Correlations of Antioxidant Activity against Phenolic Content Revisited: A New Approach in Data Analysis for Food and Medicinal Plants

Authors

  • D.A. Jacobo-Velázquez,

    1. Authors are with Dept. of Horticultural Sciences, Texas A&M Univ., Vegetable & Fruit Improvement Center, College Station, TX 77843-2133, U.S.A. Direct inquiries to author Cisneros-Zevallos (E-mail: lcisnero@tamu.edu).
    Search for more papers by this author
  • L. Cisneros-Zevallos

    1. Authors are with Dept. of Horticultural Sciences, Texas A&M Univ., Vegetable & Fruit Improvement Center, College Station, TX 77843-2133, U.S.A. Direct inquiries to author Cisneros-Zevallos (E-mail: lcisnero@tamu.edu).
    Search for more papers by this author

Abstract

ABSTRACT:  This study presents a new approach to analyze data correlating total antioxidant activity and total phenolic compounds in foods. The correlation of both variables is a common practice found in the literature. The purpose of these correlations is to determine the contribution of phenolics to the total antioxidant activity of foods. When low R2 values are obtained, the general conclusion is that other compounds have a higher relevance than phenolics in the total antioxidant activity of the samples. However, these correlations do not consider differences in the phenolic profiles that can be qualitatively (type of phenolics present) and quantitatively (the relative amounts or proportions of phenolics present) among the samples under investigation. The new approach to analyze these simple correlations presented herein takes into consideration the phenolic profiles and provides information on the effectiveness of phenolics present in the samples to neutralize free radicals. Data obtained from carrots stored under conditions of air and hyperoxia (superatmospheric oxygen) are used to exemplify how to apply this new approach.

Introduction

Fresh fruit and vegetables as well as other types of foods (that is, red wine, fruit juices, grain product, and so on) are rich in antioxidants that protect cells against the detrimental effect of reactive oxygen species (ROS) (Velioglu and others 1998; Prior 2003; Wu and others 2004). The measurement of the antioxidant activity is a common practice for the determination of the potential inhibition or scavenging capacity of foods against ROS. Several assays including 2, 2-azobis (3-ethyl-benzothialzoline-6-sulfonic acid) (ABTS), 2,2-diphenyl-1-picrylhydrazyl (DPPH), ferric reducing antioxidant power (FRAP), and the oxygen radical absorption capacity (ORAC), among others, are frequently used to estimate antioxidant capacities in fresh fruits and vegetables. Recent publications review the chemistry behind these assays (Huang and others 2005), their standardization (Prior and others 2005), validity (Yoo and others 2007), and relative values among different types of phenolic compounds (Fernandez-Panchon and others 2008; Tabart and others 2009).

It is well known that phenolic compounds exert antioxidant activity in biological systems (Husain and others 1989; Rice-Evans and others 1997). To estimate the inhibitory capacity of these compounds against ROS, data correlating antioxidant activity and phenolics concentration are commonly reported. However, when the R2 values from these correlations are low, it is concluded that phenolic compounds are not responsible of the antioxidant activity (Al-Saikhan and others 1995; Scalzo and others 2005; Deepa and others 2007; Patthamakanokporn and others 2008). Although valid in some cases, this interpretation does not consider factors such as the differences in the phenolic profiles that can be qualitatively (type of phenolics present) and quantitatively (the relative amounts or proportions of phenolics present) between samples. The samples studied can include different genotypes (Chirinos and others 2008a; Tulipani and others 2008), samples of different maturity stages (Castrejón and others 2008), or even samples exposed to different postharvest storage conditions (Rapisarda and others 2008). Several studies using correlation analysis have been conducted on fresh produce and food products without considering the phenolic profiles factor (Prior and others 1998; Kalt and others 1999; Prior and others 2005; Gu and others 2006).

The antioxidant activity of a specific phenolic compound is related with the number of available hydroxyl groups present in the chemical structure (Rice-Evans and others 1996). Therefore the manner these compounds neutralize free radicals will depend on their relative concentrations in the sample matrix. In addition, phenolic compounds can act synergistically (Cirico and Omaye 2006), additively (Heo and others 2007), or antagonistically (Peyrat-Maillard and others 2003) to inhibit reactive species. However, the typical approach used to analyze correlations between total antioxidant activity and total phenolics do not reflect these characteristics of phenolics. In recent years our research group has proposed the concept of the specific antioxidant capacity (ratio of total antioxidant capacity per total soluble phenolics) as an adequate parameter to understand the effectiveness of a mixture of phenolic compounds to neutralize free radicals (Cevallos-Casals and Cisneros-Zevallos 2003; Reyes and Cisneros-Zevallos 2003; Reyes and others 2005, 2007; Villarreal-Lozoya and others 2007; Vizzotto and others 2007; Heredia and Cisneros-Zevallos 2009). This parameter can be used by crop breeders as a tool to screen among different genotypes, maturity stages, and postharvest storage conditions to generate crops with phenolic profiles high in antioxidant activity.

In this study, we report data from the literature that support our proposal that from simple correlations between antioxidant activity, compared with phenolic content, it is possible to obtain more information related to the specific characteristics of phenolic compounds present in the sample. In addition, new data obtained from carrots stored under different atmospheric conditions (hyperoxia and air) are presented to show how to analyze the data using the proposed new approach.

Phenolic Compounds are the Major Contributors to the Antioxidant Activity of Fruits and Vegetables

The consumption of fruits and vegetables is associated with reducing the risk of degenerative diseases produced by free radicals (Ames and others 1993; Prior and Cao 2000; Wargovich 2000). In the past, this protection against different diseases was mainly attributed to their high content of vitamin C, E, and carotenoids (Block 1991; Zeigler 1991; Byers and Guerrero 1995; Vinson and others 1998). However, since clinical studies indicated that pure doses of these compounds supplemented to humans did not reduce the incidence of heart diseases, cerebral infraction, and the risk of strokes, the effectiveness of these antioxidants to scavenge free radicals was questioned (Vinson and others 2001). Since then, phenolics were identified as the compounds with the highest antioxidant capacity in fruits and several research groups directed their efforts in understanding their antioxidant potential. Nowadays it is well known that phenolics are the major contributors to the total antioxidant capacity of fruits, vegetables, and grains (Heo and others 2007).

To understand the contribution of phenolics to the total antioxidant activity of food and medicinal plants, several researchers correlate both variables and obtain a R2 value. The samples under investigation can comprise different genotypes, maturity stages, storage, and processing conditions. To exemplify the traditional approach to analyze correlations of antioxidant activity compared with phenolic content, data previously reported in the literature for crops with different genotypes were collected and plotted (Figure 1). Since the relative values of total antioxidant activity can vary depending on the assays used (that is, ABTS, DPPH, FRAP, ORAC), data were divided in 2 plots. The plotted values correspond to the reported antioxidant activities measured by either the DPPH (Figure 1A) or ABTS (Figure 1B) method. An analysis of regression among the genotypes from the same crop was performed and the R2 values were obtained. High correlations between antioxidant activity and total phenolics (R2 > 0.7) were obtained for genotypes of sweet potatoes, peaches, and mashua and oca tubers, while low correlations (R2 < 0.7) were obtained for plums and native potatoes from Perú. The common interpretation of these correlations is that phenolic compounds contribute highly to the total antioxidant activity of sweet potatoes, peaches, and mashua and oca tubers. On the other hand, it would be concluded that phenolic compounds contribute slightly to the total antioxidant activity of plums and native potatoes from Perú; however, this interpretation could be misleading since it does not consider differences in the phenolic profiles that may exist among the different genotypes.

Figure 1—.

Correlation between total phenolics and total antioxidant activity for different horticultural crops previously reported in the literature. The antioxidant activity was measured by either (A) 2,2-diphenyl-1-picrylhydrazyl (DPPH) or (B) 2,2-azobis(3-ethyl-benzothialzoline-6-sulfonic acid) (ABTS) methods. Each individual point represents different genotypes. 1Teow and others 2007, 2Cevallos-Casals and others 2006, 3Campos and others 2006.

Qualitative and quantitative differences in the phenolic profiles between genotypes have been previously reported for mashua tubers (Chirinos and others 2008a), oca tubers (Chirinos and others 2008b), peaches, plums (Tomas-Barberan and others 2001; Andreotti and others 2008), and sweet potatoes (Teow and others 2007) among other crops.

The Phenolic Profiles of Fruits and Vegetables Determine the Antioxidant Activity

Phenolics encompass a wide variety of compounds derived from the phenylpropanoid metabolism. The antioxidant activity of each phenolic compound can be different and will depend on their donor-proton capacity (Rice-Evans and others 1996, 1997; Cao and others 1997).

For flavonoids (flavonols, isoflavones, and so on), which have a diphenylpropane skeleton, their efficiency as free-radical scavengers seems to depend mainly on the number of hydroxyl groups and their position on the molecule. The antioxidant potency is related to structure in terms of electron delocalization of the aromatic nucleus. For example, the torsion angle of the B ring with respect to the rest of the molecule in flavones and flavanones compared to a planar arrangement in flavonols and flavanols with a 3-OH suggests that the planar conformation favors conjugation, electron delocalization, and an increase in phenoxyl radical stability (Van Acker and others 1996). Additionally, the presence of glycosylations on the molecule may decrease its antioxidant activity. In general, flavonoids structural arrangements are considered to impart greatest antioxidant activity (Rice-Evans and others 1996, 1997). For phenolic acids (hydroxybenzoic, hydroxyphenylacetic and hydroxycinnamic acids) and their ester derivates it is known that antioxidant activity depends on the number of hydroxyl groups in the molecule that are affected by steric hindrance from their carboxylate group (Rice-Evans and others 1996). The closeness of the carboxylate group and the hydroxyl groups on the phenolic ring in hydroxybenzoic acids negatively affects their donor–proton ability. As a result, higher antioxidant activities are usually observed on hydroxycinnamic acids (that is, coumaric, caffeic, and ferulic acid) compared to their hydroxybenzoic counterparts (Rice-Evans and others 1996).

Reyes and others (2007) reported the antioxidant capacity of a series of dilutions of quercetin (Q), chlorogenic acid (C), and their mixtures (Q/C, 75/25, 25/75). When the researchers plotted the antioxidant capacity against phenolic content of the dilutions, they found linear relationships with differing slopes for each phenolic sample tested as well as for their mixtures (Figure 2). Similar results were found by Heo and others (2007) when they evaluated the antioxidant capacity of solutions of both, isolated and combined phenolics. The researchers found that the total antioxidant capacity of the mixtures containing 2 or 3 phenolics is the result of the summation of antioxidant capacities of individual phenolics. These results demonstrate that the antioxidant activity is directly related with the phenolic profiles.

Figure 2—.

Antioxidant capacity of solutions prepared with quercetin (Q), chlorogenic acid (C), and their mixtures (Q/C) in different proportions. Reproduced from Reyes and others (2007). Each individual point represents values of a series of dilutions of pure phenolics or mixtures.

The Specific Antioxidant Activity Is an Adequate Parameter to Estimate the Effectiveness of a Mixture of Phenolics to Neutralize Free Radicals

The specific antioxidant capacity can be defined as the ratio of total antioxidant capacity per total soluble phenolics and expressed as micromol of Trolox equivalents per micromol or milligram of phenolics. This parameter provides information of the effectiveness of phenolic compounds to neutralize free radicals and will correspond to the slope of an antioxidant activity against phenolic content plot. A higher value of specific antioxidant capacity means phenolic compounds have a higher capacity to stabilize free radicals. Vizzotto and others (2007) calculated this parameter using the total phenolic values obtained by the Folin–Ciocalteau method and the antioxidant activity estimated by the DPPH radical method for different genotypes of peaches and plums. The researchers obtained specific antioxidant activity values that varied from 0.42 to 1.8 μmol of Trolox equivalents per micromol of chlorogenic acid for plums and from 0.35 to 3.12 μmol of Trolox equivalents per micromol of chlorogenic acid for peaches. Additionally, they determined that the contribution of carotenoids to the total antioxidant activity of peaches and plums is minimal when compared with phenolics. Therefore, the large variation found in the specific antioxidant activity values was attributed to differences in the phenolic profiles (qualitative and quantitative) among genotypes.

In previous research we have shown that the specific antioxidant activity of pure quercetin (Q) and chlorogenic acid (C) were 2638 and 1470 μg Trolox/mg phenolic, respectively, while mixtures of Q/C (75/25) and Q/C (25/75) resulted in specific antioxidant activity values of 2172 and 1907 μg Trolox/mg phenolic, respectively (obtained from the slopes in Figure 2) (Reyes and others 2007). Assuming an additive effect, theoretical calculations for the Q/C mixtures would have given values of 2346 and 1762 μg Trolox/mg phenolic for QC (75/25) and QC/C (25/75), respectively, suggesting that Q/C mixtures can interact antagonistically or synergistically to neutralize free radicals. This confirms that phenolic profiles affect the specific antioxidant activity values of the samples.

As phenolics are the compounds with major relevance in the total antioxidant capacity of fruits and vegetables, they can be a target for breeding programs. Selecting the genotypes with higher specific antioxidant activity would help in increasing the consumption of fruits and vegetables rich in phenolics with a high antioxidant capacity. In addition, as the phenolic profiles of horticultural crops are affected by the maturity stages (Castrejón and others 2008) and storage conditions (Rapisarda and others 2008) the specific antioxidant activity can be used to select the proper maturity stage of harvest (Brovelli and Cisneros-Zevallos 2007) as well as the adequate storage conditions to obtain highly protective crops against free radicals. This approach may be used when targeting alternative markets such as the functional foods, dietary supplements, pharmaceuticals, cosmetics, and agrochemicals (Cisneros-Zevallos 2003).

Specific Antioxidant Activity and Phenolic Profiles: Using Carrots as an Example

Our research group has demonstrated that the use of postharvest abiotic stresses (that is, wounding, methyl jasmonate, ethylene, ultraviolet radiation) in fresh fruit and vegetables increases the antioxidant activity as well as the phenolic content and induces changes in the phenolic profiles, which alters the specific antioxidant activity (Reyes and Cisneros-Zevallos 2003; Reyes and others 2007; Heredia and Cisneros-Zevallos 2009). When an abiotic stress is applied during postharvest, the phenylpropanoid metabolism of crops is activated inducing the accumulation of phenolics. Cisneros-Zevallos (2003) proposed the use of controlled postharvest abiotic stresses to stimulate the synthesis of antioxidants in fresh fruits and vegetables. The alteration of the atmospheric gas composition during storage of horticultural crops is an abiotic stress that can stimulate the synthesis of phenolics. In this context, superatmospheric oxygen has been previously reported to accumulate phenolic compounds in blueberry (Zheng and others 2003) and strawberry (Zheng and others 2007).

To demonstrate that the storage conditions of horticultural crops can affect the phenolic profiles as well as the specific antioxidant activity and to exemplify how to obtain valuable information when correlating total antioxidant activity against total phenolics, an experiment was performed using carrots as model systems. Whole carrots (Daucus carota) were stored under superatmospheric oxygen (80% oxygen and 20% nitrogen) and air for 48 h at 20 °C. Before and after storage, variables such as total soluble phenolics, HPLC phenolic profiles, and total antioxidant activity (ORAC value) were evaluated with the methods described by Swain and Hillis (1959), Heredia and Cisneros-Zevallos (2009), and Wu and others (2004), respectively.

A significant increase (P < 0.05) in the total phenolic content and total antioxidant activity of carrots was observed after their storage at 20 °C for 48 h (Figure 3). The obtained data were analyzed in the traditional way to determine if the total phenolic content is responsible for their total antioxidant activity (Figure 3A). Additionally, the data were analyzed using the new approach proposed herein (Figure 3B). The regression line shown in Figure 3A corresponds to the linear relationship between the total antioxidant activity and the total phenolic content of the carrots before and after storage under air and hyperoxia conditions. The R2 value for this correlation is 0.75. Using this criterion of data analysis, the common conclusion would be that other compounds in addition to phenolics contributes to the total antioxidant capacity of carrots. However, as observed in Figure 3, it is evident that the different treatments are grouped in specific sections of the graph and also the simple regression line does not touch any of the data points representing the phenolic content and antioxidant activity of the carrots evaluated after storage.

Figure 3—.

Correlation between total phenolics and total antioxidant activity before and after storage of carrots under air and hyperoxia (80% O2+ 20% N2) conditions for 48 h at 20 °C. Traditional method to correlate both variables (A). Proposed new approach of data analysis (B). Each individual point represents a sample replicate.

An alternative method to analyze this graph is shown in Figure 3B. The method includes the segregation of the data into groups, their analysis of regression and the calculation of the slope from the regression lines. The slope indicates the specific antioxidant activity of the group. As mentioned previously, a high specific antioxidant activity value indicates that the phenolic profile from the sample is highly effective to neutralize free radicals. As observed in Figure 3B, 3 groups were identified, corresponding to each storage condition. The regression analysis was performed and their specific antioxidant activity values were obtained from the slopes (R2 > 0.96). The samples with the highest slope are the group of carrots stored under superatmospheric oxygen. Under this hyperoxia condition, the total phenolic and total antioxidant activity increased as well as the specific antioxidant activity indicating an increase on the effectiveness of the phenolic mixture to inhibit reactive species compared to carrots before storage. On the other hand, the total phenolics content and the total antioxidant activity for air-stored carrots, also increased and were higher than carrot samples before storage, however, the slope of the regression line was lower and thus the specific antioxidant activity decreased. This means that the mixtures of phenolic compounds that are accumulated during air storage are qualitatively and quantitatively different and less efficient in neutralizing free radicals than those phenolics present in carrots before storage.

To demonstrate that the phenolics in carrots are qualitatively and quantitatively affected by both storage conditions, the HPLC profiles were obtained (Figure 4). The phenolics present in carrots before storage were chlorogenic acid (CA), ferulic acid (FA), and 3,5-dicaffeoylquinic acid (3,5-diCQA). After storage, the concentration of the 3 compounds increased in the tissue. The relative molar proportion of CA and 3,5-diCQA appreciably differed between the 3 groups of carrots (Table 1).

Figure 4—.

HPLC chromatogram at 280 nm of phenolic compounds before (A) and after storage of carrots under air (B) and hyperoxia (80% O2+ 20% N2) conditions (C) for 48 h at 20 °C. Peak assignments: 1. chlorogenic acid, 2. ferulic acid, 3. 3,5-dicaffeoylquinic acid. Phenolic compounds were identified based on their PDA spectra, retention time, and comparison with authentic standards.

Table 1—.  Relative proportions of phenolics present in carrots before and after storage under air and hyperoxia (80% O2+ 20% N2) conditions for 48 h at 20 °C.
SampleMolar proportion of phenolic compoundsA (% of the total)Specific antioxidant activity (μmol Trolox equivalents/μmol phenolics)BStoichiometric number (n) mol hydroxyl groups (OH)/mol phenolicsC (CA + FA + 3,5-diCQA)
CAFA3,5-diCQA
  1. AMolar proportion based on chlorogenic acid equivalents, BValues represent the mean ± standard error, n = 4; phenolics based on CA equivalents. CBased on the molar proportion of phenolic compounds and stoichiometric factors of 7, 6, and 2 for 3,5-diCQA, CA, and FA, respectively. Values with similar letters within columns are not significantly different (LSD test, P > 0.05), n = 4. CA = chlorogenic acid; FA = ferulic acid; 3,5-diCQA = 3,5-dicaffeoylquinic acid.

Before storage84.91 ± 0.80 a7.36 ± 0.47 a7.73 ± 1.12 c7.46 ± 0.15 b5.75 ± 0.05 b
48h-Air70.65 ± 1.87 b8.78 ± 0.77 a20.57 ± 1.42 b6.50 ± 0.16 c5.85 ± 0.05 ab
48h-Hyperoxia62.86 ± 1.06 c8.43 ± 1.14 a28.71 ± 2.10 a9.81 ± 0.36 a5.94 ± 0.03 a

In general, a decrease in the percentage of CA and a proportional increase of 3,5-diCQA affected significantly (P < 0.05) the specific antioxidant activity of carrots. The molar proportion of FA remained constant (P > 0.05). To understand how the relative amounts of these phenolics interact to inhibit free radicals, we analyzed the theoretical stoichiometric number “n” of hydroxyl groups (μmol hydroxyl groups (OH)/μmol phenolics) present in each mixture of phenolics based in the presence of CA, FA, and 3,5-diCQA (Table 1). The stoichiometric analyses can relate how the structure of the phenolic compounds influences the availability of their OH groups to react with the peroxyl radical (ROO.) affecting its specific antioxidant activity. According to Bisby and others (2008) 1 molecule of OH can stabilize 1 molecule of ROO., therefore, it would be expected that the highest specific antioxidant activity is possessed by 3,5-diCQA, followed by CA and FA with theoretical stoichiometric factors of 7, 6, and 2, respectively.

This is confirmed with previous research showing that CA and FA have specific antioxidant activities (ORAC value/μmol of phenolic) of 5.7 and 4.47 μmol of Trolox equivalents/μmol of pure compound, respectively (Dávalos and others 2004). In addition, 3,5-diCQA is known to have a higher antioxidant activity on DPPH radicals compared to FA and CA (Ohnishi and others 1994; Kweon and others 2001). Thus, using the theoretical stoichiometric factors as references and the molar fraction of 3,5-diCQA, CA, and FA in the phenolic mixtures of the whole carrots evaluated before and after storage (Table 1), we obtained theoretical “n” values of approximately 5.7, 5.8, and 5.9 mol of hydroxyl groups (OH) available to react per mol of phenolic mixture present in each sample. These similar “n” values suggest that the 3 phenolic mixtures have similar capability of reaction with the peroxyl radicals.

Comparing such values with the experimental specific antioxidant activity allows us to elucidate the nature of the interactions among the 3 phenolic compounds detected (CA, FA, and 3,5-diCQA). For example, the 48 h air samples showed a decrease in the specific antioxidant activity compared to the before storage samples despite that the theoretical “n” values are similar for the phenolic mixtures. This suggests an antagonistic interaction between the 3 phenolic compounds. Peyrat-Maillard and others (2003) reported that mixtures of phenolic compounds may have antagonistic or synergistic interactions concerning total antioxidant activity measured against peroxyl radicals.

In a similar analysis for the specific antioxidant activity of 48 h hyperoxia samples, we observe that the proportion of FA remained constant with an increase in the proportion of 3,5-diCQA and a proportional decrease of CA resulting in an increase in the specific antioxidant activity compared to before storage samples. This increase in the specific antioxidant activity can be explained in terms of a synergistic effect between the 3 phenolic compounds, since the theoretical “n” value only increases slightly for the phenolic mixture. These results support the idea that the antioxidant capacity of a phenolic mixture is not just related with the phenolic profile or the pondered theoretical stoichiometric potential, but also with the interaction between the phenolics present.

We suggest that further studies be conducted with other substrates rich in nonphenolic acids (for example, flavonoids, proanthocyanidins) to confirm the concept and the stoichiometric approach used in the present study since the degree of oxidation of the heterocyclic ring of flavonoids largely influences the capacity of the molecule to be stabilized after proton donation and in polymeric phenolics, steric hindrance may affect the ability to react of the numerous OH groups present. In addition, we suggest further studies with in vivo systems to better understand phenolic interactions since the antioxidant activity in vitro does not necessarily predict the biological effectiveness of the extracts (Danesi and others 2008).

Conclusions

Since phenolic compounds have been established as the main contributors to the antioxidant activity of fruits and vegetables, a practical and useful approach to evaluate how the phenolic profiles from samples of different genotypes, maturity stages, and storage conditions affect the antioxidant activity is of major relevance. The proposed new approach presented in this study allows the determination of the specific antioxidant activity from the slope of simple correlations of total antioxidant activity against phenolic content, giving valuable information about the effectiveness of the mixture of phenolics to neutralize free radicals.

Acknowledgments

This study is based upon research supported by the Cooperative State Research, Education, and Extension Service, U.S. Dept. of Agriculture under Agreement Nr 2008-34402-19195, “Designing Foods for Health” through the Vegetable & Fruit Improvement Center, Texas AgriLife Research. The authors also acknowledge the scholarship from the Consejo Nacional de Ciencia y Tecnología (CONACYT, México).

Ancillary