Distinguishing three Dragon fruit (Hylocereus spp.) species grown in Andaman and Nicobar Islands of India using morphological, biochemical and molecular traits

Dragon fruit (Hylocereus spp.), an important tropical fruit belonging to the family Cactaceae, is rich in essential nutrients such as vitamins, minerals, complex carbohydrates, dietary fibres and antioxidants. This study aims to distinguish three dragon fruit species well adapted to Andaman and Nicobar Island through morphological (34 quantitative and 26 qualitative traits), biochemical (5 traits) and molecular (14 ISSR primers) characterization. Morphological characterization revealed that presence of considerable amount of genetic variations among them especially for fruit characters viz., colour of peel and pulp. Cladode characters such as number of spines (3–5), length of areoles (mm) as 1–4, margin ribs of cladode (convex or concave) and its waxiness (weak or strong white waxy or light waxy) could be used for identification of three Hylocereus spp. under present study. Highest co-efficient of variation (%) obtained for pulp weight (88.7), whereas, lowest in distance of anthers belowstigma (3.3). Fruit and pulp weight (g) ranged from 26.5–419.3 and 10.3–258.8 with mean value of 204.8 and 125.3, respectively. Comparatively, high phenol (71.3–161.3) and flavonoid (26.6–508.2) content observed in peels than pulp (32.5–130.0 and 45.0–258.2) of fruit indicating higher antioxidant potential. Highest total carotenoids (µg 100 g−1), β-carotene (µg 100 g−1) and xanthophyll (µg g−1) content obtained in pulp on DGF3 (33.8), DGF4 (55.9) and DGF3 (32.7), whereas, in peel on DGF2 (24.3), DGF4 (18.5) and DGF2 (24.1), respectively. DPPH-based scavenging activity (%) revealed higher scavenging activity of peels (55.6–81.2) than pulp (36.0–75.3) extracts. Comparatively, ABTS-based scavenging activity (%) was found more than DPPH-based one. Sixteen ISSR primers screened, 14 were produced 178 reproducible amplified bands. Number of amplified bands varied from 5 in UBC887 to 19 in UBC811 with an average of 12.71 bands per primer. Range of polymorphic bands and % polymorphism observed were 1–13 and 20.0–92.8, respectively. The polymorphic information content value of ISSR marker ranged from 0.42 (UBC895) to 0.91 (UBC 856). Cluster analysis distinguished three different Hylocereus species on the basis of geographic origin and pulp colour by forming separate groups and two genotypes each showed 52% (DGF1 and DGF3) and 76% (DGF2 and DGF4) genetic similarity. Key traits identified for distinguishing three different Hylocereus species were: Pulp/ peel colour of fruits, number of spines and length of areoles in cladode. Genotypes with high carotenoid and xanthophylls content (DGF4 and DGF2) identified under present study may be of industrial importance for development of nutraceutical products to meet out the vitamin-A deficiency among humans in tropical regions needed future focus.


Results
Morphological characterization. Data recorded on 34 quantitative and 26 qualitative traits in four genotypes of three dragon fruit (Hylocereus spp.) species were presented in Tables 1 and 2. A characteristic view of three different Hylocereus species of dragon fruit on key traits of cladode, floral and fruit is presented in Figs. 1 and 2. Range of variation for dragon fruit (Hylocereus spp.) species using 34 quantitative traits under study is presented in Table 3. The highest co-efficient of variation (CV in %) was obtained for pulp weight (88.7) followed by fruit weight (85. 3 Biochemical characterization. Biochemical characterization of four genotypes of three dragon fruit (Hylocereus spp.) species with total phenol content (TPC), total flavonoid content (TFC), total carotenoid content (TCC), β-carotene, xanthophyll and colour values such as L, a, b, hue and chroma were presented in Table 4. Among four genotypes, DGF4 had highest phenolic content (mg GAE 100 g −1 ) in both peel (161.3) and pulp (130.0) extracts followed by DGF3 (118.8 and 103.8), DGF1 (42.5 and 71.3) and DGF2 (32.5 and 91.3) in pulp and peel extracts, respectively, with the coefficient of variation (CV) of 0.62. Highest flavonoid content (mg RE 100 g −1 ) was found in peel of DGF4 (508.2) followed by DGF3 (123.9), DGF1 (55.5), whereas lowest in DGF2 (26.6). In case of pulp, it was varied from 45.0 (DGF1) to 258.2 (DGF4). DGF3 pulp showed highest (33.8) carotenoids content (µg 100 g −1 ) followed by DGF2 (30.4) and DGF4 (30.0) with the CV of 0.24, whereas, in peel highest and lowest obtained on DGF2 (24.3) and DGF1 (4.82), respectively. The highest content of β-carotene (µg 100 g −1 ) was found in DGF4 (55.9 and 18.5) and DGF2 (53.2 and 16.4) than DGF3 (1.3 and 0.9) and DGF1 (1.2 and 0.2) in pulp and peel, respectively. Xanthophyll content (µg g −1 ) of DGF3, DGF2 and DGF4 pulp was found as 32.7, 29.8 and 29.5, respectively, whereas, in peel it was varied from 4.8 (DGF1) to 24.1 (DGF2). The scavenging activity (%) by DPPH and ABTS method varied between 36.0 (DGF4) to 75.3 (DGF2) and 48.3 (DGF4) to 87.9 (DGF2), respectively, in pulps and, 55.6 (DGF2) to 81.2 (DGF4) and 61.8 (DGF1) to 89.8 (DGF2), respectively, in peels of four genotypes with CV of 0.29 and 0.26, respectively (Fig. 3). The observed colour values of four dragon fruit genotypes were as: L (11.7-51.0 and 13.6-40. Molecular characterization. Details of polymorphism obtained for four genotypes of three dragon fruit (Hylocereus spp.) species subjected to 16 ISSR marker based genetic diversity study are presented in Table 5. Among 16 ISSR primers screened, 14 primers showed amplification and produced a total of 178 reproducible amplified bands. No amplification was obtained for two ISSR primers viz., UBC825 and UBC853 in all four genotypes, whereas, three ISSR primers viz., UBC815, UBC856 and UBC891 showed no amplification in two genotypes DGF2 and DGF4. The electrophoretic profile of ISSR markers study showed highly distinct and polymorphic banding pattern in primers UBC900, UBC811, UBC824 and UBC835 (Fig. 4). Number of amplified bands varied from 5 in UBC887 to 19 in UBC811 with an average of 12.71 bands per primer. Range of polymorphic bands and % polymorphism observed were 1-13 and 20.0-92.8, respectively. The polymorphic information content value of ISSR marker ranged from 0.42 (UBC895) to 0.91 (UBC 856). Dendrogram was generated by using UPGMA method of cluster analysis that differentiated all four dragon fruit genotypes into two clusters (Cluster I and II) at Jaccard's similarity coefficient value of 0.50 on the basis of geographical locations of them (Fig. 5). Among them, two genotypes each showed 52% (DGF1 and DGF3) and 76% (DGF2 and DGF4) genetic similarity.

Discussion
Dragon fruit (Hylocereus spp.) is a promising tropical fruit which can be cultivated in different tropical and subtropical parts of the world such as Southeast Asia and Central and South America. Its health benefits to human can be explained by its essential nutrients such as vitamins, minerals, complex carbohydrates, dietary fibres and antioxidants. Dragon fruit is also an essential source of betacyanin which serves as a red/ purple pigment with anti-oxidative properties. Morphological and genetic heterogeneity on fruit and other characters formed over www.nature.com/scientificreports/ the period in dragon fruit by high intra-and inter-specific hybridization made taxonomical confusion to identify them at species level. So, the present study was undertaken to identify the key traits in dragon fruit (Hylocereus spp.) using morphological, biochemical and molecular (ISSR marker) characterization for distinguishing them at species level. Thirty four quantitative and 26 qualitative traits of four dragon fruit genotypes belonging to three different species subjected to morphological characterization showed presence of considerable amount of genetic variations among them especially for fruit characters such as fruit weight, pulp weight, number of fruiting cycles, fruit yield, fruit shape, peel and pulp colour. In case of qualitative traits, all three species showed light reddish young cladode colour, edged sepal pattern, ovate shape of pericarpel bracts and milky white petals in flower. Traits such as cream coloured floral stigma lobe and medium sized seed in fruit with broad to medium fruit width of H. costariscensis could be useful in taxonomic aspects to differentiate the species with others. Cladode, floral and fruit characters of H. megalanthus such as margin ribs of cladode and its waxiness; sepal colour, colour of ring at base of reproductive organs in flower; fruit shape, position towards peel, pulp colour, peel colour and seed size in fruit are visible taxonomic traits to distinguish this species with other two Hylocereus spp., H. undatus and H. costariscensis.
Cladode characters such as cladode width (mm), distance between areoles (mm), number of spines, length of areoles (mm), margin ribs of cladode and its waxiness could be used for identification of Hylocereus spp. 18 20 ]. Natural flowering and production occurs during warmer months in dragon fruits 23,24 and the flowering season also varied between May to October across different regions of the world. Though flowering and fruiting occurs between April to November under Island condition, much variations were observed in flowering, fruiting period and fruiting cycles among three species. Fruit morphology such as size and colour of fruit is the main taxonomic evidences to differentiate among several Hylocereus spp. and also exhibits the external quality of fruit 22 . Colour of peel and pulp of fruit identified as one of the main key traits to differentiate the three different dragon fruit species as pinkish green peel with white pulp (H. undatus -DGF1), pink/ pinkish red peel with pink/ dark purple pulp (H. costariscensis -DGF2 and DGF4) and yellow peel with white pulp (H. megalanthus -DGF3) (Fig. 1). Maximum fruit set percentage (83.8%) observed for H. undatus might be due to its self-compatible nature 25 compare to other two species. The presence of high number of natural polliantors such as hawk moths and bats during night hours in the field is playing major role in fruit set in dragon fruit 26 , whereas, only honey bees observed to be the pollinators during early morning hours under Island condition. Therefore, introduction of hawk moths could ensure natural pollination and also artificial pollination may aid in increased fruit set percentage in dragon fruit genotypes. Further, the low average fruit yield of H. megalanthus -DGF3 and H. costariscensis-DGF4 could directly be linked to their number of fruiting cycles as 1 and 2, respectively with comparison to H. undatus-DGF1 (7) and H. costariscensis-DGF2 (8). Total soluble solids, being the most desirable character in view of consumers' preference, measured as °Brix, which can be affected by a set of factors such as genetic, climatic, soil, management, among others [27][28][29] . In the present study, the TSS ranged between 9.1 to 18.3 o B representing better fruit quality which evidenced by the earlier report that the TSS values between 11 to 15 o B have good market preference 30 .
Phenolic compounds are playing vital role in multiple biological activities such as anti-mutagenicity, anticarcinogenicity, anti-aging and also anti-oxidant in plant 31 . Phenolic acid (e.g. gallic acid) and polyphenol (e.g. flavonoids) are highly correlated with antioxidant activity as evidenced from earlier studies 32 . Phenol and flavanoid content at peel and pulp of fruit varied with low (H. undatus) to medium (H. megalanthus) and with medium/ high (H. costariscensis) could be used as taxonomic purposes to distinguish species at bio-molecular level. The present study revealed that inedible peels of dragon fruit had higher phenolic content as compared to edible pulps. Comparatively, higher amount of total phenolic content (mg GAE 100 g −1 ) obtained in present  Huge amount of differences in phenolic content obtained among four dragon fruits in the present study might be due to defend against or mitigate them from adverse effects of biotic and abiotic stresses of sub-tropic climate 34 . Further, high phenolic content of dragon fruit could definitely be a good source of polyphenol to be integrated www.nature.com/scientificreports/ into the human diet compare to common fruits 35 . Similar kind of results on phenol and flavonoid content of our study was reported by Ramli et al. 36 who obtained phenol content as 73.8 and 121.8 and flavonoid content as 145.9 and 510.7 in pulp and peel fractions, respectively. This might be due to the fact that non-flavonoid and flavonoid compounds found in pulps and peels, respectively 37 . Comparatively, the presence of high phenol and flavonoid content in peels than pulp of dragon fruit indicating higher antioxidant potential of peel extract in quenching free radicals. Generally, carotenoid protects the plant from photo oxidation 38 which is evidenced from its essential role in chloroplasts and chromoplasts 39 and as xanthophylls in chlorophyll 40 . Whereas, human bodies are also able to transform dietary carotenoids to biologically active vitamin A (retinol) and its derivative, one RE (retinol equivalent) corresponds to either 6 μg of dietary β-carotene or 12 μg of other dietary pro-vitamin A 41 . Dragon fruits identified with high carotenoid content such as DGF4 and DGF2 (H. costariscensis) than other two species could be an excellent source of provitamin A, because 200 g of this fruit (pulp and peel) would meet the daily vitamin A requirement of 700 and 900 μg day −1 for adult women and men, respectively 42 . So, these two genotypes may be having potential for the development of nutraceutical products to meet out the vitamin-A deficiency among humans in tropical regions.
Being most powerful scavenger of singlet oxygen, β-carotene's lower plasma levels can even lead to death 43 . In present study, the genotype with highest content of β-carotene DGF4 (55.86 µg 100 g −1 ) with dark purple pulp is www.nature.com/scientificreports/ having nearly 50 times more than white pulp fruits. Further, these white pulp fruits (DGF1 and DGF3) were low in xanthophyll content also. Therefore, the dragon fruit genotype 'DGF4′ could be an alternate good source for many common fruits such as apple (24), grape (39), Kiwi fruit (18), papaya (32) and pepper (44) 44 and underutilized fruits such as durian fruit (23) 45 , jack fruit (22) and guava (1.0) 46 . High carotenoids and xanthophylls content in DGF4, DGF2 and DGF3 may be due to high amount of chlorophyll synthesized in chloroplast as plant pigments which is responsible for colour of different capacity 47 .
The DPPH based scavenging activity (%) of four dragon fruit genotypes revealed the higher scavenging activity of peels (55.6-81.2) than pulp (36.0-75.3) extracts under present study. This values were not exactly corresponded well to levels of total phenol and flavonoid contents in both pulp and peel fractions of fruits. This non-significant differences observed in DPPH activity among genotypes, except DGF4 might be due to (1) the presence of lypophilic compounds in the fruits for TPC, (2) over-estimation of TPC by Folin-Ciocalteu reagent method 48,49 , (3) varying response of the Folin-Ciocalteu method to different phenolic compounds 50,51 and (4) removal of non-phenolic compounds (flavonoids) having antioxidant capacity during methanolic extraction. Further, the high TPC with low antioxidant capacity (AC) as observed under present study was also reported earlier in some underutilized fruits such as Garcinia, Nephelium and Syzygium fruits 52 .
Negative correlation between either phenolic or flavonoid content and ABTS activity observed under present study was also reported in earlier 53,54 . This variations could be due to (1) difference in number of phenolic groups of polyphenolic compounds which led to differently response to ABTS activity of pulp fraction 55 , and/ or (2) presence of high amount of reducing agents such as ascorbic acid, minerals and carotenoids in the fruits 48,49 , high protein content or genetic, agronomic and environmental influences 56 . In case of flavonoid compounds, both ; TCC = Total carotenoid content (µg g −1 ); β-Carotene (µg 100 g −1 ); Xanthophyll (µg g −1 ).    The white pulped fruits with low phenolic content might be due to non-betalainic phenolic compounds which could lead to lower radical scavenging activity of them and the major antioxidant capacity of pink or purple pulped fruits were due to the presence of betalains 57 .

Parameters/genotype
Morphological characterization along with molecular characterization using ISSR markers would provide strong base for unravelling the genetic diversity between the different genotypes of dragon fruit. ISSR profiling is efficient to reveal the genetic diversity in many crops [58][59][60] . In dragon fruit, utilization of ISSR markers were first reported in china by Tao et al. 61 to elucidate the genetic relationship of red pulp genotypes from white pulp ones. Range of number of polymorphic bands (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13) and % polymorphism (20-92.8%) obtained in this study was almost comparable with results of Tao et al. 61

Conclusion
Morphological, biochemical and molecular characterization of four dragon fruit (Hylocereus spp.) genotypes grown in Andaman and Nicobar Island revealed the presence of considerable amount of genetic variations among them which could be used as key traits for distinguishing three different species under present study. Cladode and fruit characters showed higher variability among morphological traits. Comparatively, the presence of high phenol and flavonoid content in peels than pulp of fruit indicating higher antioxidant potential of peel extract in quenching free radicals and was evidenced by higher DPPH-based scavenging activity of peels than pulp extracts. Comparatively, ABTS-based scavenging activity (%) found highest in DGF2 (87.9 and 89.8) and moderate in DGF1 (68.1 and 61.8) and DGF3 (56.4 and 82.4) in pulp and peel, respectively was more than DPPH-based one. ISSR-marker based clustering pattern clearly differentiated the genus Hylocereus at species level on basis of their geographic origin and pulp colour by grouping them separately. Key traits identified to  Molecular characterization. DNA samples were isolated from stem tissue of four dragon fruit genotypes using a modified CTAB method 73 . The quality and concentration of the DNA were confirmed by electrophoresis on 1% agarose gels. Selection of primers of Inter Simple Sequence Repeats (ISSR) marker were done to analyze genetic variation among four genotypes 61 and their details of primer sequence and annealing temperature are listed in Table 5. PCR reaction mixtures were prepared in a final volume of 10 μl, containing 10 µl volume containing 1 µlDNA (20 ng/µl), 5.0 µl PCR mix (Qiagen), 0.8 μl primer and 3.2 μl nuclease-free water and PCR amplification were performed in C1000 Touch Thermal Cycler (Bio-Rad) with an initial denaturation step of 5 min at 94 °C, followed by 40 cycles of denaturation for 45 s at 94 °C, annealing for 45 s at primers specific temperatures and extension for 1 min at 72 °C, and ended with a final extension step of 5 min at 72 °C. Banding pattern from each primers were resolved in 2% agarose gels using Gel Doc XR + Imager (Bio-Rad) and all reactions were repeated thrice to confirm the consistency in banding pattern before subjected to scoring.
Statistical analysis. Data observed on 34 quantitative morphological traits of dragon fruit genotypes were subjected to basic statistical analysis using software PAST 3 74 . Molecular data were observed on basis of banding patterns by scoring the presence of band as '1′ and absence as '0′ for particular fragment in lane. Polymorphism information content (PIC) was calculated following Weir 75 method according to the formula: PIC i = 1 − ∑(P ij ) 2 , where, Pij is the frequency of the jth locus for the ith marker and is summed over i loci/marker. The Un-weighted Paired Group Method with Arithmetic means (UPGMA) based Jaccard's Similarity Coefficient was used for cluster analysis. This computation was performed by using NTSYS-pc v2.2 76 .
Ethical standards. The authors declare that there is no ethical issue(s) in this study. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.