A fast and sensitive isocratic high performance liquid chromatography method for determination of guaraná (Paullinia cupana) flavan-3-ols

Graphical abstract


Background
Flavan-3-ols are a class of flavonoids, which are very abundant in guaraná (Paullinia cupana) as well as other sources of such as cocoa, green tea and apple. Despite their high concentration in foods, their bioavailability is typically low, demanding sensitive methods for their quantification in samples from human bioavailability studies as well as Caco-2 permeability assays. High performance liquid chromatography (HPLC) coupled with high-sensitivity detectors such as coulometric, fluorescence or mass spectrometry are usually the choice for the determination of flavan-3-ols (also known as catechins) in bioavailability studies. However, these detectors are expensive and not usually available in smaller labs. The ultraviolet-visible (UV) detector is the most affordable and commonly available detector for HPLC setups, but the drawback of HPLC-UV methods is the low sensitivity of catechins at their typical and most selective absorbance maximum (280 nm). Catechin and epicatechin show absorbance maxima at 280 and 210 nm. The 280 nm band is associated with electron transitions of catechin's A-ring benzoyl system, thus reasonably selective but with a low response. The 210 nm band is several times more responsive (i.e. giving high sensitivity) than the 280 nm band, however many organic compounds absorb at this wavelength, thus reducing the method's selectivity.
Samples from Caco-2 cell permeation assays of guaraná in vitro digesta usually contain low catechin concentrations, but they are relatively 'clean' (i.e. without interfering peaks at low wavelengths).Therefore, detection at 210 nm would be an option to increase the method's sensitivity. A sample of guaraná extract was previously analysed by HPLC with a photodiode array in the 190-450 nm range. The chromatogram showed that all peaks at 210 nm were from the same compounds at 280 nm (data not shown).
Our previous method [1], as well as most methods for quantification of catechins in guaraná [2][3][4] and other food matrices rely on lengthy gradients to resolve all peaks and to remove lipophilic compounds left behind in the column, and the analysis time for one sample ranges from 45 to 135 min. The present method is a fast, sensitive and simple HPLC-UV method, with a 16-minute isocratic run that deliver the high throughput needed to process a large number of samples without compromising the analyte stability. This method is suitable for the determination of catechins, procyanidins (up to degree of polymerisation 2) and caffeine in guaraná extracts, in vitro digesta, and Caco-2 cell permeates.

Material and chemicals
Guaraná seeds were purchased at a local market in Manaus (Amazonas state, Brazil) and their geographical origin was Maués (Amazonas state). Seeds were pulverised with a blender and passed through a 250 mm sieve before extraction or in vitro digestion. (+)-Catechin and (-)-epicatechin were purchased from Nacalai Tesque (Japan), Procyanidin B1 from Cayman Chemical (USA), Procyanidin B2 from Extrasynthese (France) and caffeine from FUJIFILM-Wako (Japan). Acetonitrile was of HPLC grade and purchased from Sigma-Aldrich or Honeywell. All other chemicals were of reagent grade. Ultra-pure water (resistivity higher than 18 MV cm) was used in all experiments.
Injection solvent with preservatives (diluent for working standards and samples)

Method validation
Samples from a previous study [5] were used for method validation. Samples containing guaraná catechins were: guaraná extracts, in vitro digesta of guaraná seeds and basal media (Hanks' balanced salt solution, HBSS) from Caco-2 permeability assays done with guaraná in vitro digesta.
Typical chromatograms from guaraná extracts, in vitro digesta and their Caco-2 cell permeates from bioaccessibility studies are shown on Fig. 1, and typical retention times are listed on Table 1.
Method validation was performed as described by the International Conference on Harmonization (ICH) guidelines [6]. Linearity was confirmed throughout the 10-point calibration range of all five compounds, with correlation coefficients higher than 0.999 for all tested compounds as shown in Table 1. We chose to use a zero-intercept equation) as it did not cause loss of linearity.
Limits of detection (LOD) and limits of quantification (LOQ) were calculated as the concentration equivalent to three and 10 times the baseline noise of basal media chromatograms, respectively (n = 48). LOD values for catechin, epicatechin, procyanidin B1, procyanidin B2 and caffeine were 0.019, 0.030, 0.028, 0.030, and 0.043 nmol/mL, respectively (Table 2). These values are 80-360 times lower than a previous HPLC-DAD method for guaraná catechins with detection at 280 nm [2]. The sensitivity of the present method is also a good compromise taking into account the simplicity of the HPLC-UV equipment and its low cost. Our previous method for catechin analysis with a far more expensive coulometric detection [1] is barely 3-6 times more sensitive than the present method, with a basic HPLC-UV setup.
Method accuracy (recovery) was measured after spiking blank in vitro digesta (i.e. the soluble fraction of a blank in vitro digesta) and basal medium of permeation assays (Hanks' balanced salt solution, HBSS) with 1.56, 12.5 and 50 nmol/mL of catechin, epicatechin and caffeine, and 0.78, 6.25 and 25 nmol/mL of procyanidin B1 and B2, in triplicate. Intraday and interday variation was calculated as relative standard deviation (%RSD) of spiked HBSS and diluted blank in vitro digesta (n = 9), freshly prepared and analysed in two different days by the same analyst. As shown in Table 3, a Curve plotted from 20 mL injections of 10 standard solutions across the specified range.
b Slope of the regression equation y = ax where y is the peak area and x is the standard concentration in nmol/mL.  [6].

Additional information
The present method with detection at 210 nm cannot be used when samples are dissolved in DMSO, methanol, dimethyl formamide or any other solvent with high absorption at low wavelengths. We tried analyzing Caco-2 cell extracts containing methanol, dimethyl formamide and DMSO and they all produce tailing solvent peaks that hindered the detection of catechins at 210 nm. However, only dimethyl formamide efficiently recovered catechins and procyanidins from Caco-2 homogenates. Alternatively, cell homogenates can be extracted and measured at 280 nm, however with a lower sensitivity compared to measurements at 210 nm Table 4 Repeatability (intraday RSD) and reproducibility (interday RSD) of the analysis of catechins, procyanidins and caffeine added to basal media (HBSS) and in vitro digesta, by isocratic HPLC-UV.