Diaporthe species causing stem gray blight of red-fleshed dragon fruit (Hylocereus polyrhizus) in Malaysia

This study aimed to characterize the new fungal disease on the stem of red-fleshed dragon fruit (Hylocereus polyrhizus) in Malaysia, which is known as gray blight through morphological, molecular and pathogenicity analyses. Nine fungal isolates were isolated from nine blighted stems of H. polyrhizus. Based on morphological characteristics, DNA sequences and phylogeny (ITS, TEF1-α, and β-tubulin), the fungal isolates were identified as Diaporthe arecae, D. eugeniae, D. hongkongensis, D. phaseolorum, and D. tectonendophytica. Six isolates recovered from the Cameron Highlands, Pahang belonged to D. eugeniae (DF1 and DF3), D. hongkongensis (DF9), D. phaseolorum (DF2 and DF12), and D. tectonendophytica (DF7), whereas three isolates from Bukit Kor, Terengganu were recognized as D. arecae (DFP3), D. eugeniae (DFP4), and D. tectonendophytica (DFP2). Diaporthe eugeniae and D. tectonendophytica were found in both Pahang and Terengganu, D. phaseolorum and D. hongkongensis in Pahang, whereas D. arecae only in Terengganu. The role of the Diaporthe isolates in causing stem gray blight of H. polyrhizus was confirmed. To date, only D. phaseolorum has been previously reported on Hylocereus undatus. This is the first report on D. arecae, D. eugeniae, D. hongkongensis, D. phaseolorum, and D. tectonendophytica causing stem gray blight of H. polyrhizus worldwide.

, which gradually turned into a dark-brown sunken lesion and demonstrated dampening (Fig. 3B 2 ). As the disease progressed, the lesion became apparently dry and turned gray (Fig. 3B 3 ). Then, it expanded periodically, and tiny black pycnidia appeared on the area of the lesion (Fig. 3B 4

Discussion
The present study reported on stem gray blight, which is a new emerging disease infecting H. polyrhizus plantations in Malaysia. The five species of Diaporthe, namely, D. eugeniae (group 1), D. phaseolorum (group 2), D. tectonendophytica (group 3), D. hongkongensis (group 4), and D. arecae (group 5), were identified to be the causal agents of the disease. The Diaporthe species may act as a plant pathogen or a saprophyte or an endophytic symbiont [31][32][33][34] , however, several studies have reported that it is the genus responsible for multiple destructive diseases, such as root and fruit rots, dieback, stem cankers, leaf spots, leaf and pod blights, and seed decay 31,33,[35][36][37][38][39] . A total of nine Diaporthe isolates were recovered from the blighted stem of H. polyrhizus. Based on their morphological characteristics, all the isolates produced both α-conidia and β-conidia, except for the D. arecae isolate, of which β-conidia was not observed. α-and β-conidia are the key characteristics for the identification of Diaporthe 33,40 . The formation of β-conidia can sometimes be rare or absent in certain species of Diaporthe 41 . According to Tuset and Portilla 42 and Diogo et al. 43 , for some Diaporthe species (e.g. Phomopsis amygdali), the formation of β-conidia can only be observed in pycnidia on the host but not in pycnidia in the culture plate.
Based on the similarities and differences of their macroscopic and microscopic characteristics, the isolates were assigned to five different groups. Among the groups, significant differences were observed in the number of α-conidia guttules and their size ( Table 1). Gomes et al. 34 revealed that both characteristics can be varied among the Diaporthe species. The isolates from group 1 (D. eugeniae) tended to produce bi-and multi-guttules, whereas the other isolates only produced bi-guttules of α-conidia. The size of the guttules of α-conidia varied among the groups. The isolates from groups 1 and 5 (D. eugeniae and D. arecae) produced significantly smaller guttules compared with those produced by isolates from groups 2, 3, and 5 (D. phaseolorum, D. tectonendophytica, and D. hongkongensis) ( Table 1). The guttule is defined as a small drop or particle in a spore resembling a nucleus 44 . Moreover, the morphology of α-conidia of the D. eugeniae, D. hongkongensis, and D. arecae isolates was tapered toward the ends compared with the D. phaseolorum and D. tectonendophytica isolates, the ends of which were bluntly rounded (Fig. 1). This finding was in agreement with those of Santos et al. 38 , Dissanayake et al. 45 , Doilom et al. 46 , and Lim et al. 47 . A significant difference was also observed in the length of β-conidia, of which the D. eugeniae isolates produced longer β-conidia than other isolates from different groups. Conidial mass exudation can be observed in the isolates of D. eugeniae, D. hongkongensis, and D. arecae. Contrarily, it was not observed in the isolates of D. phaseolorum and D. tectonendophytica. According to Machowicz-Stefaniak et al. 48 , the Diaporthe species require temperatures ranging from 22 to 28 °C for the optimal growth, sporulation, and rate of conidia release of conidiomata. As applied in the present study, the addition of carnation leaves to the growing medium as substrates has been recommended to improve the sporulation of the Diaporthe species 49,50 .
Aside from the microscopic characteristic, the cultural characteristics of all isolates in this study also varied among the groups. The color of the colonies ranged from whitish, grayish, brownish, to olive green. Due to this inconsistency, cultural characteristic is commonly considered as a less important criterion in distinguishing species within Diaporthe as it can be influenced by several environmental factors, such as light and temperature 34 . Based on the results obtained, morphological characteristics alone were insufficient to identify all the isolates up to the species level due to the complexity of the genus. This finding was in agreement with that of Lim et al. 47 who revealed that the morphological method alone is not informative for the species identification of Diaporthe due to pleomorphism and overlapping characteristics 43,51,52 .
With the advances in molecular techniques, DNA sequences and multigene phylogenetic analysis of ITS, TEF1-α, and β-tubulin were employed to support the morphological identification of the Diaporthe isolates in this study. The result of the BLAST search and phylogenetic inference indicated that the use of all the three genes resolved identification of the Diaporthe isolates. Aside from the present study, ITS, TEF1-α, and β-tubulin were extensively applied to delineate species within Diaporthe 46,53,54 . The ITS region served as an identification guide for the Diaporthe species 33 . It was also considered as a fungal barcode in distinguishing genera and species owing to its easy amplification and ability to provide preliminary screening of fungal classification 55,56 . However, the tree constructed based on ITS sequences alone may be doubtful and not demonstrate clear phylogenetic relationships due to the lack of interspecific variation or even deceptive in some fungi 57 . Thus, TEF1-α and β-tubulin were added to support the phylogenetic analysis of ITS in delimiting the species of the Diaporthe isolates. TEF1-α comprises an essential part of the protein translation machinery, and highly informative at the species level; moreover, non-orthologous copies have not been detected in Diaporthe 58 . β-tubulin was utilized as an alternative phylogenetic marker to specify Diaporthe as it contains fewer ambiguously aligned regions and exhibits less homoplasy among the genus 59 . Collectively, phylogenetic analysis of a combined dataset of ITS, TEF1-α, and β-tubulin was conducted in this study to overcome the ambiguity that could have emerged in the single gene analysis. Santos et al. 60 stated that the combined phylogenetic tree commonly provides a better resolution for the identification of the Diaporthe species compared with the single gene analysis.
All the tested isolates of Diaporthe exhibited varying lengths of lesion on the inoculated stems of H. polyrhizus, of which isolate DF1 (D. eugeniae) was found to be the most virulent. The fungus can act as a pathogen or a saprophyte and was reported to cause stem-end rot on mango (Mangifera indica) 47 . It also occurs as a saprophyte on cloves (Eugenia aromatica) 34 . This study discovered a new host and disease caused by D. eugeniae. The association of D. phaseolorum with dragon fruit was not new, because recently, this pathogen was reported to cause stem rot on Hylocereus undatus in Bangladesh 12 . However, the symptoms described were slightly different from those observed in the present study. It appeared as a yellow spot with a chlorotic halo in the previous report, but www.nature.com/scientificreports/ in the present study, chlorotic halo was not observed; rather, a reddish border surrounded the lesion. Similarly, gray to black pycnidia were scattered on the surface of the lesion. Aside from the dragon fruit, D. phaseolorum was reported as a causal agent of pod and stem blight, stem canker, and seed rot on soybean and trunk disease on grapevine 38,45,61,62 . It was also found to be an endophyte on Kandelia candel by Cheng et al. 63 . Similar to D. eugeniae, the present study highlighted H. polyrhizus as a new host associated with D. tectonendophytica as it causes stem gray blight. Contrarily, a study by Doilom et al. 46 demonstrated the role of D. tectonendophytica as an endophyte occurring on teak (Tectona grandis) in Thailand. The capability of D. hongkongensis to act as a pathogen is undeniable as the fungus has been reported to cause severe diseases on a number of host plants, such as stem-end rot on kiwifruit 64 , dieback on grapevine 45 , and shoot canker on pear 65 . Meanwhile, D. arecae has been reported to be pathogenic on M. indica 47 , Areca catechu 34 , and Citrus 66 . D. hongkongensis and D. arecae were first reported on H. polyrhizus worldwide especially in Malaysia.
The occurrence of the disease in two different locations in Malaysia indicates its possibility to spread worldwide. Aside from Diaporthe, dragon fruits in Malaysia also suffer from multiple diseases caused by other fungi. Among these diseases are anthracnose caused by C. gloeosporioides 22,23 and C. truncatum 24 ; stem necrosis by Curvularia lunata 25 ; stem canker by N. dimidiatum 26 ; stem rot by Fusarium proliferatum 27 and Fusarium fujikuroi 28 ; reddish brown spot by Nigrospora lacticolonia and N. sphaerica 30 ; and stem blight by F. oxysporum 29 .
This study provides overview of the five different species of Diaporthe causing stem gray blight on H. polyrhizus in Malaysia. It improves our knowledge on the symptomatology of the disease and identity of the pathogens through morphological and molecular analyses. The findings may be essential to strategize effective disease management for stem gray blight on H. polyrhizus and for quarantine restrictions.

Materials and methods
Fungal isolation. In November 2017 and July 2018, nine gray blighted stems from the different plants of H. polyrhizus were collected from Bukit Kor, Terengganu, Malaysia, and the Cameron Highlands, Pahang, Malaysia. The symptomatic samples were brought back to the laboratory for isolation. One lesion per stem exhibiting the same symptom was selected for fungal isolation. The lesion consisting of diseased and healthy parts was excised (1.5 cm 2 ) and surface-sterilized with 70% ethanol for 3 min. Then, the samples were soaked in 10% sodium hypochlorite (1% NaOCl) for 3 min and rinsed with sterile distilled water three times consecutively for 1 min each. The sterilized samples were air-dried on the sterile filter papers before being transferred to PDA plates. The inoculated plates were incubated at 25 °C ± 2 °C for 2 to 3 days. Pure cultures of fungal isolates were obtained via hyphal tip isolation and were used for morphological and molecular analyses.
Morphological identification. Each fungal isolate obtained was cultured on PDA and incubated at 25 °C ± 2 °C for 7 days. Macroscopic characteristics, such as colony appearance and pigmentation, were recorded. CLA was utilized to induce the formation of pycnidial conidiomata, and the inoculated plates were incubated at 25 °C ± 2 °C for 7 days. The morphology of α-and β-conidia was observed from the pycnidial conidiomata. The other microscopic characteristics observed were conidiophores and conidiogenous cells. The length and width of 30 randomly selected conidia and the size of the guttules of 30 randomly selected α-conidia were measured and recorded. The differences in the length and width of conidia and the size of the guttules of α-conidia were evaluated via one-way ANOVA. In addition, the means of both parameters were compared via Tukey's test (p < 0.05) using the IBM SPSS Statistics software version 24.

Molecular identification and phylogenetic analysis.
The identity of all the fungal isolates was further confirmed by molecular characterization. The isolates were grown in potato dextrose broth (PDB) and incubated at 25 °C ± 2 °C for 7 days. Fungal mycelia from PDB were homogenized under liquid nitrogen to obtain fine powder. A total of 60 mg fine powder was transferred into a 1.5 mL microcentrifuge tube, and the genomic DNA of the fungal isolates was extracted using the Invisorb Spin Plant Mini Kit (Stratec Biomedical AG, Birkenfeld, Germany), following the manufacturer's protocols. The primers of ITS5/ITS4 67 , EF1-728/EF1-986 68 , and BT2a/ BT2b 69 were used for the amplification of ITS, TEF1-α, and β-tubulin, respectively. A total of 50 µL reaction mixture was prepared, which contained 8 µL of green buffer (Promega, USA), 8 µL of MgCl 2 (Promega, USA), 1 µL of deoxynucleotide triphosphate polymerase (dNTP) (Promega, USA), 8 µL of each primer (Promega, USA), 0.3 µL of Taq polymerase (Promega, USA), 1 µL of genomic DNA, and sterile distilled water. Polymerase chain reaction (PCR) was performed using MyCycler Thermal Cycler (BioRad, Hercules, USA) under the following conditions: initial denaturation at 95 °C for 4 min, followed by 35 cycles of denaturation at 95 °C for 35 s, annealing at 54 °C (ITS)/57 °C (TEF1-α)/58 °C (β-tubulin) for 1 min, extension at 72 °C for 90 s, and final extension at 72 °C for 10 min. The PCR product was separated by running it in 1.0% agarose gel (Promega, USA) stained with HealthView Nucleic Acid Stain (Genomics, Taiwan) at 90 V and 400 mA for 90 min. The 100 bp DNA ladder (Thermo Scientific, USA) was used as a marker to estimate the size of the amplified PCR products. The PCR products were sent to a service provider (First BASE Laboratories Sdn Bhd, Seri Kembangan, Malaysia) for DNA sequencing.
The obtained sequences were aligned using the Molecular Evolutionary Genetic Analysis software (MEGA7) 70 . After pairwise alignment, the BLAST algorithm (https ://blast .ncbi.nlm.nih.gov/Blast .cgi) was used to compare the generated consensus sequences with other sequences in the GenBank database. The sequences obtained were deposited in the GenBank database.
The isolates in the present study and reference sequences used in the phylogenetic analysis are presented in Table 3. Multiple sequence alignments of fungal isolates and reference isolates were generated using the MEGA7 software. Phylogenetic analysis was conducted using the maximum likelihood (ML) method in MEGA7. The  Pathogenicity test. The pathogenicity test was conducted on 18 healthy stems of H. polyrhizus for all the obtained fungal isolates. Conidial suspension was prepared by flooding the 7-day-old PDA culture with sterile distilled water, and the concentration was adjusted to 1 × 10 6 conidia/mL using a hemocytometer (Weber, Teddington, UK). The stems were surface-sterilized with 70% ethanol, and 0.1 mL of conidial suspension was utilized for inoculation using a disposable needle and syringe. Likewise, the control points were treated with sterile distilled water. On each stem, three points were used to inoculate fungal isolate and one point for control. Each fungal isolate was tested in three replicates, and the pathogenicity tests were conducted twice. All the inoculated plants were placed in a plant house in the School of Biological Sciences, USM, and incubated at 26-32 °C for 21 days. The progression of the disease symptom was observed daily. The lesion length was measured and recorded after 3 weeks of inoculation. The differences in the lesion length were evaluated via one-way ANOVA, and the means were compared via Tukey's test (p < 0.05) using the IBM SPSS Statistics software version 24. For the fulfillment of Koch's postulates, the fungal isolates were reisolated from symptomatic inoculated stems and reidentified by morphological characteristics.