From the summer of 2021 to 2023, investigations into the incidence and severity of anthracnose were carried out in H. macrophylla growing nurseries of nine parks in Beijing, China ( Table 1 ). Disease incidence was calculated as the proportion of H. macrophylla plants exhibiting anthracnose symptoms relative to the total number of plants under evaluation. A total of 46 leaf samples of H. macrophylla exhibiting typical symptoms of anthracnose were collected for further study. Small pieces (5*5 mm) of leaf tissues were cut from the margin of anthracnose lesions, surface sterilized with 70% ethanol for 30 s, 1% NaClO for 30 s, then rinsed in sterile distilled water for three times, finally transferred onto potato dextrose agar (PDA) with lactic acid (0.1%) and incubated at 28°C for 3 days. Subsequently, the growing edges of the fungal colonies were aseptically transferred onto new PDA plates. The obtained fungal isolates were purified by single spore method and then reserved in 25% (v/v) glycerol at -80°C. The prevalence of Colletotrichum species was estimated as isolation rate (RI) and calculated using the formula RI% = (NS/NI) × 100%, Where NS is the number of isolates belonging to a specific species and NI is the total number of isolates (Vieira et al., 2014). All the 114 Colletotrichum isolates were deposited in the Laboratory of Biocontrol Microorganisms, Institute of Plant Protection, Beijing Academy of Agricultural and Forestry Sciences.

Genomic DNA of the fungal isolates was extracted using the Solarbio^® Fungi Genomic DNA Extraction Kit (Solarbio, China) according to the manufacturer’s instructions. The internal transcribed spacer region of ribosomal DNA (rDNA-ITS) was amplified using PCR primers ITS1 and ITS4 in order to screen the Colletotrichum sp. (White et al., 1990). For further precise identification, the representative Colletotrichum sp. were subjected to amplification of the partial sequences of actin (ACT) (Carbone and Kohn, 1999), beta tubulin (TUB2) (Glass and Donaldson, 1995; O’Donnell and Cigelnik, 1997), Calmodulin (CAL) (Weir et al., 2012), chitin synthase (CHS-1) (Carbone and Kohn, 1999), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Templeton et al., 1992) genes with the corresponding primers listed in Supplementary Table S1 . The amplifications were performed in a 25 μL mixture containing 10.5 μL ddH₂O, 12.5 μL 2×PCR MasterMix, 1 μL DNA template, and 0.5 μL of each primer (10 μM) as described by Weir et al. (2012). DNA sequencing was performed by Beijing B&M Biotech Co., Ltd., China, using forward and reverse primers. Sequences were subjected to BLAST searches and submitted in the National Center for Biotechnology Information (NCBI) database with accession numbers of PP709475-PP709509 for rDNA-ITS, PP768443-768477 for ACT, PP768478-768512 for TUB2, PP768513-768547 for CAL, PP768548-768582 for CHS-1, PP768583-PP768617 for GAPDH, respectively ( Table 2 ).

For phylogenetic analysis, additional reference sequences were selected based on related studies on Colletotrichum species (Weir et al., 2012; Guarnaccia et al., 2017) and retrieved from GenBank. Individual gene datasets of representative isolates were aligned using MAFFT v. 7 (https://mafft.cbrc.jp/alignment/server/) and adjusted manually with BioEdit v. 7.0.9.0 where necessary. The maximum parsimony analyses (MP) were performed based on the multi-loci alignment using PAUP v. 4.0b10. The analysis involved running 1000 replicates of a heuristic search, which utilized random sequence addition for initial tree construction followed by tree bisection reconnection branch swapping (Swofford, 2002). Bayesian inference analysis was conducted using MrBayes 3.1.2. The phylogenetic trees were visualized via TreeviewX v. 0.5.0.

For morphological characterization, mycelial discs from growing edge of the fungal cultures were transferred to fresh PDA plates and incubated at 28°C for 5 days (Cai et al., 2009). Appressoria was produced by dropping 50 μL conidial suspension (10⁶ conidia/mL) on a concavity slide containing moistened filter papers with distilled sterile water, and then incubating at 28°C in the dark for 48 h (Weir et al., 2012). The shape, color and size of conidia (n=40) and appressoria (n=40) for each test Colletotrichum isolate were observed by light microscopy, and their dimensions were examined using an Axioscope 5 microscope (Carl Zeiss Microscopy, Germany). Mycelial growth rate of the representative Colletotrichum isolates was calculated by incubating the fresh mycelia blocks on new plates at 28°C with a photoperiod of 12 h/12 h for 5 days.

Colletotrichum isolates representing different sampling sites or belonging to different Colletotrichum species were selected to conduct the pathogenicity test using mycelial plug method. Healthy leaves of H. macrophylla variety “Wujinxia” were collected, surface sterilized with 70% ethanol, washed three times with sterile distilled water, and then air dried on a sterilized tissue paper. Ten leaves per isolate with three replications were wounded by pin-pricking on both sides of the midrib with a sterilized needle, and 7-mm-diameter mycelia discs of 5-day-old cultures were inoculated, agar blocks without fungi were inoculated as control. Unwounded leaves were inoculated in the same way as described above. All the leaves were placed within a plastic box containing sterile water-soaked filter paper. The box was covered with plastic film and maintained in a growth chamber under condition of 85% relative humidity, a temperature of 28°C, and a 12/12 h light/dark photoperiod. Symptom development and lesion diameters on leaves were examined 7 days post inoculation. The fungus was re-isolated from lesions and recognized by integrated methods of morphological and molecular characteristics in order to fulfill Koch’s postulates.

Representative isolates from the dominant Colletotrichum species, namely C. gloeosporioides, C. fructicola, and C. aenigma were selected and their sensitivities to three DMIs fungicides were tested using mycelial growth rate method (Zhang et al., 2020; Kim et al., 2020). The fungicides including prochloraz, difenoconazole, and tebuconazole were dissolved and adjusted to a concentration of 10 mg/ml as the stock solution. Each fungicide was prepared separately and the stock solutions were serially diluted as follows: prochloraz (0.005, 0.01, 0.02, 0.04, 0.08, 0.1 mg/mL), difenoconazole (0.005, 0.01, 0.05, 0.1, 0.5, 1, 5 mg/mL), tebuconazole (0.01, 0.05, 0.1, 0.5, 1, 5, 10 mg/mL), and then added to the sterilized PDA (approximately 50°C) at a ratio of 1:1000 ( Supplementary Table S2 ). The margin of 5-day-old culture was used to produce 5-mm-diameter mycelial discs, which were then placed at the center of PDA plates with varying concentrations of fungicides. The diameter of each colony was measured in two perpendicular directions, after incubated at 28°C in the dark for 7 days. The percentage inhibition of mycelial growth for each Colletotrichum isolate at each test concentration (I) was also calculated as the difference between the radial growth of nonamended control (C) and the radial growth of each test concentration (T) as follows: I (%) = (C-T)/C×100. Each treatment was tested three times, with three plates for each replication. The EC₅₀ values of the fungicides were calculated and displayed as the mean values derived from 12, 9, and 6 representative isolates of C. gloeosporioides, C. fructicola, and C. aenigma, respectively.

All the data were expressed as mean ± standard deviation of three replications unless otherwise mentioned. Significance of the differences (P < 0.05) was evaluated by one-way analysis of the variance (ANOVA) using the SPSS v21 software. The EC₅₀ values were calculated by linear regression of the probit-transformed relative inhibition value on the log10- transformed fungicide concentration using the statistical algorithms.

During the investigation of H. macrophylla foliar disease from 2021-2023, severe anthracnose symptoms were observed with disease incidence ranging from 23.3%-56.7% in nine parks across Beijing, China. The disease was first observed on newly emerged leaves of H. macrophylla, the infection quickly spread to the around plants in the late growing season. Typical symptoms were initially manifested as tiny purplish-red spots, approximately the size of pinheads with a yellow halo, which subsequently transitioned to light brown or grayish white with brown margins. As the symptoms advanced, these spots ultimately enlarged and merged, resulting in the formation of extensive necrotic regions ( Figure 1 ). A total of 154 monosporic fungal isolates were recovered from symptomatic H. macrophylla leaves. In addition, certain species belonging to other genera like Pythium, Alternaria, and Fusarium were also detected during the isolation process but not shown in this study.

Based on morphology and rDNA-ITS sequence data, the remained 114 isolates resembling Colletotrichum were primarily assigned to three groups, C. gloeosporioides species complex (106 isolates), C. truncatum species complex (6 isolates), C. orchidearum species complex (2 isolates). Further identification based on sequence analysis of six gene loci indicated that, C. gloeosporioides was the most prevalent species (65 isolates, 57.0%) associated with H. macrophylla anthracnose, followed by C. fructicola (33 isolates, 28.9%), C. aenigma (8 isolates, 7.0%), C. truncatum (4 isolates, 3.5%), C. subacidae and C. sojae (2 isolates, 1.8% each). Among which, C. gloeosporioides was the prevalent species in all the parks, C. subacidae was only found in Xishan and Baiwang parks, while C. sojae was only detected in Shuguang Park and Banjing Road ( Table 1 ; Supplementary Figure S1 ).

A total of 35 representative Colletotrichum isolates from different sampling sites or belonging to different species were further subjected to multi-loci phylogenetic analysis with concatenated datasets of rDNA-ITS, ACT, TUB2, CAL, CHS-1 and GAPDH sequences ( Figure 2 ; Table 2 ). Phylogenetic analysis showed that the present Colletotrichum isolates from H. macrophylla anthracnose clearly clustered into three clades with Monilochaetes infuscans CBS 869.96 included as the outgroup ( Supplementary Table S3 ). Among which, twelve isolates within the C. gloeosporioides species complex grouped together to formed a clade with the ex-type isolate of C. gloeosporioides LF604, nine isolates clustered with ex-type isolate of C. fructicola LF130, while the remaining six isolates constituted a distinct clade along with the ex-type isolate of C. aenigma ICMP18608 and JFRL03-1005. For phylogenetic analysis of the C. truncatum species complex, four isolates were grouped together with the ex-type isolates of C. truncatum CBP002, while two isolates formed a clade in conjunction with C. subacidae NN054609. In addition, two Colletotrichum isolates from C. orchidearum complex were clustered with C. sojae ATCC62257.

Distinct morphological features including colony, conidia and appressoria of 35 representative Colletotrichum isolates ( Table 2 ) were observed for each identified Colletotrichum species after 7 days incubation on PDA ( Figure 3 ). Most isolates in C. gloeosporioides species complex developed greyish white to pale grey colonies, while the reverse sides of C. fructicola and C. aenigma were grayish green to olivaceous grey with white margin. The conidia were all cylindrical with obtuse to slightly rounded ends. Appressoria were pale brown to dark brown, subglobose or ellipsoid, and rarely irregular. The C. truncatum and C. orchidearum species complex were easily distinguishable from C. gloeosporioides species complex in terms of conidia or appressoria shape ( Table 3 ). Conidia of C. truncatum was crescent-shaped, smooth-walled, and slightly curved with parallel walls. Conidia of C. subacidae was slightly curved, acute apex, the central part was almost straight with parallel walls. The conidia of C. sojae were cylindrical with obtuse to slightly rounded ends, and their appressoria were dark brown, oval or bullet-shaped. There was a considerable variation in mycelial growth rate among the representative isolates belonging to different Colletotrichum species ( Table 3 ). The average mycelial growth rate of C. aenigma reached 12.6 ± 0.4 mm/d followed by C. gloeosporioides and C. fructicola, while the growth rate of C. sojae was only 7.1 ± 0.3 mm/d.

Pathogenicity test demonstrated that the 35 representative Colletotrichum isolates from different sampling sites or belonging to different species ( Table 2 ) exhibited varying degrees of aggression on H. macrophylla leaves. Seven days after inoculation, all the tested isolates caused symptoms on the wounded leaves. The symptoms mainly manifested as dark brown or brownish, irregular lesions with yellow halos around their peripher on the surface of leaves, consistent with the symptoms observed in field ( Figure 4 ). No lesions were induced in the control leaves inoculated with sterile PDA discs. Notably, certain species such as C. gloeosporioides JZB1040-3-1, C. subacidae JZB1490-1-4 exhibited the highest level of aggressiveness among the tested isolates. In contrast, C. truncatum JZB1564-1-2 produced small necrotic lesions. The remaining isolates had an intermediate level of aggressiveness ( Supplementary Figure S2 ). To fulfill Koch’s postulates, the Colletotrichum species were re-isolated from the lesions of inoculated leaves and identified based on integrated analysis of morphological characteristics and multi-loci sequencing data. The re-obtained isolates matched well with the original ones that were used for inoculation. Further analysis showed that lesions on the wounded leaves were much larger than those on the unwounded leaves, indicating that wound is a crucial prerequisite for the occurrence of H. macrophylla anthracnose.

A total of 27 representative isolates from the three dominant Colletotrichum species, namely C. gloeosporioides, C. fructicola, and C. aenigma ( Supplementary Table S4 ), were chosen to determine their sensitivities to fungicides using the mycelial growth method ( Supplementary Figure S3 ). Regarding the relative fungicide sensitivity of individual Colletotrichum species, we found that C. gloeosporioides, C. fructicola and C. aenigma exhibited greater sensitivity to prochloraz, since their mean EC₅₀ values were only 0.062, 0.033, and 0.023 µg/ml. Colletotrichum fructicola exhibited significantly lower EC₅₀ values against prochloraz than difenoconazole and tebuconazole. Among the three species, there were no significant differences in the EC₅₀ values respect to difenoconazole and tebuconazole ( Table 4 ). The fungicide sensitivities also varied among isolates within the same species. For prochloraz, the EC₅₀ values of C. gloeosporioides spanned from 0.004 to 0.219 µg/ml, while those of C. fructicola ranged from 0.003 to 0.074 µg/ml and those of C. aenigma ranged from 0.007 to 0.064 µg/ml. Some isolates within C. gloeosporioides demonstrated an obviously reduced sensitivity to difenoconazole and tebuconazole, with EC₅₀ values reaching 3.100 and 3.677 µg/ml, respectively ( Supplementary Table S4 ).