Variation of anthocyanins and other major phenolic compounds throughout the ripening of four Portuguese blueberry (Vaccinium corymbosum L) cultivars

Abstract Blueberries are widely recognised as one of the richest sources of bioactive compounds, among which are anthocyanins, though the ripeness of berries has been reported as affecting the phytochemical composition of fruits. Therefore, the present work aimed to evaluate the variation of anthocyanins, and other major phenolics, throughout five ripening stages in four blueberry cultivars. The results showed that the antioxidant capacity and anthocyanin content increased during ripening, reaching the highest values when the blueberries are collected from bunches comprised of 75% ripe blueberries. Antagonistically, the amount of phenolic acid decreases, while the quercetin-3-glucoside levels remain stable. Furthermore, Goldtraube blueberries appear to possess, systematically, higher amounts of phenolic compounds than the other cultivars studied. Thus, when seeking the highest yield of anthocyanins, the preferred harvest should occur in bunches that contain ca 75% of ripe blueberries and, considering the cultivars assayed, the Goldtraube cultivar appears to be the richest in phenolic compounds.


Introduction
As consumers' concern with health grew, so did the industry's attempts to explore this market through the development of functional foods. Primary food producers focussed on further developing their growth, harvest and storage protocols in an attempt to maximise their products functional potential (Gibson & Williams 2000). Blueberries are a prime example of a foodstuff that gained a strong health-promoting connotation, with general media calling it a 'superfruit' and several reports made on their rich phenolic composition. Anthocyanins, a major group of phenolic compounds found in blueberry, have been associated with a vast array of health-promoting properties such as protection against oxidative damage, reduction of hepatic damage, anti-inflammatory activity and even with weight management (li et al. 2013;Sarkar et al. 2014;Cardile et al. 2015;Cristani et al. 2016;Jiang et al. 2016). As blueberry phenolic composition varies significantly according to the ripeness of the fruit, and considering the lack of a compound-specific approach analysis of this process, the present work aimed to compare the phenolic composition of four blueberry cultivars throughout the ripening of the fruit, with a particular emphasis on their anthocyanin composition.

Results and discussion
As shown in Figure 1, the variation of the total antioxidant capacity (TAC), phenolic content (TPC) and anthocyanin content (ACY) varied differently throughout the ripening stages, though the variation profile differed from cultivar to cultivar. For monomeric anthocyanins, the ACY is significantly (p < 0.05) higher in the ripe and over-ripe stages (S3-S5) than in the unripe fruits (S1 and S2), a fact that stands in line with previous reports (Kalt et al. 2003;Castrejón et al. 2008). The TAC variation was quite similar to that of the anthocyanins for all cultivars with the exception of Bluecrop ( Figure 1). In fact, the correlation between these variables (Table 3S) showed that the TAC values observed for Duke, Goldtraube and Ozarkblue blueberries were related to their anthocyanin content as they exhibited a strong, significant (p < 0.01), positive relationship (0.980, 0.932 and 0.900, respectively). As the TAC content does not vary for Bluecrop and the ABTS method is highly susceptible to interference, it is possible that another compound/factor is responsible for the absence of correlation between these variables. As for the TPC, no overall behavioural trend could be observed, a fact that stands in line with what has been described by Castrejón et al. (2008), who reported that four different blueberry cultivars exhibited different phenolic variation trends. Considering the vast array of compounds that are capable of interfering with the Folin-Ciocalteu assay, it is hard to explain TPC variations without looking at the individual compounds (Kalt & McDonald 1996;Wang & lin 2000;vermerris & Nicholson 2007;Castrejón et al. 2008). In Figure 2, where the variation of the different anthocyanins along the ripening is expressed, it can be seen that the amount of anthocyanins increased with the ripening. This stands in accordance with the results obtained for the ACY (Figure 1). The most prevalent anthocyanins found were malvidins, which is in accordance with what has been reported by Taruscio et al. (2004). This is particularly interesting when considering the biological potential of the extracts as malvidins have been described as possessing anti-inflammatory activity and, therefore, the extracts may be used to develop nutraceuticals that aid with chronic inflammation (Huang et al. 2014). Delphinidins were found at higher levels than cyanidins which is in accordance with Taruscio et al. (2004)  Besides anthocyanins, other four phenolic compounds were identified, three phenolic acids (chlorogenic, gallic and p-coumaric acid) and a flavonol glycoside (quercetin-3-d-glucoside (Q3Glu)) ( Figure 3). Cinnamates (such as chlorogenic and hydroxycinnamic acids) have been reported to decrease (p < 0.05) as the fruit ripens, a behaviour that stands in accordance with the results observed for p-coumaric acid ( Figure 3) and for chlorogenic acid when considering Duke, Ozarkblue and Bluecrop blueberries (Kalt & McDonald 1996;Kalt et al. 2003;Castrejón et al. 2008). However, as this behaviour is not ubiquitous to all food crops, the stable values observed for Goldtraube are not unexpected. Gallic acid exhibited behaviour similar to the one observed for chlorogenic acid. As a reduction in acidity is one of the main characteristics of the ripening process, the variations observed stand in accordance with what could be expected (Ayaz et al. 2001;Kulkarni & Aradhya 2005). The Q3Glu values remained relatively stable throughout the assay. Fact that is not in accordance with what was reported by Jaakola et al. (2002) that found that quercetin levels varied throughout the ripening process, first increasing and then decreasing. Given the relatively large deviations to the mean found, it is possible that any variations are masked by the dispersion of the data.
It is interesting to note that the sum of each individual phenolic compound (as determined by HPlC) is significantly higher than the TPC or the TAC (Table 5S) which is a consequence of the increased sensitivity of the chromatographic methods by opposition to the spectrophotometric ones (lee et al. 2008).

Conclusion
Despite the various variation profiles found, it can be concluded that the ripening stage S4, i.e. blueberries collected from bunches containing 75% of ripe blueberries, allows for the overall maximisation of the anthocyanins in all blueberry cultivars tested.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
The present work was supported by the National Funds from Fundação para a Ciência e a Tecnologia