Two Newly Identified Colletotrichum Species Associated with Mango Anthracnose in Central Thailand

Anthracnose caused by Colletotrichum spp. is one of the major problems in mango production worldwide, including Thailand. All mango cultivars are susceptible, but Nam Dok Mai See Thong (NDMST) is the most vulnerable. Through a single spore isolation method, a total of 37 isolates of Colletotrichum spp. were obtained from NDMST showing anthracnose symptoms. Identification was performed using a combination of morphology characteristics, Koch’s postulates, and phylogenetic analysis. The pathogenicity assay and Koch’s postulates on leaves and fruit confirmed that all Colletotrichum spp. tested were causal agents of mango anthracnose. Multilocus analysis using DNA sequences of internal transcribed spacer (ITS) regions, β-tubulin (TUB2), actin (ACT), and chitin synthase (CHS-1) was performed for molecular identification. Two concatenated phylogenetic trees were constructed using either two-loci of ITS and TUB2, or four-loci of ITS, TUB2, ACT, and CHS-1. Both phylogenetic trees were indistinguishable and showed that these 37 isolates belong to C. acutatum, C. asianum, C. gloeosporioides, and C. siamense. Our results indicated that using at least two loci of ITS and TUB2, were sufficient to infer Colletotrichum species complexes. Of 37 isolates, C. gloeosporioides was the most dominant species (19 isolates), followed by C. asianum (10 isolates), C. acutatum (5 isolates), and C. siamense (3 isolates). In Thailand, C. gloeosporioides and C. acutatum have been reported to cause anthracnose in mango, however, this is the first report of C. asianum and C. siamense associated with mango anthracnose in central Thailand.


Introduction
Mango production has expanded to more than 100 countries, with around 44.6 M tons annually since 2018 [1]. Thailand is a major producer and contributes to almost 8% of the global mango production [1]. Because of its flavor and texture, Nam Dok Mai See Thong (NDMST) has become the most popular mango cultivar in Thailand. It is also an early-midseason cultivar, which means that it has the potential to produce fruit all year round [2,3]. Anthracnose caused by Colletotrichum spp. is a significant economic problem in both mango orchards and postharvest storage [3][4][5][6][7][8]. The pathogen infects not only fruit but also inflorescences, flowers, and leaves. During the flowering stage, especially under high humidity, a disease incidence of 100% has been observed [4,[9][10][11][12]. Young leaves that emerge during rainy periods are also prone to anthracnose infection. Anthracnose symptoms on leaves appear as small and dark brown spots often surrounded with chlorotic

Morphology-Based Identification
Based on colony morphology and conidia characteristics, the Colletotrichum isolates were classified into two groups. The first group contained five isolates that were similar to C. acutatum. The colony grown on PDA appeared white to grey and the reverse side of the colony was pale ochreous. The mycelial growth rates at day 3 ranged from 3.33-3.77 mm day −1 (average = 3.56 mm day −1 ). Spore masses were bright orange, conidia were hyaline, aseptate, straight, apex obtuse, and no setae. The sizes of conidia ranged from 3.38-6.17 µ m (average = 4.43 µ m) in width and from 10.96-19.88 µ m (average = 14.8 µ m) in length. Appressoria were clavate, long and irregular shapes, pale to dark brown in color. The diameter of appressoria ranged from 3.86-6.95 (average = 5.45 µ m) in width and ranged from 5.97-11.20 (8.36) µ m (Table S1, Figure 2).
The second group contained 32 isolates with similarity to C. gloeosporioides (some of these isolates were later classified as C. asianum or C. siamense when DNA markers were integrated, described below). The colony grown on PDA showed a great variation in color from white, greenish to grayish, pale yellowish to dark grey, while the reverse sides were dark green ( Figure 2). The mycelial growth rates also varied from 2.13 to 5.43 mm (average = 4.01 mm day −1 ). Conidia were hyaline, aseptate, straight cylindrical, rounded at the apex end and conspicuous hilum at basal end, and no setae. The sizes of conidia ranged from 3.60-7.07 µ m (average = 5.03 µ m) in width and ranged from 10.33-19.95 µ m (average = 14.24 µ m) in length (Table S1, Figure 2). Spore masses were pale, salmon-orange to bright orange colors. Various shapes of appressoria of clavate, long clavate, occasionally irregular, pale to dark brown in color were observed (Table S1, Figure 2). The size of the appressoria ranged from 4.42-8.09 µ m (average = 5.82 µ m) × 5.53-11.59 µ m (average = 8.72 µ m) (Table S1, Figure 2).

Morphology-Based Identification
Based on colony morphology and conidia characteristics, the Colletotrichum isolates were classified into two groups. The first group contained five isolates that were similar to C. acutatum. The colony grown on PDA appeared white to grey and the reverse side of the colony was pale ochreous. The mycelial growth rates at day 3 ranged from 3.33-3.77 mm day −1 (average = 3.56 mm day −1 ). Spore masses were bright orange, conidia were hyaline, aseptate, straight, apex obtuse, and no setae. The sizes of conidia ranged from 3.38-6.17 µm (average = 4.43 µm) in width and from 10.96-19.88 µm (average = 14.8 µm) in length. Appressoria were clavate, long and irregular shapes, pale to dark brown in color. The diameter of appressoria ranged from 3.86-6.95 (average = 5.45 µm) in width and ranged from 5.97-11.20(8.36) µm (Table S1, Figure 2).
The second group contained 32 isolates with similarity to C. gloeosporioides (some of these isolates were later classified as C. asianum or C. siamense when DNA markers were integrated, described below). The colony grown on PDA showed a great variation in color from white, greenish to grayish, pale yellowish to dark grey, while the reverse sides were dark green ( Figure 2). The mycelial growth rates also varied from 2.13 to 5.43 mm (average = 4.01 mm day −1 ). Conidia were hyaline, aseptate, straight cylindrical, rounded at the apex end and conspicuous hilum at basal end, and no setae. The sizes of conidia ranged from 3.60-7.07 µm (average = 5.03 µm) in width and ranged from 10.33-19.95 µm (average = 14.24 µm) in length (Table S1, Figure 2). Spore masses were pale, salmonorange to bright orange colors. Various shapes of appressoria of clavate, long clavate, occasionally irregular, pale to dark brown in color were observed (Table S1, Figure 2). The size of the appressoria ranged from 4.42-8.09 µm (average = 5.82 µm) × 5.53-11.59 µm (average = 8.72 µm) (Table S1, Figure 2).

DNA Marker-Based Identification
All DNA sequences of ACT, CHS-1, ITS, and TUB2 were subjected to BLASTn (Table S2). Sequences of ex-type or epitype strains of Colletotrichum species (Table S2) were selected for phylogenetic analysis. We firstly constructed a phylogenetic tree using two DNA markers: ITS and TUB2. A total of 47 isolates, including ten isolates of ex-type and ex-epitype or epitype strains, and Leptosphaeria veronicae CBS145.84 was used as an outgroup ( Figure 3). The topology of the ML tree and Bayesian tree were identical, therefore, only the ML tree is shown. A discrete Gamma distribution was used to estimate the divergence and evolutionary rate (+G, parameter = 7.1107) ( Figure 3). The 37 isolates were assigned to four species clades on the maximum likelihood (ML) tree. Of these, 19 isolates formed a clade that was closely related to C. gloeosporioides strain (ICMP17821), showing 0.13 posterior probability with a bootstrap value of 100%, three isolates were clustered with C. siamense

DNA Marker-Based Identification
All DNA sequences of ACT, CHS-1, ITS, and TUB2 were subjected to BLASTn (Table  S2). Sequences of ex-type or epitype strains of Colletotrichum species (Table S2) were selected for phylogenetic analysis. We firstly constructed a phylogenetic tree using two DNA markers: ITS and TUB2. A total of 47 isolates, including ten isolates of ex-type and ex-epitype or epitype strains, and Leptosphaeria veronicae CBS145.84 was used as an outgroup ( Figure 3). The topology of the ML tree and Bayesian tree were identical, therefore, only the ML tree is shown. A discrete Gamma distribution was used to estimate the divergence and evolutionary rate (+G, parameter = 7.1107) ( Figure 3). The 37 isolates were assigned to four species clades on the maximum likelihood (ML) tree. Of these, 19 isolates formed a clade that was closely related to C. gloeosporioides strain (ICMP17821), showing 0.13 posterior probability with a bootstrap value of 100%, three isolates were clustered with C. siamense (ICMP12567) with 0.01 posterior probability and a bootstrap value of 100%; ten isolates The incongruence length difference (ILD) test showed that the ACT, CHS-1, ITS, and TUB2 were homogeneous. Therefore, the concatenated sequences of four markers; ACT, CHS-1, ITS, and TUB2 were used to generate a phylogenetic tree with a total of 81 isolates, including 44 isolates of ex-type and ex-epitype or epitype strains. The topology of the ML tree was consistent with that of the Bayesian tree, and therefore only the ML tree is shown ( Figure 4). Among the 37 isolates, ten isolates were clustered with C. asianum strains (ICMP 18580, ICMP 18696, NN8, WM52, and NN19) showing 0.97 posterior probability and with bootstrap values of 99%; three isolates formed a clade with C. siamense strains  The incongruence length difference (ILD) test showed that the ACT, CHS-1, ITS, and TUB2 were homogeneous. Therefore, the concatenated sequences of four markers; ACT, CHS-1, ITS, and TUB2 were used to generate a phylogenetic tree with a total of 81 isolates, including 44 isolates of ex-type and ex-epitype or epitype strains. The topology of the ML tree was consistent with that of the Bayesian tree, and therefore only the ML tree is shown (Figure 4). Among the 37 isolates, ten isolates were clustered with C. asianum strains (ICMP 18580, ICMP 18696, NN8, WM52, and NN19) showing 0.97 posterior probability and with bootstrap values of 99%; three isolates formed a clade with C. siamense strains

Pathogenicity Test
A mycelial plug of a 5-day-old culture was inoculated on unwounded fruit and leaves. Typical anthracnose lesions were observed around the inoculation sites. The anthracnose lesions on fruit enlarged faster than those on leaves on day 5 after inoculation (Table 2 and Figure 5). No lesions were observed on the control fruit or leaves ( Figure 5). On inoculated fruit, C. asianum produced the largest lesions that differed significantly from other species with the average LD of 8.16 cm, followed by the LD means of C. gloeosporioides, C. siamense, and C. acutatum at 8.07, 7.81, and 7.61 cm, respectively (Table 2 and Figure 6). A mycelial plug of a 5-day-old culture was inoculated on unwounded fruit and leaves. Typical anthracnose lesions were observed around the inoculation sites. The anthracnose lesions on fruit enlarged faster than those on leaves on day 5 after inoculation (Table 2 and Figure 5). No lesions were observed on the control fruit or leaves ( Figure 5). On inoculated fruit, C. asianum produced the largest lesions that differed significantly from other species with the average LD of 8.16 cm, followed by the LD means of C. gloeosporioides, C. siamense, and C. acutatum at 8.07, 7.81, and 7.61 cm, respectively (Table 2 and Figure 6). Similar results were observed on inoculated leaves; C. siamense showed the largest lesions that differed significantly from other species with the average LD of 1.25 cm, whereas the LD means of C. asianum, C. gloeosporioides and C. acutatum were 0.92, 0.71, and 0.38 cm, respectively (Table 2 and Figure 6). After the pathogenicity test, all the Colletotrichum spp. were re-isolated from the infected tissues and confirmed by Koch's postulates to have identical morphological characteristics as the original isolates.

Discussion
It has been widely accepted that high genetic variation within Colletotrichum species complexes exists due to their wide host range and diverse environments [23][24][25][26][27]. Mango anthracnose has been reported to associate with different species of Colletotrichum. This study aimed to accurately identify the species of the Colletotrichum-mango system in Thailand, using a combination of morphology, multilocus sequence analyses, and Koch's postulates. A total of 37 Colletotrichum species were isolated from mango anthracnose disease from orchards located in central Thailand. When morphological characteristics such as colony color, growth rate, size, and shape of conidia and appressoria were used, 32 isolates were identified as C. acutatum, and five isolates were identified as C. gloeosporioides. It is gener- Similar results were observed on inoculated leaves; C. siamense showed the largest lesions that differed significantly from other species with the average LD of 1.25 cm, whereas the LD means of C. asianum, C. gloeosporioides and C. acutatum were 0.92, 0.71, and 0.38 cm, respectively (Table 2 and Figure 6). After the pathogenicity test, all the Colletotrichum spp. were re-isolated from the infected tissues and confirmed by Koch's postulates to have identical morphological characteristics as the original isolates.

Discussion
It has been widely accepted that high genetic variation within Colletotrichum species complexes exists due to their wide host range and diverse environments [23][24][25][26][27]. Mango anthracnose has been reported to associate with different species of Colletotrichum. This study aimed to accurately identify the species of the Colletotrichum-mango system in Thailand, using a combination of morphology, multilocus sequence analyses, and Koch's postulates. A total of 37 Colletotrichum species were isolated from mango anthracnose disease from orchards located in central Thailand. When morphological characteristics such as colony color, growth rate, size, and shape of conidia and appressoria were used, 32 isolates were identified as C. acutatum, and five isolates were identified as C. gloeosporioides. It is generally accepted that morphological characteristics alone are not sufficient for species identification since variation in traits among species can be similar under different environments, therefore the multilocus phylogenetic approach was integrated to aid in taxonomy.
Several DNA markers have been developed to identify Colletotrichum isolates. A multilocus sequence analysis using ITS, CAL, or TUB2 identified C. alienum, C. fructicola, or C. tropicale from other Colletotrichum species [27]. Similarly, a study of lupin anthracnose by Alkemade et al. [31] showed that 39 out of 50 isolates belonged to Colletotrichum lupini. The authors also used the combination of multilocus analysis (ITS, TUB2, GAPDH, and APN/MAT1) and morphological characteristics to support their taxonomic classification of Colletotrichum species complex. This study used phenotypic characters, DNA markers of ITS, ACT, CHS-1, and TUB2, and Koch's postulates to confirm that all 37 Colletotrichum isolates were mango anthracnose pathogens. In addition to the four DNA markers used in this study, GAPDH has also been used to identify C. acutatum and C. gloeosporioides complexes by Damm et al. [23] and Weir et al. [27]. In this study, two phylogenetic trees were constructed using two loci (ITS and TUB2) and four loci (ITS, TUB2, ACT and CHS-1). The topology of both trees was similar and further identified the Colletotrichum species into C. acutatum, C. gloeosporioides, C. asianum and C. siamense. These results indicated that using two loci is sufficient to distinguish these four species of Colletotrichum. Colletotrichum gloeosporioides and C. acutatum were previously reported in 1979 and 2019 [32]; however, we have uncovered for the first time that C. asianum and C. siamense caused anthracnose in mango in Thailand.
Colletotrichum asianum has been reported to be a major species causing mango anthracnose in China [33]. It has been the most common endophytic species of mango in northeastern Brazil [34] and in many countries around the world, such as Australia, China, Colombia, Japan, Malaysia, Philippines, Sri Lanka, and Taiwan [11][12][13]18,19,27,35,36]. Colletotrichum asianum is also capable of causing disease in avocado (Persea americana Mill.) in Australia [19]. In Thailand, C. asianum has been reported to cause disease in coffee (Coffea arabica L.) [37] but not in mango until this study.
Colletotrichum siamense has been reported to be a major species causing mango anthracnose in China [33] and in eastern Australia [38]. It has a wide host range in tropical and subtropical regions and can infect banana (Musa spp.), papaya (Carica papaya L.), dragon fruit (Hylocereus spp.), guava (Psidium guajava) and avocado [12,18,19,27,33,[39][40][41][42][43][44][45][46]. It was suggested by James et al. [38] that C. siamense is likely to be a common and widespread saprophyte or endophyte because the authors were able to isolate it from asymptomatic fruit of other plants, except mango and avocado. In Thailand, similar to C. asianum, C. siamense has been reported to cause disease in coffee (Coffea arabica L.) [37] but not in mango until this study.
Among the four species identified, C. gloeosporioides was a major species causing anthracnose in this study. It was found in every orchard that we isolated. The pathogenicity assay showed that all isolates appeared to be more aggressive on fruit than leaves, probably because fruit contains more sugar that serves as a carbon source for fungal growth. C. asianum and C. gloeosporioides showed a similar degree of aggressiveness on mango fruit and produced larger lesions than C. siamense and C. acutatum. The color of lesions were also slightly different on inoculated fruit, C. asianum produced dark brown spots, while the other three species produced lighter brown spots. The pathogenicity on the leaf showed that C. siamense was the most aggressive compared to other species, possibly due to species-specific host responses. Further testing using different cultivars of mango should be employed to verify this hypothesis.
The impact of this study was twofold: firstly, it has proven that a combination of morphology, DNA multilocus sequence analysis, and Koch's postulates is a robust identification approach to identify species of Colletotrichum, and secondly, our approach enables the discovery of previously unreported anthracnose caused by C. asianum and C. siamense, in Thailand. This discovery raises a concern regarding the cross-infection potential where the two species can infect different hosts. Therefore, accurate diagnosis is a crucial first step for disease control and prevention. It supports the current quarantine regulations as well as establishes strategies for integrated management of anthracnose disease between orchards. It should be noted that fungal pathogens, although in the same genus, will respond to fungicides differently. Further study is essential to determine the fungicide sensitivity of these four Colletotrichum species to help implement the fungicide management strategy.

Fungal Isolation
The infected samples of 25 fruit, 10 leaves, and two inflorescences with typical anthracnose symptoms were collected from eight orchards located in central Thailand: Chachoengsao, Phichit, and Ratchaburi, in 2016-2017. The locations of orchards, geographic coordinates, codes of the isolates, and types of infectious tissues collected are provided in Table 1. Thirty-seven isolates of Colletotrichum spp. were recovered. The pathogens were isolated and cultured on potato dextrose agar (PDA) using a tissue transplantation technique. To obtain pure isolates, mycelial plugs were sub-cultured on fresh PDA, and incubated at 25 • C under a photoperiod of 12 h light/12 h dark for 5 days [8], followed by single spore isolation on water agar (WA). The single spore from each isolate was transferred to a new PDA plate and incubated under the same condition. Each colony served as a single genetic source for further analysis.

Morphological Characteristics
Each isolate was inoculated with a 6-mm-diameter plug taken from an actively growing edge of a 5-day-old culture on a PDA plate. The culture was incubated under the same condition mentioned above. Fungal growth and colony diameter were recorded daily until there were no changes in diameter (day 3). Fifty conidia were randomly selected for measurement of their length and width at day 5 (conidia were fully matured) under an Olympus CX31 binocular compound microscope at 400× magnification with the Olympus CellSens standard software version 1.16 (Olympus Co., Ltd., Tokyo, Japan). Appressoria were induced using a slide culture technique [47]. Briefly, the isolates were transferred onto 25.4 × 76.2 mm sterile microscope slides (7101 microscope slides, Shandong Harmowell Trade Co., Ltd., Shanghai, China) and covered with 22 × 22 mm coverslips (Menzel Gläser, Thermo Fisher Scientific Co., Ltd., Waltham, MA, USA) incubated in a petri dish at 25 • C until maturation, usually for 4 days. Lengths and widths of 30 appressoria per isolate were measured under a microscope. This experiment was performed in five replicates.

DNA Extraction and Molecular Identification
Genomic DNA was extracted from a 5-day-old fungal colony, according to Pongpisutta et al. [48]. A PCR mixture contained genomic DNA (20 ng), primers (0.48 µM each, Table 3), Taq polymerase buffer (1×, Thermo Fisher Scientific, Waltham, MA, USA), MgCl 2 (2.4 mM), dNTPs (10 µM each), Taq polymerase (1 U, Thermo Fisher Scientific, Waltham, MA, USA). PCR was performed in a thermal cycler (Sensoquest GmbH, Göttingen, Germany). The PCR was programmed as follows: 94 • C for 2 min, 35 cycles of denaturation at 94 • C for 30 s, annealing at 55-58 • C (Table 3) for 30 s, and extension at 72 • C for 1 min, with a final extension step of 72 • C for 10 min. The PCR products were visualized on 1.2% agarose gels stained with GelStar ® and GeneRuler, and 100 bp Plus DNA Ladder (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine the size of DNA fragments. The PCR products were sequenced by the 1st Base Laboratory Co., Ltd., Seri Kembangan, Malaysia. A basic local alignment search, BLASTn (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 22 April 2021), was performed to analyze nucleotide sequences in comparison to reference sequences available in the National Center for Biotechnology Information (NCBI) database.

Phylogenetic Analyses
Multiple sequence alignments were performed using ClustalW alignment [52] implemented in MEGA version X [53] and were manually adjusted to allow maximum sequence similarity. Bayesian inference (BI) was used to reconstruct the phylogenetic trees using MrBayes version 3.2.7 [54] implemented in the CIPRES cluster (https://www.phylo.org/ portal2/home.action, accessed on 12 December 2021). The nucleotide substitution model was determined by jModelTest v. 2.1.7 [55]. Following Drummond and Rambaut [56], 1,000,000 generations (four chains, four independent runs) were set up, and the analyses were sampled every 1000 generations, with the first 25% of the samples discarded. Maximum likelihood analyses were conducted by the MEGA version X [53] using a TN93+G substitution model based on 1000 bootstrap replicates.

Pathogenicity Test
Harvested mature mango fruit and leaves were washed under running water, immersed in 1.2% sodium hypochlorite solution for 2 min, rinsed twice with sterile distilled water, and allowed to dry under a laminar flow hood. Thirty-seven isolates of Colletotrichum were used for the pathogenicity test. Each Colletotrichum isolate was inoculated on unwounded fruit and leaves as described in Pongpisutta et al. [8]. Briefly, a mycelial plug from the growing edge of 5-day-old PDA culture was placed onto the fruit and leaf surface of the NDMST cultivar. A control treatment was performed using a non-colonized agar plug. The inoculated samples were placed on trays lined with sterile moist paper towels and kept in sealed plastic bags. Five days after inoculation at room temperature, evaluation of the virulence was performed by measurement of lesion diameter (LD). This experiment was performed in a completely randomized design (CRD) with ten replicates. One-way ANOVA was performed using R software version 3.5.2 [57] with the agricolae package (Statistical procedures for agricultural research) [58]. The means of the LDs were compared by the least significant difference (LSD) test. Additionally, variation of disease lesion diameters on NDMST mango fruit and leaves was compared amongst four Colletotrichum species by using box plot analysis which was created in R software likewise.

Conclusions
A combination of multilocus phylogenetic analysis, phenotypic characters, and Koch's postulates, provides an effective strategy to overcome the problem of identification and characterization of fungal species. We showed that molecular analysis of at least two loci (ITS and TUB2) provides accurate identification of Colletotrichum species causing anthracnose disease in mango from central Thailand. This study represents the first report that C. asianum and C. siamense were found to be causative agents of mango anthracnose in Thailand.