Samples of leaves exhibiting disease symptoms were collected from E. sagittatum in Bibo Town, Kaili City, Guizhou Province (26°30'38"N, 107°37'23"E). Pathogenic fungi were isolated from symptomatic leaves using single-spore and tissue isolation methods. If visible conidial masses were observed on the leaves, the conidia were retrieved under a microscope and transferred to sterile water to create a conidial suspension, which was then evenly spread on PDA plates in a laminar flow hood (Senanayake et al. 2020). After 24 h, the mycelium was transferred to fresh PDA plates to obtain pure cultures.

For leaves without visible conidial masses, the tissue isolation method was used to isolate the pathogenic strains. Leaf tissue pieces of approximately 0.5 × 0.5 cm were excised from the margin between healthy and diseased areas. The tissue pieces were immersed in 75% ethanol for 30 s for disinfection, followed by three washes with sterile water for a total of 30 s. The leaf pieces were then placed on sterile filter paper to dry before being transferred to PDA plates (Yao et al. 2024). All PDA plates were incubated in the dark at 28 °C for 1–2 days. Once colonies emerged, a small amount of mycelium was picked from the colony edge and transferred to fresh PDA medium for further cultivation. Colonies were purified at least twice until pure cultures were obtained. Type specimens were deposited in the Herbarium of the Department of Plant Pathology, Agricultural College, Guizhou University (HGUP). Ex-type cultures were deposited in the Culture Collection at the Department of Plant Pathology, Agriculture College, Guizhou University, P.R. China (GUCC).

The purified pathogenic fungi were inoculated onto PDA plates and incubated at 28 °C for 7 days. The colony morphology, including shape and colour, was observed and recorded according to the colour map of Rayner (1970). A small amount of mycelium from the colonies was sampled and examined using a Carl Zeiss AGAxiomo microscope to record structural details such as conidia, setae, and asci. Appressoria germination was induced by slide culture, and the conidial masses from PDA plates were transferred to sterile water to prepare a conidial suspension. A drop of this suspension was placed on a microscope slide and incubated at 28 °C for 12–24 h (Cai et al. 2009). Subsequently, the slides were observed using a Carl Zeiss AGAxiomo microscope to document appressoria shape, colour, and size. Thirty conidia and appressoria were selected randomly to measure their lengths and widths.

DNA extraction: After culturing the pathogenic fungi on PDA for 10 days, the mycelium was scraped from the plates using a sterile surgical scalpel and placed into a 2 mL centrifuge tube for storage. DNA of pathogenic fungi was extracted using a Fungal DNA (Biomiga) reagent kit and stored at -20 °C for future use.

Polymerase chain reaction (PCR) amplification was performed using the extracted DNA samples as templates. The gene sequences of the following genomic regions were amplified: rDNA internal transcribed spacer (ITS), actin (act), chitin synthase (chs-1), β-tubulin (tub2), glyceraldehyde-3-phosphate dehydrogenase (gapdh), histone H3 (his3), and calmodulin (cal). Primers used for the amplification are listed in Table 1. Each PCR reaction mixture had a total volume of 25 μL, comprising 1 μL each of forward and reverse primer, 12.5 μL of 2× PCR Master Mix, 9.5 μL of deionised water (ddH₂O), and 1.0 μL of the DNA template (Liu 2023).

The amplification protocol for the ITS region was as follows (Woudenberg et al. 2009): first, initial denaturation at 95 °C for 4 min; second, 35 denaturation cycles at 95 °C for 30 s, annealing at 52 °C for 30 s, and extension at 72 °C for 45 s; lastly, single extension at 72 °C for 10 min. The annealing temperatures for the other genes were as follows: 58 °C for act, his3, and chs-1; 55 °C for tub2; 59 °C for gapdh; and 57 °C for cal.

A 1.2% agarose gel was used for electrophoresis, which was stained with 0.5 g/mL ethidium bromide for 10 min. Visualisation was performed using a BIO-RAD gel imaging system. Subsequently, PCR products were sent to the Shanghai Bioengineering Company for sequencing.

Phylogenetic analysis was performed using DNA sequence data obtained from GenBank (https://www.ncbi.nlm.nih.gov/, accessed on 23 October 2024) (Tables 2, 3) and following previous publications (Liu et al. 2022; Zhang et al. 2023; Li et al. 2024). Multiple sequence alignments were performed using MAFFT (Miller et al. 2010); aligned sequences were manually adjusted using BioEdit v7.0.5 software (Rambaut 2009); using SequenceMatrix 1.8 to assemble multiple gene sequences (Vaidya et al. 2011). Phylogenetic trees were constructed using MrBayes and RAxML on the CIPRES Science Gateway V.3.3 website (https://www.phylo.org/portal2/login.action) (Miller et al. 2010; Weir et al. 2012). For the maximum likelihood (ML) analysis, RAxML-HPC2 on XSEDE v.8.2.12 was employed with a PHYLIP-formatted sequence alignment file under the GTR+GAMMA nucleotide substitution model. Branch support values were estimated through 1,000 bootstrap replicates, and the final ML tree retained branch lengths. For the Bayesian inference (BI) analysis, MrBayes v.3.2.6 was executed using a NEXUS-formatted alignment file. The optimal nucleotide substitution model was selected via MrModeltest v.2.3. Markov Chain Monte Carlo (MCMC) simulations were run with sampling every 1,000 generations. To ensure convergence, the first 25% of samples were discarded as burn-in, and a majority-rule consensus tree was generated from the remaining post-burn-in samples. Branch credibility was assessed using posterior probabilities (PP). The resulting tree files were visualised and resized using FigTree v1.4.0 (Rambaut et al. 2018) and then edited with Adobe Illustrator CS5.

The pathogenic fungus was isolated and purified from diseased Epimedium plants before performing a Koch’s postulate reinoculation experiment. The strains were cultured in a 28 °C incubator for 10–25 days until conidia were produced. Using a punch, fungal cakes were harvested and transferred to conical flasks containing potato dextrose broth (PDB). These flasks were then placed on a shaker set to 220 rpm at 28 °C and cultured for 5 days to prepare a conidial suspension. The conidia concentration was determined using a hemocytometer, and the conidial suspension was adjusted to 1 × 10⁶ conidia/mL using sterile water (Guo 2023). Subsequently, this conidial suspension was sprayed onto the leaves of healthy, injured, and uninjured Epimedium plants for pathogenicity assessment (Than et al. 2008; Cai et al. 2009), with three replicates per treatment and spraying plants with sterile water as controls. The inoculated plants were covered with plastic bags to maintain humidity and placed in a controlled environment chamber (28 °C, 12 h light/12 h dark, 80% humidity) for observation. Regular assessments of disease development were conducted, and the results were recorded. After disease onset, the pathogen was isolated and purified from the diseased tissue.

The disease typically starts in April, persisting until October, with a peak incidence from June to August, when the disease incidence reaches up to 25%, and in severe cases, up to 90%. Early symptoms manifest as small light brown or brown circular spots on the middle or edges of the leaves. Gradually, these lesions expand into circular, elliptical and irregular shapes, often accompanied by irregular concentric rings. The centre of the lesions eventually turns greyish-white or grey-brown, and the margins turn dark brown and are surrounded by a yellow halo (Fig. 1A, B). In the late stages of the disease, the lesions become thin and prone to cracking, with distinct fruiting bodies; conidial heads can be observed microscopically (Fig. 1C, D).

Figure 1. 

Symptoms of field anthracnose in Epimedium sagittatum.

A total of 16 C. epimedii and 9 C. sagittati isolates were obtained in this study. Three representative strains from each species were selected for pathogenicity re-inoculation experiments. Symptoms began to appear four days after inoculation with C. epimedii, characterised by localised leaf yellowing and discolouration and accompanied by irregular brown spots. Over time, the lesions resembled the symptoms observed in the natural fields (Figs 1A, B, 2A, B). Non-treated control plants remained healthy without any symptoms (Fig. 2C). Similarly, C. sagittati could infect healthy plants but exhibited weaker pathogenicity, requiring wound inoculation to induce disease development. The resultant symptoms resembled field observations (Fig. 2D, E). Negative controls inoculated with sterile water remained asymptomatic (Fig. 2F).

Figure 2. 

Pathogen inoculation and symptom (10 days) A, B symptoms resulting from inoculation with Colletotrichum epimedii C control D, E symptoms resulting from inoculation with Colletotrichum sagittati F control.

Following pathogenicity assays, lesion margin tissues from inoculated E. sagittatum leaves were subjected to re-isolation of both Colletotrichum species. Morphological characterisation of the re-isolated pathogens revealed identical conidial dimensions and colony characteristics to those of the original isolates. These findings were consistent with initial isolation data, confirming C. epimedii and C. sagittati as the causal agents of anthracnose in E. sagittatum.

Twenty-five strains of Colletotrichum, isolated from leaves of E. sagittatum, were identified based on phylogenetic analyses of six or seven loci. In the phylogenetic analysis of the C. spaethianum species complex, a total of 2327 characters, including gaps, were identified (ITS: 538, act: 237, chs-1: 251, gapdh: 255, his3: 373 and tub2: 673). Similarly, the phylogenetic analysis of the C. boninense species complex yielded a total of 2583 characters, including gaps (ITS: 554, act: 254, chs-1: 251, gapdh: 242, his3: 375, tub2: 500, and cal: 407). The topology of Bayesian analysis of cascading datasets is almost the same as the ML consistency tree.

In the phylogenetic tree (Figs 3, 4), the isolates from this study formed two distinct, well-supported clades and, thus, were considered to represent two previously unknown species. C. epimedii GUCC 24-0190, GUCC 24-0191 and GUCC 24-0192 without the DNA base differences in six loci amongst strains (ITS, gapdh, act, his3, chs-1 and tub2) form an independent branch with strong support (ML = 95, PP = 1) sister to C. incanum (Fig. 3). Similarly, in the phylogenetic tree (Fig. 4), C. sagittati GUCC 24-0193, GUCC 24-0194 and GUCC 24-0195 also form an independent branch with strong support (ML = 94, PP = 1).

Figure 3. 

Phylogenetic tree of the Colletotrichum spaethianum complex based on multi-gene sequences (ITS, act, tub2, gapdh, chs-1 and his3). Support values at the nodes indicate a maximum likelihood (ML) of > 60% and Bayesian posterior probability (BYPP) of > 0.70. The outgroup is C. destructivum CBS 136228 and C. destructivum AKCC 47638. The strains used were GUCC 24-0190, GUCC 24-0191 and GUCC 24-0192. The scale bar represents 0.02.

Figure 4. 

Phylogenetic tree of the Colletotrichum boninense complex constructed using multi-gene sequences (ITS, act, tub2, gapdh, chs-1, his3 and cal). Support values at the nodes indicate a maximum likelihood (ML) of > 60% and Bayesian posterior probability (BYPP) of > 0.70. The outgroup is C. euphorbiae CBS 134752, COAD 3300, and Col-170. The strains used are GUCC 24-0193, GUCC 24-0194 and GUCC 24-0195, with a scale bar of 0.04.

The authors have declared that no competing interests exist.

No ethical statement was reported.

This work was supported by the National Key Research and Development Program under grant number 2021YFD1601001.

Data curation: ZL, XZ, KJ. Formal analysis: XZ, ZL, SL. Funding acquisition: ZL. Investigation: WZ, XC. Writing - original draft: KJ. Writing - review and editing: KJ.

All of the data that support the findings of this study are available in the main text.