In the 1st stage of the study, an optimal ratio of 1-PeCSO and 2-PeCSO for blue–green pigment formation was investigated to develop a model reaction system that consisted of 1-PeCSO, 2-PeCSO, alliinase, and glycine. Different ratios of 1-PeCSO, 2-PeCSO, alliinase, and glycine were mixed and evaluated using a spectrophotometer. Pink to blue pigments developed in the model solutions (Table 2). Pigment was formed in solutions containing only 1-PeCSO, but no absorbance maximum was detected in a 100% 2-PeCSO solution. Blue pigments developed only in combinations containing both 1-PeCSO and 2-PeCSO, and the absorbance at 580 nm increased with increasing 1-PeCSO until reaching a maximum value. The highest value obtained (0.53) was for a 30% 1-PeCSO and 70% 2-PeCSO solution. However, there was no absorbance at 580 nm in solutions of 60% to 100% 1-PeCSO and resulting colors were violet to pink with no significant difference. Kubec and others (2004) reported 1-propenyl derivatives form pink, pink-red, and magenta compounds, whereas those containing allyl groups give rise to blue products after reacting with glycine. Imai and others (2006b) confirmed that a vivid-blue color could be produced using a highly defined model reaction system comprised only of isolated 1-PeCSO, 2-PeCSO, pure glycine, and purified garlic alliinase. Our results are consistent with these investigations.
On the basis of absorbance at 580 nm, we conducted the next investigation with a 30% 1-PeCSO and 70% 2-PeCSO solution. All 22 amino acids were reacted with the thiosulfinate solution which comprised a 30% 1-PeCSO, 70% 2-PeCSO solution and alliinase. HPLC spectra were compared with naturally greened extracts from garlic-onion juice. Because blue–green pigments are easily generated by mixing juice from heated white onions (Imai and others 2006b), a good source of 1-PeCSO, we added white onions to the garlic juice. Peaks in HPLC spectra were detected at 440 nm and 580 nm in the visible range with different retention times: 2.7 to 5.6 min and 16.6 to 28.3 min, respectively (Figure 2). These spectra indicated that the green color was a mixture of yellow pigment (440 nm) and blue pigment (580 nm). It was previously assumed that at least 2 species of pigment existed in “Laba” garlic pickling solution using spectrophotometer with 2 maxima absorptions at 440 and 590 nm: one species with a yellow color and a second with a blue color. The combination of both species would therefore result in the observed green color of the pickling solution (Bai and others 2005; Wang and others 2008). In this study, we separated yellow and blue pigments using HPLC and diode array detector and we confirmed this discoloration was blueing, not greening. Figure 2 showed 8 blue peaks detected at 580 nm in natural garlic-onion juice. Their retention times were 16.6 (1), 19.7 (2), 20.8 (3), 21.7 (4), 23.5 (5), 25.1 (6), 26.9 (7), and 28.3 (8). Each peak was matched with an amino acid-thiosulfinate reaction solution. As shown in Figure 3, different kinds of blue pigments developed when different amino acids were reacted with thiosulfinate solution. In contrast to Kubec and others (2004), who reported that pink pigment is developed with only 1-PeCSO and alliinase, pink pigment did not develop in the absence of amino acid. Most of the amino acids formed blue pigments by reaction with thiosulfinate except Cys, Pro, Met, and h-Pro. However, peaks at 580 nm were detected only when mixed with Gly, Arg, Lys, Ser, Ala, Arp, His, Asn, Gln, and Tyr (Figure 4) and the greatest color-generating ability was exhibited by Gly, followed by Lys and Asn (data not shown). Compared with the natural greened extract, Arg matched with (1) and (2), Lys matched with (3), (4), and (7), Asn matched with (5), (7), and (8) and Gln matched with (5), (6), and (7), Gly and Ser matched with (7), and Asp matched with (8) (Table 3). Although glycine showed the greatest color-generating ability and most of previous researcher concluded glycine was the major amino acid involved in discoloration reaction (Shannon and others 1967a,b; Kubec and others 2004; Imai and others 2006b; Kubec and Velíšek 2007), the spectrum of glycine did not match well with the spectrum of greened garlic juice. Major peaks of glycine were detected at 24.9, 27.2, and 30.8 min, matching only 7, which is a minor peak in the natural greened extract. On the other hand, Arg, Asn, and Gln had spectra that were the most similar to the spectra of natural green garlic-onion juice. According to reports on the amino acid composition of garlic, Arg, Gln, Asn, and Glu are typically the most abundant free amino acids (Montaño and others 2004; Lee and others 2005) and these amino acids also have the ability to form blue pigments. Moreover, Gly was one of the minor amino acids and some varieties were not even detected glycine (Lee and Harnly 2005), although Gly showed the strongest blue color when reacted with thiosulfinate solution. Based on these findings, candidate amino acids for forming blue pigments in green garlic juice were Arg, Asn, and Gln rather than Gly.
In the last step of this study, we compared spectra and chromatograms at 580 nm of blue pigments formed by the 3 major candidate amino acids: Arg, Asn, and Gln, with natural green juice to confirm retention time and spectra of each peak. Figure 5 shows that for addition of the candidate amino acids to both natural green juice and thiosulfinate solution each peak had the same retention time as one of the peaks in the natural green juice and no extra peaks developed due to treatment of the amino acid.