (v1.23)33 to correct base-level errors and enhance the overall accuracy of the Canu assembly. Quast (v5.0.2)34 was used to determine assembly statistics and genome coverage (Table 1).

The Funannotate (v1.8.18)35 pipeline was used for genome annotation following the recommended instructions (https://funannotate.readthedocs.io) for gene prediction. All assemblies were cleaned and masked using Funannotate default settings, which utilize Tantan (v40)36 (https://gitlab.com/mcfrith/tantan) for repeat masking. Subsequently, Benchmarking Universal Copy Orthologs (BUSCO (v2.0.0))37 was run on all genomes using the dikarya_odb9 dataset and its results were used to train GlimmerHMM (v3.0.4)38, SNAP (2006-07-28)39, and Augustus v3.3.340 for ab initio gene prediction. Gene predictions were generated using the aforementioned trained gene models and GeneMark-ES (v4.59)41. Weights were assigned to each prediction, based on the source and the strength of each prediction, and passed into EVidenceModeler42, which was used to generate a consensus gene model. Finally, tRNAscan-SE43 was used to generate tRNA predictions.

Functional Annotation was performed using the funannotate wrapper through which the results of numerous analyses were combined. Secondary metabolic gene clusters were predicted using antiSMASH (v6.0)44. The number of each category of biosynthetic gene clusters, found by antiSMASH are listed in Table 2. Interproscan (v5)45, eggnog-mapper (v2)46 (utilizing eggNOG (v5.0)47), HMMer3 searches against Pfam (v32.0)48, dbCAN (v8.0)49; and diamond searches against MEROPS (v12.0)50, uniprot and CAZYmes51 databases were used to annotate protein functional domains. Phobius (v1.01)52 was used to predict protein secretion and transmembrane domains, while signal peptides were predicted using SignalP (v6.0)53. The number of annotations added by these steps is summarized in Table 3.

Table 2 Counts of different biosynthetic gene clusters found from antiSMASH, in each genome.
Full size table
Table 3 Protein counts with functional annotations from each of the listed databases.
Full size table

Genomes were visualized using Circos diagrams to assess quality (Fig. 1). GC skew and GC content were calculated over a sliding window of size 100 kbp. Counts of the canonical telomeric sequences”TTAGGG”, and its reverse complement “CCCTAA” were found over a 50 kbp window. Spikes at the ends of scaffolds likely indicate the presence of telomeres. Scaffolds with telomere count spikes on either side can be found in Paraconiothyrium Brasiliense strain M42-189, Nothophoma quercina strain M97-236 and Diplodia seriata strain FDS-637, indicating that some of the scaffolds in these assemblies may be chromosomes sequence telomere-to-telomere. Functional annotations was assessed by evaluating the number of Gene Ontology (GO) terms (Fig. 2 and Tables S1–S3) and the Clusters of Orthologous Groups of proteins (COGs) present in each annotation, as shown in Fig. 3 and Table S4. Relatively similar percents of different COG groups and GO annotation terms were found between genomes of the same species.

Fig. 1
figure 1

Circos plots for all fungal assemblies in this study, each with 7 rings lettered A to G. Only scaffolds greater than 100 kbp are shown. Ring A illustrates scaffolds in the assembly, with smaller ticks are placed at intervals of 100 kbp, and larger ticks are placed at intervals of 1Mbp. Ring B highlights the coding sequences (CDSs) on the forward strand, while Ring C does the same for the reverse strand and Ring D depicts the tRNAs. Ring E presents the direction and magnitude of GC skew, with black fill signifying an increase in GC skew relative to the average of the entire genome, while gray represents a decrease. Similarly, Ring F presents the GC content, with olive fill representing an increase in GC content over the average of the assembly, while magenta represents a decrease. Ring G plots the counts of telomeric sequence.

Full size image
Fig. 2
figure 2

A series of plots representing the number of GO ontology terms in each annotation. In order of top to bottom, the first plot shows a line graph for the number of instances of the three different categories of GO terms: GO process, referring to the biological processes in which a particular gene may play a part; GO function, the activity the product of the gene may possess; and GO component, the location where the gene product is active. The next three plots (in order: GO function, GO component, and GO process) represent the relative percentages of each of the top 15 most common GO terms from their respective category, while terms not in the top 15 are summarized in the “other” group.

Full size image
Fig. 3
figure 3

Plot of COGs in each annotation. Each bar represents the percent number of each of the COGS (the corresponding y-axis is on the left). The line graph shows the total number of COG annotations in each assembly (with the corresponding y-axis on the right).

Full size image

Phylogenomic analyses

Genomes of the 12 fungal strains sequenced, assembled, and annotated in this study, along with 90 fungal genomes downloaded from NCBI using the get-assemblies tool (GitHub - davised/get_assemblies), were included in the phylogenetic analysis (Fig. 4). All genomes were retrieved by using all genera names of the sequenced organisms in this study as queries, such as “Didymella”, “Paraconiothyrium”, and more. Sixty-one core genes, listed in Table 4, were extracted, aligned and concatenated using the Universal Fungal Core Genes (UFCG) pipeline version 1.0.554. Maximum Likelihood-based phylogenetic analysis of the concatenated sequences was performed using IQ-Tree (v2.2.6)55, with 1,000 bootstrap replicates processed to determine the best-scoring Maximum Likelihood tree. The phylogenetic tree was rooted with Venturia oleaginea Yun35.

Fig. 4
figure 4

Maximum Likelihood phylogenetic tree generated using a concatenated alignment of 61 core genes extracted from the genomes of Dothideomycetes and Sordariomycetes species sourced from NCBI. The 12 fungal genomes obtained in this study are highlighted in bold. The tree was rooted to Venturia oleaginea Yun35. The numbers at each node represent bootstrap support, expressed as percentages.

Full size image
Table 4 List of the 61 core genes extracted from fungal genomes using the UFCG pipeline to generate the phylogenetic tree, along with their corresponding functions and identifiers sourced from the Saccharomyces Genome Database (SGD, www.yeastgenome.org) and UniProt (www.uniprot.org).
Full size table

Data Records

The datasets have been deposited in the Sequenced Read Archive (SRA) under accession numbers: SRX14291869, SRX14155424, SRX14046716, SRX14003295, SRX13989866, SRX13979984, SRX13961038, SRX13940527, SRX13921100, SRX13894123, SRX13883208, SRX13877378, SRX1344580056. All assemblies and annotations can be found on NCBI GenBank under BioProject PRJNA79001357, with the following accession numbers: JAJVDB00000000058, JAJSPL00000000059, JAKOOP00000000060, JAKNSF00000000061, JAKJXP00000000062, JAKJXN00000000063, JAKEKT00000000064, JAJVDA00000000065, JAJVCZ00000000066, JAJVDC00000000067, JAKIXB00000000068, JAKJXO00000000069.

Technical Validation

BUSCO (v5.8.2) employing the Ascomycota_odb12 dataset was used to assess whether core genes were adequately captured in the genomes sequenced. All genomes were of high quality (Table 5) and had greater than 97% of BUSCOs complete.

Table 5 Benchmarking Universal Copy Orthologs (BUSCO) assessment results.
Full size table