Phytochemical profiles and classification of Agave syrups using 1H‐NMR and chemometrics

Abstract Background Agave syrups are natural sweeteners that are highly desirable for human consumption because they have low glycemic index. In this work, we explored the potential of 1H‐NMR‐Chemometrics as a useful tool in the identification and differentiation of Agave syrups. Also, we evaluated the phytochemical screening and antioxidant capacity of Agave syrup compared to other natural sweeteners. Results The phytochemical screening stands out for Agave syrups containing higher levels of metabolites with antioxidant activity, mainly saponins, glycosides, and terpenoids. Agave syrup antioxidant activity was in a range from 10% to 53%, while the total phenolic content was from 24 to 300 EAG/100 g, and condensed tannins were between 240 and 1,900 mg CE/g. Additionally, 1H‐NMR spectroscopy was used to characterize syrup profiles and chemometrics. PCA group analyses allowed the sweeteners’ classification by origin and kind of Agave. Conclusion Thus, we conclude that 1H‐NMR and chemometrics can be used for identifying, differentiating, and classifying Agave syrups. Besides, Agave syrups contain significant amounts of antioxidative components and can be considered as an effective source of antioxidant.

This carbohydrate composition gives Agave syrups a low glycemic index and makes it sweeter than other syrups containing appreciable levels of glucose and/or sucrose, such as corn and sugarcane (Willems & Low, 2012). Besides fructose and glucose, some fructooligosaccharides (FOS) are also present in some Agave syrups in very smaller amounts as result of incomplete agavin hydrolysis (Mellado-Mojica, Seeram, & López, 2016).
Nowadays, carbohydrate fingerprinting is a very useful molecular marker of authenticity, adulterants detection, quality, and origin of natural sweeteners; therefore, the determination of glucose, fructose, and sucrose contents, and the oligosaccharides profiles became a method for the determination of quality in honey and syrups (Bueno et al., 2015;Peshev & Van den Ende, 2014;Rizelio et al., 2012).
Due to their relative newness in the market, it is of great importance to study the chemical and phytochemical composition of Agave syrups in order to establish their phytochemical composition and metabolite richness, as well as the development of new strategies for the differentiation of Agave syrups from other natural sweeteners. In addition, the knowledge on their phytochemical composition might be of great importance to human health issues.
In natural sweeteners, phenolic compounds, flavonoids, and carotenoids are reported to be responsible for the antioxidant activity in honeybee and maple syrup (Phillips, Carlsen, & Blomhoff, 2009).
In Agave syrups, there is only one report on their antioxidant activity; however, the above investigation was limited to four Agave syrups samples, where three of them were artisanal samples (Olvera, Cardador, & Martín, 2014). Therefore, exploring the potential of commercial Agave syrups as new antioxidant food supplements still needs to be undertaken. Saponins, terpenoids, flavonoids, and tannins have been identified in Agave sisalana, Agave impressa, Agave ortnithobroma, A. tequilana, Agave angustifolia, and Agave americana species (Ahumada et al., 2013;Dias, Sales, Weingart, & Zarur, 2013;Hamissa et al., 2012) and all or some of these phytochemicals might also be present in Agave syrups.
Nuclear magnetic resonance (NMR) spectroscopy has been widely applied to identify compounds in a wide diversity of food samples such fruit juices, wines, and honeys because it is nondestructive, selective, and capable of detecting a great number components in complex mixtures (Consonni, Cagliani, & Cogliati, 2012;Kosir & Kidric, 2001;Vlahov, Chepkwony, & Ndalut, 2002). Italian honeys were successfully classified based on their NMR spectra and multivariate statistical analysis (Beretta, Caneva, Regazzoni, Bakhtyari, & Facino, 2008;Lolli, Bertelli, Plessi, Sabatini, & Restani, 2008;Schievano, Peggion, & Mammi, 2010). It was suggested that the sugar profile could be used to characterize particular sweeteners used in some samples. NMR spectroscopy coupled to principal component analysis (PCA) can therefore be applied to construct an "identity card" of saccharides for each floral source (Belton et al., 1998;Beretta, Granata, Ferrero, Orioli, & Facino, 2005;Bertelli et al., 2010). The relevance of PCA is mainly to reduce the original data sets to a smaller number of independent variables. PCA can identify trends or characteristics within the NMR data (Beretta et al., 2005;Bertelli et al., 2010).
The aim of this work was to establish the phytochemical screening, total phenolic, and tannin contents, as well as the antioxidant activity of Agave syrups. In addition, 1 H-NMR-PCA was used to differentiate Agave syrups among other natural sweeteners.

| Syrupextracts
Extracts from each natural sweetener were prepared for phytochemical screening analysis, antioxidant activity, total phenol contents, and condensed tannins (protoantocyanidins) determination.
Syrup extracts were obtained according to the protocol described by Chaikham and Prangthip (2014). Briefly, 200 mg of each syrup were placed in a 1.5-ml Eppendorf tube along with 2 ml of distilled water and 9 ml of absolute ethanol. Samples were mixed for 5 min and then centrifuged at 22,000g, for 10 min. Supernatants were separated and collected in a new sample reservoir. Extracts were stored in darkness until their analyses. | 5 VELÁZQUEZ RÍOS Et aL.

| Testforsaponins
To 2 ml of syrup extract, 2 ml of distilled water was added and shaken in a graduated cylinder for 15 min length wise. Formation of a 1-cm layer of foam indicates the presence of saponins.

| Testforflavonoids
To 2 ml of syrup extract, 1 ml of 2 N sodium hydroxide was added.
Presence of yellow color indicates the presence of flavonoids.

| Testforquinones
To 1 ml of extract, 1 ml of concentrated sulfuric acid was added.
Formation of red color indicates the presence of quinones.

| Testforglycosides
To 2 ml of extract, 3 ml of choloroform and 10% ammonia solution were added. Formation of pink color indicates the presence of glycosides.

| Testforcardiacglycosides
To 0.5 ml of extract, 2 ml of glacial acetic acid and few drops of 5% ferric chloride were added. This was under layered with 1 ml of concentrated sulfuric acid. Formation of a brown ring at the interface indicates the presence of cardiac glycosides.

| Testforterpenoids
To 0.5 ml of extract, 2 ml of chloroform were added and concentrated sulfuric acid was added carefully. Formation of red brown color at the interface indicates the presence of terpenoids.

| Testforcoumarins
To 1 ml of extract, 1 ml of 10% NaOH was added. Formation of yellow color indicates the presence of coumarins.

| Syrups'color
The syrups' color designation was determined according to the United States Standards for Grades of Extracted Honey approved color Pfund scale (USDA, 1985). The sweeteners' samples were diluted in water in 50% p/v ratio, then heated in a water bath at 50°C. Subsequently, they were centrifuged at 22,000g for 3 min to precipitate particles. Absorbance was measured on a BIORAD Mark 10360 microplate reader at λ = 635 nm. The results were expressed according to the Pfund scale (water white, white extra, white, extra light amber, light amber, amber, and dark amber) according to the following formula:

| AntioxidantactivityandIC50from naturalsweeteners
Antioxidant activity of the natural sweetener was determined according to the α-diphenylβ-picrylhydrazyl (DPPH) free radical scavenging method. The DPPH reagent (2,2-diphenyl-1-picrylhydrazyl) was prepared at 150 μM concentration. One hundred and fifty miroliters of each extract were added to 150 μl of DPPH solution placed in a 96-well microplate. Absolute ethanol was used as a blank. It was subsequently incubated at room temperature in the absence of light and the absorbance was measured on a BIORAD Mark 10360 microplate reader at λ = 517 nm. Readings were taken at 0, 30, and 60 min (Canadanovic et al., 2014).

The antioxidant activity was expressed as:
The IC50 is the inhibitory activity expressed by the amount of sample per milliliter (g/ml) of solution that achieves 50% inhibition of DPPH free radicals. Three different concentrations 0.1, 0.2, and 0.5 g/ml of syrup from ethanol-water solution 82% and 18%, respectively, of each extract were prepared; 0.1 ml of each solution was mixed with 3.9 ml of 0.1 mM DPPH. Immediately, zero time and after 30 min reading were made, at λ = 515 nm.

| Totalphenolcontentsinsweeteners
The total phenol content was determined according to the methodology reported by Pelvan, Alasalvar, and Uzman (2012)  were expressed as milligrams of gallic acid equivalents per gram of extract (mg of GAE/g of extract).

| Condensedtanninsinnaturalsweeteners
The condensed tannin contents were estimated according to a modified protocol employed by Tili et al. (2015). Ten microliters of ethanolic extract were mixed with 197 μl of 4% ethanol-vanillin solution and 99 μl of concentrated sulfuric acid incubated at room temperature and subsequently the absorbance was measured at λ = 490 nm.
Condensed tannins were expressed in milligrams of catechin equivalents per gram of sample (EC)/g.

| 1 H-NMRspectroscopy-PCAfrom naturalsweeteners
Nuclear magnetic resonance spectra of syrup samples were acquired with a Varian/Agilent 600 MHz AR Premium COMPACTTM spectrophotometer; all NMR experiments were performed at 300 K.
The 1 H-NMR spectra were measured at 300 K and 599.77 MHz frequency using D 2 O as solvent and as an internal reference; the residual HOD signal was used at 4.9 ppm. The pulse of π/2 employed was 8.7 μs, relaxation time of 15 s with 16 repetitions. Once the data were collected, the main components were analyzed in the statistical software STATGRAPHICS Plus 5.1, analyzing sweetener and later all the data together.

| Statisticaldataanalysis
Analysis of variance was realized in the statistical software Infostat, with a confidence level of 95%. The means comparison was made with the Mean Significant Difference (LSD Fisher) method with α = 0.05, with a confidence level of 95% (p < 0.05). Each test was performed in triplicate.

| Phytochemicalscreening
Phytochemical compounds such as saponins, flavonoids, quinones, glycosides, cardiac glycosides, terpenoids, and coumarins are known as nutraceutical compounds due to their medicinal importance.   The above findings showed the potential of Agave syrups as a new phytochemical and natural source of these valuable compounds.
The presence of natural products in the four different sweetener groups was identified, and in some cases, higher presence of these natural products was observed than in others. In general, the samples presented positive results for each of the seven compound groups analyzed in this work, except for two samples of ATS (ATS1 and ATS2) and two samples of honey (HB4 and HB5) that gave negative results for flavonoids and terpenoids (  (Table 1). Nevertheless, CS was unique by the high content of cardiac glycosides observed.
In general, the four different sweeteners presented potential as source of natural components. Agave syrups might be highlighted due to the greatest amounts of saponins, cardiac glycosides, and terpenoids with respect to other evaluated sweeteners.
Finally, several natural products have been found in Agave species, for example, the presence of saponins with a content from 1.17 to 100 g in samples of Agave obtained from Agave atrovirens during fermentation (Can et al., 2015).

| Color
The color in syrups is partly related to their content of phytochemicals with antioxidant activity such carotenoids and flavonoids (Chaikham & Prangthip, 2014). Syrups were diluted to 50% (w/v) with distilled water and their absorbance was determined. The absorbance of the used dilution presented values from 164 to 16 mm/Pfund. A. salmiana syrups being the sweeteners with greater color intensity were classified as "Dark amber" and the corn syrups with lowest intensity were classified as "Extra white" ( Table 2). The colors in all these natural sweeteners were in accordance with data described by Mellado-

| AntioxidantactivityandIC50
The antioxidant activity of natural sweeteners was determined as the percent of free radical scavenging compared to the free radical scavenging of the DPPH reagent. The antioxidant activity of natural sweeteners was in average as described below: ATS 23.56%, ASS syr-  (Table 3).
Agave tequilana syrups showed great variability with respect to this parameter; the above could be probably due to the botanical origin of the sweeteners and processing, and also due to problems of product standardization, meaning quality. However, ATS as well as ASS may be good antioxidant syrup candidates.
The IC50 in general agreed with the results obtained on the antioxidant activity determination by the percentage of free radical scavenging DPPH since the sample with the highest antioxidant activity, ATS-16 with 52.20% free radical uptake, was the sweetener needing less sample amount to reduce free radicals DPPH to 50% (0.199 g) according to the results obtained in the IC50. On the other hand, CS-2 sample had the lowest antioxidant activity (8.71%) and therefore, 30.82 g syrup is required to reduce free radicals DPPH to 50%. This proportion is conserved for most of the samples analyzed (Table 3).
In honey bees, the antioxidant capacity was attributed to the combined activity of a wide range of natural compounds such phenolics, organic acids, Maillard reaction products, and probably other minor components (Daglia, 2012). Thereby, the antioxidant activity of Agave syrups could be due to similar natural compounds, so it would be of great importance to characterize and identify these phytochemical compounds in these new sweeteners.

| Totalphenols
The natural sweeteners displaying the darkest color also showed the highest total phenolic compound amounts as well as major antioxidant capacity, A. salmiana and sugarcane syrups (253.30 and 185.22 EAG/100 g, respectively) ( Table 3). These results suggest that phenolic compounds in these natural sweeteners might be responsible for the antioxidant activity observed (Table 3).
In some studies where the antioxidant activity and phenolic content were determined, such as berries, fruit wines, and liqueurs, no relationship was found between the two parameters (Ruiz et al., 2012). This information agrees with our results; we did not find statistically relationship (data no show) between phenolic and antioxidant activity. On the other hand, it is very useful to associate individual phenolic compounds with antioxidant activity because their particular structural characteristics are able to neutralize free radicals more easily (Ruiz et al., 2012).

| Determinationofcondensedtannins
Proanthocyanidins are compounds composed of flavyl units, containing carbohydrates and amino acid residues, and also having different degrees of condensation and called condensed tannins.  López (2013, 2015) reported the physicochemical properties and carbohydrate profiles of Agave syrups. The authors describe that glucose, fructose, and sucrose amounts were the main differences between syrups from different Agave species. They also mentioned that oligosaccharide type and profile play an important role in the differentiating A. tequilana and A. salmiana syrups. TA B L E 3 Antioxidant activity, total phenol content, condensed tannin content, and IC50 of natural sweeteners

| 1 H-NMRspectroscopy-PCAof naturalsweeteners
In this work, we analyzed the 1 H-NMR spectra of the Agave syrups and other natural sweeteners (Figure 1). 1 H-NMR spectra reveal differences among natural sweeteners; however, some signals are also common ( Figure 1) It is known that sugars are the main component of honey and syrups and the possibility to analyze these components would help establish the qualitative characteristics and authenticity of these products (Consonni, Cagliani, & Cogliati, 2013;Kortesniemi et al., 2016). The PCA 1-3 plot (Figure 3b) shows a clearer grouping of these natural sweeteners depending on their origin, differentiating even among Agave species, demonstrating the potential of 1 H-NMR for a good differentiation and classification of foods. The global sugar composition in reference to other natural sweeteners played an important role in the discrimination (Kortesniemi et al., 2016;Lolli et al., 2008).
The 1 H-NMR spectra coupled with principal component analyses reveals that there is a difference in carbohydrate profiles between Agave syrups relative to other sweeteners; there are even differences among Agave syrups according to the species. showing a difference among these two species; it was observed that Agave syrups differ in composition and carbohydrate content.

| CON CLUS ION
Agave syrups showed a greater phytochemical potential than other sweeteners due to the presence of more natural compounds with antioxidant activity. They were also different in carbohydrate profile with respect to other sweeteners, even in species of Agave. The differences in the chemical composition also occur within the same samples of A. tequilana syrups; this difference is due to different times used in the cooking process of the Agave.
Among the Agave syrups, ASS showed the highest phytochemical potential due to their higher antioxidant activity, higher content of total phenols, and proanthocyanidins compared to ATS. In addition, ASS also showed greater homogeneity in the color of the samples.
The color of the sweeteners was related to the content of pigments with antioxidant activity since darker sweeteners had higher antioxidant activity, content of phenols, and proanthocyanidins.
The 1 H-NMR spectra of Agave syrups and comparative sweeteners were obtained. 1 H-NMR data coupled to multivariate methods such as principal component analysis (PCA) allowed the identification and classification of Agave syrups as well as differentiation with respect to other sweeteners.

ACK N OWLED G M ENT
The authors wish to thank the financial support provided by PRODEP; this allowed us to conduct the study.

CO N FLI C TO FI NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A LS TATEM ENT
This study does not include any animal or human tests.