The following standards were used to determine sugars, organic acids, soluble polyphenol content, carotenoids and antioxidant capacity: sucrose, fructose and glucose; citric acid, malic acid, succinic acid and fumaric acid; gallic acid; α/β-carotene, lutein, β-cryptoxanthin, violoxanthin, zeaxanthin; trolox solution from Sigma-Aldrich Chemie (Steinheim, Germany).

Persimmon fruits from seven astringent cultivars (cv. Rojo Brillante; cv. Tone Wase; cv. Giboshi; cv. Kaki Tipo; cv. Aizumishirazu-A; cv. Giombo; cv. Hachiya) and three non-astringent cultivars (cv. O’gosho; cv. Hana Fuyu; cv. Jiro) were harvested from October to November in the persimmon germplasm collection hosted by the Instituto Valenciano de Investigaciones Agrarias (IVIA, Valencia, Spain). Harvesting was carried out when fruits reached a homogeneous full orange or reddish-orange colour depending on the cultivar. In all the cultivars, fruits presented a crisp texture with firmness values over 30 N.

One homogeneous lot of 20 persimmon fruits from each cultivar was evaluated immediately after harvest. A second lot of 20 fruits from astringent cultivars was subjected to deastringency treatment under standard conditions (95% CO₂ at 20˚C and 90% RH for 24 h). After deastringency treatment, fruits were kept at room temperature in an air atmosphere for 24 h to evaluate the changes associated with treatment application.

At harvest and after deastringency treatment (with astringent cultivars), besides measuring external colour and firmness, the following parameters were evaluated: soluble polyphenol content, total antioxidant capacity, total soluble solids, total and individual sugars, organic acids and carotenoids compounds. The astringency level was sensory-evaluated by a semi-trained panel. With astringent cultivars, acetaldehyde concentration was analysed at harvest and after deastringency treatment.

External skin colour and flesh firmness were determined on 20 fruits. External colour was evaluated with a Minolta colorimeter (Model CR-300 Ramsey, NY, USA); Hunter parameters (“L”, “a”, “b”) were measured and the results were expressed as a colour index: CI = 1000a/Lb [11] . Fruit firmness was determined by a Texturometer Instron Universal Machine model 4301 (Instron Corp., Canton, MA, USA) using an 8-mm plunger and breaking the flesh in each fruit on 180˚ sides after removing peel.

Acetaldehyde concentration was analysed by headspace gas chromatography, as described by Salvador et al. [11] and the results were expressed as mg 100 mL⁻¹.

The sensory evaluation of astringency was performed on composite samples of three (peeled and sliced) fruits. A semi-trained panel of 8 - 10 people, who were familiar with persimmon fruits, was asked to evaluate astringency. A 4-point scale was used, where 0 was “no astringency” and 3 was “very high astringency”. Samples were presented to panel members on trays labelled with random 3-digit codes and served at room temperature (25˚C ± 1˚C). Judges had to taste several segments of each sample to compensate, as far as possible, the biological variation of the material. Milk was provided for palate rinsing between samples.

The soluble polyphenol content of each cultivar was measured as gallic acid equivalents (GAE) using the Folin-Ciocalteau’s phenol reagent (FC reagent) according to Taira [12] . The results were expressed as milligrams of GAE per 100 g of dried weight (dw).

Total antioxidant capacity was determined as the antiradical activity of methanolic extracts, where 20 mg of sample were homogenised with 2 mL of methanol (80% v/v). It was spectrophotometrically tested by measuring absorbance at 515 nm of free radical DPPH, adapted from Novillo et al. [13] . The obtained values were compared to the concentration-response curve of the standard Trolox solution expressed as micromoles of Trolox equivalents (TE) per 100 mg of dw.

The extraction and determination of sugars and organic acids were conducted from 50 mg of sample and extracted with 1 mL of bi-distilled water. The extracted sample was centrifuged twice at 14,000 rpm for 15 min at 4˚C. The supernatant was filtered through 0.45 µm filters and purified in a Sep-Pak C18 column. An analysis of sugars was performed in a Thermo Separation Products HPLC. Separation of sugars was done isocratically with Millipore water as a mobile phase, at a flow of 0.5 mL・min^{− 1 in a fast carbohydrate column (Aminex HPX87-C column), 100} × 7.8 mm i.d. (Bio Rad Laboratories, Hercules, CA, USA), which was preceded by a micro-guard cartridge, Carbo-C 30 × 4.6 mm i.d. (Bio Rad Laboratories, Hercules, CA, USA) maintained at 75˚C and attached to a refractive index detector (RID). Separation of the organic acids analysis was carried out in an Alliance liquid chromatographic system (Waters, Barcelona, Spain) equipped with a 2695 separation module, coupled to a 2996 photodiode array detector and a ZQ2000 mass detector. A thermostat column oven, an ICSep ICE-COREGEL 87H3 column (Transgenomic), an ICSep ICE-COREGEL 87H guard kit and an automatic injector were used for chromatographic separation. The Empower 2 software was used for data acquisition. Sample temperature was 5˚C and column temperature was 35˚C. Capillary voltage was 3.0 kV, cone voltage was 23 V, source temperature was 100˚C, desolvation temperature was 200˚C and desolvation gas flow was 400 L/Hr. Full data acquisition was performed by scanning 100 to 400 uma in the centroid mode. The solvent system was an isocratic mobile phase of 0.1% H₂SO₄ solution. The total run time was 20 min at 0.6 mL・min⁻¹. The injection volume was 5 µL. The results were expressed as µg・mg⁻¹・dw. Sucrose, glucose and fructose sugars were identified by comparing their retention time with a standard and they were quantified using an external calibration curve. Organic acid compounds were identified on the basis of comparing their retention times, UV-Vis spectra and mass spectrum data with the corresponding authentic standards. Concentrations were determined using an external calibration curve with citric acid (rT = 7.7 min; [M-H] + 191 m/z), malic acid (rT = 9.1 min; [M-H] + 133 m/z) and succinic acid (rT = 11.2 min; [M-H] + 117 m/z).

The total carotenoid content determination was conducted according to Koca et al. [14] , where 50 mg of dried sample were homogenised with 2 mL of hexane:acetone mixture (7:3). Carotenoid content was determined spectrophotometrically at 450 nm from the standard curve and expressed as micrograms of β-carotene equivalents (β-CE) per 100 mg of dw.

Individual carotenoids were determined from 50 mg of ground tissue and extracted in 1 mL cold ethanol and 0.8 mL hexane. The extraction mixture was centrifuged at 10,000 rpm for 5 min at 4˚C. The upper organic phase was transferred to a new microcentrifuge tube. In order to avoid losses, saturated sodium chloride (0.5 mL) and hexane (0.8 mL) were added twice to the pellet and the mixture was recentrifuged. The organic phase was free of visible particulates and was, therefore, not filtered. The organic phase was evaporated under a nitrogen stream. For the saponification of persimmon carotenoids, residues were dissolved in 1 mL of methanolic potassium hydroxide ( 2 M ) and 0.1% butylated hydroxytoluene was used to avoid oxidation. The mixture was left 3 h in the dark. It was transferred to a separatory funnel. Then 4 mL of 10% sodium chloride, plus 2 mL of hexane, were poured over to remove potassium hydroxide. The hexane layer was also washed with bi-distilled water until the rinse had a neutral pH. Once again 2 mL of hexane were added to the potassium hydroxide to collect any remaining carotenoids. The combined hexane extracts were evaporated under a nitrogen stream. Samples were dissolved in 300 µL of ethyl acetate and 10 µL of were injected in reverse phase HPLC (Agilent 1200 series HPLC system (Santa Clara, CA)), including a model G1311A quaternary pump, a model G1367B autosampler, a model G1316A column oven and a model G13150 photodiode array detector. The column was an Agilent ZORBAX Eclipse XDB-C18, 5 µm bead size, 4.6 mm × 150 mm , connected with an Eclipse XDB-C18 guard column. Column temperature was controlled at 30˚C during HPLC runs. Data were processed by the Agilent ChemStation software. Separation of carotenoids pigments was achieved by modifying a previously established gradient elution programme [15] . The flow rate was 1 mL・min⁻¹. The mobile phases were acetonitrile:H₂O:trie- thylamine (900:99:1, v/v/v) (A) and ethyl acetate (B). The gradient elution programme was: 0 - 5 min, 100% - 75% A; 5 - 10 min, 75% - 30% A; 10 - 13 min, 30% - 0% A; 13 - 14 min, 0% - 100%; 14 - 15 min, 100% A. Data were collected at 440 nm, 477 nm and 296 nm. Several carotenoids were identified based on retention time and spectrum compared to commercially available authentic standards. Their quantity was extrapolated from standard curves of authentic compounds and was corrected for extraction efficiency based on the β-apo-8’-ca- rotenal internal recovery standard. Concentrations were expressed as µg・kg⁻¹・dw.

The vitamin A value in fruits was calculated as retinol activity equivalents (RAE) using the following conversion according to [16] : RAE = [(β-cryptoxanthin/12) + (α-carotene/12) + (β-carotene/6)].

Data were subjected to an analysis of variance, and the multiple comparisons between means were determined by the least significant difference (LSD) test (P ≤ 0.05) using the Statgraphics Plus 5.1 software application (Manugistics Inc., Rockville, MD, USA).

External colour is the most widely used parameter as a non-destructive index for harvesting persimmon. In the present research, the fruits from all the cultivars were harvested in the commercial maturity stage, with an external colour index that ranged from 15 to 20, and corresponded to homogeneous skin coloration from full orange to reddish-orange, depending on the cultivar (data not shown). Firmness at harvest was above 30 N for all the studied cultivars, values which allow persimmons with a crisp texture to be marketed. The application of deastringency treatment with CO₂ to cultivars that were astringent at harvest neither modified fruit firmness nor skin colour. However, the treatment brought about a sharp rise in acetaldehyde content from values below 0.5 mg 100 mL⁻¹ at harvest to values between 1 and 3 mg 100 mL⁻¹ (data not shown). This increase was due to exposing fruits to a high-CO₂ atmosphere since acetaldehyde accumulates under partially or totally anaerobic conditions [17] .

The sensory evaluation corroborated absence of astringency at harvest in the fruits of non-astringent cultivars (data not shown). It also revealed that deastringency treatment was effective for removing astringency in all the astringent cultivars, except “Hachiya”, in which panellists detected slight astringency after applying treatment.

In persimmon, the soluble polyphenols that belong to the proanthocyanidin class are responsible for typical astringency since they interact with salivary proteins and give rise to the astringency sensation [18] . It is known that persimmons from both astringent and non-astringent cultivars are very astringent, with high soluble polyphenol content when fruit are small and immature. Nevertheless, while the fruits from astringent cultivars remain strongly astringent even when fruits are fully coloured, those from non-astringent cultivars lose astringency while growing on trees. It has been claimed that tannin cell development is continuous to late fruit growth stages in astringent cultivars, while tannin cells development stops in early fruit growth stages in non-astringent ones. Thus natural loss of astringency on trees is thought to be due to the dilution of tannins concentration on flesh as fruits grow [1] .

In the present study, the fruits of the cultivars that were astringent at harvest showed high soluble polyphenol content, which ranged from 650 mg 100 g^{− 1 in} “Rojo Brillante” to 1220 mg 100 g^{− 1 in “Giombo” (}Figure 1(a)).

Major differences in the soluble polyphenol contents among cultivars have been previously reported by Veberic et al. [8] . In contrast, the soluble polyphenol content of non-astringent cultivars was much lower, and they all showed values below 185 mg 100 g⁻¹.

As expected, when persimmons from astringent cultivars were subjected to deastringency treatment with a high CO₂ concentration, soluble polyphenols drastically dropped, and similar values to those of non-astringent cultivars were given. The decline in soluble polyphenols in astringent cultivars ran in parallel to the sharp rise in acetaldehyde content observed after deastringency treatment. It is known that the insolubilisation of tannins, and consequently astringency removal, are mediated by the acetaldehyde generated under anaerobic conditions [13] [19] .

After deastringency treatment, the concentration of soluble polyphenols vastly varied among cultivars; the highest values, 200 mg 100 g⁻¹, were observed in cultivar “Hachiya” which agrees with the sensory astringency detected by the panellists. Currently, there is no established threshold of soluble tannins that ensures lack of astringency in persimmon fruits. In fact, perception of astringency has been shown to be influenced by presence of other compounds, like acids and sugars [20] [21] and it is probably affected by the specific tannin composition of each cultivar [22] . However, it must be noted that after applying deastringency treatment to astringent cultivars, the concentration of soluble tannins lowered to similar values to those of naturally non-astringent cultivars.

The higher antioxidant capacity exhibited by astringency cultivars at harvest compared to non-astringent ones (Figure 1(b)) was related directly to content of polyphenols, which possess strong radical scavenging activity. It has been reported that polyphenols are the major antioxidant compounds in persimmon pulp [23] . Non-astrin- gent cultivars gave values below 6 µmol TE 100 mg⁻¹. However, the content among astringent cultivars varied vastly and presented values that ranged from 25 to 60 µmol TE 100 mg⁻¹ (Figure 1(b)). Similar results were obtained by Katsube et al. [24] , who measured antioxidant activity by the DPPH radical method, and obtained more than one order magnitude higher in astringent persimmons than in non-astringent ones.

As observed with soluble polyphenol content, total antioxidant capacity also significantly lowered after subjecting astringent cultivars to deastringency treatment. Thus after CO₂ treatment, most cultivars showed similar total antioxidant capacity values to those observed by non-astringent cultivars. The highest values, 10 µmol TE 100 mg⁻¹, were detected in cultivar “Hachiya” while the lowest values, below 2 µmol TE 100 mg⁻¹, were given by “Rojo Brillante”, “Tone Wase” and “Kaki Tipo” (Figure 1(b)). These results agree with those of Del Bubba et al. [9] , who reported that the deastringency treatment with CO₂ applied to cultivars “Kaki Tipo” and “Rojo Brillante” led to a 10-fold drop in antioxidant capacity.

Table 1 shows the individual and total sugars in the 10 studied persimmon cultivars. The main sugars found in the flesh of all persimmon cultivars were sucrose, glucose and fructose, which agrees with previously reported studies [8] [9] [25] .

After taking into account the individual and total sugars of astringent cultivars at harvest, “Tone wase” and “kaki Tipo” presented the highest content of glucose (509.9 - 394.4 µg・mg⁻¹) and fructose (412.1 - 335.7 µg・mg⁻¹) and total sugars (1086 - 996.7 µg・mg⁻¹). The highest sucrose content value went to “Giombo” (401.8 µg・mg⁻¹), “Rojo Brillante” (396.9 µg・mg⁻¹) and “Kaki Tipo” (356.5 µg・mg⁻¹), while “Tone Wase” obtained the lowest value (74.7 µg・mg⁻¹). Cultivars “Aizumishirazu-A”, “Hachiya” and “Giboshi” showed the lowest sugar content values, similarly to that observed in non-astringent cultivars. The total sugar content in non-astringent cultivars was lower, and ranged from 433.6 to 609.4 µg・mg⁻¹. Cultivar “Hana Fuyu” presented a higher content of reducing sugars (glucose and fructose) than sucrose, while sucrose was found at the highest concentration, followed by glucose and fructose in “Jiro” and “O’gosho”.

The effect of deastringency treatment on total sugar content strongly depends on the cultivar. While the total sugar content of “Aizumishirazu-A”, “Hachiya” and “Giboshi” was not affected by deastringency treatment, in “Tone Wase”, “Rojo Brillante”, “Kaki Tipo” and “Giombo”, the cultivars that presented the highest sugars level at harvest, deastringency treatment caused an important drop in the total sugar content to about half the original value. So when the fruits of all the astringent cultivars were ready to be marketed (after treatment deastringency) exhibited similar total sugars content to that of non-astringent cultivars.

Reduced total sugars after deastringency treatment was generally due to the sharp drop in sucrose in most cultivars, which even became undetectable in cultivars such as “Hachiya”, “Tone Wase”, “Kaki Tipo” and “Giboshi”. It is noteworthy that cultivar “Aizumishirazu-A” was the only cultivar that showed no differences in sucrose

^aValues are the mean of three independent determinations. For each sugar, values with different letters indicate significant differences (P < 0.05). nd: not detected.

content between values at harvest and those recorded after applying CO₂ treatment. Glucose and fructose were much less affected by deastringency treatment. Glucose and fructose contents remained almost constant in most cultivars, except cultivars “Rojo Brillante” and “Tone Wase”. Glucose in “Rojo Brillante” significantly decreased after applying treatment (from values of 317.7 µg・mg⁻¹ at harvest to values of 158.8 µg・mg⁻¹ after treatment), while “Tone Wase” showed a more reduced decline for glucose (from values of 509.9 µg・mg⁻¹ to 323.5 µg・mg⁻¹) and fructose (from 412.1 µg・mg⁻¹ to 301.5 µg・mg⁻¹). The results obtained in the present work agree with Ittah [25] , who reported that deastringency treatment in CO₂ atmospheres reduces the total sugars content and diminishes sucrose starts when CO₂ treatment commences, while changes in glucose and fructose content are very gradual. This reduction in total sugars due with deastringency treatment can be explained by a possible glycoside formation between soluble tannins and soluble sugars.

The reduced sucrose content in most cultivars after deastringency treatment could be due to induced invertase activity through the tannin insolubilisation process during deastringency. In previous studies, the decline of sucrose that occurred in some astringent persimmon fruits during fruit growth and ripening have been attributed to an increase of invertase activity coupled with a decrease of soluble tannins [9] as an in vitro study has demonstrated the strong inhibition of this enzyme by gallic and tannin acids [26] . It has also been reported that fruit enzymatic systems, such as sucrose synthetase, invertase, fructose isomerase, cellulose and enzymes from the gluconeogenesis pathway, are involved with fruits under anaerobic conditions [27] [28] .

Citric acid, malic acid and succinic acid were the main organic acids found in the present study, while fumaric acid was found in trace amounts (Table 2). At harvest, the total organic acids content ranged from 15.8 µg・mg⁻¹ to 30.9 µg・mg⁻¹. Citric acid content was over 5.3 µg・mg⁻¹ among all the cultivars, except non-astringent cultivars “Hana Fuyu” and “O’gosho”, which presented lower values. Malic acid content was over 10 µg・mg^{− 1 in astringent cultivars such as “Rojo brillante”, “Tone wase”, “Kaki tipo”, and also in non-astringent cultivars “Jiro” and “O’gosho”. The lowest value of this acid was shown for “Giombo” (4.3 µg・mg−1}). It is noteworthy that succinic acid content was higher in astringent cultivars, and “Giombo” is highlighted because it showed the highest values (19.3 µg・mg⁻¹). Non-astringent cultivars gave values below 5 µg・mg⁻¹ (Table 2).

Although the effect of deastringency treatment on organic acids has not been previously addressed, a reduction

^aValues are the mean of three independent determinations. For each organic acid, values with different letters indicate significant differences (P < 0.05). nd: not detected.

in their content was expected since it is known that high CO₂ atmospheres induce changes in organic acid metabolism in other fruits, such as strawberry or cherimoya [29] [30] . Nevertheless in the present study, deastringency treatment did not affect the total organics acid content, and slightly modified individual organic acids.

The organic acids content found in the present study was higher than those reported by Veberic et al. [8] , which is not surprising given the maturity stage of the fruit evaluated in each case. Fruits were analysed herein in the “ready-to-eat” maturity stage, while the evaluations of Veberic et al. [8] were made with fully ripened fruit. It has been reported that during fruit ripening, metabolic processes led to a considerable loss in organic acid in persimmon fruits [31] .

Figure 2 depicts the total carotenoid content at harvest and after submitting astringent cultivars to the CO₂ treatment. At harvest the total content greatly varied among the ten cultivars; cultivar “Hachiya” had the highest content (above 100 µg β-CE 100 g⁻¹), while the lowest values went to cultivar “Rojo Brillante” and “O’gosho”, with 20 and 23.5 µg β-CE 100 g⁻¹, respectively. All the other cultivars showed values that ranged from 40 to 80 µg β-CE 100 g⁻¹.

The individual carotenoids detected by HPLC in persimmon flesh were lutein, violoxanthin, zeaxanthin, β-cryptoxanthin and β-carotene (Table 3). β-cryptoxanthin was the predominant carotenoid found at harvest in all the studied cultivars. This xanthophyll has also been reported in previous studies as the main carotenoid in persimmon flesh [8] [10] [32] - [34] . The relative levels among zeaxanthin, lutein and β-carotene were cultivar-dependent. Violoxanthin was the lowest carotenoid found in all the cultivars.

Deastringency treatment led to an increase in the total carotenoids content compared to the values recorded at harvest in some cultivars, such as “Aizumishirazu-A”, “Giombo” and “Giboshi” (Figure 2), which coincides with the changes observed in individual carotenoids (Table 3). Thus the content of all the individual carotenoids content in cultivars “Aizumishirazu-A” and “Giombo” significantly increased after CO₂ treatment, while the content of β-cryptoxanthin and β-carotene in “Giboshi” mainly increased. Very few studies have addressed the effect of deastringency treatment on carotenoid content in persimmon. Plaza et al. [35] observed that the carotenoids content of persimmons subjected to high pressure remained unchanged when applied to fruits in a mid-season maturity stage. Yet when this treatment was assayed in fruits in an advanced maturity stage, it led to

^aValues are the mean of three independent determinations. For each organic acid, values with different letters indicate significant differences (P < 0.05). nd: not detected.

increased carotenoids content, probably due to modifications in their bioaccessibility. The retinal activity equivalent (RAE) provided information about dietary intake of provitamin A carotenoids. In the present study, cultivars “Jiro” and “Hachiya” presented the highest RAE values (102.14 and 126.99 µg・kg⁻¹, respectively), mainly due to their high β-cryptoxanthin content.