Improving red radish anthocyanin yield and off flavor removal by acidified aqueous organic based medium

Abstract

In view of the high content of the highly stable anthocyanin in red radish roots, the plant is considered as a potent source of natural anthocyanins. However, the presence of off flavor in their anthocyanin extracts is the only limitation as far as their application in food and pharmaceutical industries is concerned. The aim of this work was to develop a suitable method for extracting the off flavor free red radish root anthocyanin, while conserving high anthocyanin yield. Results showed that phosphoric and citric acidified hexane (pH 2.5) extracts yielded maximum anthocyanin levels of 208.37 mg/100 g and 202.89 mg/100 g fresh radish, respectively with lower percentage polymeric color and higher color density, chroma and hue angle compared to other extracts at different pH values. Tentative anthocyanin identification by UPLC-TQ-MS showed 12 glycosylated anthocyanins substituted at C3 and C5 in phosphoric acidified extracts and only 10 in citric acidified extracts. In addition, analysis of total glucosinolates (GSLs) revealed significant inhibition of glucosinolate degradation in acidified hexane extracts (pH 2.5) compared to the other extracts. The off flavor evaluated by GC-MS revealed a significant decrease of isothiocyanates, sulfides and nitrogen containing compounds in acidified hexane (pH 2.5) compared to other extracts. Furthermore, dehydration and rehydration of extracts showed that over 70% of off flavor compounds could be removed in two phases (aqueous and hexane) during hexane removal. Results of sensory evaluation confirmed that the red radish root anthocyanin from the phosphoric acidified hexane (pH 2.5) was closer to the color and odor characteristics of the commercial red radish anthocyanin. Therefore, the phosphoric acidified hexane method could be a suitable method to extract off flavor free anthocyanin from red radish roots.

---

1. Introduction

Recently, the use of safe and natural colorings in food products has attracted the attention of many researchers as alternatives to synthetic dyes because of both legislation and consumer health concerns.¹ Food colors from new plant sources that have been used or suggested for potential applications in food systems, include anthocyanins from red grape skins,² roselle,³ strawberry⁴ and red radish.⁵

Chinese red radish (Raphanus sativus L.), uniquely characterized by green colored skin and red flesh, is a potential economic plant due to the significant amounts of natural food colors it possesses. Widely known for its use as a natural pigment in food, Chinese red radish has the advantage of not only being a natural source of food color but also has high stability and attractive red hue characteristics, similar to synthetic Food Red no. 40.⁶ The major limitation in the use of red radish anthocyanin products as food colorants or nutraceuticals is the distinct undesirable flavor associated with its preparations. The off-flavors result from the degradation of glucosinolates (sulphur-containing compounds in cruciferous vegetables) by myrosinase enzyme. Glucosinolate degradation is affected by processing procedures and storage conditions.^7,8

Various studies have been carried out to improve the color intensity and the yield of anthocyanins without affecting their functional properties. Patil et al. reported that the mixture of 50% ethanol and acidified water resulted in maximum anthocyanin yield with better chroma and hue angle and further used osmotic membrane extraction for de-alcoholization and concentration of the anthocyanin extract at ambient temperature and atmospheric pressure.⁹ Rodriguez-Saona et al. studied the concentration of anthocyanin from red radish extract by employing centritherm evaporation or direct osmotic process separately, as well as in combination with each other.¹⁰ Gachovska et al. reported that pulse electric field (PEF) treatments enhanced total anthocyanin extraction in water from red cabbage by 2.15 times with a higher proportion of nonacylated forms than the control.¹¹ Gilewicz-Łukasik et al. employed ultrafiltration and nanofiltration for the concentration of anthocyanin from aronia (black chokeberry) fruit.¹² However, the use of some organic solvent during anthocyanin extraction has raised safety concerns for human consumption due to potential toxic effects arising from the residual solvents. Commonly used solvents for the extraction of natural colors include ethanol, methanol, and acetone. Metivier et al. noted that methanol was the best extractant, being 20% more effective than ethanol and 73% more effective than water.¹³ Hydrochloric acid was most effective with ethanol, but is also very corrosive. Of the organic acids, citric acid was the most effective with methanol and acetic acid with water. Hexane is another solvent used to extract anthocyanins and polyphenols.¹⁴ It is easily vaporized when exposed to air or nitrogen gas. Such property of hexane of easy vaporization could be considered as good potential and safer solvent in extraction of natural anthocyanin colorant used in food industries. The combination of n-hexane with acidified water has not yet been used in extraction of anthocyanins.

Recently, different techniques have been used to improve the red radish anthocyanin yield and reduce the undesirable off flavor characteristics including membrane processes such as ultrafiltration and osmosis which have been found to alleviate the undesirable aroma compounds, but do not completely solve the off-flavor problem.¹⁰ However, these approaches are generally more expensive for removing the sulfur containing components from radish anthocyanins. The inhibition of myrosinase activity by ferrous ion at different pH showed that ferrous ion was not involved in the enzymatic liberation of glucose from glucosinolate but in the subsequent degradation of the aglucon.^15,16 Gao et al. found that the use of adsorptive chitosan could reduce glucosinolate content resulting in the removal of off flavor from red radish.¹⁷ Kottman extracted color from premium radish skin by abrasive action or steam then treated the extracted colored juice by filtration, evaporation, salt and acid addition followed by heat.¹⁸ The results of sensory evaluation showed that samples extracted with salt, acid or steam had lower aroma intensity but with lower pigment yields. Though all these techniques have been utilized, there is still a long way to go in removing the off-flavor from red radish pigment extracts without affecting the anthocyanin yield and structure. The assessment of the flavor compounds from red radish anthocyanin extracts is of importance in the process of making a colored extract that would be suitable for industry uses and acceptable to consumers. Different techniques have been utilized to evaluate the flavor compounds in food products as well as in extracted red radish pigments including descriptive sensory evaluation by trained panel and a combination of gas chromatography and mass spectrometry (GC-MS).

There is a growing demand for developing suitable extraction techniques for more efficient and effective extraction of available active matters from the plant materials. Therefore, the aim of this study was to develop a suitable method for removing the off flavor from red radish anthocyanin while simultaneously maintaining the stability and higher anthocyanin yield. This method could be useful in future to produce off flavor free red radish pigment in food industries and nutraceutical companies.

2. Material and methods

2.1. Materials

Red radishes were obtained from a farm in Chongqing Province and subsequently stored at −40 °C prior to their use. Food grade citric acid and phosphoric acid were purchased from Sigma (Shanghai, China). 1,2-Dichloro-benzene, dimethyl disulfide, dimethyl trisulfide, 3-(methylthio)propyl isothiocyanate, 4-(methylthio)butenyl isothiocyanate, 3-buten-1-yl-isothiocyanate and 2-phenyl isothiocyanate, internal standards (Gas Chromatography grade) were purchased from ANPEL Laboratory Technology (Shanghai) Inc. All other reagents used in this study were of analytical grade.

2.2. Aqueous extraction of anthocyanins pigment from red radish

Twenty-five grams of frozen radish were quickly thawed and cut into small pieces before immersing into 25 mL of citric and/or phosphoric acidified water at pH 4. The samples were homogenized in fruit juicer (Joyoung JYL-D020, Joyoung Co., Ltd, China) into radish juice and 50 mL of acidified water was added to make the mixture to 100 mL. The pH of the mixture was adjusted with concentrated citric acid or phosphoric acid (6 N) to pH 2.5, 3.5, and 4.5. The red radish anthocyanins pigments were extracted at 50 °C for 2 h. After extraction, the mixture was filtered by Whatman no. 1 filter paper through a Buchner funnel and stored at 4 °C prior to further analysis.

2.3. Solvent extraction and concentration of anthocyanins pigment from red radish

The two phases red radish anthocyanin pigment extraction consisted of first by enzyme inhibition at low pH using different acid as described above (aqueous phase) followed by extraction in organic phase with n-hexane. The anthocyanin pigment extracts were firstly extracted in citric and/or phosphoric acidified water for 2 h at 50 °C and then were mixed with n-hexane in ratio 1 : 1 (v/v) according to preliminary analysis (data not shown). The mixture was extracted at 45 °C for 1 h. After the extraction, the mixtures were cooled at room temperature. The extract was first separated on a separating funnel and the aqueous phase (low phase) was collected and concentrated in a rotary vacuum evaporator at >40 °C to ensure the solvent was totally removed.

2.4. HPLC-MS anthocyanins identification

Prior to HPLC-MS analysis, anthocyanin containing extracts were passed through a (1 g, 6 mL) LC-C18 SEP cartridge (ANPEL Laboratory Technology, Shanghai Inc.), previously activated with 3.8 mL of methanol followed by 4.8 mL of acidified water (0.5% TFA) through the sorbent bed.¹⁹ Anthocyanins and other phenolics were absorbed onto the C-18 cartridge, and sugars, acids and other water-soluble compounds were eluted with 20 mL of acidified water (0.5%, v/v, TFA). The non-anthocyanin phenolics were eluted from the cartridge using 20 mL of ethyl acetate, and anthocyanins were subsequently eluted with 40 mL of acidified methanol (0.5% v/v, TFA). The methanolic extracts were then concentrated using a vacuum rotary evaporator at 40 °C. Mass spectrometry (MS) analysis was performed on UPLC-TQ-MS (Waters, Tokyo, Japan). Mass spectrometry was performed with a Waters TQD system for data collection. Compounds were analyzed under the positive ion (PI) mode with electrospray ionization (ESI) source. The major MS parameters were as follows: mass range 100–1000 m/z, ion source temperature 200 °C, heated block temperature 200 °C, nebulization gas flow 1.5 L min⁻¹, detector voltage 1.75 kV. Positive and negative ion modes were performed. Data acquisition and processing were conducted using the LCMS Solution software (Mass Lynx V4.1, Waters). Anthocyanins were identified by comparing the mass data with published data.²⁰

2.5. Determination of total anthocyanin content

Total anthocyanin content was determined using the pH-differential method described by Lee et al.²¹ using two buffer systems: potassium chloride buffer, pH 1.0 (0.025 M), and sodium acetate buffer, pH 4.5 (0.4 M). A 0.1 mL aliquot of the pigment extract was transferred to a 10 mL volumetric flask and made up to 10 mL with corresponding buffer, and the absorbance was measured at 512 and 700 nm. The dilution factor (DF) was determined by diluting each extract in 0.025 M potassium chloride (Fisher Scientific, Fair Lawn, NJ) buffer at pH 1.0, until the absorbance was within the appropriate spectrophotometer range; greater than 0.1 and less than 1.2. Total anthocyanins were calculated as cyanidin-3-glucoside according to the following equation:

Total anthocyanins (mg/100 g FW) = A × MW × DF × 1000/(ε × l),

where A (absorbance) = (A₅₁₂ − A₇₀₀)pH 1.0 − (A₅₁₂ − A₇₀₀)pH 4.5; the MW and ε were calculated as pelargonidin-3-glucoside, where the MW = 433.2 g with an extinction coefficient of 31 600 cm⁻¹ mg⁻¹; DF = dilution factor; l = path length in cm; 1000 = conversion from g to mg. All analyses were done in triplicate (n = 3).

2.6. Determination of color density and percentage polymeric color of anthocyanins

A bisulfate bleaching method was used to determine percentage the polymeric color of the extracted red radish anthocyanins. Two hundred microliters of 20% K₂O₅S₂ and 0.2 mL of distilled water were added to two different 2.8 mL samples of extracted anthocyanins separately to make bleached and unbleached samples. Then, a spectrophotometer A360 (AOE Instrument (Shanghai) CO., Ltd, China) was used to read and record the absorbance of bleached and unbleached samples of anthocyanin at the λ_max 512 nm, 420 nm and at 700 nm (for correction of turbidity). The percentage of polymeric color of anthocyanin extracts was calculated using the following equations:

Color density = (λ_{512 nm} − A_{700 nm}) + (A_{420 nm} − A_{700 nm}) × DF

% polymeric color = (polymeric color/color density) × 100

2.7. Determination of color properties

Color was evaluated with a WSC-S color difference meter (Shanghai, China). For color measurements of radish red pigment in aqueous system, 2 mL of samples were placed in 2.5 cm diameter test tube for measurement. The CIELAB parameters (L*, a*, b*, C, h) were determined. The ΔE, C, h representing the total color differences between the L*, a*, and b*, chroma and hue angle of the sample respectively was shown as followings

ΔE = ((ΔL*)² + (Δa*)² + (Δb*)²)^1/2

C = ((a*)² + (b*)²)^1/2

h = tan⁻¹(b*/a*)

2.8. Determination of total glucosinolates

2.8.1. Extraction of GSLs. Glucosinolates were extracted according to the method described by Ishida et al. with slight modification.²² Fifty milligrams of lyophilized red radish anthocyanin extracts powder were placed in glass tubes, and each test tube was pre-incubated in a hot water bath at 75 °C for 1 min. After cooling, 4.8 mL of 80% (v/v) methanol was added and the GSLs extraction was carried out at room temperature. The tubes were kept at 25 °C for 30 min and then shaken reciprocally (120 rpm) for 30 min in a shaker. The tubes were centrifuged at 1600 × g for 10 min. The supernatant was used as a crude extract.

2.8.2. Palladium colorimetric analysis of the total GSLs content. Colorimetric analysis of the total GSLs content was performed by simplifying the method described by Møller et al.²³ Purification with ion-exchange chromatography was omitted.^24,25 To 0.2 mL of crude GSL extract, 0.3 mL of distilled water and 3 mL of 2 mM palladium(II) chloride were added and mixed. After incubation at 25 °C for 1 h, absorbance at 452 nm was measured using a spectrophotometer A360 (AOE Instrument (Shanghai) CO., Ltd, China). Absorbance was shown by an average of three measurements. After subtraction of value of a blank, as obtained using 0.2 mL of distilled water as a sample, differences of absorbance were used for estimation of total GSL contents according to the equation of linear regression relating the concentration of sinigrin (standard) to absorbance y = 2.6365x + 0.0228 (R² = 0.9948) and the concentration was estimated as μmol sinigrin per g (DW).

2.9. Headspace solid-phase microextraction (HS-SPME)

The volatile compounds of red radish extracts before and after n-hexane treatment were investigated by headspace solid-space microextraction (HS-SPME) combined with gas chromatography-mass spectrometry according to the method of Eric et al.²⁶ The SPME device for manual extraction included a holder assembly with a 50/30 μm divinyl-benzene/carboxen/polydimethylsiloxane (DVB/CAR-PDMS) fibre (Supelco, Madrid, Spain). The fibre was equilibrated at 250 °C for 30 min prior to use as recommended by the manufacturer. Radish extracts (5 mL) were transferred into vials (15 mL) and 50 μL of internal standard were added. A proper mixing of the solutions during the SPME was achieved with a magnetic stirrer. The syringe assembly unit was lowered into the vial with the fibre suspended in the headspace above the liquid layer of extracts for 30 min at 40 °C. Then, the fibre was immediately retracted back into the needle and transferred immediately to the injection port of gas chromatograph. A time period of 8 min was adopted for desorption and conditioning at the desorption temperature of 250 °C.

2.10. Gas chromatography/mass spectrometry (GC/MS)

The GC-MS was performed with an SCION-SQ-456-GC-MS, (Bruker Daltonics, Billerica, MA, USA). An injector in a splitless mode was maintained at 250 °C. The separation was achieved on a DB-Wax fused silica capillary (50 m × 0.32 mm, 1 μm film thickness) supplied from J&W Scientific, Folsom, CA, USA. Helium was used as the carrier gas at a flowrate of 1.0 mL min⁻¹. The inlet temperature was 250 °C. The column temperature was initially maintained at 40 °C for 1 min, increased at the rate of 5 °C min⁻¹ to 200 °C, maintained at 200 °C for 5 min, increased at the rate of 10 °C min⁻¹ to a final temperature of 240 °C and maintained at 240 °C for 5 min. The MSD transfer-line and ion source temperatures were 230 and 250 °C, respectively, with an electron-impact ionization potential of 70 eV. A mass range from m/z 35–400 was recorded in a full scan mode. Data was collected and processed using a HP-Chem station system.

2.10.1. Identification and quantitative determination of components. Individual peaks were identified by comparing their retention indices (LRI) and mass spectra (MS) with spectra in our homemade library, as well as by computer matching against the Wiley 275-library spectra database, NIST Mass Spectral Search Program (National Institute of Standards and Technology, Washington, DC, USA) and comparison of the mass spectra with literature data.^27,28 Spectra and LRI values of sulfides and isothiocyanates were compared with those of authentic compounds obtained commercially.

2.10.2. Determination of linear retention indices (LRI). A standard mixture of n-alkane (C6–C20) in ethanol was analyzed each day before GC runs to allow a check of the instrument performance and the calculation of retention indices of each component in the samples. The standard (1.0 μL) was injected to the trap and the solvent was removed by purging with oxygen-free nitrogen (40 mL min⁻¹) for 5 min. These alkanes were used as external standard references in Linear Retention Indice (LRI) calculations. LRI of each compound was calculated from the standard alkane retention time and the peak retention time using the following formula:
where: LRI = linear retention index, RT_x = retention time of compound, RT_n = retention time of n-alkane before peak, RT_n+1 = retention time of n-alkane after peak, n = carbon number of n-alkane before peak.

2.11. Sensory evaluation

The citric and/or phosphoric acidified (aqueous phase) red radish anthocyanin extracts pH 2.5; 3.5 and 4.5 and extracts from two-phase mixture (aqueous and organic phase) were diluted with spring water to a final anthocyanin content of 20 mg/100 L. The diluted solutions (8 mL) were placed into 60 mL amber-colored glass bottles capped with Teflon caps. Enough samples were prepared in order to allow for the headspace to rebuild between panelist usages. Thirteen trained panelists from food science and technology school (Jiangnan University) participated in the test and they were asked to rate the overall visual color and off flavor intensity. A multi sample difference test was performed. The color and off flavor intensity of the pigment extracts were rated using a 10-point hedonic scale. Samples were labeled with 3-digit random numbers and presented to panelists simultaneously in a randomized order. The commercial Chongqing red radish pigment extracts were used as control; the visual color and aroma intensity of the control were rated (10) and (5), respectively. The sensory evaluation was carried out in two separated days to avoid panelist's fatigue. The first day, the panelists assessed the samples from aqueous extraction and the second day, they assessed sample prepared from combined aqueous and organic phase. All the assessments were carried out in triplicate. The highly rated samples were finally compared to design the best of the red radish pigment extract with acceptable visual color and aroma intensity.

2.12. Statistical analysis

Statistical analysis was performed using Microsoft Excel 2016 (Microsoft, Redmond, WA, USA) and SPSS 24 (IBM, Armonk, NY, USA). Analysis of variance was used to compare means with Tukey multiple range tests for post-hoc analysis. P ≤ 0.05 or P ≤ 0.01 was considered significant. Data reported in this work were the means of duplicate (GC-MS) or triplicate (anthocyanin yield, color change and sensory evaluation) experiments.

3. Results and discussions

In order to know the suitable extracting medium for the maximum extraction and removal of off flavor from red radish roots anthocyanin pigments, different experiments were carried out involving two extraction media; citric and phosphoric acidified water at different pH (2.5; 3.5; 4.5) and acidified organic solvent (n-hexane). Details about each one of these processes have been explained in following sections.

3.1. Anthocyanin identification by UPLC-TQ-MS

As shown in Table 1, a tentative identification of anthocyanins from acidified water and hexane extracts was carried out by UPLC-TQ-MS analysis. A total of 12 anthocyanins were identified with a maximum visible wavelength of absorbance around 520 nm. Fragmentation patterns of individual peaks showed clearly that glycosylated substitutes at positions C3 and C5 of the flavylium ring were cleaved, which is consistent with previous reports on the fragmentation pattern for triglycosidated anthocyanins.^20,29 All 12 anthocyanins were detected in phosphoric acidified extracts while only 10 were identified in citric acidified extracts. They were all identified as pelargonidin derivative anthocyanins. Otsuki et al. identified 12 acylated anthocyanins in red radish among which pelargonidin was the major anthocyanidin.⁵ With respect to the acyl group, all anthocyanins were characterized as 3-mono- or dihydroxycinnamoyl (p-coumaric, caffeic, and/or ferulic acid)-diglucoside-5-glucoside, 3-mono- or dihydroxycinnamoyl (p-coumaric, caffeic, and/or ferulic acid)-diglucoside-5-malonylglucoside of pelargonidin. The acyl fragmentation patterns are in accordance with previous studies on red radish anthocyanin structures.²⁰ Giusti and Wrolstad reported that red radish major anthocyanidin was pelargonidin, which mono-acylated with p-coumaric acid or ferulic acid up to 30%, while 70% of di-acylation occurred with p-coumaric and malonic acids or ferulic and malonic acids.³⁰ Eight of the twelve anthocyanins detected with m/z 919; 933; 989; 1005; 1019; 1165; 1181 and 1195 in the present study were previously identified by Wu and Prior.³¹

Table 1 Qualitative analyses of anthocyanins presenting in red radish extracts as affected by the extraction medium^a

Compounds	Fragmentation	CA and CAH		PA and PAH
a Pg: pelargonidin, Diglu: diglucoside, Glu: glucoside, Mal: malonic acid.
Pg-3-(caffeoyl)diglu-5-glu	919[M]^+, 757[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+			+	+
Pg-3-(feruloyl)diglu-5-glu	933[M]^+, 771[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+			+	+
Pg-3-(p-coumaroyl)diglu-5-(malonyl)glu	989[M]^+, 741[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+
Pg-3-(caffeoyl)diglu-5-(malonyl)glu	1005[M]^+, 757[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+
Pg-3-(feruloyl)diglu-5-(malonyl)glu	1019[M]^+, 771[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+
Pg-3-(caffeoyl)(p-coumaroyl)diglu-5-glu	1065[M]^+, 903[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+	+	+	+	+
Pg-3-(feruloyl)(p-coumaroyl)diglu-5-glu	1079[M]^+, 917[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+	+	+	+	+
Pg-3-(caffeoyl)(feruloyl)diglu-5-glu	1095[M]^+, 933[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+	+	+	+	+
Pg-3-(feruloyl)(feruloyl)diglu-5-glu	1109[M]^+, 947[M − glu]^+, 433[Pg + glu]^+, 271[Pg]^+	+	+	+	+
Pg-3-(feruloyl)(p-coumaroyl)diglu-5-(malonyl)glu	1165[M]^+, 917[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+
Pg-3-(caffeoyl)(feruloyl)diglu-5-(malonyl)glu	1181[M]^+, 933[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+
Pg-3-(feruloyl)(feruloyl)diglu-5-(malonyl)glu	1195[M]^+, 947[M − glu − mal]^+, 519[Pg + glu + mal]^+, 271[Pg]^+	+	+	+	+

---

3.2. Effect of extraction medium on total monomeric anthocyanin

A quantitative analysis of anthocyanin composition was performed on red radish pigment extracts that were richer in red color as affected by acids at different pH (Fig. 1A). A significant difference was observed between anthocyanin content of red radish extracted with citric and phosphoric acidified water at different pHs. As shown in Fig. 1A, the anthocyanin content decreased with increasing pH (2.5 to 4.5) of acidified water as well as acidified hexane. The anthocyanin content was higher at pH 2.5 in both citric and phosphoric acidified water accounting 185.07 mg/100 g FW and 177.53 mg/100 g FW respectively. The anthocyanin content decreased by 41% and 38% in citric and phosphoric acidified water at pH 4.5 respectively. The higher red radish anthocyanin yield at lower pH in different acidified water might be due to the stabilization of the pigments, which lowered the pH to a level where the absorbance of the anthocyanins was at their maximum. However, acidified organic solvent improved the anthocyanin extraction compared to that of acidified water. The organic solvent was acidified at different pHs (2.5; 3.5; 4.5) in ratio (1 : 1, v/v) prior to the extraction. Phosphoric acidified hexane extracts at pH 2.5 accounted for the highest anthocyanin content (208.37 mg/100 g FW) followed by citric acidified hexane (202.89 mg/100 g FW). These results support the findings of Patil et al. who observed that the use of combined acidified water and alcohol (50 : 50, v/v) resulted in the maximum extraction of radish peels anthocyanins with a better chroma and hue angle.⁹ Jing et al. found the anthocyanin concentrations of three Chinese radish cultivar YZ2, TXH and XLM were 160.74 mg, 144.56 mg and 63.77 m in 100 g fresh radish roots respectively.²⁰ Giusti et al. reported that the highest pigment content was obtained with Spring cvs Fuego and Winter cvs Red Meat radish root (16.5 to 76.9 mg ACN/100 g root).³² Compared to previous findings, the acidified hexane extraction could be a novel effective method for red radish anthocyanin extraction. On the other hand, these differences in anthocyanin content might be attributed to the source of red radish roots and the agricultural area from which the red radish roots were harvested. Giusti et al. reported that the red radish pigment content was dependent on cultivar, root weight and location.³² The losses of monomeric anthocyanins might be most likely due to the formation of anthocyanin polymers.

Fig. 1 Effect of extraction media on (A) anthocyanin yield (mg/100 g FW), (B) percentage polymeric color (%) and (C) color density. CA: acidified with citric acid, PA: acidified with phosphoric acid, CAH: citric acidified hexane, PAH: phosphoric acidified hexane. Values represent the means of three replicate samples with standard error of the means. Values followed by the same letter are not significantly different at P ≤ 0.05.

According to Wrolstand et al., Delgado-Vargas and Paredes anthocyanin pigments are labile compounds that can undergo a number of degradative reactions.^33,34 Their stability is highly variable depending on their structure and the composition of the matrix in which they exist. The change in percentage polymeric anthocyanin in red radish are shown in Fig. 1B. According to the figure, the percentage content of the polymeric anthocyanin significantly increased with increasing pH from 2.5 to 4.5 in both extraction media with percentage yield of variation between 2 and 7% in all extracts. The percentage polymeric anthocyanin was higher in acidified water extracts compared with acidified hexane extracts. However, that of acidified water extracts increased rapidly after pH 3.5 in contrast with the extracts of acidified hexane. CA-4.5 and PA-4.5 showed the highest percentage polymeric anthocyanin, recording 6.52% and 6.42% respectively. The change in percentage polymeric color is an indicator of the anthocyanins polymerization during extraction. These results indicate that the extraction medium has a marked role in anthocyanin extraction and stability. The possible mechanism for anthocyanin polymerization involves condensation reactions of anthocyanins with other phenolic compounds, including flavan-3-ols or polyflavan-3-ols,³⁵ that can be mediated by acetaldehyde³⁶ and furfural³⁷ or occur via direct anthocyanin–tannin reactions.³⁸ The lower percentage of polymeric anthocyanin in acidified hexane extracts in contrast with that of acidified water might be related to the destruction of hydrogen bonds that bound these macromolecules to monomeric anthocyanins. These results are in agreement with the findings of Cao et al. who indicated that the hydrogen bonds could be destroyed by acidified methanol during the extraction process.³⁹ On the other side, the color density (CD) exhibited a significant difference after anthocyanin extraction. CD decreased with increasing pH from 2.5 to 4.5 (Fig. 1C). Acidified hexane extracts showed higher color density compared with acidified water extracts. The change in CD was consistent with the change of total monomeric anthocyanins.

3.3. Effect of extraction medium on instrumental color parameters

The color parameters of red radish pigments extracted in acidified water and acidified hexane media are shown in Fig. 2. The lightness (L) significantly increased with increasing pH from 2.5 to 4.5 regardless of the extraction medium. The lesser lightness of the extracts at pH 2.5 indicated that the extracts were darker and more vivid. These results are in good agreement with those of Jing et al. who reported that the increase in pH values from 1.5 to 4.5 caused greater lightness, indicating that the color gradually became lighter and duller.²⁰ However, the chroma (C) and the hue angle (h) significantly decreased with increasing extraction medium pH. The acidified hexane extracts showed higher chroma and hue angle than those of acidified water extracts. These results indicated that the change in hue angle and chroma of the extracts were dependent on the chemical nature of aqueous phase. These observations of the present study support the findings of Giusti and Wrolstad who reported that when the pH was raised to 3 in 0.1 M citrate buffer, the hue angles decreased from 40° to 20° and the hue of the radish pigment diacylated with cinnamic acids decreased continuously as the pH was increased, towards a more purplish hue.³⁰ In addition, Rodriguez-Saona et al. reported that the increase in anthocyanin content of red radish might be correlated to the increase of the pigment chroma and hue angle and low lightness of the red radish pigment extracts at pH 3.^6,10 The rise in pH values from 1.5 to 4.5 caused greater lightness and smaller chroma, indicating that the color became lighter and duller in a gradual manner. However, color with smaller lightness and greater chroma appeared to be darker and more vivid when pH > 4.5.

Fig. 2 Effect of extraction media on (A) lightness, (B) chroma and (C) hue angle. CA: acidified with citric acid, PA: acidified with phosphoric acid, CAH: citric acidified hexane, PAH: phosphoric acidified hexane. Values represent the means of three replicate samples with standard error of the means. Values followed by the same letter are not significantly different at P ≤ 0.05.

3.4. Effect of extraction medium on total glucosinolates content

Glucosinolates are sulphur-containing compounds, which can possibly be hydrolyzed by myrosinase enzymes to give off-flavors during mastication or processing.⁴¹ Due to the onerous and time-consuming pre-treatment and analysis for the desulfation method by HPLC, the use of a palladium colorimetric method has been developed for analysis of GSLs contents. The GSLs in crude extracts are bound with palladium to form a stable complex, resulting in a color change from light brown to dark brown.⁴⁰ In this method, absorbance at 450 nm is influenced strongly by inhibitory substances in the GSLs crude extracts. The extracts from radish roots are not only colorless but also transparent, which make them present few interfering substances in colorimetric analysis resulting in slight variation of the GSLs composition. Therefore, we applied the simplified method of Møller et al. to evaluate the effect of the extraction solvents on total glucosinolates levels.²³ The total glucosinolate content of red radish before pigment extraction was 140.26 μmol g⁻¹ DW sinigrin equivalent. Ishida et al. reported that the total glucosinolates from Japanese red radish cv. ‘Karami 199’ varied from 94.1–97.6 μmol g⁻¹ DW.²² The total glucosinolates content significantly decreased after anthocyanin extraction (Fig. 3). Acidified hexane extracts at pH 2.5 showed higher total glucosinolate content compared with acidified water extracts. GSLs decreased by 36% and 32% in citric and phosphoric acidified hexane (pH 2.5) respectively. While, it decreased by 94% and 92% in citric and phosphoric acidified water at pH 4.5 respectively. These results confirmed the inhibition of myrosinase enzymes at low extraction pH (2.5). The optimum pH for myrosinase enzyme activity is 6 and its activity decreased by 70% when the pH was decreased to around 3. These results indicated that acidified organic solvent might have highly inhibited the myrosinase enzyme activity compared to that of acidified water. Kiddle et al. reported that GSLs was poorly hydrolyzed by myrosinases in the presence of denaturants such as methanol.⁴⁹ In addition, phosphoric acidified extracts showed higher glucosinolate content than citric acidified extracts. This indicates that acid type might have different effect on inhibition of myrosinase enzyme activity.

Fig. 3 Effect of extraction medium on total glucosinolate content (μmol sinigrin per g (DW)). CA: acidified with citric acid, PA: acidified with phosphoric acid, CAH: citric acidified hexane, PAH: phosphoric acidified hexane. Values represent the means of three replicate samples with standard error of the means. Values followed by the same letter are not significantly different at P ≤ 0.05.

3.5. Effect of extraction medium on off flavor removal

From the above results, the effect of anthocyanin medium of extractions showed significant effect on total glucosinolate content degradation, resulting in off flavor development in extracts. It is known that the red radish anthocyanin off flavors are attributed to the enzymatic breakdown of glucosinolates to generate sulfur and nitrogen containing compounds during processing.²⁹ During this study, the effect of different medium on removal of off flavor from red radish anthocyanin extracts was carried out using GC-MS analysis. As shown in Table 2, fifty-one volatile compounds were detected from red radish anthocyanin extracts including sulphur-containing compounds (4), isothiocyanates (4), nitrogen containing compounds (3), alcohols (9), aldehydes (15), esters (5), ketones (4), carboxylic acids (3), furans (1) and others (3). Isothiocyanates, sulfide, disulfides and nitrogen containing compounds are the major compounds that contribute to the pungency and off-flavor of red radish anthocyanin extracts. They are generated upon the hydrolysis of glucosinolates by myrosinase enzyme during processing. During this study, the extraction of red radish anthocyanins in different medium showed significant effect on inhibition of glucosinolate degradation (Fig. 3), being the reason for the reduction of off flavor development in red radish anthocyanin extracts. According to Iori et al. myrosinase enzymes are susceptible to various conditions; they can be inhibited and usually have a specific preferred environment.⁴² Myrosinase activity was reduced to about 40% at a pH of 3 and below 40% at a pH of 10 while their optimum activity was achieved at pH 6. The extraction medium pH showed a significant effect on reduction of off flavor development during extraction of red radish anthocyanins. As shown in Table 2, the concentration of isocyanates, sulfide, disulfide and nitrogen containing compounds increased significantly with increasing pH from 2.5 to 4.5. These results are in accordance with the findings of Uda et al. who reported that the pH value of the hydrolysis system was important for the hydrolytic reaction and had direct effect on the yield of the isothiocyanates.¹⁶ In addition, Vaughn and Berhow who reported that the isothiocyanates were easily formed at pH equal to 7.0 in the breakdown process of glucosinolates by myrosinase.²⁸

Table 2 Effect of anthocyanin extraction medium on red radish off flavor characteristics (ng g⁻¹ FW)^a

LRI	Compounds	CA-2.5	CAH-2.5	CA-3.5	CAH-3.5	CA-4.5	CAH-4.5	PA-2.5	PAH-2.5	PA-3.5	PAH-3.5	PA-4.5	PAH-4.5
a Reliability of the identification proposal was indicated by the following: ^Amass spectrum and Kovats index according to literature; ^Bmass spectrum compared with NIST98 and Wiley mass spectral databases. KI Kovats index calculated for the DB-Wax capillary column. nd: not detected. Values are represented as means ± SD (n = 2). Values in the same line not sharing a common letter are significantly different at P ≤ 0.05.
Sulfides
1043	Dimethyl disulfide^A,B	1.08 ± 0.01^d	0.63 ± 0.04^c	3.27 ± 0.17^e	0.76 ± 0.05^cd	5.02 ± 0.04^g	3.77 ± 0.31^f	0.08 ± 0.00^ab	ND	ND	ND	0.43 ± 0.00^bc	ND
1354	Dimethyl trisulfide^A,B	0.31 ± 0.06^abc	0.15 ± 0.00^abc	1.70 ± 0.24^d	0.54 ± 0.07^c	2.85 ± 0.12^e	1.73 ± 0.31^d	0.02 ± 0.00^a	0.03 ± 0.00^ab	0.15 ± 0.00^abc	0.10 ± 0.00^abc	0.05 ± 0.00^ab	0.50 ± 0.00^bc
1625	Heptyl methyl sulfide^B	ND	ND	ND	0.08 ± 0.00	ND	ND	ND	ND	ND	ND	ND	ND
1892	Benzothiazole^B	ND	ND	ND	ND	ND	ND	0.15 ± 0.00^a	ND	0.19 ± 0.00^b	ND	0.23 ± 0.01^c	ND
Isothiocyanates
1456	3-Buten-1-yl-isothiocyanate^A,B	0.30 ± 0.00^b	ND	4.41 ± 0.00^d	ND	8.13 ± 0.00^e	ND	0.05 ± 0.02^a	ND	0.37 ± 0.00^c	ND	ND	ND
1814	4-Isothiocyanato-1-(methylthio)-1-butene^A,B	ND	ND	ND	ND	ND	ND	ND	0.02 ± 0.00^a	ND	ND	ND	ND
1868	3-(Methylthio)propyl isothiocyanate^A,B	0.61 ± 0.03^c	0.46 ± 0.00^b	1.10 ± 0.02^d	0.68 ± 0.01^c	5.03 ± 0.04^e	1.05 ± 0.03^d	0.02 ± 0.00^a	ND	ND	ND	ND	ND
1879	4-(Methylthio)-3-butenyl isothiocyanate^A,B	0.43 ± 0.00^d	0.19 ± 0.00^bc	1.44 ± 0.04^g	0.64 ± 0.00^e	5.22 ± 0.07^i	5.87 ± 0.03^j	0.03 ± 0.00^a	0.07 ± 0.00^a	0.29 ± 0.00^c	0.11 ± 0.00^ab	1.85 ± 0.00^h	1.20 ± 0.00^f
1976	2-Phenethyl isothiocyanate^A,B	0.25 ± 0.01^b	0.136 ± 0.00^a	0.41 ± 0.00^c	ND	4.70 ± 0.03^d	ND	0.26 ± 0.00^b	ND	ND	ND	ND	ND
Nitrogen compounds
1200	Pyridine^B	8.25 ± 0.04^f	6.73 ± 0.01^d	7.72 ± 0.02^e	10.94 ± 0.39^g	12.74 ± 0.09^h	13.33 ± 0.38^i	2.42 ± 0.00^a	2.47 ± 0.00^ab	2.78 ± 0.00^ab	3.73 ± 0.00^c	2.94 ± 0.11^b	3.94 ± 0.00^c
1838	2-Mercaptopyridine^A,B	3.16 ± 0.29^e	6.92 ± 0.00^h	6.88 ± 0.00^h	4.10 ± 0.00^f	4.36 ± 0.00^g	4.19 ± 0.00^f	0.14 ± 0.00^c	0.67 ± 0.00^d	0.06 ± 0.00^b	0.02 ± 0.00^a	0.02 ± 0.00^a	0.04 ± 0.00^ab
1864	5-(Methylthio)pentanenitrile^A,B	2.07 ± 0.58^e	ND	ND	0.25 ± 0.00^c	0.26 ± 0.00^d	0.19 ± 0.00^b	0.02 ± 0.00^a	ND	ND	ND	ND	ND
Aldehydes
1058	Hexanal^A,B	1.94 ± 0.05^d	0.74 ± 0.04^b	3.19 ± 0.27^e	2.11 ± 0.10^d	ND	1.29 ± 0.01^c	0.11 ± 0.00^a	ND	ND	ND	ND	ND
1156	Heptanal^B	ND	1.11 ± 0.01^b	ND	ND	ND	ND	0.22 ± 0.00^a	ND	ND	ND	ND	ND
1263	Octanal^A,B	0.25 ± 0.00^d	0.71 ± 0.01^f	0.21 ± 0.00^c	0.08 ± 0.00^b	0.47 ± 0.02^e	0.23 ± 0.02^cd	0.06 ± 0.00^b	ND	ND	ND	ND	0.01 ± 0.00^a
1371	Nonanal^A,B	1.07 ± 0.07^e	0.12 ± 0.00^b	0.89 ± 0.00^d	0.88 ± 0.00^d	0.85 ± 0.02^d	1.02 ± 0.04^e	0.30 ± 0.00^c	0.11 ± 0.00^ab	0.15 ± 0.00^b	0.02 ± 0.00^a	0.19 ± 0.00^b	0.12 ± 0.00^ab
1428	2-Octenal^A,B	ND	ND	ND	ND	ND	ND	0.22 ± 0.00^b	ND	0.17 ± 0.00^a	ND	0.15 ± 0.00	ND
1499	Benzaldehyde^A,B	0.65 ± 0.02^f	0.37 ± 0.01^e	0.28 ± 0.00^d	0.12 ± 0.00^b	0.15 ± 0.01^bc	0.15 ± 0.01^c	0.02 ± 0.00^a	0.01 ± 0.00^a	0.02 ± 0.00^a	0.02 ± 0.00^a	0.01 ± 0.00^a	0.02 ± 0.00^a
1505	Decanal^B	ND	ND	ND	ND	ND	ND	1.26 ± 0.00^b	1.34 ± 0.00^b	1.50 ± 0.00^c	ND	0.67 ± 0.00^a	ND
1514	(E)-2-Nonenal^B	0.11 ± 0.00^c	0.09 ± 0.00^c	0.05 ± 0.00^b	ND	0.02 ± 0.00^a	0.10 ± 0.00^c	0.35 ± 0.00^e	ND	0.36 ± 0.00^e	ND	0.22 ± 0.00^d	ND
1603	(E)-4-Nonenal^B	ND	ND	ND	ND	ND	ND	ND	ND	0.08 ± 0.02	ND	ND	ND
1646	Octadecanal^B	ND	0.28 ± 0.02	ND	ND	ND	ND	ND	ND	ND	ND	ND	ND
1670	(E)-2-Decenal^A,B	ND	ND	ND	ND	ND	ND	0.34 ± 0.00^c	0.31 ± 0.01^bc	0.28 ± 0.00^ab	0.29 ± 0.01^ab	0.27 ± 0.02^a	ND
1682	Dodecanal^B	0.14 ± 0.01^a	ND	ND	0.14 ± 0.00^a	ND	ND	0.28 ± 0.00^c	ND	0.20 ± 0.00^b	ND	0.15 ± 0.00^a	ND
1717	2-Undecenal^B	0.03 ± 0.00^b	ND	0.02 ± 0.00^a	ND	ND	ND	0.31 ± 0.00^f	0.28 ± 0.01^e	ND	0.24 ± 0.00^d	0.20 ± 0.00^c	0.20 ± 0.00^c
1764	2,4-Dimethyl benzaldehyde^A,B	0.84 ± 0.02^h	0.59 ± 0.02^g	0.43 ± 0.00^e	0.41 ± 0.00^e	0.50 ± 0.00^f	0.33 ± 0.02^d	0.13 ± 0.00^b	0.09 ± 0.00^a	0.08 ± 0.00^a	0.19 ± 0.00^c	0.11 ± 0.00^ab	0.08 ± 0.00^a
1858	3-Phenylprop-2-enal^A,B	0.65 ± 0.04^e	0.62 ± 0.02^e	0.31 ± 0.02^b	0.14 ± 0.01^a	0.13 ± 0.01^a	0.11 ± 0.00^a	0.73 ± 0.00^f	0.50 ± 0.00^d	0.40 ± 0.00^c	0.26 ± 0.00^b	0.26 ± 0.01^b	0.25 ± 0.00^b
Alcohols
1183	Eucalyptol^B	1.02 ± 0.00^d	0.98 ± 0.00^d	1.16 ± 0.00^e	0.41 ± 0.00^c	1.30 ± 0.00^g	1.24 ± 0.00^f	0.21 ± 0.00^b	0.16 ± 0.00^a	0.10 ± 0.00^a	0.12 ± 0.00^a	0.13 ± 0.00^a	0.15 ± 0.00^a
1326	1-Hexanol^A,B	ND	ND	0.09 ± 0.00^a	0.25 ± 0.00^b	ND	ND	ND	ND	ND	ND	ND	ND
1525	Linalool^B	0.23 ± 0.02^b	0.33 ± 0.02^c	0.50 ± 0.01^d	0.12 ± 0.00^a	0.13 ± 0.01^a	0.16 ± 0.00^a	1.70 ± 0.00^f	1.77 ± 0.00^g	ND	ND	1.14 ± 0.00^e	ND
1526	1-Octanol^A,B	0.29 ± 0.01^d	0.36 ± 0.03^e	0.23 ± 0.00^c	0.16 ± 0.01^b	0.19 ± 0.00^b	0.34 ± 0.02^e	ND	0.01 ± 0.01^a	0.01 ± 0.01^a	ND	0.01 ± 0.01^a	ND
1635	L-Alpha-terpineol^B	0.06 ± 0.00^a	ND	0.05 ± 0.00^a	0.05 ± 0.00^a	ND	ND	ND	ND	ND	ND	ND	ND
1961	5-Methyl-2-propan-2-yl-phenol^B	3.20 ± 0.00^f	2.42 ± 0.00^e	1.36 ± 0.00^d	0.57 ± 0.00^b	0.58 ± 0.00^c	0.39 ± 0.00^a	ND	ND	ND	ND	ND	ND
2003	2,4-Di-tert-butyl phenol^B	0.48 ± 0.01^d	0.67 ± 0.03^e	0.54 ± 0.02^d	0.34 ± 0.02^c	0.47 ± 0.04^d	0.39 ± 0.01^c	0.11 ± 0.00^a	0.09 ± 0.00^a	0.10 ± 0.00^a	0.19 ± 0.00^b	0.07 ± 0.00^a	0.08 ± 0.00^a
2020	Farnesol^B	ND	ND	ND	ND	ND	ND	ND	0.82 ± 0.00^a	ND	ND	ND	ND
2027	2,4-Di-tert-butylphenol^B	ND	ND	ND	ND	ND	ND	1.28 ± 0.00^d	2.00 ± 0.00^e	0.85 ± 0.00^a	1.01 ± 0.00^b	1.22 ± 0.00^c	1.09 ± 0.00^b
Carboxylic acids
1426	Acetic acid^A,B	ND	0.45 ± 0.03^f	0.15 ± 0.01^c	0.06 ± 0.00^b	0.13 ± 0.00^c	0.30 ± 0.01^e	0.07 ± 0.00^b	0.02 ± 0.00^a	0.24 ± 0.00^d	0.01 ± 0.00^a	0.75 ± 0.00^h	0.54 ± 0.00^g
1572	Hexanoic acid^A,B	0.19 ± 0.01^d	0.26 ± 0.00^e	0.10 ± 0.00^c	0.05 ± 0.00^b	0.06 ± 0.00^b	0.05 ± 0.01^b	0.01 ± 0.00^a	0.47 ± 0.00^i	0.37 ± 0.00^h	0.32 ± 0.01^g	0.28 ± 0.00^f	0.32 ± 0.00^g
1946	Nonanoic acid^A,B	ND	0.08 ± 0.00^a	ND	ND	ND	ND	0.71 ± 0.00^e	0.42 ± 0.00^d	0.71 ± 0.00^e	0.36 ± 0.00^c	0.46 ± 0.00^d	0.26 ± 0.00^b
Furan
1440	Furfural^A,B	ND	0.16 ± 0.00^d	ND	ND	ND	ND	0.03 ± 0.00^c	ND	ND	0.01 ± 0.00^a	0.02 ± 0.00^b	ND
Ketones
1299	Methyl heptenone^A,B	ND	0.30 ± 0.01^c	ND	ND	ND	0.38 ± 0.00^d	0.07 ± 0.00^a	ND	ND	ND	0.13 ± 0.00^b	ND
1820	(E)-Geranyl acetone^A,B	ND	ND	ND	ND	ND	ND	0.07 ± 0.00^a	0.38 ± 0.00^d	0.07 ± 0.00^b	ND	0.14 ± 0.00^c	ND
1834	3-Nonen-2-one^A,B	2.36 ± 0.00^b	1.11 ± 0.00^a	3.27 ± 0.00^c	ND	ND	ND	ND	ND	ND	ND	ND	ND
1913	1-Phenylhexan-1-one^A,B	2.34 ± 0.00^b	1.16 ± 0.00^a	3.51 ± 0.00^c	1.360 ± 0.00^a	ND	ND	ND	ND	ND	ND	ND	ND
Esters
874	Ethyl acetate^B	22.26 ± 0.05^d	9.49 ± 0.07^c	22.34 ± 0.83^d	31.00 ± 0.71^e	34.67 ± 0.19^f	35.47 ± 0.00^f	5.57 ± 0.00^b	6.12 ± 0.00^b	5.83 ± 0.06^b	2.14 ± 0.03^a	6.08 ± 0.02^b	6.28 ± 0.01^b
1346	Hexyl formate^B	ND	0.18 ± 0.00^b	ND	ND	0.10 ± 0.01^a	ND	ND	ND	ND	ND	ND	ND
2016	Ethyl hexadecanoate^B	ND	ND	ND	ND	ND	ND	0.44 ± 0.00^e	0.62 ± 0.01^f	0.28 ± 0.00^d	0.22 ± 0.00^b	0.24 ± 0.00^c	0.17 ± 0.00^a
2110	Diisobutyl phthalate^A,B	2.31 ± 0.00^e	1.51 ± 0.01^d	2.59 ± 0.00^f	0.07 ± 0.00^b	0.12 ± 0.00^c	9.35 ± 0.01^g	ND	ND	ND	0.02 ± 0.00^a	ND	ND
2136	Benzyl butyl phthalate^A,B	ND	ND	ND	ND	ND	ND	ND	ND	0.257 ± 0.02^b	ND	0.211 ± 0.00^a	ND
Others
1341	1-Methoxy-4-[(E)-prop-1-enyl]benzene^B	0.14 ± 0.01^b	0.15 ± 0.01^b	0.08 ± 0.00^a	ND	ND	ND	ND	ND	ND	ND	ND	ND
1549	Camphene^B	1.58 ± 0.79^d	ND	0.80 ± 0.01^c	ND	0.84 ± 0.01^c	0.66 ± 0.00^b	0.08 ± 0.00^a	ND	ND	ND	ND	ND
1750	Naphthalene^A,B	ND	9.00 ± 0.02^g	6.15 ± 0.07^f	4.50 ± 0.02^d	4.22 ± 0.07^c	5.45 ± 0.11^e	ND	ND	ND	ND	0.45 ± 0.00^b	0.03 ± 0.00^a

---

Four isothiocyanates were identified including 3-buten-1-yl isothiocyanate, 3-(methylthio)propyl isothiocyanate, 4-(methylthio)-3-butenyl isothiocyanate, and 2-phenethyl isothiocyanate. Their respective concentrations drastically increased with increasing pH from 2.5 to 4.5 in acidified water but the increase became gradual in the case of acidified hexane. These results indicate that acidified hexane medium significantly inhibits the degradation of glucosinolates in contrast with that of acidified water. On the other hand, phosphoric acidified extracts showed lower isothiocyanate concentration compared to citric acidified extracts. These results are in accordance with the findings of glucosinolates content (Fig. 3). Among isothiocyanates identified, 4-(methylthio)-3-butenyl isothiocyanate was previously identified as the product of myrosinase rapid hydrolysis of 4-methylthio-3-butenylglucosinolate (glucoraphenin) in red radish and it is known to be responsible for red radish pungent principle flavor.⁴³ 3-(Methylthio)propyl isothiocyanate (radish isothiocyanate) was also identified among methylthio-substituted compounds and was considered as products of enzymatic hydrolysis of glucoside progenitors in various cruciferae.⁴³ It was also identified in fresh Japanese and Kenyan radish.⁴⁴ According to Friis and Kjaer radish isothiocyanate and 4-isothiocyanato-1-(methylthio)-1-butene might have a common biogenetic pathway for the corresponding parent thioglucosides.⁴³ 2-Phenethyl isothiocyanate has strong aroma of watercress and tingling sensation. It was derived from 2-phenylethyl glucosinolate and was first isolated in wasabi powder. It was also identified in Kenyan radish.⁴⁴

Sulfides were other sulphur containing compounds identified during this study. Dimethyl disulphide and dimethyl trisulphide are derived from (+)-S-methyl-L-cysteine sulphoxide, an amino acid found in Brassica.²⁷ They can also be derived from subsequent degradation of some volatiles derived from glucosinolates via oxidation of methanethiol.²⁷ Dimethyl disulphide and dimethyl trisulphide are unpleasant oxidized sulfurous off-flavor compounds with low detection threshold in water and air. Their highest concentration might be responsible for the sulfurous and cabbage-onion like odor characteristics of red radish anthocyanin extracts.

The extraction of red radish anthocyanin in acidified hexane (acidified organic solvent) significantly reduced the off flavor of the extracts compared to acidified water extraction. The acidification of medium with phosphoric acid was more effective to reduce off flavor in red radish anthocyanin extracts than the medium acidified with citric acid. Furthermore, phosphoric acidified hexane medium at pH 2.5 reduced isothiocyanate and sulfides compounds by 76% and 90%, respectively, while, citric acidified hexane at pH 2.5 reduced isothiocyanate and sulfite compounds by 51% and 44%, in that order. This might be due to different pK_a of two acids (2.16 for phosphoric acid and 3.13 for citric acid). The lower concentration of isothiocyanate and sulfide compounds in acidified hexane (organic solvent) might be attributed to the thermolability of isothiocyanate and sulfite compounds, which might be easily vaporize in environment during anthocyanin extraction and hexane removal under vacuum. These results are in good agreement with the findings of Van Eylen et al. who reported isothiocyanates as thermolabile compounds and pressure stable.^45,46

Apart from isothiocyanate and sulfur containing compounds, the nitrogen containing compounds were also identified including 5-(methylthio)pentanenitrile, 2-mercaptopyridine, and pyridine. 5-(methylthio)pentanenitrile might be derived from the same 4-(methylthio)-3-butenyl glucosinolate.²⁷ 2-Mercaptopyridine, and pyridine were first detected in red radish anthocyanin extracts. Contrary to isothiocyanate and sulfide compounds, their concentration were higher in acidified hexane extracts than that of acidified water. At pH 2.5, nitrogen compounds increased by 13% and 18% in citric acidified hexane and phosphoric acidified hexane respectively. However, the concentration of nitrogen compounds was significantly lower in phosphoric acidified extracts compared to citric acidified extracts. Nitrogen compounds might also be generated from the breakdown of glucosinolates during processing.

These extracts, except for the above mentioned compounds, contained compounds without nitrogen and sulphur, mostly aldehydes, fatty acids, esters and alcohols; the major components being nonanal, benzaldehydes, (E)-2-nonenal, 2,4-dimethyl benzaldehyde, 3-phenylprop-2-enal, ethyl acetate, diisobutyl phthalate in citric acidified extracts, ethyl hexadecanoate in phosphoric acidified extracts, acetic and hexanoic acids, eucalyptol, 1-octanol, and 2,4-di-tert-butyl phenol were abundantly detected in most extracts. Alcohols and aldehydes are common volatiles in processed vegetables, and are formed by oxidative breakdown of free fatty acids by the action of lipoxygenase.^47,48 1-Hexanol and hexanal might impart the leaf-green odor to the red radish anthocyanin extracts.

To confirm the removal of off flavor responsible compounds, further study was carried out on GC-MS analysis with water anthocyanin extracts, acidified water extracts (pH 2.5) and acidified hexane extracts (pH 2.5) (Table 3). The responsible off flavor compounds (sulphur and isothiocyanates) were identified and quantified by comparing with their respective internal standards. Samples were first dehydrated under vacuum at low pressure and then rehydrated with respective medium. Results showed that the major off flavor compounds were with the aqueous or solvent phases during vacuum dehydration. A larger number of off flavor compounds was removed with acidified hexane for which the water phase recorded 15% of off flavor compounds while 64% was detected in hexane phase. Following dehydration, 9% off flavor compounds were detected in the vaporized water in the case of samples extracted with water and that of acidified water extract yielded 25% in the vaporized phase. The higher concentration of off flavor compounds (sulfur and isothiocyanates) in hexane phase indicate that these compounds are more lipid soluble and could be easily removed by extracting with organic phase (hexane). These results confirmed the above finding, which showed over 70% off flavor compounds could be removed during red radish extraction with acidified hexane.

Table 3 Verification of red radish off flavor removal by GC-MS^a

KI	Compounds	Water extract pigment	Water extract vaporized phase	Acidified water extract	Acidified water extract vaporized phase	Acidified hexane	Vaporized hexane phase	Aqueous phase from acidified hexane
a Reliability of the identification proposal was indicated by the following: ^Amass spectrum and Kovats index according to literature; ^Bmass spectrum compared with NIST98 and Wiley mass spectral databases. ^CSpectra and LRI values of authentic compounds obtained commercially. KI Kovats index calculated for the DB-Wax capillary column. ND: not detected. Values are represented as means ± SD (n = 2). Values in the same line not sharing a common letter are significantly different at P ≤ 0.05.
Sulfides
1040	Dimethyl disulfide^A,B,C	0.88 ± 0.01^d	1.02 ± 0.06^e	0.38 ± 0.03^b	0.49 ± 0.03^c	ND^a	ND^a	ND^a
1338	Dimethyl trisulfide^A,B,C	0.06 ± 0.05^c	0.12 ± 0.04^d	0.44 ± 0.04^f	0.23 ± 0.05^e	0.03 ± 0.01^b	0.03 ± 0.00^b	0.01 ± 0.01^a
Isothiocyanates
1456	3-Buten-1-yl-isothiocyanate^A,B,C	0.02 ± 0.00^b	0.06 ± 0.01^c	0.06 ± 0.00^d	0.09 ± 0.01^f	ND^a	0.07 ± 0.02^e	0.12 ± 0.01^g
1839	3-(Methylthio)propyl isothiocyanate^A,B,C	0.29 ± 0.04^b	0.50 ± 0.04^d	0.47 ± 0.02^c	0.65 ± 0.03^e	ND^a	ND^a	ND^a
1945	4-Methylthio-3-butenyl isothiocyanate^A,B,C	0.61 ± 0.01^f	0.36 ± 0.02^c	0.38 ± 0.00^d	0.84 ± 0.04^g	0.23 ± 0.01^a	0.27 ± 0.03^b	0.52 ± 0.02^e
1952	2-Phenethyl isothiocyanate^A,B,C	0.12 ± 0.02^b	0.15 ± 0.01^e	0.11 ± 0.01^a	0.18 ± 0.01^f	0.12 ± 0.00^c	0.15 ± 0.00^d	0.51 ± 0.05^g

---

3.6. Sensory evaluation

Results of the sensory evaluation showed that the extraction medium has a significant effect on the sensory attributes of red radish anthocyanin extracts (Table 4). The color and off flavor intensity sensory attributes showed a significant (P ≤ 0.05) difference in acidified water and hexane extracts at different pHs. Compared with commercially available synthetic red color anthocyanin, the panelist assessment confirmed that the color intensity of both acidified water and hexane extracts decreased with increasing extraction media pH. Panelists observed that color intensity of acidified water extracts at pH 2.5 was not significantly different to the color intensity of the commercial products, however, they have higher off flavor intensity than commercial product. On the other side, panelists revealed no significant differences between commercial anthocyanin extracts with acidified hexane extracts (pH 2.5). Both acidified water and hexane extracts at pH 4.5 showed lowest color intensity and highest off flavor intensity compared to the commercial product. Their low anthocyanin content might be the reason for the poor color intensity while their higher concentration in sulphur and isocyanates compounds might explain the higher off flavor intensity. These results correlated well with the data of both anthocyanin yield and GC-MS.

Table 4 Effect of extraction medium on red radish anthocyanin sensory characteristics^a

Samples	Color intensity	Off-flavor	Samples	Color intensity	Off-flavor
a Mean scores (listed in ascending order) for each attribute within a column with different letters are significantly different (P ≤ 0.05) using one-way ANOVA comparison test (n = 36; 12 panellists with 3 replications).
Control	10.00 ± 0.00^d	5.00 ± 0.00^a	Control	10.00 ± 0.00^d	5.00 ± 0.00^a
CA-2.5	8.67 ± 1.08^c	7.58 ± 0.67^b	CAH-2.5	9.00 ± 0.60^c	5.08 ± 0.79^a
CA-3.5	6.83 ± 0.72^b	8.50 ± 1.00^bcd	CAH-3.5	6.50 ± 0.52^b	6.33 ± 0.78^b
CA-4.5	4.67 ± 1.30^a	8.83 ± 0.72^cd	CAH-4.5	4.50 ± 0.52^a	7.17 ± 1.11^bc
PA-2.5	8.83 ± 0.67^c	8.00 ± 0.85^bc	PAH-2.5	9.17 ± 0.72^cd	5.33 ± 0.98^a
PA-3.5	7.00 ± 0.78^b	8.67 ± 1.15^cd	PAH-3.5	6.67 ± 0.78^b	7.17 ± 0.72^bc
PA-4.5	5.00 ± 0.60^a	9.17 ± 0.72^d	PAH-4.5	4.83 ± 1.11^a	7.83 ± 0.72^c

---

4. Conclusions

The study successfully established an alternate efficient technique for extraction and removal of anthocyanin and off flavor, respectively from red radish root. Results showed that acidified hexane extraction (pH 2.5) yielded higher anthocyanin compared to that of acidified water. Phosphoric acidified hexane extracts (pH 2.5) showed lower percentage polymeric color, higher color density with higher chroma and hue angle explaining good stability during extraction. Twelve anthocyanins tentatively identified from UPLC-MS were found in phosphoric acidified extracts. In addition, results showed that the off flavor intensity was significantly decreased in phosphoric acidified hexane extracts (pH 2.5) followed by citric acidified hexane (pH 2.5) in comparison with other extracts. The results of sensory evaluation confirmed the good color and lower off flavor intensity of phosphoric acidified hexane (pH 2.5) anthocyanins comparable to the commercial red anthocyanins. These results indicated that the acidification of extraction medium with phosphoric acid was more effective to conserve red radish anthocyanins compared with citric acid. These anthocyanins could be considered as safe natural colorings in food industry due to the easy vaporization coupled with the non-toxic nature of hexane.

Acknowledgements

The research was supported in part by the National Program of China (2013AA102204) and Postdoctoral Science Foundation of China (2016M590143) and program of “Collaborative innovation center of food safety and quality control in Jiangsu Province”. It was also founded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

1. M. M. Giusti and R. E. Wrolstad, J. Food Sci., 1996, 61, 688–694.
2. M. E. Camire, A. Chaovanalikit, M. P. Dougherty and J. Briggs, J. Food Sci., 2002, 67, 438–441.
3. D.-L. Aurelio, R. G. Edgardo and S. Navarro-Galindo, Int. J. Food Sci. Technol., 2008, 43, 322–325.
4. V. T. Šaponjac, A. Gironés-Vilaplana, S. Djilas, P. Mena, G. Ćetković, D. A. Moreno, J. Čanadanović-Brunet, J. Vulić, S. Stajčić and M. Vinčić, RSC Adv., 2015, 5, 5397–5405.
5. T. Otsuki, H. Matsufuji, M. Takeda, M. Toyoda and Y. Goda, Phytochemistry, 2002, 60, 79–87.
6. L. E. Rodriguez-Saona, M. M. Giusti and R. E. Wrolstad, J. Food Sci., 1999, 64, 451–456.
7. L. E. Force, T. J. O'Hare, L. S. Wong and D. E. Irving, Postharvest Biol. Technol., 2007, 44, 175–178.
8. I. Sarvan, R. Verkerk, M. van Boekel and M. Dekker, Innovative Food Sci. Emerging Technol., 2014, 25, 58–66.
9. G. Patil, M. C. Madhusudhan, B. Ravindra Babu and K. S. M. S. Raghavarao, Chem. Eng. Process., 2009, 48, 364–369.
10. L. E. Rodriguez-Saona, M. M. Giusti, R. W. Durst and R. E. Wrolstad, J. Food Process. Preserv., 2001, 25, 165–182.
11. T. Gachovska, D. Cassada, J. Subbiah, M. Hanna, H. Thippareddi and D. Snow, J. Food Sci., 2010, 75, E323–E329.
12. B. Gilewicz-Łukasik, S. Koter and J. Kurzawa, Sep. Purif. Technol., 2007, 57, 418–424.
13. R. P. Metivier, F. J. Francis and F. M. Clydesdale, J. Food Sci., 1980, 45, 1099–1100.
14. R. Baharfar, R. Azimi and M. Mohseni, J. Food Sci. Technol., 2015, 52, 6777–6783.
15. H. Liang, Q. Yuan and Q. Xiao, J. Mol. Catal. B: Enzym., 2006, 43, 19–22.
16. Y. Uda, T. Kurata and N. Arakawa, Agric. Biol. Chem., 1986, 50, 2735–2740.
17. R. Gao, P. Jing, S. Ruan, Y. Zhang, S. Zhao, Z. Cai and B. Qian, Food Chem., 2014, 146, 423–428.
18. S. D. Kottman, Master of Science, Ohio State University, 2011.
19. X. Sui, X. Dong and W. Zhou, Food Chem., 2014, 163, 163–170.
20. P. Jing, S.-J. Zhao, S.-Y. Ruan, Z.-H. Xie, Y. Dong and L. Yu, Food Chem., 2012, 133, 1569–1576.
21. J. Lee, R. W. Durst and R. E. Wrolstad, J. AOAC Int., 2005, 88, 1269–1278.
22. M. Ishida, T. Kakizaki, T. Ohara and Y. Morimitsu, Breed. Sci., 2011, 61, 208–211.
23. P. Møller, A. Plöger, H. Sørensen and H. Sørenson, Advances in the production and utilization of cruciferous crops, 1985, pp. 97–110.
24. M. Ishida, M. Nagata, T. Ohara, T. Kakizaki, K. Hatakeyama and T. Nishio, Breed. Sci., 2012, 62, 63–70.
25. Y. Hu, H. Liang, Q. Yuan and Y. Hong, Nat. Prod. Res., 2010, 24, 1195–1205.
26. K. Eric, L. V. Raymond, S. Abbas, S. Song, Y. Zhang, K. Masamba and X. Zhang, Food Res. Int., 2014, 57, 242–258.
27. I. Blažević and J. Mastelić, Food Chem., 2009, 113, 96–102.
28. S. F. Vaughn and M. A. Berhow, Ind. Crops Prod., 2005, 21, 193–202.
29. M. M. Giusti, L. E. Rodríguez-Saona, D. Griffin and R. E. Wrolstad, J. Agric. Food Chem., 1999, 47, 4657–4664.
30. M. M. Giusti and R. E. Wrolstad, Biochem. Eng. J., 2003, 14, 217–225.
31. X. Wu and R. L. Prior, J. Agric. Food Chem., 2005, 53, 3101–3113.
32. M. M. Giusti, L. E. Rodriguez-Saona, J. R. Baggett, G. L. Reed, R. W. Durst and R. E. Wrolstad, J. Food Sci., 1998, 63, 219–224.
33. R. E. Wrolstad, R. W. Durst and J. Lee, Trends Food Sci. Technol., 2005, 16, 423–428.
34. F. Delgado-Vargas and O. Paredes-López, Natural colorants for food and nutraceutical uses, CRC Press, 2002.
35. J. D. Reed, C. G. Krueger and M. M. Vestling, Phytochemistry, 2005, 66, 2248–2263.
36. N.-E. Es-Safi, H. Fulcrand, V. Cheynier and M. Moutounet, J. Agric. Food Chem., 1999, 47, 2096–2102.
37. N.-E. Es-Safi, V. Cheynier and M. Moutounet, J. Agric. Food Chem., 2000, 48, 5946–5954.
38. S. Remy, H. Fulcrand, B. Labarbe, V. Cheynier and M. Moutounet, J. Sci. Food Agric., 2000, 80, 745–751.
39. X. Cao, Y. Zhang, F. Zhang, Y. Wang, J. Yi and X. Liao, J. Sci. Food Agric., 2011, 91, 877–885.
40. W. Thies, Fette, Seifen, Anstrichm., 1982, 84, 338–342.
41. A. M. Bones and J. T. Rossiter, Phytochemistry, 2006, 67, 1053–1067.
42. R. Iori, P. Rollin, H. Streicher, J. Thiem and S. Palmieri, FEBS Lett., 1996, 385, 87–90.
43. P. Friis and A. Kjaer, Acta Chem. Scand., 1966, 20, 698–705.
44. A. Kjær, J. Øgaard Madsen, Y. Maeda, Y. Ozawa and Y. Uda, Agric. Biol. Chem., 1978, 42, 1715–1721.
45. D. Van Eylen, N. Bellostas, B. W. Strobel, I. Oey, M. Hendrickx, A. Van Loey, H. Sørensen and J. C. Sørensen, Food Chem., 2009, 112, 646–653.
46. D. Van Eylen, M. Hendrickx and A. Van Loey, Food Chem., 2006, 97, 263–271.
47. P. Schreier, Chromatographic studies of biogenesis of plant volatiles, Huthig, 1984.
48. H. Tian, X. Yang, C.-T. Ho, Q. Huang and S. Song, Food Chem., 2013, 141, 131–138.
49. G. Kiddle, R. N. Bennett, N. P. Botting, N. E. Davidson, A. A. Robertson and R. M. Wallsgrove, Phytochem. Anal., 2001, 12, 226–242.

---