Regularities of Anthocyanins Retention in RP HPLC for “Water–Acetonitrile–Phosphoric Acid” Mobile Phases

The influence of exchange of HCOOH (System 2) by phosphoric acid (System 1) for acidification of the “acetonitrile–water” mobile phases for reversed-phase HPLC of anthocyanins was investigated in the framework of relative retention analysis. The differences and similarities of anthocyanins separation were revealed. It has been shown that some common features of the quantitative relationships may be used for preliminary anthocyanins structure differentiation, according to the number of OH-groups in anthocyanidin backbone as well as to a number of saccharide molecules in glycoside radicals in position 3 of the anthocyanin without MS detection.


Introduction
Anthocyanins are powerful water-soluble antioxidants of flavonoids class with health promoting effect [1,2]. The coloured flavylium form of anthocyanins is a reason to regard them as natural food colorants [2]. The latter explains a high scientific and technological interest to the substances. Anthocyanins are synthesized in plant tissue, mainly in fruits, flowers and for some species in leaves as a rule as a complex mixture of compounds with different structures [3]. Anthocyanins are glycosides of anthocyanidins ( Figure 1), with great varieties of more than 600 anthocyanin structures found in plant sources [4] though only six structures of the latter cover the majority of the structures due to glycosylation type variability [5].
Reversed-phase HPLC is a common method for analysis of complex mixtures of plant anthocyanins [6,7]. The specificity of the method is usage of rather strong acidic mixtures of water and organic water-miscible solvent. Acidification is necessary to transfer the substances into charged and coloured flavylium form, due to the fact that anthocyanins may be easily detected at the presence of large amounts of other colourless substances.
The retention of substances in HPLC depends upon type and even trademark of stationary reversed phase, composition of mobile phase, and temperature as well as solute structure. Mobile phases of water mixtures with acetonitrile or methanol are acidified with HCOOH [8], acetic [9], phosphoric [10], and trifluoroacetic [11] acids as well as the mixtures (without any explanation) of some of them [12][13][14] for anthocyanins separation. However, as far as we know only "HCOOH-acetonitrile-water" mobile phases were investigated extensively to elucidate the regularities of anthocyanins retention [8,[15][16][17]. According to Snyder' selectivity triangular water, acetonitrile and HCOOH are solvents of different groups, VIII, VIb, and IV correspondingly [18], so the withdrawal or exchange of some solvents may lead to alteration of solutes separation selectivity.

Extraction and Partial Purification of Anthocyanins.
Plant material was dispersed in 0.1 M solution of HCl in distilled water with blander and macerated overnight. Then supernatant was decanted and applied for solid phase extraction by DIAPAC C18 cartridges (BioChemMak ST, RF).

Semipreparative Isolation of Individual Anthocyanins.
The isolation of individual anthocyanins was performed by Shimadzu equipment LC-20 with spectrophotometric detection on chromatographic column 10 × 250 mm SUPEL-COSIL C18 (5 mcm) in eluents of water-acetonitrile, 10 vol. % HCOOH system. System 1: solutions of acetonitrile in distilled water (8-14 vol. %), acidified with phosphoric acid (0.5 vol. %). The concentration of phosphoric acid (0.5 vol. %) in System 1 was chosen to match the pH close to that for 10 vol. % of HCOOH (1.45-1.50) of System 2. It should be mentioned that the decrease of acid content leads to increase of mobile phase pH and to decrease of number theoretical plates of the chromatographic system under investigation [19].

Spectral Characteristics.
Electronic spectra of the anthocyanin' peaks were recorded in DAD cell with a range step of 0.50 nm. Mass spectra were recorded at ESI-mode with column 2.1 × 150 mm Kromasil 100-3.5C18, mobile phase 10 vol. % of HCOOH and 8 vol. % of CH 3 CN in distilled water, 150 mcl/min. Fragmentor voltage of 100 V was applied to get molecular ions and 150 or 200 V to get fragmented ions of corresponding anthocyanidins.

Common Consideration.
The dependence of retention of sample anthocyanin cyanidin-3-glucoside, Cy3Glu (Figure 2), upon volume fraction of CH 3 CN in a mobile phase is rather linear function according to Snyder equation [20] (1) for System 1 ( 2 = 0.9998) as well as for System 2 ( 2 = 0.998): where ( ) is retention factor of Cy3Glu at a given , volume fraction of CH 3 CN in a mobile phase, and log ( ) and ( ) are linear approximation constants. The replacement of 10 vol. % of HCOOH by 0.5 vol. % of phosphoric acid results in pronounced increase of Cy3Glu retention, so that ∼3.5 vol. % of CH 3 CN must be added to compensate the increase. Moreover, this means that elution power of HCOOH is three times lower than that of CH 3 CN. Meanwhile for the eluent Systems 1 and 2 parameters ( ) are 0.162 and 0.150, respectively. The relatively small difference between negative slopes ( ( )) becomes valuable for the calculated values for the stoichiometric displacement model [20] (2) on the same basis: where ( ) is retention factor of Cy3Glu at a given c(CH 3 CN), molar concentration of CH 3 CN in a mobile phase, and and are linear approximation constants; corresponds to the number of CH 3 CN molecules that are released into a mobile phase during solute sorption.
For System 1 and System 2 parameter is 4.50 and 2.60, respectively. The results indicate a valuable role of HCOOH in sorption-desorption processes and its withdrawal may influence the solute separation selectivity.

Solutes with Different Anthocyanidin Structures and the Same Glycosylation
Type. The separation map in the framework of relative retention analysis [17] with pelargonidin-3glucoside as a reference solute is presented in Figure 3. Each point on the plot has coordinates , logarithm of capacity factor of Pg3Glu and , that for corresponding solute in the same mobile phase. Points for the same solute and different mobile phase compositions settle down on straight lines according to equation of relative retention (3) ( Table 1): Pg3Glu was taken as a reference solute for the simplest ring B structure. Parameters of (3) are the valuable characteristics of corresponding solutes. For example, addition of OH-group into position 3 of ring B (for a transfer from Pg3Glu to Cy3Glu) leads not only to decrease of retention (and parameter ) because of solute hydrophilicity increase but also to increase of parameter as a consequence of additive van der Waals interactions of these O and H atoms with stationary phase atoms. It is easy to see that addition of OH-and CH 3 O-groups to positions 3 and 5 of ring B leads to close to additive increase of capacity factor logarithm. This property is true for the same solute retention in System 2 [15]; thus the sequence of anthocyanins' elution on the chromatograms of mixtures of 3-glucosides delphinidin, cyanidin, petunidin, pelargonidin, peonidin, and malvidin for both systems remains unchanged (Figure 4). Meanwhile the exchange of phosphoric acid by HCOOH one results in slight selectivity alterations: some decrease of relative retention for OH-substitutions and an increase of that for OCH 3 -substitutions are evident ( Figure 5).
Thus, points for Mv3Glu and Pn3Glu for System 2 settle down above the lines for the same anthocyanins for System 1, while the opposite case is found for relative retention of Dp3Glu and Cy3Glu. Accordingly, System 2 has somewhat higher selectivity for separation of substances with different flavylium ions hydrophilicity.
The lines of relative retention approximated to the region of zero points [21] (left down corner of the plot on Figure 3) are differentiated according to the number of OH-groups    in ring B of anthocyanin structure. Moreover, according to our observation the place of OH-addition has no meaningthe particular anthocyanins of Alstroemeria flowers [22] with OH-groups at carbon atoms number 6 have retention close to that of isomers with OH-groups in ring B. This is in a full agreement with the property of reversed-phases chromatography having a low selectivity for isomers separation. Accordingly, the position of lines of relative retention (3) in the region of zero points may be utilised for tentative chromatographic estimation of the number of OH-groups in solute molecules.

Solutes with the Same Anthocyanidin Structure and
Different Glycosylation Type. The separation map of relative retention of some cyanidin-3-glycosides with Cy3Glu as a reference solute in System 1 is presented in Figure 6. It becomes evident that not only the absolute retention but also relative retention of different cyanidin-3-glycosides depends not only upon solute structure, but also upon mobile phase composition; coelution of some solute pairs including the reversal of the elution order may occur after alteration of In the case of System 1 parameter of (3) may be utilised for preliminary estimation of complexity of glycoside structure in position 3 of anthocyanidin backbone: has a value in the region 1.000 ± 0.020 for monoglycosides (Cy3Gala and Cy3Ara), 1.122 ± 0.011 for diglycosides (Cy3Sopho, Cy3Sam, Cy3AGlu, and Cy3Rut), and 1.300 ± 0.030 for triglycosides (Cy3GRut, Cy3XRut) ( Table 2). The values are close to that reported for retention of cyanidin-3-glycosides in RP HPLC in solvent System 2 [17,23] proving the property to be a common regularity of the solute chromatographic behaviour at least for the systems under investigation.
Finely, the exchange of HCOOH by phosphoric acid also leads to slight decrease of relative retention of diand trisaccharides, Figure 7, but no significant alterations of solutes separation selectivity were found.

Conclusions
The exchange of HCOOH (System 1) for phosphoric acid (System 2) leads to substantial increase of anthocyanins retention in mobile phases, acidified mixtures of water and acetonitrile.
Selectivity of resolution of the same glycosides of six common anthocyanidins is only slightly greater in System 2, by the way, though the sequence of elution remains the same for reasonable solutes retention times.
Analysis of anthocyanins relative retentions on the separation map (in the region of zero points) may be explored for estimation of the number of OH-groups in anthocyanidin backbone.
Selectivity of resolution of the different glycosides of the same anthocyanidin (an example of the most common natural aglicone, cyanidin) is also close to that for System 1.
But the advantage of relative retention analysis is a sensitivity to structure of carbohydrate radicals; parameter for mono-, di-, and trisaccharides is differing enough permitting determination of complexity of glycosyl radical in the 3 positions of anthocyanidin without utilization of MS detection. Moreover, parameter is highly sensitive to sugar isomers structures. Thus, eluent systems under investigation have close properties, though System 2 seems to be somewhat more efficient.