Accepted Manuscript

Comprehensive evaluation of antioxidant activity: a chemometric approach us‐

ing principal component analysis

Dorina Casoni, Costel Sârbu

PII: S1386-1425(13)00968-2

DOI: http://dx.doi.org/10.1016/j.saa.2013.08.094

Reference: SAA 10953

To appear in: Spectrochimica Acta Part A: Molecular and Biomo‐

lecular Spectroscopy

Received Date: 22 May 2013

Revised Date: 27 July 2013

Accepted Date: 24 August 2013

Please cite this article as: D. Casoni, C. Sârbu, Comprehensive evaluation of antioxidant activity: a chemometric

approach using principal component analysis, Spectrochimica Acta Part A: Molecular and Biomolecular

Spectroscopy (2013), doi: http://dx.doi.org/10.1016/j.saa.2013.08.094

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers

we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and

review of the resulting proof before it is published in its final form. Please note that during the production process

errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 

 

Comprehensive evaluation of antioxidant activity: a chemometric approach using principal

component analysis

Dorina Casoni, Costel Sârbu*

Department of Chemistry, Faculty of Chemistry and Chemical Engineering, Babeş-Bolyai

University, Arany Janos Str., No. 11, RO-400028, Cluj-Napoca, România

Corresponding author (Costel Sârbu): Tel.: +40-264-93833; Fax: 140-264-590818

E-mail addresses: csarbu@chem.ubbcluj.ro (C. Sârbu)

casoni_dorina@yahoo.com (D. Casoni)

Abstract

A novel chemometric approach is described for evaluating the radical scavenging activity of

biogenic amine related compounds by using the 2,2-diphenyl-1-picrylhydrazyl (DPPH•)

procedure and principal component analysis (PCA) tool. By a comprehensive chemometric

investigation of variations in the radical scavenging profiles provided by the full-range UV-

Vis spectra for different test duration and different relative concentrations (different molar

ratio - [AH]/[DPPH•]) of the investigated compounds, new antioxidant activity evaluation

parameters were proposed. The new proposed parameters (PC1, mPC1, maxPC1) are in good

agreement with the reference DPPH• results (% RSA and IC50 derived from the reference

DPPH• test), obtained for the investigated amines and reference antioxidants. Much more, the

PCA profiles are better patterns for the comprehensive characterization of radical scavenging

2 

 

activity of compounds, allowing visualization of complex information by a simple graphical

representation and underlying the (dis)similarity of compounds related both to the reaction

kinetics and compounds concentration.

Keywords: Radical scavenging activity, Biogenic amine related drugs, UV-Vis spectra, PCA

profiles, Antioxidant activity

1. Introduction

The range of both naturally occurring and synthetic antioxidants applied in clinical settings

has appreciably expanded since aging and degenerative diseases were related to the oxidation of

biological components induced by reactive oxygen species (ROS). Also different experimental

methods have been developed to evaluate the antioxidant activity effectiveness and a lot of data

have been accumulated to describe kinetic process of radical-induced oxidation on biological

experimental materials [1]. Besides research under real circumstances including in vivo studies,

the in vitro convenient methods were also used to test the antioxidant effectiveness under

relatively simple and controlled circumstances [2] and the chemical principles for determining

antioxidant capacities were investigated [3].

In vitro antioxidant activity evaluation methods show extreme diversity, some of them

involving a distinct oxidation step followed by measurement of the outcome but in other

instances, there is no clear distinction between the various steps in the procedure [4]. Most of the

reported methods are based on measurements using individual procedures as accelerated stability

tests [5,6], peroxide value [7-9], diene conjugation [10,11], measurements of free radicals [12-

3 

 

16] or FRAP (ferric reducing antioxidant power) assay [17,18]. Tests including a large group of

antiradical activity determinations still arouse scientist’s interest as they are very useful even for

preliminary analysis. In this area, much effort have been devoted to the development of chemical

methods to evaluate antioxidant ability based on reaction with different radical species of

biological significance (such as O2•-, OH•, NO• or lipid peroxyl radicals) [2] but tests using

radical species such as 2,2’-azinobis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS•+), 2,2-

diphenyl-1-picrylhydrazyl (DPPH•) and N,N-dimethyl-p-phenylendiamine (DMPD) are also

quite popular [19,20]. These tests have been used for many decades to study the mechanism of

hydrogen-atom donation from certain substances or the antioxidant activity of compounds with –

SH, -OH and –NH groups [21].

Methods of expressing antioxidant activity results appear also to be as varied as the methods

of measurement [22]. The large diversity of radical scavenging assays that differ from each other

in terms of reaction conditions or in the form that results are expressed make difficult the

comparison of results. As a consequence, efforts have been made during the last years to

standardize analytical methods and provide valid guidelines that should be pursued by future

researchers. In this order, new methodologies and parameters were proposed and different

regression tools and multivariate mathematical approaches were applied to find insights into the

antiradical process and generate a complete antioxidant profile of different classes of compounds

[23-30]. Therefore new chemometric derived parameters for expressing results are trying to be

used which more or less serve the same purpose. Most of them form the basis of the newer test

methods such as the ABTS/TEAC (trolox equivalent antioxidant capacity) and different DPPH•

assays [31, 32]. While the spectrophotometric technique was intensively used in the DPPH•

scavenging assays, however, to our knowledge, no research has been reported as yet to

4 

 

implement this technique based through the chemometric analysis of the full-range spectrum

information (within the wavelength of 200–800 nm). Moreover, the short history of antioxidants

employed in the treatment of central nervous system (CNS) disorders [33] still receives

increasing attention. In this field efforts are likely to be directed at finding more powerful

methods to investigate the free-radical-scavenging profiles and antioxidant properties of existing

drugs in order to enhance their applicability as antioxidants [34] and improve their therapeutic

applications [35]. According to recent studies, some biogenic amines (important compounds

involved in usual metabolic processes and supporting the diagnosis of many diseases) and related

drugs proved to have effective in vitro antioxidant and radical scavenging activity [36-38] but the

complex profiles of these properties are still unclear.

In view of the above considerations, the present study aims to use the chemometric tools in

order to find new potential antioxidant activity evaluation parameters based on the analysis of the

information within the full-range UV-Vis spectra. The graphical approach to exploratory data

analysis is also described and illustrated with data obtained on complex antiradical profiles of

biogenic amines, related drugs and reference antioxidants.

2. Materials and methods

2.1 Reagents and solutions

All chemicals were of analytical reagent grade. Standard biogenic amines (tyramine, 3-

methoxytyramine hydrochloride, dopamine hydrochloride, (±)-octopamine hydrochloride, (-)-

norepinephrine, (-)-epinephrine, isoprenaline hydrochloride, (-)-3,4-dihydroxy-norephedrine

5 

 

(levonordefrin), 3',4'-dihydroxy-2-(methylamino)acetophenone hydrochloride (adrenalone),

3,4-dihydroxy-D-phenyl-alanine (D-dopa), S-(-)-carbidopa and D-tyrosine), phenolic

compounds (phenol, pyrocatechol (catechol), resorcinol, hydroquinone, pyrogallol, gallic acid

and caffeic acid), flavonols (kaempferol and quercetin) and some reference antioxidants

(ascorbic acid, (±)-6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (trolox) and (+)-

α-tocopherol) were purchased from Sigma-Aldrich (Steinheim, Germany). 2,2-diphenyl-1-

picrylhydrazyl free radical (95%) (DPPH•) was from Alfa Aesar (Karlsruhe, Germany).

Analytical-grade methanol was purchased from Chemical Company (Iaşi, Romania).

For the DPPH• assay a 0.15 mM solution was prepared by dissolving appropriate required

amount of DPPH• in methanol. This solution was prepared daily and protected from light

throughout the analysis time in order to minimize the loss of free radical activity. Stock

solutions of 1x10-1 M concentration were prepared for each of the investigated compounds by

dissolving appropriate required amount of standard in 100 mL methanol in all cases. Working

solutions of different concentrations (in range 2x10-5 M - 7x10-5 M) were prepared daily by

rigorous dilution of the stock solution in all cases.

2.2 Reference DPPH• method

To evaluate the new proposed methodology, the results obtained by the reference DPPH• test

[21] (with some minor modifications) were used. Also, for a comparative purpose, in the study,

we used a dataset of antioxidant activity evaluation parameters (% RSA and IC50 values for a set

of amines, related drugs and reference antioxidants), experimental determined by us and already

published [38].

6 

 

2.3 New DPPH• method

The effect of biogenic amines, related drugs and reference antioxidants on the DPPH• radical

absorbance profiles was monitored by recording full-range UV-Vis spectra (in the range 200-800

nm) for different relative concentrations (the molar ratio [AH]/[DPPH•] = 0.07 (C1); 0.10 (C2);

0.13 (C3); 0.17 (C4) and 0.23 (C5) mole/mole) by adding 1 mL standard solution prepared in

methanol to 2 mL of DPPH• methanolic solution (0.15 mM) in all cases. To monitor the reaction

kinetics, the spectra were also recorded in all cases at different time intervals (T1 – T7) (1, 5, 10,

15, 20, 25 and 30 minutes after the reaction was started) until the reaction reached an

equilibrium. Methanol was used as a blank solution and DPPH• solution (2 mL) with additional

methanol (1 mL) served as reference for the standard DPPH• spectra. All measurements were

performed in duplicate and the reactions were carried out at room temperature (22o C).

2.4 Spectra acquisition

A Jasco V-550 UV-VIS spectrophotometer with double beam system with single

monochromator (Tokyo, Japan), in absorbance mode, was used for spectra acquisition in the

range of 200–800 nm. The acquired spectra were stored after the smoothing process. The Spectra

Manager for Windows 95/NT version 1.53.04 (1995-2002, Jasco Corporation) software package

was used for the spectra acquisition control, smoothing process, storage and spectral data

digitization.

7 

 

2.5 Spectra processing and PCA analysis

Spectra processing are commonly used to improve the performance of the spectral data. In

our case the spectral processing Savitzky-Golay smoothing (SGS) algorithm, that fits a

polynomial equation to the data points, was implemented using the Spectra Analysis option

included in Spectra Manger for Windows 95/NT version 1.53.04 software package. In addition,

before the chemometric analysis, the original data set provided by spectra digitization was

standardized in relation to the DPPH• radical variables since the absorbance values for various

samples are dependent on the initial DPPH• solution absorbance. For this, the values for each

variable of recorded [AH]/[DPPH•] full-range UV-Vis spectra were divided by the variables

values provided from the DPPH• radical spectra in absence of any antioxidant in the reaction

cuvette. The mean of two identical measurements values corresponding to each of the spectrum

variables were subtracted using spectral data for various reaction periods and different

concentrations. As a consequence, complex data matrices (25 samples x 601 variables) were

obtained for each of the investigated concentrations, after the spectra digitization and

processing.

Among various computational chemometric methods, especially those classified as

multivariate exploratory techniques are used to extract systematic information often dispersed

over large sets of data. The principal component analysis (PCA), a linear dimensionality

reduction technique, which identifies orthogonal directions of maximum variance in the original

data, and projects the data into a lower-dimensionality space formed of a sub-set of the highest-

variance components, is the most preferred. This is because of its capacity to reduce the

dimensionality of the original dataset by retaining the maximum variability of a large number of

8 

 

variables (in terms of the variance-covariance structure) by few underlying factors (principal

components - PCs) without losing the important information. Upon application of PCA, the

number of variables in a data set is reduced by finding linear combinations of those variables

which explain most of the data variability [39]. In the PCA application, there are several ways to

preprocess a data set. The most popular (and perhaps the simplest) is unit variance scaling which

applies a linear transformation in which every variable in the data set is converted to a new one

with unit variance. Another technique used is mean centering, which involves defining a new

point of origin for the data given by the multivariate mean. It provides a convenient new point of

reference for the transformed variables from which further analyses can be carried out. In

addition, some software that has been developed, including our personal version, can perform the

PCA analysis using the initial data unchanged, without any scaling or centering preprocessing. In

all cases, the obtained PCs are useful tools for examining the relationships between compounds

and properties/descriptors, looking for similarities or sorting out the outliers.

Based on these advantages, some important antioxidant activity related information or new

antioxidant activity indices could be obtained by applying PCA directly to the data values

provided by digitization of full-range UV-Vis spectra obtained for the [AH]/[DPPH•] reactions.

In this order, new descriptive antioxidant activity parameters were obtained by applying the PCA

technique on the complex matrices of data obtained for different concentrations of compounds

and different periods of reaction with DPPH• and compared with classical ones.

The computation was performed by using Statistica 8.0 (StatSoft, Inc. 1984-2007, Tulsa,

USA) software package. The PCA analysis was performed using also Statistica 8.0 and our

personal software package developed manly for fuzzy methods.

9 

 

3. Results and discussion

Currently, efforts are likely to be directed at finding and standardize reference methods for

the investigation of the radical scavenging and antioxidant properties of the more powerful

antioxidants. As a consequence, the most recent studies proposed the involvement of different

chemometric tools for a complete antioxidant profile evaluation of various classes of

compounds [25, 27, 40]. According to our previous study [38], some important compounds (as

biogenic amines and their related drugs) involved in usual metabolic processes proved to have

effective in vitro radical scavenging activity. For a comprehensive evaluation of their radical

scavenging profiles, in the present study, a new DPPH• spectrophotometric method combined

with a chemometric approach was proposed and evaluated. While the reference DPPH•

technique employs a decrease of the radical absorbance proportional to the concentration and

antioxidant activity of the compound added, the research in the field generally used only the

maximum absorbance values (close to 516 nm in methanol) and did not considered an optimal

wavelength selection or the analysis of the full-range spectrum information. In this order, no

further analysis was carried out on which spectral wavelengths were relevant to the antioxidant

activity and are critical for its quantitative determination. For this reason, the proposed study

was conducted for the first time to select informative wavelength variables from the full-range

UV-Vis spectra in order to asses a comprehensive evaluation of the antioxidant profile of

compounds. Also, a number of thirty-five UV-Vis spectra were recorded for each of the

investigated amines, related drugs and also for some reference antioxidants taking in to account

different concentration compound/DPPH• and different reaction periods.

10 

 

3.1 PCA analysis of the UV-vis spectra

By a careful examination of the full-range UV-Vis spectra (Fig. 1) obtained in case of

reaction of biogenic amines, related drugs and respectively some reference antioxidants with

DPPH•, there may be appreciated that two specific ranges show variations in the absorbance

spectra profiles: the first one ((I) between 200-440 nm) with a maximum variation at 327 nm

and the second one ((II) between 440-800 nm) with a maximum variation at 516 nm. Also, for

a comprehensive investigation of the spectra profiles, different data processing methodologies

(using the scaling, centering and initial unchanged data) were applied in the PCA analysis of

the data set obtained by digitization of the full-range UV-Vis spectra (25 samples x 601

variables) and also on the data provided by the specified two ranges ((I) 25 samples x 241

variables and (II) 25 samples x 360 variables respectively).

The PCA analysis, representing the information contained in all variables as a new set of

coordinates called principal components (PC that are a linear combination of original

variables) showed that in all cases the first five PCs explain more than 99.99 % of the total

variance of the initial data. The most relevant results were obtained when the initial normalized

data, without any transformation, were used for the PCA analysis. In this case, the first

component (PC1) explained more than 98.70 % of the total variance in the data set while the

first two components (PC1 and PC2) explained more than 99.80 % of the total variance in all

cases. The best results (Table 1) were obtained when the 440-800 nm range spectra ((II) with a

maximum variation at 516 nm) was used. In this case, for certain of the investigated

concentrations (the first three concentrations), PC1 explained more than 99.66% of the total

variance in the data set while PC2 explained no more than 1.10% in all cases. The cumulative

11 

 

explained variance for the first two PCs was higher than 99.82% of the total variance in the

data set. In addition, the obtained results showed that the range 440-800 nm spectral profile

can provide additional information in order to assess a comprehensive evaluation of the radical

scavenging activity of amine related compounds and reference antioxidants.

According to the new PCs parameters (Table 2), the biogenic amines related drugs (D-dopa

(12), adrenalone (10) and carbidopa (13) are the most potent antioxidants and levonordefrin (9)

and 3-methoxytyramine (3) the less ones. These findings are supported by the IC50 values if

compounds with a very weak (octopamine (4)) or with no activity (phenylephrine (7)) in the

reasonable range of concentration are omitted from the discussion.

Also, the results for the investigated phenolic acids, flavonols and other reference

antioxidants, revealed that quercetin (22) and gallic acid (19) are the most powerful

antioxidants and kaempferol (21) and α-tocopherol (25) the less ones. The activity order of the

investigated antioxidants (quercetin (22) > gallic acid (19) > pyrogallol (18) > caffeic acid (20)

> catechol (15) > trolox (24) > ascorbic acid (23) > kaempferol (21) > tocopherol (25))

obtained by using the new proposed parameters (mPC1 and maxPC1), is almost the same as

reported by literature [38, 41]. By a comparative evaluation, the catecholamine related drug D-

dopa (12) is more active than quercetin (22) and adrenalone (10) and carbidopa (13) are more

active than gallic acid (19). The activities of biogenic amines neurotransmitters (dopamine (2),

norepinephrine (5), epinephrine (6) and their related drug isoprenaline (8) were comparable

with the activity of pyrogallol (18) and caffeic acid (20) respectively. All these findings are in

good agreement with the results obtained by reference DPPH• method [38].

12 

 

3.2 Evaluation of the new proposed radical scavenging parameters

Using the new PCs parameters as potential antioxidant activity descriptors, it was observed

that, in all cases, the scores corresponding to the firs principal component (PC1) are highly

correlated with the reference antioxidant activity evaluation parameters % RSA (r > 0.85) and

1/IC50 (r > 0.84) derived from the classical DPPH• test [38]. Much more, using mPC1 (mean of

scores corresponding to PC1) and maxPC1 (maximum of PC1 scores) (Table 2) as new

parameters derived from spectra corresponding to seven different reaction periods (1, 5, 10, 15,

20 25 and 30 minutes respectively), a strong statistical correlation was observed (Table 3) with

the reference antioxidant activity indices for all of the investigated concentrations. As it could

be observed, no significant differences were highlighted in the statistical correlations when the

second range (440-800 nm) or full-range UV-Vis spectra were used.

Based on the above findings the score plot profile variations obtained for the 440-800 nm

range spectra were investigated for the tested amines, related drugs and proposed reference

antioxidants. The new profiles described by the scores corresponding to PC1 (Fig. 2)

highlighted the possibility to simultaneously evaluate the radical scavenging activity

similarities among different compounds, different concentrations and also different reaction

time periods. From these representations some interesting observations can be made. Firstly,

the results indicate that the measured radical scavenging capacity and reaction kinetics of

different pure substances can depend significantly on the applied sample concentration. As a

second observation, the more antioxidant compounds (with the lowest IC50 values) have the

highest PC1 scores while the less active ones have the lowest PC1 scores. The first three

compound/DPPH• concentrations investigated seems to be more relevant for the

13 

 

comprehensive evaluation of the radical scavenging activity profiles of the investigated

compounds. In addition, the PCA profiles provide additional insights on the radical scavenging

activity characterization by allowing visualization of a complex dataset, related both to the

concentration and also the kinetic profiles of reaction with DPPH•, by a simple graphical

representation. As we can see from this pattern, the natural occurring catecholamine

(compounds 2, 5, 6) and their related drugs (compounds 10, 12, 13) show the same kinetic

reaction profiles as the strong reference antioxidants gallic acid (19) and quercetin (22) while

the profiles of synthetic catecholamine isoprenaline (8) is more similar with that of other

antioxidants as caffeic acid (20), ascorbic acid (23), trolox (24) or α-tocopherol (25). Moreover

the maxPC1 parameters showed an increasing of radical scavenging activity with the increase

of compounds concentration, generally for the phenols and some common reference

antioxidants (ascorbic acid, caffeic acid, kaempherol, trolox, α-tocopherol) and a different

behavior for the investigated amines (with the exception of dopamine) and the most powerful

antioxidants (quercetin and gallic acid) (Fig. 3).

Using the new mPC1 values representation for the first three compound/DPPH•

concentrations investigated (Fig. 4), a pertinent classification of the investigated compounds, in

accord with the previous findings, was made based on similarities on their radical scavenging

activity and reaction kinetics with DPPH•.

4. Conclusion

A chemometric approach for evaluating the radical scavenging activity of biogenic amines,

related drugs and reference antioxidants based on an advanced DPPH• assay was proposed. On

14 

 

the basis of the obtained results the potential of the spectrophotometric method and multivariate

data analysis to allow an easy interpretation of similarity and differences in radical scavenging

activity of compounds was assessed. It was found that the PCA analysis of the digitized UV-

Vis radical scavenging profiles leads to a good agreement between activity of the investigated

compounds determined by the new proposed methodology and results provided by the

reference experimental method. In addition, the new proposed approach allows a

comprehensive evaluation of the variations in the antiradical profile of compounds and their

reaction kinetics in the same time.

By supporting all the above mentioned statements related to the radical scavenging

activity of amines and reference antioxidants, the new applied methodology could be a possible

powerful tool in the evaluation of the antioxidant activity of compounds by revealing new

insights in this field. Besides of the advantage of simplicity and easy to perform, the application

of the PCA analysis directly on the experimental data values provided by UV-Vis spectra, some

of the processing data errors (as, for example, the extrapolation error in case of IC50

determination) are eliminated.

Acknowledgment

This work was possible with the financial support of the Sectorial Operational Programme for

Human Resources Development 2007–2013, co-financed by the European Social Fund, under

the project number POSDRU 89/1.5/S/60189 with the title “Postdoctoral Programs for

Sustainable Development in a Knowledge Based Society” and PN-II-ID-PCE-2011-3-0366

project.

15 

 

References

[1] N. Etsuo, Free Rad. Biol. Med. 49 (2010) 503–515.

[2] L. Zai-Qun, Chem. Rev. 110 (2010) 5675–5691.

[3] D. Huang, B. Ou, R.L. Prior, J. Agric. Food Chem. 53 (2005) 1841-1856.

[4] M. Antolovich, P.D. Prenzler, E. Patsalides, S. McDonald, K. Robards, Analyst 127 (2002)

183–198.

[5] E.N. Frankel, Trends Food Sci. Technol. 4 (1993) 220-225.

[6] C. Liang, K. Schwarzer, J. Am. Oil Chem. Soc. 75 (1998) 1441–1443.

[7] F.R. Van de Voort, A.A. Ismail, J. Sedman, J. Dubois, T. Nicodemo, J. Am. Oil Chem. Soc.

71 (1994) 921-926.

[8] E.M. Marinova, N. Yanishlieva, Food Chem. 56 (1996) 139-145.

[9] J.A. Larrauri, P. Ruperez, F.S. Calixto, J. Agric. Food Chem. 45 (1997) 4028-4031.

[10] I.M. Heinonen, P.J. Lehtonen, A.I. Hopia, J. Agric. Food Chem. 46 (1998) 25-31.

[11] Y. Miyake, K. Yamamoto, Y. Morimitsu, T. Osawa, J. Agric. Food Chem. 45(1997) 4619-

1623.

[12] O.I. Aruoma, J.P.E. Spencer, D. Warren, P. Jenner, J. Butler, B. Halliwell, Food Chem. 60

(1997) 149-156.

[13] Y.H. Chu, C.L. Chang, H.F. Hsu, J. Sci. Food Agric. 80 (2000) 561-566.

[14] D.D. Beer, E. Joubert, W.C.A. Gelderblom, M. Manley, J. Agric Food Chem. 51 (2003)

902-909.

[15] O. Erel, Clin. Biochem. 37 (2004) 277-285.

16 

 

[16] M. Philpott, K.S. Gould, C. Lim, L.R. Ferguson, J. Agric. Food Chem. 52 (2004) 1511-

1513.

[17] P.T. Gardner, T.A.C. White, D.B. McPhail, G.G. Duthie, Food Chem. 68 (2000) 471-474.

[18] I.F.F. Benzie, J.J. Strain, Anal. Biochem. 239 (1996) 70–76.

[19] D. Villaño, M.S. Fernández-Pachón, M.L. Moyá, A.M. Troncoso, M.C. García-Parrilla,

Talanta 71 (2007) 230–235.

[20] A. Tai, T. Sawano, F. Yazama, H. Ito, Biochim. Biophys. Acta 1810 (2011) 170–177.

[21] N. Nenadis, M. Tsimidou, J. Am. Oil Chem. Soc. 79 (2002) 1191-1195.

[22] K. Robards, P.D. Prenzler, G. Tucker, P. Swatsitang, W. Glover, Food Chem. 66 (1999)

401-436.

[23] F.C. Lavarda, Int. J. Quantum Chem. 95 (2003) 219–223.

[24] K.C. Weber, K.M. Honόrio, S.L. da Silva, R. Mercadante, A.B.F. da Silva, Int. J. Quantum

Chem. 103 (2005) 731–737.

[25] M. Hossain, A. Patras, C. Barry-Ryan, A. Martin-Diana, N. Brunton, J. Funct. Food 3

(2011) 179-189.

[26 ] A. Pérez-Garrido, A.M. Helguera, J.M. Morillas Ruiz, P.Z. Rentero, Eur. J. Med. Chem. 49

(2012) 86-94.

[27] Q. Chen, Z. Guo, J. Zhao, Q. Ouyang, J. Pharm. Biomed. Anal. 60 (2012) 92– 97.

[28] D. Wu, J. Chen, B. Lu, L. Xiong, Y. He, Y. Zhang, Food Chem. 135 (2012) 2147–2156.

[29] T. Wang, R. Jonsdottir, G. Olafsdottir, Food Chem. 116 (2009) 240–248.

[30] M. Cam, H. Yasar, D. Gokhan, Food Chem. 112 (2009) 721–726.

[31] N. Nenadis, L.-F. Wang, M. Tsimidou, H.–Y. Zhang, J. Agric. Food Chem. 52 (2004) 4669-

4674.

17 

 

[32] A. Om, J.H. Kim, J. Med. Food 11 (2008) 29–37.

[33] A.N. Erin, N.V. Gulyaeva, E.V. Nikushkin, Bull. Exp. Biol. Med. 118 (1994) 1045-1049.

[34] S. Betigeri, A. Thakur, K. Raghavan, Pharm. Res. 22 (2005) 310-317.

[35] J.G. Baker, S.J. Hill, R.J. Summers, Trends Pharmacol. Sci. 32 (2011) 227-234.

[36] İ. Gülçin, Chem. Biol. Interact. 179 (2009) 71-80.

[37] T. Kawashima, K. Ohkubo, S. Fukuzumi, J. Phys. Chem. B 14 (2010) 675-680.

[38] C. Sârbu, D. Casoni, Cent. Eur. J. Chem. 11 (2013) 679-688.

[39] A. Nasal, A. Bucinski, L. Bober, R. Kaliszan, Int. J. Pharm. 159 (1997) 43–55.

[40] A.C. Cabrera, J.M. Prieto, Food Chem. 118 (2010) 141–14.

[41] J. Tabart, C. Kevers, J. Pincemail, J.-O. Defraigne, J. Dommes, Food Chem. 113 (2009)

1226-1233.

18 

 

Figure Captions

Fig. 1. The spectra profiles corresponding to the reactions of investigated amines and reference

antioxidants (4 x 10-5 M, solutions in methanol) with DPPH• (0.15 x 10-3 M, solution in

methanol) after 30 minutes reaction period.

Fig. 2. Profiles of the PC1 parameters of the investigated amines, related drugs and reference

antioxidants for different compound/DPPH• concentrations (C1: 0.07 mole/mole; C2: 0.10

mole/mole; C3: 0.13 mole/mole; C4: 0.17 mole/mole; C5: 0.23 mole/mole) and different

reaction periods.

Fig. 3. Profiles of the maxPC1 parameters of the investigated amines, related drugs and some

reference antioxidants against the compound/DPPH• concentrations.

Fig. 4. Clustering of the investigated compounds according to their similarities in radical

scavenging activity against DPPH• and using the new proposed mPC1 evaluation parameter.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

1

Table 1

Eigenvalues of covariance matrix and related statistics (λ = 440-800 nm).

Conc/PCs Eigenvalue % Total Variance Cumulative (%)

C1

PC1 9.2340 99.740 99.7384

PC2 0.0170 0.180 99.9217

PC3 0.0056 0.060 99.9820

PC4 0.0010 0.010 99.9927

PC5 0.0003 0.003 99.9955

C2

PC1 18.630 99.670 99.6690

PC2 0.0480 0.260 99.9272

PC3 0.0120 0.060 99.9894

PC4 0.0010 0.006 99.9955

PC5 0.0003 0.002 99.9974

C3

PC1 28.5108 99.687 99.6866

PC2 0.0549 0.192 99.8785

PC3 0.0314 0.110 99.9882

PC4 0.0021 0.007 99.9957

PC5 0.0004 0.002 99.9973

C4

PC1 34.6181 98.872 98.8718

PC2 0.3365 0.961 99.8327

PC3 0.0481 0.137 99.9701

PC4 0.0071 0.020 99.9903

PC5 0.0022 0.006 99.9966

C5

PC1 40.8328 98.722 98.7216

PC2 0.4548 1.099 99.8214

PC3 0.0611 0.148 99.9688

PC4 0.0085 0.021 99.9895

PC5 0.0031 0.008 99.9971

1

Table 2

The antioxidant activity indices (IC50) and the new PCs parameters of the investigated compounds for different compound/DPPH•

concentrations (C1: 0.07 mole/mole; C2: 0.10 mole/mole; C3: 0.13 mole/mole; C4: 0.17 mole/mole; C5: 0.23 mole/mole).

Name of Compound [38]

IC50 (μ M) New antioxidant activity parameters

C1 C2 C3 C4 C5

Amines mPC1 max PC1 mPC1 max PC1 mPC1 max PC1 mPC1 max PC1 mPC1 max PC1

1 Tyramine - -14.95 -14.44 -13.26 -12.37 -11.45 -10.11 -8.11 -6.52 -3.91 -1.26

2 Dopamine 10.50±0.51 -12.69 -12.02 -9.72 -8.87 -7.79 -6.36 -5.81 -3.72 -3.35 -0.99

3 3-Methoxytyramine 63.16±0.91 -16.97 -16.25 -16.48 -15.52 -15.93 -14.90 -16.15 -15.11 -15.34 -13.91

4 Octopamine 455.04±2.26 -18.70 -18.59 -18.75 -18.64 -18.47 -18.41 -18.17 -18.12 -18.27 -18.17

5 Norepinephrine 10.87±0.12 -12.93 -12.70 -10.76 -9.81 -8.93 -7.27 -7.65 -5.42 -6.68 -3.90

6 Epinephrine 11.27±0.01 -13.53 -13.27 -10.97 -10.35 -9.08 -7.72 -6.31 -3.70 -6.61 -4.37

7 Phenylephrine > 1000 -18.62 -18.07 -18.25 -18.22 -18.91 -18.82 -18.41 -18.37 -18.51 -18.49

8 Isoprenaline 11.17±0.01 -13.21 -13.17 -10.14 -10.09 -7.06 -6.96 -4.18 -3.62 -2.48 -1.55

9 Levonordefrin 35.18±0.99 -16.62 -15.33 -16.04 -14.45 -15.66 -13.87 -15.23 -13.23 -14.80 -11.98

10 Adrenalone 6.58±0.16 -10.24 -9.44 -6.98 -5.77 -4.47 -2.86 -3.03 -1.83 -2.24 -1.22

11 D-Tyrosine 44.91±0.62 -18.88 -18.84 -18.79 -18.73 -18.34 -18.30 -18.40 -18.32 -18.23 -18.18

12 D-Dopa 5.67±0.01 -7.92 -7.04 -3.70 -2.02 -3.26 -0.74 -4.15 -0.63 -5.07 -1.04

13 S-Carbidopa 7.90±0.16 -11.61 -10.99 -7.73 -6.61 -4.23 -2.79 -2.18 -1.80 -1.82 -1.77

Phenols

14 Phenol > 1000 -18.43 -18.37 -18.64 -18.58 -19.11 -19.06 -18.67 -18.63 -18.67 -18.63

15 Catechol 13.92±0.05 -14.95 -14.44 -13.26 -12.37 -11.44 -10.11 -8.11 -6.52 -3.91 -1.26

16 Resorcinol > 1000 -18.65 -18.62 -18.31 -18.29 -19.02 -18.95 -17.91 -17.88 -17.26 -17.16

17 Hydroquinone 18.42±0.16 -16.02 -15.77 -13.22 -12.79 -13.59 -13.02 -11.68 -10.76 -7.77 -6.26

18 Pyrogallol 10.43±0.05 -12.90 -12.67 -10.61 -10.02 -7.77 -6.96 -5.52 -4.42 -2.23 -1.93

Phenolic acids

19 Gallic Acid 8.15±0.07 -11.20 -10.80 -8.01 -7.49 -5.16 -4.33 -3.18 -2.22 -1.43 -0.95

20 Caffeic Acid 11.99±0.06 -13.68 -13.49 -11.41 -11.13 -8.94 -8.51 -6.04 -5.52 -2.15 -1.57

Flavonols

21 Kaempferol 22.30±0.20 -16.36 -16.33 -14.63 -14.61 -13.59 -13.57 -12.00 -11.98 -8.99 -8.96

22 Quercetin 5.96±0.01 -9.85 -8.40 -6.01 -3.94 -2.54 -1.28 -0.81 -0.63 -0.68 -0.64

Antioxidants of references

23 Ascorbic Acid 20.08±0.03 -16.21 -16.15 -14.69 -14.60 -12.71 -12.69 -10.64 -10.60 -7.79 -7.77

24 Trolox 19.62±0.11 -15.73 -15.65 -14.09 -14.00 -13.08 -13.06 -10.96 -10.92 -7.77 -7.57

25 α-Tocopherol 27.89±0.06 -16.40 -16.36 -15.47 -15.43 -14.32 -14.23 -13.44 -13.32 -10.96 -10.79

1

Table 3

Correlations of the new antioxidant activity parameters with the indices derived from the reference DPPH• method for different relative

concentrations (C1: 0.07 mole/mole; C2: 0.10 mole/mole; C3: 0.13 mole/mole; C4: 0.17 mole/mole; C5: 0.23 mole/mole).

Antioxidant

activity

parameters

Full-range UV-Vis spectra

C1 C2 C3 C4 C5

mPC1 minPC1 maxPC1 mPC1 minPC1 maxPC1 mPC1 minPC1 maxPC1 mPC1 minPC1 maxPC1 mPC1 minPC1 maxPC1

C1 0.99 0.94 1.00 0.98 0.91 0.99 0.96 0.82 0.98 0.90 0.70 0.93 0.78 0.59 0.82

%RSA C3 0.99 0.97 0.98 0.99 0.96 0.99 0.99 0.90 1.00 0.96 0.80 0.98 0.87 0.71 0.90

C5 0.87 0.91 0.84 0.89 0.93 0.87 0.92 0.89 0.91 0.95 0.84 0.95 0.98 0.82 1.00

1/IC50 0.98 0.95 0.97 0.98 0.94 0.98 0.97 0.86 0.97 0.92 0.76 0.94 0.82 0.66 0.84

The 440-800 nm range of the spectra

%RSA

C1 0.99 0.94 1.00 0.98 0.90 0.99 0.96 0.81 0.98 0.89 0.69 0.92 0.77 0.58 0.82

C3 0.99 0.97 0.98 0.99 0.96 0.99 0.99 0.89 1.00 0.96 0.79 0.97 0.87 0.70 0.89

C5 0.87 0.81 0.84 0.89 0.93 0.86 0.91 0.88 0.90 0.95 0.84 0.96 0.98 0.82 1.00

1/IC50 0.98 0.96 0.98 0.98 0.93 0.98 0.97 0.85 0.98 0.92 0.74 0.93 0.82 0.64 0.84

Highlights

►A novel DPPH•

assay combined with a chemometric approach was proposed.

►Relevant wavelengths for the radical scavenging activity evaluation were selected.

►New potential antioxidant activity indices were proposed and evaluated.

►The radical scavenging profile of biogenic amines and related drugs was evaluated.

►The study offers the possibility of significant reduction of experimentally work.

.