Analysis of the Thermal Degradation of the Individual AnthocyaninComponents of Black Carrot (Daucus carota L.) – A New Approach UsingHigh-Resolution 1H NMR Spectroscopy

IOANNA ILIOPOULOU,ᵻ DELPHINE THAERON,‡ ASHLEY BAKER, ‡ ANITA JONES,ᵻ NEIL ROBERTSON*ᵻ

ᵻEaStCHEM School of Chemistry, Joseph Black Building, David Brewster Road, Edinburgh, United Kingdom EH9 3FJ,

‡Macphie of Glenbervie, Stonehaven, United Kingdom, AB39 3YG

*Author to whom correspondence should be addressed. Telephone +44 131 6504755; E-mail:

1

The black carrot dye is a mixture of cyanidin molecules, the NMR spectrum of which shows a highly overlapped aromatic region. In this study, the 1H NMR (800 MHz) aromatic chemical shifts of the mixture were fully assigned by overlaying them with the characterised 1H NMR chemical shifts of the separated components. The latter were isolated using RP-HPLC and their chemical shifts were identified using 1H and 2D COSY NMR spectroscopy. The stability of the black carrot mixture to heat exposure was investigated at pH 3.6, 6.8 and 8.0 by heat-treating aqueous solutions at 100oC and the powdered material at 180oC. By integrating high-resolution 1H NMR spectra it was possible to follow the relative degradation of each component, offering advantages over the commonly used UV/Vis and HPLC approaches. UV/Vis spectroscopy and CIE colour measurements were used to determine thermally induced colour changes, under normal cooking conditions.

KEYWORDS: Anthocyanins; Cyanidin, Thermal degradation; NMR Integration; acylation; UV/Vis spectroscopy; CIE colour measurements

1

INTRODUCTION

The colour of a food or beverage is of paramount importance, as it is the first characteristic to be noticed and one of the main ways of visually assessing the food before consuming it. The perceived colour provides an indication of the expected taste of food and the quality of a food is also first judged from its colour 1.

Many raw foods, such as fruits and vegetables, have vibrant, attractive colours. However, upon processing, their colour may fade or be completely lost. Most natural colours are highly labile towards temperature, pH, oxygen and light during processing and storage. The thermal impact during pasteurisation, sterilisation or concentration enhances the formation of degradation products and the concomitant colour loss. Consequently, the food products may cease to be attractive to consumers 2. Thus, it is important to understand the conditions governing colourant degradation in order to establish measures to avoid its occurrence 3.

Research over the past decades has produced incontrovertible evidence of the health benefits arising from the consumption of many fruits and vegetables. Many researchers have tried to identify the health- promoting ingredients of flavonoids, a class of phenolic. Most prominent amongst the flavonoids are the anthocyanins, one of the most abundant constituents responsible for the attractive red, blue and purple colours in many fruits and vegetables. They are widely found in berries, dark grapes, cabbages, red wine, cereal grains and flowers 4-6.

Anthocyanins are derivatives of salts called anthocyanidins7; they occur in nature as glycosides of anthocyanidins and may have aliphatic or aromatic acids attached to the glucosidic molecules 7-9. Anthocyanins are responsible for the intensive red colour of black carrot. The anthocyanin profile of black carrot has been analysed in the past and found to consist mainly of cyanidin-based dyes 10-13 (Table 1).

Thermal treatment can result in pigment breakdown and/or a variety of degradation species, depending on the nature of the anthocyanins and the severity of the heat-treatment 4. Sadilova et al. (2007) identified, by HPLC, different thermal degradation compounds which depend on the nature of the natural dyes 3.

Previous studies suggest possible mechanisms for the degradation of anthocyanins. Amongst them, opening of the pyrylium ring and formation of chalcone as an initial step was proposed by Hrazdina (1971) and Markakis (1957) 14-15. On the other hand, hydrolysis of the glycosidic moiety and formation of aglycon was suggested by Adams (1973)16. This study also confirmed that anthocyanins are degraded during heating into a chalcone structure which in a second step involves transformation into a coumarin glycoside with a B-ring loss. Von Elbe and Schwartz (1996) also suggested that coumarin 3,5-diglycosides are common degradation products for anthocyanin 3,5 diglycosides 4,17.

During heat exposure, the stability of anthocyanins depends on the composition and the characteristics of the medium, with pH playing an important role. Anthocyanins adopt different chemical structures which exist in pH-dependent equilibrium 18-20.

Some studies indicate that acylated anthocyanins, mainly those with planar aromatic substituents, exhibit greater stability, especially when kept in aqueous solutions, and play an important role in increasing the thermal stability of the dye compared to the non-acylated counterparts. It is believed that the aromatic residues of the acyl groups stack with the pyrylium ring of the flavylium cation which reduces the likelihood of the hydration reaction in the vulnerable C-2 and C-4 positions 21-25.

Black carrot consists of a high ratio of mono-acylated anthocyanins 10-13.The question that arises is; are black carrot acylated anthocyanins more stable compared to the non-acylated ones? In other words, does the structure of anthocyanins affect the stability?

Previous studies of black carrot have typically used HPLC and UV/Vis analysis for quantification of the components, but there are some weaknesses to these techniques. For HPLC, there are concerns about pH-dependent variation in the wavelengths of absorption maxima and the values of absorption coefficients, leading to unreliable quantification. Using UV/Vis spectroscopy alone, it is impossible to resolve and quantify the separate components; only the total anthocyanin concentration can be approximated.

In the present study, 1H NMR spectroscopy and signal integration was used to investigate the thermal degradation of the individual anthocyanin components of a commercial black carrot concentrate, and the effect of pH on this degradation. Complementary UV/Vis spectroscopy and CIE colour space measurements26 were used to follow colour degradation. Separation of the mixture into individual components, assignment of each component followed by integration of 1H NMR signals in spectra of the mixtures were the steps used. The resulting understanding of the relative stabilities of the different components, in particular the role of structure (acylation) on the stability should be valuable in developing future strategies to enhance the stability of this commercially available natural colourant.

MATERIALS AND METHODS

Plant Materials. Commercial concentrate of black carrot (Daucus carota L.) was supplied by Naturex Ltd (manufacturer’s code: COPG4167, sample code: G00017). The concentrate was stored at -18oC.

Solvents and Reagents. Deuterated NMR solvents were purchased as follows: methanol-d4 from Sigma Aldrich (USA) trifluoroacetic acid-d from Sigma Aldrich (USA) or Cambridge Isotope Laboratories (CIL) (USA). Hydrochloric acid S.G. 1.18 (≈37%), sodium hydroxide (97%) and sodium dihydrogen orthophosphate dihydrate were purchased from Fisher Scientific (UK). Disodium hydrogen phosphate dihydrate and citric acid monohydrate were obtained from Sigma Aldrich (USA). Acetonitrile and water for HPLC were purchased from VWR International. All the HPLC solvents were of analytical grade. C-18 Cartridges Vac. 35cc (10g) (WAT043345)) purchased from Waters (Ireland, U.K).

Sample Preparation. A two-step extraction process was applied to the black carrot sample to remove non-anthocyanin components. 100g of black carrot concentrate was mixed with 150 mL of chloroform in a separating funnel and left overnight. The aqueous phase was collected and further purified by solid-phase extraction 27, using mini columns (C-18 Cartridges Vac. 35cc (10g) (WAT043345)) purchased from Waters (Ireland). The eluent of the extraction (methanolic mobile phase), was concentrated in a rotary evaporator (IKA® RV 10 basic) at 25oC and further dried under vacuum, using liquid nitrogen, yielding 5g of powder.

High Performance Liquid Chromatography. 100 mg of the extracted black carrot powder (see section 2.3) were dissolved in 1 mL of distilled water. Semi-preparative reverse-phase high-performance liquid chromatography (RP–HPLC) was performed on an HP1100 system equipped with a semi-preparative C18 Agilent column Eclipse XDB-C18 (9.4x250mm i.d., 5 μm) at a constant temperature of 20oC and a flow rate of 2 mL/min. A mobile phase gradient was used for elution; eluent A consisted of water with 0.1 mL formic acid and eluent B of acetonitrile (ACN) and water with 0.1% formic acid (1:1). The elution profile was 10% B at 0min, 35% of B at 10min, 50% of B at 35min, 80% of B at 40min and 10% of B at 45min. The injection volume was 20μL and the detector was set at 520 nm. The fractions were transferred into vials and mass spectrometric analysis performed on an Agilent Series 1100 HPLC system fitted with an electrospray ionization (ESI) source. Repeated injections were performed, and the isolated fractions were combined until a mass of 2-5mg per fraction was obtained. The purified fractions were frozen using dry ice and acetone, and then dried under vacuum on a freeze-drier.

Heating Experiments.A domestic oven (Dēlonghi E012001W) was used to heat-treat aqueous solutions with pH values of 3.6, 6.8 and 8.0. Citric acid/ phosphate and phosphate buffers were used to adjust the pH. Hydrochloric acid and sodium hydroxide were also used where necessary for adjusting the pH, to avoid salts which interfere with the HPLC column. The samples were heated in an oven at around 180oC, to maintain the aqueous sample temperature at 100 oC, for periods up to 100 minutes. A 60-minute heat-treatment was also applied to samples for which the pH ranged between 3.4 and 8.2.

Powder samples of black carrot were prepared by dissolving black carrot in aqueous solutions, adjusting the pH to 3.6 and to 6.8, followed by freeze-drying. The powders were then exposed to heat in a high performance furnace (CARBOLITE® (UK), at 180oC).

NMR Spectroscopy.A mixed solvent consisting of 10 g MeOH-d4:0.5 mL trifluoacetic acid-d (TFA-d) was used for all NMR measurements.The structures of the compounds isolated by RP-HPLC were determined using 1D 1H-NMR analysis on a Bruker 500 MHz spectrometer (10 mg in 0.8 mL MeOH/TFA), in combination with two-dimensional COSY NMR to assign aromatic peaks. High-resolution 1H NMR spectra of the purified (unseparated) black carrot (10 mg in 0.8 mL of MeOH/TFA) were acquired on a Bruker 800 MHz spectrometer.In addition,1H NMR spectra (800 MHz) used for integration of the samples exposed to heat were also acquired (10 mg in 0.8 mL of MeOH/TFA).

UV/Vis Spectroscopy.Absorption spectra in the visible region (300-800 nm) were recorded using a Jasco V-670 series spectrophotometer. Solutions (40 μL of sample solution in 3 mLcitric acid/phosphate buffer) were contained in a quartz cell (d = 1cm) and the data were collected using Spectra ManagerTM II Software.

Colour Measurements. The Spectra ManagerTM II Software was used to calculate the CIE lab coordinates using the Jasco V-670 series spectrophotometer (Tokyo, Japan). Chroma value C*[C* = a*2 + b*2)1/2] and hue angle ho [ho = arctan (b*/a*)] were calculated from parameters a* (from green to red) and b* (from blue to yellow) values. The hue angles were expressed on a 360o colour wheel, in which 0 and 360o represent red, 90o yellow, 180o green and 270o blue. The illuminant was D65 and the observer angle was 10o. The change in colour produced by heat treatment was calculated using the ΔE* = [(ΔL*) 2 + (Δa*) 2 + (Δb*) 2] 1/2 equation at pH 3.6, 6.8 and 8 for several time intervals 26.

RESULTS AND DISCUSSION

Isolation and Structure Characterisation – NMR Studies.As shown in Figure 1, even after purification by solid-phase extraction, the NMR spectrum of the black carrot mixture contains many overlapping peaks, preventing the assignment of individual components. The extraction process has simplified the aromatic region (6.2 to 9 ppm) as shown by comparison with Figures S1 and S2, but further information on the individual components is needed to enable assignment of the aromatic protons.

Montilla (2011) describes that for different black carrot species the composition of anthocyanins can vary 11. Using RP-HPLC, five major anthocyanin components of black carrot were isolated, as shown in the chromatogram in Figure 2. Mass spectrometric analysis indicated that each fraction corresponds to a particular anthocyanin molecule (identifiable as the molecular ion), as summarised in Table 1 (Figures S3-S7). These results are consistent with previous studies of the composition of black carrot 10-13.

1D 1H-NMR and 2D COSY NMR analysis (Figures S3- S7) enabled assignment of each aromatic proton of each compound (Tables 1 and 2). The chemical shifts are consistent with the ones assigned for the anthocyanins from cell suspension culture of Daucus carota L. in Gläβgen’s study (1992) 28. The NMR results from the present study (Table 2) show that compounds 1 and 2 are the non-acylated components and compounds 3, 4 and 5 are the acylated anthocyanin compounds, confirming results from previous studies. It can also be seen that the presence of sinapic, ferulic and coumaric acids on the glucose moiety has an effect on the chemical shift of the cyanidin protons. For example, the chemical shifts of H-4 in the non-acylated compounds are in the range δ 9.016 -9.010 while those for the acylated compounds appear at δ close to 8.540. The same effect is seen on the chemical shifts of the H-6 and H-8 protons. In Gläβgen’s study (1992) a low-frequency shift of H-4 protons of black carrot anthocyanins acylated with sinapic, ferulic and coumaric acids compared to the non-acylated counterparts was also noted 28. Also, Dougall (1998) described a marked effect of cinnamic or benzoic acids on the chemical shifts of the cyanidin H-4, H-6 and H-8 protons in the acylated compounds providing evidence for NMR shifting caused by acylation 29.

The aromatic region of the black carrot mixture could be assigned completely with reference to the NMR spectra of the five separated components. Each signal in the spectrum of the mixture can be identified with a single signal or with overlapping signals from the spectra of the five compounds, as illustrated in Figure 3. In the region between δ = 6.0 and 7.5 ppm there is considerable overlap between peaks of the individual components in the spectrum of the mixture. On the other hand, in the region between δ = 7.8 and 9 ppm the individual component peaks are generally well resolved. The small intensity doublets in the region δ = 7.55, highlighted in red, and δ = 6.25 (overlapped), are attributed to minor impurities and also appeared in the spectrum for compound 5.

Thermal Degradation Studies. The clear assignment of the NMR peaks in the black carrot mixture enables the fractional concentration of compounds 1 – 5 to be unambiguously determined and to be followed during thermal degradation. The well-resolved region between δ = 8.4 – 9 ppm was used for integration to determine the composition of the mixture and to quantify the degradation of each component. Specifically, the integrals were determined for the combined H-4 proton signals of components 1 and 2 at δ = 9 ppm and the individual H-4 proton signals of the other three components in the region around 8.5 ppm. Overall, the expected general trend was noticed; the longer the exposure to heat, the more the integrated NMR signals of the compounds were reduced. However, it was also apparent that the integrals of the individual components were decreasing at different rates, resulting in variation in the composition of the mixture during thermal degradation.

Before examining the NMR results in detail, we first describe the general degradation behaviour observed in the UV/Vis spectra of the anthocyanin mixture in solution and as a solid powder.

UV/Vis Spectroscopy and Colour Measurements. Black Carrot in Solution. Exposure to heat at pH 3.6 for 100 minutes resulted in an increase in lightness, L*, by 7.99 units, insignificant change in the hue angle ho and a decrease in the chroma, C*, by 16.13 units (Table S1). This indicated that the colour of the anthocyanins was fading with slight change of the hue. The UV-Visible spectra (Figure 4a) showed a decrease in absorbance but the λmax was not notably shifted. Heat-treatment at pH 6.8 for 100 minutes also resulted in an increase in lightness (8.65 units) and decrease in chroma (22.26 units), but there was also an increase in the hue angle by 18.12 units (Table S1). Therefore, in neutral conditions, the colour not only faded but the hue also changed from red to a more orange shade. The UV-visible spectra (Figure 4b) showed both a decrease in absorbance and a slight bathochromic shift in the λmax. To explore the pH-dependence of these effects, the UV/Vis spectra of the black carrot heat-treated for 60 minutes at a range of pH values, from 3.4 to 8.2, were recorded. As shown in Figure 5, with increasing pH, the λmax shifted bathochromically andthe absorbance maximum was noticeably decreased. Observing the CIE lab parameters; the lightness gradually enhanced by 25.89 units, the ho increased dramatically to 70.43 units and the colour of neutral and more basic solutions decolourised and gradually turned brown. The colour saturation decreased 40.52 units (Table S2). It is clear from the UV/Vis spectra that there is more rapid degradation of the anthocyanins at higher pH values. This can be related to the different forms of the anthocyanins present as a function of pH; in the case of the acidic solution, more of the stable flavilium cationic form of anthocyanin is present, whereas in neutral conditions the percentage of less stable chalcone, carbinol and quinonoidal form will be increased.

Black Carrot Powders. To assess the effect of pH on solid-state samples, the pH was adjusted to 3.6 and 6.8 before freeze-drying. After 60 min of heat-treatment, the colours showed a slight increase in lightness (L*) by 2.73 and 5.03 units, respectively. The hue value change was negligible for both cases and the chroma (C*) value decreased by 7.92 and 12.47 units, respectively (Table S1). The difference in the absolute absorbance between the two samples is also negligible in the λmax of theUV-Vis spectra though (Figure 6 (a) and (b)). The UV-Vis spectra show that, at pH 3.6, the thermal stabilities of the solution and powder samples are similar; however, at pH 6.8 the powder shows much higher stability than the solution. Comparing the two powder samples, it is evident that the pH prior to freeze-drying has little effect on the stability.

NMR spectroscopy.Integration of the H-4 proton peaks in the aromatic region of NMR spectrum (δ = 8.4 – 9), as indicated in Figure 3 enabled the percentage of each component in the anthocyanin mixture to be quantified as a function of heating time, over a range of pH. The results are presented in Figures 7 to 11. As shown in the insets of the Figures, the overall degradation of the total anthocyanin content determined from the NMR integrals follows a similar trend to that derived from the UV/Vis absorbance. Notably, however, the absorbance measurements imply a lesser extent of degradation than that quantified by the NMR data. This can be attributed to residual absorbance, in this spectral region, by the decomposition products, which persists after degradation of the primary anthocyanin components. Thus, the NMR data are able to give a better quantitative measurement of degradation than the commonly used UV/Vis data.