Molecular phylogeny-based identification of Colletotrichum endophytica and C. siamense as causal agents of avocado anthracnose in Sri Lanka

Avocado (Persea americana) is a sub-tropical fruit with high nutritional value and numerous health benefits. Among the postharvest fungal diseases that affect ripe avocados, anthracnose is one of the most destructive disease worldwide, causing significant postharvest fruit losses and limiting shelf life. Over 15 Colletotrichum species have been reported as causing avocado anthracnose from avocado growing countries in the world. In the present study, 35 Colletotrichum isolates were obtained from ripe avocados showing anthracnose symptoms, collected from the Central and North Western Province of Sri Lanka. Fifteen randomly selected isolates were subjected to DNA sequence analysis using ITS, TUB2, and GAPDH regions. Species affiliations and identities of the resulting sequences were determined through similarity-based searches of the NCBI GenBank Database. Based on the combined phylogentic analysis of three gene regions, nine and six isolates were identified as C. endophytica and C. siamense respectively, both belonging to the C. gloeosporioides species complex. Of the two species, C. endophytica is reported as a causal agent of avocado anthracnose for the first time.


INTRODUCTION
Avocado (Persea americana Mill, Lauraceae) is native to Central America and southern Mexico and believed to have originated about 12,000 years ago, based on archeological evidence. The avocado is botanically classified in to three races, West Indian (WI), Mexico (XX), and Gautemalan (G). Systematic studies have classified more than 500 cultivars worldwide and there is a great variability in fruit traits not only between races but also among cultivars within races. The peel of some cultivars (e.g. Hass) changes from green to black or purple. The pericarp, which is the fruit tissue proper excluding the seed, comprises the rind (exocarp), the fleshy edible portion (mesocarp), and a thin layer next to the seed coat (endocarp) (Biale and Young, 1971).
Avocado is a fruit with high nutritional value and numerous beneficial health effects (Meyer and Terry, 2010). The fruit is a rich source of fats, particularly of monounsaturated fatty acids. The most abundant fatty acid is oleic acid that is known to reduce inflammation, a risk factor for cardiovascular diseases, and beneficial effects on cancer (Yoneyama et al., 2007). Health benefits of avocados are due to the presence of numerous bioactive phytochemicals (Adikaram et al., 1992;Tabeshpour et al., 2017). The fruit contains rare sugars of high carbon number and is relatively rich in certain vitamins, dietary fibre, and minerals. The fruit has high oil content and low sugar, hence recommended as a high energy food for people with diabetics. Avocado is a climacteric fruit, with a marked rise in respiration rate, followed by a decline.
Genus Colletotrichum is composed of plant pathogens of worldwide importance, particularly causing anthracnose in several tropical fruit species. Anthracnose disease (Sivanathan and Adikaram, 1989), caused by Colletotrichum species, and the stem-end rot (Madhupani and Adikaram, 2017) incited by several fungal pathogens, including Lasiodiplodia theobromae, are major constraints to the avocado industry, causing heavy fruit losses after harvest and limiting their marketing potential and shelflife.
Anthracnose disease was believed to be caused by Colletotrichum gloeosporioides (Sivanathan and Adikaram, 1989) and C. acutatum (Hartill, 1991) for decades in the 19 th century. More recent molecular studies, have revealed the association of over 15 Colletotrichum species with the anthracnose disease from both avocado growing and marketing countries of the world. Among them, the most significant number of species recorded, from a single country, was Israel where multi-locus phylogenetic analyses using ITS, act, ApMat, cal, chs1, GAPDH, GS, HIS3, TUB2 gene/ markers, identified eight previously described species, C. aenigma, C. alienum, C. fructicola, C. gloeosporioides sensu stricto, C. karstii, C. nupharicola, C. siamense, C. theobromicola, and a novel species, C. perseae, as causing avocado anthracnose, confirming their pathogenicity (Sharma et al., 2017). Talhinhas et al. (2002) were the first to carry out multilocus-based phylogenetic analysis for Colletotrichum species. Using multiple sequence alignment, past phylogenetic analyses have revealed that the genus Colletotrichum comprises of eleven species complexes and 23 singletons where the C. gloeosporioides species complex is collective of C. gloeosporioides s.s and 51 closely related species (Weir et al., 2012;Jayawardena et al., 2020). Similarly, C. acutatum is now considered a species complex consisting of 41 species that include C. acutatum s.s and its close relatives (Jayawardena et al., 2020).
The present study re-evaluated the Colletotrichum species associated with avocado anthracnose in Sri Lanka by multigene DNA sequence approach, using 35 isolates from diseased fruits collected in two major avocado producing provinces and also the semi-systemic nature of internal symptom development.

Isolation of Colletotrichum
Ripe avocado fruits showing characteristic symptoms of anthracnose disease were collected from wholesale fruit stores or retail outlets in two main avocado-producing and distributing areas, Kandy (Central Province), and Kurunegala (North-Western Province) Districts, of Sri Lanka, over two fruit seasons in 2015 -2016. Diseased fruits were brought in sealed polythene bags to the Plant Pathology laboratory at the Department of Botany, University of Peradeniya, Sri Lanka.
Colletotrichum was isolated from anthracnose lesions on 35 infected avocado fruits. Segments (5 × 5 mm 2 ) of infected tissues, cut from the advancing margin of anthracnose lesions in the fruit peel, were surface sterilized in 1% sodium hypochlorite (Clorox, USA) for 1 -3 min followed by rinsing twice in sterile distilled water (SDW). The excess liquid in tissue segments was removed by placing them on sterile filter papers. Tissue pieces (4 per plate) were aseptically transferred onto PDA medium, supplemented with 50 µg mL -1 tetracycline to suppress bacterial growth. The plates were incubated at 28 ℃ for 5 -7 days. The 35 isolates obtained were sub-cultured by transferring discs (6 mm diameter) of mycelium onto fresh PDA plates and allowed to grow at 28 ℃ for 14 days.

Preparation of mono-conidial cultures
A suspension of conidia of each isolate was prepared by suspending the mycelium scraped from 10 -day old cultures in sterile distilled water (SDW) and filtering through sterile glass wool. A loop-full of each suspension was streaked over thin tap water agar plates. After incubation the plates for 18 h at 28 ℃, a small piece of agar with a single germinated conidium, located by moving the objective lens (× 25) of a light microscope (Olympus CX 22) along the streak line, was cut and transferred onto fresh PDA. The plates were incubated for seven days. Pure cultures were maintained in microcentrifuge tubes (1.5 mL) containing 800 µL sterile PDA at 15 ℃ (Prihastuti et al., 2009) to be used in subsequent studies.

DNA extraction, PCR amplification and sequencing
Fifteen isolates, selected randomly from the initial 35 isolates, were used for molecular studies. DNA was extracted using the protocol described by Živković et al. (2010). Aerial mycelium (0.5 g), scraped from seven days old cultures, using a sterile inoculation loop, was placed in a sterile microcentrifuge tube (1.5 mL) containing 300 µL of extraction buffer (0.2 M Tris-HCl, 0.25 M NaCl, 25 mM EDTA, and 2% SDS, pH 8.5) and crushed well. Uncapped tubes were then placed in a boiling water bath for 5 min and allowed to cool to 25 ℃. Aliquots (200 µL) of phenol, equilibrated with the extraction buffer (vol/vol), and chloroform (200 µL) were added. The tubes were vortexed for 2 -3 min and centrifuged at 7,647 g for 5 min. The supernatant was transferred into a new 1.5 mL microcentrifuge tube containing 200 µL of chloroform and vortexed for 30 s followed by centrifugation at 7,647 g for 15 min. The supernatant was pipetted out into a new 1.5 ml tube and 200 µL of ice-cold isopropanol was added. Tubes were inverted several times for DNA to precipitate and centrifuged at 7,647 g for 15 min. The pellet was retained and washed with 400 µL of ice-cold ethanol and centrifuged at 7,647 g for 5 min. The pellet was air-dried for 10 min and re-suspended in 50 µL in low-TE buffer (10 mM Tris-HCl and 0.1 mM EDTA, pH 8.5) to dissolve DNA and stored at -20 ℃.
All PCR amplifications were carried out, as described by Weir et al. (2012). The PCR products were sequenced for both directions using Applied Biosystems, 3500 Genetic Analyzer at the Department of Molecular Biology and Biotechnology, Faculty of Science, University of Peradeniya, Sri Lanka.

Pathogenicity test
Anthracnose lesions in ripe avocado fruits collected in the study were examined, and the symptoms were recorded. Isolates of C. endophytica and C. siamense were grown on pure culture. Freshly harvested fruits of uniform size, devoid of blemishes or any disease symptoms, were chosen for artificial inoculation. Suspensions of conidia of an isolate each of C. endophytica and C. siamense were prepared by scraping mycelium, suspending them in sterile distilled water and filtering through glass wool. The concentration of conidia was adjusted to 1 x 10 6 mL -1 . Four drops (20 µL) of conidia from each isolate were applied on to four equally distanced sites along the fruit surface, from the stem-end to the blossom-end. Six replicate fruits were used for each isolate. The fruits treated with drops of SDW were maintained as controls. Inoculated and control fruits were incubated in separate trays, lined with moistened tissues, and covered with glass plates, at 28 -30 ℃. The fruits were examined daily and the symptoms, when appeared, were compared with those of the original diseased fruits in which the disease was initially observed. The pathogens were reisolated from symptomatic fruits on PDA. Morphological features of the colonies and, asexual reproductive stages of the isolate, were compared with those of the original isolates used for inoculation.

Data analysis
The species affiliations and identities were determined through similarity-based searches of the NCBI GenBank Database (htttp://www.ncbi.gov). Based on the identifications that resulted from the BLAST search, a combined phylogenetic analysis for ITS, TUB2, and GAPDH was conducted including the authenticated sequences of the members belonging to the C. gloeosporioides complex obtained from the GenBank (Table 1). Bayesian inference analysis was performed for the combined matrix. The best fitting substitution model was determined with jModelTest v.2 (Darriba et al., 2012) using the Akaike information criterion. The nucleotide substitution model General Time Reversible was selected. Bayesian inference was conducted to obtain posterior probabilities using MrBayes ver. 3.2.6 (Huelsenbeck and Ronquist, 2001;Ronquist and Huelsenbeck, 2003) with 10,000,000 generations Markov chain Monte Carlo chains with a sampling frequency of every 1,000 generations. The initial 25% samples from each run were discarded as burning. A majority rule consensus tree was calculated using the remaining trees to obtain the posterior probabilities for each node. The resulting tree was visualized and edited in FigTree ver. 1.4.3 (Rambaut and Drummond, 2016). Colletotrichum hippeastrum (isolate CBS 241.78) was used as the out group. All the sequences, generated during the study and used in multi-gene analyses, were deposited in GenBank and the accession numbers are given in Table 1.

Isolation of pathogen
Colletotrichum infections in ripe fruits appeared as blackish brown, and circular lesions of different sizes with slightly irregular margins scattered over the peel of ripe fruit. Salmon colored, sticky conidia masses, resembling slimy droplets, were seen in the centre of older lesions (Figure 1). Lesions enlarged in size widening their diameter, up to 3 -5 cm or more. Multiple infections in closer proximity tended to coalesce forming larger diseased areas.
Thirty five isolates were obtained from the anthracnose lesions in fruits collected from different locations in the Central Province where avocados are mostly produced and, also from the North Western Province of Sri Lanka. All 35 isolates produced oblong conidia, and the colonies of majority of the isolates consisted of pink conidial masses. The isolates were identified to the Genus Colletotrichum from their cultural and conidial characteristics.

Phylogenetic analyses
The combined data set for ITS, TUB2, and GAPDH sequences consisted of 1286 bps. The phylogenetic tree that resulted from the Bayesian analysis is given in the Figure 2. All members of the C. gloeosporioides complex formed a monophyletic group while all the Sri Lankan Colletotrichum isolates identified as C. siamense and C. endophytica formed a separate monophyletic clade. However, the both clades received low support, posterior probability of 0.84 and 0.73 respectively and the clades are unresolved.

Morphological characteristics of Colletotrichum siamense
The colonies on PDA first appeared white and turned pale yellow to grey with time. Aerial mycelium was greyish white, dense, wooly, or cottony with very few conidial masses at the center. Sectoring was observed in some cultures. Conidia were cylindrical with slightly rounded ends and sometimes tapering towards one end and measured 20.8 -30.4 µm × 7.0 -8.4 µm. Appressoria were ovoid or irregularly lobed, 9.2 -11.1 µm in diameter. Some cultures produced both ovoid and lobed appressoria while others produced only ovate. Appressoria colour ranged from brown to dark brown (Figure 4).

Morphological characteristics of Colletotrichum endophytica isolate
Colonies on PDA first appeared white and the center of colony of some cultures became grey to ash color with time. Aerial mycelium at the periphery was white, dense, wooly or cottony with numerous conidia masses. Some isolates produced sectoring after sub-culturing. Conidia  Table 1: Accession numbers of authenticated sequences of the C. gloeosporioides species complex obtained from the GenBank and the sequences generated from the present study, used for phylogenetic analysis.    were cylindrical with slightly rounded ends. Length and breadth of the conidia varied from 20.7 -32.6 µm (length) and 6.9 -9.9 µm (breadth). Appressoria were ovoid or irregularly lobed. All cultures produced both ovoid and lobed appressoria. Appressoria were initially pale brown color and later turned dark brown ( Figure 5).

Pathogenicity test
The two Colletotrichum species could be repeatedly isolated from diseased avocados in the study, indicating their consistent presence in infected fruits. Healthy fruits, artificially inoculated with the two fungi, developed typical anthracnose symptoms that were observed originally, 6 -7 days after inoculation. The control fruits did not develop any disease symptoms. Morphological characteristics of the colony and conidia of the two fungi re-isolated were similar to those of the original isolates used for inoculation.

DISCUSSION
Colletotrichum gloeosporioides (Sivanathan and Adikaram, 1989) and C. acutatum (Silva-Rojas and Ávila-Quezada, 2011) have been believed for decades to be the pathogens causing anthracnose disease in avocado and, certain other tropical and sub-tropical fruit species. The two species show morphological similarities. The conidia morphology of C. acutatum being the only distinguishable, but often inconsistent, character between them. Morphology and the development of reproductive structures have been utilized in the characterization of the genus Colletotrichum and its teleomorph, Glomerella, by taxonomists. The variability of morphological characters with changing environmental and growth conditions makes them unreliable as taxonomical criteria. Molecular-based methods are presently considered advantageous in the species level identification of the genus Colletotrichum (Weir et al., 2012) than morphological features. In general, morphological differences among isolates within the genus Colletotrichum do not correlate with molecular differences.
The present study used multigene sequence analysis with two coding genes, TUB2, GAPDH, and the nuclear ITS region that contributes to a higher resolving ability for species level identification of Colletotrichum causing avocado anthracnose in Sri Lanka. Two species, C. endophytica and C. siamense, were identified as casual agents. ITS region has been useful only for the identification of Colletotrichum isolates into the species complex level (Prihastuti et al., 2009). Colletotrichum endophytica, belonging to the C. gloeosporioides species complex, was isolated as an endophyte in Pennisetum purpureum. Colletotrichum endophytica was later reported as causing anthracnose disease in Camellia sinensis  and chili in China, black pepper in India (Chethana et al., 2015) and more recently in mango also in Southern China (Li et al., 2019).
TUB2 sequences, generated for C. endophytica in the present study and deposited in the GenBank, would therefore be a valuable source of reference sequence material for future studies of Colletotrichum. The present study did not encounter either C. gloeosporioides or C. gigasporum that were previously identified to be associated with the avocado anthracnose in Sri Lanka (Hunupolagama et al., 2015).
Interestingly, the authenticated isolate of C. endophytica [CAUG28 (Diao et al., 2017)] was also grouped, within the C. endophytica clade, together with the Sri Lankan isolates. However, the authenticated C. siamense, which is also an ex-type (ICMP_18578) Weir et al. (2012), grouped in the main clade with the other species of the C. gloeosporioides complex, separate from the Sri Lankan C. siamense isolates.  Based on multi-locus phylogenetic analyses, eight previously described species and a novel species (C. perseae) were identified as avocado anthracnose pathogens in Israel (Sharma et al., 2017). In addition, several more Colletotrichum species were reported causing anthracnose disease in avocado from countries worldwide, raising the total number of species to over fifteen. The inconsistency of the Colletotrichum species reported warrants further studies on the avocado-Colletotrichum pathosystem in avocado producing and marketing countries of the world.
Colletotrichum siamense was first described as a causal agent of coffee berry anthracnose from Northern Thailand (Prihastuti et al., 2009). The species was later recorded on many hosts across tropical and subtropical regions without any host specificity, peach (Yang et al., 2009), mango (Phoulivong et al., 2010;Udayanga et al., 2013), custard apple, Cerbera sp., figs, and papaya (Rampersad, 2011;Udayanga et al., 2013) and Pongamia pinnata (Dwarka et al., 2016). The understanding of C. siamense is still in a state of confusion. The present study identified variable cultural, conidial and appressorial characters within C. siamense suggesting that C. siamense might not be a single species. Molecular analysis of 85 Colletotrichum isolates from fruit crops in India, using ApMat marker, resolved C. siamense to be a species complex (Sharma et al., 2015). However, Liu et al. (2016), following a molecular analysis based on Genealogical Concordance Phylogenetic Species Recognition (GCPSR) and coalescent methods, concluded that C. siamense sensu lato is a single species rather than a species complex.
The present phylogenetic analyses have shown a separation of the Sri Lankan C. siamense isolates from the authenticated C. siamense which may supports the idea that C. siamense might well be a species complex (Sharma et al., 2015) rather than a single species. Similarly, the authenticated C. endophytica together with the Sri Lankan C. endophytica isolates are placed outside the main C. gloeosporioides complex, indicating the diversity of these species at the molecular level. The separation of the Sri Lankan Colletotrichum isolates from the rest of the species of the C. gloeosporioides complex indicates that the Sri Lankan isolates are genetically diverse. While studying the DNA sequence alignments, a notable difference of all 15 Sri Lankan isolates was an INDEL of six base pairs, either CACACG or CACATG, that was unique only to the local isolates in the GAPDH region (Figure 3). This INDEL was also shared by C. perseae, a novel species that was recently identified and described as causing avocado anthracnose disease in Israel (Sharma et al., 2017).
The present study reports C. endophytica for the first time from avocado anthracnose disease. This would necessitate new disease management strategies for avocado anthracnose since C. endophytica is new to avocado. This may also increase its importance as a quarantine pathogen (Yan et al., 2015). These re-iterate once again the importance of accurate identification of causal agents in designing disease management strategies.

CONCLUSION
In conclusion, the present study identified C. endophytica and C. siamense as pathogens avocado anthracnose where this is the first report of C. endophytica from avocado. The study also reports for the first time the semi-systemic nature of symptom development in the disease.