Genomic Resources of Four Colletotrichum Species (C. fioriniae, C. chrysophilum, C. noveboracense, and C. nupharicola) Threatening Commercial Apple Production in the Eastern United States

The genus Colletotrichum comprises nine major clades containing 252 species and 15 major phylogenetic lineages, also known as species complexes (Cannon et al. 2012; Talhinhas and Baroncelli 2021). Colletotrichum spp. are one of the top fungal plant pathogens (Dean et al. 2012), causing anthracnose and pre- and postharvest fruit rot on a wide range of economically important crops (Cannon et al. 2012). Apple (Malus domestica Borkh.) is a major agricultural product cultivated in temperate regions worldwide and significantly contributes to the human nutrition and global economy. Orchards can suffer devastating losses by bitter rot of apple, which is a serious disease caused by several species of Colletotrichum (Sutton 2014). Bitter rot was also first described as a major postharvest pathogen in the Mid-Atlantic region, and it was found that C. fioriniae caused bitter rot in commercial apple storage facilities (Kou et al. 2014), leading to 2 to 14% unmarketable fruit. The incidence of apple bitter rot in the Mid-Atlantic United States has risen in the last 15 years and shifted from minor to highly problematic, with consistent rot problems of up to 100% loss in organic systems and over 80% in conventional orchards with inadequate fungicide spray programs (Khodadadi et al. 2020). Symptoms on apple fruit appear as small flat and circular spots that turn into large sunken lesions with concentric rings covered with salmon to black-colored masses of numerous asexual spores.

Dominant species causing apple bitter rot in the Mid-Atlantic are C. fioriniae, from the Colletotrichum acutatum species complex (CASC), and C. chrysophilum and C. noveboracense, from the C. gloeosporioides species complex (CGSC) (Khodadadi et al. 2020; Martin et al. 2021). C. fioriniae is the dominant species causing apple bitter rot in the Northeastern and Mid-Atlantic regions of the United States (Khodadadi et al. 2020; Kou et al. 2014; Martin and Peter 2021; Martin et al. 2021; Munir et al. 2016; Wallhead 2016). While C. chrysophilum was first identified as an independent novel evolutionary lineage on banana (Vieira et al. 2017) and cashew (Veloso et al. 2018), in 2020, it was detected as the second most dominant species causing apple bitter rot in the Mid-Atlantic (Khodadadi et al. 2020). Using multi-loci phylogenetic analyses, C. noveboracense MB 836581 was identified as a novel species in the CGSC causing apple bitter rot in the Mid-Atlantic region (Khodadadi et al. 2020). C. nupharicola is a sister group to C. fructicola and C. noveboracense. C. nupharicola is easily distinguished within the CGSC morphologically and is reported to cause bitter rot on apple (Leonberger et al. 2019). Species within the CASC and CGSC differ in crucial characteristics such as lifestyle, temperature optimum, pathogenicity, and fungicide sensitivity, indicating important differences in their genetic makeup. Correct species-level detection and identification is critical for developing effective disease control practices in apple orchards in the United States. In this study, we performed whole-genome sequencing, de novo assembly, and annotation of two isolates of C. fioriniae (ACFK5, ACFK16), three isolates of C. chrysophilum (AFK154, AFK26, PMKnsl-1), and two isolates of C. noveboracense (PMBrms-1, AFKH109), collected from bitter rot–infected apple fruit in New York, Pennsylvania, and Virginia (Khodadadi et al. 2020). The C. noveboracense Coll940 isolate was collected from Juglans nigra in Oklahoma, and two isolates of C. nupharicola (CBS470 and Coll922) were obtained from yellow waterlily (Nuphar lutea) in Washington and New Jersey, respectively (Doyle et al. 2013).

Total genomic DNA was extracted based on the protocol described by Yelton et al. (1984), with some modifications. We used fungal mycelia collected from 5-day-old single-spore cultures grown on potato dextrose agar at room temperature in complete darkness. Except for C. noveboracense isolate AFKH109 (European Nucleotide Archive [ENA] H1091258; Mycobank as 386581), construction of libraries, whole-genome sequencing, and quality control were performed by the Beijing Genomics Institute (BGI), using the DNBSEQ short read platform for 350-bp libraries with the paired-end 150-bp sequencing strategy, as described in the DNBseq De Novo service overview manual (BGI). Sequence adapter trimming and quality control were conducted with Trimmomatic v0.39 and a phred score cut-off at 20 (Bolger et al. 2014). Genome assemblies were constructed using SPADES 3.15.2 (isolates Coll940 and AFK154) and megahit 1.2.9 (isolates CBS470, ACFK5, ACFK16, PMKnsl-1, AFK26, PMBrms-1, and Coll922).

For C. noveboracense AFKH109, the sequencing library was prepared using the Ion Torrent Ion Xpress Plus fragment library kit (ThermoFisher Scientific) and the Ion Plus fragment library kit (ThermoFisher Scientific). Genomic DNA was fragmented to 400 bp, using the Covaris M220 Sonicator (D-Mark Biosciences Inc.). Library fragment size selection was performed using the Pippin Prep (Sage Science) with the 2% DF Marker L cassette, with the target size set to 300 to 450 bp. Final library amplification was performed using 12 amplification cycles. A total of 37 pM of sequencing library was loaded onto the S5 Chef instrument (ThermoFisher Scientific), using the Ion 530 kit (ThermoFisher Scientific), and was sequenced on the Ion Torrent S5 sequencer (ThermoFisher Scientific) (Tremblay and Bilodeau 2022). Sequencing reads were checked for their quality and to provide graphical guidance for filtering, trimming, and reformatting, using PRINSEQ-LITE v.0.20.4 and PRINSEQ-graphs v.0.6 software packages (Schmieder and Edwards 2011). Identification and trimming of sequencing adapters on raw reads were performed using the AdapterRemoval version 2 package (Schubert et al. 2016). Following adapter removal, reads were trimmed using a sliding window approach (window size 7, shift 1 base). From the 3′ end, read bases with the average quality score Q of 7 bases above Q = 10 were kept. Reads were additionally filtered by complexity score and minimum length (100 bp, DUST complexity threshold = 7). Genome assembly of the processed reads was performed using Newbler v.2.6 with default parameters (Margulies et al. 2005). Genome annotation was performed as described by Giroux and Bilodeau (2020), with modifications of the reference organisms retrieved from the National Center for Biotechnology Information used for initial gene predictions within the MAKER pipeline (C. gloeosporioides GSE41844, Augustus species = Fusarium graminearum, repeat_protein modeling using C. fioriniae GCA_000582985.1, C. fruticola GCA_000319635.1, C. graminicola GCA_000149035.1, C. orchidophilum GCA_001831195.1) and when using BUSCO (benchmarking universal single-copy orthologs) to produce a trained, species-specific hidden Markov model for use in MAKER (lineage Sordariomyceta).

For all genome annotations reported in this study, the completeness of the genome assemblies was assessed using BUSCO v5.3.2 with the fungi_odb10 lineage dataset (created June 28, 2021) (Simão et al. 2015), and the genome annotation metrics were calculated using AGAT v0.8.1 (Dainat 2022). Summaries of the assembly and annotation metrics for the species described in this announcement are provided in Table 1. An average of 16,621 gene models per species assembly were predicted. While the genome assembly metrics varied for each of the Colletotrichum species sequenced, all genomes had BUSCO completeness predictions over 95%, except the C. noveboracense AFKH109 genome, which was 87.2%. The GC content of the genomes sequenced ranged from 49.33 to 53.94%, with C. fioriniae and C. chrysophilum having the upper and lower ranges, respectively. The estimated genome sizes ranged from 49.33 to 59.08 Mb. The estimated sequencing coverage ranged from 46× to 103×, with the lowest and highest coverages obtained for C. nupharicola and C. fioriniae isolates, respectively.

Table 1. Colletotrichum species draft genome assembly and completeness summary statistics

	C. nupharicola		C. fioriniae		C. chrysophilum			C. noveboracense
Statistics	CBS470	Coll922	ACFK16	ACFK5	PMKnsl-1	AFK26	AFK154	Coll940	AFKH109	PMBrms-1
Genome assembly metrics
Genome size (Mb)a	58.75	51.67	49.33	49.44	56.06	55.97	57.85	58.17	58.01	59.08
GC%b	51.03	50.76	52.36	51.94	53.94	52.61	53.21	52.34	52.2	52.59
Coveragec	87×	46×	103×	103×	91×	91×	78×	77×	99×	86×
Contigs	3,696	39,421	514	169	1,954	1,029	1,702	1,007	4,510	2,959
N50 contig length (kb)	137,264	34,056	742,368	688,115	255,107	288,101	268,763	168,995	27,185	176,901
Longest contig length (kb)	938,454	597,000	3,179,133	3,363,293	1,344,106	1,480,977	1,515,687	1,555,225	139,303	1,645,399
Genome annotation metrics
BUSCO completeness (%)	99.2	99	98.8	98.6	98.7	98.9	99.1	99	87.2	99.1
Predicted gene modelsd	13,838	14,983	12,747	12,254	13,816	14,530	14,434	14,420	15,301	13,803
Mean gene length (bp)e	1,510	1,402	1,553	1,542	1,513	1,504	1,485	1,505	1,559	1,517
Genome covered by genes (%)	34.52	25.07	38.97	37.09	36.19	37.97	35.95	36.26	40.06	34.42
Number of genes	13,425	14,546	12,377	11,886	13,409	14,123	13,999	14,011	14,903	13,397

^aGenome size calculated from the final processed reads total sequence length.

^bGC% calculated from final assembly of processed reads.

^cGenome sequencing coverage calculated using the mean read length of the final processed reads (286.74), the calculated genome size, and the number of read sequences using the coverage/read count calculator (https://stephenturner.shinyapps.io/covcalc/).

^dPredicted gene model value includes predicted transfer RNA (tRNA).

^eMean gene length was calculated from coding sequences, excluding predicted tRNA.

View as image HTML

All strains were deposited to the Agricultural Research Service Culture Collection (NRRL) at the National Center for Agricultural Utilization Research in Peoria, Illinois, while those for AFKH109 were deposited in the ENA as H1091258. The genome sequences for the above strains have been deposited in GenBank. This work provides new genomic resources for Colletotrichum spp. that were previously not available and significantly contributes to the expansion of existing resources for the Colletotrichum genus, whose species complexes make designing species-specific diagnostic assays challenging. Thus, the whole-genome sequencing data, obtained from these dominant disease-causing species, will provide the genome resources needed to identify species-specific gene targets for molecular marker development to complement existing diagnostic and detection tools. It is envisioned that these tools will also facilitate many comparative genomics studies that aim to understand the nature of host-pathogen interactions and factors impacting disease resistance in the host.

Data Availability

Genome accession numbers in the National Center for Biotechnology Information GenBank database for the sequenced isolates listed in Table 1 are as follows: CBS470: GCA_026319135.1, Coll922: GCA_026319225.1, ACFK16: GCA_026319165.1, ACFK5: GCA_026319145.1, PMKnsl-1: GCA_026319215.1, AFK26: GCA_026319265.1, AFK154: GCA_026319245.1, Coll940: GCA_026319155.1, AFKH109: GCA_946151405.1, PMBrms-1: GCA_026319125.1.