Branches of Picea glauca were collected from the trees with typical symptoms of CC in the Niagara region, Ontario, Canada, in 2020. The isolates were first obtained using the single spore isolation technique (Phookamsak et al. 2015). To isolate species of Cytospora from plant tissue, small pieces (0.5–1 cm) of cankered wood were surface sterilized with 70% ethanol for 30 s, following sterilization with 0.5% sodium hypochlorite for 2 min, rinsed three times with sterile water, and plated on malt extract agar (MEA). Purification of Cytospora-like colonies was performed by transferring hyphal tips to new MEA plates. The representative isolate obtained from necrotic plant tissue clearly resembling a Cytospora colony was selected for further analysis. The isolate EI-19(A) was identified based on morphological characteristics and sequence data analysis. To correctly identify the isolate EI-20, both detailed morphological and phylogenetic analyses were conducted. The specimens (branches with fruiting structures, wood samples, and living cultures) were deposited into the Herbarium of the Nature Research Centre (BILAS), Institute of Botany, Vilnius, Lithuania, the Canadian National Mycological Herbarium (DAOM), and the Canadian Collection of Fungal Cultures (DAOMC), Ottawa, Canada.

The description of the new species was carried out using the pure culture of the representative isolate. Radial colony growth and color were assessed after 7 and 14 days of fungus incubation under room temperature in the dark on MEA, respectively. Morphological characterization and measurements of asexual reproductive structures (conidiomata (n = 10), conidiophores (n = 20), and conidia (n = 50)) were performed with dissecting (AmScope SE306R-PZ) and compound (AmScope B120C-E5) microscopes. Pictures were taken with a 12 MP digital AmScope camera MD1200A supplied with AmScopeX software (AmScope, Irvine, CA, USA).

Total genomic DNA (gDNA) was extracted from 8-10-day-old pure cultures using the DNeasy PowerSoil Pro Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. The internal transcribed spacer (ITS) region was amplified with the primer pair ITS1/ITS4 (White et al. 1990) using a C-1000 thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) under the conditions described in the references for the region. The quality of PCR products was examined using electrophoresis in 1% agarose gel. Sanger sequencing was carried out at the Genome Quebec Innovation Centre (Montreal, QC, Canada). The partial protein-coding gene (act, rpb2, tef1-α, tub2) sequence data were obtained from assemblies of the strains EI-19(A) and EI-20. The available sequences of the strain C. piceae CFCC 52842 were run against the strains’ genome assemblies to locate corresponding gene regions with the BLASTn search algorithm. The nucleotide sequences were further retrieved using BioEdit v.7.0 (Hall 1999).

The initial identification was performed employing the BLASTn tool against the GenBank nucleotide database of the National Center for Biotechnology Information (NCBI). Sequence data of the related reference strains (Lin et al. 2024) were downloaded from the GenBank database. The sequences were initially aligned employing CLUSTAL-X2 v.2.1 (Thompson et al. 1994). Some characters were trimmed from both ends of the alignments to approximate the size of the obtained sequences to those included in the dataset with MEGA-X (Kumar et al. 2018). Phylogenetic analyses were executed using randomized accelerated maximum likelihood (RAxML) v. 8.0 (Stamatakis 2014) for maximum likelihood (ML) analysis. Bayesian posterior (BP) probabilities were defined using MrBayes v.3.2.7 (Huelsenbeck and Ronquist 2001) with the TrEase web server (Mishra et al. 2023). The ML analysis was performed using the general time-reversible (GTR) substitution model with a gamma-distributed rate of heterogeneity and a proportion of invariant sites selected with ModelTest-NG v.0.1.7 (Darriba et al. 2020). The statistical support values were estimated with bootstrapping of 1,000 replicates (Hillis and Bull 1993). The GTR model was also chosen for the BI analysis. The Markov chain Monte Carlo (MCMC) algorithm was used to estimate Bayesian posterior probabilities (BPP). Six simultaneous Markov chains were run for 1,000,000 generations. A burn-in was implemented by discarding the first 30% of generated trees. The phylograms were visualized using FigTree v. 1.4.4 (Rambaut 2018). The newly generated sequences were deposited in GenBank (Table 1). The master alignment used in the analyses was submitted to TreeBase (www.treebase.org; ID: S31874).

The obtained gDNA of the strains EI-19(A) and EI-20 was quantified employing the Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies, Carlsbad, CA, USA). Libraries were prepared with the NEBNext Ultra II DNA Library Preparation Kit for Illumina (New England Biolabs, Ipswich, MA, USA). Whole genome sequencing was performed using the NovaSeq 6000 platform with a 150 bp pair-end sequencing strategy. This option was selected as being more cost-effective for obtaining draft genome sequences compared to third-generation sequencing. All procedures were performed at the Genome Quebec Innovation Centre (Montreal, Canada). The raw reads were quality-filtered using Trimmomatic v. 0.38.1 (Bolger et al. 2014) with the implementation of Nextera (pair-ended) and sliding window settings. The obtained reads were assembled into scaffolds using SPAdes v. 3.14.1 with K-mer values 21, 33, 45, 57, 69, 81, 93, 105, and 117. The genome assembly characteristics were assessed using QUAST v. 5.2.0 (Mikheenko et al. 2018). Repeat sequence content was estimated with RepeatMasker v. 4.1.5 (Smit et al. 2013), employing the Repbase library of repeats for fungi (Bao et al. 2015). The tRNA regions were predicted using the tRNA scan tool v. 0.4 (Lowe and Eddy 1997). The completeness of assembly was estimated with BUSCO v.5.7.1 (Simão et al. 2015) based on the dataset ascomycota_odb10 (Manni et al. 2021).

Gene prediction from assemblies of the strains EI-19(A) and EI-20 was performed using Augustus v. 3.4.0 (Stanke and Morgenstern 2005) with the gene models of C. mali 03-8 employed to train the tool. For consistency, the genes of other strains included in the analysis were predicted using Fusarium graminearum as a model organism with default parameters. The eggNOG mapper tool v. 2.1.9 (Huerta-Cepas et al. 2019) was employed for functional annotation of the genes predicted. The annotation of CWDEs was performed using dbCAN3 (Yin et al. 2012) with the HMMER-based classification tool. BSGCs were predicted with antiSMASH v. 6.1.1 (Blin et al. 2021). Secretomes were identified using the SECRETOOL pipeline (Cortázar et al. 2014) with default cut-off values and further classified (Miyauchi et al. 2020). EffectorP-fungi v.3.0 (Sperschneider and Dodds 2022) was employed to predict effector proteins. The PHI-base dataset v. 4.17 (Urban et al. 2017), containing 9,056 protein sequences, was used to identify virulence-related genes. A local BLASTp search was performed based on the following parameters: e-value 1 × 10⁻⁵, minimum query coverage per HSP 100%, and percent identity cut off ≥ 65%.

The single-copy orthogroups (SCOs) within all species included in the analysis were identified with Orthofinder v. 2.5.4 (Emms and Kelly 2015). Multiple sequence alignment was performed using MAFFT v. 7.310 (Katoh and Standley 2013). ModelFinder (Kalyaanamoorthy et al. 2017) was used to select a model for each alignment. The phylogenetic tree was produced based on concatenated alignments employing IQ-Tree v. 2.1.3 (Nguyen et al. 2015). Genome sequences were aligned with MashMap v. 2.0 (Jain et al. 2018) and displayed as dot-plot graphs using D-Genies (Cabanettes and Klopp 2018). The RealTime tool implemented in MEGA-X (Kumar et al. 2018) was employed to estimate divergence times. The secondary calibration nodes for the splits of Juglanconis juglandina and Diaporthe citrichinensis (mean: 55.9 MYA, standard deviation: 1 MYA) and Diaporthe citrichinensis and Cytospora chrysosperma (mean: 42.3 MYA, standard deviation: 1 MYA) were used for divergence time estimations based on the TimeTree dataset (Kumar et al. 2017). Gene family expansion and contraction events were revealed using CAFE5 (Mendes et al. 2020) with a p-value of 0.01 as the threshold.

The ML and BP analyses of the combined ITS, act, rpb2, tef1-α, and tub2 sequence data produced phylogenetic trees with highly similar topologies. The best-scoring ML tree with a log-likelihood value of −63932.107826 is shown in Fig. 2. Estimated base frequencies were as follows: A = 0.245570, C = 0.2245652, G = 0.261603, T = 0.247175; substitution rates: AC = 1.770329, AG = 2.617916, AT = 1.724382, CG = 1.002700, CT = 5.573700, GT = 1.000000. The strain of the novel species C. piceicola EI-20 was nested separately, forming a well-supported clade with C. globosa and C. pinastri (95% ML). Notably, C. piceicola was also grouped (80% ML) with other species isolated from conifer trees (C. mougeotii, C. piceae, C. saccardoi, and C. verrucosa).

Figure 2. 

Phylogram of the RAxML tree generated based on the analysis of combined ITS, act, rpb2, tef1-α, and tub2 sequence data of the Cytospora genus. Bootstrap support values for ML ≥ 50% and BP ≥ 0.90 are shown as ML/BP above or below the nodes. Ex-type strains are in bold. Strains obtained in this study are in blue. The tree is rooted to Diaporthe vaccinii (CBS 160.32).

Cytosporaceae Fr., Systema Orbis Vegetabilis 1: 118 (1825)

Cytospora Ehrenb., Sylvae mycologicae Berolinenses: 28 (1818)

 Cytospora piceae X.L. Fan, in Pan, Zhu, Tian, Alvarez & Fan, Phytotaxa 383(2): 188 (2018)

Fig. 3

Description.

Sexual morph : not observed. Asexual morph: Conidiomata pycnidial, immersed in bark, erumpent, ostiolated, with multiple irregularly arranged circular or ovoid locules, 1,050–1,400 μm in diam. Conceptacle absent, ostiole conspicuous, circular, dark gray, at the same level as the disc. Conidiophores semimacronematous, hyaline, filamentous, mainly unbranched or branched at base, elongated, smooth, thin-walled, (16.0–)17.7–22.2(–23.5) μm. Conidiogenous cells enteroblastic, polyphialidic. Conidia abundant, single, hyaline, aseptate, curved, allantoid, thin-walled (5.0–)5.4–6.1(–6.5) × (1.0–)1.2–1.5 μm.

Figure 3. 

Asexual morph of Cytospora piceae A habit of conidiomata on twig of P. glauca B transverse section of conidioma C conidiophores and conidiogenous cells D conidia E adverse and reverse view of seven-day-old culture on MEA. Scale bars: 1 mm (A, B); 5 μm (C, D).

Culture characteristics.

Colonies on MEA initially white, becoming beige with dense aerial mycelium, slow-growing (17 mm in diameter) after 7 days of incubation. Hyphae hyaline, smooth, branched, and septate.

Material examined.

Canada • Ontario, Lincoln, 43°06'38.2"N, 79°19'17.5"W, branches of Picea glauca (Moench) Voss, pycnidia (conidiomata) formed on cankered branches and twigs, April 2020, E. Ilyukhin (BILAS 51883, DAOM 985023, DAOMC 256987).

Notes.

Morphologically, two isolates of C. piceae are very similar, but EI-19(A) has slightly longer conidiophores and conidia than CFCC 52841. Originally, C. piceae was described as a pathogen associated with the canker disease of Picea crassifolia in China (Pan et al. 2018). But the species was isolated from different hosts (including non-conifers) in other parts of Canada based on ITS sequence data (ON352565, PQ671332, PQ671333, PQ666762) available in GenBank.

 Cytospora piceicola Ilyukhin & Markovsk., sp. nov.

MycoBank No: 856480

Fig. 4

Etymology.

The name refers to the host genus, Picea, from which the fungus was first isolated.

Figure 4. 

Asexual morph of C. piceicola A isolation source (cankered branches of P. glauca) B adverse and reverse view of seven-day-old culture on MEA C pycnidia (with locules) on the surface of the colony after 25 days of incubation D conidiogenous cells E conidia. Scale bars: 1 mm (C); 5 μm (D, E).

Holotype.

Canada • Ontario, Lincoln, 43°06'39.0"N, 79°19'15.4"W, isolated from cankered wood (branches) of Picea glauca, April 2020, E. Ilyukhin (holotype BILAS 51884, ex-holotype living culture BILAS 51886=EI-20, isotype DAOM 985024, DAOMC 256985).

Description.

Sexual morph : not observed in culture. Asexual morph: Conidiomata appearing after 25 days of incubation on MEA, rare, pycnidial, solitary, globose to subglobose, dark grey to black when dry, with few ovoid locules, (610–)824–1071(–1380) μm diam. Conidiophores micronematous, hyaline, smooth-walled, reduced to unbranched conidiogenous cells. Conidiogenous cells enteroblastic, phialidic, lageniform, or ampulliform (7.5–)8.8–10.6(–13.0) × (1.0–)1.3–1.7(–2.0) μm. Conidia abundant, relatively small, single, hyaline, aseptate, slightly curved, allantoid, thin-walled (3.5–)3.8–4.9(–5.5) × (1.0–)0.8–1.3(–1.5) μm.

Culture characteristics.

Colonies on MEA white to light brown with short aerial mycelium tufts in the center, becoming darker with age, relatively slow-growing (28 mm in diameter) after 7 days of incubation. Hyphae hyaline, smooth, branched, and septate.

Notes.

Based on ITS sequence data, C. piceicola is 99% similar to C. globosa MFLU:16-2054 (554/559, 3 gaps) and C. pinastri CBS 113.81 (540/545, 0 gaps). But combined multi-gene phylogenetic analysis clearly distinguished C. piceicola from these two species (ML/BI = 95/-). The new species, C. piceicola, differs from C. globosa (4–6.5 × 1–2 µm) and C. pinastri (4–7 × 1–1.3 μm) by having shorter conidia and clearly lageniform or ampulliform conidiogenous cells (Hayova and Minter 2013; Li et al. 2020). The culture characteristics cannot be used for species discrimination as they are either unavailable (C. pinastri) or different media has been used (C. globosa). Thus, C. piceicola is considered a novel species based on both molecular data and morphological characteristics.

Synteny analysis employed in comparative genomics is crucial to understanding molecular-level similarities and differences in species diversity and genome evolution (He et al. 2023). Despite the close phylogenetic relationship between the studied species and the same sequencing strategy used, only 49.08% of synteny blocks in C. piceae EI-19(A) matched those in C. piceicola EI-20 with minor gaps and few inversions (Fig. 5A). The low-scale synteny linked to genome rearrangements might facilitate species-specific evolution of pathogenicity-related genes and contribute to ecological niche adaptation (e.g., host switch) (Sillo et al. 2015; Zeng et al. 2017). The analysis between assemblies of two strains of C. piceae revealed that there was no match for 6.51% of aligned sequences while 93.49% of them were more than 75% similar with large gaps, multiple inversions, and some repeats (Fig. 5B). It indicated a high degree of homology between the two genomes, typical for different strains of the same species (Hao et al. 2021). However, some assembly characteristics of the C. piceae strains appeared to differ (Table 2). The genome of C. piceae CFCC 52841 consisted of 21 scaffolds with N50 of 2.94 Mbp, which indicated better assembly quality than C. piceae EI-19(A) (105 scaffolds with N50 of 1.22). In addition, the assembly of the former contained notably fewer tRNAs (176) compared to C. piceae EI-19(A) (203). 8.82% of repeat classes were detected in the assembly of C. piceae CFCC 52841 based on multiple repeat masking tools (Zhou et al. 2021). This is higher than the average repeat content (5.18%) revealed for a large set of ascomycete genomes sequenced with an Illumina platform (Yu et al. 2022). A nearly complete (99.7%) genome assembly was reported for the strain C. piceae CFCC 52841 based on the fungi_odb9 dataset (Zhou et al. 2021). When employing the more specific ascomycota_odb10 dataset (Manni et al. 2021), assemblies of both strains were found to have lower completeness (C. piceae EI-19(A) (97.6%), C. piceae CFCC52841 (97.4%)). The genome of C. piceicola EI-20 was 98.1% complete. Among the analyzed Cytosporaceae species, the assembly of C. chrysosperma CFL2056 sequenced with a PacBio platform was 93.7% complete, whereas the completeness of the C. leucostoma SXYLt’s genome obtained with Illumina short reads was 98.6%.

Figure 5. 

Dot-plot diagram showing genome sequence alignments of C. piceae EI-19(A) and C. piceicola EI-20(A), and C. piceae EI-19(A) and C. piceae CFCC 52841(B).

Another study reported similar BUSCO estimates (97.8%–99.2%) for a set of the eleven Diaporthe species assemblies obtained with both the second and third-generation sequencing technologies (Hilário et al. 2022). Considering some limitations such as higher error rates (Alkan et al. 2011; Stoler and Nekrutenko 2021), Illumina technology can still produce high-quality reads, allowing for the assembly of nearly complete genomes that are reliable for downstream analyses.

CWDEs are enzymes involved in carbohydrate synthesis or breakdown (Cantarel et al. 2009). They are enriched in many fungal species associated with plant hosts. Carbohydrate enzyme (CAZy) profiles of such fungi can provide clues to reveal their lifestyles and pathogenicity (Zhao et al. 2013). Functional modules of these enzymes are grouped into the classes: glycosyl transferases (GTs), glycoside hydrolases (GHs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), enzymes for auxiliary activities (AAs), and carbohydrate-binding modules (CBMs) (Huang et al. 2018). The rich content of CWDE homologs in the genomes of hemibiotrophic and necrotrophic plant pathogens indicates their important role in cell wall breakdown. The reduced number of CAZy superfamilies (GHs, PLs, and CEs, in particular) can be attributed to biotrophy when organisms derive nutrients from living plant cells (O’Connell et al. 2012).

The number of secreted CWDEs in Diaporthales species included in the analysis ranged from 177 (C. leucostoma CXYLt) to 439 (D. vochysiae LGMF1583) (Fig. 6). A relatively low number of CAZy enzymes was identified in C. piceicola EI-20 (196 secreted out of a total of 498) and C. piceae EI-19(A) (209 secreted out of a total of 505). The strain of C. piceae CFCC 52841 contained a reduced set of GHs and a slightly larger number of AAs than C. piceae EI-19(A). The total number of CAZy detected in both studied species was slightly lower compared to those annotated for the model biotrophic fungus Melampsora larici-populina (526) and hemibiotrophic species Zymoseptoria tritici (537) (Wang et al. 2022). Colletotrichum spp., known as hemibiotrophic plant pathogens mainly, also possessed larger arsenals of secreted CWDEs (from 246 (Col. falcatum Cf671) to 512 (Col. fructicola Cg38)) (Chen et al. 2022). C. piceae EI-19(A) and C. piceicola EI-20 were clustered with other Cytospora species and Celoporthe dispersa CMW9976. This species of Celoporthe was found to be a non-pathogen or potential pathogen of Myrtales (Nakabonge et al. 2006) and carried 220 CAZy, while a highly virulent strain of C. mali 03-8 causing canker disease of Malus spp. (Wang et al. 2011) had a content of 207 such enzymes.

Figure 6. 

Contents of CWDEs and BSGCs predicted in genomes of Diaporthales species included in the analysis. The species are clustered based on the richness of each CWDE or BSGC group.

The genus Diaporthe had diverse sets of CWDEs across the species, from 325 for D. helianthi 7–96 to 439 for D. vochysiae LGMF1583. Interestingly, the former was well studied as a causal agent of sunflower stem canker (Baroncelli et al. 2016), but the latter was first isolated as an endophyte of Vochysia divergens (Noriler et al. 2019). Based on the obtained evidence, it can be assumed that C. piceae EI-19(A) and C. piceicola EI-20 have a biotrophy-like relationship with the corresponding plant host. It can also be concluded that CWDE contents in both studied species allow for causing symptomatic canker disease that may unlikely lead to tree collapse.

Secondary metabolites (SM) consist of low-molecular-weight compounds that can play an important role in species pathogenesis. Nonribosomal peptides (NRP), polyketides (PKS), terpenes, and hybrid metabolites are synthesized by BSGC pathways (Keller et al. 2005). Produced chemical compounds (mycotoxins, in particular) are directly involved in fungus interactions with related hosts (Howlett 2006; Pusztahelyi et al. 2015).

The Diaporthaceae species (incl., Melanconium sp. and Stenocarpella maydis) had the richest complement of SMs (from 77 (D. ampelina DA912) to 122 (D. nobilis DJY16A 5-1)) compared to Schizoparmaceae (from 36 (Co. lustricola B22-T-1) to 42 (Co. vitis QNYT13637)) (Fig. 6). Significant variability in the number of these gene clusters (PKS, especially) was previously noticed on the genus level for Diaporthe species (Hilário et al. 2022). Despite the relatively low number of BSGCs, Coniella species were found to be pathogens of some economically valuable crops (Çeliker et al. 2012; Zambounis et al. 2024). The Cytospora species shared similar proportions of different SMs: from 49 predicted for C. chrysosperma CFL2056 to 60 identified in C. leucostoma CXYLt. It is worth noting that the smallest content of CWDEs within the order was detected for the latter. There were 50 and 54 gene clusters involved in the secondary metabolism identified in C. piceae EI-19(A) and C. piceicola EI-20, respectively. C. piceicola EI-20 had a nearly identical repertoire of BSGCs as C. mali var. pyri SXYL 134, a variety of C. mali with reduced pathogenicity (Wang et al. 2011). Both species carried 30 polyketides, which were found to be involved in fungal virulence (Baker et al. 2006; Ruocco et al. 2018). That was notably higher than those predicted for C. piceae EI-19(A) (24). Generally, the Cytospora species possessed nearly the same number of SMs as many of the analyzed Cryphonectria species (e.g., 51 for Cr. parasitica ES15 and 55 for Cr. carpinicola M9290). The Cryphonectria species were previously related to hemibiotrophic or necrotrophic fungi and considered latent pathogens similar to some endophytes (Stauber et al. 2020). However, some were associated with important plant diseases such as chestnut blight (Rigling and Prospero 2018) and hornbeam decline (Cornejo et al. 2021). Functional genomics and transcriptomic studies have recently revealed an important role of NRPs in C. mali’s pathogenesis (Ma et al. 2016; Feng et al. 2020). This well-studied Cytospora species was previously classified as both necrotrophic (Yin et al. 2015) and hemibiotrophic (Stauber et al. 2020). 17 and 19 NRPs were predicted for C. piceicola EI-20 and C. piceae EI-19(A), respectively, smaller than those detected in C. mali 03-8 (23). However, the strain of C. piceae CFCC 5284 carried 22 of these SMs in its genome. Considering a reduced number of BSGCs identified in C. piceae EI-19(A) and C. piceicola EI-20, it is assumed that the lifestyle of both studied species can be multitrophic (closer to hemibiotrophic) with a capability to infect plant tissue and cause the CC disease quickly. The analysis also suggested that the different strains of C. piceae may have distinct pathogenicity and virulence characteristics.

A phylogenetic tree with estimated divergence times inferred based on the set of 1,706 SCOs is depicted in Fig. 7. The tree topology mainly supported the taxonomic positions of the families and genera within Diaporthales (Senanayake et al. 2017). The analysis showed the placement of S. maydis A1-1 within Diaporthaceae (Lamprecht et al. 2011). The strain of Melanconium sp. NRRL 54901 of Melanconidaceae (Castlebury et al. 2002) was also clustered with species of this family. The family-level classification (as Cytosporaceae) was confirmed for both Cytospora species analyzed in this study. The obtained divergence times matched those previously estimated for some Diaporthales families (e.g., 29.6 MYA for Cryphonectriaceae and 35.1 MYA for Cytosporaceae) (Guterres et al. 2018). The clade of C. picea and C. piceicola originated at 11-10 MYA, which might be considered a relatively late divergence compared to other Diaporthales such as Co. vitis QNYT13637 and Co. lustricola B22-T-1 or D. amygdali DUCC20226 and D. ilicicola FPH2015-502.

Figure 7. 

An ML tree showing phylogenomic relationships of C. piceae EI-19(A), C. piceicola EI-20, and other species of Diaporthales. Bootstrap support for all clades is 100. Blue and red numbers on nodes indicate the number of expanded and contracted gene families, respectively. Circles represent the number of expanded and contracted gene families for the top eight COG categories. Abbreviations: Q—Quaternary, N—Neogene, PG—Paleogene, K—Cretaceous, MYA—Million Years Ago.

A total of 7,579 expanded and 13,984 contracted gene families were identified in the species of Diaporthales included in the analysis. The studied species harbored notably different numbers of expanded and contracted gene families (C. piceae (+46/-109), C. piceicola (+108/-319)), implying a moderate rate of gene contraction (+45/-82) (Fig. 7). For example, C. leucostoma CXYLt experienced a significantly higher rate of gene loss (+107/-638). In contrast, some species within the order have undergone gene family expansion (e.g., D. amygdali (+336/-250), Oph. clavigignenti-juglandacearum (+485/-369), Cr. nitschikei (+81/-16)). Gene family expansion is considered more beneficial for fungal pathogenesis. It may be linked to broader host adaptation, different reproductive strategies, and larger genomes carrying more pathogenicity-related genes (Ma et al. 2010). The species of Diaporthaceae harboring larger genomes compared to other Diaporthales experienced significant gene family expansion (+462/-84), while the reverse was true for the Cytosporaceae (+47/-701) and Schizoparmaceae (+23/-639) families. Host specificity and biotrophy characteristics were observed for the species after gene family contraction events (Zhang et al. 2018; Rogers et al. 2022). It was also found that the plant pathogenic species of Dothideomycetes contained more contracted gene families compared to saprophytes (Yu et al. 2022). This evidence can additionally support the biotrophic-like relationship of C. piceae EI-19(A) and C. piceicola EI-20 with Picea spp. specifically. On the other hand, gene family contraction may drive lifestyle differentiation and host shift in the studied species (Zhang et al. 2018).

The gene families with unknown function (S) were most frequently observed among the analyzed taxa of Diaporthales. Almost all the species experienced gene contraction of this category with some exceptions (e.g., D. caulivora D57 (+121/-95) or Cr. nitschkei CBS 109776 (+36/-7). The studied species showed moderate contraction of the S-categorized gene families (C. piceae EI-19(A) (+18/-53), C. piceicola EI-20 (+37/-154) than the other members of Cytosporaceae (C. leucostoma (+41/-269) or Diaporthaceae (D. helianthi 7–96 (+36/-349)). The secondary structure (Q) and carbohydrate metabolism and transport (G) gene families were also found to be contracted in C. piceae EI-19(A) and C. piceicola EI-20. But the latter underwent a notable loss (+7/-37) of the G category gene families compared to the former (+3/-8). A minor expansion (+9/-5) of the Q-categorized genes was observed for the strain C. piceae CFCC52841. This funding points out that the closely related species or different strains of the same species may differently affect corresponding hosts in terms of pathogenicity and virulence.

Fungal plant pathogens secrete various proteins and metabolites to facilitate host infection. Among them are effectors secreted to reprogram host cells and modulate plant immunity (Stergiopoulos and de Wit 2009; Shao et al. 2021). The predicted secretome and effectome sizes varied significantly among 46 genomes of Diaporthales species included in the analysis. The number of secreted proteins ranged from 347 (C. leucostoma CXYLt) to 928 (D. vochysiae LGMF1583), or 3.45–5.55% of their respective proteomes (S/P). Effectome size ranged from 82 to 290, that is, 0.82–1.74% of the corresponding proteomes (E/P) (Table 3).

C. piceae EI-19(A) and C. piceicola EI-20 harbored very similar secretomes and effectomes, accounting for 3.63% (1.01%) and 3.68% (1.02%), respectively. It is worth noting that the strain C. piceae CFCC 52841 carried a smaller secretome (3.46%) with a higher number of effectors (1.03), which separated it from the studied strains (Fig. 8). Principal component analysis using the ratios of secreted proteins and effectors to respective proteomes grouped C. piceae EI-19(A) and C. piceicola EI-20 with the species of Cryphonectriaceae and Cytosporaceae. The closest species were C. chrysosperma CFL2056, C. malicola (C. schulzeri) 03-1, Ch. austroafricana CMW 2113, and Ch. deuterocubensis CMW 8650, the well-known canker-causing necrotrophic pathogens (Wang et al. 2011; Kepley et al. 2015; Dahali et al. 2023; Suzuki et al. 2023). Interestingly, Melanconium sp. NRRL 54901 and S. maydis A1-1 clustered with the Diaporthaceae species based on the contents of CWDEs and BSGCs had different secretome and effectome parameters.

Figure 8. 

PCA plot showing secretome and effectome sizes of C. piceae EI-19(A), C. piceicola EI-20, and other Diaporthales species.

A distinct separation of some Diaporthales taxa (Diaporthaceae, especially) can indicate genetic diversity variations of secreted proteins and effectors (Wang et al. 2022; Jia et al. 2023). This feature can be employed as a tool to reveal species lifestyles and pathogenicity potential of fungal species associated with plant diseases.

Virulence genes encode proteins related to counteraction of host defense mechanisms that eventually lead to pathogen spread (Toth et al. 2003). The search of such genes was performed against the database of experimentally verified pathogenicity, virulence, and effector genes from fungal, oomycete, and bacterial pathogens of different hosts (PHI-base). The frequency of virulence-associated genes relative to their proteomes was nearly the same for the studied species (2.58–2.61%) (Table 4). The strain C. piceae CFCC 52841 had a relatively lower number of these genes (2.48%) in its proteome. The majority of the hits with PHI-base accessions for all three strains belonged to the categories with reduced virulence (144–148). The genes with unaffected pathogenicity (47–50) and loss of pathogenicity (32–35) were also defined in the analyzed species (Suppl. material 1). It showed that the main part of the identified genes (reduced virulence and loss of pathogenicity) was directly involved in fungal pathogenesis. The strains of C. piceae and C. piceicola EI-20 shared 238 virulence-associated genes. Notably, C. piceicola EI-20 carried 24 unique genes that was significantly more than C. piceae EI-19(A)(6) and C. piceae CFCC 52841(5). The findings additionally confirm that different strains or phylogenetically close species may have distinct pathogenicity and virulence characteristics. The content of virulence-related genes can be a strong indicator of fungus pathogenicity potential. For instance, the less pathogenic C. mali var. pyri SXYL134 (Wang et al. 2011) carried twice the lower number of these genes than the highly aggressive strain C. mali 03-8 (66 and 121, respectively) (Sun et al. 2023). This kind of analysis was also performed for other ascomycetous fungi associated with plant diseases such as Diaporthe (Hilário et al. 2022), Neonectria (Salgado-Salazar et al. 2021), etc. Because of different search parameters employed, the obtained results cannot be properly compared to those from the previous studies.

Host-induced gene silencing (HIGS) is a powerful alternative to traditional (e.g., chemical) treatments employed to protect plants from pathogenic organisms. The technology allows for the downregulation of the target genes in organisms associated with hosts that are not recalcitrant to genetic modifications (Hartmann et al. 2020; Koch and Wassenegger 2021). Target gene selection is the first and crucial step in HIGS. The pathogenicity and virulence factors identified in a pathogen isolated from a specific host significantly contribute to this process. For example, suppression of the effector genes (e.g., AGLIP1, Avra10) resulted in a reduction of fungus growth and disease development caused by Rhizoctonia solani and Blumeria hordei in such economically important crops as barley, rice, and wheat (Nowara et al. 2010; Mei et al. 2024). Overall, properly selected candidate genes for HIGS enhance host resistance against a pathogen and help with disease control. The genes (effectors and virulence-associated genes, specifically) annotated for two Cytospora species associated with canker disease of spruce can be potential candidates for HIGS and other related gene engineering technologies.

The authors have declared that no competing interests exist.

No ethical statement was reported.

Funded by PNURSP2025R31, Princess Nourah bint Abdulrahman University, Saudi Arabia, and A1098531023601245, University of Electronic Science and Technology of China.

Conceptualization: EI. Data curation: EI, YC, SM. Formal analysis: EI, YC. Funding acquisition: AS, SSNM: Methodology: EI, YC, SM, SSNM. Writing—original draft: EI, SSNM.

Evgeny Ilyukhin https://orcid.org/0000-0002-2358-0023

Yanpeng Chen https://orcid.org/0000-0002-2554-5272

Svetlana Markovskaja https://orcid.org/0000-0003-3111-6949

Ashwag Shami https://orcid.org/0000-0002-5336-038X

Sajeewa S. N. Maharachchikumbura https://orcid.org/0000-0001-9127-0783