Enhancement of phenolics content and biological activities of longan (Dimocarpus longan Lour.) treated with thermal and ageing process

This study is the first to compare the chemical compositions and biological activities of a conventional dried Dimocarpus longan with a novel black D. longan that underwent a thermal ageing process. Pericarp, aril, and seed of both D. longan were macerated in 95% v/v ethanol. Their chemical compositions were investigated using a Folin–Ciocalteu assay, aluminum chloride assay, and high-performance liquid chromatography. Antioxidant activities were evaluated in terms of radical scavenging and iron (III) reduction capacity. An enzyme inhibition assay was used to evaluate the hyaluronidase inhibition. Inflammatory cytokine secretion was evaluated with an enzyme-linked immunosorbent assay. After being exposed to a heating and ageing procedure, gallic acid and ellagic acid content were increased tenfold, while the corilagin content was doubled. Black D. longan seed extract was the most potent anti-hyaluronidase and antioxidant with the strongest free radical scavenging and reduction power, while black D. longan aril extract resulted in the highest inhibition of inflammatory cytokine secretion. Black D. longan contained more biologically active compounds and possessed more potent biological activities than conventional dried D. longan. Therefore, thermal ageing treatment is suggested for producing black D. longan, for which seed extract is suggested as a cosmeceutical active ingredient and aril extract for anti-inflammation.

than the dried D. longan extracts. The color of pericarp extracts was the darkest, followed by the extract from seed and aril. Yields of each D. longan extract are shown in Fig. 2. The aril yielded the highest extract content, followed by pericarp and seed. The highest yield of the extract was obtained from black D. longan aril (21.6% w/w), followed by dried D. longan aril (17.6% w/w), dried D. longan pericarp (13.8% w/w), black D. longan aril (11.0% w/w), black D. longan seed (6.6% w/w), and dried D. longan seed (3.6% w/w). It was highlighted that black D. longan yielded higher extract content compared to dried D. longan in aril and seed but not in pericarp. The explanation is that the pericarp of dried D. longan lost more water content than that of black D. longan, which underwent a heating and ageing process in conditions that kept the relative humidity constant at 75%. Therefore, the mass of initial pericarp material of dried D. longan was lower and led to a higher extract yield since the yield was calculated from the initial pericarp material used in the extraction process.

Chemical compositions of dried D. longan and black D. longan extracts. Dimocarpus longan
extracts were investigated for contents of total phenolic compounds, total flavonoid, gallic acid, corilagin, and ellagic acid. Gallic acid and corilagin are natural polyphenolic compounds which belong to hydrolysable tannin, whereas ellagic acid belongs to a flavonoid group 19,20 . Among different parts of dried D. longan, pericarp extracts contained the significantly highest total phenolic content (p < 0.05) and the highest total flavonoids as shown in Fig. 3. The results agreed well with a previous study, which reported that polyphenolic compounds are abundant in pericarp and seed of D. longan compared to the D. longan aril 1 . The total phenolic content of pericarp, seed, and aril extracts from dried D. longan, which were 967.6 ± 31.5, 739.3 ± 62.3, and 229.5 ± 2.6 µg GAE per g extracts, were found to be in agreement with a previous study, which reported that the total phenolic content of D. longan was in the range of 22.09-132.47 mg gallic acid equivalent (GAE/100 g), which was equivalent to 220.9-1324.7 µg GAE per g extract 1 . Interestingly, the dramatic increase in total phenolic content was observed in black D. longan seed extract. The ethanolic extract from black D. longan seed contained as much as 1827.1 ± 73.1 µg GAE per g extracts, which was much higher than previously reported 1 . On the other hand, there was no significant difference between the total phenolic content of dried and black D. longan extract from pericarp and aril (p > 0.05).
In addition to the total phenolic content, black D. longan seed extract also contained the significantly highest flavonoid contents (p < 0.05). Among various dried D. longan extracts, the pericarp contained the significantly highest flavonoid content of 2.8 ± 2.4 µg QE per g extract as shown in Fig. 3 (p < 0.05). The results were in accordance with a previous study, which reported that the quercetin content of D. longan pericarp was 3.12 ± 0.76 mg/ kg, which was equivalent to 3.12 ± 0.76 µg/g extract 21 . D. longan pericarp has been reported to contain slightly higher content of flavonoids than D. longan seed and aril 22 . Obviously, the flavonoid content of black D. longan seed extract, which was as high as 13.6 ± 2.5 µg QE per g extract, was dramatically enhanced and was about four times higher than previously reported.
Although the thermal ageing process of D. longan did not affect the phenolic and flavonoid contents of D. longan pericarp and aril, the total phenolic and flavonoid contents of D. longan seed were obviously enhanced after the production process of black D. longan. The likely explanation might be the formation of biological compounds, which were not originally present in the D. longan seed, during the thermal ageing process such as resistant starch 23,24 . During the temperature/time-controlled incubation, the starch inclusion complexes were generated by interaction with other components in the seeds 24 .
As shown in Fig. 4, gallic acid (1), corilagin (2), and ellagic acid (3) were detected in D. longan extract. The peak of gallic acid, corilagin, and ellagic acid in the HPLC chromatograms were detected at around 3.7, 9.9, and 19.2 min, respectively. The results agreed with the previous studies, which identified these compounds as major polyphenolic components of D. longan pericarp and seed 25-27 . In the present study, the contents of these polyphenolic components were investigated in the context of a comparison between black D. longan extract and dried D. longan extracts. The amounts of each compound in D. longan extracts are shown in Fig. 5. The findings were in line with the total phenolic and total flavonoid contents  www.nature.com/scientificreports/ since black D. longan seed extract contained the significantly highest quantities of polyphenolic compounds and flavonoid contents (p < 0.05). Among various parts of dried D. longan fruit, seeds contained the significantly highest content of gallic acid, corilagin, and ellagic acid (p < 0.05) with 5.3 ± 0.0, 8.9 ± 0.1, and 1.9 ± 0.2 mg/g extract, respectively. Interestingly, the these phenolic and flavonoid contents were significantly enhanced after the production process of black D. longan (p < 0.05). The gallic acid, corilagin, and ellagic acid content of black D. longan seed extract were as high as 53.6 ± 0.9, 19.8 ± 2.9, and 24.5 ± 0.7 mg/g extract, respectively. After being exposed to a heating and ageing procedure, the amount of gallic acid and ellagic acid of D. longan were increased by around 10-fold, while the quantity of corilagin was doubled. The reason might be the liberation of free polyphenolic compounds and flavonoids from the bound forms (i.e., esterified and glycosylate) or the decline in enzymatic oxidation involving in the antioxidant compounds in the raw fruit 28 . The results of black D. longan were in accordance with those of black garlic, as the total phenolic and total flavonoid contents of the garlic subjected to the thermal processing steps were significantly higher than those of fresh garlic 23,28 . The previous study reported that the phenolic content was increased by about 4-10-fold in the black garlic cloves compared with the fresh garlic 23 .
Apart from the findings showing differing content of the biologically active component in various parts of D. longan fruit, different methods used in the drying process also affected their bioactive compounds. The thermal ageing process was hence proposed for the enhancement of bioactive compounds in D. longan. Antioxidant activities of dried D. longan and black D. longan extracts. The antioxidant activities of dried and black D. longan extracts were investigated by two assays with different mechanisms of action. The ABTS assays measure the electron transfer reaction and represent the radical scavenging activity of the tested samples, while the FRAP assay is concerned with the ion reduction process, which represents the ability of the tested compound to convert ferric ions (Fe 3+ ) to ferrous ions (Fe 2+ ) 29 . The ferric reducing antioxidant power (EC 1 ) and Trolox equivalent antioxidant capacity (TEAC) values of dried and black D. longan extracts are shown in Fig. 6.
The TEAC values of black D. longan extracts were not significantly different from those for the dried D. longan extracts, except in the aril. The dried D. longan aril extract had no antioxidant activity, whereas the black D. longan aril extract possessed some antioxidant activity with a TEAC value of 4.1 ± 1.4 µg Trolox/mg extract. A probable explanation lies in the greater Maillard reaction, which occurs in the aril as compared with the others. As D. longan aril is composed of glucose, fructose, and various types of amino acids, such as γ-aminobutyric acid, it tends to undergo Maillard reactions, which are the chemical reactions between an amino acid and a reducing www.nature.com/scientificreports/ sugar that occur in the presence of heat 30 . These non-enzymatic browning reactions gave black D. longan a darker color and resulted in the formation of some antioxidant compounds 28 .
On the other hand, black D. longan pericarp and seed extracts possessed the same radical scavenging activity as those from dried D. longan. A likely explanation might be the degradation of some oxidative compounds during the heating process, although some free polyphenolic compounds and flavonoids were liberated from the bound forms 23 . Interestingly, the TEAC values of pericarp and seed extracts from both dried and black D. longan were comparable to ascorbic acid, gallic acid, and corilagin (p > 0.05). Ellagic acid was remarked as the most potent radical scavenger (TEAC = 23.4 ± 0.3 µg Trolox/mg), followed by ascorbic acid (TEAC = 12.3 ± 0.0 µg Trolox/mg), gallic acid (TEAC = 12.8 ± 0.2 µg Trolox/mg), and corilagin (TEAC = 12.7 ± 0.1 µg Trolox/mg). Thereby, ellagic acid was found to be the main compound responsible for the free radical scavenging activity of D. longan extracts together with gallic acid and corilagin 31 . Although the previous study reported that among various polyphenolic compounds, tannins demonstrated the strongest ABTS·+ radical scavenging activity 32 , in the present study it was observed that ellagic acid, which belongs to a flavonoid group, was more potent than gallic acid and corilagin, which belongs to hydrolysable tannin 1 . Furthermore, D. longan extracts from both pericarp and seed could therefore be considered as natural extracts with potent radical scavenging activity.
Aside from radical scavenging activity, D. longan extracts also possessed a reduction ability as shown in Fig. 6. The reduction ability of D. longan extracts was in accordance with their phenolic and flavonoid contents. Gallic acid possessed the significantly highest EC 1 value of 237.0 ± 1.6 mM FeSO 4 /mg, which was comparable to that of ascorbic acid (238.3 ± 0.2 mM FeSO 4 /mg), followed by corilagin (226.2 ± 2.9 mM FeSO 4 /mg) and ellagic acid (192.3 ± 0.7 mM FeSO 4 /mg). However, both phenolic compounds and flavonoids were responsible for their reduction capacity 33 . The black D. longan seed extract, which contained the highest total phenolic, total flavonoid, gallic acid, corilagin, and ellagic acid contents, thus possessed the significantly highest reduction ability with an EC 1 value of 150.0 ± 1.0 mM FeSO 4 /mg extract (p < 0.05). Consequently, the black D. longan seed extract was suggested as the most potent antioxidant extract with the strongest free radical scavenging and reduction ability. Because of its potent antioxidant effect, black D. longan seed has been proposed as a natural antioxidant source for use in food and cosmetic products. Since the portions of pericarp and seed account for 30% of the whole fruit dry weight 34 , the utilization of these by-products would not only reduce the agricultural waste product but also increase its value. activities against the secretion of IL-6 and TNF-α, which are key players involved in the age-related inflammatory process 35 , of dried and black D. longan extracts were investigated. RAW 264.7 macrophage cells were used in the present study since they can secret these cytokines after the stimulation of LPS. The RAW 264.7 macrophage cell viability after treatment with dried and black D. longan extracts is shown in Table 1. No cytotoxicity was detected in any of the D. longan extracts since the cell viability was more than 100%. Dexamethasone, corilagin, gallic acid, and ellagic acid were also found to be safe for RAW 264.7 macrophage cells. The IL-6 and TNF-α inhibitory activities of dried and black D. longan extracts are shown in Fig. 7. TNF-α is known as an indicator of chronic inflammatory processes related to ageing, whereas IL-6 has been associated with poor physical performance and muscle weakness by geriatricians and could predict the onset of disability 36,37 . Among various parts of D. longan fruit, aril of both dried and black D. longan was predominant in IL-6 and TNF-α inhibition. Gallic acid was suggested to be the main compound responsible for both IL-6 and TNF-α inhibitory activities. In contrast, corilagin was responsible only for TNF-α inhibition. Although D. longan extracts and their major chemical components exhibited only low to moderate anti-inflammatory activities compared to dexamethasone, a corticosteroid used in the treatment of inflammation, they could be consumed as natural anti-inflammatory supplements with no steroidal side effects.
Anti-hyaluronidase activities of dried D. longan and black D. longan extracts. Hyaluronidase, a homologous enzyme that hydrolyzes or depolymerizes hyaluronan, plays an important role in the modulating activity of many pathological processes 38 . Hyaluronan plays a pivotal role in the maintenance of the elastoviscosity of liquid connective tissues and controls the water transportation related to the tissue hydration 39 . The degradation of hyaluronan resulting in the production of breakdown products, including lower molecular mass polymers. These breakdown products of hyaluronan exhibited distinct biological properties from the larger www.nature.com/scientificreports/ precursor molecules 40 . The hyaluronan depolymerization occurs in tissue injury and initiates the inflammatory response 38 . Additionally, hyaluronan is known as a lubricant and shock-absorber in joints and connective tissues 41 . Its degradation hence leads to the deterioration of the viscoelastic properties of the synovial fluid 42 . The inhibitory activities against hyaluronidase of dried and black D. longan extracts are shown in Fig. 8. Although D. longan extracts exhibited low anti-hyaluronidase activity, the inhibitory effect of black D. longan seed was significantly enhanced compared to the dried D. longan seed extract. Since the anti-hyaluronidase activity  www.nature.com/scientificreports/ of black D. longan seed extract (18.4 ± 2.0%) was the most significantly potent (p < 0.05), black D. longan seed extract could be suggested to have anti-hyaluronidase activity in addition to its antioxidant activities.
In conclusion, black D. longan was successfully developed after undergoing a heating and ageing procedure at a controlled temperature of 70 °C and a relative humidity of 75%. A novel black D. longan contained a larger quantity of biologically active compounds and possessed more potent biological activities than a conventional dried D. longan. The ethanolic extract from the seed of black D. longan contained the most significantly abundant of biologically active compounds, including total phenolic, total flavonoid, gallic acid, corilagin, and ellagic acid content (p < 0.05). Furthermore, it possessed the most significantly potent antioxidant and anti-hyaluronidase activities (p < 0.05). Since oxidative stress is known to be related to ageing and skin wrinkles, black D. longan  www.nature.com/scientificreports/ seed extract with a potent antioxidant activity (p < 0.05) was suggested for further topical use as a cosmeceutical ingredient for anti-skin ageing. On the other hand, degradation of hyaluronan in the skin resulted in the loss of the skin's natural moisturizing factor, while the loss of hyaluronan from the synovial fluid in joint resulted in joint pain and several conditions. Therefore, black D. longan seed extract, which significantly inhibited hyaluronidase activity (p < 0.05), is suggested for both topical use for anti-skin ageing and joint pain relief. On the other hand, the aril of D. longan, which possessed the significantly highest anti-inflammatory activities, is suggested as a natural edible anti-inflammatory agent.

Material and methods
Chemical material. l-Ascorbic acid, aluminum chloride (AlCl 3 ), 2,2′-azino-bis3-ethylbenzothiazoline-  www.nature.com/scientificreports/ where X is GAE or µg of gallic acid per g of the D. longan extracts and Y is an absorbance of each sample tested with the Folin-Ciocalteu assay. The experiments were performed in triplicate.
Total flavonoid content determination. Total flavonoid content of each D. longan extract was investigated using the aluminum chloride method, which has been previously described, with some modifications 45 . Firstly, 20 μL of the ethanolic solution of D. longan extracts (1 mg/mL) was mixed with 80 μL of AlCl 3 aqueous solution (0.1 g/ mL, 10% w/v) and 100 μL of CH 3 COOK aqueous solution (98.15 g/l, 1 M). After the resulting mixtures were incubated for 30 min in the dark, they were measured for an absorbance at 415 nm using a multimode detector (SPECTROstar Nano, BMG Labtech, Offenburg, Germany). Quercetin was applied as a standard compound to construct a calibration curve. Finally, the results were presented as quercetin equivalent (QE) values, which represented a µg of quercetin per g of the D. longan extracts. QE was calculated following Eq. (2); where X is QE or µg of quercetin per g of the D. longan extracts and Y is an absorbance of each sample tested in the aluminum chloride assay. The experiments were performed in triplicate.
Determination of gallic acid, corilagin, and ellagic acid content by high performance liquid chromatography (HPLC). The quantitative analysis of gallic acid, corilagin, and ellagic acid was performed using an HP 1100 chromatographic system (Hewlett-Packard, Waldbronn, Germany). A gradient mobile phase system composed of two phases was used, including phase A (0.05% formic acid in acetonitrile) and phase B (0.05% formic acid aqueous solution). The program was set for gradient elution of 10% where X2 is the concentration of corilagin, A is the AUC of the corilagin peak detected around 10 min, and C is the concentration of the respective sample solution.
where X3 is the concentration of ellagic acid, A is the AUC of the ellagic acid peak detected around 20 min, and C is the concentration of the respective sample solution. where X is TEAC value and Y is an absorbance of each sample tested in ABTS assay. l-Ascorbic acid was used as a positive control. The experiments were performed in triplicate.

Antioxidant activity determination of dried
Ferric reduction/antioxidant power (FRAP) assay. The reduction capacity of D. longan extracts, gallic acid, corilagin, and ellagic acid were investigated by means of a ferric ion reduction assay 44  www.nature.com/scientificreports/ ing the ferric-TPTZ reduction capacity, which is equivalent to 1 mg of the D. longan extract. The EC 1 values were calculated following Eq. (7); where X is EC 1 value and Y is an absorbance of each sample tested in the FRAP assay. l-Ascorbic acid was used as a positive control. The experiments were performed in triplicate.

Anti-inflammatory activity determination of dried D. longan and black D. longan extracts.
Murine monocytemacrophage (RAW 264.7) cells (American Type Culture Collection, ATCCTIB-71) treated with LPS were used to investigate the effect of D. longan extracts and their chemical compositions on the pro-inflammatory cytokine secretion (IL-6 and TNF-α ). Cells were cultured according to a method previously described with some modifications 47,48 . Briefly, RAW 264.7 cells with a density of 2 × 10 6 cells per well in 750 μL of DMEM supplemented with GlutaMAX™-I, inactivated FBS (10%), penicillin (100 U/mL), streptomycin (100 μg/mL), and amphotericin B (0.25 μg/mL) were seeded in wells of 24 well-plates and incubated in a CO 2 incubator (37 °C, 5% CO 2 in air, 90% humidity) for 24 h. Thereafter, 1 µL of the D. longan extracts or dexamethasone (100 µg/mL) was added and further incubated in a CO 2 incubator (37 °C, 5% CO 2 in air, 90% humidity). Three replicates per sample were performed. After 24 h of the extract treatment, 250 μL of LPS in DMEM (4 µg/mL) was treated in each well and incubated in a CO 2 incubator (37 °C, 5% CO 2 in air, 90% humidity) for another 24. After that, the treated cells along with the supernatant were divided into two parts. The medium (500 μL) was collected for cytokine dosages, while the attached cells were subjected to the viability assay using MTT dye. The collected medium was then centrifuged for 10 min at 13,500×g, and the supernatant was investigated for cytokine secretion using an enzyme-linked immunosorbent assay (ELISA) following the manufacturer's protocol (R&D Systems). The remain medium, which was left over in the wells, was investigated for cell viability using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. To reduce variation due to cell density differences, secretion of IL-6 and TNF-α from RAW 264.7 cells was normalized to MTT levels 48 . RAW 264.7 cells without lipopolysaccharide treatment served as a negative control, while 100% of cytokine secretion was from the positive control of RAW 264.7 cells treated with LPS. The inhibitory activities of each sample were calculated following Eq. (8); where A is the cytokine secretion. Dexamethasone served as a positive control for both IL-6 and TNF-α secretory inhibition. The experiments were performed in triplicate.
Anti-hyaluronidase activity determination of dried D. longan and black D. longan extracts. The hyaluronidase inhibitory activity of D. longan extracts, gallic acid, corilagin, and ellagic acid was evaluated by measuring a product from the cleavage of sodium hyaluronate by hyaluronidase 49 . Prior to the experiment, the enzyme activity of hyaluronidase was determined. More than 90% enzyme activity was used in the anti-hyaluronidase activity determination. Firstly, 20 μL of the ethanolic solution of D. longan extracts (1 mg/mL) was mixed with hyaluronidase (15 unit/mL). After incubation at 37 °C for 10 min, hyaluronic acid solution (0.03 % w/v) in phosphate buffer pH 5.35 was added and further incubated for 45 min. Immediately after the addition of acid bovine serum albumin, the extracts were measured for an absorbance at 600 nm using a multimode detector (SPECTROstar Nano, BMG Labtech, Offenburg, Germany). The hyaluronidase inhibitory activity was calculated according to Eq. (9); where X is the absorbance of the mixtures with sample; Y is the absorbance of the mixtures without sample. Oleanolic acid was used as a positive control. The experiments were performed in triplicate.

Statistical analysis.
All the values are given as means ± standard deviations and were analyzed. The statistical analysis used was one-way analysis of variance (ANOVA) followed by Tukey's post-hoc tests using the SPSS software (SPSS Statistics 21.0, IBM Corporations, New York, NY, USA). A value of p < 0.05 was accepted as significant. www.nature.com/scientificreports/