Response to organic cultivation of heirloom Capsicum peppers: Variation in the level of bioactive compounds and effect of ripening

Peppers (Capsicum spp.) are one of the most important vegetables and their double use (vegetable or spice) and two commercial stages (unripe and fully ripe) contributed to their use in many recipes and fast diffusion from America. Nowadays, Spain is a center of diversity for C. annuum, with many landraces, offering a great opportunity for adaptation to organic cultivation. Furthermore, Capsicum peppers contain many bioactive compounds, essential to provide high added-value to these cultivars, especially for organic markets, although knowledge about the effect of organic cultivation on Capsicum fruit quality is still scarce. Here, 37 accessions of Spanish landraces and foreign materials from C. annuum and other species were grown under organic and conventional conditions and evaluated for ascorbic acid (AAC), total phenolics (TP) and total red and yellow/orange carotenoids, considering both ripening stages. A large genotypic variation was found within each ripening stage and growing condition for the studied traits. Also, both stages showed high levels, although fully ripe fruits were the richest. Organic conditions enabled higher levels in fully ripe fruits of AAC and TP on average (135 vs 117 mg·100 g-1 and 232 vs 206 mg·100 g-1) and in most accessions, although the genotype×growing conditions interaction also contributed, but at lower extent, to the observed variation. Significant genotype×ripening stage and growing conditions×ripening stage interactions were also found, suggesting that the magnitude of the increase with ripening depends on the accession and growing conditions. By contrast, there were no differences between growing conditions for carotenoids and differences were mainly due to the genotype factor. Finally, the large genotypic variation and favourable organic conditions allowed identifying several materials from different types and uses with very high levels of bioactive compounds for organic cultivation, in both ripening stages but particularly at fully ripe stage (>500 mg·100 g-1).


Introduction
The increase of human population and changes in their needs are causing a considerable impact on the use of land for cultivation and drives to an unsustainable use of agricultural but also as protectors against photooxidation and providing health benefits as antioxidants, vitamin A precursors or antitumoral agents [26][27][28][29][30]. Carotenoids are mainly present in the fully ripe fruits of Capsicum, while their level are comparatively negligible (or even nil) in unripe fruits [31]. Therefore, these traits are of paramount importance for breeding peppers and, at the same time, for promoting the use of traditional varieties of C. annuum by increasing their added value in the frame of organic production and markets. Despite there are some interesting reports on the levels of bioactive compounds in C. annuum under organic practices, most of them were based on a scarce or nil varietal diversity and/or only considered one ripening stage [32][33][34]. Thus, the knowledge about the range of responses expected from the great variation available in the traditional varieties of this crop, as well as the opportunities for exploiting the genotype-by-environment interaction at each ripening stage, remains almost nil. The objective of the present work is to assess, in a comprehensive collection of C. annuum and other exotic Capsicums, the response of different genotypes to organic practices in comparison to conventional practices in terms of the levels of their main bioactive factors, i.e. ascorbic acid, phenolics, and carotenoids, considering also the two main commercial stages, unripe and fully ripe. The effect of ripening stage and growing conditions, as well as the interactions of cultivars (genotype) with these factors and the opportunities of selecting materials for specific growing conditions are also discussed. To our knowledge, this is the first comprehensive comparative study in Capsicum for adaptation to organic conditions by exploiting their variation in bioactive compounds.

Plant material
A collection of thirty-seven accessions of Capsicum was evaluated in this work, encompassing a wide diversity of varietal types and origins, including many Spanish traditional varieties or ecotypes from PDOs and PGIs and several peppers and chillies from other countries (Table 1, S1 and S2 Figs). This collection mainly included C. annuum accessions (30 accessions), but also included C. chinense (4), C. frutescens (1) and South American ají C. baccatum (2).

Growing conditions
Plants materials were grown open field in two growing conditions, organic and conventional, in the spring-summer growing season of years 2015 and 2016. To work with large numbers of accessions, individuals and fruit analyses is tedious and time-consuming and also involves several technical and logistics difficulties. For this reason, our experiment and materials were conducted in two years and a group of 16 accessions were evaluated in both years as controls of the year effect (Table 1). In both growing conditions plants were transplanted at the 4-5 true leave stage, in April 2015 and 2016 and at planting frame 1×0.5 m. Both organic and conventional experimental plots were located in the area of Sagunto (northern Valencia city, Spain). The organic plot was located in the Marxal dels Moros protected area (UTM coordinates X: 734494.88, Y: 4390434.86), while the plot managed under conventional conditions (control) was close to the organic plot to get the same conditions in climate, water, and irrigation time, and similar soil main properties (see S1 Table) (UTM coordinates X: 732900.40/ Y: 4391754.37). Plants were irrigated by surface irrigation every 8-10 days, depending on the evapotranspiration, from the irrigation ditch that bordered the plots. In this way, both growing systems differed mainly in management and historical soil conditions. Both plots were managed by technicians and farmers from the Unió de Llauradors i Ramaders (LA UNIÓ ), one of the most relevant farmers association in organic production in the Region of Valencia, following their usual organic or conventional practices. The organic management was based in a soil with rotation of crops every four years. Also, fertilization was based on organic sheep manure (4 kg/m 2 ) at the beginning of the season. External treatments for pest control were not necessary as natural predators maintained the microfauna balance. Adventitious plants were controlled by hand tools monthly and three thermal applications (i.e. gas burners) per season, applied on emerging plantlets. The conventional management included fertilization based on one application of vegetable humus (4 kg/m 2 ) and one application of a mix of nitrogen, phosphorus and potassium (15-15-15) (50 g/m 2 ) before transplanting. After transplanting calcium nitrate was applied at three times (one at 20 g/L and two at 10 g/L) and iron chelate was applied once at a dose of 3 kg/1000 m 2 . Pest and diseases were controlled by applying chlorpyrifos (48%, EC) and abamectin (1.8%, EC) as pesticides, combined with copper oxychloride (58.8% WP) as fungicide. Copper oxychloride (100 cc) combined with chlorpyrifos (50 cc) in 20 L of water were applied six times along the growing season, while copper oxychloride (100 cc) combined with abamectin (30 cc) in 20 L of water were applied three times. Adventitious plants were controlled in the same way as described for the organic plot.

Experimental design and sample preparation
For the design of both plots a random distribution model was arranged. Each accession was represented within each growing system by ten plants distributed in two blocks of five plants randomly distributed the plot. In this way, the position effect in each experimental plot was mitigated and each variety was subject to a generalized environmental distortion within the plot, randomly distributing the position effect. Each plot was surrounded by a bordure of plants to prevent the border effect. A total of five samples per accession × growing condition × ripening stage combination were prepared per year. Each sample was prepared with fruits from the two plants of each block (n = 5) and therefore the 10 plants per accession were sampled. Fruits were harvested twice per week at both stages of commercial maturity: i) unripe, i.e. final size, firm fruit but still green, ii) fully ripe, i.e. final size, fully covered of carotenoids and still firm) from the end of June to the end of October. Fruits were harvested at both stages when fitted these criteria according to our experience managing diverse Capsicum germplasm and the commercial expertise of LA UNIÓ staff. Each sample was divided into two subsamples i) 15 g of fresh fruits were used for the ascorbic acid analysis and ii) 30 g of fresh fruits were lyophilized and preserved in dark and dry conditions for the subsequent analysis of phenolics and carotenoids. Due to their very low or nil content in unripe fruits, the analyses of carotenoids were only considered in fully ripe samples, while ascorbic acid and phenolics were estimated in both commercial stages. Thus, a total of 1060 subsamples were prepared (530 unripe stage and 530 from fully ripe stage) for each trait under study in the two years of the experiment, and a total of 2650 analyses were done (1060 unripe stage and 1590 fully ripe stage).

Analytical methods
Ascorbic acid content (AAC), total phenolics (TP) and total red carotenoids and yellow/orange carotenoids (C R and C Y/O ) were estimated following different spectrophotometrical methods with high accuracy and adapted to peppers, which enable to process large amounts of fresh and dried samples, essential for breeders in order to evaluate and compare accurately large collections of cultivars [35,36].
Ascorbic acid content (ACC) was determined reflectometrically, based on the reduction of yellow molybdophosphoric acid to phosphomolybdene blue. The equipment used was a reflectometer RQflex plus and a test of ascorbic acid Reflectoquant (Merck, Darmstadt, Germany). Samples of 15 g of fresh fruit pulp were homogenized with an industrial blender by adding distilled water to facilitate blending. After the homogenization the liquid was filtered with a mesh (0.5 mm) in a 100 mL test tube and it was filled with distilled water up to 50 mL for unripe fruits and 100 mL for fully ripe fruits. The results were expressed in mg of ascorbic acid per 100 g of fresh weight, according to the following equation: Total phenolics (TP) were estimated according to Folin-Ciocalteu methodology, based on a colorimetric reaction quantified by measuring absorbance at 750 nm, and results were referred according to a standard curve of chlorogenic acid as reported by [36]. This analysis is used to determinate total phenolic antioxidants by measuring the reducing capacity [36,37]. Prior to the analysis, each sample was submitted to extraction according to the following protocol: 125 mg of lyophilized fruit were placed in 15 mL Falcon centrifuge tubes and 5 mL of extraction solution (700 mL acetone, 5 mL glacial acetic acid and 295 mL Milli-Q1 water) were added. Then samples were incubated and stirred for 24 hours at room temperature. Afterwards, the samples were centrifuged (3500 rpm for 3 min) and 1.5 mL of the supernatant was collected in a microcentrifuge tube and stored at -80˚C until analysis. For the analyses, samples were centrifuged again (10000 rpm for 5 min) and to 65 μL of the sample were added 500 μL of Folin-Ciocalteu reagent (diluted with water at 10%, v/v) and were incubated for 5 min at room temperature. After that, 500 μL of a saturated sodium carbonate solution (60 g/L) were added and the mixture was incubated for 90 min in darkness at room temperature. Then, an aliquot (200 μL) of each sample or standards was placed in a 96-well microplate, and absorbance was measured at 750 nm using a microplate reader (Bio-Rad iMark TM , Herts, England). Based on the dry matter content of each sample, results were expressed as mg of chlorogenic acid per 100 g of fresh weight.
Determination of carotenoids was carried out according to a modification of the spectrophotometric method developed by [38], which estimates accurately the sum of the main red (TC R = capsanthin + capsorubin) and yellow-orange (TC Y/O = β-carotene + β-criptoxanthin + lutein + violaxanthin + zeaxanthin) isochromic carotenoid fractions in Capsicum fruits. The extraction was made from 100 mg of lyophilized samples in a 100 mL Erlenmeyer flask with 20 mL of acetone and shaking for 1 hour in darkness at room temperature. Then the sample was filtered with filter paper (Watman n˚2) and the filtered dilution was filled up to 25 mL in a volumetric flask with acetone. A sample of 3 ml of this dilution was introduced in glass cells with an optical path length of 10 mm and the absorbance of the extract was measured at 472 and 508 nm using a visible-UV spectrophotometer (UviLine 9400, SI Analytics, Weilheim, Germany) and comparing to acetone as blank. Lambert-Beer equation for multicomponent mixtures was applied according to the following equations: Finally, as for TP, carotenoid content was expressed in mg per 100 g fresh weight.

Statistical analysis
The values of the 16 accessions evaluated both years were used to estimate and correct the year effect. The significance of the year effect was estimated by ANOVA and the data were corrected by subtracting the corresponding mean year deviation (D y ) to the data from all accessions. Mean year deviation was estimated as follows: D y = μy-μ T , where D y is the mean deviation of year y, μ y is the total mean of the 16 accessions in year y and μ T is the total mean of both years. The results of two years were expressed as a mean for each combination of genotype × ripening stage × growing system. Normality was evaluated by W of Shapiro-Wilks test and homocedasticity was analysed by Bartlett and Levene's tests, both performed by Statgraphics Centurion XVI software (StatPoint Technologies, Inc; Warrenton, Virginia, USA). Required transformations were performed to adapt the data to a normal distribution model [39]. Then the analysis of variance ANOVA of main factors genotype, growing system, ripening stage and their interactions were carried out, in both original and transformed data, in order to verify the statistical interpretation. The linear model used for the general ANOVA was X ijkl = μ + a i + b j +c k + (α × β) ij + (α × γ) ik + (β × γ) jk + e ijk(l) , where X ijkl is the value for fruit sample l of genotype i, growing system j and ripening stage k; μ is the general mean; a i is the effect of genotype i; b j is the effect of growing system j; c k is the effect of ripening stage k; (α × β) ij is the interaction between genotype i and growing system j; (α × γ) ik is the interaction between genotype i and ripening stage k; (β × γ) jk is the interaction between growing system j and ripening stage k and e ijk(l) is the effect of fruit sample l from the combination of genotype i, growing system j and ripening stage k (i.e. the error term). Also specific ANOVA considering separately unripe and fully ripe stages were performed and the linear model used was X ijk = μ + a i + b j + (α × β) ij + e ij(k) , where X ijk is the value in one specific ripening stage for fruit sample k of genotype i and growing system j; μ is the general mean; a i is the effect of genotype i; b j is the effect of growing system j; (α × β) ij is the interaction between genotype i and growing system j and e ij(k) is the effect of fruit sample k from the combination of genotype i and growing system j. No differences were found between the original and transformed data in terms of the results of the ANOVA and, therefore, our results were showed and discussed on the basis of the original nontransformed data to facilitate the understanding. Genotype×environment (i.e. accession×growing system) interactions were studied by means of regression analysis [35]. Regression coefficients (β) of each variety were calculated from the average contribution of each growing system according to the formula β i = (μ ij −μ ijk ) / environmental mean = (μ ij −μ ijk ) / (μ j −μ jk ); where β i is the regression coefficient value for each specific trait and ripening stage for genotype i; μ ij is the mean of genotype i in growing system j; μ ijk is the mean of genotype i for both growing systems j and k, μ j is the mean for all the genotypes in growing system j, μ jk is the mean for all the genotypes in both growing systems j and k. The genotypes with β values that did not differ significantly from 0 were considered stable against the growing system effect [40], i.e. they did not show significant differences between the growing systems for the trait under study.

Analysis of variance in ascorbic acid, phenolics and carotenoids
The year effect had significant contribution to the variation observed in all traits, which indicates that all the traits studied experienced significant changes in their levels from one year to the other, as can be expected in open field trials in agriculture. For that reason and in order to use all the data from both years, the values were corrected considering the year effect. The general ANOVA revealed that the effect of the genotype and particularly the ripening stage were significant and contributed considerably to the variation of AAC and TP according to the magnitude of their mean squares ( Table 2). By contrast, the effect of the growing system was only significant for the observed variation in ACC. In addition, the interactions between main factors were also significant and, particularly, those including the ripening stage (Table 2).
Furthermore, mean square values showed a remarkable contribution of the ripening stage in both AAC and TP, not only as a main factor but also as part of interactions with the other factors, which could be biasing the real magnitude of the contribution of the genotype and the growing system, specially the latter. For this reason, further ANOVA considering separately data from unripe fruits and from fully ripe fruits were performed.
Thus, the results of the ANOVAs for each ripening stage showed a significant contribution of both the genotype and growing system factors at both stages, with the only exception of AAC at the unripe stage (Table 2). Furthermore, the contribution of the growing system factor, on the basis of mean square values, increased considerably in comparison to that of the general ANOVA, confirming that the effects of the interactions of the ripening stage with both the genotype and the growing system hidden the real effect of the main factor growing system. Also, the genotype × growing system interaction was found to be significant for AAC and TP at both ripening stages, suggesting that such interactions could be exploited for selection in these traits. On the basis of these results, the descriptive study of the results will be then displayed in the next sections considering separately both the unripe and fully ripe fruits.
Finally, the analysis of variance for carotenoids (only at fully ripe stage) showed a high significant contribution of the genotype as it was observed for AAC and TP, while by contrast both the growing system and its interaction with the genotype did not provide significant contribution to the observed variation in red or yellow/orange carotenoids (Table 2). Therefore, these results suggest that the growing system (organic vs. conventional) had no mean effect on the level of carotenoids, and that the opportunity for exploiting the interaction genotype × growing system is much lower for these compounds than that for AAC and TP.

Ascorbic acid content (AAC)
A broad range of variation for AAC in both growing systems and ripening stages was found in this experiment (Table 3). Thus, mean values of studied accessions at the unripe stage ranged from 14 to 121 mg�100 g -1 of fresh weight for Doux Long des Landes and Petit Marsellaise, respectively, in organic conditions, while values were comprised between 10 to 93 mg�100 g -1 for Jalapa F1 and BOL-144 under conventional conditions (Table 3). At the fully ripe stage AAC values increased considerably with values in organic ranging from 48 to 208 mg�100 g -1 in Ají dulce and Mojo Palmero, respectively, and from 48 to 196 mg�100 g -1 in Serrano and Petit Marsellaise, respectively, in conventional conditions (Table 3). In addition to these wide ranges of variation, it should be noted that the observed AAC values could be considered very high in general. Thus, the recommended daily intake of vitamin C in adults is 60-70 mg [35] and consequently many of our materials may satisfy this threshold with an intake of 50-100 g of fresh fully ripe fruit. Even several accessions at the unripe stage, where the content was clearly lower, also showed very high levels (� 60 mg�100 g -1 ). All these findings demonstrate the richness of Capsicum fruits in this antioxidant. In fact, other vegetables like tomato (25 Table 2. General ANOVA for the content of ascorbic acid (AAC) and total phenolics (TP) and specific ANOVA corresponding to each ripening stage for AAC, TP, total red carotenoids (TCR), and total yellow/orange carotenoids (TCY/O). mg�100 g -1 ), kiwifruit and oranges (50 mg�100 g -1 ) have in general levels considerably lower than those from our collection [41]. Moreover, in comparison to previous reports of peppers in organic [32,42,43], we observed that our ranges of variation were considerably higher, supporting the advantage of using wide collections to assess the real performance of a crop under different growing conditions, which also offers the opportunity of selecting very high-vitamin C genotypes. Moreover, it has always believed that fleshy, sweet, high water-content bell peppers have, in general, lower levels of ascorbic acid than hot/spicy thinner-mesocarp genotypes [41]. However, our results indicate that such acceptance is not so obvious. In fact, in most cases, within each ripening stage and growing condition, the group of bell peppers showed similar or higher AAC average than the other groups with smaller and thin-flesh fruits (other C. annuum and other Capsicums) ( Table 3).
Regarding the effect of the growing system, at the unripe stage we observed as a whole similar mean AAC values in both organic and conventional conditions, around 45 mg�100 g -1 (Table 3), which is in agreement with the nonsignificant contribution of this factor for AAC at the unripe stage detected by ANOVA (Table 2). Nevertheless, this lack of effect in the general mean must be explained in detail, as there was a significant contribution of the growing system but depending on its interaction with the genotype and groups of genotypes. Thus, remarkable differences were found comparing the three Capsicum groups, as well as the performance of individual accessions. Thus, the group of bell peppers C. annuum and other Capsicums showed average values higher in conventional conditions than in organic, while the contrary was true in the other C. annuum group (Table 3). Even more, depending on the growing system, a very variable response could be observed in the different genotypes, which confirmed the significant interaction genotype × growing system detected in the ANOVA for this ripening stage ( Table 2). By contrast, a different behaviour was found at the fully ripe stage, where the total AAC mean was higher in organic than conventional (135 vs. 117 mg�100g -1 ), and the same was found in C. annuum (142 vs. 121 mg�100g -1 in bell peppers and other C. annuum) ( Table 3). Also, most studied accessions showed higher values in organic and only accessions from the other Capsicum species showed an erratic performance (Table 3). All these results were reflected in a high contribution of the genotype×growing system interaction to the variation in AAC at fully ripe fruits ( Table 2).
In this regard, despite ACC average levels at both growing systems were similar at the unripe stage, the considerable contribution of the genotype×environment interaction provided significant differences among cultivars from one growing system to the other. Thus, within the unripe stage and according to the regression coefficient, fifteen accessions showed significantly higher AAC values under organic (β significantly > 0), being Chimayo, Mojo Palmero, Padron, Petit Marsellaise and Piquillo the cultivars with the highest values and increases due to this growing system. By contrast, eighteen genotypes had a better response under conventional practices (β significantly < 0), of which Bola, BOL-37R, BOL-144, California breeding line and Organic cultivation and ripening effects on Capsicum peppers quality Gernika were the ones with the best behaviour in these growing conditions. Finally, four accessions showed a stable behaviour between growing systems (β = 0). In the case of fully ripe fruits, where organic conditions provided on average higher ACC, most cultivars followed this trend and showed significant higher values under organic than under conventional practices, although accessions differed considerably in the magnitude of such differences, which was mainly due to the genotype × environment interaction. Thus, Chimayo, Espellete, Mojo Palmero, Numex Big Jim and Serrano Criollo showed the highest values in organic conditions and considerably higher than their values under conventional conditions (Table 3). Despite this general trend, the genotype × environment interaction also enabled a few accessions to have better AAC values under conventional practices like Ancho, Berbere and Petit Marsellais. Finally, several genotypes did not show significant differences between growing systems, like Jalapenos, Di Senise, Doux Long des Landes and Piquillo and some C. baccatum and C. chinense accessions, which could be considered stable under both systems (Table 3).

Total phenolics (TP)
As it occurred in ascorbic acid, TP values were highly variable depending on the genotype and particularly growing conditions (Table 4). Regarding the genotype effect, at unripe stage and considering organic conditions, TP values were comprised between 68 and 626 mg chlorogenic acid equivalents (c.a.)�100 g -1 fresh weight for California breeding line and BOL-144 respectively, and between 87 and 506 mg�100 g -1 for the same accessions, respectively, under conventional conditions (Table 4). Similarly to AAC, TP values increased in most cases with ripening. Thus, at fully ripe stage, values ranged in organic between 109 and 367 mg�100 g -1 for Aji Dulce and Najerano, respectively, and between 96 and 438 mg�100 g -1 in conventional for Jalapeno Candelaria and BOL-144 respectively (Table 4). In comparison, these TP values in both unripe and fully ripe stages were considerably higher than those from other reference species as tomato, broccoli or banana, with average contents of 68 and 88 mg catechol equiv-alents�100 g -1 fresh weight, and 36 mg gallic acid equivalents�100 g -1 fresh weight, respectively [44,45], are in the lower limits of our TP intervals and therefore reinforces the idea of peppers as fruits very rich in antioxidants.
The growing system had a significant effect in the phenolic values for both unripe and fully ripe stage as indicated by the ANOVA (Tables 2 and 4). Thus, at the unripe stage, total mean was higher in conventional conditions than organic, with values of 151 and 126 mg�100 g -1 , respectively, and the same trend was found in C. annuum (80 and 113 mg�100 g -1 in organic and conventional in bell peppers; 123 and 147 mg�100 g -1 in other C. annuum), while means were similar in other Capsicum species (about 240 mg�100 g -1 ) ( Table 4). Moreover, considering further the genotype×environment interaction, in terms of both the magnitude of the increase/decrease of TP and the sense of this behavior and according to the regression study, twenty-five accessions showed higher TP values in the conventional system, being Berbere, Doux Long des Landes, Gernika and several genotypes from other Capsicum species, particularly PI-152225, the accessions with the highest values (200-500 mg�100 g -1 ), while seven accessions had higher TP values under organic cultivation, being Jalapeno Candelaria, Padrón, BOL-37R, BOL-144 and ECU-973 the ones with the highest values (190-600 mg�100 g -1 ), and five accessions were stable (Table 4).
Contrary to these results, fully ripe fruits showed, on the whole, higher TP levels under organic conditions than under conventional conditions (total means of 232 vs 206 mg�100 g -1 , respectively) and the same was found in most cultivars, particularly in C. annuum (226 vs 185 mg�100 g -1 ) ( Table 4), similarly to AAC. Nevertheless, other Capsicum species showed a less consistent trend, highly depending on the accession, and means from organic and Table 4. Average content in total phenolics content (TP, mg chlorogenic acid equivalents�100 g -1 fresh weight) estimated in the unripe and fully ripe fruits from the peppers studied, evaluated under organic and conventional conditions and regression coefficient (β) based on environmental means.

Genotype
TP unripe stage (mg�100 g -1 ) TP fully ripe stage (mg�100 g -1 ) conventional conditions were very similar (about 230 mg�100 g -1 ). This fact confirmed again the high contribution of the ripening stage and their interactions to the observed variation in TP, as well as in AAC (Table 2). Regarding the individual behavior of the different accessions, twenty-two accessions had better behaviour in organic, where BOL-58, Guindilla Ibarra, Numex 6-4, Najerano and Petit Marsellaise showed the highest values (300-360 mg�100 g -1 ), while only six accessions showed a better performance in conventional conditions, being Gernika and BOL-144 the most outstanding materials (>300 mg�100 −1 ), and nine accessions were stable (Table 4).

Total red carotenoids and yellow/orange carotenoids (TC R ,TC Y/O ).
The range of variation in carotenoids in fully ripe Capsicum fruits was very wide, similar to that observed in AAC and TP, and it confirmed the high contribution of the genotype factor to the observed variation (Tables 2 and 5). Organic conditions registered values of red carotenoids (TC R ) comprised between 0 and 257 mg�100 g -1 of fresh matter from yellow-fruited cultivars like Cuneo or Petit Marsellais and brownish-fruited cultivar Pasilla, respectively, and TC Y/O between 2 and 147 mg�100 g -1 for Jalapeno Espinalteco and BOL-58 respectively (Table 5). Within the conventional growing system, TC R values ranged again between 0 and 324 mg�100 g -1 in the same genotypes than organic system, and TC Y/O were comprised between 1 and 166 mg�100 g -1 for Jalapeno Espinalteco and Gernika respectively (Table 5). These values of total carotenoids were relatively high in comparison to other vegetables and fruits like broccoli, tomato, carrot, apricot or squash (around 5, 7, 15, 17 and 41 mg�100 g -1 , respectively) [46][47][48] and makes clear that Capsicum pods are a rich source of this bioactive compounds. In addition, by contrast to AAC and TP results, bell peppers had, in general, considerably lower amounts of carotenoids than the other C. annuum and other Capsicum species. Thus, the average mean of bell peppers was 18 and 11 mg�100 g-1 in TC R and TC Y/O , respectively, and only Bierzo (65 and 30 mg�100 g -1 ) reached the usual levels of the other groups, which showed a thin mesocarp and were less fleshy than bell peppers, with the only exception of Arnoia and Jalapeno types. Such results can be explained by the high correlation between carotenoids and dry matter content in peppers [41].
Moreover, the level in carotenoids showed by several accessions was very interesting. Thus, Pasilla was the cultivar with the highest values TC R and very high TC Y/O , regardless the growing system (Table 5). This fact suggests paradoxically that the chlorophyll retainer (cl) mutation, characterized by inhibition of chlorophyll degradation during fruit ripening and causing brown fruits (due to the combination of chlorophylls and red carotenoids) [49], also provides high levels of carotenoid pigments [29]. In addition, other peppers like Bola, Espelette or Mojo Palmero also showed very high TC R levels (Table 5), which explain why they were mainly selected to be used as food colorants in different food preparations (e.g. ground powder, sauces, spicy pork sausages and chorizo) [8]. Moreover, it should be also pointed out the performance of cultivars like Gernika, Guindilla de Ibarra or Padron, which have been selected Organic cultivation and ripening effects on Capsicum peppers quality Table 5

. Average content in total red carotenoids and yellow/orange carotenoids (TC R and TC Y/O , mg�100 g -1 fresh weight) estimated in the unripe and fully ripe fruits from the peppers studied, evaluated under organic and conventional conditions and regression coefficient (β) based on environmental means.
Genotype TC R fully ripe stage (mg�100 g -1 ) TC O/Y fully ripe stage (mg�100 g -1 ) Organic for consumption at the unripe stage (i.e. roasted or pickled at very early unripe fruit stage), had very high average measures of TC R , suggesting that their use at the fully ripe stage can be promoted as, in fact, it has been partly started in the Basque Country, where they are called Chorizeros (to make chorizo) when fully ripe [8]. Finally, Di Senise, Numex Big Jim and Piquillo, despite used when fully ripe but not as food colorants they also showed high TC R levels (>100 mg�100 g -1 ), as well as the exotic BOL-58 and ECU-994 (100-200 mg�100 g -1 ) ( Table 5), indicating that materials not selected originally as food colorants can be also fully ripe selected for high red carotenoid content.
In terms of yellow/orange carotenoids, we found a high correlation with red carotenoids and therefore those accessions with the highest TC R values mentioned before were also the ones with the highest TC Y/O values (70-170 mg�100 g -1 ). These levels in yellow and red carotenoids (including β-carotene and β-cryptoxanthin) indicates that these accessions can provide high levels of provitamin-A as reported in previous works [29,31] while yellow-fruited accessions Cuneo and Petit Marsellais showed very low TC Y/O values (Table 5). Such fact can be explained because red-fruited peppers contain both red and yellow/orange carotenoids, and the red pigments (capsanthin and capsorubin) are synthetized from yellow and orange precursors, feeding back the new synthesis of yellow/orange carotenoids [50].
Total average was similar in both growing systems for both TC R and TC Y/O confirming the nonsignificant effect of growing system detected in the ANOVA (Tables 2 and 5). Thus, TC R total average was 93 and 103 mg.100 g -1 for organic and conventional systems respectively, and around 58 mg.100 g -1 for TC Y/O in both systems, and the same was found considering the different groups of peppers (Table 5). Moreover, in comparison to AAC and TP very few accessions showed significant differences for TC R and TC Y/O between organic and conventional conditions, which explains the lack of significance of the genotype × growing system interaction (Table 2). Thus, for TC R only three accessions showed higher levels in organic conditions compared to conventional (ECU-994, Guindilla Ibarra and PI-152225), while the contrary was found in seven accessions ( Table 5). The behavior in TC Y/O was similar with only two accessions significantly better under organic conditions (Guindilla Ibarra and Numex Big Jim) and five under conventional conditions (Ancho, Gernika, Jalapa F1, Mojo and Pasilla) ( Table 5). This fact was previously reported by some authors [51,52], although other studies found significant differences due to growing conditions [32,34]. Moreover, it has been described that the content in carotenoids depended on genotype, soil type and grower-specific inputs as fertilizers and methods of pests control [53]. From our experiment, performed with a large diversity of varietal types, we can conclude that the factors that enhance in general AAC and TP under organic conditions do not have this effect in carotenoids. Probably the antioxidant nature of ascorbic acid and phenolics drives to increase their levels in plant tissues in response to free-radicals accumulation under more stressing-low input conditions of our Organic cultivation and ripening effects on Capsicum peppers quality organic trials. By contrast, the primary role of carotenoids as photoreceptors and photoprotectors did not provide any response.

Ripening process in AAC and TP under organic and conventional practices
Ripening process favoured the accumulation of AAC in fully ripe fruits. Thus, all the varieties showed higher AAC at fully ripe stage than at unripe stage (slope fully ripe/unripe > 1, Fig 1). These findings agree with reports from other authors [19,25,54], although the large number of accessions used in our study showed a wide variation in the magnitude of such increase, mainly depending on the genotype but also modified by growing conditions. In this regard, some accessions showed an increase in AAC during ripening comprised between 1% and 100% (1< slope � 2), although in most accessions the ripening process increased AAC more than 100% (slope > 2) (Fig 1). In these last cases, there was a range of increases with ripening and even some accessions showed levels at the fully ripe stage 4-8 fold from those at the unripe stage (Fig 1).
In comparison, organic conditions enabled higher increases in AAC than conventional conditions, and data from organic grouped mainly in the areas of higher slopes (Fig 1). This was in agreement with the results observed on Table 3, where AAC total means increased with Fig 1. Comparative study of the content of ascorbic acid (AAC, upper, mg�100 g -1 fresh weight) and total phenolics (TP, bottom, mg�100 g -1 fresh weight), in organic (�) and conventional (Δ) growing systems. The oblique lines included in the graph represent slope 1 (continuous line, ratio fully ripe/unripe = 1, increase 0% with ripening) and slope 2 (discontinuous line, ratio TP fully ripe/TP unripe = 2, increase 100% with ripening). https://doi.org/10.1371/journal.pone.0207888.g001 Organic cultivation and ripening effects on Capsicum peppers quality ripening but at higher extent in organic, and also the increase of most accessions was higher under organic.
As observed for AAC, the accumulation of TP increased with the ripening process in general and in most genotypes (slope fully ripe/unripe � 1 , Fig 1), although this effect was not so consistent and even some accessions, five in organic and ten in conventional, had lower values at the fully ripe stage (Fig 1). Our results agree with a review of other reports. Thus, the accumulation of phenolic compounds with ripening has been described but also the decrease and even depending on the cultivars [55][56][57].
The studied accessions showed different magnitudes of TP increase with ripening, as genotype × ripening stage interaction showed in the ANOVA (Table 2) as can be observed in the range of slopes displayed in Fig 1. It should be also noted that the increase in TP with ripening could be modified by growing system conditions (i.e. growing system × ripening stage interaction, Table 2). In this sense, within the organic growing system eleven accessions had slopes comprised between 1 and 2 and twenty-one accessions had slopes between 2 and 5 ( Fig  1). By contrast, within the conventional system twenty-four accessions showed slopes comprised between 1 and 2, while only three accessions had slopes ranged between 2 and 3 (Fig 1). Thus, similarly to AAC, the organic growing system appeared to increase the accumulation of TP with the ripening process at a higher extent than conventional conditions. These results suggest that, in general, organic conditions may strengthen the increase of AAC with the ripening process in a more intense way than conventional conditions. In this regard, the different behaviour detected between unripe and fully ripe fruits in terms of TP and AAC should be considered. Thus, the levels of both antioxidants in fully ripe fruits were higher (in general and in most accessions) in organic cultivation, while the contrary was true in unripe fruits. These findings suggest that some factors present in organic conditions may increase the accumulation of antioxidants in the fruits of most Capsicum genotypes, but mostly in fully ripe fruits. Perhaps, fruits during ripening are more sensitive to factors from soils, like those from high fertilized conditions typical of conventional practices, in comparison to less fertilized (organic) conditions. In this regard, some studies [58][59][60][61] reported that high rates of specific mineral fertilization could lead to nutrient dilutions and interactions. Thus, conventional soil management could cause interactions that disturb the balance of soil nutrients, essential in the metabolic pathways of biosynthesis, which could be one of the reasons that leads to a lower accumulation of ascorbic acid and phenolics at the end of ripening.

Potential selections of materials for organic cultivation
Based on the results of our trials, a potential selection of plant material for organic cultivation would be possible considering high mean levels of bioactive factor, coefficient of regression, and varietal groups, as well as each ripening stage due to the importance of this factor for the commercial use of fruits.
Thus, within the unripe stage, C. annuum accessions such as Chimayo, Gernika, Jalapeno Candelaria, Mojo Palmero, Padrón, Pasilla, Petit Marsellais and Serrano Criollo would be the best options as their sum of AAC and TP reaches 200-300 mg�100 g -1 , as well as Pimiento Valenciano among the fleshy bell peppers group (Fig 2). Also, many accessions from the exotic species like C. chinense Aji Dulce, ECU-973 and ECU-994, C. baccatum BOL-37R and particularly C. frutescens BOL-144 showed an outstanding performance, which offers an opportunity to these materials in new markets with the added value of their exotic appearance and levels of bioactive compounds under organic practices.
Considering the fully ripe stage, several materials could be selected for organic cultivation. Thus, within C. annuum accessions like Bola, di Senise, Doux Long des Landes, Espelette, Gernika, Guindilla de Ibarra, Mojo Palmero, Numex 6-4, Numex Big Jim, Pasilla, Padron, Petit Marsellais and Piquillo, as well as Najerano, and even Bierzo, among the bell peppers group, offered very high total levels of bioactive compounds (500-800 mg.100 g -1 ) (Fig 3). Even more, C. annuum Chile de Arbol and Chimayo and C. chinense PI-152225 showed high levels of AAC and TP and Petit Marsellais, despite of very low carotenoids levels, is a good alternative among yellow/orange peppers. In addition, C. baccatum BOL-58, C. chinense ECU-994 and C. frutescens BOL-144 showed high total levels and could be selected among the exotic species for a niche of exotic products within the organic market, especially BOL-58 and ECU-994, because of their particular aroma and flavour at the ripe stage as reported by [62].
Finally, most of the mentioned materials also showed a high performance in conventional conditions, which suggests a wide adaptation to other growing conditions, including different organic production units (Figs 2 and 3). Even more, several materials like Chimayo, Disenise, Gernika, Guindilla Ibarra, Mojo Palmero, Padrón, Pasilla, Petit Marsellais, Piquillo, and the Organic cultivation and ripening effects on Capsicum peppers quality exotic C. chinense ECU-994 and C. frutescens BOL-144 offered remarkable levels of bioactive compounds regardless the ripening stage (Figs 2 and 3) and, therefore, its cultivation could be advised for producing high-quality fruits both unripe and fully ripe stages.

Conclusions
In conclusion, our findings enabled to assess the impact of factors like the genotype, the ripening stage and their interactions on the response to organic cultivation (compared to conventional) for the levels of the main bioactive compounds in peppers, founded on a comprehensive varietal diversity of Capsicum. A wide range of variation was found among the studied accessions, regardless the ripening stage and growing conditions. Also, fully ripe fruits were found considerably richer than unripe fruits in the studied bioactive factors, with the only exception of very few accessions in phenolics; and organic cultivation provided, as a whole, Organic cultivation and ripening effects on Capsicum peppers quality higher levels of ascorbic acid and phenolics at this stage. Moreover, the remarkable genoty-pe×environment interaction allowed the selection of several accessions with a remarkable performance under organic conditions at both ripening stages. By contrast the effects of growing conditions and the genotype×growing conditions were nil on the levels of carotenoids, whose variation was mainly due to the genotype factor. The results will be very helpful for breeders and scientists aimed at breeding high-added vegetables for adaptation to organic production systems.